Identification of candidate genes determining chemosensitivity to anti-cancer drugs of gastric cancer cell lines.

In order to efficiently develop improved cancer therapies it is important to predict the efficacy of anti-cancer drugs. In this regard, identification of genes that are related to drug sensitivity is vital. We previously established a panel of 39 human cancer cell lines (JFCR39) and a panel aiming for organ-specific analysis of 45 human cancer cell lines (JFCR45). Here, we focus on 20 human gastric cancer cell lines from JFCR45, a panel of human cancer cell lines to predict genes that determine chemosensitivity to anti-cancer drugs. We measured both chemosensitivity to a range of anti-cancer drugs as well as changes in gene expression profile. We then identified genes in which expression is related to chemosensitivity by using a Pearson correlation. As a result, anti-cancer drugs that have similar mechanisms of action showed similar fingerprints against a gastric subpanel of human cancer cell lines, as was the case with JFCR39 and JFCR45. Furthermore, we identified many candidate genes related to the sensitivity of gastric cancer cells to anti-cancer drugs.

Evaluation of action mechanisms of toxic chemicals using JFCR39, a panel of human cancer cell lines.

We previously established a panel of human cancer cell lines, JFCR39, coupled to an anticancer drug activity database; this panel is comparable with the NCI60 panel developed by the National Cancer Institute. The JFCR39 system can be used to predict the molecular targets or evaluate the action mechanisms of the test compounds by comparing their cell growth inhibition profiles (i.e., fingerprints) with those of the standard anticancer drugs using the COMPARE program. In this study, we used this drug activity database-coupled JFCR39 system to evaluate the action mechanisms of various chemical compounds, including toxic chemicals, agricultural chemicals, drugs, and synthetic intermediates. Fingerprints of 130 chemicals were determined and stored in the database. Sixty-nine of 130 chemicals ( approximately 60%) satisfied our criteria for the further analysis and were classified by cluster analysis of the fingerprints of these chemicals and several standard anticancer drugs into the following three clusters: 1) anticancer drugs, 2) chemicals that shared similar action mechanisms (for example, ouabain and digoxin), and 3) chemicals whose action mechanisms were unknown. These results suggested that chemicals belonging to a cluster (i.e., a cluster of toxic chemicals, a cluster of anticancer drugs, etc.) shared similar action mechanism. In summary, the JFCR39 system can classify chemicals based on their fingerprints, even when their action mechanisms are unknown, and it is highly probable that the chemicals within a cluster share common action mechanisms.