Faculty Bibliography

Six novel polymorphic short sequence repeats were identified and localized on the linkage map of human chromosome 21 by genotyping the CEPH reference pedigrees. One of these markers, the tetrameric (AAAG)n repeat D21S1245, was found to be hypermutable. In the DNAs from lymphoblastoid cell lines of members of the 40 CEPH families a total of 18 new alleles were detected. These new alleles, sometimes appearing in mosaic forms, arose equally in paternal and maternal DNAs, and could be equally larger or smaller than the alleles from which they were derived. The larger alleles of D21S1245 are more prone to be converted to new alleles. None of the new alleles with mosaicism were present in the corresponding genomic blood DNA, and therefore originated during or after the establishment of the lymphoblastoid cell lines; half of the new alleles without mosaicism were also found in genomic blood DNA of the appropriate CEPH individuals. The range of germline mutation rate observed in the 716 meioses examined was 0.56-1.4 x 10(-2); the range of somatic mutations observed in the 405 cell lines examined was 1.96-3.46 x 10(-2). This is one of the most hypermutable microsatellite repeat polymorphism in the human genome detected to date. D21S1245, is highly polymorphic (heterozygosity of 0.96) and maps between D21S231 and D21S198.

We conducted family studies and segregation analyses of rheumatoid arthritis (RA) that were based on consecutive patients with RA ascertained without regard to family history or known risk factors. First-degree relatives from 135 simplex and 30 multiplex families were included in the analyses. A highly penetrant recessive major gene, with a mutant allele frequency of .005, was identified as the most parsimonious genetic risk factor. Significant evidence for heterogeneity in risk for RA was observed for proband gender but not for proband age at onset. Kaplan-Meier risk analysis demonstrated significant evidence for differences in the distribution of risk among first-degree relatives. These analyses demonstrated that both proband gender and age at onset are important risk factors but that proband gender appears to be the more important determinant of risk, with relatives of male probands having the greatest cumulative risk for RA. In addition, log-linear modeling identified proband gender, familiality (multiplex or simplex), and an interaction term between these two variables as being adequate to define the distribution of risk in families. The pattern of risk for RA among susceptible individuals and its inheritance is thus heterogeneous. For future genetic analyses, families with an excess of affected males having a young age at onset may be the most informative in identifying the putative recessive gene and its modifiers.

Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are distinct mental retardation syndromes caused by paternal and maternal deficiencies, respectively, in chromosome 15q11-q13. Approximately 70% of these patients have a large deletion of approximately 4 Mb extending from D15S9 (ML34) through D15S12 (IR10). To further characterize the deletion breakpoints proximal to D15S9, three new polymorphic microsatellite markers were developed that showed observed heterozygosities of 60%-87%. D15S541 and D15S542 were isolated from YAC A124A3 containing the D15S18 (IR39) locus. D15S543 was isolated from a cosmid cloned from the proximal right end of YAC 254B5 containing the D15S9 (ML34) locus. Gene-centromere mapping of these markers, using a panel of ovarian teratomas of known meiotic origin, extended the genetic map of chromosome 15 by 2-3 cM toward the centromere. Analysis of the more proximal S541/S542 markers on 53 Prader-Willi and 33 Angelman deletion patients indicated two classes of patients: 44% (35/80) of the informative patients were deleted for these markers (class I), while 56% (45/80) were not deleted (class II), with no difference between PWS and AS. In contrast, D15S543 was deleted in all informative patients (13/48) or showed the presence of a single allele (in 35/48 patients), suggesting that this marker is deleted in the majority of PWS and AS cases. These results confirm the presence of two common proximal deletion breakpoint regions in both Prader-Willi and Angelman syndromes and are consistent with the same deletion mechanism being responsible for paternal and maternal deletions. One breakpoint region lies between D15S541/S542 and D15S543, with an additional breakpoint region being proximal to D15S541/S542.

We have used a half-YAC containing the human chromosome 21 long-arm telomere to clone, map, and characterize a new dinucleotide repeat polymorphism (D21S1575) close to 21qter. This marker is < 120 kb from the telomeric (TTAGGG)n sequences and is the most distal highly polymorphic marker on chromosome 21q. This marker has a heterozygosity of 71% because of a variable (TA)n repeat embedded within a long interspersed element (LINE) element. Genotyping of the CEPH families and linkage analysis provided a more accurate determination of the full length of the chromosome 21 genetic map. A highly significant difference was detected between male and female recombination rates in the telomeric region: in the most telomeric 2.3 Mb of chromosome 21q, recombination was only observed in male meioses.

Hirschsprung disease (HSCR), or congenital aganglionic megacolon, is the most common cause of congenital bowel obstruction with an incidence of 1 in 5000 live births. Recently, linkage of an incompletely penetrant, dominant form of HSCR was reported, followed by identification of mutations in the RET receptor tyrosine kinase. To determine the frequency of RET mutations in HSCR and correlate genotype with phenotype, we have screened for mutations among 80 HSCR probands representing a wide range of phenotypes and family structures. Polymerase chain reaction (PCR) and single-strand conformation polymorphism (SSCP) analysis of RET's 20 exons for mutations among probands revealed eight putative mutations (10%). Sequence changes, which included missense, frameshift and complex mutations, were detected in both familial and isolated cases, among patients with both long- and short-segment HSCR and in three kindreds with other phenotypes (maternal deafness, talipes and malrotation of the gut, respectively). Two mutations (C609Y and C620R) we identified have previously been associated with multiple endocrine neoplasia type 2A (MEN2A), medullary thyroid carcinoma (MTC) and, on rare occasions, HSCR. Thus, while HSCR family members may be at risk for developing neuroendocrine tumors, it follows that identical mutations in RET may be able to participate in the pathogenesis of distinct phenotypes. Our data suggest that: (i) the overall frequency of RET mutations in HSCR patients is low and therefore, other genetic and/or environmental determinants contribute to the majority of HSCR susceptibility, and (ii) at present, there is no obvious relationship between RET genotype and HSCR phenotype.

Hirschsprung disease (HSCR) is a congenital disorder of unknown etiology characterized by the absence of enteric ganglia in the distal colon. We have ascertained a large, inbred, Mennonite kindred which demonstrates a high incidence of Hirschsprung disease (HSCR). Genealogical analysis of all kinship relationships identified a single common ancestral couple for all parents of affected offspring. Segregation analysis yielded a segregation ratio of 10.67% for males and 5.45% for females. We searched for locations of the gene(s) responsible for HSCR in this pedigree by genotyping three small multicase families and locating genomic regions demonstrating identity-by-descent followed by linkage disequilibrium analysis of 28 additional nuclear families. Based on this novel strategy, we report the mapping of a new locus for HSCR to chromosome 13q22. Nine microsatellite markers spanning 10 cM in this region were genotyped on thirty-one nuclear families. Significant nonrandom association was detected with alleles at markers D13S162, D13S160, D13S170, and AFM240zg9. In addition, our studies reveal preliminary evidence for a genetic modifier of HSCR in this kindred on chromosome 21q22.

Hirschsprung's disease (HSCR) is characterized by an absence of enteric ganglia in the distal colon and a failure of innervation in the gastrointestinal tract. We recently mapped a recessive susceptibility locus (HSCR2) to human chromosome 13q22, which we now demonstrate to be the endothelin-B receptor gene (EDNRB). We identified in HSCR patients a G-->T missense mutation in EDNRB exon 4 that substitutes the highly conserved Trp-276 residue in the fifth transmembrane helix of the G protein-coupled receptor with a Cys residue (W276C). The mutant W276C receptor exhibited a partial impairment of ligand-induced Ca2+ transient levels in transfected cells. The mutation is dosage sensitive, in that W276C homozygotes and heterozygotes have a 74% and a 21% risk, respectively, of developing HSCR. Genotype analysis of patients in a Mennonite pedigree shows HSCR to be a multigenic disorder.

We have identified a polymorphic compound dinucleotide repeat sequence in intron 1 of the beta-amyloid precursor protein (APP) gene on chromosome 21. Using polymerase chain reaction (PCR) amplification of the locus, designated APPivs1, we detected 13 alleles in the CEPH family members (heterozygosity = 0.69). Lod score analysis showed complete linkage of the marker to the loci D21S210 and D21221.

High-resolution genetic linkage maps are indispensable for positional cloning of disease genes. Current procedures for map construction, although aided considerably by many existing computer programs, require extensive user-intervention at each of many repetitive steps. This is time consuming, labour intensive and increases the chance of error. We have developed an expert system computer program, MultiMap, which automates this step-by-step procedure. MultiMap is based on a novel map construction algorithm and allows investigator control of marker locus characteristics, such as informativeness, scorability or distance to nearest neighbours. We used MultiMap to construct a human genetic map at an average resolution of 6 cM, using published genotypes at 1266 microsatellite markers, and further extended this map by adding 397 VNTR and polymorphic gene markers.