Abstract
We have engineered yeast cytosine deaminase (CD) to be activated by the tumor marker, hypoxia‐inducible factor 1‐alpha (HIF‐1α) ‐ turning on its ability to convert 5‐fluorcytosine (5FC) into the anticancer drug 5‐fluoruracil (5FU). HIF‐1α is overexpressed in many types of cancer but is virtually undetectable in normal tissue due to rapid degradation. We sought a “tumor‐sensing” protein whose prodrug converting activity is activated in the presence of HIF‐1α. Such a protein could be used in a therapeutic treatment similar to gene‐directed enzyme prodrug therapy (GDEPT). However, our approach would have an advantage over GDEPT because the “suicide” gene could be delivered using very efficient methods for gene delivery, whether or not they were selective, since the selectivity occurs at the protein level and not the gene delivery level.
To create our desired protein, we constructed a library of fusion proteins between CD and the CH1 domain of p300 ‐ a domain known to interact with HIF‐1α. This library was subjected to a two‐tiered genetic selection in E. coli designed to identify genes that conferred increased sensitivity to 5FC when HIF‐1α was coexpressed. We identified two library members (#3 and #59), which conferred to E. coli sensitivity to 5FC equivalent to wildtype CD but only in the presence of coexpressed HIF‐1α. In the absence of HIF‐1α, #3 conferred a 5‐fold higher LC50 to 5FC and #59 did not confer any sensitivity to 5FC. Both proteins were purified and characterized in vitro for their activity in the presence and absence of HIF‐1a. Protein #3 exhibited a 4‐fold increase in its CD activity in the presence of HIF‐1α, whereas protein #59's activity did not have a significant dependence on HIF‐1α. Currently, the two genes are being examined in human cancer cell lines for their ability to confer sensitivity to 5FC in the presence of HIF‐1α.
Using advanced protein engineering techniques, we have created two proteins that confer sensitivity to 5FC in the presence of the tumor‐marker, HIF‐1α. We believe that these genes have therapeutic potential for the treatment of cancers that exhibit elevated HIF‐1α levels. Our strategy circumvents a limitation of GDEPT by not requiring selective delivery of the gene to cancer cells.
Citation Information: Mol Cancer Ther 2009;8(12 Suppl):C245.