Between 1970 and 1994, cancer claimed the lives of about 500,000 Americans every year. According to the most recent statistics, it is estimated that 1.3 million new cases of cancer will be diagnosed and 555,500 people will die from cancer in the United States in the year 2002. Cytotoxic and cytostatic molecules designed to disrupt the DNA of cancer cells or its transcription have long been used to treat this disease despite their often serious side-effects. The identification of new targets at the genomic level and therapeutically-relevant leads to increase their selectivity would therefore be of major clinical utility for oncologist. In our recent dossier "Histone deacetylase inhibitors: Redefining pharmaceutical approaches to the treatment of cancer" we report a second highly exciting approach. The consequences of DNA damage are likely to greatly depend on damage location in the genome. It may be sufficient to hit a few critical regions to kill cancer cell. Unfortunately, the majority of clinical DNA-reactive anticancer drugs show little discrimination in their DNA binding. Damage to non-crucial regions may contribute little to anticancer effects but is likely to enhance adverse effects in normal tissues. Dr Jan Woynarowski of the University of Texas Health Science Center at San Antonio and colleagues, show for the first time that small molecular weight drugs (bizelesin and analogs) are capable of region-specific damage to cellular DNA. These studies led to the identification of islands of AT-rich DNA with the properties of Matrix Associated Regions (MARs) as a promising new target. AT islands seem to be a critical target, since less than 10 DNA bizelesin lesions/cell are sufficient to block cell growth. Adducts formed by AT-specific, but non-region specific drug, tallimustine, are markedly less lethal (~200/cell at 50% growth inhibition). Non-region specific classical drugs need to form several thousands lesions/cell to inhibit cell growth. These data suggest that AT islands, being inherently genetically unstable minisatellites, can vary in their organization and abundance among various tumors and between tumor and normal cells. This variability includes AT-rich fragile sites, domains whose instability (expansion) has been implicated in various cancers, in particular leukemias. These researchers believe that targeting AT islands, or their specific subsets, may provide a novel unexplored basis for tumor versus normal cell selectivity. Furthermore in the course of these studies, they have established an original approach to identify cellular targets for DNA-reactive antitumor drugs at the genomic level. This approach allows the prediction of region specificity of various agents based on long-range analysis of their binding motifs in silico. Quantitative PCR stop assay, which is used to monitor the levels of damage in individual regions of cellular DNA, and other methods verify the reliability of the in silico predictions. These predictions can suggest, for instance, new analogs that would be even more region-specific for AT islands than bizelesin. Examining region-specificity of DNA damage at the genomic level can also shed new light on the molecular basis of cell killing by other DNA-reactive drugs. Importantly, the developed in silico bioinformatics approach to identify drug targets and optimize drug binding motifs is applicable not only to cancer but also to targeting the genomes of pathogenic microorganisms.

Interested in collaborating with this group? Contact leaddiscovery@bioportfolio.co.uk