The selectivity problem
When chemists design an antibiotic, they exploit selective toxicity: the bacterium is a separate organism with machinery you simply do not have, so a drug can strike the microbe and leave you untouched. Cancer offers no such gift. A tumour is built from your own cells, carrying almost the same genes, the same enzymes, the same targets. The agonising question behind every cancer drug is therefore not just how do I kill this cell? but how do I kill *this* cell and spare its near-identical neighbour?
Because tumour and host overlap so heavily, the therapeutic window in oncology is famously narrow. Many cancer drugs are dosed close to the level that makes patients sick, and the side effects we accept — hair loss, nausea, low blood counts — are the price of that narrow window. Every advance in this field can be read as one long campaign to widen it: to find some difference, however small, between the cancer and the rest of you, and to aim a molecule at exactly that difference.
Two strategies, one goal
Historically the field split into two philosophies. The first is cytotoxic chemotherapy: hit anything that divides fast. Cancer cells multiply relentlessly, and the classic agents poison cell division itself — they damage DNA or jam the machinery of replication. The selectivity here is crude. It rests only on the fact that tumours divide faster than most healthy tissue, which is why the tissues that *also* divide fast — hair follicles, gut lining, bone marrow — take collateral damage.
The second philosophy is targeted therapy: find a specific molecular *cause* of this cancer and block that. Many tumours are driven by a single broken protein — a mutated kinase stuck in the on position, a fusion gene, an overactive receptor. If the cancer leans on that one crutch to survive, a drug that knocks the crutch away can be devastating to the tumour and gentle on everything else. This is the dream of mechanism-based design, and it is where most of modern cancer chemistry now lives.
Killing on purpose: apoptosis
A good cancer drug rarely blows the cell apart. Instead it pushes the cell to kill *itself* through apoptosis — a tidy, programmed self-destruct that every cell carries built in. Healthy cells trigger apoptosis when they are damaged beyond repair; cancer cells survive precisely because they have learned to mute that signal. Many oncology drugs work by re-arming the suicide switch: enough damage or enough stress, and the tumour cell finally does what a normal cell would have done long ago.
Keep this frame in mind for the rest of the track. Whether a drug shatters DNA, freezes a kinase, or drags a protein to the shredder, the endpoint is usually the same: tip a cancer cell over the edge into apoptosis, while leaving as many normal cells as possible standing on safe ground.