TITLE:

MICROARRAY GENE EXPRESSION STUDIES IN CARDIAC CELLS (No. 11)

DATE:

Friday, February 14th, 2003

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Paul Paolini, Department of Biology, San Diego State University

ABSTRACT:

Our laboratory group has recently investigated the potential beneficial effects on the heart of a synthetic antidiabetic drug from the thiazolidinedione (TZD) family of insulin-sensitizing compounds used in the treatment of type 2 diabetes. Cell signaling pathways involving a receptor (PPAR?) affected by this drug can be explored by measuring changes in expression levels of the genes controlling these pathways. Microarray technology was used to obtain comprehensive profiling of expression and detect up-and down-regulation of genes produced by an experimental perturbation (here, treatment with the TZD compound over time). Rat genomic microarray chips for the study were generously provided by Motorola/ Amersham.

Primary ventricular cardiomyocytes were harvested from neonatal rats and plated out on culture medium; RNA extracted and purified from these cells were hybridized with cDNA probes (single-stranded oligonucleotides derived from the rat genome) on the microarray chips. Microarrays were exposed to a storage phosphor screen to read fluorescent label intensities; globally averaged background was subtracted from each spot. Gene expression levels (mRNA concentrations) vary over several orders of magnitude, with some signals buried in noise, and others oversaturated. The Motorola/ Amersham microarrays probed for 9,911 rat genes plus 68 bacterial genes, 18 positive, and 50 negative control probes; many of the genes are annotated mRNA from the NCBI public database, while others are ESTs. The spots/wells on the array are mapped and labeled, corrected for background noise, and normalized using housekeeping genes (Goryachev et al., J. Theor. Biol. 2001. 8(2): 443-61). Signals that are <= background are defined as zero (?silent? genes).

Determining whether a gene is up- or down-regulated requires that its differential expression be consistent and statistically significant. The current high cost of this technology limits the number of redundant arrays that can be employed, thereby increasing chances of obtaining false positives. The number of false positives depends dramatically on number of redundant measurements made, e.g. 3 microarray slide repeats for each measurement can yield >5.0% while 6 repeats can lower this to < 0.2% (Liu et al., Circ. Res. 2001. 88: 1231-8),

Making sense of expression data from 104 genes represents an enormous challenge; as a first step, the rat genes in the data set can be characterized using one of several published functional classification schemes. Application of one such scheme, for example, might yield six main categories and 43 subcategories of genes covering the entire genome. For each of the main categories identified, one of several methods can then be applied to identify a few genes of interest for further experimental investigations (for example, using quantitative PCR). Meta-analysis can be employed using a consistent search strategy to collect, code, and interpret experimental results reported in the literature for these genes of interest. Results are transformed to a common measurement so that different investigations, using different experiment designs, can be compared.

HOST:

Jose Castillo

DOWNLOAD: