UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Clinical and molecular analysis of Williams-Beuren syndrome : a search for factors determining clinical… Wang, Michael Steve 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1999-0391.pdf [ 6.26MB ]
JSON: 831-1.0089058.json
JSON-LD: 831-1.0089058-ld.json
RDF/XML (Pretty): 831-1.0089058-rdf.xml
RDF/JSON: 831-1.0089058-rdf.json
Turtle: 831-1.0089058-turtle.txt
N-Triples: 831-1.0089058-rdf-ntriples.txt
Original Record: 831-1.0089058-source.json
Full Text

Full Text

CLINICAL AND MOLECULAR ANALYSIS OF WILLIAMS-BEUREN SYNDROME: A SEARCH FOR FACTORS DETERMINING CLINICAL VARIABILITY IN A CONTIGUOUS DELETION SYNDROME by MICHAEL STEVE WANG B.Sc, The University of British Columbia, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Medical Genetics Graduate Programme, Department of Medical Genetics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA May 1999 © Michael Steve Wang, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Wy[Cf<L g^U^ TlCS The University of British Columbia Vancouver, Canada Date dlUS£ 2 4 ^ -DE-6 (2/88) Abstract Williams-Beuren syndrome (WBS) is clinically characterized by a distinctive "elfin" facial appearance, short stature, mental retardation, a unique cognitive profile, friendly outgoing personality, and various congenital heart malformations with supravalvular aortic stenosis (SVAS) being the most common. Clinical features seen with less frequency include infantile hypercalcaemia, hoarse voice, hyperacousis, renal hypoplasia, clinodactyly of the 5th finger, and teeth hypoplasia. The syndrome arises as a consequence of a 2 cM deletion of 7qll.23 in 90 - 95% of all clinically typical cases reported to date. This region includes a number of genes that include elastin (ELN) and LEVI kinase 1 (LTMK1). ELN mutations have been found in isolated cases of SVAS while LIMK1 has been implicated in the unique WBS cognitive profile. However, not all patients present with congenital cardiovascular malformations and the severity to which cognition is impaired in WBS individuals varies with each patient. This variability of clinical phenotype presentation in clinically typical cases has not been explained. In order to identify the underlying cause of this variability, clinical and molecular data were analyzed from 108 clinically diagnosed WBS cases. A deletion of 7qll.23 was identified in 85 cases and extended from D7S489U to D7S1870 (including these markers) in the majority of cases. Only two cases were found to have a more distal deletion that also encompassed D7S489L. In these 85 cases, no influence on phenotypic variability was found for deletion extent, parental origin of deletion, ELN and LLMK1 polymorphisms, or gender using a Bonferroni corrected level of significance (oc'M^xlO"4). Nonetheless, a number of interesting 'suggestive' associations with p<0.05 were identified. ii Correlations between clinical features of WBS were examined to identify those features which may have a common etiology. Identifying correlates may also identify specific phenotypic features that can be used as a prognostic tool. This has revealed significant correlations between: 1) low birth weight ^ICP centile) and low weight at the time of physical examination (<10 centile at the time of examination; p=9.0xl0" ), and 2) hypercalcemia and presentation of a stellate iris pattern (p=1.0xl0"5). While not fulfilling the strict criterion (a'=3.3xl0"4) necessary to reach an experiment-wide false positive rate of less than 5%, a number of interesting suggestive correlations were observed including: 1) stellate and blue irides (p=1.2xl0"3), 2) stellate irides and a typical WBS personality (p=6.8xl0"3), and 3) renal/urogenital abnormalities and cardiovascular malformations (p=7.9xl0"3). These identified associations provide an important starting point for the understanding of the underlying etiology of WBS and will be of interest to confirm in an independent data set. iii Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures ix List of Abbreviations x Acknowledgements xii Chapter 1 - Introduction 1 1.1 Opening remarks 1 1.2 Deletions and duplications give rise to an abnormal phenotype 1 1.3 Microdeletion syndromes 3 1.3.1 Deletion extent 5 1.3.2 Hemizygosity: the effects of the inherited copy 6 1.3.3 Imprinting 8 1.3.4 Genetic Background 9 1.3.5 Mosaicism 10 1.3.6 Environmental effects 10 1.3.7'Non-deletion'cases 11 1.4 Williams-Beuren syndrome: a microdeletion syndrome 12 1.4.1 Clinical description 12 1.4. La Distinctive characteristics of WBS 12 1.4.1 .b Other characteristics of WBS 15 1.4.2 Historical look at cytogenetic and molecular analyses of WBS 16 iv 1.5 Thesis obj ective 18 Chapter 2 - Materials and Methods 19 2.1 Patients 19 2.2 Molecular methods 19 2.2.1 Blood lysates and D N A extraction 20 2.2.2 Identifying polymorphic markers within the deletion region 20 2.2.3 P C R and multiplex P C R amplification 23 2.2.4 Restriction digests 23 2.2.5 Polyacrylamide gel electrophoresis and allele sizing 23 2.2.6 Screening for a deletion by microsatellite analysis 26 2.2.7 Genetic mapping 28 2.3 Statistical analysis 30 2.3.1 Contingency table 30 2.3.2 Chi- square (x2) analysis 31 2.3.3 Fisher's exact test 32 2.3.4 Significance levels 33 2.3.5 Estimating power and necessary sample size 34 2.3.6 Cluster analysis 36 2.3.7 Odds ratio and its confidence interval 41 Chapter 3 - Results 43 3.1 Genetic Mapping 43 3.2 Molecular-clinical correlations 46 3.2.1 Molecular screening 46 3.2.2 7ql 1.23 deletion cases 46 3.2.37qll.23 non-deletion cases 5 8 3.3 Clinical correlations 63 3.4 Clinical difference between deletion and non-deletion cases 66 Chapter 4 - Discussion 68 4.1 Genetic mapping 68 4.2 Statistical power 68 4.3 Determining the cause for phenotypic variability 70 4.4 Clinical correlations 79 4.4.1 Significant correlation between low birth weight and poor postnatal weight gain: evidence for deletion of a growth factor gene? 80 4.4.2 Correlations with a stellate iris pattern 81 4.4.3 Other associations: renal/urogenital and cardiovascular abnormalities 83 4.5 Deletion versus non-deletion WBS cases: discussing heterogeneity and phenocopies 86 4.6 Conclusions 88 Chapter 5 - References 93 Appendix I Dataform 116 Appendix II Binary data for cluster analysis 118 Appendix III Estimated necessary sample sizes to detect a true association with 120 study-wide oc=0.05 and [3=0.10 vi List of Tables From Chapter 2 - Materials and Methods 2.1 Sample 2x2 contingency table 30 2.2 Most extreme table of second tail 33 2.3 2x2 table giving the number of 1-1, 1-0, 0-1, and 0-0 matches 37 2.4 Sample resemblance matrix with a total of 5 patients using the Jaccard coefficient of similarity. 38 2.5 Cophenetic matrix converted from sample cluster tree 40 2.6 Unstrung list of resemblance and cophenetic matrix values 40 2.7 Sample 2x2 table used to arrange data in order to calculate odds ratio 42 From Chapter 3 - Results 3.1 CEPH recombinant mapping panel 45 3.2 Pair-wise comparisons between parental origin of deletion and clinical outcome 50 3.3 Expected and observed ELNil allele frequencies with expected values based on those published by Urban et al. (1997) 51 3.4 Correlation of ELN and LLMK1 genotypes with clinical outcome of WBS cases with a proven 7ql 1.23 deletion 52 3.5 Expected and observed ELNil8 allele frequencies with expected values based on those published by Foster et al. (1993) 53 3.6 Expected and observed Helg 15/16 allele frequencies with expected values based on those published by Mari et al. (1995) 54 3.7 Expected and observed LLMK1-GT allele frequencies with expected values based on those published by Mari et al. (1998) 54 3.8 (A) Linkage disequilibrium betweenELNilS allele 7 and Helg 15/16 allele 2 (B) Correlation found between absence of "7-2" haplotype and presence of typical WBS personality 56 3.9 Effect of gender on the phenotypic outcome in WBS cases with a proven 7ql 1.23 deletion 57 3.10 E L N and LIMK1 polymorphism allele groups and "genotypes" used in correlating with clinical data of non-deletion WBS cases 61 3.11 Effect of gender on the phenotypic outcome in WBS cases without a 7ql 1.23 deletion 62 3.12 Analysis for associations between WBS clinical features in 85 7ql 1.23 deletion cases 64 3.13 Analysis for associations between WBS clinical features in 23 non-deletion cases 65 3.14 Summary of pair-wise comparisons of WBS phenotypic features between 85 7ql 1.23 deletion cases and 23 non-deletion cases 67 From Chapter 4 - Discussion 4.1 Sample calculations of statistical power and required sample size to detect true associations for a given a 69 4.2 Renal/urogenital and cardiovascular abnormalities observed in 12 WBS patients with positive correlation of traits 85 viii List of Figures From Chapter 1 - Introduction 1.1 Formation of a microdeletion and microduplication by unequal meiotic exchange 4 1.2 Individual with Williams-Beuren syndrome 16 From Chapter 2 - Materials and Methods 2.1 Map of Williams-Beuren syndrome deletion region including genes known to be deleted 22 2.2 Results for E L N and L L M K l genotypes resolved on polyacrylamide gels 25 2.3 Microsatellite analysis for deletion 27 2.4 Genetic mapping 29 2.5 Sample cluster analysis output tree 39 From Chapter 3 - Results 3.1 Molecular screening in confirmed 7ql 1.23 deletion cases 48 3.2 Summary of 7ql 1.23 and 22ql 1 screening for WBS v72 59 List of Abbreviations AS: Angelman syndrome ASD: atrial septum defect BCCH: BC Children's Hospital CGS: contiguous gene syndrome cM: centimorgan CMT1A: Charcot-Marie-Tooth disease type IA COL2 A l . collagen type 2 gene DGS: DiGeorge syndrome ELN: elastin gene ELNil: elastinintron 1 microsatellite ELNil8: elastin intron 18 microsatellite FBN1: fibrillin-1 gene FISH: fluorescence in situ hybridization GDB: Genome Database Helg 15/16: elastin exon 20-intron 20restriction fragment length polymorphism HLA: human leukocyte antigen HNPP: hereditary neuropathy with liability to pressure palsies IC: imprinting control ILS: isolated lissencephaly sequence LTMK1: LEVI kinase 1 gene LIMKl-GT: LEVI kinase 1 intron 13 microsatellite LIS1. lissencephaly-1 gene Mb: megabase MDS: Miller-Dieker syndrome MRI: magnetic resonance imaging OCA2: oculocutaneous albinism type 2 OFC: head circumference PCR: polymerase chain reaction PMP22: peripheral myelin protein-22 gene PPS: peripheral pulmonary stenosis PWS: Prader-Willi syndrome RFLP: restriction fragment length polymorphism SAS: segmental aneusomy syndrome SMS: Smith-Magenis syndrome SRS: Silver-Russell syndrome SVAS: supravalvular aortic stenosis SVPS: supravalvular pulmonary stenosis UPD: uniparental disomy UPGMA: unweighted pair-group method using arithmetic averages UTR: untranslated region VCFS: Velo-cardio-facial syndrome VNTR: variable number tandem repeats VSD: ventricular septum defect WBS: Williams-Beuren syndrome Acknowledgements I would like to thank first and foremost, Jesus Christ, for being my savior and blessing me with the opportunity to work with so many wonderful people at UBC. I sincerely thank my supervisor, Dr. Wendy P. Robinson, for all that she has done in guiding me through this project. I also thank my committee members Dr. Carolyn J. Brown and Dr. Jan M. Friedman for their time and suggestions. And finally, I thank my mum, dad, Mary, and friends for their support and believing in me throughout my studies. xii CHAPTER 1: Introduction 1.1 Opening remarks Subtle clinical differences often exist in the phenotypic presentation of a genetic disorder with a similar underlying cause, such as a cytogenetic rearrangement. It may seem obvious that variable presentation of phenotypic characteristics is likely caused by the differences in rearrangement breakpoints where the breakpoints can differ by megabases (Mb) such as cytogenetically detectable deletions and duplications. However, it is not entirely clear what the cause is in seemingly identical cytogenetic rearrangements such as microdeletions and microduplications. Possible causes of phenotypic variability are considered with respect to microdeletions specifically and are tested in this study using Williams-Beuren syndrome (WBS) as a model microdeletion syndrome. 1.2 Deletions and duplications give rise to an abnormal phenotype An abnormal clinical phenotype could arise from duplications and deletions of chromosomal segments, resulting in segmental trisomy and monosomy for the region duplicated or deleted. A duplication or deletion of genetic material can result in an abnormal phenotypic outcome if gene(s) within the duplicated or deleted region are dosage-sensitive. For the present thesis, I mean by 'dosage-sensitivity' that the phenotype will be influenced by whether there are 1, 2, or 3 copies present. In the case of a deletion, the term haploinsufficiency is used to indicate one copy of the dosage-sensitive gene is insufficient to give you the normal phenotype typically found when two copies of the gene are present and active. Genes that are not dosage-sensitive could also contribute to l an abnormal phenotype as the effects of a recessive mutation could be revealed through a deletion. This would have the same effect as an individual without a deletion being homozygous for the recessive mutation since only a mutated product is present that has altered function and/or structure. Therefore, the underlying cause of an abnormal phenotype in this situation is the presence of a mutant product. These two possible effects of a deletion and how they result in clinical abnormality will be discussed further in the following Section 1.3.2, "Hemizygosity: the effects of the inherited copy". A pattern of phenotypic features may sometimes be identified if a particular region is prone to recurrent deletions or duplications. Such characteristic phenotypic patterns identified for these and other chromosomal rearrangements as well as single gene alterations are termed "syndromes". Deletion and duplication syndromes are defined based on how they were first identified. Those that are identified first cytogenetically or concomitant with clinical recognition are termed "classical deletion and duplication syndromes" (Schmickel, 1986; Greenberg, 1993). These syndromes, such as Cri-du-chat syndrome and 18p deletion syndrome, are generally the result of large rearrangements easily identifiable using conventional cytogenetic techniques. Those syndromes that are recognized as a clinical entity prior to identification of their associated cytogenetic location, are generally the results of a cryptic deletion or duplication that require molecular techniques to detect (Schmickel, 1986; Greenberg, 1993). Such syndromes owing to cryptic deletions and duplications are known by a number of different names that include microdeletion/duplication, contiguous gene, and segmental aneusomy syndromes. 1.3 Microdeletion syndromes Disorders that were clinically recognized first that arise from the deletion or duplication of adjacent genes were originally named 'contiguous gene syndromes' (CGSs; Schmickel, 1986). It was initially suggested that clinical manifestation of CGSs would be dictated by the extent of the chromosome involved and clinical features of the syndrome could be inherited as an independent feature (Schmickel, 1986). However, it has since been suggested that this is not an entirely appropriate term as not all genes within the duplicated or deleted region, contribute to the phenotype. The alternative term, 'segmental aneusomy syndrome' (SAS), was therefore introduced to imply that the phenotype results from dosage imbalance of critical genes within the deleted or duplicated segment (Budarf and Emanuel, 1997). Yet another proposed alternative, the terms 'microdeletion' and 'microduplication' syndromes, was suggested to emphasize the pathogenesis of the disorder through a deletion or duplication event (Schinzel, 1988). For this reason, microdeletion and microduplication will be used to describe syndromes arising from cytogenetically undetectable deletions and duplications. The specific factors affecting the formation of microdeletions and microduplications are not known. While unique microdeletions and microduplications have been reported throughout the genome, the mechanism of formation is likely to be distinct from recurring microdeletions and microduplications, which occur with greater frequency in specific regions of the genome. It is not clear whether this tendency for recurrence is the result of preferential survival of duplications and deletions of regions without dosage lethal genes or the result of structural features of the genome that make it susceptible to rearrangement or both. One such structural feature common to recurring 3 microdeletions and microduplications, is the presence of flanking repetitive sequences. These flanking repeats may facilitate unequal exchange between homologous chromosomes resulting in a duplication on one homologue and a corresponding deletion on the other (Figure 1.1). A well documented example of such deletion/duplication formation through unequal meiotic exchange, comes from a 1.5 M b microdeletion and microduplication at 17pl 1.2. This 17pl 1.2 microduplication results in Charcot-Marie-Tooth disease 1A ( C M T 1 A ) , a peripheral demyelinating neuropathy resulting from a dosage effect of the peripheral myelin protein gene, PMP22, which is included within the duplication. A microdeletion of that same region results in hereditary neuropathy with liability to pressure palsies (HNPP), another peripheral neuropathy (Lupski et al., 1991; Patel et al., 1992; Pentao et al., 1992). Figure 1.1: Formation of a microdeletion and microduplication by unequal meiotic exchange. Duplication and deletion by unequal meiotic exchange. Unequal exchange Normal Duplication Deletion Normal 4 With multiple genes present within the deleted/duplicated region, with each potentially having a function in multiple tissues, one might expect a complex clinical profile in microdeletion and microduplication syndromes. The clinical phenotype however, is not uniform, even within families (Morris et al., 1993; Sadler et al, 1993; Devriendt et al., 1997). This is especially evident in cases of monozygotic twins both carrying identical deletions but discordant for clinical features (Pankau et al., 1993; Castorina et al., 1997). A number of suggestions have been made that might explain the cause for clinical variability. The following sections will introduce these suggestions and will focus on their possible influence in microdeletions specifically. 1.3.1 Deletion extent Phenotypic consequences of microdeletion syndromes are thought to be the result of the additive effects of hemizygosity of dosage-sensitive genes. Thus, it is possible that there could be a phenotypic spectrum for each disorder that is directly related to the extent of the deleted region. Fluorescence in situ hybridization (FISH) and molecular analyses offer superior resolution over conventional cytogenetics and have been used to delineate the deletion extent in several studies of various microdeletion syndromes. This has revealed some evidence in support of deletion extent accounting for deletion variability. For instance, a deletion of 17pl3.3 has been found to be associated with Miller-Dieker syndrome (MDS). Probes isolated from the MDS critical region have been used to identify smaller deletions of 17pl3.3 in a number of patients with isolated lissencephaly sequence (ILS), a feature also present in MDS (Ledbetter et al., 1992; Chong et al., 1997). It has since been determined that the smaller deletion contains the 5 gene, LIS1, responsible for ILS but cannot account for many other features of MDS (Lo Nigro et al., 1997). Similar to 17pl3.3 deletions, a deletion of 7qll.23 manifests as Williams-Beuren syndrome (WBS) but a mutation or intragenic deletion of ELN, which resides within the WBS deletion, results in isolated cases of supravalvular aortic stenosis (SVAS), also a feature of WBS (Ewart et al., 1993a; Ewart et al., 1993b). However, there is also some evidence to suggest deletion size has little or no bearing on phenotypic variability in certain microdeletion syndromes. Perhaps the most apparent example, is the 22q deletion syndrome. Microdeletions of 22qll result in a large number of overlapping phenotypic abnormalities that include DiGeorge syndrome (DGS), velo-cardio-facial syndrome (VCFS), conotruncal facial abnormality, and sporadic cardiac defects. Studies delineating deletion size at 22qll reveal high variability of the deletion extent but no evidence of a correlation between deletion size and phenotype severity (Morrow et al., 1995; Carlson et al., 1997). 1.3.2 Hemizygosity: the effects of the inherited copy There are two ways the intact copy of a gene can influence the phenotype in individuals carrying a deletion. First of all, many developmental pathways are controlled by a rate limiting step or involved in protein complexes requiring exact stoichiometry (Fisher and Scambler, 1994). Therefore, variability in the expression level of the intact copies of genes within a deletion region may explain variable clinical outcome. For example, although each patient has roughly 50% normal gene expression, this may show variability between individuals. It may be that 60% normal gene expression is enough to prevent disease status whereas 50% might not be. Decreased allelic expression has been 6 found to be associated with disease phenotypes. For instance, a type 2 collagen gene (COL2A1) sequence polymorphism was identified with one allele being associated with expression at <12% normal levels in 11% of patients presenting with osteoarthritis (Loughlin et al., 1995). Another study found decreased expression of fibrillin (FBN1) in 3 patients with Marfan syndrome. This decreased expression was associated with a 3' untranslated region (UTR) restriction fragment length polymorphism (Hewett et al., 1994). Decreased expression and susceptibility to type I diabetes has also been found to be associated with polymorphic variable number tandem repeats (VNTR). Such is the case observed with insulin expression levels being associated with 5' minisatellite alleles (Kennedy et al., 1995). As mentioned previously, hemizygosity by deletion could also reveal the effects of one or more recessive alleles on the retained homologue, resulting in a range of phenotypic effects. One would expect to find these recessive disorders to be more frequent in a population of individuals with a deletion of the gene than the normal population as homozygosity would not be required. Therefore, one might expect to find presentation of this recessive abnormality in a population carrying a deletion, to be equal to half the frequency of heterozygous carriers for rare recessive disorders since there is also a 50% chance of the abnormal allele being deleted. This is the case for type 2 oculocutaneous albinism (OCA2), a congenital hypopigmentary disease, which results from a recessive mutation of the P gene, located in 15qll-ql3 (Lee et al., 1994a). Albinism affects 1/36,000 Caucasians (p=0.9947, q=0.0053) but is found in -1% of patients with Prader-Willi (PWS) or Angelman syndrome (AS), conditions that can arise from the deletion 15qll-ql3 of paternal and maternal origin, respectively (Rinchik et al., 7 1993; Lee et al., 1994b). Studies of the P gene in a number of PWS individuals with albinism have demonstrated mutations within the P gene (Rinchik et al., 1993; Lee et al., 1994b). The increase in frequency of albinism found in PWS and AS is consistent with half the expected carrier frequency in the normal population. 1.3.3 Imprinting Although parent of origin effects had been noted as long as 3000 years ago in Asia Minor by mule breeders, it was not proven that a normal chromosome complement was required from both maternal and paternal origin until pronuclear transplantation experiments were conducted in mice (Savory, 1970; McGrath and Solter, 1984; Surani et al., 1984). Fusion of two female pronuclei produced gynogenetic zygotes. These showed normal development of the embryo proper but poor development of the placenta and membranes whereas development of androgenetic zygotes (fusion of two male pronuclei) was abnormal in the embryo but normal in the extraembryonic tissue. These observations provided conclusive evidence for the necessity of biparental genetic contribution to the offspring in mammals in order for normal development. It has since been postulated that the requirement for biparental inheritance is due to a phenomenon known as 'imprinting' (Moore and Haig, 1991). Genetic or genomic imprinting is defined as the parent-specific expression of a subset of genes. Paternal inheritance only of a required paternally imprinted gene or vice versa, would therefore result in an abnormal phenotype. Evidence for imprinting being associated with microdeletion syndromes was first noted in PWS where a de novo microdeletion of 15qll-ql3 is detected in 60 - 70% of the patients. These deletions were 8 noted to be of paternal origin exclusively (Butler and Palmer, 1983; Nicholls et al., 1989a). Subsequent analysis of non-deleted PWS cases revealed that inheritance of two maternal copies of chromosome 15 only, uniparental disomy [UPD(15)], could also give rise to the disease phenotype (Nicholls et al., 1989b). Alternatively, a maternal deletion of 15qll-ql3 or paternal UPD(15) gives rise to a clinically distinct abnormality, Angelman syndrome (AS; Knoll et al., 1989; Malcolm et al., 1991). Based on the fact that the phenotypic outcome is dependent upon parental origin of deletion at 15qll-ql3, it is possible that the parental origin could also affect phenotypic variability in other microdeletion syndromes. 1.3.4 Genetic background Genetic background must certainly play a role in phenotypic variability in any genetic disorder including microdeletion syndromes. For instance, Waardenburg syndrome (WS), a condition resulting in hearing, craniofacial, limb and pigmentation disorders, can arise by a deletion or mutation of PAX3 (Pasteris et al., 1992; Pasteris et al., 1993; Tassabehji et al., 1994). Studies of the Pax3 mouse model, Splotch, have revealed that severity and penetrance of the phenotype depend on the genetic background, including the sex (Asher et al., 1996). In humans, there are a number of examples of disorders requiring two unlinked loci to interact or have mutations. For instance, families with retinitis pigmentosa, a hereditary retinal dystrophy, have been identified with mutations in two unlinked photoreceptor-specific genes, ROM1 and peripherin/RDS (Kajiwara et al., 1994). Similarly, presentation of a specific clinical feature in a microdeletion syndrome could require an additional mutation at a different 9 locus. However, identifying such additional modifiers is a difficult task as large kindreds are required but microdeletion syndromes are generally sporadic. 1.3.5 Mosaicism Mosaicism can potentially lead to phenotypic diversity. For instance, mosaics for Down syndrome (46/47,+21), Klinefelter syndrome (46,XY/47,XXY), and Turner syndrome (45,X/46,XX) are commonly discovered when clinically atypical patients of these conditions are investigated for cytogenetic anomalies. Their phenotype generally lies somewhere between normal and that typical of the full constitutional abnormality (Strachan and Read, 1996). It has been suggested that severity of phenotype may be dependent upon when mosaicism arose in development, as earlier origin of an abnormal line could result in multiple organ systems being affected. Survival itself may be dependent upon presence of a normal cell line being present early in development in certain chromosomal trisomies (Robinson et al, 1995; Robinson and Kalousek, 1996; Cantu et al., 1996). Mosaicism could conceivably contribute to phenotypic diversity in microdeletion syndromes as well, but has not been proven convincingly in published studies thus far (Robinson et al., 1998). 1.3.6 Environmental effects Clearly, development is affected by environmental factors, including the intrauterine growth environment. These can give rise to monozygotic twins, discordant for phenotypic features in cases of microdeletion syndromes (Pankau et al., 1993; Castorina et al., 1997; Yamagishi et al., 1998) and even in the normal population. 10 However, the effects of any of a large number of teratogens identified to date perhaps best illustrate environmental effects. These include factors such as exposure to infectious agents such as rubella, insufficiency of necessary nutrient intake such as folic acid, and maternal intake of drugs such as alcohol during pregnancy. The timing of such environmental insults is also critical to phenotypic outcome, as there may be only a narrow window of time during which development can be affected. Folic acid, for example, is required during the early stages of development, usually before the pregnancy is even recognized, for normal development of the neural tube (reviewed by Butterworth and Bendich, 1996). However, it is still unknown whether any environmental factors play a role in manifestation of specific phenotypic features of microdeletion syndromes. 1.3.7 'Non-deletion' cases All cases of clinically typical PWS have some kind of detectable abnormality of chromosome 15, although only 75% of these cases have a detectable microdeletion (Nicholls et al, 1998). This provided hope that similar defects may be identified in the non-deletion cases associated with other microdeletion syndromes, such as WBS (~90% of clinically typical cases with a 7qll.23 deletion; Nickerson et al, 1995), VCFS (>90% deleted at 22qll; Carlson et al, 1997), and MDS (also >90% deleted at 17pl3.3; Kuwano et al, 1991). One study of Smith-Magenis syndrome (SMS) has reported 100% of cases studied carried a deletion at 17pll.2, but the authors noted that the syndrome is likely under-diagnosed due to its relatively recent description and is only given a diagnosis of SMS if a deletion had been confirmed (Juyal et al, 1996; Elsea et al, 1997). The question, therefore, lies in whether the remaining 'non-deletion' cases actually do 11 carry a smaller microdeletion that was not detected or whether the patients do not carry a deletion. Those that do not could be cases of genetic heterogeneity, where a similar phenotype is conveyed by a different genetic mechanism(s). Alternatively, non-deletion cases could also be phenocopies, which are phenotypically similar but arise due to some environmental interaction with the genotype. The fact that the majority of cases with a microdeletion syndrome do actually carry a deletion would indicate that phenocopies are rare. However rare, these clinically diagnosed but non-deletion cases may potentially have clinical differences with patients with a confirmed deletion as well. 1.4 Williams-Beuren syndrome: a microdeletion syndrome 1.4.1 Clinical description Two independent researchers first characterized the Williams-Beuren syndrome. Williams (1961) described a syndrome of supravalvular aortic stenosis (SVAS), mental retardation, and a distinctive facial appearance, and Beuren (1962) noted the same features but also added dental abnormalities and peripheral pulmonary stenosis (PPS). Over the years, additional clinical features have been recognized to be manifestations of this multisystemic syndrome. 1.4.1.a Distinctive characteristics of WBS Often characterized as elf-like, WBS facies feature a broad forehead, periorbital fullness, a flat nasal bridge, anteverted nares, epicanthic folds, full cheeks and lips, a long smooth philtrum, and a wide mouth (Figure 1.2). These features become coarser with age but the wide and full lips remain a characteristic facial feature (Preuss, 1984; Morris et 12 al., 1990). There are also ocular findings associated with WBS that include esotropic strabismus, hyperopia, and blue irides with a stellate or lacy appearance (Greenberg and Lewis, 1988). At birth, patients with WBS are generally small for gestational age in terms of both length and weight. A number of complications that include feeding difficulty, failure to thrive, and colic, arise during early infancy in the majority of cases. Also reported are chronic otitis media, vomiting, chronic constipation, and hypercalcaemia (Burn, 1986; Morris et al., 1988), but the most severe complication noted in infancy is the presence of a congenital heart defect. This is found in the majority of WBS cases. Supravalvular aortic stenosis (SVAS) is the most common heart defect found in WBS patients, resulting in narrowing of the ascending aorta. Also inherited as an isolated autosomal dominant trait, SVAS can be corrected by surgery, but failure to do so could potentially lead to heart failure and death. Histopathology of stenotic regions has revealed loss of elastic fiber parallel orientation, hypertrophy of smooth muscle cells, and proliferation of smooth muscle in fibroblasts (Perou, 1961; O'Connor, 1985). Stenosis can also affect a number of other blood vessels including the supravalvular pulmonary artery (SVPS) and peripheral pulmonary artery (PPS). Although the pathology of stenosis is not completely understood, it is believed that vascular obstruction may increase hemodynamic damage to the endothelium of inelastic vessels and result in proliferation of intimal smooth muscle cells and fibroblasts, resulting in further narrowing of the blood vessel (Keating, 1997). Such constriction of blood vessels may lead to hypertension, a feature common to many WBS patients, especially in those approaching adulthood (Morris et al., 1990; Lopez-Rangel et al, 1992). Other structural cardiac abnormalities 13 found in WBS patients include atrial and ventricular septum defects (ASD and VSD, respectively) and mitral valve prolapse (Keating, 1997). Another distinctive characteristic of WBS is the neurological, linguistic, and cognitive profde of the syndrome. The majority of WBS cases are classified as mildly to moderately mentally retarded, with a mean IQ of 57, but some have borderline normal intelligence or severe mental retardation (Udwin et al., 1987; Morris et al., 1988). Magnetic resonance imaging (MR1) of WBS patients reveals uneven reduction in volume of the cerebrum (80%) and cerebellum (99%) but no characteristic lesions (Bellugi et al., 1990; Jernigan and Bellugi, 1990; Brinkmann et al., 1997). Emotionally and behaviorally, WBS children exhibit certain abnormal features with higher frequency when compared with other mentally handicapped and normal control children. This manifests as extreme anxiety and fear when placed under stressful conditions as well as hypersensitivity to loud noises. However, WBS children have also been described as hyperactive, "unusually friendly", and "pleasant and affectionate" with even strangers (Preuss, 1984). Cognitive and linguistic profiles of WBS children are most striking as they present with an uneven pattern of reduced functioning, resulting in a marked weakness in visuospatial constructive cognition and strength in language and auditory skills (Morris et al., 1988; Udwin and Yule, 1990; Karmiloff-Smith et al., 1995). Reduction in visuospatial cognition is easily demonstrated by their inability to recreate an object when given the required components, even when the object is as simple as a checkerboard consisting of four cubes (Frangiskakis et al., 1996). Their speech, on the other hand, has been described as being reminiscent of "cocktail party speech" or loquaciousness. This is characterized by incessant chatter at a superficial level and overuse of stereotyped phrases and cliches (Udwin et al., 1987). 1.4.1.b Other characteristics of WBS A number of other organ systems are also affected in WBS patients, including musculoskeletal, dental, and renal abnormalities, among others. Joint hyperelasticity is seen in children with WBS, but progressive joint tightening is characteristic, resulting in mild kyphosis, lordosis, or scoliosis in some cases. An awkward gait is also sometimes observed during early childhood due to tightening of the heel cords and hamstrings. Additionally, muscle tone is generally poor in WBS children (Burn, 1986; Morris et al., 1988). Dental abnormalities are noted to affect both primary and permanent dentition. This includes enamel hypoplasia, microdontia, malocclusion, and excessive interdental spacing. A high frequency of caries is also noted in WBS patients, but no single dental finding has been found to be pathognomonic of WBS (Burn, 1986; Morris et al., 1988; Hertzberg et al., 1994). A broad range of kidney and urinary tract abnormalities are found in WBS patients. These include renal agenesis, duplicated kidney, renal structural abnormality, bladder diverticulum, vesicoureter reflux, and recurrent urinary tract infections, as well as renal artery stenosis (Morris et al., 1988; Pankau et al., 1996). Again, there is no single renal or urinary tract abnormality that is distinctive of WBS. 15 Additionally, patients with WBS often have a hoarse deep voice, premature graying of the hair as adults, and hyperacousis. The cause of these findings is not known (Preuss, 1984; Burn, 1986). Figure 1.2: Individual with Williams-Beuren syndrome. From Ounap K, Laidre P, Bartsch O, Rein R, and Lipping-Sitska M (1998). Familial Williams-Beuren syndrome. Am. J. Med. Genet. 80: 491 - 493. 1.4.2 Historical look at cytogenetic and molecular analyses of WBS Early studies of WBS patients reported normal karyotypes, but researchers hypothesized that a common genetic etiology of WBS and isolated SVAS could exist as SVAS is a common feature of WBS. Linkage was found to exist between markers in chromosome 7q and SVAS in two unrelated families, and a third family was identified where a translocation t(6:7)(p21.1;ql 1.23) was found to co-segregate with SVAS (Ewart 56 et al., 1993a; Curran et al., 1993). This translocation was shown to disrupt ELN, suggesting that abnormalities of ELN resulted in the vascular disease. The researchers hypothesized that WBS could be a deletion syndrome at 7qll.23 that encompassed multiple genes including ELN. In a subsequent study, both sporadic cases and a familial case of WBS were analyzed by 1) quantitative Southern analysis using cosmid probes cELN-272 and cELN-HD which flank ELN; 2) polymerase chain reaction (PCR) amplification of a polymorphic Bfal endonuclease restriction site within ELN; and 3) fluorescence in situ hybridization (FISH) using cosmid cELN-272. These indicated presence of only one ELN in all cases analyzed, implicating ELN hemizygosity in the pathogenesis of WBS (Ewart et al., 1993b). Subsequent studies have indicated a 7qll.23 deletion of consistent size in -90% of clinically typical WBS cases (Kotzot et al., 1995; Nickerson et al., 1995; Perez Jurado et al., 1996; Joyce et al., 1996; Brendum-Nielsen et al., 1997; Wu et al., 1998). No clear evidence has been found for any one factor besides ELN having a clear influence on phenotypic outcome, despite conflicting reports of parental origin of deletion having an effect (Perez Jurado et al., 1996; Wu et al., 1998). In addition, a recent study has also implicated 22q in WBS. One suspected case of WBS was reported in a study of 52 patients who carried a ring chromosome 22 (Joyce et al., 1996). 22qll has also been implicated in cases of both isolated and syndromic pulmonary stenosis and VSD, heart conditions also found in some WBS cases (Momma et al., 1997; Recto et al., 1997). Such an overlap in clinical features may account for some atypical WBS cases harboring a proximal 22q rearrangement, specifically a microdeletion. 17 1.5 Thesis Objective Subtle phenotypic differences arise in WBS cases despite a seemingly similar deletion size. The purpose of this thesis was, therefore, to analyze clinically diagnosed WBS cases for deletion extent, parental origin-of-deletion, inherited hemizygous genotype and haplotype, and gender for their influence on phenotypic outcome. In the event patients did not carry a deletion, it had to be determined whether differences in clinical outcome existed between deletion and non-deletion patients and whether there were any cases with a 22ql 1 deletion in the 7ql 1.23 non-deletion WBS patients. Thirdly, clinical traits were compared in a pair-wise fashion to determine whether any correlations existed as these may indicate a common etiology for co-occurring traits. Clinical correlations may also provide a clinical indicator for diagnosis of a more severe complication arising from WBS. 18 CHAPTER 2: Materials and Methods 2.1 Patients A total of 108 patients with a clinical diagnosis of WBS were included in this study. Of these, 27 were referred to the Provincial Medical Genetics Department at BC Children's Hospital (BCCH), 16 were seen at the Pediatrics Department at the University of Saskatchewan, and 7 were diagnosed at the Pediatrics Department at the University of Manitoba. An additional 7 were analyzed at the Institute for Child Health, Athens, Greece, and 51 at the University of Zurich, Zurich, Switzerland. At least partial clinical data were recorded on patients seen in person by collaborating physicians (outside Vancouver) or collected retrospectively from patient files at BCCH. Clinical data on some of the patients ascertained in Switzerland have been previously published (Kotzot et al., 1995a; Dutly and Schinzel, 1996; Baumer et al., 1998). Clinical data collected was based on features contained in a clinical diagnosis dataform developed by Dr. A. Schinzel (Appendix n) and analyzed using Microsoft EXCEL™ 97. 2.2 Molecular methods 2.2.1 Blood lysates and DNA extraction Peripheral blood samples were available from all BC and Saskatchewan patients except BC patients v86 and vl28. Blood was also available from some of their parents for the study. DNA samples from 38 affected individuals and their parents seen in Zurich, Switzerland, were made available for molecular study by Dr. A. Schinzel. Neither blood nor DNA was made available from patients seen in Greece or Manitoba. 19 DNA was extracted from 7cc unfrozen peripheral blood by salt-extraction (Laitinen et al., 1994). This was accomplished by transferring the peripheral blood sample to a 50 mL Falcon tube along with 30 mL of ice-cold lysis buffer (155 mM NH4CI; 10 mM K H C O 3 ; 0.1 mM EDTA; pH 7.4) and incubating at 4°C for 30 minutes or until the blood was no longer cloudy. The solution was then centrifuged at 1500 rpm in a Jouan CR412 bench-top centrifuge at 4°C for 15 minutes. The resulting supernatant was decanted and the remaining leukocyte pellet was washed by resuspending in 15 mL ice-cold lysis buffer and centrifuging a second time under the same conditions. After decanting again, the cell pellet was suspended in 5 mL of SE digest buffer (75 mM NaCl; 25 mM EDTA; pH 8.0) along with 50 uL proteinase K (10 mg/mL) and 500 uL 10% SDS and was incubated at 37°C overnight. Following incubation, an additional 5 mL of SE buffer was added followed by a 5 minute incubation at 55°C. This was immediately followed by addition of 3 mL of 6 M NaCl and vortexing for 30 seconds. The mixture was then centrifuged at 3000 rpm at room temperature for 15 minutes and the clear supernatant was poured into a fresh 50 mL Falcon tube. 2 volumes of 95% ethanol were added to precipitate the DNA. Precipitated DNA was washed twice in ice-cold 70% ethanol and allowed to air dry. The extracted DNA was then resuspended in 500 uL of TE buffer (10 mM Tris/HCl pH 8.0; 1 mM EDTA) for storage and diluted to 30 ng/uL for microsatellite analysis by PCR. 2.2.2 Identifying polymorphic markers within the deletion region Polymorphic markers within and flanking the WBS common deletion region were identified utilizing published maps. This revealed centromeric flanking microsatellites 20 D7S645, D7S672, and D7S653 as well as telomeric flanking microsatellites D7S675 and D7S669 (Tsui et al., 1995). Nine microsatellite polymorphisms were identified within the deletion region: D7S489U and D7S489L (Robinson et al., 1996), also named D7S489B and D7S489A respectively (Perez Jurado et al., 1996); D7S613, D7S2472, and D7S2476 (Perez Jurado et al., 1996); D7S1870 (Gilbert-Dussardier et al., 1995); ELN intron 1 (CCTT)n (ELNil; Urban et al., 1997); ELN intron 18 (CA)n (ELNil8), originally placed in intron 17 due to ambiguity in numbering of exons (Foster et al., 1993; Tassabehji et al., 1997a); and (GT)n located in intron 13 of LIMK1 (LIMK1-GT; Mari et al., 1998). In addition, an Mval restriction fragment length polymorphism (RFLP) was identified in the exon 20-intron 20 junction of ELN (Helgl5/16; Tromp et al., 1991). Markers were ordered according to published maps (Figure 2.1). These 7qll.23 markers were screened to determine whether they were deleted in our sample of 108 WBS patients. How this screening is accomplished is discussed in further detail in section 2.2.6. In cases that were not deleted at 7qll.23, I also screened for a deletion at proximal 22qll as the region has been implicated in WBS in a previous study (Joyce et al., 1996). To do so, the following published VCFS/DGS region microsatellite markers were tested for a deletion in order from most centromeric to most telomeric: D22S427 - D22S1638 - D22S941 - D22S944 - D22S1623 - D22S264 -D22S311 - D22S306 (Morrow et al., 1995; Carlson et al., 1997). Both 7qll.23 and 22qll marker information is also available online from the Genome Database (GDB) at http://www.gdb.org. 21 Figure 2.1: Map of Williams-Beuren syndrome deletion region including genes known to be deleted. Markers in bold represent polymorphic microsatellites except Helg 15/16, which is an RFLP. Italicized markers represent duplicated sequences flanking the common WBS deletion region. 1.8 c M 5.1 c K i D7S645 D7S672 2 c M D7S653 IcM PMS2L/IB291/D7S1778 D7S489M/D7S1490 PMS2L/IB291/D7S1778 D7S489U/D7S1490 FZD3 D7S2476 STX1A ELN LIMK1 D7S613 RFC2 D7S2472 D7S1870 D7S489LID7S1490 . PMS2LP1/1B291/D7S1778 D7S675 D7S669 ELNil ELNil8 Helg 15/16 LIMK1-GT Williams-Beuren syndrome deletion region (2cM) Variably deleted 1 c M NB: Map is not drawn to scale 22 2.2.3 PCR and multiplex PCR amplification Polymorphic markers were amplified under conditions described in the primary research papers cited previously (Section 2.2.2) using an MJ Research PTC-60 thermocycler with a heated lid. Two pairs of microsatellite polymorphisms were PCR amplified together as multiplex reactions: D7S613/D7S2476 and LIMK1 -GT/ELNi 18. PCR products were resolved on polyacrylamide electrophoretic gels after amplification except Helg 15/16 products, which had to be digested with a restriction enzyme. All PCR primers are available from Research Genetics except Helg 15/16, which was synthesized elsewhere. 2.2.4 Restriction digests Helg 15/16 PCR amplified products were precipitated at -20°C for five hours with 2.5 volume 95% ethanol and 0.1 volume 3M sodium acetate (C2H302Na), washed with ice-cold 70% ethanol, and air dried. The dried PCR products were then resuspended in 10 pL Mval restriction digest cocktail consisting of 1 unit Mval and 1 uL lOx Buffer 'H' from Boehringer Mannheim and then incubated at 37°C for 6 hours. Digested PCR products were resolved on polyacrylamide electrophoretic gels. 2.2.5 Polyacrylamide gel electrophoresis and allele sizing Equal volumes of PCR amplified/restriction enzyme digested products and urea loading buffer were loaded onto denaturing polyacrylamide gels for size separation. Amplified products larger than 200 bp (ELNil) were resolved on 5% polyacrylamide while the remaining smaller products were resolved on 6% polyacrylamide gels. 23 Specific allele sizing of ELNil, ELM18, Helg 15/16, and LIMK1-GT was accomplished by initially running molecular weight ladders alongside a number of PCR amplified products from WBS patients and non-WBS control DNA samples. Standard allele controls were established that correspond with published allele sizes. These controls were then run with a molecular ladder alongside all subsequent PCR products of the polymorphism in question (Figure 2.2). 24 2.2.6 Screening for a deletion by microsatellite analysis The presence of heterozygosity was interpreted as the marker not being deleted when each allele is found in at least one of the parents; the lack of inheritance of a maternal or paternal band was evidence for a deletion of that marker; and presence of a single allele in the patient that was also present in both parents was considered uninformative. All markers were screened to avoid missing detection of smaller deletions than the previously reported WBS deletion extent (Perez Jurado et al., 1996). In patients for whom no parental blood was available, presence of multiple heterozygous markers within the deletion region were interpreted as non-deletion cases (Figure 2.3). 26 Figure 2.3: Microsatellite analysis for deletion. Top row illustrates alleles seen on polyacrylamide gels for a) non-deleted, b) deleted, and c) uninformative cases when both parental and patient DNA is available. Bottom row illustrates alleles seen on polyacrylamide gels for d) non-deleted and e) uninformative cases when only patient DNA is available. a) o b) c) a ab be cc Biparental inheritance: No deletion ac b be Paternal inheritance: Maternal deletion be bb ab Homozygosity: Uninformative d) e) ab Biparental inheritance: No deletion aa Homozygosity: Uninformative 27 2.2.7 Genetic mapping To determine the position of markers within the WBS deletion region, a crossover mapping panel for the WBS deletion region was constructed based on work started by a summer student, Julie Waslynka. To do so, CEPH family parental haplotypes were constructed for markers from D7S645 to D7S669 based on CEPH online mapping data (http://www.cephb.fr/cephdb/). This was accomplished by identifying grandparent-specific inheritance of the markers. Parental haplotypes were then used to identify parental and recombinant haplotypes in their offspring. Recombinant haplotypes were compared to control parental haplotypes to identify where the recombination had occurred to determine the most-likely order of the markers (Figure 2.4). Markers not included online by CEPH were typed by us (JW or MSW) in families showing recombinant haplotypes in the WBS deletion region. 28 6Z 9. \ o a o > o "A w > o o a » n n a O W U > o. sa a* s» o ca a a * a 65 O O O i 68 >• O O W K> H - j c r co a * c a n o § O to ti n n a Cl CO a > W >• O O r> r> r> CO CO > o co O > > 03 H tt n 60 I o o a* 2t CM C ft n 3 $ fD 3 •1 re * & 3 3 CO T3 ft c CO fD C CO I fD CO C CO fD 3 fD O 3 p fo 3 SL eT S. !T* to n » S-1. fD OQ "13 fD OQ o fa 2. &• co O g B f? a. P ° a-ft sr . co 3 1. CO 3 -2.3 Statistical analysis 2.3.1 Contingency table Data on multiple variables are often collected together in a study. To determine whether the occurrence of two distinct variables are correlated, the data are arranged in contingency tables with r rows and c columns, with a total of r x c "cells". The table then tests the null hypothesis that the observed data are found by chance with the numbers observed for rows not being correlated with those of the columns (Zar, 1984). Each cell of the table, Xy, is filled in with the number of cases observed, f]j, with the corresponding two variables. For instance, the following example, Table 2.1, correlates presence/absence of a congenital heart defect with gender of the affected. Table 2.1: Sample 2x2 contingency table. Congenital Heart Defect Gender Present Absent Row Total Male xn: fn xi2: fn Ri Female x2i: h\ X22: f22 R2 Column Totals Ci c2 Total (n) A From the observed numbers, % the expected number, % can be calculated under the null hypothesis in a 2 x 2 table with the following formula: A T where: xy- = cell identity (i = 1, 2; j = 1, 2); fij = observed number of cases in cell xy- (i = 1, 2; j = 1, 2); Q = the total for column j (j = 1, 2); 30 Ri = the total for row i (i = 1, 2); n = the total number of cases. To determine whether there was a significant difference between the observed and expected values, either Fisher's exact test or the chi-square test was utilized. Contingency tables were employed in this study to look for correlations between molecular and clinical variables as well as between different clinical variables. With the clinical data being presented in binary format- presence or absence of a feature, analysis of only 2 x c contingency tables were necessary. 2.3.2 Chi-square (%2 ) analysis The chi-square analysis is the most common way to determine whether the observed data are significantly different from the expected in contingency tables. The test gives a value that can be compared to tables giving estimates for their probability given a certain degree of freedom. This calculation is made using the formula: A A where: fj = the expected number of cases in cell xy (i = 1, 2; j = 1,2); fij = the observed number of cases in cell xy- (i = 1, 2; j = 1,2); E = the sum of calculation for each cell Xy . However, the chi-square can only give an estimate to the probability of observing such a data set as it describes a continuous distribution. The data collected in this study, however, represents a discrete distribution. This does not present a problem unless there is only one degree of freedom in which case, Yate's correction for continuity can be employed. This was not necessary in this study as the chi-square was only used to test goodness of fit for contingency tables larger than 2x2. 31 2.3.3 Fisher's exact test The Fisher's exact test was used to calculate the probability of 2 x 2 contingency tables. The advantage of using this test over the chi-square test lies in its calculation of the exact probability rather than an estimate. In addition, the Fisher's test is not affected by smaller sample sizes as the chi-square test is (Zar, 1984). The test simultaneously calculates the probability of observing by chance: 1) fn from a total of Ri cases; and 2) f i 2 from a total of R2 cases. This is accomplished by calculating the sum of the probabilities of observing the given data table and all tables more extreme using the following equation: RiJRaiCiJC2! P= nj fll!fl2!f2l!f"22! where: R! = the factorial of the total number of cases in row 1 or 2; C! = the factorial of the total number of cases in column 1 or 2; fy! = the factorial of the observed number of cases in cell xy (i = 1, 2; j = 1, 2); n! - the factorial of the total number of cases; and P = the probability of seeing the original data set if the null hypothesis is true. This would provide the calculation for a one-tailed hypothesis. That is, there is an a priori expectation for a correlation to exist between the variables. But if there is no a priori expectation, a two-tailed test is required. This is accomplished by finding the sum of P-values for both tails. To determine the P-value of the second tail, mi and m2 need to be determined, 'mi' is defined as the smallest of either Ri of Q (the marginal totals) of the 2x2 table, while 'm2' is defined as the smaller marginal total of the other axis. For instance, if Ci = 32 mi, then m2 would be the smaller of either Ri or Rj. Then, the value seen in the corresponding cell, Xy, is subtracted from mi, giving a value that then replaces the existing Xy to create new 2x2 matrix and acts as the most extreme case seen in the other tail. For instance, if m2 = R2, then the most extreme table of the second tail would have X12 replaced by a value, mi - X12 = xi2' (Table 2.2). Table 2.2: Most extreme table of second tail. Congenital Heart Defect Gender Present Absent Row Total Male X n xn' Ri Female x2i X22 R2 Column Totals Ci C 2 Total (n) The probability of observing such a table is calculated with the Fisher's formula listed previously and for all tables less extreme (where value of x i 2 ' is then decreased so as to make it less extreme by a unit of one progressively). P-values for all 2 x 2 tables of the 2nd tail, less than that obtained for the most extreme table are then summed to determine the probability of the second tail. Two-tailed calculations were calculated for all pair-wise comparisons in this study. 2.3.4 Significance levels The significance level, a, is defined as the criterion for rejection of the null hypothesis, that is, there is no correlation. This is usually set at a probability of 5% or less. Due to the large number of comparisons made in this study however, the chance of a 33 type I error, or the probability of rejecting the null hypothesis when it is in fact true, increases linearly with the number of comparisons. Therefore, Bonferroni's correction for multiple comparisons was used to set a significance level for each comparison made with an overall a of 0.05 so that a false correlation would not be reported. This was defined by the formula: a' = 005 k where: a' = significance level for each comparison; k = number of tests made. Five a's were established for the three components of the study: 1) 108 comparisons between molecular and clinical features of deletion cases (a' = 4.6X10-4); 2) 61 comparisons between molecular and clinical features of non-deletion cases (a' = 8.2X10"4); 3) 153 pair-wise comparisons between clinical features of deletion cases (a' = 3.3xl0"4); 4) 90 pair-wise comparisons between clinical features of non-deletion cases (a' = 5 .6X10" 4 ) and 5) 27 comparisons of clinical features between deletion and non-deletion cases (a' = 1.9xl0'3). Due to the inverse relationship between type I and II errors, such stringency increases P, the probability of type II errors. That is, not rejecting the null hypothesis when, in fact, it should be, or failing to report a true correlation. This, in turn, reduces the power of the test, or the ability to detect a true correlation since power is defined by 1 - p. For this reason, all p-values less than 0.05 are reported as they may indicate potential associations worthy of investigation in additional data sets. 2.3.5 Estimating power and necessary sample size 34 As mentioned in the previous section, the power of a statistical test is defined as the ability of the test to detect a true association. To estimate power of the Fisher's test, one normally incorporates the Z approximation in the calculation (Zar, 1984). Power = PJZ < - Z ^ V pq/Ri + pq7Rl=jfejUzj^'l+ PJZ > Z ^ V p q7R, + pq/RzzJ^^g) Vpiqi/Ri + p2q2/R2 J I Vpiqi/Ri + p2q2/R2 where: p = xn + xr[ & q = 1.0 - p ; Ri + R2 xn, X 2 1 , Ri, and R2 are defined in Table 2.1; Za(2) = to(2)iiif = critical values of the t distribution for a 2-tailed test and required significance level of a; pi = frequency of trait in population 1; p2 = frequency of trait in population 2. As standard t distribution tables only list critical values for a as low as 0.001, this to.ooi(2) value was used in the calculation despite the fact the required level for significance for some comparisons in this study was substantially lower in some comparisons. When planning data collection, equal sample sizes from the two groups being compared is preferred. However, this is not always practical and so estimation of the required sample sizes from each group can be calculated based on ratio between two groups being compared: r = R2/R, where: Ri and R2 are defined in Table 2.1. Therefore, the required sample size for RI is calculated by: 4r5 35 where: B = \L<2) V (r + 1) p'0q'0 + Zm V rpiqi + p2q2]2; P = 1 - power; 5 = |pi - P2I; p'o = £i + r p2 ; r+ 1 q'0= 1 - p'0; pi = frequency of trait in population 1; qi = 1 - pi = frequency without trait in population 1; p2 = frequency of trait in population 2; q2 = 1 - p2 = frequency without trait in population 2. It should be noted that the calculation for power and desired sample size call for use of the population frequency (i.e. pi, qi, p2, and q2). These true population frequencies are not known in this study and the observed frequencies of the traits were therefore used in place of the population frequencies. This introduces a bias in the calculations that is directly influenced by both the observed frequency and also the observed frequency difference of the trait between the two groups being compared. 2.3.6 Cluster analysis Cluster analysis is a method to estimate similarity or differences between pairs of objects. In this study, cluster analysis was applied to determine whether deletion and non-deletion WBS cases were 2 distinct populations. To do so, the clinical data were reduced to 10 factors with each factor described in binary format such that presence is represented by the value "1" and absence, by "0". These were then arranged into a data matrix with the following factors: 1) small at birth (where either weight or length is less than 10 centile); 2) small at subsequent physical examination (where weight or height is less than 10 centile); 3) hypercalcaemia; 4) cardiac abnormality; 5) typical facies; 6) blue and stellate iris pattern; 7) strabismus; 8) typical personality; 9) clinodactyly; and 10) dys/hypoplastic nails (Appendix HI). 36 From the data matrix, the similarity between each patient was calculated in a pair-i wise manner. This was accomplished by first creating a 2 x 2 table for each pair of patients that summarizes the total 1) 1-1 matches (features for which both patients are coded "1"), represented by "a"; 2) 1-0 matches (features for which patient j is coded "1" and k is coded "0"), represented by "b"; 3) 0-1 matches (patient j is coded "0" and k is coded "1"), represented by "c"; and 4) 0-0 matches (both patients are coded "0"), represented by "d" (Table 2.3). Table 2.3: 2x2 table giving the number of 1-1,1-0, 0-1, and 0-0 matches. Patient k 1 0 Patient j 1 a b 0 c d The Jaccard similarity coefficient, C, which indicates the overall resemblance between each pair of patients, is then calculated using the following formula: Qk = a a + b + c where: 1.0>Cjk>0.0; Cjk = 0.0 indicates no resemblance; and Cjk =1.0 indicates complete resemblance. This calculation does not include values of "d" as it has been argued that joint lack of features between patients should not be allowed to contribute to their similarity (Romesburg, 1984). 37 A resemblance matrix was then created, listing the calculated coefficient for each pair of patients. A sample matrix is indicated in the following Table 2.4: Table 2.4: Sample resemblance matrix with a total of 5 patients using the Jaccard coefficient of similarity. First patient in pair 1 2 3 4 5 1 - - - - -Second patient in pair 2 0.13 - - - -3 0.84 0.21 - - -4 0.65 0.68 0.61 - -5 0.28 0.71 0.67 0.72 -These values of similarity were then used to create a tree that represents the similarity between the patients. To do so, each patient is initially considered a "cluster", defined as one or more patients we are willing to call similar to each other. With each step in clustering, the 2 patients with the highest similarity become grouped together and become a new single cluster that is subsequently compared to other clusters and the number of clusters is reduced by one. A new resemblance was then created such that both rows and columns were reduced by one since the most similar patients, 1 and 3, were clustered to create a new cluster "13". The similarity coefficients between 2, 4, and 5 remained the same in the new matrix but when compared to cluster "13", a new similarity was calculated using the "unweighted pair-group method using arithmetic averages" (UPGMA). This was accomplished by calculating the averages of the values between the individual objects in one cluster and those in the other cluster. That is, the similarity between "13" and 2, is the 38 average of the similarity between 1 & 2, and 3 & 2. With the above given sample data, C o 2 = Vi x (C12+C32) = V2 x (0.13+0.21) = 0.17. Such a process was repeated until only one cluster remained. A tree can then be created from these resemblance matrices with the position of horizontal lines indicating level of similarity when two clusters were merged such as the example provided in Figure 2.5. Figure 2.5: Sample cluster analysis output tree. m O O o 3 0.0-0.2-0.4-0.6-0.8-1.0-0.84 0.72 0.37 0.70 Object While the tree represents intersubject similarities, it may not be necessarily representative of the original data matrix after applying UPGMA. Therefore, the degree to which the tree and resemblance matrix "say the same thing" must be measured. 39 However, a direct comparison between the tree and data matrix is not feasible and, therefore, the tree is compared to the resemblance matrix instead. This is accomplished by first converting the tree into a cophenetic matrix, with similarity values representing the degree of similarity at which the two subjects were joined into the same cluster on the tree. From Figure 2.5, this would give the following cophenetic matrix: Table 2.5: Cophenetic matrix converted from sample cluster tree (Figure 2.5). First patient in pair 1 2 3 4 5 1 - - - - -Second patient in pair 2 0.37 - - - -3 0.84 0.37 - - -4 0.37 0.70 0.37 - -5 0.37 0.70 0.37 0.72 -This cophenetic matrix is then compared with the resemblance matrix (Table 2.4) by calculating the cophenetic correlation coefficient, rx, Y- To do so, the values in each corresponding (j, k) cell from both matrices are first listed where the X list contains values from the resemblance matrix and the Y list contains values from the cophenetic matrix, such as the following sample table: Table 2.6: Unstrung list of resemblance and cophenetic matrix values. Cell: (2,1) (3,1) (4,1) (5,1) (3,2) (4,2) (5,2) (4,3) (5,3) (5,4) X: Y: 0.13 0.37 0.84 0.84 0.65 0.37 0.28 0.37 0.21 0.37 0.68 0.70 0.71 0.70 0.61 0.37 0.67 0.37 0.72 0.72 From this, rx, Y is then calculated using the following formula: 40 rx,v= Sxv-fl/nXSxXSy) , {[Ex2 - (l/n)(Ex)2][Ey2 - (l/n)(Ey)2]}1/2 where: 1.0 > rx,Y 2:-1.0; n = total number of cells in matrices; x = similarity value listed for rowX; y = similarity value listed for row 7; and xy = the product of similarity values for corresponding cell (j, k). This gives the following calculations: Exy = (0.13)(0.37) + (0.84)(0.84) + ... +(0.67)(0.37) + (0.72)(0.72) = 3.14 Ex = 0.13 +0.84+ ... +0.67 + 0.72 = 5.5 Ey = 0.37+ 0.84+ ... +0.37 + 0.72 = 5.18 Ex2 = (0.13)2 + (0.84)2 + ... + (0.67)2 + (0.72)2 = 3.57 Ey2 = (0.37)2 + (0.84)2 + ... + (0.37)2 + (0.72)2 = 3.87 rx,Y= 3.14-(1/10X5.S)(5.18) = 0.36 {[3.57 - (1/10)(5.5)2][3.87 - (1/10)(5.18)2]}1/2 A cophenetic correlation coefficient value of 1.0 indicates perfect concordance while a value of 0.0 indicates no concordance, and -1.0 indicates negative concordance. With rx, Y = 0.36, our example indicates that there is low concordance between the output tree and the resemblance matrix as a result of data distortion by cluster analysis. Generally, rx, Y ^ 0.8 is used indicate that using cluster analysis has not distorted the data significantly (Romesburg, 1984). 2.3.7 Odds Ratio and its confidence interval In this study, an odds ratio was calculated to determine whether the presence of specific clinical features would increase the probability that the individual carried a 7q 11.23 deletion. This was accomplished by arranging the data into a 2x2 table such as the following Table 2.7: 41 Table 2.7: Sample 2x2 table used to arrange data in order to calculate odds ratio. Presence of phenotypic feature Present Absent Presence of 7qll.23 deletion Present a b Absent c d From the given 2x2 table, the odds ratio was calculated with the formula: O.R. = ad be where: if OR. = 1.0, then presence of the phenotypic feature has no effect on whether the individual has a 7qll.23 deletion. A 99% confidence interval was also calculated. This represents the interval within which we can be certain of our odds ratio with 99% confidence. This utilizes the formula: C.I. = fadl exp J±Z0.99(i) / ! +1 +1 +1 \ [bej \ V a b c d J , where: a,b,c, and d are defined in Table 2.7; exp = e (2.718) to the power of; Zo.99(i) = 2.32635. 42 CHAPTER 3: Results 3.1 Genetic mapping Published maps of the WBS region initially showed a discrepancy of markers from and including ELN to D7S613 arising from studies by Perez Jurado et al. (1996) and Osborne et al. (1996). A genetic map was therefore created based on recombination analysis of CEPH individuals to help clarify this confusion. However, the discrepancy in marker order has since been resolved (Wu et al., 1998; Tassabehji et al., 1999). CEPH individuals listed in Table 3.1 were identified as being recombinant in the WBS deletion region based on haplotype information available online from the CEPH Database. Marker data was available for D7S672, D7S653, D7S675, and D7S669. After further typings, results for 141607 and 133203 were consistent with the previously published marker order of D7S672, D7S653, and D7S489U. Recombination events identified in individuals 10204, 136209, 88406, 133409, 141304, 141306, 141315, 141316, and 88411 are consistent with published marker orders of D7S2472, D7S1870, D7S489L, D7S675, and D7S669. Only 133404 carried a recombination event between ELN and D7S2472, but the data cannot resolve the discrepancy of marker order from and including ELN to D7S613. Analysis also indicated CEPH individual 133203 carries two recombination events with one between D7S672 and D7S653 and a second between D7S653 and D7S1870. This is almost certainly an error in genotyping of D7S653, which was typed elsewhere (CEPH Database). This result cannot be verified as attempts at PCR amplification of D7S653 and further markers within the second recombination interval 43 failed due to failure to amplify. With little DNA sample of 133203, the supply was exhausted and further DNA could not be obtained. It is also interesting that many recombinants were identified that span the distal deletion breakpoint. This would seem to be in agreement with the common WBS distal breakpoint being a site of instability and therefore prone to increased frequency of both recombination and deletions. 44 9v U) w .b. H o\ a g g g 3 3 3 -g 03 pa P 3 3 A ti & & a a K <• SI, SI, h to fai &, O OT Cu P . w O O OT OT Cu CU j O O O I OT OT pu Cu OT O OT I O OT Cu Cu OT O O O OT OT j O W Cu Cu I. 3 5 r° B u> o O o 8 Cu -S I-o 8 ? o o 3 cn a 4 ° g 3 r * cn § g-co g 8* 2- a o ui cn cr a tn g CD 3.2 Molecular-clinical correlations 3.2.1 Molecular Screening Of the 108 cases included in this study, all patients contributed from Greece (7) and from Manitoba (7) were confirmed for a deletion locally, either by FISH or by molecular screening. These results were provided by Drs. M. Petersen and B. Chodirker, respectively. Patient DNA was not made available for further analysis of deletion size here in this study. Deletion was also confirmed locally for all 51 patients contributed from Zurich but DNA was available for only 38 of these cases for further delineation of deletion extent. Screening for presence of a deletion of remaining WBS cases revealed 12/25 BC cases and 6/16 Saskatchewan cases carried a deletion. DNA was not available for BC cases v86 and vl28 but these cases were confirmed for a deletion at the BCCH Cytogenetic Diagnostics Laboratory by FISH. 3.2.2 7q 11.23 deletion cases In total, 85 WBS cases were confirmed to carry a deletion by either FISH or molecular methods in this study. Deletion cases with DNA available (56) were screened for three factors: 1) deletion extent; 2) parental origin of deletion; and 3) one LIMK1 and three different ELN genotypes. These three factors were analyzed in order to look for associations with clinical features. The gender of confirmed deletion cases was also compared with clinical features. Screening for the deletion extent revealed a constant deletion size of approximately 2 cM, with a centromeric breakpoint between D7S489U and D7S2476 in all cases for which the markers were informative. The telomeric breakpoint was found to 46 be variable with it lying between D7S1870 and D7S489L in 14 cases for which markers were informative, and between D7S489L and D7S675 in two cases, WBS 9 and 24. D7S489L was uninformative in 45 cases and therefore, the breakpoint could theoretically lie anywhere between D7S1870 and D7S675. The results of deletion extent in confirmed deletion cases are summarized in Figure 3.1. This figure does not include results for Saskatchewan cases (WBS v90 - v96) and BC cases v86 and vl28, for which DNA was not available, nor 5 of the 7 Greece cases (WBS vl06 - vllO), for which no details of deletion extent were provided. As deletion extent was consistent, with only cases WBS 9 and 24 being detectably larger, comparison of clinical features was made only between these two patients and the remaining deletion cases. These revealed that WBS 9 and 24 did not present with any unusual clinical features not found in other WBS deletion cases and were considered phenotypically typical. Therefore, no evidence was found to support a correlation between deletion extent and clinical outcome. 47 817 ; I 3- P • u a ! I <v o 'z c o < p S (1 O 3 m B p S o ^ 5' §• 3 & (1 ft s £. 2 o " e £• 2 °* & TP. 3, ? WBS Deletion Region R ES O ti) 6 H o 8-a •o a a r - j - j - j xn c« Vi cfir 4* to 55 00 00 4^  t—» - a - J L/i r o to 9 ON ffl ffl to 4*. 4*-00 O 3 g a rt I 3 O 3. <B. 3. H • m u • • • u • m P El • m m • n El • m • M • • m m • El El • m • M El m m • El m a m • M • • m m El El m • m • M m • m m m • m • m • P m • m m m • m m • M m El m • El U m El m m m El m • m • M m • m m H M • El m m M m • m u m • m • m m M m El m • m • m • m H P • El m • u El m n m • M • m El m n m M m El m m m El m n m B U m El m m • El M • m m P m • • • • • m n m m P m El • n m El m • m • M • • • • • El • • m m P • • • m m El • m m P m • • u m El • m u • P ® • El m * El El m m El M m • 0 m * m El m m M m • • m m m El m m • P m • m u • m m m P m El El m m m • m m m M m El • m • n • m u m P m • • m • m • m m • M m • • m m • • m m El P a El m m m El • m m • P m • m m m • El • m • M • El m m m • El n m • M m El m m m El H • m El P m El m m m • • m • 0 M m • El • • M m • m El • m El M • • m m m El m m m El P m El • m El 0 * m El P El m m El m m U m • u m m El m u m El M m • * m • El m m m El M m • El • m m • M • • m m El • m m m M n El m m El El El m U • • m m El El • m m • M m • m m El El • m m El M m El m u El El • • m El P • M • m P • M • • m El m M • El m m • El H n m m P • El m m El El n m m U m • El m M m El u • m m P m El m m m • • m m M • • m m m El El m m • M El • • m • M El El m m m • • • u M • • m • m El P m El m m m El n m • • P m • m m m El m • • M • • m m ® El • m • El P n El u m m El El u • El M • • P • M El 0 m m El m n m U El • • m u • m n m El P Ei El m • m m H P El • • u m • P *0 O 3 < 11 a 0 rT 2 S 0\ Q 3 3 1 p. to < g ^ S H - ^ a 00 a. w tyj >^ fa „ fa a ft> p cn Using molecular methods, a non-significant bias was revealed when parental origin was analyzed in cases where DNA was available (38 maternal compared to 27 paternal). In addition, 6 FISH-diagnosed cases were also screened using molecular methods but were uninformative for all markers tested. These cases were designated "Uninformative" for parental origin of deletion in Figure 3.1. Pair-wise comparisons were made between parental origin and 17 WBS clinical traits with the 65 informative deletion cases. As clinical data were not available for all traits analyzed for each patient, all comparisons resulted in samples smaller than 65. Comparisons indicated deletion cases of paternal origin were more likely to have a hoarse voice (p=0.047) and a small head circumference (OFC <10th centile, p=4.4xl0"3) but these associations are not significant after Bonferroni's correction for multiple comparisons (a' = 4.5xl0"4). Summary of pair-wise comparisons is listed in Table 3.2. 49 Table 3.2: Pair-wise comparisons between parental origin of deletion and clinical outcome. Required level of significance, a - 4.6X10"4 Parental origin of deletion p Maternal (38) Paternal (27) (a'=4.6xlO~A) At birth: Weight (<10%) 6/26 23.1% 10/21 47.6% n.s. Length (<10%) 4/17 23.5% 5/16 31.3% n.s. At examination: Weight (<10%) 9/24 37.5% 10/20 50.0% n.s. Height (<10%) 10/23 43.5% 14/22 63.6% n.s. OFC (<10%) 5/22 22.7% 14/21 66.7% 4.0x10" Cardiovascular defect 17/26 65.4% 18/23 78.3% n.s. Typical facies (>4/7 features)3 16/23 69.6% 14/19 73.7% n.s. Strabismus 15/25 60.0% 14/20 70.0% n.s. Stellate irides 13/25 52.0% 10/20 50.0% n.s. Blue irides 12/24 50.0% 8/20 40.0% n.s. Hands: clinodactyly 14/25 56.0% 14/22 63.6% n.s. dys/hypoplastic nails 7/24 29.2% 9/22 40.9% n.s. Infantile hypercalcaemia 6/16 37.5% 5/13 38.5% n.s. Hoarse voice 8/12 66.7% 12/12 100.0% 0.047 Renal/urogenital abnormalities 2/20 10.0% 7/21 33.3% n.s. Dental abnormalities 12/22 54.5% 15/18 83.3% 0.054 Typical personality (>2/4 features) 16/24 66.7% 9/17 52.9% n.s. 3 Correlation between parental origin of deletion and each of the 7 facial features also did not reveal associations, (n p>0.05) Genotyping 56 deletion-confirmed patients with DNA available for the tetranucleotide repeat, ELNil, revealed 5 of the 11 possible alleles to be present in this WBS sample. Alleles of size 286, 282, 276, 272, 258, and 254 bp were absent. However, the allele frequencies observed were not significantly different from expected values when tested with chi-square (Urban et al., 1997) and are summarized in Table 3.3. Table 3.3: Expected and observed ELNil allele frequencies in deletion cases with expected values based on those published by Urban et al. (1997). Allele 1 2 3 4 5 6 7 8 9 10 11 Size (bp) 286 282 278 276 274 272 270 266 262 258 254 Expected frequency (%) 0.2 2.3 8.6 0.2 23.0 0.4 25.9 22.8 11.1 4.8 0.6 Observed frequency (%) 0 0 8.9 0 33.9 0 25.0 16.1 16.1 0 0 Comparison of ELNil alleles with clinical outcome of deletion-confirmed cases revealed an association with alleles smaller than 272 bp (alleles 7-11) and presence of a congenital heart defect (one or more of supravalvular aortic stenosis, peripheral aortic stenosis, atrial septum defect, ventricular septum defect, or mitral valve prolapse; Table 3.4a). The p-value of 0.023 was not significant after correction for multiple comparisons. Adjacent alleles were grouped together rather than non-adjacent alleles because similar sized alleles are more likely to be on related haplotypes. When the ELNil alleles were not grouped, no associations were observed with presence of congenital heart defects (Table 3.4b) or any other WBS clinical feature. <E O ra 1 I §• Q co CO O o* P •£> o I & M s co "'I co 13 i I o I I co* I s ! I' Bt P. g o o 3 o 3 O o I s. cn O 13 r o I g NJ 00 o UJ N ' B co co r§" N i I* 4^  i V NJ 00 NJ 00 VO o © U l an > /—s CB CO* co* CO (TO* h—> U l V—» Cs NJ os NJ a* •o NJ -£>-4*. NJ c r T3 U l UJ a* •a 8 l l 4>-bs X K o *—-f o © v—» -4 N> as © NJ O Sfi fc U l ' 1—1 1 - 1 I 5; * ri* Os as oo 'as T3 o i v O I 1 a. CO - 4 00 as 3^ oo NJ U i O 2? fc NJ -4 © NJ Os OS NJ OS NJ NJ U l 00 NJ U i I ? OS vn X * o NJ 00 as H I* NJ 00 NJ ( v i NJ 00 UJ NJ Os NJ -4 4^  NJ - J NJ as 0 n 1 CL NJ © NJ © © NJ UJ EH fc CO CO NJ 00 OS NJ -4 NJ NJ © i NJ U l 4*-4*-Os 13 © i v p a r n o - S /—\ v D 2, 89 a a B r OQ I — cro KJ\ re a o <2 •a rt 00 CTs ft 1 — 1 — Q 5' c*' 2 S. 3 re f 2, o ^ •— OQ" rs 3. S re o v P £ • p » if -a 3 ^ re o B a re" co a. o I-I o a Analysis of the ELN intron 18 polymorphism, ELNil8, revealed 6 out of the 8 known alleles in our deletion-confirmed sample at frequencies comparable to previously published results (Foster et al., 1993). The two rare alleles (175 and 161 bp) were not identified in our sample of 56 cases. Observed and expected frequencies are listed in Table 3.5. Table 3.5: Expected and observed ELNil8 allele frequencies in deletion cases with expected values based on those published by Foster et al. (1993). Allele 1 2 3 4 5 6 7 8 Size (bp) 175 173 171 169 167 165 163 161 Expected frequency (%) 0.7 3.3 2.8 12.8 19.8 4.1 56.1 0.6 Observed frequency (%) 0 3.6 1.8 14.3 16.1 3.6 60.7 0 Comparison of ELNil8 alleles with clinical features of WBS revealed an association between alleles of size 165 to 175 bp (alleles 1-6) and low birth weight ^lO* centile) in the deletion-confirmed sample (p=0.020; Table 3.4c). As for ELNil, no associations were detected when ELNil8 alleles were not grouped (data not shown). The ELN RFLP Helg 15/16 was genotyped in the deletion group, revealing all 3 alleles expected at frequencies comparable to those previously reported in a study of 60 unrelated WBS cases (Mari et al., 1995). This marker was also genotyped in additional Zurich cases for which DNA was not available in Vancouver, thereby increasing the sample to 61 cases. Allele frequencies can be found in Table 3.6. 53 Table 3.6: Expected and observed Helg 15/16 allele frequencies in deletion cases with expected values based on those published by Mari et al. (1995). Allele 1 2 3 Size (bp) 262 244 175 Expected frequency (%) 40.0 44.0 16.0 Observed frequency (%) 32.8 47.5 19.7 An association was found in the deletion cases between the Helg 15/16 alleles and "typical WBS personality" (defined by presence of at least 3 of the features highly active, highly sociable, overly anxious, and loquacious). However, this was not significant after Bonferroni's correction (p<0.05; Table 3.4d). Analysis of the LTMK1 dinucleotide polymorphism, LIMKl-GT, revealed 7 of the 8 alleles (the 132 bp allele not being found in our sample) at frequencies comparable to those obtained in a study of 82 unrelated, healthy Caucasians (Mari et al., 1998). The genotyping results of 56 deletion-confirmed WBS cases are summarized in the following Table 3.7. Table 3.7: Expected and observed LEVIK1-GT allele frequencies in deletion cases with expected values based on those published by Mari et al. (1998). Allele 1 2 3 4 5 6 7 8 Size (bp) 150 148 146 144 142 140 132 124 Expected frequency (%) 1.7 5.3 7.1 48.2 10.7 8.9 5.3 12.5 Observed frequency (%) 1.8 5.4 12.5 53.6 7.1 7.1 0 12.5 54 LIMK1 is expressed highest in fetal and adult brain but is also expressed above basal levels in the peripheral nervous system, olivary nucleus, the cerebellar ependymal layer, and part of the inner ear (Frangiskakis et al., 1996). Thus pair-wise comparisons of the inherited LIMK1-GT allele with clinical features were restricted to nervous system-related traits. As in depth psychological evaluation data was not available for this study, degree of mental retardation could not be correlated with the LIMK1-GT genotype. Only the previously mentioned 4 patient personality features, based on subjective clinician evaluation, could be correlated with the molecular data. This identified only loquaciousness as a possible associate with the LIMK1-GT genotype, although this again was not significant after Bonferroni's correction (p=0.037; Table 3.4e). The LIMK1 and three ELN polymorphisms were also examined for linkage disequilibrium in the deletion cases, revealing disequilibrium between ELNil8 allele 7 and Helg 15/16 allele 2 (p=7.0xl0"4; Table 3.8a). This "7 - 2" haplotype was compared with presence or absence of phenotypic features of WBS. A negative correlation was found between the haplotype and presence of the typical WBS personality (Table 3.8b). 55 Table 3.8: (A) Linkage disequilibrium between ELNil8 allele 7 and Helg 15/16 allele 2. (B) Correlation found between absence of "7 - 2" haplotype and presence of typical WBS personality. Required level of significance, a'=4.6xl0"4. (A) Linkage disequilibrium Haplotype 7-2 7 - not 2 not 7 - 2 not 7 - not 2 Observed frequency 0.404 0.193 0.088 0.316 Expected frequency 0.288 0.319 0.187 0.206 (B) Correlation with clinical feature Haplotype 7-- 2 not "7 - 2" p (a'=4.6xl0-4) Typical personality (>2/4) 5/15 (33.3%) 23/33 (70.0%) 0.201 The study population consisted of 40 affected males and 37 females (not statistically significant) with a proven 7qll.23 deletion with the gender of the remaining 14 cases not indicated on the dataforms contributed by the referring clinicians. Pair-wise comparisons between gender and clinical presentation indicated affected males were more likely to have congenital heart anomalies (p=0.043), a "typical" facial appearance defined as presence of more than 4 of the 7 facial characteristics listed in Appendix II (p=0.037), and nail dysplasia (p=0.006). However, these results are not significant when multiple comparisons are taken into account (Table 3.9). 56 Table 3.9: Effect of gender on the phenotypic outcome in WBS cases with a proven 7qll.23 deletion. Required level of significance, a-4.6xl0"4. Gender of Affected P Male (40) Female (37) (a'=4.6xl0"4) At birth: Weight (<10%) 11/20 55.0% 13/30 43.3% n.s. Length (<10%) 5/21 23.8% 8/20 40.0% n.s. At examination: Weight (<10%) 15/29 51.7% 13/28 46.4% n.s. Height (<10%) 15/28 53.6% 18/30 60.0% n.s. OFC (<10%) 13/25 52.0% 12/29 41.4% n.s. Cardiovascular defect 26/31 83.9% 19/31 61.3% 0.043 Typical facies (>4/7 features)3 21/27 • 77.8% 12/24 50.0% 0.037 Strabismus 17/27 63.0% 13/26 50.0% n.s. Stellate irides 13/27 48.1% 13/27 48.1% n.s. Blue irides 12/24 50.0% 10/25 40.0% n.s. Hands: clinodactyly 16/27 59.3% 13/28 46.4% n.s. dys/hypoplastic nails 15/25 60.0% 7/30 23.3% 6.0xl0"3 Infantile hypercalcaemia 8/21 38.1% 10/21 47.6% n.s. Hoarse voice 10/12 83.3% 10/15 66.7% n.s. RenalAirogenital abnormalities 6/23 26.1% 5/29 17.2% n.s. Dental abnormalities 16/25 64.0% 16/24 66.7% n.s. Typical personality (>2/4 features) 13/24 54.2% 14/32 43.8% n.s. 3 Correlation between gender and each of the seven facial characteristics also did not reveal any associations, (n.s. p>0.05)57 3.2.3 7qll.23 non-deletion cases Molecular screening revealed a total of 23 cases not deleted at 7qll.23, of which 13 were from BCCH and 10 were from Saskatchewan (data not shown). Screening these cases for a 22qll deletion revealed one case, WBS v72, which carried an apparent paternal deletion of the dinucleotide marker, D22S944. Inheritance of this marker is also compatible with a null allele segregating. Two proximal markers, D22S1638 and D22S941, were uninformative while all other VCFS/DGS deletion region markers tested were not deleted (Figure 3.2). This patient, a girl diagnosed at BCCH, had facial features suggestive of WBS including a periorbital fullness, full cheeks and lips, a flat nasal bridge, and upturned nares but she did not have the characteristic wide mouth. She also presented with esotropic strabismus but was corrected by surgery at 1 year 11 months. Clinical details at birth including weight and length were not indicated in the patient chart. At the age of 2 years 11 months, she was above the mean for weight (11.7 kg = 75 centile) and OFC (49.5 cm = 50th centile) but was short for age (87.4 cm = 5th centile). Serum calcium levels were found to be elevated (2.79 mM at 2 years 4 months) but returned to within normal limits (2.45 mM at 2 years 8 months) following 4 months of a vitamin D and calcium-reduced diet. However, no renal abnormalities were detected. Examination of the heart did not reveal any abnormalities. Personality assessment of the girl indicated that she did not possess typical "cocktail behavior" but was "reasonably active and energetic". No other unusual clinical features were indicated in this patient's file. 58 6£ • • . M p Z c o < n CD M O CD* 3 WBS Deletion Region a a a a - J CA! ca cn 4> to C7> oo 4- 1— -J r o to • • • VCFS/DGS Deletion Region a 0 c a a a a a -o (O to to to to (O a" to N) so to to to to to men cn C/l cn cn GO CO cn cn men UJ UJ Ki —i so vD • 4-men o z o> _ JO Os to men Os — to JO :—* UJ -o UJ 0 0 X a w a O n cs r r- - J - J h cfq* z; cn cn a — > to g -fc- •00 o * ~ J SO me * OS C me p. » * o CD re 3 • • • • i l • • C to to cn to -O O o to to cn UJ o o t o t o cn o to to cn to Cs 4 -C ro to cn OS to WJ o & M CD QJ Sri « P 3 S a g 3. o & OQ o ^ FT e e &. s o » 3 s (a a. S £ 1 — 1 r* 3 S I- = q >* ° 1 NJ •< NJ < vo «, £ ^ S. M s. CO W 3 2» SL 5. S. 3 1. 5* o a _ CO ro S >— co - t OL a i-h g. £? 5? g — to I-. OJ B ° *r e s ^ 3 q Genotyping 23 non-deletion WBS cases for ELNil, ELM18, Helg 15/16, and LIMK1-GT revealed allele frequencies consistent with published frequencies, indicated previously in Section 3.2.2 (data not shown). To compare ELN and LEvTKl genotypes with clinical data, the alleles were grouped due to the small sample size and the predominance of 1 allele for each polymorphism. That is, ELNil alleles were grouped into 2 groups: A) alleles 1-6 and B) alleles 7-11; ELM18 alleles were grouped into A) allele 7 and B) not allele 7; and LIMK1-GT was grouped into A) allele 4 and B) not allele 4. Helg 15/16 alleles were grouped in 2 different ways because when comparisons were made with clinical features, %2 tables with cell values of zero were sometimes created. Therefore, Helg 15/16 alleles were grouped as either A) allele 1 and B) not allele 1; or A) allele 2 and B) not allele 2 (Table 3.10). %2 analysis revealed no correlations between either ELN or LIMK1 grouped genotypes, and any WBS clinical feature in the non-deletion cases (data not shown). Genotypes could not be compared with presence or absence of infantile hypercalcaemia, a hoarse voice, or with low birth length (<10 centile) as data on these traits was available for only 5, 2, and 3 cases, respectively. 60 Table 3.10: ELN and LIMK1 polymorphism allele groups and "genotypes" used in correlating with clinical data of non-deletion WBS cases. Two different allele groups were used for Helg 15/16, denoted "a" and "b", respectively. Marker Allele Group Alleles included Allele size (bp) ELNil A 1-6 286 - 272 B 7-11 270 - 254 ELNil A 7 163 B 1-6&8 175 - 165 & 161 Helg 15/16a A 1 262 B 2&3 244 & 175 Helg 15/16b A 2 244 B 1 &3 262 & 175 LIMK1-GT A 4 144 B 1-3, 5-8 150- 146, 142-124 Clinical data were also compared with the gender of the affected individual. This revealed that affected males were more likely to have renal/urogenital abnormalities in i the non-deletion WBS group (Table 3.11). This was not significant after the correction for the 61 comparisons made, requiring a significance level of 5.6X10"4 No other association with p<0.05 was found between gender and presence of a clinical trait but sample sizes for some features were too small to make a comparison meaningful and were therefore not included in calculating the corrected level of significance. 61 Table 3.11: Effect of gender on the phenotypic outcome in WBS cases without a 7qll.23 deletion. Required level of significance, a-S.lxlO'4. Gender of Affected p Male (13) Female (10) (^=^.2x10^) At birth: Weight (<10%) 1/9 11.1% 2/8 25.0% n.s. Length (<10%) 1/2 50.0% 0/1 0.0% -At examination: Weight (<10%) 4/9 44.4% 3/10 30.0% n.s. Height (<10%) 4/9 44.4% 4/10 40.0% n.s. OFC (<10%) 2/8 25.0% 1/8 12.5% n.s. Cardiovascular defect 2/7 28.6% 2/9 22.2% n.s. Typical facies (>4/7 features)3 2/6 33.3% 4/8 50.0% n.s. Strabismus 3/8 37.5% 1/8 12.5% n.s. Stellate irides 2/6 33.3% 4/8 50.0% n.s. Blue irides 4/6 66.7% 5/7 71.4% n.s. Hands: clinodactyly 3/7 42.9% 3/7 42.9% n.s. dys/hypoplastic nails 1/5 20.0% 1/6 16.7% n.s. Infantile hypercalcaemia 0/2 0.0% 2/3 66.7% -Hoarse voice 0/1 0.0% 0/1 0.0% -Renal/urogenital abnormalities 3/5 60.0% 0/8 0.0% 0.035 Dental abnormalities 4/7 57.1% 4/7 57.1% n.s. Typical personality (>2/4 features) 3/6 50.0% 2/6 33.3% n.s. Correlation between gender and each of the seven facial characteristics also did not reveal any associations, (n.s. p>0.05)62 3.3 Clinical correlations Pair-wise comparisons between clinical features were conducted for both deletion and non-deletion cases to determine whether a common etiology exists for certain traits and all associations with a probability less than 5% are given in Table 3.12. In WBS cases with a confirmed 7qll.23 deletion, two significant co-occurrences were found: 1) low birth weight and poor postnatal weight gain (<10th centile at the time of examination; p=9.0x!0"5); and 2) transient infantile hypercalcaemia and stellate eye patterns (p=1.0xl0"5). Other potential associations between presence of both 1) heart and renal/urogenital abnormalitites (p=7r9xl0"3); 2) a stellate eye pattern and characteristic behavioral patterns that include loquaciousness, anxiety, sociability, and hyperactivity (p=6.8xl0"3); and 3) a stellate eye pattern and blue iris (p=1.2xl0"3) were not significant using an oc^JxlO"4 In the non-deletion cases, meaningful pair-wise comparisons could not be made with presence or absence of infantile hypercalcaemia and of a hoarse voice or with short birth length as the sample sizes were too small. Comparisons of remaining clinical traits resulted in 90 comparisons with an ot'=5.6xl0"4 and revealed no significant associations (Table 3.13). However, p-values less than 0.05 were found between 1) low birth weight and poor postnatal weight gain; and 2) co-presence of blue and stellate patterned irides, as was found in the WBS deletion group. A p-value of 0.046 was found between presence of stellate irides and absence of a typical WBS personality, which is contrary to that seen in the deletion group. The non-significant co-occurrence of a small OFC and low weight at the time of physical examination in the WBS non-deletion group (p=0.036) was not 63 found in the deletion group, although small O F C was correlated with short stature at the time of physical examination (p=0.036). Table 3.12: Analysis for associations between WBS clinical features in 85 7qll.23 deletion cases. Clinical data was not available for all traits analyzed resulting in samples <85 in some comparisons. Required level of significance, a'=3.3xlO"4 Birth weight (<10%) Cardiovascular defects Weight at time of examination <10% 17/27 25/29 >10% 4/31 20/31 p (a'=3.3xl0"4) 9.0xl0"5a 0.049 Infantile hypercalcaemia Blue irides Typical personalityb Stellate irides Present 14/19 16/23 16/19 Absent 1/19 7/29 12/27 p (oo'=3.3xl0"4) 1.0xl0"5a 1.2xl0"3 6.8xl0"3 Cardiovascular defects Renal/urogenital abnormalities Present 12/12 Absent 26/42 p (a'=3.3xl0"4) 7.9xl0"3 O F C <10% Stellate irides Height at time of examination <10% 19/32 10/30 >10% 8/25 15/24 p(a'=3.3xl0"4) 0.036 0.031 Dental abnormalities Typical personalityb Blue irides Present 17/21 16/20 Absent 13/27 12/26 p(a'=3.3xl0"4) 0.020 0.020 Typical facies0 Typical personalityb Clinodactyly Present 25/32 21/29 Absent 11/21 8/19 p (a'=3.3xl0"4) 0.049 0.036 Dental abnormalities Dys/hypoplastic nails Hoarse Voice Present 15/21 11/22 Absent 1/7 0/8 p (a'=3.3xl04) 0.013 0.013 3 p-values are statistically significant. b Typical personality defined as >2/4 WBS personality trais. c Typical facies defined as >4/7 WBS facial characteristics. 64 £9 E. ft i p. s. g V to X) <t i-i co O 3 if o t/3 I 5: > 5? 8 S o o o to as a! a! o 4* ON 4^  O 3s D; R ll" U l ON X o IV A i—' t-^ o o es as I-1 O t o £ C*> Os 3 ft i-l S» t= re u> • s p CO co p N fD Cu fD CO C ©• •* &» CO CO O O 5' co C. " O 9 co P fD O" co re fD* Ig to ~ co 2 * s 5' £. co S* o 5. 3 « s e. P CO fD CO re fD 1. fD Cu 5* 2-o >-+> CO oo' 3. th o o fD re CO W s o 9 i a re o s o as CO re CO O R. H ON 3. o p. D-O . ' * fD i-t fD a o 3 2. I rT 3.4 Clinical difference between deletion and non-deletion cases Twenty-seven WBS clinical feature comparisons were made in a pair-wise fashion between the 85 deletion and 23 non-deletion cases in order to identify whether phenotypic differences exist between these groups. Deletion cases were much more likely to have a congenital heart abnormality (p=5.0xl0"4), to be highly sociable (p=5.5xl0"3), and to have a wide mouth (p=7.2xl0"3) than non-deletion cases. When all three features were considered simultaneously, deletion cases were also found to be much likely to present with all three features (p=6.0xl0"4). Using an a'=1.9xl0"3, only the first was statistically significant. Comparisons between the 2 groups are summarized in Table 3.14. In an effort to further compare deletion and non-deletion cases, cluster analysis was applied using SAS/STAT software. However, complete clinical data were available for only 10 of the 108 cases (WBS z82, v84, v86, v87, z89, vl06, vl07, vllO, v l l l , and vl20). This resulted in SAS/STAT excluding 98 cases from the analysis. Attempts to remove clinical features with the most missing values (hypercalcaemia, dental anomalies, and strabismus) from the data matrix could not increase the total sample analyzed by SAS/STAT beyond 17 individuals (data not shown). This method was therefore uninformative in regards to identifying clinical similarities and differences between deletion and non-deletion WBS cases in this study. 66 Table 3.14: Summary of pair-wise comparisons of WBS phenotypic features between 85 7qll.23 deletion cases and 23 non-deletion cases. Required level of significance, a-l^xlO"4 7qll.23 Deletion P Present (85) Absent (23) (a'=1.9xl0"3) At birth: Weight (<10%) 24/62 38.7% 3/17 17.6% n.s. Length (<10%) 13/44 29.5% 1/4 25.0% -At examination: Weight (<10%) 29/60 48.3% 7/19 36.8% n.s. Height (<10%) 35/62 56.5% 8/19 42.1% n.s. OFC (<10%) 27/58 46.6% 3/16 18.8% 0.040 Cardiovascular defect 48/65 73.8% 4/16 25.0% 5 . 0 X 1 0 - 4 3 Typical facies (>4/7 features): 37/55 67.3% 6/14 37.5% n.s. Broad forhead 31/54 57.4% 7/15 46.7% n.s. Periorbital fullness 47/60 78.3% 10/14 71.4% n.s. Epicanthic folds 34/59 57.6% 7/14 50.0% n.s. Flat nasal bridge 36/59 61.0% 8/14 57.1% n.s. Long smooth philtrum 44/59 74.6% 8/14 57.1% n.s. Full cheeks and lips 58/62 93.5% 12/15 80.0% n.s. Wide mouth 55/61 90.2% 8/14 57.1% 7.2xl0"3 Strabismus 33/57 57.9% 4/16 25.0% 0.020 Stellate irides 27/57 47.4% 6/14 42.9% n.s. Blue irides 23/53 43.4% 9/13 69.2% n.s. Hands: Clinodactyly 32/60 53.3% 6/14 42.9% n.s. Dys/hypoplastic nails 22/59 37.3% 2/11 18.2% n.s. Infantile hypercalcaemia 19/44 43.2% 2/5 40.0% -Hoarse voice 22/29 75.9% 0/1 0.0% -Renal/urogenital abnormalities 12/54 22.2% 3/13 23.1% n.s. Dental abnormalities 34/52 65.4% 8/14 57.1% n.s. Typical personality (>2/4 features): 29/48 60.4% 5/12 41.7% n.s. Hyperactive 23/46 50.0% 7/16 43.8% n.s. Highly sociable 47/54 87.0% 9/17 52.9% 5.5xl0"3 Anxious 33/51 64.7% 9/12 75.0% n.s. Loquacious 34/50 68.0% 9/15 60.0% n.s. Cardiovascular defect, wide mouth, and highly sociable: 28/55 50.9% 1/17 5.9% 6 . 0 X 1 0 - 4 3 i-value is statistically significant after correction for number of comparisons. CHAPTER 4: Discussion 4.1 Genetic mapping The original discrepancy in marker order from and including ELN to D7S613 had already been resolved in previously published studies (Wu et al., 1998; Tassabehji et al., 1999). But the construction of a genetic map of the WBS common deletion region using recombination data on CEPH individuals is still worthwhile as it provides a framework with which to determine the location of newly found markers. The recombination data from this study (Table 3.1) would not have been able to resolve the original discrepancy but are still compatible with the marker order determined in more recent studies (Wu et al., 1998; Tassabehji et al., 1999). The inability to resolve the discrepancy by this study, however, is an indication of the maximum resolution of this method as recombination likelihood decreases with increasing proximity of the markers. For this reason, it is also unlikely to observe two recombination events within the WBS deletion region, making the double recombination event observed in CEPH individual 133203 likely to be a genotyping error. 4.2 Statistical power Typically in a statistical study, one sets a=0.05 to indicate that the type I error, or probability of reporting a 'false positive', is less than 5%. In a study with multiple comparisons, it is important to consider use of a corrected, lower level of a to keep the experimental-wide probability of a type I error at 5%. In studies such as the present, where many comparisons are made with a limited amount of data, it may be difficult to reach statistical significance. Furthermore, using a strict criterion for a necessarily results 68 in a loss of power. That is, the chance of missing a true association by using too strict a criterion increases (greater value of P). It is often a value judgement as to find a balance between the risk of false positives (type I errors) and the risk of missing what are in fact, true associations (type II errors). It is therefore important to report, not only those associations which reach a criterion accorded to an experiment-wide a=0.05, but to report all potential associations (p<0.05) which might be interesting to follow-up in an additional study. A two-tiered strategy, whereby a small set of data is used to identify the most interesting associations and a second study is used to follow-up these associations, is the most efficient design strategy. The present WBS study can thus be considered to be step one of such a two-tiered design. To demonstrate the effect of using an extremely strict a however, let us consider the following examples of approximate power and required sample size calculations. These were determined using formulas described in section 2.3.5: Table 4.1: Sample calculations of statistical power and required sample size to detect true associations for a given a. Population frequencies of the trait and sample size are represented for groups 1 and 2 by pi, p2, and ni, n2, respectively. Probability of a type II error, P, is determined by P=l -power. Needed sample sizes to detect an association at P=0.10 are represented by Ni & N 2 for the two groups. Group 1 Group 2 a= 0.001 a=0.05 7qll.23 Deleted Non-deleted P needed P needed Clinical trait pi ni P2 n2 N i & N 2 N i & N 2 a. Cardiovascular 0.738 65 0.250 16 0.345 107 & 0.031 57 & defect 26 14 b. OFCat 0.466 58 0.188 16 0.936 293 & 0.480 150 & examination 81 41 c. Renal/urogenital 0.222 54 0.231 13 0.999 256,831 & 0.948 129,045 & defects 61,830 31,066 69 From the sample calculations given in the above Table 4.1, it is apparent that use of a strict criterion for significance (a=1.0xl0"3) severely reduces the power of the test to detect a true association unless the difference between the two groups is large. Even more apparent, is the requirement for an enormous sample size to detect associations, should it be desired to increase the power of the test to 0.90 and keep the significance level fixed at a=1.0xl0" . Lowering the required significance level to oc=0.05 while maintaining the power at 0.90 dramatically reduces the required sample size to detect an association. It is difficult to calculate exact power for each comparison in the present study however, as the true 'p's and 'q's are not known. These values can only be estimated in this study by substituting the observed 'p's and 'q's. One might be able to use a bootstrapping method to achieve a better estimate of these values, but again will be different for each comparison. Additionally, critical t-values do not go as low as that needed for this study in standard tables, as they are only available for ct>1.0xl0"3. Thus, calculations can only give rough estimates of the power and necessary sample size for each pair-wise comparison conducted in the present study. 4.3 Determining the cause for phenotypic variability Four major factors were considered to determine whether they had an effect on phenotypic variability in WBS. Examination of deletion extent in 108 cases of clinically diagnosed cases, screened by molecular methods, revealed 85 cases with a deletion. While smaller deletions have been reported to be associated with the presentation of fewer WBS clinical features (Frangiskakis et al., 1996; Tassabehji et al., 1999), no similar small deletions were identified in this study. Identifying these rare cases of WBS 70 with an atypical smaller deletion will be important for narrowing down a critical deletion region required for presentation of the syndrome. The common deletion was found to have a centromeric breakpoint lying proximal to D7S489U and a distal breakpoint lying between D7S1870 and D7S489L, similar to that reported in the literature (Perez Jurado et al., 1996). However, our finding of two cases with a more distal breakpoint that included deletion of D7S489L has not been observed in two other studies in which D7S489L was analyzed (Perez Jurado et al., 1996; Wu et al., 1998). An influence of this larger deletion was not observed in these two individuals suggesting an absence of dosage-sensitive gene(s) or regulatory elements within this region. As the map indicates, D7S489L is one of a series of repeats mapping on either side of the deletion (Figure 2.1). Presence of these repeats leads to the possibility of unequal meiotic exchange resulting in a deletion, as mentioned in the introduction. This has been supported by studies of meiotic exchange which led to the deletion found in WBS individuals. These studies showed that the majority of cases studied have a recombination between the grandparental chromosomes at the site of the deletion that occurred during meiosis in the parent from whom the deleted chromosome originated (Urban et al., 1996; Dutly and Schinzel, 1996; Baumer et al., 1998). The variable inclusion of D7S489L in the deletion found in this study might simply reflect variability in the precise site of unequal exchange, mediated by the series of repeats flanking the common deletion. For instance, unequal exchange between the repetitive sequences PMS2L (proximal to D7S489U) and PMS2LP1 (distal to D7S489L) would result in the inclusion of all sequences within and including D7S489U and D7S489L in the deletion. Screening WBS patients with newly identified polymorphic markers at the ends of the 71 common deletion region and sequencing the breakpoints will be necessary to help determine the precise location of the unequal exchange which occurred and help determine whether these breakpoints are indeed consistent between patients. While some variability in deletion size may still exist, the fact that the deletion region identified in this and previous studies is consistent in the large majority of WBS cases, strongly suggests deletion size does not play a major role in phenotypic variability in clinically typical WBS cases. This is similar to that reported in other microdeletion syndromes such as VCFS where the majority of patients have a 3 Mb deletion and a smaller proportion harbor a 1.5 Mb deletion. There are also rare cases of VCFS having even smaller deletions but no correlation has been identified between phenotypic severity and deletion size in these cases (Carlson et al., 1997). Additionally, studies of PWS and AS cases arising from deletions has also revealed that two common proximal breakpoints exist in both paternal and maternal origin of deletions that occur with similar frequencies (Knoll et al., 1990; Christian et al., 1995). There have been no reports that significant clinical differences exist between the two deletion sizes. Furthermore, initial screening of SMS patients has revealed no apparent correlation between the deletion size at 17pl 1.2 and presentation of a typical phenotype (Elsea et al., 1997). Therefore, it might be said that deletion size does not play a major role in phenotypic variability in microdeletion syndromes so long as the minimal critical region is deleted. For instance, the distal breakpoint of the 17pl3.3 deletion resulting in MDS always lies telomeric to LIS1. But distal deletion breakpoints that lie within LIS1 have only been reported to be associated with cases of ILS (Chong et al., 1997). Looking at parental effects as a cause of phenotypic variability, a previous study has suggested paternally imprinted genes may exist in the WBS deletion region (Perez Jurado et al., 1996). Specifically, WBS patients with maternal deletions were more likely to present with microcephaly (p=7.0xl0"4, n=22) and growth retardation (p=7.0xl0~3, n=29). However, the level of significance was not corrected for multiple comparisons. A total of 24 characteristics were correlated with parental origin, requiring a significance level of 2.0xl0"3. Therefore, only the association between maternal origin and OFC remains significant when using this stricter criterion. A second study has also shown a tendency for retarded growth in maternal deletions but was not, in itself, statistically significant (Brandum-Nielsen et al., 1997). A more recent study of 51 WBS cases, however, found no evidence of parental effects (Wu et al., 1998). Results from the present study also do not reveal any significant correlations between parental origin and clinical outcome. Moreover, the presence of a small OFC was more commonly found in cases of paternal origin (p=4.4xl0'3). Given that this tendency is towards the opposite parent of origin than that reported by Perez Jurado et al., it is likely that the previously reported correlation was due to chance. The authors also suggested that the maternal deletions could be larger than the paternal deletions in their study, thereby affecting additional gene(s) involved in growth control. Although this is unlikely since the deletion is probably the consequence of unequal meiotic exchange between flanking repeats, a parental origin-specific deletion size difference cannot be ruled out entirely since the size estimate of the deletion is still approximate. In addition, all publications to date including this study, report a maternal and paternal origin of deletion with approximately equal frequency (114 maternal: 93 73 paternal; Kotzot et al., 1995a; Perez Jurado et al., 1996; Brendum-Nielsen et al., 1997; Wu et al., 1998). Parent-specific origin of deletion is observed for microdeletion syndromes where genes within the deletion region are imprinted. For example, a maternal deletion of a paternally imprinted gene at 15qll-ql3 gives rise to AS while a paternal deletion of maternally imprinted genes gives rise to PWS (Nicholls et al., 1989a; Knoll et al., 1989). With WBS, both parental origin of deletion occur with approximately equal frequency and do not result in any statistically significant clinical phenotypic differences. This would then argue against imprinting in the WBS deletion region, as one would expect differential phenotypes or subtle differences in growth restriction, based on parental origin of deletion if there were imprinted genes in the region. It is therefore unlikely that there are imprinted genes within the WBS common deletion region. Furthermore, phenotypic variability is most probably not influenced significantly by parental origin of deletion in WBS. Comparing phenotypic outcome in our sample of WBS patients with ELN polymorphisms revealed several associations. It is intriguing that cardiovascular defects were amongst the stronger associations found with ELN polymorphisms in the deletion cases of this study as point mutations within, as well as microdeletions including, ELN have been shown to result in isolated cases of SVAS (Olson et al., 1995; Li et al., 1997; Tassabehji et al., 1997a). Cardiovascular abnormalities in WBS are presumably due to haploinsufficiency of ELN, but are present in only 75% of WBS cases. It is possible that expression of the intact ELN copy has a modifying effect. I therefore hypothesized that polymorphisms within ELN may themselves alter or be in linkage disequilibrium with mutations that result in subtle differences in ELN expression. Because of sample size in 74 each allele class, I was forced to lump together alleles into "small" and "large" which could obscure an effect if only one or two specific alleles were associated. A potential association was also found between ELNil8 and low birth weight. There are, to my knowledge, no known c/s-acting factors that reside within introns 1 or 18 of ELN that would alter expression patterns, although intron 18 bears high homology with Alu-like sequences (Olson et al., 1995; Tassabehji et al., 1997a). Intronic Alu sequences are known to affect expression of some genes by splice-mediated incorporation into the mRNA (Makalowski et al., 1994). Incorporation of Alu sequences can introduce a premature stop codon or create a novel sequence, thereby resulting in a mutant ELN protein (Makalowski et al., 1994). Should the intact copy of ELN be mutated or be prematurely truncated in WBS cases, presentation of SVAS or even lethality would most certainly be expected (Curran et al., 1993; Morris et al., 1993; Ewart et al., 1994; Olson et al., 1995). Therefore, it is unlikely the intron 18 Alu sequence plays a role in producing aberrant ELN transcripts in the WBS patients. It is also unlikely that the association identified between ELNil8 and low birth weight is a direct consequence of ELN. This is because mutations and decreased levels of ELN have only been associated with presentation of cardiovascular malformations and cutis laxa, a rare group of disorders characterized by lax inelastic skin (Beighton, 1979; Tassabehji et al., 1998). However, ELN may still theoretically be responsible for poor growth indirectly as mutated or decreased ELN could result in presentation of a cardiac malformation, which in turn, could lead to poorer growth. Therefore, a correlation of poor postnatal growth and a cardiac malformation might be expected and was indeed observed in this study. The association between ELNil8 and low birth weight could also be explained as a result 75 of linkage disequilibrium between ELNil8 and mutations affecting expression levels of some yet-to-be-identified growth factor. Correlation of the Helg 15/16 polymorphism with clinical outcome did not reveal significant correlations in a previous study by Perez Jurado et al. (1996). The authors suggested this was an indication that the protein polymorphism neither imparts any functional differences nor is in linkage disequilibrium with a polymorphism affecting WBS phenotype. While the present study also did not find any statistically significant correlations with the Helg 15/16 polymorphism using our strict criterion, a negative correlation with presentation of a typical WBS personality was identified. This is particularly interesting because a negative correlation was also identified between a typical WBS personality and the ELNil8 and Helg 15/16 "7 - 2" haplotype. Since ELN is expressed only at basal levels in the brain (Frangiskakis et al., 1996), it seems very unlikely that presentation of the typical personality is a direct consequence of these ELN polymorphisms. It is also unlikely that the haplotype is in linkage disequilibrium with a mutation that results in decreased ELN levels, resulting in the typical personality. It is more probable that the negative correlation observed is due to linkage disequilibrium between the non-"7 - 2" haplotype and a mutation in either LIMK1 or another gene expressed in the brain that produces the distinct personality. Alternatively, the WBS personality may be caused by haploinsufficiency but the "7 - 2" haplotype may be associated with some allele that inhibits the manifestation of the trait. No crossovers were observed in this study between D7S489U and LIMK1-GT in the CEPH reference pedigrees. Low recombination between the two loci of interest is a requirement for linkage disequilibrium and therefore supports the possibility of linkage disequilibrium 76 between the "7 - 2" haplotype with an allele that affects behavior . This association between the "7 - 2" haplotype and a seemingly unrelated clinical feature is comparable to the associations observed between certain human leukocyte antigen (HLA) haplotypes and presentation of specific diseases where the antigens do not actually play a role in the pathology of the disease (Thomson, 1995). The lack of statistical significance between ELN polymorphisms and phenotype in this study, particularly between ELNil and presentation of a cardiovascular malformation, does not exclude that meaningful variability in ELN expression exists due to the lack of statistical power, as described in the previous section 4.2. Analysis of further data sets with attention to same trends would be necessary to exclude this possibility as well as to confirm identified potential associations. The fact that these subsequent studies will only need to examine the 3 potential associations identified between the inherited ELN allele and presence of a clinical feature in the present study (with p<0.05), will reduce the necessary significance level to a'=0.017. Therefore, a sample of 56 individuals is required to detect a positive association with a probability of a type II error (P) of 0.10, compared to 160 individuals being needed in the present study with a'^^xlO"4 and P =0.10. Another approach would be to directly examine synthesis of ELN protein in vivo to see if it shows significant inter-individual variation in expression. LTMK1 is responsible for inactivating cofilin by phosphorylation (Arber et al., 1998; Yang et al, 1998). Cofilin is required for regulation of the actin cytoskeleton by depolymerizing actin, and thereby allowing the actin to be recycled (reviewed by Rosenblatt and Mitchison, 1998). This is a necessary function for cell movement such as 77 the growth of neurons. Haploinsufficiency or point mutations of LIMK1 would presumably allow buildup of cofilin and, therefore, retard actin recycling. This, in turn, could restrict neuronal growth during development. Frangiskakis et al. (1996) first identified LIMK1 hemizygosity in a family of patients presenting with SVAS and the WBS cognitive profile but lacking other WBS traits. Affected family members were shown to carry an 83 kb deletion that encompassed LTMK1 and ELN which suggested that LIMK1 is dosage-sensitive and contributes to visuospatial cognitive perception. However, another research group has found the contrary, identifying two cases with isolated SVAS that are hemizygous for LIMK1 but do not show the WBS cognitive profile (Tassabehji et al., 1997b; Tassabehji et al., 1999). Due to its implication in the WBS cognitive profile and since LIMK1 is expressed primarily in the brain, LIMKl-GT was correlated only with presence of WBS personality traits in this study. The absence of any statistically significant associations in both WBS cases with and without a 7q 11.23 deletion could indicate: 1) LIMK1 does not have an effect dn personality characteristics (as defined in Appendix IT); 2) lack of statistical power to detect a correlation; 3) inaccuracy in the assessment of these traits due to their subjective nature; 4) LIMKl-GT is not in linkage disequilibrium with a phenotype-conveying polymorphism; or 5) LIMKl-GT is not correlated with expression levels of LIMK1, assuming variability of expression results in phenotypic variability. With the cases identified in this study, a total of 83 males and 70 females with a confirmed deletion (n.s. different from 1:1) have been reported in the literature (Elcioglu et al., 1998; Wu et al., 1998). While not significant with our strict criteria, males were noted to present with cardiovascular defects, a typical facies, and nail dysplasia 78 (p=6.0xl0"3) more commonly than females. Such tendencies have not been observed in two other studies (Perez Jurado et al., 1996; Wu et al., 1998) although Perez Jurado et al. (1996) did observe a slight tendency for females to be more growth retarded. The tendency for presentation of the above three phenotypic features by males was not duplicated in the sample of non-deletion cases in this study. While it is possible these observations could suggest males have a more "typical" appearance for WBS than females, it is more likely that the observations made in this study are artifacts of sampling or chance deviations in frequencies. Therefore, no differences in prognosis can be predicted based on gender of the affected individual. However, this will need to be verified in a subsequent study comparing gender with presence of these 3 clinical features. With only 3 comparisons necessary, the required significance level is a'=0.017. Assuming the observed frequencies of the traits are reflective of the true population frequencies, a sample size of 148 individuals (74 each of males and females) would be necessary to determine whether an association exists with p=0.10. This is substantially more powerful test compared to the present study which requires 338 WBS cases (169 each of males and females) with a'=4.6xl0"4 to detect a true association. 4.4 Clinical correlations A number of clinical features were found to be correlated with one another in WBS individuals with a deletion at 7qll.23. Those significantly correlated after correction for number of comparisons, as well as others with p<0.05, are discussed in terms of the common underlying cause. 4.4.1 Significant correlation between low birth weight and poor postnatal weight gain: evidence for deletion of a growth factor gene? Both low birth weight (<!()* centile) and low postnatal weight ( l^O* centile) have been reported to be variably presenting features of WBS. Frequency of presentation range from 37% (n=24; Brondum-Nielsen et al., 1997) to over 70% of cases (n=42; Morris et al., 1988) for low birth weight and approximately 50% of cases for poor weight gain postnatally (Borg et al., 1995; Perez Jurado et al., 1996). Co-occurrence of low birth weight and poor postnatal weight gain was found in this study. While this phenomenon may not be surprising even in a normal population, the correlation could also represent the presence of a yet-to-be-identified dosage-sensitive gene within the common WBS deletion that contributes to both pre- and postnatal growth. Hemizygosity of this gene could reveal the effects of an allele that conveys poor weight gain. Alternatively, the deleted growth gene may not be necessarily polymorphic. Instead, the variability in presentation of growth restriction in WBS may be accounted for by the fact that growth is a multifactorial trait. Therefore, the influence of environmental factors such as nutrition and genetic background are certain to play a role in growth (Ayala and Kiger, 1984). Extensive efforts by other research groups in cloning genes within the WBS commonly deleted region may still reveal this putative growth factor gene. It is interesting to note that a condition characterized by growth retardation, Silver-Russell syndrome (SRS), is associated with maternal uniparental disomy of chromosome 7 (UPD 7) in 10% of cases (Kotzot et al., 1995b; Eggermann et al., 1997; Bernard et al., submitted). This association between UPD 7 and an abnormal phenotype could suggest the presence of one or more imprinted genes on chromosome 7 with 80 biparental inheritance of chromosome 7 required for a normal phenotype. The growth genes for SRS and WBS are likely to be different for two reasons. Firstly, a candidate gene for SRS, PEG1/MEST, has been mapped to 7q31-34 near D7S649, outside the WBS common deletion (Kobayashi et al., 1997). Secondly, this candidate gene shows paternal expression in early embryonic development, which is consistent with the observation of maternal UPD 7 association with SRS, but there is no evidence for imprinting in WBS. 4.4.2 Correlations with a stellate iris pattern A stellate iris pattern was found to be associated with blue irides and with transient infantile hypercalcaemia, with only the latter being significant after correction for multiple comparisons. The finding of blue irides in 43% of WBS cases in this study is somewhat lower than previous studies which have reported frequencies of 51 - 79% (Holmstrom et al., 1990; Winter et al., 1996). The lower frequency found in this study could potentially be due to differences in ethnic background of the samples used in each study. While data on ethnicity were not reported in the previous reports nor recorded in the data collection of this study, the previous studies were based in England and Germany, respectively, where the incidence of blue irides might be expected to be higher. Stellate irides were detected in 47% of cases in the present study. This frequency in deletion confirmed cases ranged from 14 - 67% in previous (Elcioglu et al., 1998; Perez Jurado et al., 1996). Such a variability in the frequency of stellate irides may be attributable to the clinician's degree of experience in detecting the feature as the pattern is more difficult to detect in dark irides (Holmstrom et al., 1990; Winter et al., 1996). Winter et al. suggested using slit-lamp microscopy to analyze dark irides as they reported 81 greater success in detecting the stellate pattern. This difficulty in observing the stellate pattern in dark irides would likely account for the association between blue and stellate patterned irides detected. Previous studies of the eyes of WBS patients were designed to determine whether hypercalcaemia manifested as any ocular features (Greenberg and Lewis, 1988; Winter et al., 1996). To do so, the authors looked for presentation of unusual traits, such as calcium deposits in the conjunctiva, calcific band keratopathy of the cornea, and punctate opacities in the lens or retina in hypercalcaemic patients. None of these ocular features was found in association with hypercalcaemia in WBS. Review of patient files in the present study also did not reveal any unusual ocular manifestations. However, conjunctival biopsies were not performed in any of the patients to confirm the absence of calcium deposits. While it has not been looked for in previous studies, it is interesting that an association between transient infantile hypercalcaemia and presence of stellate irides was found to be highly significant in this study. Stellate irides have been described in a number of different manners in the past. These include "stromal hypoplasia" (Morris et al., 1988), "lateral displacement or absence of the iris collarette" (Greenberg and Lewis, 1988), and "abnormality anterior to the iris" (Jones and Smith, 1975). The exact biology of this feature, however, is not known. A review of the literature on iris development has not revealed any obvious involvement of calcium. Further investigation of stellate iris biology and development will be necessary to determine the significance of this potential association. Interestingly, anterior iris hypoplasia has also been reported in a patient with Menkes disease (Ferreira et al., 1998). Menkes syndrome is an X-linked recessive disorder that results in 82 progressive neurodegeneration caused by a mutation in a copper-binding P-type ATPase gene (Dierick et al., 1995). This results in copper transport failure and subsequently, decreases activity of many copper-dependent enzymes. Also, formation of some structural proteins is inhibited, so an effect on elastin is possible (reviewed by Vulpe and Packman, 1995). It is tempting to speculate that manifestation of stellate irides in WBS patients is the result of both elastin haploinsufficiency and infantile hypercalcaemia. WBS cases presenting with stellate irides but described as normocalcaemic could be accounted for by the transient nature of hypercalcaemia in WBS. 4.4.3 Other associations: renal/urogenital and cardiovascular abnormalities The co-occurrence of renal or urogenital abnormalities and cardiovascular defects has not been looked for in previous studies of WBS. While this correlation is not significant using our stringent criteria, it is still worthy of attention as these abnormalities represent the more severe clinical manifestations of WBS. It is also interesting as congenital cardiac malformations are the most common malformations associated with congenital solitary kidney abnormalities (Emanuel et al., 1974). For instance, individuals with unilateral renal agenesis, a feature found in 13% of WBS cases with a renal or urogenital abnormality (Pankau et al., 1996), have an incidence of cardiovascular malformations 20 times greater than expected by chance alone (Robson et al., 1995). Overall, renal abnormalities have been reported in 4% (Perez Jurado et al., 1996; Brendum-Nielsen et al., 1997) to over 20% (Borg et al, 1995; Elcioglu et al., 1998) of WBS cases with a proven 7qll.23 deletion. In our study, 12 individuals (22%) with a 7ql 1.23 deletion had a renal or urogenital abnormality, all of whom also had a congenital 83 cardiac malformation (Table 4.2). The fact that all 12 patients presented with a cardiovascular defect could suggest that renal abnormality is characteristic of the more extreme phenotype in WBS. x Two of the three cases with calcium deposits in the kidney also had documented hypercalcaemia, while calcium levels in the third child were not looked at. All renal and urogenital abnormalities found in our sample have also been reported in previous studies. There was no particular renal abnormality that occurs with significant frequency that was characteristic of WBS in our study, a finding consistent with previous studies (Pankau et al., 1996). 84 Table 4.2: Renal/urogenital and cardiovascular abnormalities observed in 12 WBS patients with positive correlation of traits. Patient Renal/urogenital abnormality Cardiovascular malformation 20 Inguinal hernia PPSa, VSDa 25 Not provided by collaborator VSD 50 Not provided by collaborator SVASa z82 Not provided by collaborator SVAS, PPS z84 Not provided by collaborator Mild aortic stenosis v96 Retractile testes Mild SVAS v97 Renal dysplasia PPS vl06 Bladder diverticula PPS, VSD vllO Hydronephrosis, cystic calcification SVAS vll2 Nephrocalcinosis VSD vl23 Nephrocalcinosis Mild PPS vl25 Unilateral renal dysplasia PPS aPPS: Perip Ventricular leral pulmonary stenosis; SVAS: Supravalvular aortic stenosis; VSD: septum defect. The underlying cause of renal abnormalities in WBS is not known at this time. While the deletion of ELN is known to cause the cardiovascular defects seen in WBS, neither mutations nor a deletion of ELN has been found to be associated with renal abnormalities despite the fact ELN is expressed in kidney. This could suggest another gene(s) is responsible for the presentation of renal abnormalities that could have a modifying effect on cardiovascular structure. Genetic background is likely also to affect presentation of both traits, and therefore complete concordance would not be observed. However, there have been two reports of monozygotic twins with WBS discordant for renal or urogenital abnormalities (Pankau et al., 1993; Castorina et al., 1997). This would 85 seem to indicate that environmental factors or chance also play a role in presentation of renal/urogenital abnormalities. 4.5 Deletion versus non-deletion WBS cases: discussing heterogeneity and phenocopies While 85 cases in our study carry a deletion of 7qll.23, all patients contributed from Zurich, Athens, and Winnipeg were confirmed for a deletion by local screening before they were sent for further analysis in this study. Of those cases tested by us, 20/43 (47%) cases from BC and Saskatchewan were found to carry a deletion. Overall, a deletion has been found in approximately 90% of all clinically typical WBS cases reported in the literature to date, suggesting WBS is genetically heterogenous. This lower deletion frequency in the present study is easily accounted for by the clinical criteria for inclusion of the patients. For instance, Joyce et al. (1996) divided test subjects into two groups with a clinical diagnosis that was either "classical WBS" or "suspected WBS". Screening for a deletion by FISH revealed 22/23 (96%) cases carried a deletion in the former of the two groups while only 2/22 (9%) cases carried a deletion in the latter. The criteria in the present study were less stringent and included both clinical diagnoses of "WBS" and "suspected WBS". This would suggest that our 23 cases without a deletion are likely a heterogeneous population consisting of both WBS cases due to a common underlying 7q 11.23 defect, and genocopies or phenocopies, i.e. phenotypically similar conditions that have a different genetic or non-genetic basis. Our identification of deletion cases presenting with cardiovascular defects significantly more often than our non-deletion cases would seem to support the non-deletion cases are different from deletion cases of WBS. It has been argued that the large majority of clinically typical WBS cases have some mutation of 7qll.23 (Mari et al., 1995). The authors suggested patients without an identifiable deletion would be one of two possibilities: Firstly, non-deletion cases could have a smaller cryptic deletion within the common deletion region. Such a small cryptic deletion could affect multiple genes within the deletion region if it encompasses a common regulatory element. Point mutations of the common regulatory element could also silence multiple genes within the WBS critical region. For instance, both point mutations and deletions of 10 - 100 bases of a putative imprinting control (IC) element for PWS have been found in patients without the typical 4 Mb 15q deletion but are typical PWS (Sutcliffe et al., 1994; Buiting et al., 1995; Saitoh et al., 1996). Similar deletions clustering further upstream from the PWS IC region have defined a putative IC region for AS (Buiting et al., 1995; Saitoh et al., 1996) despite that the UBE3A gene implicated in AS is located as much as 1 Mb downstream (Jiang et al., 1998). Thus, it is can be argued that non-deletion cases of WBS could still be "true" WBS (genetically due to an aberration at 7qll.23 with either point mutations or deletions of a common regulatory element, or a deletion of the common WBS deletion region). This will require screening of clinically typical non-deletion cases for decreased expression of ELN and other genes within the WBS common deletion region, relative to normal controls. Patients identified with decreased expression can then be screened for smaller microdeletions within the typical WBS deletion region. These atypical microdeletions 87 may define a regulatory element region that can be screened in other non-deletion cases for point mutations once the region has been cloned and sequenced. The remaining non-deletion cases without any 7q 11.23 aberration comprise the second possibility. These cases may represent phenocopies or genocopies of WBS that would require further investigation to determine their etiology. This could be accomplished first by a linkage study of familial cases of SVAS where no ELN deletions or mutations have been identified. Any identified locus could then be screened for mutations in these WBS cases without a 7ql 1.23 mutation. It is possibility that some of these WBS patients without any 7qll.23 aberration do not fully satisfy the clinical criteria for WBS. To support this suggestion, the Mari et al. (1995) pointed to 6 individuals in their study who were initially diagnosed with WBS but tested negative for a deletion by FISH. These 6 individuals were subsequently re-evaluated by other physicians and a clinical diagnosis of WBS could not be confirmed. Therefore, it might be said that the non-deletion cases identified could be a heterogeneous group comprised of both "true" WBS patients, as defined previously, and also individuals with similar phenotypic features but owing to a different genetic or environmental etiology. 4.6 Conclusions This study has looked at 108 clinically diagnosed cases of WBS for the purpose of identifying molecular and clinical correlations. Molecular screening identified 85 cases (79%) with and 23 cases (21%) without a 7qll.23 deletion that were significantly different from each other as deletion cases present with cardiovascular malformations more often than cases without a deletion. Cases with a deletion were also more likely to 88 have the typical wide mouth and extreme sociability. An odds ratio of 16.6 (99% confidence interval 1.39 - 197.88) was calculated when all three features were considered simultaneously. This strongly suggests that an individual with a clinical diagnosis of WBS who presents with all three features, is likely to harbor a deletion at 7qll.23. The cases without a deletion are probably a heterogeneous population consisting of WBS with either undetectable deletions or mutations of a common regulatory element, phenocopies, or genetically distinct disorder with a phenotype similar to that of WBS. The use of WBS as a model system for understanding the underlying cause for clinical variability in microdeletion syndromes revealed several important facts. First of all, a good indication that unequal recombination is the primary mechanism to deletion formation is the presence of repetitive elements flanking the common deletion region. This has been observed in at least 3 other microdeletion syndromes, PWS/AS (Buiting et al., 1998; Ji et al., 1999), SMS (Chen et al., 1997), and HNPP (Patel et al., 1992; Pentao et al., 1992; Lupski et al., 1991), and 1 microduplication syndrome, CMT1A (Patel et al., 1992; Pentao et al., 1992; Lupski et al., 1991). Where there is evidence for unequal recombinant exchange between flanking repeats as the primary mechanism for deletion formation, variability in size of the deletion is extremely rare. A total of 83 cases out of 85 with a deletion were identified to carry a deletion that corresponded in size with that reported in 2 previous studies (Perez Jurado et al., 1996; Wu et al., 1998). With very little variability in deletion size, there is no evidence that differences in deletion size play a major role in clinical variability in microdeletion syndromes formed primarily by unequal recombination. This study also does not find evidence for imprinting in WBS as parental origin of the deletion does not appear to play a significant role in phenotypic variability. This is especially important as a previous report (Perez Jurado et al., 1996) has suggested imprinted gene(s) may exist within the WBS deletion region. Data from this study also provide some evidence that ELN and LIMK1 polymorphisms may affect expression levels of dosage-sensitive genes. This may be the consequence of either the polymorphisms directly or the polymorphisms being in linkage disequilibrium with another mutation that affects expression level. The use of strict statistical criteria, coupled with a low sample size, resulted in low statistical power and subsequently, an absence of statistically significant associations. But the notable potential association detected between an ELN polymorphism and presentation of cardiovascular malformations is of great importance for two reasons. Firstly, ELN is known to be associated with cardiovascular defects. Therefore, this could potentially be extrapolated to encompass other deletion syndromes where a correlation between a clinical feature and a polymorphism within a deleted gene could indicate the involvement of that gene with the etiology of the clinical trait. Secondly, this could suggest that correlations between polymorphisms of deleted genes and a clinical trait in other deletion syndromes are the result of that polymorphism also being associated with decreased expression levels of the gene in question. I can hypothesize, based on these results, that genetic background must play a significant role in phenotypic variability in WBS and presumably in other microdeletions. WBS is not an ideal choice as a model to find secondary modifying loci as WBS is predominantly a sporadic disorder and family studies are necessary to find modifying 90 loci. One might consider creating mouse knockouts where deletion size can be manipulated. The creation of several lines with different deletion sizes could reveal the effects of hemizygosity for each individual gene in the common deletion region and also help to better understand the etiology of each clinical trait. However, there may not be homology between humans and mice for the phenotypic outcome due to the genetic background. For instance, a deletion of PAX3 in humans results in Waardenburg syndrome type I, characterized by dystopia canthorum, pigmentary abnormalities of the skin, hair, and eyes, and hearing loss. However, loss of this gene in mice results only in a white belly splotch (Read et al., 1997). Even more complex, is trying to decipher the intricacies of the complex WBS behavioral and cognitive profile as mouse models will be of limited use in understanding this aspect of WBS. Instead, the ascertainment and study of WBS individuals with smaller deletions will be necessary. Thus, WBS provides a good model of how to study complex cognition and behavior in a microdeletion syndrome. Furthermore, pair-wise comparisons of clinical features of WBS indicated that this method is valuable for identifying clinical features that may have a common underlying etiology. A number of correlations were identified of which, two were significant even after statistical correction for multiple comparisons: 1) low birth weight and continued poor weight gain postnatally; and 2) co-presence of hypercalcaemia and a stellate iris pattern. Other associations identified that were not significant after Bonferroni's correction include: 1) presence of a stellate and blue iris; and 2) presence of renal abnormalities and cardiovascular malformations. Such correlations provide a framework with which, follow-up investigations can focus on and could provide invaluable data regarding the biology and pathology of specific clinical traits that manifest in a microdeletion syndrome. It should be emphasized that the large number of comparisons made in this study and the low a' required for each individual test make it extremely difficult to identify all true associations. As a result, the many suggestive associations (i.e. 0.05>p>a') identified in this study are therefore of interest to confirm in a follow-up study. Sample sizes of approximately 50 to 320, depending on which factors are being compared (Appendix IV), will be required if the study-wide a=0.05 and 3=0.10. Such follow-up studies with improved statistical power will clarify the true associations in WBS. 92 CHAPTER 5: References Arber S, Barbayannis FA, Hanser H, Schneider C, Stanyon CA, Bernard O, and Caroni P (1998). Regulation of actin dynamics through phosphorylation of cofilin by LIM-kinase. Nature 393: 805-809. Asher JH Jr, Harrison RW, Morell R, Carey ML, Friedman TB (1996). Effects of Pax3 modifier genes on craniofacial morphology, pigmentation, and viability: a murine model of Waardenburg syndrome variation. Genomics 34: 285 - 298. i Ayala FJ and Kiger JA (1984). Modern Genetics, 2nd Ed. Benjamin/Cummings Publishing, Menlo Park. Baumer A, Dutly F, Balmer D, Riegel M, Tukel T, Krajewska-Walasek M, and Schinzel AA (1998). High level of unequal meiotic crossovers at the origin of the 22qll.2 & 7qll.23 deletions. Hum. Mol. Genet. 7: 887 - 894. Beighton P (1979). Cutis laxa. Birth Defects Compendium, 2nd Ed. Alan R Liss, New York. 93 Bellugi U, Bihrle A, Jernigan T, Trauner D, and Doherty S (1990). Neuropsychological, neurological, and neuroanatomical profile of Williams syndrome. Am. J. Med. Genet. 6: 115-125. Bernard LE, Penaherrera MS, Van Allen MI, Wang MS, Yong S-L, Gareis F, Langlois S, and Robinson WP (submitted). Clinical and molecular findings in two patients with Russell-Silver syndrome and UPD7: comparison with non-UPD7 cases. Beuren AJ, Apitz J, and Harmjanz D (1962). Supravalvular aortic stenosis in association with mental retardation and certain facial appearance. Circulation 26: 1235 - 1246. Blum H, Beier H, and Gross HJ (1987). Improved silver staining of plant proteins RNA and DNA in polyacrylamide gels. Electrophoresis 8: 93 - 99. Borg I, Delhanty JD, and Baraitser M (1995). Detection of hemizygosity at elastin by FISH analysis as a diagnostic test in both classical and atypical cases of Williams syndrome. J. Med. Genet. 32: 692 - 696. Brinkmann G, Heller M, Partsch C-J, Gosch A, and Pankau R (1997). Magnetic resonance imaging of the brain in Williams-Beuren syndrome. Am. J. Med. Genet. 68: 243. 94 Brondum-Nielsen K, Beck B, Gyftodimou J, Horlyk H, Lijenberg U, Peterson MB, Pederson W, Peterson MB, Sand A, Skovby F, Stafanger G, Zetterqvist P, and Tommerup N (1997). Investigation of deletions at 7q 11.23 in 44 patients referred for Williams-Beuren syndrome using FISH and four DNA polymorphisms. Hum. Genet. 99: 56-61. Budarf ML and Emanuel BS (1997). Progress in the autosomal segmental aneusomy syndromes (SASs): single or multi-locus disorders? Hum. Mol. Genet. 6: 1657 - 1665. Buiting K, Saitoh S, Gross S, Bittrich B, Schwartz S, Nicholls RD, Horsthemke B (1995). Inherited microdeletions in the Angelman and Prader-Willi syndromes define an imprinting centre on human chromosome 15. Nat. Genet. 9: 395 - 400. Buiting K, Gross S, Ji Y, Senger G, Nicholls RD, and Horsthemke B (1998). Expressed copies of the MN7 (D15 S3 7) gene family map close to the common deletion breakpoints in the Prader-Willi/Angelman syndromes. Cytogenet. Cell Genet. 81: 247 - 253. Burn J (1986). Williams syndrome. J. Med. Genet. 23: 389 - 395. Butler MG and Palmer CG (1983). Parental origin of chromosome 15 deletion in Prader-Willi syndrome. Lancet 1: 1285 - 1286. 95 Butterworth CE Jr and Bendich A (1996). Folic acid and the prevention of birth defects. Annu. Rev. Nutr. 16: 73 - 97. Cantu ES, Eicher DJ, Pai GS, Donahue CJ, Harley RA (1996). Mosaic vs. nonmosaic trisomy 9: report of a liveborn infant evaluated by fluorescence in situ hybridization and review of the literature. Am. J. Med. Genet. 62:330-335. Carlson C, Sirotkin H, Pandita R, Goldberg R, McKie J, Wadey R, Patanjali SR, Weissman SM, Anyane-Yeboa K, WarburtonD, Scambler P, ShprintzenR, Kucherlapati R, and Morrow BE (1997). Molecular definition of 22qll deletions in 151 Velo-Cardio-Facial syndrome patients. Am. J. Hum. Genet. 61: 620 - 629. Castorina P, Selicorni A, Bedeschi F, Dalpra L, and Larizza L. (1997). Genotype-phenotype correlation in 2 sets of monozygotic twins with Williams syndrome. Am. J. Med Genet.69: 107-111. Chance P, Abbas N, Lensch M, Pentao L, Roa B, Patel P, and Lupski J (1994). Two autosomal dominant neuropathies result from reciprocal DNA duplication/deletion of a region on chromosome 17. Hum. Mol. Genet. 3: 223 - 228. Chen K-S, Manian P, Koeuth T, Potocki L, Zhao Q, Chinault AC, Lee CC, and Lupski JR (1997). Homologous recombination of a flanking repeat gene cluster is a mechanism, for a common contiguous gene deletion syndrome. Nat. Genet. 17: 154- 163. 96 Chong SS, Pack SD, Roschke AV, Tanigami A, Carrozzo R, Smith ACM, Dobyns WB, and Ledbetter DH (1997). A revision of the lissencephaly and Miller-Dieker syndrome critical regions in chromosome 17pl3.3. Hum. Mol. Genet. 6: 147 - 155. Christian SL, Robinson WP, Huang B, Mutirangura A, Line MR, Nakao M, Surti U, Chakravarti A, and Ledbetter DH (1995). Molecular characterization of two proximal deletion breakpoint regions in both Prader-Willi and Angelman syndrome patients. Am. J. Hum. Genet. 57: 40 - 48. Curran ME, Atkinson DL, Ewart AK, Morris CA, Leppert MF, and Keating MT (1993). The elastin gene is disrupted by a translocation causing supravalvular aortic stenosis. Cell 73: 159- 168. Devriendt K, Van Hoestenberghe R, Van Hole C, Devlieger H, Gewillig M, Moerman Ph, Van den Berghe H, and Fryns JP (1997). Submicroscopic deletion in chromosome 22ql 1 in trizygous triplet siblings and their father. Clin. Genet. 51: 246 - 249. Dierick HA, Ambrosini L, Spencer J, Glover TW, and Mercer JF (1995). Molecular structure of the Menkes disease gene (ATP7A). Genomics 28: 462 - 469. 97 Dutly F and Schinzel A (1996). Unequal interchromosomal rearrangements may result in elastin gene deletions causing the Williams-Beuren syndrome. Hum. Mol. Genet. 5: 1893 - 1898. Eggermann T, Wollmann HA, Kuner R, Eggermann K, Enders H, Kaiser P, and Ranke MB (1997). Molecular studies in 37 Silver-Russell syndrome patients: frequency and etiology of uniparental disomy. Hum. Genet. 100: 415 - 419. Elcioglu N, Mackie-Ogilvie C, Daker M, and Berry AC (1998). FISH analysis in patients with clinical diagnosis of Williams syndrome. Acta Paediatr. 87: 48-53. Elsea SH, Purandare SM, Adell RA, Juyal RC, Davis JG, Finucane B, Magenis RE, and Patel PI (1997). Definition of the critical interval for Smith-Magenis syndrome. Cytogenet. Cell Genet. 79: 276 - 281. Emanuel B, Nachman R, Aronson N, and Weiss H (1974). Congenital solitary kidney. A review of 74 cases. J. Urol. 111: 394 - 397. Ewart AK, Morris CA, Einsing GJ, Locker J, Moore C, Leppert M, and Keating MT (1993a). A human vascular disorder supravalvular aortic stenosis maps to chromosome 7. Proc. Natl. Acad. Sci. USA 90: 3226 - 3230. 98 Ewart AK, Morris CA, Atkinson D, Jin W, Sternes K, Spallone P, Stock AD, Leppert M, and Keating MT (1993b). Hemizygosity at the elastin locus in a developmental disorder Williams-Beuren syndrome. Nat. Genet. 5: 11-16. Ewart AK, Jin W, Atkinson D, Morris CA, and Keating MT (1994). Supravalvular aortic stenosis associated with a deletion disrupting the elastin gene. J. Clin. Invest. 93: 1071 -1077. Ferreira RC, Heckenlively JR, Menkes JH, and Bateman JB (1998). Menkes disease. New ocular and electroretinographic findings. Ophthalmology 105: 1076 - 1078. Foster K, Ferrell R, King-Underwood L, Povey S, Attwood J, Rennick R, Humphries SE, and Henney AM (1993). Description of a dinucleotide repeat polymorphism in the human elastin gene and its use to confirm assignment of the gene to chromosome 7. Ann. Hum. Genet. 57: 87 - 96. Frangiskakis JM, Ewart AK, Morris CA, Mervis CB, Bertrand J, Roninson BF, Klein BP, Ensing GJ, Everett LA, Green ED, Proschel C, Gutowski NJ, Noble M, Atkinson DL, Odelberg SJ, and Keating MT (1996). LIM-kinase 1 hemizygosity implicated in impaired visuospatial constructive cognition. Cell 86: 59 - 69. Gilbert-Dussardier B, Bonneau D, Gigarel N, Merrer ML, Bonnet D, Philip N, Serville F, Verloes A, Rossi A, Ayme S, Weissenbach J, Mattei MG, Lyonnet S, and Munnich A 99 (1995). A novel microsatellite DNA marker locus D7S1870 detects hemizygosity in 75% of patients with Williams syndrome. Am. J. Hum. Genet. 56: 542 - 544. Greenberg F and Lewis RA (1988). The Williams syndrome. Spectrum and significance of ocular features. Ophthalmology 95: 1608 - 1612. Greenberg F (1993). Contiguous gene syndromes. Growth Genet. Hormones 9: 5 - 10. Hertzberg J, Nakisbendi L, Needleman HL, and Pober B (1994). Williams syndrome-oral presentation of 45 cases. Pediatr. Dent. 16: 262 - 267. Hewett D, Lynch J, Child A, Firth H, and Sykes B (1994). Differential allelic expression of a fibrillin gene (FBN1) in patients with Marfan syndrome. Am. J. Hum. Genet. 55: 447 -52. Holmstrom G, Almond G, Temple K, Taylor D, and Baraitser M (1990). The iris in Williams syndrome. Arch. Dis. Child 65: 987 - 989. Jacobs PA and Melville M (1974). A cytogenetic survey of 11,680 newborn infants. Ann. Hum. Genet. 37: 359-368. Jarvela I, Savukoski M, Ammala P, and von Koskull H (1998). Prenatally detected paternal uniparental chromosome 13 isodisomy. Prenat. Diagn. 18: 1169-1173. 100 Jernigan TL and Bellugi U (1990). Anomalous brain morphology on magnetic resonance images in Williams syndrome and Down syndrome. Arch. Neurol. 47: 529 - 533. Ji Y, Walkowicz MJ, Buiting K, Johnson DK, Tarvin RE, Rinchik EM, Horsthemke B, Stubbs L, and Nicholls RD (1999). The ancestral gene for transcribed, low-copy repeats in the Prader-Willi/Angelman region encodes a large protein implicated in protein trafficking, which is deficient in mice with neuromuscular and spermiogenic abnormalitites. Hum. Molec. Genet. 8: 533 - 542. Jiang Y-H, Tsai T-F, Bressler J, and Beaudet AL (1998). Imprinting in Angelman and Prader-Willi syndromes. Curr. Opin. Genet. Develop. 8: 334 - 342. Jones KL and Smith DW (1975). The Williams' elfin face syndrome. J. Pediatr. 86: 718 -723. Joyce CA, Zorich B, Pike SJ, Barber JCK, and Dennis NR (1996). Williams-Beuren syndrome: phenotypic variability and deletions of chromosomes 7, 11, and 22 in a series of 52 patients. J. Med. Genet. 33: 986 - 992. Juyal RC, Figuera LE, Hauge X, Elsea SH, Lupski JR, Greenberg F, Baldini A, and Patel PI (1996). Molecular analyses of 17pll.2 deletions in 62 Smith-Magenis syndrome patients. Am. J. Hum. Genet. 58: 998 - 1007. 101 Kajiwara K, Berson EL, and Dryja TP (1994). Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci. Science 264: 1604 - 1608. Karmiloff-Smith A, Klima E, Bellugi U, Grant J, and Baron-Cohen S (1995). Is there a social module? Language, face processing, and theory of mind in individuals with Williams syndrome. J. Cognitive Neurosci. 7: 196 - 208. Keating MT (1997). On the trail of genetic culprits in Williams syndrome. Cardiovasc. Res. 36: 134- 137. Kennedy GC, German MS, and Rutter WJ (1995). The minisatellite in the diabetes susceptibility locus IDDM2 regulates insulin transcription. Nat. Genet. 9: 293 - 298. Knoll JH, Nicholls RD, Magenis RE, Graham JM, Lalande M, and Latt SA (1989). Angelman and Prader-Willi syndromes share a common chromosome 15 deletion but differ in parental origin of the deletion. Am. J. Med. Genet. 32: 285 - 290. Knoll JH, Nicholls RD, Magenis RE, Glatt K, Graham JM Jr, Kaplan L, and Lalande M (1990). Angelman syndrome: three molecular classes identified with chromosome 15qllql3-specific DNA markers. Am. J. Hum. Genet. 47: 149 - 155. 102 Kohn B, Kleyman SM, Conte RA, Macera MJ, Glassberg K, and Verma RS (1997). Characterization of an isodicentric Y-chromosome for the long arm in a newborn with mixed gonadal dysgenesis. Ann. Genet. 40: 10-13. Kotzot D, Bernasconi F, Brecevic L, Robinson WP, Kiss P, Kosztolanyi G, Lurie IW, Superti-Furga A, and Schinzel A (1995a). Phenotype of the Williams-Beuren syndrome associated with hemizygosity at the elastin locus. Eur. J. Pediatr. 154: 477 - 482. Kotzot D, Schmitt S, Bernasconi F, Robinson WP, Lurie IW, Ilyina H, Mehes K, Hamel BC, Often BJ, Hergersberg M, Werder E, Schoenie E, and Schinzel A (1995b). Uniparental disomy 7 in Silver-Russell syndrome and primordial growth retardation. Hum. Mol. Genet. 4: 583 - 587. Kuwano A, Ledbetter SA, Dobyns WB, Emanuel BS, and Ledbetter DH (1991). Detection of deletions and cryptic translocations in Miller-Dieker syndrome by in situ hybridization. Am. J. Hum. Genet. 49: 707 - 714. Laitinen J, Samarut J, and Holtta E (1994). A nontoxic and versatile protein salting-out method for isolation of DNA. Biotechniques 7: 316, 318, 320 - 322. Ledbetter SA, Kuwano A, Dobyns WB, and Ledbetter DH (1992). Mcrodeletions of chromosome 17pl3 as a cause of isolated lissencephaly. Am. J. Hum. Genet. 50: 182 -189. 103 Lee ST, Nicholls RD, Schnur RE, Guida LC, Lu-Kuo J, Spinner NB, Zackai EH, and Spritz RA (1994a). Diverse mutations of the P gene among African-Americans with type II (tyrosinase-positive) oculocutaneous albinism (OCA2). Hum. Molec. Genet. 3: 2047 -2051. Lee ST, Nicholls RD, Bundey S, Laxova R, Musarella M, and Spritz RA (1994b). Mutations of the P gene in oculocutaneous albinism, ocular albinism, and Prader-Willi syndrome plus albinism. New Engl. J. Med. 330: 529 - 534. Li DY, Toland AE, Boak BB, Atkinson DL, Ensing GJ, Morris CA, and Keating MT (1997). Elastin point mutations cause an obstructive vascular disease supravalvular aortic stenosis. Hum. Mol. Genet. 6: 1021 - 1028. Lo Nigro C, Chong CS, Smith AC, Dobyns WB, Carrozzo R, and Ledbetter DH (1997). Point mutations and an intragenic deletion in LIS1 the lissencephaly causative gene in isolated lissencephaly sequence and Miller-Dieker syndrome. Hum. Mol. Genet. 6: 157 — 164. Lopez-Rangel E, Maurice M, McGillivray B, and Friedman JM (1992). Williams syndrome in adults. Am. J. Med. Genet. 44: 720 - 729. 104 Loughlin J, Irven C, Athanasou N, Carr A, and Sykes B (1995). Differential allelic expression of the type II collagen gene (COL2A1) in osteoarthritic cartilage. Am. J. Hum. Genet. 56: 1186-93. Lowery MC, Morris CA, Ewart A, Brothman LJ, Zhu XL, Leonard CO, Carey JC, Keating M, and Brothman AR (1995). Strong correlation of Elastin deletions, detected by FISH, with Williams syndrome: evaluation of 235 patients. Am. J. Hum. Genet. 57: 49 -53. Lupski JR, de Oca-Luna RM, Slaugenhaupt S, Pentao L, Guzzetta V, Trask BJ, Saucedo-Cardenas O, Barker DF, Killian JM, Garcia CA, Chakravarti A, and Patel PI (1991). DNA duplication associated with Charcot-Marie-Tooth disease type 1A. Cell 66: 219 -232. Makalowski W, Mitchell GA, and Labuda D (1994). Alu sequences in the coding regions of mRNA: a source of protein variability. Trends Genet. 10: 188 - 193. Malcolm S, Clayton-Smith J, Nicholls M, Robb S, Webb T, Armour JA, Jeffreys AJ, and Pembrey ME (1991). Uniparental paternal disomy in Angelman's syndrome. Lancet 337: 694 - 697. 105 Mari A, Mingarelli R, Giannotti A, Sebastio G, Colloridi V, Novelli G, and Dallapicolla B (1995). Analysis of the elastin gene in 60 patients with clinical diagnosis of Williams syndrome. Hum. Genet. 96: 444 - 448. Mari A, Amati F, Conti E, Bengala M, Novelli G, and Dallapiccola B (1998). A highly polymorphic CA/GT repeat (LIMK1GT) within the Williams syndrome critical region. Clin. Genet. 53: 226-227. McGrath J and Solter D (1984). Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 37: 179- 183. Moore T and Haig D (1991). Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 7: 45 - 49. Morris CA, Demsey SA, Leonard CO, Dilts C, and Blackburn BL (1988). Natural history of Williams syndrome: physical characteristics. J. Pediatr. 113: 318 - 326. Morris CA, Leonard CO, Dilts C, Demsey SA (1990). Adults with Williams syndrome. Am. J. Med. Genet. 6: 102 - 107. Morris CA, Thomas IT, and Greenberg F (1993). Williams syndrome: autosomal dominant inheritance. Am. J. Med. Genet. 47: 478 - 481. 106 Morrow B, Goldberg R, Carlson C, Das Gupta r, Sirotkin H, Collins J, Dunham I, O'Donnell H, ScambTer P, Shprintzen R, and Kuckerlapati R (1995). Molecular definition of the 22qll deletions in Velo-cardio-facial syndrome. Am. J. Hum. Genet. 56: 1391 - 1403. Nicholls RD, Knoll JH, Glatt K, Hersh JH, Brewster TD, Graham JM Jr, Wurster-Hill D, Wharton R, and Latt SA (1989a). Restriction fragment length polymorphisms within proximal 15q and their use in molecular cytogenetics and the Prader-Willi syndrome. Am. J.Med. Genet. 33: 76-77. Nicholls RD, Knoll JH, Butler MG, Karam S, and Lalande M (1989b). Genetic imprinting suggested by maternal heterodisomy in nondeletion Prader-Willi syndrome. Nature 342: 281 -285. Nicholls RD, Saitoh S, and Horsthemke B (1998). Imprinting in Prader-Willi and Angelman syndromes. Trends Genet. 14: 194 - 199. Nickerson E, Greenberg F, Keating M, McCaskill C, and Shaffer L (1995). Deletions of the elastin gene at 7q 11.23 occur in -90% of patients with Williams syndrome. Am. J. Hum. Genet. 56: 1156- 1161. O'Connor W (1985). Supravalvular aortic stenosis: clinical and pathologic observations in six patients. Arch. Pathol. Lab. Med 109: 179 - 185. 107 Olson TM, Michels W , Urban Z, Csiszar K, Christiano AM, Driscoll DJ, Feldt RH, Boyd CD, and Thibodeau SN (1995). A 30 kb deletion within the elastin gene results in familial supravalvular aortic stenosis. Hum. Mol. Genet. 4: 1677 - 1679. Osborne LR, Martindale D, Scherer SW, Shi X-M, Huizenga J, Heng HQ, Costa T, Pober B, Lew L, Brinkman J, Rommens J, Koop B, and Tsui LC (1996). Identification of genes from a 500 kb region at 7ql 1.23 that is commonly deleted in Williams syndrome patients. Genomics 36: 328 - 336. Osborne LR, Herbrick J-A, Greavette T, Heng HHQ, Tsui L-C, and Scherer SW (1997). PMS2-related genes flank the rearrangement break- points associated with Williams syndrome and other diseases on human chromosome 7. Genomics 45: 402 - 406. Pankau R, Gosch A, Simeoni E, and Wessel A (1993). Williams-Beuren syndrome in monozygotic twins with variable expression. Am. J. Med. Genet. 47: 475 - 477. Pankau R, Partsch C-J, Winter M, Gosch A, and Wessel A (1996). Incidence and spectrum of renal abnormalities in Williams-Beuren syndrome. Am. J. Med Genet. 63: 301 -304. Patel PI, Roa BB, Welcher AA, Schoener-Scott R, Trask BJ, Pentao L, Snipes GJ, Garcia CA, Francke U, Shooter EM, Lupski JR, and Suter U (1992). The gene for the peripheral 108 myelin protein PMP-22 is a candidate for Charcot-Marie-Tooth disease type 1A. Nat. Genet. 1: 159- 165. Pasteris NG, Trask B, Sheldon S, and Gorski JL (1992). A chromosome deletion 2q35-36 spanning loci HuP2 and COL4A3 results in Waardenburg syndrome type III (Klein-Waardenburg syndrome). Am. J. Hum. Genet. 51 (suppl.): A224. Pasteris NG, Trask B, Sheldon S, and Gorski JL (1993). Discordant phenotype of two overlapping deletions involving the PAX3 gene in chromosome 2q35. Hum. Molec. Genet. 2: 953 - 959. Pentao L, Wise CA, Chinault AC, Patel PI, Lupski JR (1992). Charcot-Marie-Tooth type 1A duplication appears to arise from recombination at repeat sequences flanking the 1.5 Mb monomer unit. Nat. Genet. 2: 292 - 300. Perez Jurado LA, Peoples R, Kaplan P, Hamel BCJ, and Francke U (1996). Molecular definition of the chromosome 7 deletion in Williams syndrome and parent-of-origin effects on growth. Am. J. Hum. Genet. 59: 781 - 792. Perez Jurado LA, Wang Y-K, Peoples R, Coloma A, Cruces J, and Francke U (1998). A duplicated gene in the breakpoint regions of the 7qll.23 WBS deletion encodes the 109 initiator binding protein TFII-I and BAP-13 5 a phosphorylation target of BTK. Hum. Mol. Genet. 7: 325 - 334. PerouM (1961). Congenital supravalvular aortic stenosis. Arch. Pathol. 71: 113 - 126. Preuss M (1984). The Williams syndrome: objective definition and diagnosis. Clin. Genet. 25: 422-428. Read AP and Newton VE (1997). Waardenburg syndrome. J. Med. Genet. 34: 656 - 665. Rinchik EM, Bultman SJ, Horsthemke B, Lee ST, Strunk KM, Spritz RA, Avidano KM, Jong MTC, and Nicholls RD (1993). A gene for the mouse pink-eyed dilution locus and for human type II oculocutaneous albinism. Nature 361: 72-76. Robinson WP, Binkert F, Bernasconi F, Lorda-Sanchez I, Werder EA, and Schinzel AA (1995) . Molecular studies of chromosomal mosaicism: relative frequency of chromosome gain or loss and possible role of cell selection. Am. J. Hum. Genet. 56: 444 - 451. Robinson WP and Kalousek DK (1996). Mosaicism most likely accounts for extended survival of trisomy 22. Am. J. Med. Genet. 62: 100 - 101. Robinson WP, Waslynka J, Bernasconi F, Wang M, Clark S, Kotzot D, and Schinzel A (1996) . Delineation of 7qll.2 deletions associated with Williams-Beuren syndrome and 110 mapping of a repetitive sequence to within and to either side of the common deletion. Genomics 34: 17-23. Robinson WP, Dutly F, Nicholls RD, Bernasconi F, Penaherrera M, Michaelis RC, Abeliovich D, and Schinzel AA (1998). The mechanisms involved in formation of deletions and duplications of 15qll-ql3. J. Med. Genet. 35: 130 - 136. Robson WL, Leung AK, and Rogers RC (1995). Unilateral renal agenesis. Adv. Pediatr. 42: 575 - 592. Romesburg HC (1984). Cluster Analysis for Researchers. Lifetime Learning Publications, Belmont. Rosenblatt J and Mitchison TJ (1998). Actin, cofdin and cognition. Nature 393: 739 -740. Sadler LS, Robinson LK, Verdaasdonk KR, and Gingell R (1993). The Williams syndrome: evidence for possible autosomal dominant inheritance. Am. J. Med Genet. 47: 468 - 470. Saitoh S, Buiting K, Rogan PK, Buxton JL, Driscoll DJ, Arnemann J, Konig R, Malcolm S, Horsthemke B, and Nicholls RD (1996). Minimal definition of the imprinting centre 111 and fixation of a hromosome 15qll-ql3 epgenotype by imprinting mutations. Proc. Natl. Acad. Sci. USA 93: 7811 -7815. Savory TH (1970). The mule. Sci. Am. 223: 102 - 109. Schinzel A (1988). Microdeletion syndromes, balanced translocations, and gene mapping. J. Med Genet. 25: 454 - 462. Schmickel RD (1986). Contiguous gene syndromes: a component of recognizable syndromes. J. Pediatr. 109: 231 -241. Strachan T and Read AP (1996). Human Molecular Genetics. BIOS Scientific Publishers, New York. Surani MAH, Carton SC, and Norris ML (1984). Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308: 548 - 550. Sutcliffe JS, Nakao M, Christian S, Orstavik KH, Tommerup N, Ledbetter DH, and Beaudet AL (1994). Deletions of a differentially methylated CpG island at the SNRPN gene define a putative imprinting control region. Nat. Genet. 8: 52 - 58. Tassabehji M, Newton VE, Leverton K, Turnbull K, Seemanova E, Kunze J, Sperling K, Strachan T, Read AP (1994). PAX3 gene structure and mutations: close analogies 112 between Waardenburg syndrome and the Splotch mouse. Hum. Molec. Genet. 3: 1069 -1074. Tassabehji M, Metcalfe K, Donnai D, Hurst J, Reardon W, Burch M, and Read AP (1997a). Elastin: genomic structure and point mutations in patients with supravalvular aortic stenosis. Hum. Mol. Genet. 6: 1029 - 1036. Tassabehji M, Gladwin AJ, Karmiloff-Smith A, Donnai D, Dennis D, Splitt M, Read AP, and Metcalfe K (1997b). LJJVTKl hemizygosity does not contribute to the Williams syndrome cognitive profile. J. Med. Genet. 34 (Suppl. 1): S56. Tassabehji M, Metcalfe K, Hurst J, Ashcroft GA, Kielty C, Wilmot C, Donnai D, Read AP, and Jones CJP (1998). An elastin gene mutation producing abnormal tropoelastin and abnormal elastic fibres in a patient with autosomal dominant cutis laxa. Hum. Molec. Genet. 7: 10210- 1028. Tassabehji M, Metcalfe K, Karmiloff-Smith A, Carette MJ, Grant J, Dennis N, Reardon W, Splitt M, Read AP, and Donnai D (1999). Williams syndrome: use of chromosomal microdeletions as a tool to dissect cognitive and physical phenotypes. Am. J. Hum. Genet. 64: 118- 125. Thomson G (1995). HLA disease associations: models for the study of complex human genetic disorders. Crit. Rev. Clin. Lab Sci. 32: 183 - 219. 113 Tromp G, Christiano A, Goldstein N, Indik Z, Boyd C, Rosenbloom J, Deak S, Prockop D, and Kuivaniemi H (1991). A to G polymorphism in ELN gene. Nucleic Acids Res. 19: 4314. Tsui L-C, Donis-Keller H, and Grzeschik K-H (1995). Report of the second Chromosome 7 mapping workshop. Cytogenet. Cell Genet. 71: 1-31. Udwin O and Yule W (1990). Expressive language of children with Williams syndrome. Am. J. Med. Genet. 6: 108 - 114. Urban Z, Helms C, Fekete G, Csiszar K, Bonnet D, Munnich A, Donis-Keller H, and Boyd C (1996). 7qll.23 deletions in Williams syndrome arise as a consequence of unequal meiotic crossover. Am. J. Hum. Genet. 59: 958 - 962. Urban Z, Csiszar K, Fekete G, and Boyd CD (1997). A tetranucleotide repeat polymorphism within the human elastin gene (ELNil). Clin. Genet. 51: 133 - 134. Vulpe CD and Packman S (1995). Cellular copper transport. Annu. Rev. Nutr. 15: 293 -322. Williams JCP, Barrat-Boyes BG, and Lowe JB (1961). Supravalvular aortic stenosis. Circulation 24: 1311-1318. 114 Wilson D, Burn J, Scambler P, and Goodship J (1993). DiGeorge syndrome: part of CATCH 22. J. Med Genet. 30: 852 - 856. Winter M, Pankau R, Amm M, Gosch A, Wessel A (1996). The spectrum of ocular features in the Williams-Beuren syndrome. Clin. Genet. 49: 28 - 31. Wu Y-Q, Sutton VR, Nickerson E, Lupski JR, Potocki L, Korenberg JR, Greenberg F, Tassabehji M, and Shaffer LG (1998). Delineation of the common critical region in Williams syndrome and clinical correlation of growth, heart defects, ethnicity, and parental origin. Am. J. Med. Genet. 78: 82 - 89. Yamagishi H, Ishii C, Maeda J, Kojima Y, Matsuoka R, Kimura M, Takao A, Momma K, and Matsuo N (1998). Phenotypic discordance in monozygotic twins with 22qll.2 deletion. Am. J. Med. Genet. 78: 319 - 321. Yang N, Higuchi O, Ohashi K, Nagata K, Wada A, Kangawa K, Nishida E, and Mizuno K (1998). Cofdin phosphorylation by LIM-kinase 1 and its role in Rac-mediated actin reorganization. Nature 393: 809 - 812. Zar JH (1984). Biostatistical Analysis, 2nd Edition. Prentice-Hall Inc, Englewood Cliffs. 115 APPENDIX I: Dataform WBS ID: Name: Medg #: Date of examination: Date of birth: Parents Birth Date Height Cardiac Malformation Dysmorphic Features Mother Father Birth: Weeks gestation Weight: Length: OFC: Complications: Early Infancy ( 0 - 2 yrs): Ca levels: Constipation/diarrhea: Feeding problems: At examination: Age: y mo Weight: ( %) Height: ( %) OFC: ( %) Cardiac Malformation: Supravalvular aortic stenosis: Peripheral Pulmonary stenosis: Other: Eyes: Strabismus: Stellate pattern: Colour: Teeth: Partial anodontia: Enamel hypoplasia: Facial Dysmorphism: Broad forehead: ( %) ( %) ( %) 116 Periorbital fullness: Epicanthic folds: Flat nasal bridge: Long and smooth philtrum: Full cheeks and lips: Wide mouth: Ears: Hands: Total hand length: ( %) Middle finger length: ( %) Clinodactyly: Hypoplastic or dysplastic nails: Feet: Total foot length: ( %) Clinodactyly: Hypoplastic or dysplastic nails: Hallux Valgus: Kidney/urogenital abnormalities: Psychomotor development: Walking: Speech: Low Normal High Activity Sociability Anxiety Loquaciousness Elastin Deletion: Parental Origin of Deletion: Other Comments: 117 APPENDIX H: Binary data for Cluster Analysis WBS rrenatai <10% fostnatal <10% nyper-Ca uaraiac Mai form i-acies >4/7 stel «t Diue iris Strabis eenavior >2/4 D5 Clino uys/nypopi Nails 3 0 1 ? 1 0 1 1 1 1 1 5 0 1 1 1 1 1 1 1 1 1 6 0 1 ? 0 0 0 1 0 0 7 0 1 ? 1 1 1 1 1 0 1 9 0 1 0 1 1 0 1 1 0 11 1 ? 0 0 1 0 1 ? 0 0 12 1 1 0 1 0 0 0 1 0 0 15 1 1 0 1 0 0 0 0 0 18 1 ? 0 1 0 0 0 1 0 0 19 ? 1 ? 0 1 0 0 0 0 0 20 1 1 0 1 0 0 0 ? 0 0 21 ? 1 0 0 0 0 1 0 0 0 22 1 0 1 1 0 1 1 0 1 24 ? 1 ? 1 1 0 1 0 0 0 25 ? 1 7 1 1 0 0 0 0 0 31 0 0 0 1 0 1 0 1 0 34 0 1 ? 0 0 0 0 0 1 35 1 0 1 1 0 0 1 1 1 38 1 1 ? 1 1 1 1 1 1 1 39 0 1 0 1 1 ? ? 1 1 41 0 1 0 1 1 0 1 1 1 44 1 0 1 1 1 1 1 1 0 46 1 1 0 1 1 0 1 1 1 1 50 0 1 0 1 1 0 1 0 1 52 0 1 0 1 1 0 1 ? 1 1 68 0 1 1 0 0 0 1 1 1 70 0 1 ? 0 0 1 1 1 1 1 z73 ? 1 0 0 0 0 0 0 0 v76 ? 1 1 0 0 0 0 1 1 0 v79 ? ? ? 1 ? ? 1 1 0 ? v80 0 1 1 1 0 0 1 ? 1 1 z82 1 1 1 1 1 1 0 1 0 0 v84 1 1 1 1 1 1 1 1 1 z84 1 1 ? 1 ? 0 0 1 0 1 v86 1 1 0 1 1 0 0 1 1 1 v87 0 1 1 1 1 1 0 1 1 0 v88 1 1 ? 1 ? ? ? ? ? ? z89 0 1 0 1 1 0 1 1 1 0 v90 0 1 0 1 ? ? 0 ? ? ? z90 1 1 7 1 ? 0 1 0 0 0 v91 1 1 0 ? ? ? ? 0 1 z91 0 0 1 ? 1 1 1 1 1 v92 7 1 1 0 ? 7 ? 7 7 7 v93 ? 1 1 0 1 ? ? ? ? ? v94 0 1 0 1 0 0 1 ? 7 ? v95 1 1 1 1 0 ? ? ? 1 0 v96 0 1 1 1 0 ? 0 ? 0 1 v97 1 1 1 1 1 0 0 ? 0 0 vlOO 1 1 ? 1 0 0 ? 1 1 0 v103 0 ? 0 1 1 0 1 1 0 v106 1 1 0 1 1 0 0 1 1 0 v107 1 0 0 1 0 0 0 1 0 v108 ? 1 ? 0 1 1 1 1 1 0 v109 0 0 ? 1 1 0 1 1 1 0 v110 0 0 1 1 1 0 1 1 1 0 v112 ? 0 1 1 1 0 1 0 1 0 v115 1 0 ? 0 1 0 0 0 1 1 v116 ? 1 1 1 1 1 0 0 0 v120 0 1 1 1 1 1 1 1 0 0 v123 ? 0 1 1 1 1 1 ? 0 0 118 APPENDIX H: Binary data for Cluster Analysis Prenatal <10% postnatal <10% 1 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 1 1 0 1 1 1 1 0 0 uardiac Malform h a c i e 8 >4/7 1 1 uys/nypopi Nails WBS v126 v126 v127 v128 v132 v71 v72 v73 v74 v78 v81 v82 v89 v98 v99 v105 v111 v113 v114 v117 v118 v119 v121 v122 v130 nyper-Ca Stel & blue iris 0 0 0 ? ? Strabis 0 1 1 1 0 0 0 1 0 1 0 0 0 1 1 0 1 0 0 0 0 •senavior >2/4 D5 Clino 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 119 APPENDIX HJ: Estimated necessary sample sizes to detect a true association with study-wide a=0.05 and (3=0.10 Comparison ('n' number of comparisons with p<0.05 Sample size required Total identified in the present study) <x'=0.05/n in each group sample Parental origin of deletion (2) - 0.025 Maternal Paternal N OFC <10% at exam 27 26 53 Hoarse voice 24 24 48 Gender (3) - 0.017 Male Female N Cardiovascular defect 122 122 244 Typical facies (>4/5 features) 99 88 187 Dys/hypoplastic nails 52 62 114 Pair-wise clinical comparisons (14) 3.6xl0"3 1st trait present 1st trait absent N Weight <10% at exam - cardiovascular defect 153 163 316 Clinodactyly - typical facies 168 110 278 Height <10% at exam - OFC <10% at exam 149 117 266 Clinodactyly - typical personality (>2/4 features) 132 86 218 Height <10% at exam - stellate irides 115 92 207 Blue irides - dental abnormalities 76 97 173 Blue irides - typical personality (>2/4 features) 72 93 165 Renal/urogenital abnormalities - cardiovascular defect 29 102 131 Stellate irides - typical personality 50 70 120 Stellate irides - blue irides 42 53 95 Hoarse voice - dys/hypoplastic nails 57 21 78 Hoarse voice - dental abnormalities 56 19 75 Weight <10% at exam - weight <10% at birth 34 39 73 Stellate irides - infantile hypercalcaemia 19 19 38 7q 11.23 deletion (5) - 0.010 Deleted Non-deleted N OFC <10% at exam 210 58 268 Strabismus 160 45 205 Wide mouth 132 30 162 Highly sociable 109 34 143 Cardiovascular defect 78 19 97 120 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items