An international database and integrated analysis tools for the study of cancer gene expression.

Researchers working collaboratively in Brazil and the United States have assembled an International Database of Cancer Gene Expression. Several strategies have been employed to generate gene expression data including expressed sequence tags (ESTs), serial analysis of gene expression (SAGE), and open reading-frame expressed sequence tags (ORESTES). The database contains six million gene tags that reflect the gene expression profiles in a wide variety of cancerous tissues and their normal counterparts. All sequences are deposited in the public databases, GenBank and SAGEmap. A suite of informatics tools was designed to facilitate in silico analysis of the gene expression datasets and are available through the NCI Cancer Genome Anatomy Project web site (http://cgap.nci.nih.gov).

In silico analysis of cancer through the Cancer Genome Anatomy Project.

The Cancer Genome Anatomy Project (CGAP) was designed and implemented to provide public datasets, material resources and informatics tools to serve as a platform to support the elucidation of the molecular signatures of cancer. This overview of CGAP describes the status of this effort to develop resources based on gene expression, polymorphism identification and chromosome aberrations, and we describe a variety of analytical tools designed to facilitate in silico analysis of these datasets.

Human hepatocellular carcinoma is characterized by a highly consistent pattern of genomic imbalances, including frequent loss of 16q23.1-24.1.

Comparative genomic hybridization (CGH) analysis was used to identify chromosomal imbalances in 52 human primary hepatocellular carcinomas (HCCs). The most prominent changes were gains of part or all of chromosome arms 8q (83% of cases) and 1q (73%) and loss of 16q (63%). Other commonly overrepresented sites were 5p, 7q, and Xq. Recurrent sites of DNA sequence amplification included 8q23--24 (five cases) and 11q13--14 (four cases). Other frequently underrepresented sites were 4q, 8p, 16p, and 17p. Taken collectively, these findings and data from other CGH studies of HCCs define a subset of chromosome segments that are consistently over- or underrepresented and highlight sites of putative oncogenes and tumor suppressor genes, respectively, involved in hepatocellular oncogenesis. Loss of heterozygosity analysis with a panel of polymorphic microsatellite markers distributed along 16q defined a minimal region of chromosomal loss at 16q23.1--24.1, suggesting that this region harbors a tumor suppressor gene whose loss/inactivation may contribute to the pathogenesis of many HCCs.

High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry.

We describe here a system for the rapid identification, assay development, and characterization of gene-based single nucleotide polymorphisms (SNPs). This system couples informatics tools that mine candidate SNPs from public expressed sequence tag resources and automatically designs assay reagents with detection by a chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry platform. As a proof of concept of this system, a genomewide collection of reagents for 9,115 gene-based SNP genetic markers was rapidly developed and validated. These data provide preliminary insights into patterns of polymorphism in a genomewide collection of gene-based polymorphisms.

Construction of reference genetic maps.

This unit details the specialized resources and procedures used for constructing reference genetic maps. Construction of such maps in humans represents a subset of the linkage analysis problem. Objectives include the addition of a new locus to an established map, development of a detailed map of loci in a localized area, and construction of de novo maps. Conceptually, the procedures for updating a reference map through the addition of a subset of new loci are similar to those used in establishing linkage for a disease locus. However, construction of new maps of multiple loci is most efficiently accomplished using different family resources that permit the use of accumulated typing resources and alternative, highly efficient statistical methods.

Expression-based genetic/physical maps of single-nucleotide polymorphisms identified by the cancer genome anatomy project.

SNPs (Single-Nucleotide Polymorphisms), the most common DNA variant in humans, represent a valuable resource for the genetic analysis of cancer and other illnesses. These markers may be used in a variety of ways to investigate the genetic underpinnings of disease. In gene-based studies, the correlations between allelic variants of genes of interest and particular disease states are assessed. An extensive collection of SNP markers may enable entire molecular pathways regulating cell metabolism, growth, or differentiation to be analyzed by this approach. In addition, high-resolution genetic maps based on SNPs will greatly facilitate linkage analysis and positional cloning. The National Cancer Institute's CGAP-GAI (Cancer Genome Anatomy Project Genetic Annotation Initiative) group has identified 10,243 SNPs by examining publicly available EST (Expressed Sequence Tag) chromatograms. More than 6800 of these polymorphisms have been placed on expression-based integrated genetic/physical maps. In addition to a set of comprehensive SNP maps, we have produced maps containing single nucleotide polymorphisms in genes expressed in breast, colon, kidney, liver, lung, or prostate tissue. The integrated maps, a SNP search engine, and a Java-based tool for viewing candidate SNPs in the context of EST assemblies can be accessed via the CGAP-GAI web site (http://cgap.nci.nih.gov/GAI/). Our SNP detection tools are available to the public for noncommercial use.

Reliable identification of large numbers of candidate SNPs from public EST data.

High-resolution genetic analysis of the human genome promises to provide insight into common disease susceptibility. To perform such analysis will require a collection of high-throughput, high-density analysis reagents. We have developed a polymorphism detection system that uses public-domain sequence data. This detection system is called the single nucleotide polymorphism pipeline (SNPpipeline). The analytic core of the SNPpipeline is composed of three components: PHRED, PHRAP and DEMIGLACE. PHRED and PHRAP are components of a sequence analysis suite developed to perform the semi-automated analysis required for large-scale genomes (provided courtesy of P. Green). Using these informatics tools, which examine redundant raw expressed sequence tag (EST) data, we have identified more than 3,000 candidate single-nucleotide polymorphisms (SNPs). Empiric validation studies of a set of 192 candidates indicate that 82% identify variation in a sample of ten Centre d'Etudes Polymorphism Humain (CEPH) individuals. Our results suggest that existing sequence resources may serve as a valuable source for identifying genetic variation.

Constitutional genetic variation at the human aromatase gene (Cyp19) and breast cancer risk.

The activity of the aromatase enzyme, which converts androgens into oestrogens and has a major role in regulating oestrogen levels in the breast, is thought to be a contributing factor in the development of breast cancer. We undertook this study to assess the role of constitutional genetic variation in the human aromatase gene (Cyp19) in the development of this disease. Our genotyping of 348 cases with breast cancer and 145 controls (all Caucasian women) for a published tetranucleotide repeat polymorphism at intron 4 of the Cyp19 gene revealed the presence of six common and two rare alleles. Contingency table analysis revealed a significant difference in allelic distribution between cases and controls (chi2 5df = 13.52, P = 0.019). The allele measuring 171 bp was over-represented in cases; of 14 individuals homozygous for this allele, 13 were cases. These individuals had a higher incidence of cancer in family members and an earlier age at diagnosis than other cases. In sequencing Cyp19's coding exons and regulatory regions, we discovered a perfect association between a silent polymorphism (G-->A at Val80) and the high-risk genotype. Our conclusion is that constitutional genetic variation at the Cyp19 locus is associated with the risk of developing breast cancer, with the 171-bp allele serving as the high-risk allele.

Asthma and bronchial hyperresponsiveness linked to the XY long arm pseudoautosomal region.

We examined the long arm XY pseudoautosomal region for linkage to asthma, serum IgE, and bronchial hyperresponsiveness. In 57 Caucasian families multipoint nonparametric analyses provide evidence for linkage between DXYS154 and bronchial hyperresponsiveness (P = 0.000057) or asthma (P = 0.00065). This genomic region is approximately 320 kb in size and contains the interleukin-9 receptor gene. These results suggest that a gene controlling asthma and bronchial hyperresponsiveness maybe located in this region and that the interleukin-9 receptor is a potential candidate.

Identification of inbred mouse strains harboring genetic modifiers of mammary tumor age of onset and metastatic progression.

Metastasis is one of the most important and complex processes in human neoplastic disease. A large number of both positive and negative events must occur to permit a tumor cell to colonize a distant site successfully. To identify mouse strains that harbor dominant genetic modifiers of this process, a strain survey was initiated utilizing a transgenic mouse mammary tumor model that exhibits a high incidence of pulmonary metastases. The transgenic animal was bred to 27 different inbred strains of mice and scored for the metastatic organ tropism and metastatic density. Thirteen strains were identified that had a statistically significant reduction in the numbers of pulmonary metastases. In addition, 10 strains were identified that altered the kinetics of induction of the primary mammary tumor. These strains will likely provide useful model systems for the analysis of genetic interactions in the initiation and progression of mammary adenocarcinomas.

Interleukin 9: a candidate gene for asthma.

Asthma is a complex heritable inflammatory disorder of the airways associated with clinical signs of atopy and bronchial hyperresponsiveness. Recent studies localized a major gene for asthma to chromosome 5q31-q33 in humans. Thus, this segment of the genome represents a candidate region for genes that determine susceptibility to bronchial hyperresponsiveness and atopy in animal models. Homologs of candidate genes on human chromosome 5q31-q33 are found in four regions in the mouse genome, two on chromosome 18, and one each on chromosomes 11 and 13. We assessed bronchial responsiveness as a quantitative trait in mice and found it linked to chromosome 13. Interleukin 9 (IL-9) is located in the linked region and was analyzed as a gene candidate. The expression of IL-9 was markedly reduced in bronchial hyporesponsive mice, and the level of expression was determined by sequences within the qualitative trait locus (QTL). These data suggest a role for IL-9 in the complex pathogenesis of bronchial hyperresponsiveness as a risk factor for asthma.

Defining etiologic heterogeneity in breast cancer using genetic biomarkers.

Most breast cancer has a complex, multifactorial etiology. One consequence of this multifactorial phenomenon is that etiological heterogeneity may exist. This heterogeneity implies simply that two or more groups of breast cancer cases in the general population may have been caused by different sets of events. The ability to define etiologically heterogeneous subgroups in the population may facilitate a number of research and clinical issues. Studying etiologically homogeneous subgroups in the general population may improve the ability to identify etiologic agents. Identification of a homogeneous group of breast cancer cases may also aid breast cancer diagnosis or treatment, and may allow a more effectively application of cancer prevention and control strategies. Defining etiologic heterogeneity in the general population is one initial step in the process of understanding cancer etiology. Using knowledge such as that provided in the two examples presented here, formal case-control or cohort studies can be undertaken to examine whether the factors that define etiologic heterogeneity are involved in etiology. Furthermore, the results of studies of etiologic heterogeneity can point toward potential gene-gene or gene-environment interactions. The type of studies presented here can therefore serve a useful role in leading to more formal molecular epidemiological analyses.

mtDNA analysis shows common ancestry in two kindreds with X-linked recessive hypoparathyroidism and reveals a heteroplasmic silent mutation.

Two kindreds residing in eastern Missouri and exhibiting X-linked recessive idiopathic hypoparathyroidism have been described. Genealogical records extending back five generations revealed no common ancestor. To investigate the possibility of relatedness, the DNA sequence of the mitochondrial D-loop was compared among several individuals in both kindreds. The mtDNA D-loop was amplified from the total DNA of individuals by use of nested PCR reactions, and the resulting 430-bp fragment was sequenced. The mtDNA sequence was identical among affected males and their maternal lineage for individuals in both kindreds. Conversely, the mtDNA sequence of the fathers of the affected males differed from that of the maternal lineage at three to six positions. These results demonstrate that the two kindreds exhibiting X-linked recessive hypoparathyroidism are indeed related and that an identical gene defect is responsible for the disease. A further feature of the inheritance pattern was examined when a unique point mutation was identified in the mtDNA of one branch of one of the kindreds. This mutation appears to be de novo and segregates in subsequent generations without obscuring relatedness. In addition, the results of our study of mtDNA analysis indicate that this approach may be of importance in investigating common ancestry in other X-linked disorders.

Studies of the candidate genes TGFB2, MSX1, TGFA, and TGFB3 in the etiology of cleft lip and palate in the Philippines.

Population-based candidate-gene studies can be an effective strategy for identifying genes involved in the etiology of disorders where family-based linkage studies are compromised by lack of access to affected members, low penetrance, and/or genetic heterogeneity. We evaluated association data for four candidate genes using a population from the Philippines that is genetically separate from previously studied Caucasian populations. Case ascertainment was made possible by collaboration with Operation Smile, a volunteer medical organization, which facilitated identification of a large number of cases for study. A new allelic variant of transforming growth factor-beta 3 was identified to use in these studies. After exclusion of syndromic cases of cleft lip and palate, no evidence for association with previously reported allelic variants of transforming growth factor-beta 2 (TGFB2), homeobox 7 (MSX1), or transforming growth factor-alpha (TGFA), or with the new TGFB3 variant was detected. Previous association studies using Caucasian populations of nonsyndromic cleft lip and/or palate (CL/P) and cleft palate only (CPO) have strongly suggested a role for TGFA in the susceptibility of clefting in humans. Exclusion of significant association in a non-Caucasian population for TGFA suggests that TGFA plays less of a role than it does in Caucasians. This may be due to multiple or different genetic and/or environmental factors contributing to the etiology of this most common cranio-facial anomaly in the Philippine population.

Clinical and epidemiologic studies of cleft lip and palate in the Philippines.

Clinical and epidemiologic studies of defined geographic populations can serve as a means of establishing data important for genetic counseling and as a first step in identifying strategies best suited for identification of causes. Under the sponsorship of Operation Smile International, clinical, genetic, and epidemiologic studies were carried out at six sites within the Philippines between 1989 and 1996. Patients who were being evaluated for surgical repair of craniofacial anomalies (primarily clefts of the lip and palate) were briefly examined for the presence of associated anomalies, and a family history was obtained to look for the frequency of cleft lip and palate in siblings. Birth records of 47,969 newborns over an 8-year period at one hospital in Bacolod City in the province of Negros Occidental were reviewed. Medical records of infants born with clefts of the lip and/or palate and other major anomalies were reviewed and birth prevalence rates calculated. Findings include a birth prevalence of 1.94 per 1000 live births for cleft lip with/without palate in the Philippines. Recurrence rates in siblings for nonsyndromic clefts of the lip and palate were 23 per 1000 for cleft lip with or without cleft palate, and 14 per 1000 for cleft palate only. The percentage of clefts associated with multiple anomalies was 21% at birth and 6% for individuals examined during the screening process, providing evidence for a high postnatal death rate. These data provide groundwork for additional etiologic studies including segregation analysis and molecular genetic studies involving linkage or association, as well as for studies of environmental contributions to clefting such as vitamin deficiencies. Preliminary molecular analysis using an association approach is reported in a companion paper. The findings suggest a high incidence of cleft lip and palate in native-born Filipinos.

Variability in loss of constitutional heterozygosity across loci and among individuals: association with candidate genes in ductal breast carcinoma.

Genes involved in the metabolic activation or detoxification of environmental carcinogens may contribute to breast cancer susceptibility by influencing rates of somatic mutation. To examine this hypothesis, we studied the association between loss of constitutional heterozygosity (LOH) in ductal breast tumors and allelic variability in genes that regulate the metabolism of environmental carcinogens. LOH was measured by typing the tumor and normal tissue of 28 breast cancer cases at 33 chromosomal loci by using highly polymorphic tetranucleotide repeat markers. Genotypes in non-tumor tissue were also measured at the cytochrome P450 1A1 (CYP1A1), cytochrome P450 2D6 (CYP2D6), glutathione-s-transferase mu (GSTM), epoxide hydrolase (EH), and NAD(P)H:quinone oxidoreductase (NQO1) loci. The observed proportion of LOH was 11% overall and ranged from 0% to 37% across loci. LOH greater than 20% was observed on chromosomes 1p, 2p, 10q, 11q, 17p, and 18q. The observed proportion of LOH ranged from 0% to 67% among individuals. An elevated proportion of LOH was observed for genotypes at CYP2D6 (17% for the 1/1 and 1/2 genotypes vs 8% for the 2/ 2 genotype), NQO1 (13% for the 1/2 and 2/2 genotypes vs 8% for the 1/1 genotype), and GSTM (15% for the null genotype vs 7% for the wild-type genotype). No elevated proportion of LOH was observed for genotypes at CYP1A1 (12% for the 1/2 genotype vs 10% for the 1/1 genotype) or EH (11% for the 1/1 genotype vs 10% for the 1/2 genotype). There was no correlation of LOH with any other tumor characteristic such as estrogen- or progesterone-receptor status or number of positive lymph nodes. These results suggest that the proportion of LOH varies substantially across loci and among individuals. Interindividual variability in LOH may thus be explained in part by genes that regulate the metabolism of environmental carcinogens.