Appearance
1.8 — How Cells Talk, and How They Die on Command
A car pulls out in front of you. Within about a second your heart is beating harder, your airways have widened, your pupils have opened, blood has been redirected away from your gut and toward your muscles, and your liver has started dumping glucose into your bloodstream. You did none of that deliberately, and the organs involved are nowhere near each other.
What happened is that a gland above your kidney released a molecule into your blood, and cells all over your body that carry the right receptor read it and acted. The signal was tiny — adrenaline circulates at concentrations around a billionth of a mole per litre — and the response was enormous. How a whisper produces a shout is the first half of this page.
The second half is about the opposite instruction. Between 50 and 70 billion of your cells kill themselves every day, on command, quietly and without inflammation. That is not damage. It is maintenance, and a cell that refuses the order is on its way to becoming a tumour.
Four ranges of signalling
A signalling molecule is called a ligand — from Latin ligare, to bind, because binding is all it does. The receptor is what turns binding into action. Signals are grouped by how far they travel.
Endocrine signals are hormones released into the blood, reaching the whole body. Slow to arrive, long-lasting, and they act only on cells that carry the matching receptor. Insulin, thyroid hormone and cortisol are endocrine, and Part 12 covers all of them.
Paracrine signals act on neighbours — released, diffusing a short distance, and broken down quickly so they do not spread. Most inflammation works this way, and so do the growth factors that coordinate wound healing.
Synaptic signalling is paracrine taken to an extreme: a neuron releases neurotransmitter into a gap about 20 nanometres wide, so the message is delivered to one specific target cell in under a millisecond. Chapter 11.2.
Autocrine signals act on the cell that released them. Useful for a cell confirming its own state, and dangerous when a cancer cell starts producing its own growth factor and telling itself to divide forever.
Direct contact signalling needs no released molecule at all — a surface protein on one cell binds a receptor on another that is touching it. This is how a killer T cell verifies that the cell in front of it is infected before destroying it (Chapter 13.2).
Four kinds of receptor
Where the receptor sits determines everything about how fast the response is and how long it lasts.
1. G-protein coupled receptors — the largest family in your body
These receptors thread through the membrane seven times and are coupled on the inside to a G protein, so called because it binds guanine nucleotides — GDP when off, GTP when on.
Follow the cycle. The ligand binds the receptor's outer face. The receptor changes shape, and its inner face now grips a nearby G protein. The G protein releases the GDP it was holding and picks up GTP instead, which activates it. Its alpha subunit breaks away, slides along the inner membrane surface, and switches on an effector — usually an enzyme that makes a second messenger.
The switch-off is built into the switch. The alpha subunit is itself a slow enzyme that chops its own GTP back to GDP, after which it reassembles with its partners and goes quiet. The signal has a timer and does not need to be actively cancelled. That single design feature is why the cholera story of Chapter 1.4 is so destructive: cholera toxin chemically modifies one G protein alpha subunit so it can no longer destroy its own GTP. The switch never turns off, so the chloride channel never closes, and litres of fluid pour out of the bowel.
Humans have around 800 GPCR genes, and they handle a striking range of jobs: adrenaline, most neurotransmitters, many hormones, the sense of smell, and much of the sense of taste. Roughly a third of all approved drugs act on a GPCR. Beta-blockers, opioid painkillers, antihistamines, salbutamol inhalers and most antipsychotics are all in this category.
2. Enzyme-linked receptors, mostly kinases
Here the receptor is an enzyme. The commonest kind is a receptor tyrosine kinase: ligand binding makes two receptor molecules pair up, and once paired they attach phosphate groups to each other's tyrosine amino acids. Those phosphates become docking points for a queue of internal signalling proteins.
The insulin receptor works this way, which is why Chapter 18.7 on diabetes returns here. So do the growth factor receptors — and that is why they matter in cancer. HER2 is a growth factor receptor that is over-produced in about 15 to 20 percent of breast cancers, and the drug trastuzumab is an antibody that binds it and shuts it down. That drug turned HER2-positive breast cancer from the worst subtype into one of the more treatable ones, and it only exists because the receptor was understood first.
3. Ligand-gated ion channels
The receptor is a channel that opens when the ligand binds. These are the fastest — the response is a flow of ions within a millisecond — and they are how nerves and muscles work.
The nicotinic acetylcholine receptor at the junction between nerve and muscle opens to let sodium in, which is the trigger for every voluntary movement you make (Chapter 6.2). The GABA-A receptor in the brain opens to let chloride in, which makes a neuron harder to fire and therefore calms brain activity. Benzodiazepines such as diazepam, most general anaesthetics, and alcohol all act at this receptor, which explains both why they are sedating and why combining them is so often fatal.
4. Intracellular receptors
Steroid and thyroid hormones are small and fat-soluble, so they cross the membrane and find their receptor inside the cell. The hormone–receptor pair then travels into the nucleus and binds DNA directly, switching genes on or off.
This route is slow and long-lasting, because it works by changing which proteins get made — hours to take effect, and days for the effect to fade. That is exactly why a steroid inhaler does not stop an asthma attack in progress while a salbutamol inhaler does. Salbutamol hits a GPCR and relaxes airway muscle in minutes; the steroid is reducing inflammation over days. Chapter 22.10 covers why a person with asthma needs both and why confusing them is dangerous.
Amplification: how a whisper becomes a shout
Follow adrenaline into a liver cell and count.
One adrenaline molecule binds one beta-adrenergic receptor. While bound, that receptor activates not one G protein but tens of them. Each activated G protein switches on an enzyme called adenylyl cyclase, and each of those, for as long as it is on, converts many molecules of ATP into the second messenger cyclic AMP — hundreds of them. Each cAMP molecule helps activate protein kinase A, and each activated kinase attaches phosphates to many target proteins. Those targets include the enzyme that releases glucose from glycogen, and each of those enzymes then liberates many glucose molecules.
Multiply the steps and one adrenaline molecule causes the release of something in the order of a hundred million glucose molecules. Each stage multiplies, so the total is the product of the stages, not the sum. This is why hormones work at concentrations too low to measure by ordinary chemistry, and it is why a small change in hormone level produces a large change in the body.
Earl Sutherland discovered cyclic AMP and received the 1971 Nobel Prize. Alfred Gilman and Martin Rodbell received it in 1994 for the G proteins, and Robert Lefkowitz and Brian Kobilka in 2012 for the receptors themselves. Four Nobel Prizes for one pathway is a fair indication of how much of medicine sits on it.
The other second messengers
Calcium is the most versatile. Cells keep their internal calcium extraordinarily low — about ten thousand times lower than outside — precisely so that opening a calcium channel produces a huge proportional change. A calcium spike triggers muscle contraction, neurotransmitter release, and secretion of hormones.
IP₃ and DAG are made together by splitting a membrane phospholipid. IP₃ diffuses to the endoplasmic reticulum and opens its calcium stores; DAG stays in the membrane and activates another kinase.
Cyclic GMP relaxes smooth muscle in blood vessel walls, and it produces one of the neatest drug stories in medicine.
Nitric oxide, cGMP, and two drugs that must never be combined. The lining of a blood vessel releases nitric oxide, a gas that diffuses into the muscle of the vessel wall and switches on an enzyme that makes cGMP. cGMP relaxes the muscle, the vessel widens, and blood flows more freely. The signal is then ended by an enzyme called phosphodiesterase, which destroys the cGMP.
Sildenafil blocks phosphodiesterase type 5, the version found mainly in the blood vessels of the penis. cGMP is not destroyed, relaxation persists, and blood flow increases. The drug was originally developed for angina, performed unremarkably in that trial, and the effect that made it famous emerged as a side effect the volunteers reported.
Nitroglycerin, the classic angina treatment, works on the same pathway from the other end — it releases nitric oxide, raising cGMP throughout the circulation and widening vessels everywhere. Give both together and you get sustained, body-wide vasodilation with nothing to stop it: blood pressure can collapse to a fatal level. This is one of the few absolute drug contraindications in medicine, and it is why an ambulance crew asks a man with chest pain whether he has taken anything for erectile dysfunction in the last 24 to 48 hours before giving nitrates. It is not an awkward question; it is the pharmacology of this paragraph. Chapter 22.13.
Receptors adapt, and that has consequences
A receptor exposed to constant stimulation stops responding. Two mechanisms: desensitisation, where the receptor is chemically modified within seconds to minutes so it no longer couples to its G protein, and downregulation, where the cell physically removes receptors from its surface over hours to days.
This is what tolerance to a drug is. It is why a person on long-term opioids needs escalating doses, and why continuous salbutamol becomes less effective.
And it produces one genuinely dangerous situation. A person on a beta-blocker has their beta receptors blocked continuously, so the cell responds by making more of them. If the drug is stopped suddenly, all those extra receptors are exposed at once to normal circulating adrenaline, and the heart is now far more sensitive than before treatment. The result can be a surge in heart rate and blood pressure severe enough to cause angina, a heart attack or an arrhythmia. Beta-blockers are therefore tapered over one to two weeks, never stopped abruptly. Chapter 22.7.
Programmed cell death
Now the other instruction. There are two fundamentally different ways for a cell to die, and telling them apart is one of the most useful distinctions in pathology.
Necrosis is death by injury. The cell loses its energy supply, the sodium pump of Chapter 1.4 stops, water floods in, the cell swells and finally bursts, spilling its contents — including lysosomal enzymes — into the surrounding tissue. Neighbouring cells are damaged, the immune system arrives, and inflammation follows. Necrosis is always messy, always affects groups of cells together, and always causes inflammation. A heart attack, frostbite and a burn all produce necrosis.
Apoptosis is death by instruction. The word is Greek for the falling of leaves from a tree, chosen deliberately for the image of an orderly, seasonal, unremarkable departure. The cell shrinks rather than swells, its chromatin condenses, its DNA is cut into neat fragments, and the cell breaks into membrane-wrapped parcels that neighbouring cells swallow whole. Nothing leaks. There is no inflammation. It happens to single cells scattered among healthy ones, and it is over in an hour or two.
The machinery
The demolition crew is a family of enzymes called caspases — proteases that cut their targets specifically after an aspartate amino acid. They are made in an inactive form and sit dormant in every cell, so the tools are always present and only the order is missing. When activated, they cut the nuclear lamins so the envelope collapses, cut the cytoskeleton so the cell rounds up, and switch on an enzyme that chops the DNA between the histone spools.
Two routes activate them.
The intrinsic pathway starts inside, in response to DNA damage, oxygen starvation or the p53 signal of Chapter 1.7. The decision is made at the mitochondrial membrane by a family of proteins that oppose each other: Bcl-2 and its relatives hold the membrane shut and keep the cell alive, while Bax and its relatives punch pores in it. When the pro-death side wins, cytochrome c — the same small protein that ferries electrons between Complexes III and IV in Chapter 1.6 — leaks out of the mitochondrion into the cytosol. In the cytosol it has an entirely different job: it assembles a platform that activates the first caspase. One molecule, two completely unrelated functions depending on which side of a membrane it is on, which is a good demonstration of how little biology wastes.
The extrinsic pathway starts outside, when a death ligand from another cell binds a death receptor on the surface. This is how a killer T cell executes an infected or cancerous cell — it does not poison it, it instructs it to die.
What apoptosis is for
It sculpts. Your hands begin as solid paddles in the embryo, and fingers appear because the cells in between are told to die. When that instruction fails, the fingers stay joined — the condition is called syndactyly.
It edits the immune system. As T cells develop in the thymus, any that react against your own tissues are ordered to die. Chapter 13.6 covers what happens when that filter leaks: autoimmune disease.
It removes the damaged. Sunburn is a good everyday example. Ultraviolet light damages DNA in skin cells; p53 detects the damage; the cells that cannot be repaired are ordered to die. The peeling a few days later is the corpses. The peeling is the defence working — those are exactly the cells that would otherwise have become skin cancers.
And it deletes the surplus. The nervous system produces far more neurons than it needs, and the ones that fail to make useful connections die.
When it goes wrong
Too little apoptosis is cancer. Follicular lymphoma is the textbook case: a chromosomal rearrangement moves the BCL2 gene next to a heavily used antibody gene, so the cell over-produces Bcl-2 — the protein that keeps the mitochondrial membrane shut. The cells do not divide unusually fast; they simply refuse to die, and they accumulate. The drug venetoclax, approved in 2016, is a molecule that fits into Bcl-2's binding pocket and disables it, releasing the block. It works remarkably well in some leukaemias, and it is a direct product of understanding this pathway.
Too much apoptosis is degeneration. Neuron loss in Parkinson's and Alzheimer's disease is partly apoptotic, and HIV destroys CD4 T cells largely by triggering their apoptosis rather than by bursting them.
And most cancer treatment works by pushing cells into apoptosis, not by dissolving them. Chemotherapy and radiotherapy damage DNA; the damage is detected; p53 orders death. Which is precisely why tumours that have lost p53 resist both, and why that mutation is such bad news.
Two more ways cells die
Autophagy — self-eating — is not primarily a death mechanism but a recycling one. The cell wraps a worn-out organelle or a patch of cytoplasm in a membrane and delivers it to a lysosome to be broken down into reusable parts. It runs constantly, and it steps up sharply during starvation, when the cell begins consuming its own less essential parts to survive. Yoshinori Ohsumi received the 2016 Nobel Prize for working out the machinery. Much of what is claimed about fasting and autophagy in popular writing runs well beyond the evidence in humans, and Chapter 24.4 separates the two.
Pyroptosis is deliberately inflammatory death, used against bacteria that have got inside a cell. The cell bursts on purpose, destroying the bacterium's hiding place and releasing alarm signals. It is useful in infection and destructive in sepsis, where it runs out of control (Chapter 17.11).
What the next page fixes
Everything in Part 1 so far describes cells as they exist now — signalling, dividing, dying, running chemistry of extraordinary complexity that nothing simpler could have assembled in one step. Chapter 1.9 asks how any of this started: what the evidence actually shows about the origin of life, what has been demonstrated in a laboratory, what remains genuinely unknown, and why the honest answer is more interesting than the confident ones.