Cells that defect on the body

Why does a body of cooperating cells keep producing cells that destroy it for all of them?

Cancer is usually framed as damage: something broke, and the repair failed. That framing explains a mutation. It does not explain why the failure takes this particular shape — a cell that grows faster, ignores stop signals, refuses to die, and consumes more than its share — rather than a thousand other ways a cell could malfunction.

So try a different question. What exactly is a body asking of its forty trillion cells, and what would it look like for one to stop complying?

List what a cell must give up to be part of a body. It must divide only when the tissue calls for it, even though it is fully equipped to divide continuously. It must kill itself on command when it is damaged or surplus. It must do one specialised job and forgo others. It must take its share of glucose and oxygen rather than all it can get. And it must help maintain the shared surroundings — the matrix, the local chemistry, the blood supply — rather than remodel them for itself.

Every one of those is a restraint on the cell for the benefit of the whole. Multicellularity is therefore not merely an arrangement of cells; it is a cooperative arrangement, and cooperative arrangements have a characteristic vulnerability. Ask the classic question of any cooperative system: what happens to the individual who takes the benefits and declines the restraint?

It does better. A cell that divides without waiting outgrows its neighbours. A cell that ignores the suicide signal outlasts them. A cell that recruits blood vessels toward itself is better fed than they are. Now note where this happens: inside the body, among cells that are dividing, mutating, and being differentially culled. That is a population under selection, and selection has no preference for the body's interest — it favours whatever lineage leaves more descendant cells this month. The cheat is not a malfunction of evolution but an instance of it, running at the wrong level.

That reframes the puzzle entirely. It was never "why does cancer arise", but "why does it not arise constantly", given that the payoff for defecting is available to every cell every day. The answer is that a body is thick with countermeasures: checkpoints that halt division when DNA is damaged, forced suicide programmes, limits on how many times a lineage may divide, immune cells that patrol for the altered, and tissue architecture that physically confines a lineage to its compartment. These are policing mechanisms, and they are why a cheat needs several independent mutations rather than one.

Then ask the sharper evolutionary question: why is the policing not perfect? Two reasons, both honest. Selection acts through reproduction, so defences are shaped mainly by what happens before and during reproductive life; failures that appear at seventy are nearly invisible to it. And the defences cost something — suppressing division too aggressively impairs wound healing and tissue renewal, so the setting is a trade-off rather than a maximum. There is good evidence the dial can be turned further when it pays: cancer incidence does not rise with body size across species as naive expectation says it should, a discrepancy known as Peto's paradox, and elephants carry roughly twenty copies of the tumour-suppressor gene TP53 where we carry one. Exactly how much of the paradox that explains is still an open research question.

Three refinements. First, "cheating" is a description of the behaviour's structure, not an attribution of intent or of malice. Nothing in the cell wants anything; the word simply marks that the trait benefits the lineage carrying it at the collective's expense.

Second, the dead-end point has real exceptions, and they prove the framing rather than undermining it. Where a cancer lineage can reach a new host, it stops being a dead end and becomes a parasite in its own right: the facial tumour disease transmitted between Tasmanian devils by biting, the venereal tumour passed between dogs for thousands of years, and transmissible leukaemias in shellfish are all cell lineages that outlived their original bodies.

Third, this explains a clinical fact that pure damage-thinking cannot. Because a tumour is a population under selection, treatment is a selection pressure, and killing the sensitive cells hands the survivors an emptier field. Resistance is not bad luck; it is the expected result of applying strong selection to a diverse population — which is the reasoning behind adaptive therapy, which deliberately holds a tumour in check rather than pushing for maximum kill.

• Peto's paradox: how large, long-lived animals actually suppress cancer, and what medicine might borrow from them.
• Adaptive therapy, and the evidence for treating a tumour as a population to be managed rather than eradicated.
• The other transitions in individuality — genes into genomes, cells into bodies, insects into colonies — and the cheats each one had to suppress.

Every term the collection defines is gathered in the glossary.