The emergence of resistance to chemotherapy by cancer cells, when combined with metastasis, is the primary driver of mortality in cancer and has proven to be refractory to many efforts. Theory and computer modeling suggest that the rate of emergence of resistance is driven by the strong selective pressure of mutagenic chemotherapy and enhanced by the motility of mutant cells in a chemotherapy gradient to areas of higher drug concentration and lower population competition. To test these models, we constructed a synthetic microecology which superposed a mutagenic doxorubicin gradient across a population of motile, metastatic breast cancer cells (MDA-MB-231). We observed the emergence of MDA-MB-231 cancer cells capable of proliferation at 200 nM doxorubicin in this complex microecology. Individual cell tracking showed both movement of the MDA-MB-231 cancer cells toward higher drug concentrations and proliferation of the cells at the highest doxorubicin concentrations within 72 h, showing the importance of both motility and drug gradients in the emergence of resistance.
Cancer cells evolve drug resistance to chemotherapy within the tumor microenvironment. Although it is widely accepted that the tumor microenvironment provides a sequential selective pressure for preexisting mutants within the population (1-3), an additional contribution to rapid cancer evolution is mutagenic stress response followed by the emergence of adaptive phenotypes (4, 5). Further, mutagenic drug gradients in the tumor microenvironment lead to a spatially dependent fitness landscape of the cancer cells and can further accelerate the evolution of drug resistance if the cells are motile across the gradient (5, 6). We recently demonstrated using a bacteria model how a spatial gradient of antibiotic concentration in a metapopulation accelerated the evolution of antibiotic resistance (7). We would expect similar processes to occur in cancer cell metapopulations as well. Because cancer cells have a much longer doubling time (âŒ1 d) compared with that of bacteria (âŒ30 min), similar experiments with cancer cells take an order of magnitude more time (days vs. hours) than those for bacteria. This presents two experimental challenges: (i) creation of a drug stable gradient and (ii) creation of an environment hospitable for healthy cell growth. Once these conditions are established, it is possible to probe in an in vitro system the complex driving forces of resistance in systems that are in vivo.
ResultsMicrofluidic devices have become a versatile platform to provide precise concentration gradient control for understanding various biological systems and controlling the population size (8-10). Gradient-generating devices can be classified as (i) static generators, which are based solely on diffusion (11, 12), and (ii) constant-flow generators (13-16). In this paper we adopt the constant-flow approach because it is capable of creating timeindependent stable gradients. However, to date, it has been challenging to grow mammalian cells in such platforms (17,18)...