Simulating cancer growth using cellular automata to detect combination drug targets

Abstract

Cancer treatment is a fragmented and varied process, as "cancer" is really hundreds of different diseases. The "hallmarks of cancer" were proposed by Hanahan and Weinberg in 2000 and gave a framework for viewing cancer as a single disease-one where cells have acquired ten properties that are common to almost all cancers, allowing them to grow uncontrollably and ravage the body. We used a cellular automata model of tumour growth paired with lattice Boltzmann methods modelling oxygen flow to simulate combination drugs targeted at knocking out pairs of hallmarks. We found that knocking out some pairs of cancer-enabling hallmarks did not prevent tumour formation, while other pairs significantly prevent cancer from growing beyond a few cells (p=0.0004 using Wilcoxon signed-rank adjusted with Bonferroni for multiple comparisons). This is not what would be expected from models of knocking out the hallmarks individually, as many pairs did not have an additive effect but either had no effect or a multiplicative one. We propose that targeting certain pairs of cancer hallmarks, specifically cancer's ability to induce blood vessel development paired with another cancer hallmark, could prove an effective cancer treatment option. © 2014 Springer International Publishing Switzerland.