A cross-study comparison of gene expression studies for the molecular classification of lung cancer.

PURPOSE: Recent studies sought to refine lung cancer classification using gene expression microarrays. We evaluate the extent to which these studies agree and whether results can be integrated. EXPERIMENTAL DESIGN: We developed a practical analysis plan for cross-study comparison, validation, and integration of cancer molecular classification studies using public data. We evaluated genes for cross-platform consistency of expression patterns, using integrative correlations, which quantify cross-study reproducibility without relying on direct assimilation of expression measurements across platforms. We then compared associations of gene expression levels to differential diagnosis of squamous cell carcinoma versus adenocarcinoma via reproducibility of the gene-specific t statistics and to survival via reproducibility of Cox coefficients. RESULTS: Integrative correlation analysis revealed a large proportion of genes in which the patterns agreed across studies more than would be expected by chance. Correlation of t statistics for diagnosis of squamous cell carcinoma versus adenocarcinoma is high (0.85) and increases (0.925) when using only the most consistent genes identified by integrative correlation. Correlations of Cox coefficients ranged from 0.13 to 0.31 (0.33-0.49 with genes selected for consistency). Although we find genes that are significant in multiple studies but show discordant effects, their number is approximately that expected by chance. We report genes that are reproducible by integrative analysis, significant in all studies, and concordant in effect. CONCLUSIONS: Cross-study comparison revealed significant, albeit incomplete, agreement of gene expression patterns related to lung cancer biology and identified genes that reproducibly predict outcomes. This analysis approach is broadly applicable to cross-study comparisons of gene expression profiling projects.

Nuclear genes involved in mitochondria-to-nucleus communication in breast cancer cells.

BACKGROUND: The interaction of nuclear and mitochondrial genes is an essential feature in maintenance of normal cellular function. Of 82 structural subunits that make up the oxidative phosphorylation system in the mitochondria, mitochondrial DNA (mtDNA) encodes 13 subunits and rest of the subunits are encoded by nuclear DNA. Mutations in mitochondrial genes encoding the 13 subunits have been reported in a variety of cancers. However, little is known about the nuclear response to impairment of mitochondrial function in human cells. RESULTS: We isolated a Rho0 (devoid of mtDNA) derivative of a breast cancer cell line. Our study suggests that depletion of mtDNA results in oxidative stress, causing increased lipid peroxidation in breast cancer cells. Using a cDNA microarray we compared differences in the nuclear gene expression profile between a breast cancer cell line (parental Rho+) and its Rho0 derivative impaired in mitochondrial function. Expression of several nuclear genes involved in cell signaling, cell architecture, energy metabolism, cell growth, apoptosis including general transcription factor TFIIH, v-maf, AML1, was induced in Rho0 cells. Expression of several genes was also down regulated. These include phospholipase C, agouti related protein, PKC gamma, protein tyrosine phosphatase C, phosphodiestarase 1A (cell signaling), PIBF1, cytochrome p450, (metabolism) and cyclin dependent kinase inhibitor p19, and GAP43 (cell growth and differentiation). CONCLUSIONS: Mitochondrial impairment in breast cancer cells results in altered expression of nuclear genes involved in signaling, cellular architecture, metabolism, cell growth and differentiation, and apoptosis. These genes may mediate the cross talk between mitochondria and the nucleus.

Mucinous cancers have fewer genomic alterations than more common classes of breast cancer.

Mucinous cancers of the breast are distinguished histologically by their abundant pools of mucin and low degree of nuclear pleomorphism. Relative to the more common breast cancers of no distinctive type (ductal carcinoma), mucinous cancers have a relatively favorable prognosis. In a study of chromosomal changes in mucinous cancers, we evaluated the extent of loss of heterozygosity (LOH) at chromosomal regions commonly deleted in usual infiltrating ductal carcinoma, including markers on chromosomal arms 1p, 1q, 3p, 6q, 8p, 9p, 11p, 11q, 13q, 16q, 17p, and 17q. Remarkably, we found an average frequency of LOH of only 1.9 of these 12 chromosomal arms in 18 cases of mucinous carcinoma, compared to an average frequency of LOH of 6.4 of these same chromosomal arms in cases of infiltrating ductal cancer. In three of the 18 cases of mucinous carcinoma studied, including one case with regional lymph node metastases, no LOH was seen at any of the 12 chromosomal regions studied. We considered the possibility of other chromosomal loci being more commonly affected in mucinous cancers and conducted comparative genomic hybridization on six of the cases. These studies demonstrated a low overall frequency of genomic copy number changes (mean of 3.1 changes per case) and failed to reveal any other chromosomal locus with frequent losses that had not been evaluated by microsatellite analysis. Together, these data indicate that mucinous cancers of the breast do not have the extensive genomic alterations that are typically found in more common variants of breast cancer. Thus, mucinous cancers most likely have less genetic instability than most other forms of breast cancer and the molecular pathogenesis of this form of breast cancer is likely to be substantially different than that of usual ductal breast cancer.

Microarray Assistant clone organizer and array simulator.

Microarrays are extensively used in molecular biology experiments. While several vendors offer microarrays on a variety of platforms, many researchers prefer to use custom microarrays with a selected list of clones for their experiments. Many research centers have established core facilities for the production of custom microarrays. Microarray production involves a number of steps, including maintaining a master list of stock clones, selecting required clones for custom microarrays, subculturing selected clones, amplifying inserts, recording results, and identifying the orientation of clones in the microarray. We have created a simple, user-friendly, and versatile Microsoft Excel spreadsheet-based software, Microarray Assistant, which can assist the user in all the steps of microarray design and synthesis. In addition, the program gives options to insert, delete, or interchange clones during various steps. The program also gives a visual picture of the locations of the clones in the plates, as well as in the microarray. The program can also be used to assist in the transfer of clones between plates of different configuration.

A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer.

Aberrant hypermethylation of gene promoters is a major mechanism associated with inactivation of tumor-suppressor genes in cancer. We previously showed this transcriptional silencing to be mediated by both methylation and histone deacetylase activity, with methylation being dominant. Here, we have used cDNA microarray analysis to screen for genes that are epigenetically silenced in human colorectal cancer. By screening over 10,000 genes, we show that our approach can identify a substantial number of genes with promoter hypermethylation in a given cancer; these are distinct from genes with unmethylated promoters, for which increased expression is produced by histone deacetylase inhibition alone. Many of the hypermethylated genes we identified have high potential for roles in tumorigenesis by virtue of their predicted function and chromosome position. We also identified a group of genes that are preferentially hypermethylated in colorectal cancer and gastric cancer. One of these genes, SFRP1, belongs to a gene family; we show that hypermethylation of four genes in this family occurs very frequently in colorectal cancer, providing for (i) a unique potential mechanism for loss of tumor-suppressor gene function and (ii) construction of a molecular marker panel that could detect virtually all colorectal cancer.