Today I will speak about the relationship between game theory and cancer. Specifically, I will talk about why we need game theory in cancer research.

The first question I usually get when I introduce my research is: How are these two fields connected? They are researched by very different people, with very different mindsets, using very different methods, so how can we make a connection?

In this talk, not only do I claim that a connection exists, that game theory can be relevant as a means to understand cancer, I claim that it is useful as a tool to combat cancer. Even stronger, I am going to convince you that we can’t afford not to introduce game theory into how we view cancer before we can cure it. Except a miracle happens, and cancer is cured tomorrow, but, you don’t have to worry about that.

First, let me introduce game theory. Game theory is the mathematical discipline of decision making. It has three main components, players, let this be you against everyone else in the room; strategies, in my story all of us can be either cooperative or selfish, and payoffs. Payoffs tell you how much you stand to win in each possible outcome. If you are an economist with no imagination, you can think of payoffs as money. If you’re more into psychology, you can call it happiness. In my story, and in most biological applications, we view these numbers as a measure of your ability to reproduce. These things are of course not independent, money may buy happiness, happiness may cause you to be more eligible for reproduction, and so on.

For the model it doesn’t matter how you interpret the numbers. You don’t even have to like the exact numbers I put up here, the only thing that matters for my story is that a cooperative population (Luke and the rebels) is doing better than a selfish one (Boba Fett and Jabba’s goons), being selfish when everyone else is cooperative is fantastic (Han Solo showing up for the last ten seconds of the battle against the Death Star and gaining half the credit for its destruction) while being cooperative when everyone else is selfish is terrible (Princess Leia chained to Jabba the Hutt and not exactly thinking about raising a child).

Once we have our game we can talk about how the population evolves. You, as the individual, will have incentives to either be cooperative or selfish. Morals, altruism, our institutions, and all that other obscure stuff that my behavioral colleagues research will make you more cooperative. The advantage you get from being selfish of course makes you more selfish. If everyone is playing the same game, then these effects work for everyone in the population, making the whole population more cooperative or selfish. Intuitively, if I increase the advantage you get from being selfish, both you and the whole population will become more selfish.

Now, let's talk about cancer. The very simplified story of how you get cancer starts in your normal cells. As a cell duplicates, its version of your DNA, called the genome, is copied to produce two cells. The copying process is imperfect, which leads to errors. If the errors are superficial or happen on a part of the genome which the cell doesn’t use, it has no effect on the daughter cell. If the errors are too severe, the daughter cell is destroyed immediately. Rarely, however, the errors are not grave enough to kill the daughter cell but enough to change the way the cell functions, usually for the worse. This is called mutation. The new cell may not be able to fulfill its intended function. It may not understand the instructions that your body is sending. Crucially, these mutations are fixed in the new cell’s genome, so when it multiplies, the errors are inherited. If cells accumulate enough mutations, they become cancerous. They keep multiplying when the body tells them to stop, they consume more nutrients than they need and process them in a way that releases toxins, producing tumors. If the tumors become malignant, the cancer spreads through your body, causing your organs to fail as your normal cells to literally starve to death. This is called metastatic cancer, and our ability to treat it has not increased significantly since the 1970s. If the disease reaches this point, the cancer cells’ genome is irreparably broken so it will undergo more and more mutations faster than your normal cells.

Mutations that lead to cancer can occur randomly for no reason at all, they may be caused by our own crappy habits, they can be genetic, or they may occur because you’re old and your cells’ genome is already damaged. We differentiate types of cancer by which one of your original cells mutated. The most common and deadly cancer types are lung cancer, some subtypes of which are caused by cigarette smoke, prostate cancer, which affects up to 50-80% of all men, breast cancer, whose awareness has seen a much needed, massive increase in recent years, and colorectal cancer.

One of the most dangerous aspects of cancer, what makes it really hard to treat is its fast rate of mutation thanks to its broken genome. As many of you know, the standard therapy for many types of metastatic cancer is chemotherapy. Chemotherapy is a blanket term for a number of cytotoxic drugs. Cyto means cell, so this drug is bad for cells. Not just cancer cells, all types of cells. If you know a person treated by chemo, you know that it is miserable. The drug builds up in cells that multiply quickly, like cancer, so you may lose your hair, you may lose your fingernails, and you generally feel terrible. Cancer cells respond by mutating again, developing little pumps on their surface, called transmembrane pumps. They use these pumps to empty the chemo and dump it outside into the environment. What we typically see in chemotherapy is a good initial response, since the drug is killing most of your cancer cells that do not yet have the pump. However, once those are dead, the cancer will come back, and this time all cancer cells have the pumps, so chemo becomes ineffective.

Since mutation happens in everyone, if this was the end of the story, we would all have cancer and it would be incurable. Fortunately, our body is able to destroy cancerous cells. This is done by our immune system, which I represent by blue colored police cells, called killer T-cells. Normally, killer T-cells seek out and kill cancer cells in a process we call immune reaction.

Why can you still get cancer, despite your immune system? Mutation again. The cancer cells may mutate to evade destruction by the T-cells. They do this by two ways, the same two ways that you or I would use when being chased by the police, or by zombies, or by my PhD supervisors. Cancer cells can either hide or fight back.

Hiding is pretty straightforward. The cancer cell just puts on a mask and pretends to be a normal cell. They can do this by manipulating their surface proteins in a way that when the killer T-cell makes contact, it mistakes it for a non-cancerous, normal cell.

Fighting is a more complicated story. If you are allergic, then you know that sometimes your immune system goes on a rampage and starts fighting your own body. What can happen is that you killer T-cells go wild and start killing everything they meet, mostly your normal cells. This is called an autoimmune reaction. If this starts happening, your normal cells call the bosses of your immune system and tell them that there’s a rampage going on. The bosses are called T-regulator cells and their only job is to protect you against autoimmune diseases. Once they receive the call, they go around and shut down your killer T-cells, and that stops the rampage. This is called immunosuppression. Cancer does something so evil, so sinister, that it is kind of awesome: they develop the ability to call the bosses and convince them that there’s a rampage going on. The T-regulator cells then shut down the immune system which was just doing its job stopping cancer. This strategy is what I call fighting.

In cancer therapy they try to overcome these strategies by a new, cool type of therapy called the immunotherapy of cancer. Specifically, I’m focusing on a therapy that allows your immune system to recognize and kill the cancer that is in hiding. This idea is so new and so cool that the masterminds, Allison and Honjo, got a Nobel Prize for it a month ago. In some cases, this treatment works like a miracle, and every news outlet celebrates that we finally “cured cancer”. Sadly, in some much less reported cases, immunotherapy fails spectacularly for seemingly no reason. We simply have no idea why these therapies work or don’t work. Almost certainly, the problem is that cancer picks up some new mutations that we didn’t prepare for and that kills the patient.

So, the obvious question is, why don’t we design a therapy that keeps these mutations in mind? This new paradigm is called evolutionary enlightenment, developed by Bob Gatenby. This is not about inventing new drugs or new forms of therapy, this is about paying attention to cancer’s evolutionary dynamics. The tool we use to anticipate mutations, the only tool we can use, is game theory.

In my work we build a game from the cancer cells’ strategies, hide and fight. Take a tumor of cells who are all hiding. These cells are doing fine, but they have to contend with a healthy and strong immune system. A tumor made out of fighters, who all cooperate to suppress the immune system is going to do much better. But here’s the catch:

If I add a fighter into a tumor made out of hiders, that cell is going to get killed in a millisecond. It isn’t hiding like all the others, and it has no hope of fighting the immune system by itself. However, if I add a hider into a tumor of fighters, that cell is going to do fantastic! Not only is the immune system weak thanks to all the others, it even has additional protection.

So, we’re playing the same game I showed you in the beginning. We have a cooperative strategy, fight, and a selfish one, hide. A population of fighters is doing better than a population of hiders, but individually, it is always better to hide than to fight. It’s clear that cancer cells who are hiding have an advantage in reproduction. Now comes the scary part. If you introduce the therapy I talked about, your immune system becomes better at finding the cells who hide. You make hiding less effective. You introduce a private incentive for the cancer cells to fight back. All cancer cells are playing the same game, so, as we discussed earlier, the same will happen on the whole population. Therapy makes your tumor more cooperative, and you can end up making the tumor worse. Simply by not paying attention to the evolutionary dynamics.

So, why should we use game theory against cancer? I listed three reasons. First, it helps with treatment selection. In my game we find that it is much better to treat against the fighter cells, than the hiding cancer cells. Second, game theory determines the conditions in which I need to apply treatment. For example, if the immune system is strong, we should treat against hiders, otherwise probably should not. Finally, game theory gives us the tools to anticipate how the tumor will behave after treatment. It is important to know that cancer will always adapt faster than the therapy we give to the patient, so we simply cannot afford not to prepare for the adaptive dynamics when we start the therapy.

Thank you for reading. I would like to extend my gratitude to the other three speakers, one of the chief organizers, Martijn Stroom, and to my wife Nóri for feedback on the presentation. I thank my wonderful coauthors, Joel and Katerina, for introducing me to the world of cancer research and writing the paper with me, and my wonderful sister, Ada, for preparing the illustrations on the slides.

December 2, 2018.

Link to the paper: https://www.sciencedirect.com/science/article/pii/S0022519318303382.

Ada's other artwork on Instagram: https://www.instagram.com/clockwork.cherry/.

Comments should be addressed to peter.bayer7@gmail.com.