A repetitive mutation and selection system for bacterial evolution to increase the specific affinity to pancreatic cancer cells.

It is difficult to target and kill cancer cells. One possible approach is to mutate bacteria to enhance their binding to cancer cells. In the present study, Gram-negative Escherichia coli and Gram-positive Bacillus subtilis were randomly mutated, and then were positively and negatively selected for binding cancer vs normal cells. With repetitive mutation and selection both bacteria successfully evolved to increase affinity to the pancreatic cancer cell line (Mia PaCa-2) but not normal cells (HPDE: immortalized human pancreatic ductal epithelial cells). The mutant E. coli and B. subtilis strains bound to Mia PaCa-2 cells about 10 and 25 times more than to HPDE cells. The selected E. coli strain had mutations in biofilm-related genes and the regulatory region for a type I pilus gene. Consistent with type I pili involvement, mannose could inhibit the binding to cells. The results suggest that weak but specific binding is involved in the initial step of adhesion. To test their ability to kill Mia PaCa-2 cells, hemolysin was expressed in the mutant strain. The hemolysin released from the mutant strain was active and could kill Mia PaCa-2 cells. In the case of B. subtilis, the initial binding to the cells was a weak interaction of the leading pole of the motile bacteria. The frequency of this interaction to Mia PaCa-2 cells dramatically increased in the evolved mutant strain. This mutant strain could also specifically invade beneath Mia PaCa-2 cells and settle there. This type of mutation/selection strategy may be applicable to other combinations of cancer cells and bacterial species.