Game Theory in Tumor Microenvironment
It is the conventional wisdom for chemical and biomedical engineers to think of the human body as a factory when it comes to absorbing nutrients or substances and outputting products, all the while using signals to communicate and decide how to do those necessary chemical conversions on the microscopic/cellular scale. This resulted in the extensive study of signaling and blocking the transmission and/or reception of these signals. When it comes to cancer, the same approach had taken place, using chemotherapy to destroy existing cancerous cells (and regular (stromal) cells through collateral damage) and monoclonal antibodies to block receptors to prevent future growth. However, it’s a paradigm shift to consider cancerous cells and stromal cells as a population with survival/evolutionary game theory on the cellular scale, as this paper suggests, in contrast to the a priori approach as signaling circuits.
Using microfluidic technology, the authors have created a drug gradient to test the fitness and proliferation of cancer cells vs. time. Specifically, a population gradient was created, creating local competition for space and metabolic resources such as glucose or oxygen. This drives the cells to head towards the drug source, where the strongest cancer cells then develop drug resistance. In essence, this competition creates drug resistant cells. Along this population gradient, a “death galaxy” is formed in which connected microhabitats accelerate emergence of drug resistance, and through this galaxy, the dynamics of cancerous and stromal cell microenvironments were studied under simulated chemotherapy treatment.
They treated the system as a well-mixed population and evaluated the model as a spatial-independent evolutionary game model. They created the same matrix that we use in class, with the players being cancerous cells (CC) and stromal cells (ST). The combinations were CC vs. ST, CC vs. CC, ST vs. CC, and ST vs. ST. Each player on the left of those combinations had a payoff value, but this paper was trying to determine what these payoffs were, specifically in a tumor microenvironment under chemotherapy conditions. The results they found were surprising in terms of the payoffs and differed greatly from the previous publications on stromal and cancerous cells.
As expected, in ST-rich conditions, ST cells form a population gradient that is inversely proportional to the drug gradient (more drugs, less ST cells) and CC cells also dropped linearly. Under CC-rich conditions, the CC count initially increases for a few days with the idea of garnering more payoffs in an attempt to stave off toxins and starvation from the chemotherapy, but ultimately decline. The surprising part is what happens at low dosage of chemotherapy. It was believed a priori that ST cells help promote cancer survival and proliferation by blocking signals to trigger the death of CCs. However, at the low dose/concentration of the drug, there is active competition between the CCs and STs, which means there is cooperation or competition depending on the ST population levels relative to the CC population. Specifically, cells need energy to pump out the chemotherapy drug to survive. Thus, at low concentrations, the two were competing for nutrients to provide that aforementioned energy, in addition to the heavy crowding of that environmental space.
Taking this back to pure game theory, if the CC and STs were independent of each other, natural selection would favor the type of cell that was more resistant to the therapeutic drug. However, as this paper shows, it’s more complex than that as CC becomes less sensitive to the drug when grown in proximity to STs. This makes the ratio of the STs to CCs the most powerful force that causes the switching between the competition and cooperation modes of the microenvironment. Also, as the drug concentration varies, so would the payoffs, which was a linear gradient in proportion to the linear drug gradient. Thus, in higher drug concentrations, ST cells grow more slowly while only the most resistant CCs survive and the weaker ones die off; again both are maximizing their respective payoffs (STs to be able to adapt and keep growing under the protection of the most powerful CCs while STs block cell death signaling in the CCs) and both cooperate. However, at low drug concentrations, this mechanism changes as the payoffs are somewhat similar between the STs and CCs, thus resulting in competition between the two for nutrients to pump out the toxins in the drug. Therefore, evolutionary game theory accounts for the ratios of the two cell types and the frequency of their contacts with each other, thus making the cancerous microenvironment another situation to which game theory can be applied.
link: http://rsfs.royalsocietypublishing.org/content/4/4/20140028.full.pdf+html