Cancer: When Cell Division Loses Its Brakes

The most misunderstood diagnosis#

Say the word "cancer" and most people picture something alien — an invader that arrives from outside, takes hold, and spreads. Almost every part of that picture is wrong, and the correct picture is stranger and, oddly, less frightening once you see it clearly.

Cancer is not one disease. It is an umbrella term for hundreds of them, and they have in common a single broken behaviour: a cell of your own body dividing when it should not. Not a foreign organism, not a poison, not a thing that got in. Your own cell, running your own DNA, doing the one thing cells are built to do — copy themselves — with the controls that normally regulate that copying no longer working.

To see how those controls fail, we first have to see what they are.

The cell cycle and its checkpoints#

Every time one cell becomes two, it runs through an ordered sequence called the cell cycle. It grows (G1), copies its entire genome (S phase, a feat of DNA replication across three billion letters), grows a little more and checks its work (G2), and then divides (M phase, mitosis). Left alone, a cell does not simply race around this loop. It stops at gates.

These gates are the checkpoints, and they are the reason your body is not a tumour. At the G1 checkpoint the cell asks: is my DNA intact, do I have the resources to divide, and has anything told me to hold? At the G2 checkpoint it asks: did I copy everything correctly? A cell that finds a problem does one of two things. If the damage looks fixable, it pauses and calls in the repair machinery. If the damage is too severe, it triggers apoptosis — programmed cell death, a clean and deliberate self-destruct that removes the cell before it can pass its errors on.

This is worth sitting with, because it inverts the usual intuition. Your cells acquire DNA damage constantly — from ultraviolet light, from ordinary metabolism, from the sheer statistics of copying billions of letters. The remarkable thing is not that damage happens; it is that almost all of it is caught and either repaired or deleted. Damaged cells being destroyed is the system working, not failing.

Two kinds of gene run this control system, and it helps to think of a car. There are accelerators — genes that say "divide now." In their normal form these are called proto-oncogenes, and they are essential: growth, healing, and development all depend on them. And there are brakes — tumour-suppressor genes like TP53 and RB1 that halt the cycle for repair, or order apoptosis when repair is hopeless. A healthy cell divides when the accelerators are pressed at the right moment and the brakes are released, and stops otherwise.

Why one mutation is not enough#

Here is the misconception that does the most damage: that cancer is a single disease caused by a single mutation, one unlucky hit and it is over. It is not, and understanding why is the key to the whole subject.

A car does not run away because you tap the accelerator once. To lose control you need the accelerator stuck down and the brakes cut. A cell is built the same way, deliberately, with redundancy. To become cancerous a cell lineage typically has to accumulate several mutations that cooperate: some that jam the accelerator — turning a proto-oncogene into a permanently-on oncogene — and others that cut the brakes, disabling tumour-suppressors like TP53 and defeating apoptosis. This is the multi-hit model. A stuck accelerator alone usually just triggers the brakes: a cell dividing when it should not is exactly what a checkpoint is built to catch, and TP53 will often send such a cell to apoptosis. Only when the brakes are also gone does the accelerator win.

If becoming cancerous requires several independent rare events landing in the same cell lineage, we can reason about the odds. Suppose each required mutation has a small probability $p$ of occurring in a given cell over a given time, and suppose $k$ of them are needed. Very roughly, the chance that one lineage has collected all of them scales like

$$P(\text{cancer in a lineage}) \sim p^{\,k},$$

a tiny number when $k$ is, say, five or six. But you have tens of trillions of cells dividing over a lifetime, and $p$ itself rises with every carcinogen and every division that copies an existing error. Multiply a minuscule per-lineage probability by an enormous number of lineages accumulating hits over decades, and the steep rise of cancer risk with age falls straight out of the arithmetic. Cancer is largely a disease of time because collecting a full set of cooperating mutations takes time.

Your own cells, and why cancer is not contagious#

The second great misconception is that cancer cells are foreign, or that cancer is something you can catch. In humans it is, with vanishingly rare exceptions, neither. A cancer cell carries your DNA — a corrupted copy of it, but yours. It is not a bacterium or a parasite. This is also why cancer is not contagious: there is no organism to transmit, and another person's immune response would in any case recognise transplanted cells as foreign and destroy them. (Some viruses, such as HPV, can raise cancer risk by damaging these control genes over years — but it is the accumulated genetic damage that causes cancer, not a transmissible cancer itself.)

So if the cell is your own and starts from a single ancestor, why does a tumour become such a formidable, shifting thing? Because what happens next is evolution by natural selection, running inside your body.

The ingredients Darwin needs are all present in a growing mass of dividing cells. There is variation: every division risks new mutations, so the population is not uniform. There is heritability: a daughter cell inherits its parent's mutations. And there is differential reproduction: a cell that happens to divide a little faster, or ignore a "stop growing" signal, or slip past the immune system, leaves more descendants than its neighbours. Selection does the rest. Over many generations the faster, tougher lineages come to dominate — not by design, but because they simply out-divide the others.

This is why tumours are heterogeneous — a patchwork of competing clones rather than a uniform lump. It is why they can trigger angiogenesis, coaxing the body to grow new blood vessels to feed them, and why some cells eventually acquire the ability to break away, travel, and seed new colonies elsewhere — metastasis, the property that makes cancer dangerous. Each of these is a trait that got selected because the cells carrying it left more descendants. A tumour is not a plan. It is what natural selection produces when it is pointed, accidentally, at your own tissue.

What this understanding buys us#

Framing cancer this way is not just tidy; it directs treatment. If cancer is division without brakes, then some therapies aim squarely at dividing cells — classic chemotherapy hits cells in the act of copying and splitting, which is why it also affects fast-renewing healthy tissues and produces familiar side effects. If a particular cancer is driven by one specific stuck accelerator, a targeted therapy can be designed to block that exact oncogene's protein. And because a tumour is a population your immune response is supposed to police, immunotherapies work by releasing the brakes the tumour put on your own immune cells, letting them recognise and attack the cancer.

The evolutionary view also explains treatment's hardest problem. If you kill 99% of a heterogeneous tumour, you may have simply selected for the resistant 1% — the same logic that drives antibiotic resistance in bacteria. This is why combination therapies, and strategies that try to manage rather than maximally attack a tumour, are areas of active research.

None of this is medical advice, and cancer biology is vastly richer than one article can hold. But the core idea is honest and, in its way, hopeful: cancer is not an unknowable curse or an alien invader. It is a comprehensible failure of a control system we understand better every year — and understanding the mechanism is exactly how the mechanism gets outmatched.