DNA checkpoints: Hercules in the Hydra's den

Many of you will be familiar with an old argument in biology, that of the unit of selection. This is one of those arguments where each side pours in far more effort than the controversy merits, and where half the problem is caused by individuals working in closely related fields having slightly different perspectives on the situation. Molecular and population geneticists, for example, define “gene” differently. In the case of the unit of selection, the persistence of the argument is aided by the presence of Richard Dawkins, who seems to prompt automatic disagreement from certain quarters. In his first book, The Selfish Gene, Dawkins made the case for the gene as the unit of selection, rather than the individual. Natural selection favours genes that are good at maximising their copy numbers, through any of a wide variety of strategies. Genes and natural selection have no foresight, and therefore the good of the individual, population or species are not always considered.

I was reminded of this when reading Wodarz and Komarova's review, “Can loss of apoptosis protect against cancer?”^[1] on the train the other day. It's a review paper, so it's not quite so dry and technical as to be unreadable by those without a relevant degree. It's interesting enough to warrant a summary here though.

 

Much of the cell's content does not have to be present in exact quantities: during cell division, the cell expands, increases the number of mitochondria, and other such organelles, and then splits down the middle roughly symmetrically. The compliment of DNA, however, must (with a few exceptions that we can ignore in this story) be the same in all of the cells in the body. We must therefore exactly duplicate all of the DNA, and make sure that one copy, and one copy only, ends up in each daughter cell. The process by which the DNA is shared out is mitosis (or, in the special case of sperm and eggs, meiosis), a process that is roughly the same throughout eukaryotes (animals, plants, fungi, etc).

Between the DNA duplication and mitosis there is a pause while the DNA is “proofread” or passed through a series of “checkpoints”. Minor errors and damage are repaired, but if something more serious has occurred, the cell commits suicide: “apoptosis” or “programmed cell death”. Many of the gene variants affecting one's predisposition to cancer that you hear about on the news are components of this error detection and repair system. A variety of mutations may cause cancer (the initial event is known as “transformation”), and further mutations accumulate as the cancer progresses. These can be classified by the molecular systems that they disrupt, the most important of these being the cell cycle. In a cell line in which the proofreading machinery has been lost, more mutations will be maintained per cell division.

Cancer makes a lot of sense when viewed from the point of view of the selfish gene, and natural selection. Genes and selection are, after all, lacking in foresight. Most of our cells are part of dead-end lines. The only cells in our body with any potential long-term future are the gametes – sperm or eggs. The rest of our cells cooperate and allow the gametes their privileged position as a result of millions of years of selection events. However, as long as there is a population of limited size that reproduces with high but imperfect accuracy, there will be selection. Each (or most) population of specialised cells in the body conforms to these conditions. The success of an individual in such a population is determined by its relative fidelity, fecundity and longevity. Fidelity must be high, to prevent the creation of too many non-viable offspring, but not necessarily perfect (after all, how else would the variation arise in the first place?). When the population is limited, there are two strategies for trying to occupy the maximum number of positions: pour your efforts into keeping the positions you get as long as possible; or producing lots of offspring in the hope that some of them succeed in finding a vacant position, or capturing an occupied one. At the organism scale, humans have gone for the former – producing a powerful and versatile brain, and a durable body with immune systems, wound healing systems, and so on, in order to hang on to their spot while they produce a small number of offspring. Mayflies have gone for the latter, producing huge numbers of stupid and delicate offspring in their short lives.

Tumours have limited sizes, and therefore limited populations of cells: just like humans and mayflies, tumours are limited by space and available nutrition. Tumours grow until they use up this space, or run out of nutrition, and then stop growing. Some then go on to metastasise, leading to secondary tumours, analogous somewhat to an insect colony producing a daughter nest, or the spread of humanity out of Africa; others manipulate blood vessels to gain new supplies, and therefore grow larger, analogous perhaps to the adoption of agriculture.^[2] Others just stop growing and do nothing for years. The route the tumour takes depends on whether it happens to acquire a random mutation that allows it to move on to one of the next stages.

This is where the checkpoints come into the story. In the tumour, a side effect of the checkpoints is to keep the tumour size less than maximum, reducing the limit on fecundity imposed by the limits on space and nutrition: every time the checkpoint triggers cell death, a new cell division can occur. The more cell divisions occur, the more mutations occur, and the greater the chance that the tumour will acquire a mutation that enables it to metastasise, or to manipulate the blood vessels. So, while loosing the checkpoints allows more mutations to occur per cell division, the overall mutation rate may fall. If your cells are healthy, you don't want to be without your checkpoints; but if they turn bad, hope that the checkpoints are the first to go.

Footnotes

1. ^  Dominik Wodarz and Natalia Komarova, 2007. Can loss of apoptosis protect against cancer? Trends in Genetics 23 232-237. 10.1016/j.tig.2007.03.005.
2. ^  Bad analogies: the staple of science communication.