I will soon be attending a workshop in Barcelona entitled “Running to stand still” – Evolution and management of drug pesticide resistance in healthcare and agriculture. Most speakers will be scientists working on the control of pests and bacteria. As a result they have worked for some time on trying to control, rather than eradicate, an invading population in an ecosystem. We have learned that there are a lot of lessons (and mathematical tools) that cancer researchers could borrow from these other communities. But also that there are also important differences in the populations we want to control. This is mainly as a result of pests and bacteria being foreign and cancer cells being our own cells and usually difficult to unambiguously single out from their well-behaved sisters.
Paul Neve and Silvie Huijben, the meeting organisers, have asked us to come with a few questions that could be addressed in a short time-frame (5-10 years) to better understand the evolution of drug and pesticide resistance. These are some points, not quite questions, I have come up with:
- Heterogeneity. If you know Sandy Anderson you already know that intra-tumour heterogeneity drives resistance. We need to understand the heterogeneity in a patient’s tumour and how treatments impact this heterogeneity. A treatment can kill many cells and yet hinder the efficacy of later treatments. New technologies like CRISPR combined with mathematical modelling will help to figure this out.
- Microenvironment. Usually overlooked because of its complexity. But we know that some tumour phenotypes interact with the physical microenvironment and with normal cells to confer protection to themselves and other tumour cells. Drugs to prime the microenvironment for later treatment could improve the efficacy of treatments.
- Plasticity vs adaptation. How much of the change in a tumour population in response to a treatment is due to stochasticity, how much the treatment selecting against certain phenotypes and how much to phenotypic plasticity. The latter makes the problem more complex as tumour cells can effectively switch phenotypes in short time scales and are more flexible adapting to change.
- Cost of resistance. Controlling a heterogeneous tumour population is easier if resistant phenotypes have a lower fitness than non-resistant ones in the absence of the treatment. If neutral mutations confer resistance then there is little hope that we can stop or control resistant phenotypes.
- Mapping treatments to phenotypes. We know that treatments select for resistance but how else do they shape the tumour population? Can we predict how the population will change in response to a treatment? and can we use that information to schedule treatments to optimise outcomes?
Controlling an unwanted population is a great start. And given our success (or lack thereof) with certain types of cancer, a much more sensible approach that aiming to kill as many tumour cells as possible. But if you read between the lines in these questions you will find a pattern. That understanding the cancer as an ecosystem could enable us not only to control cancer but to engineer its evolution to our advantage.