OnCo

Mechanics of cancer

How cancer works, drawn stage by stage: from the defences a tumour has to breach, through the machinery it hijacks to copy itself, to the ways it feeds, hides, spreads and outlasts treatment. Every stage is a diagram, and every diagram is tied to the drugs, technologies and targets that act on it.

9
chapters
56
stages
70
pathway diagrams
166
products and targets placed

Read it top to bottom as a story, or jump to a chapter. In each diagram, violet boxes are druggable targets; pick a product above a diagram to see which nodes it hits and where the escape routes are. The resistance atlas continues chapter nine class by class, and the pathway index lists every diagram alone.

Chapter 1 · 8 stages

The body's defences

Cancer is not the default. A body runs about 37 trillion cells and copies its DNA billions of times a day; almost none of it turns malignant because layer upon layer of defence stands in the way. Every one of these layers has to be breached before a tumour exists, and every one is a place a treatment can help.

  1. 1.1 Tissue architecture and the basement membrane
  2. 1.2 Immune surveillance
  3. 1.3 DNA damage checkpoints
  4. 1.4 Apoptosis: the self-destruct switch
  5. 1.5 Senescence: permanent retirement
  6. 1.6 Tumour suppressors: p53 and RB
  7. 1.7 Contact inhibition
  8. 1.8 Telomere limits
Stage 1.1

Tissue architecture and the basement membrane

Organs are built like walled towns. Lining cells sit on a dense protein sheet, the basement membrane, and hold hands through junctions. A growth that stays above the sheet is 'in situ' and almost always curable; the disease becomes cancer proper when it cuts through.

Polarised epitheliumE-cadherin junctionsBasement membraneIntegrins / hemidesmosomesCarcinoma in situMMPs, uPA, invadopodiaStroma, vesselsInvasive carcinomaMyoepithelial layeractivatesinhibitsdruggable target (click)hit by selected productescape route
Basement membrane & tissue barriersEvery organ keeps its lining cells behind a thin, dense sheet of protein called the basement membrane. A tumour that has not crossed it is 'in situ' and essentially curable; crossing it is the moment cancer becomes invasive.

A shop floor with a locked glass floor beneath it. Staff (epithelial cells) can be unruly upstairs and it is still contained; the emergency begins when someone cuts through the glass into the building services below, where the plumbing (blood and lymph vessels) runs.

Contact, stiffness, GPCRsNF2 (Merlin)MST1/2 → LATS1/2YAP/TAZTEAD transcriptionGrowth, EMT, drug toleran…activatesinhibitsdruggable target (click)hit by selected productescape route
Hippo–YAP/TAZThe pathway that tells organs when to stop growing. Cancers disable it so YAP and TAZ stay in the nucleus driving growth; in mesothelioma, NF2 loss does exactly that, and the first drugs against the YAP–TEAD switch are in trials.

A building inspector (Hippo) who checks that the block is full and stops new floors. Cancers fire the inspector, and the architect (YAP/TAZ) keeps adding storeys.

Stage 1.2

Immune surveillance

The immune system removes abnormal cells all the time. Seven steps have to work: antigen released, picked up, T cells trained, dispatched, let in, target recognised, target killed. Tumours that exist are the ones that broke a step.

Light up a product:+12 more via the Connected tab
1 Antigen release2 DC capture (cDC1)3 Priming (CD28 / CTLA-4)4 Trafficking (CXCL9/10)5 Infiltration6 Recognition (MHC-I)7 Killing (PD-1 brake)Escape at any stepactivatesinhibitsdruggable target (click)hit by selected productescape route
The cancer-immunity cycleSeven steps the immune system must complete to kill a tumour: release of antigens, pick-up by dendritic cells, priming of T cells in lymph nodes, travel, entry into the tumour, recognition, and killing. Every immunotherapy pushes on one step; every escape blocks one.

A relay of seven runners. The race is only won if every baton is passed. Cancers usually drop only one or two batons, so the treatment that works is the one that fixes the step that actually failed, which is why the same drug cures one patient and does nothing for the next.

Light up a product:+1 more via the Connected tab
Proteasome → peptidesTAP transportMHC-I / B2M loadingPeptide–MHC on surfaceCD8 T cell (TCR)IFN-γ → JAK1/2 → STAT1PD-L1 inductionDendritic cell cross-pres…Escape: B2M/HLA loss, JAK…activatesinhibitsdruggable target (click)hit by selected productescape route
Antigen presentation & immune editingHow the immune system sees cancer, and how cancer learns to hide. Tumours display fragments of their proteins on MHC molecules; T cells kill the ones they recognise; the survivors are the ones that stopped showing fragments or switched on brakes.

Wanted posters: the cell pins fragments of everything it makes onto its surface. Immune police recognise criminals' faces. Cancers that survive have taken down the posters (lost MHC) or bribed the police (checkpoints).

Open questions
  • Why do some people with the same mutations get cancer and others not: is it immune genetics (HLA) or luck?
  • Can surveillance be strengthened in healthy people at high risk (interception vaccines)?
Stage 1.3

DNA damage checkpoints

Before a cell copies or divides, sensors check the DNA. Damage halts the cycle until repair crews finish; severe damage triggers suicide. Cancers cut the sensors and, in doing so, become dependent on the few repair routes they have left.

Light up a product:
DNA damageATM / ATRp53MDM2p21CDK4/6 – cyclin DRBE2FS phase (DNA replication)WEE1 (G2/M)activatesinhibitsdruggable target (click)hit by selected productescape route
p53 / RB / cell-cycle checkpointThe p53 and RB checkpoints are the cell's brakes. p53 senses damage and stops the cell from copying itself; RB holds the cell at the G1 gate until CDK4/6 unlocks it. Cancers cut these brakes.

A checkpoint at a border: p53 is the inspector who halts traffic when something looks wrong, MDM2 is the manager who keeps sending the inspector home, RB is the barrier arm, and CDK4/6 is the motor that lifts it. Cancers bribe the inspector (TP53 mutation) or hot-wire the motor (cyclin D amplification).

Light up a product:
Single-strand breakPARP1Replication fork collapseATR / CHK1Double-strand breakBRCA1/2 – RAD51 (HR)NHEJ / POLQ (error-prone)Accurate repairGenomic collapse / deathactivatesinhibitsdruggable target (click)hit by selected productescape route
DNA damage response & homologous recombinationThe DNA damage response is the cell's set of repair crews. Single-strand breaks are patched by PARP; double-strand breaks by BRCA-dependent homologous recombination. Lose one crew and the cell survives; lose both and it dies. That is how PARP inhibitors work.

Two repair crews for a road: PARP fixes potholes, BRCA rebuilds collapsed bridges. A town that has lost its bridge crew (BRCA mutation) survives as long as potholes are fixed before they become bridge collapses. Block the pothole crew (PARP inhibitor) and the bridges fall.

Open questions
  • How wide is the therapeutic window for ATR, WEE1 and CHK1 inhibitors given that normal dividing tissue uses the same checkpoints?
  • Can checkpoint status be read from a routine biopsy or blood test?
Stage 1.4

Apoptosis: the self-destruct switch

Every cell carries a demolition kit. Internal alarms (damage, oncogene stress) or external orders (from immune cells) set it off. Cancer cells wire the kit shut with proteins such as BCL-2, and venetoclax proved that cutting that wire can put leukaemia into remission.

Light up a product:
DNA damage, oncogene stre…BH3-only (BIM, PUMA, NOXA)BCL-2 / BCL-XL / MCL-1BAX / BAKMitochondrial permeabilis…Cytochrome c → caspase-9Caspase-3/7 → apoptosisactivatesinhibitsdruggable target (click)hit by selected productescape route
Intrinsic apoptosis (BCL-2 family)Intrinsic apoptosis is the cell's self-destruct switch. BCL-2 holds it shut; BAX and BAK pull it open. Venetoclax pries BCL-2 off so the switch can fire.

A dam (mitochondrial membrane) held by guards (BCL-2, MCL-1) against demolition crews (BAX/BAK). Cancer hires extra guards. Venetoclax fires the BCL-2 guards, and the dam breaks.

Light up a product:
CTL / NK cellFASL, TRAILFAS, DR4/DR5Decoy receptors, c-FLIPDISC: FADD, caspase-8Perforin / granzyme BtBID → mitochondriaXIAP (IAPs)Caspase-3/7 → deathactivatesinhibitsdruggable target (click)hit by selected productescape route
Extrinsic apoptosis (death receptors)Immune cells kill by touch: they present FAS ligand or TRAIL to a target cell, whose death receptors then trigger self-destruction from the outside in. Tumours cut this wire by deleting the receptors or over-producing decoys and blockers.

A doorbell wired to a self-destruct switch: immune cells ring it. Some tumours rip out the doorbell (FAS loss), some stuff the wiring with insulation (c-FLIP), and some install a second doorbell that rings nowhere (decoy receptors).

Open questions
  • Can MCL-1 and BCL-XL be drugged without heart and platelet toxicity?
  • Does BH3 profiling of a patient's cells predict which BH3 mimetic will work?
Stage 1.5

Senescence: permanent retirement

A stressed cell can stop dividing for good but stay alive. That protects against cancer in the short run, but retired cells leak inflammatory signals that help neighbouring tumours grow and resist treatment.

Light up a product:
Oncogene / therapy stressp53 → p21p16 → RBStable arrestSASP (IL-6, IL-8, MMPs)Inflammation, relapse, re…Senolytics (BCL-XL, uPAR …activatesinhibitsdruggable target (click)hit by selected productescape route
Cellular senescenceDamaged cells can stop dividing permanently instead of dying. That protects against cancer at first, but senescent cells linger, secrete inflammatory signals, and after chemotherapy can help tumours relapse, so removing them (senolytics) is a new strategy.

Retired workers who refuse to leave the office. They no longer do the job, but they shout, clutter the corridors, and sometimes hire back the very people who were fired.

Open questions
  • Should senescent cells be cleared after chemotherapy (the one-two punch), and with what?
  • Do CDK4/6 inhibitors work mainly through senescence or arrest?
Stage 1.6

Tumour suppressors: p53 and RB

Two master brakes. p53 coordinates the response to any emergency; RB holds the door to DNA copying shut. Half of cancers break p53; most of the rest disable it indirectly. RB is bypassed by over-active CDK4/6 or lost outright.

Light up a product:
DNA damage (ATM/ATR)Oncogene stress → ARFHypoxia, ribosome stressMDM2 / MDMXp53TP53 mutation (~50%)p21 → arrestPUMA, NOXA → apoptosisSenescence, repairactivatesinhibitsdruggable target (click)hit by selected productescape route
The p53 network (guardian of the genome)p53 is the cell's emergency coordinator. Damage, oncogene stress, or lack of oxygen switch it on; it then pauses division, orders repairs, or triggers suicide or permanent retirement. MDM2 keeps it switched off in healthy cells. Half of all cancers break p53 outright; many of the rest over-produce MDM2.

A fire marshal who is normally kept locked in a cupboard (by MDM2). When alarms sound, the cupboard opens and the marshal stops work, calls repairs, and if the building is beyond saving, orders evacuation (apoptosis) or condemns it (senescence). Cancers either sack the marshal (TP53 mutation) or weld the cupboard shut (MDM2 amplification).

Open questions
  • Can mutant p53 be reactivated at scale (rezatapopt is the first proof for one mutation)?
  • Why do MDM2 inhibitors cause so much marrow toxicity, and can dosing solve it?
Stage 1.7

Contact inhibition

Normal cells stop dividing when they touch neighbours or feel a stiff, crowded tissue. The Hippo pathway relays that signal by locking YAP/TAZ out of the nucleus. Tumours ignore the crowd.

Open questions
  • Are TEAD inhibitors tolerable and active beyond NF2-mutant mesothelioma?
  • How much of drug tolerance is YAP-driven and reversible?
Stage 1.8

Telomere limits

Chromosome ends shorten with every division, a built-in counter that retires cells after roughly 50 divisions. Cancers reset the counter by switching telomerase back on.

Each division shortens te…Senescence / crisisTERT reactivation (promot…ALT (ATRX/DAXX loss)Replicative immortalityATR dependence (ALT)activatesinhibitsdruggable target (click)hit by selected productescape route
Telomere maintenance & replicative immortalityNormal cells can divide only so many times because the protective caps on their chromosomes, telomeres, wear down. Cancers switch the cap-rebuilding enzyme telomerase back on, or find another way (ALT), so they can divide forever.

The plastic tips on shoelaces fray a little every time you tie them; when they are gone the lace unravels and the shoe is thrown out. Cancer cells carry a machine that keeps re-tipping the laces.

Open questions
  • Imetelstat works in MDS; will telomerase inhibition work in solid tumours before the tumour outgrows the patient?
  • Can ALT-positive tumours be targeted through their ATR dependence?
Chapter 2 · 6 stages

How a cell becomes cancer

Cancer is evolution inside a body. Damage writes mutations, a few of them give a cell an edge, its descendants compete, and epigenetic and environmental pressures shape which lineage wins. Viruses and chronic inflammation shortcut the process. Reading this history in a tumour's genome now guides treatment.

  1. 2.1 Mutation and mutational signatures
  2. 2.2 Drivers versus passengers
  3. 2.3 Clonal evolution
  4. 2.4 Epigenetic reprogramming
  5. 2.5 Field cancerisation
  6. 2.6 Viral and inflammatory causes
Stage 2.1

Mutation and mutational signatures

Every cause of DNA damage leaves its own fingerprint. Sunlight, tobacco, a missing repair enzyme, even a gut bacterium, each write a recognisable pattern into the genome. Those patterns say what caused a cancer and which repair crews it lacks.

Light up a product:
Exogenous: UV, tobaccoEndogenous: APOBEC, ROSReplication errorsDNA lesionsRepair: MMR, HR, BER, NERFixed mutationsSignature (SBS, ID, CN)HRD, MSI, TMB biomarkersDrivers, neoantigensactivatesinhibitsdruggable target (click)hit by selected productescape route
Mutagenesis & mutational signaturesEvery cause of DNA damage leaves its own fingerprint in the genome: sunlight, tobacco, a faulty repair enzyme, a gut bacterium. Reading these fingerprints tells you what caused a cancer and which repair crews it is missing, which in turn predicts which drugs will work.

Footprints in snow. A fox, a dog and a child each leave a distinct print; you can tell who crossed the garden without having seen them. Cancer genomes are snowfields, and each mutagen and each broken repair crew leaves its own print.

Stage 2.2

Drivers versus passengers

Of the thousands of mutations in a tumour, only a handful drive it: a stuck accelerator (oncogene, one hit) or cut brakes (tumour suppressor, both copies). The rest are passengers, though they make the tumour visible to the immune system.

Light up a product:
Proto-oncogeneMutation, amp, fusionOncogene ON (1 hit)Tumour suppressor1st hit (germline/somatic)2nd hit: LOH, methylationBrake lost (2 hits)Passengers, neoantigensClonal expansionactivatesinhibitsdruggable target (click)hit by selected productescape route
Drivers, passengers & the two-hit modelOf the thousands of mutations in a tumour, only a handful (typically 2-8) actually drive it. Drivers either jam an accelerator on (oncogenes, one hit is enough) or remove a brake (tumour suppressors, both copies must go). Everything else is a passenger along for the ride.

A car with a stuck accelerator (oncogene: one fault is enough) and cut brake lines (tumour suppressor: both lines must fail because they are duplicated). The scratches on the paintwork (passengers) did not cause the crash but they tell you what road it drove on.

Open questions
  • How many drivers hide among 'variants of uncertain significance', and can AI variant-effect models classify them?
  • Which passengers are collateral vulnerabilities (MTAP loss) worth drugging?
Stage 2.3

Clonal evolution

A tumour is a population, not a clone. Subclones compete, therapy selects the fittest, and relapse is usually a minority that was there all along. Chromosome shuffling speeds evolution; blood tests can follow it in real time.

Light up a product:
Truncal driver cloneSubclone ASubclone B (resistant)Therapy (selection)MRD (ctDNA)Relapse dominated by BAdaptive / combination th…activatesinhibitsdruggable target (click)hit by selected productescape route
Clonal evolution & minimal residual diseaseA tumour is a population that evolves by natural selection. Treatment kills the sensitive cells and selects the rest, which is why resistance is the rule; measuring the surviving population (MRD) and adapting therapy is the counter-strategy.

Clonal evolution is like weeding a field with one herbicide year after year: the field fills with the one weed that shrugs it off. Rotating herbicides and leaving some susceptible weeds to crowd out the resistant ones is the evolutionary alternative.

Light up a product:
Ageing HSCsDNMT3A, TET2, ASXL1, PPM1…CHIP clone (VAF ≥2%)Therapy-related MDS/AMLInflammation → cardiovasc…False-positive ctDNAChemo, PARPi, radioligand…activatesinhibitsdruggable target (click)hit by selected productescape route
Clonal haematopoiesis (CHIP)As we age, blood stem cells with cancer-like mutations quietly expand in most people. These clones raise leukaemia and heart disease risk, are accelerated by chemotherapy, and confuse blood tests for cancer DNA.

Weeds in the blood's seed bank. Harmless individually, they spread with age and after chemical sprays, some turn into leukaemia, and their DNA litters the blood, so a test for tumour DNA can mistake weeds for cancer.

Open questions
  • Can evolution be steered (adaptive therapy) rather than only outrun?
  • How early in life do the first driver clones arise, and could they be intercepted?
Stage 2.4

Epigenetic reprogramming

Cells can change behaviour without changing their DNA sequence, by rewriting the chemical tags that decide which genes are read. Cancers silence brakes and antigens this way, and switch identity under drug pressure. Unlike mutations, tags can be erased.

Light up a product:
DNA methylation (DNMT, TE…Histone marks (EZH2, KMT2…Readers (BET) & remodelle…Chromatin stateSilenced tumour suppresso…Active oncogenic programm…Menin–KMT2A scaffoldactivatesinhibitsdruggable target (click)hit by selected productescape route
Epigenetic reprogrammingCancer changes not just its genes but how they are read: chemical tags on DNA and histones silence guardians and awaken growth programmes. Unlike mutations, these changes are reversible, which is the hope behind epigenetic drugs.

The genome is the book; epigenetics is the highlighting and the pages stapled shut. Cancer staples shut the safety chapters and highlights the growth chapters. Epigenetic drugs pull staples.

Light up a product:
BAF / PBAF / ncBAFNucleosome repositioningEnhancer access, differen…ARID1A lossSMARCA4 loss → SMARCA2 de…SMARCB1 loss → EZH2PRC2 (EZH2) antagonismactivatesinhibitsdruggable target (click)hit by selected productescape route
SWI/SNF chromatin remodellingA machine that opens and closes DNA so genes can be read. One in five cancers has a broken part (ARID1A, SMARCA4, PBRM1), and losing one part often creates a dependence on its twin, which is the basis for new synthetic-lethal drugs.

A librarian who unlocks shelves on request. When one librarian is fired the other covers both shifts; fire the second and the library stops working. That second librarian is the drug target.

Stage 2.5

Field cancerisation

Years before a tumour, whole patches of normal-looking tissue are already colonised by mutant clones (a smoker's airway, sun-exposed skin, Barrett's oesophagus). The cancer is the one clone that got further; the field explains second cancers and local recurrence.

Light up a product:
Carcinogen exposure, agei…Mutant clones in normal t…Clonal competitionProgression to dysplasia …Second primaries, local r…Interception: chemopreven…activatesinhibitsdruggable target (click)hit by selected productescape route
Field cancerisationCancer often arises from a whole region of tissue that already carries mutations, not from one rogue cell. Sun-exposed skin, smokers' airways, and Barrett's oesophagus are patchworks of mutant clones competing long before a tumour appears.

A lawn where many patches have already turned to weeds. Any single tumour is one patch that took over; mowing it leaves the rest of the lawn ready to sprout again.

Open questions
  • Which mutant clones in normal tissue matter, given that healthy skin and oesophagus are full of them?
  • Can a field be treated (chemoprevention, vaccines) without harming the organ?
Stage 2.6

Viral and inflammatory causes

One cancer in eight is caused by a virus, and many more by chronic inflammation. HPV and hepatitis B carry master keys to the cell's brakes; long-running inflammation supplies growth signals and mutagens. Vaccines and anti-infectives are among the most effective anti-cancer drugs ever made.

Light up a product:+2 more via the Connected tab
HPV E6 / E7EBV LMP1, EBNAHBV / HCVp53 degradedRB inactivatedNF-κB, immortalisationChronic inflammation, cir…Virus-driven cancerViral antigens → IO respo…Vaccination, antiviralsactivatesinhibitsdruggable target (click)hit by selected productescape route
Oncogenic virusesAbout one cancer in eight worldwide is caused by a virus. HPV, hepatitis B and C, Epstein-Barr, HTLV-1, KSHV and Merkel cell polyomavirus each hijack the same brakes cancer normally has to mutate, which is why vaccines against HPV and HBV are among the most effective anti-cancer drugs ever made.

A burglar who does not need to pick the locks because he carries a master key: E6 and E7 are keys that open p53 and RB directly, saving the virus the years of mutation a spontaneous cancer needs. The upside is that the burglar's face is on every camera, so the immune system can be taught to spot him.

Light up a product:
Infection, injury, obesityTNF, IL-1β, IL-6IKK → IκB degradationNF-κBIL-6 → STAT3Survival, proliferation, …BCR → BTK (lymphoma)COX-2 → PGE2activatesinhibitsdruggable target (click)hit by selected productescape route
Inflammation & NF-κBChronic inflammation is soil for cancer: it feeds growth signals, DNA damage, and immune suppression. The NF-κB switch inside cells is the master relay, and colitis, hepatitis, and H. pylori gastritis are the clinical proof.

A wound that never heals: the repair crews keep pouring in growth signals and clearing away rubble, and a wound that is always being rebuilt is a wound where mistakes accumulate.

Chapter 3 · 8 stages

Replication and growth machinery

Cancer cells use the same engine as normal cells, only stuck at full throttle. The cell cycle, DNA copying, cell division, growth-factor relays, gene transcription, protein synthesis and protein disposal are all ordinary machinery; most drugs in oncology work by jamming one of these parts harder in the cells that lean on it most.

  1. 3.1 The cell cycle and CDKs
  2. 3.2 DNA replication and replication stress
  3. 3.3 Mitosis and chromosome segregation
  4. 3.4 Growth-factor signalling: RTKs and RAS/RAF/MEK
  5. 3.5 PI3K/AKT/mTOR
  6. 3.6 Transcription and MYC
  7. 3.7 Translation
  8. 3.8 Protein homeostasis and the proteasome
Stage 3.1

The cell cycle and CDKs

Division runs on a clock of cyclins and CDKs firing in order. Cancers flood the first step (cyclin D) or remove the throttle (p16). CDK4/6 inhibitors slow the clock and changed breast cancer treatment.

Light up a product:
Mitogens (ER, RTK, RAS)Cyclin D – CDK4/6p16 (CDKN2A)RB → E2F releasedCyclin E – CDK2p21 / p27S phase (cyclin A)Cyclin B – CDK1WEE1 / PKMYT1Mitosisactivatesinhibitsdruggable target (click)hit by selected productescape route
The cell-cycle engine (cyclins & CDKs)Cell division runs on a clock made of cyclins and their kinases (CDKs), each pair firing in order: D-CDK4/6 to leave rest, E-CDK2 to start copying DNA, A-CDK2 to finish, B-CDK1 to divide. Cancers speed the clock; CDK inhibitors slow it.

An engine with four cylinders that must fire in sequence. Cyclins are the fuel injected into each cylinder in turn and burned away; CDKs are the pistons. p16 and p21 are the hand on the throttle. Cancers flood the first cylinder (cyclin D) or remove the throttle hand (CDKN2A).

Open questions
  • Which patients need a CDK4/6 inhibitor at all, and who is cured by endocrine therapy alone?
  • Will CDK2 inhibitors overcome cyclin E-driven resistance safely?
Stage 3.2

DNA replication and replication stress

Copying three billion letters exactly once means licensing thousands of start points and firing them in waves. Cancers fire too many with too little ink; the forks stall and break. Most classic chemotherapy starves or jams this machinery.

Light up a product:
ORC, CDC6, CDT1MCM2-7 loaded (licence)Geminin, CRL4-CDT2CDK2 / DDK firingCMG helicase + Pol ε/δdNTP supply (RNR)Replication forkTOP1 (chemo, ADC payloads)MYC, cyclin E: excess ori…Replication stressactivatesinhibitsdruggable target (click)hit by selected productescape route
DNA replication & origin licensingBefore a cell divides it must copy three billion letters of DNA exactly once. It does this by 'licensing' thousands of start points in advance and then firing them in waves. Cancers fire too many too fast, and many chemotherapies work by starving or jamming the copying machinery.

Printing a 3,000-page book on thousands of small presses at once. Licensing hands out one ticket per press before printing begins; firing starts them in waves. Cancer starts far too many presses with too little ink (nucleotides), so pages smear and tear, and many chemotherapies simply cut off the ink supply or jam the rollers.

Light up a product:
Oncogenes (MYC, cyclin E,…Excess origin firing, sho…Stalled forks, ssDNA gapsATR → CHK1WEE1 / PKMYT1 restrain CD…G2/M checkpointFork collapse → DSBsMitotic catastropheTP53 (lost)activatesinhibitsdruggable target (click)hit by selected productescape route
DNA replication stressCancers copy their DNA too fast and with broken checkpoints, so replication forks stall and collapse. They survive only by leaning on emergency repair kinases such as ATR, CHK1, and WEE1, which is why blocking those kinases can be selectively lethal.

A photocopier running at triple speed with the paper-jam sensor removed. It keeps working only because a technician (ATR/CHK1/WEE1) constantly clears jams. Remove the technician and it destroys itself.

Open questions
  • Can replication stress be measured in a biopsy to pick patients for ATR/CHK1/WEE1 drugs?
  • Why do TOP1-payload ADCs work after topotecan or irinotecan failed?
Stage 3.3

Mitosis and chromosome segregation

A scaffold of microtubules pulls one copy of each chromosome to each daughter, and a checkpoint holds the split until every chromosome is hooked on. Taxanes freeze the scaffold; cells without p53 slip through with the wrong number of chromosomes.

Light up a product:+6 more via the Connected tab
Centrosomes (Aurora A, PL…Spindle microtubulesKinetochore attachmentTaxanes, vincas, MMAE, DM1SAC: MAD2, BUBR1, MPS1APC/C–CDC20Anaphase (separase)Mitotic death (BCL-XL↓)Slippage → tetraploidp53 arrestactivatesinhibitsdruggable target (click)hit by selected productescape route
Mitosis & the spindle assembly checkpointWhen a cell divides, a scaffold of microtubules (the spindle) pulls one copy of each chromosome to each side. A checkpoint holds the split until every chromosome is hooked on. Taxanes and vinca alkaloids freeze the spindle so the cell is stuck at this checkpoint until it dies.

A tug-of-war where the referee (spindle checkpoint) will not blow the whistle until every player has a grip on the rope. Taxanes glue the rope so nobody can pull; the match never starts and the players eventually collapse. Cells without p53 sneak off the pitch with the wrong number of players, which is how aneuploidy begins.

Light up a product:+6 more via the Connected tab
Mitotic errors, WGDChromosomal instabilityAneuploidy / karyotype he…Micronuclei → cytosolic D…cGAS–STINGecDNA oncogene amplificat…Clonal evolution, resista…Dependencies: KIF18A, SAC…TP53 loss permitsactivatesinhibitsdruggable target (click)hit by selected productescape route
Chromosomal instability & aneuploidyMost cancers have the wrong number of chromosomes and keep shuffling them at every division. This chaos fuels evolution and drug resistance, but it also stresses the cell and can trigger immune alarms, a double edge that researchers are trying to exploit.

Chromosomal instability is a library that reshuffles and duplicates random shelves every night. Most rearrangements are useless, some ruin the building, but occasionally one yields a book the librarian needs to survive a new rule, and the mess itself keeps the fire alarms twitching.

Open questions
  • Is chromosomal instability itself a druggable vulnerability (KIF18A) or too heterogeneous?
  • Why did Aurora and PLK1 inhibitors show so little activity as single agents?
Stage 3.4

Growth-factor signalling: RTKs and RAS/RAF/MEK

Antennas on the surface pair up when a signal lands and switch on the 'divide' relay: RAS to RAF to MEK to ERK. Cancers glue the antennas on or jam RAS. Most targeted pills work here, and so do the escape routes.

Light up a product:+34 more via the Connected tab
Ligand (EGF, HGF, NRG)RTK dimer (EGFR, HER2, ME…Mutation, amp, fusionCBL → degradationGRB2/SOS → RASPI3K → AKTSRC, STAT3, PLCγProliferation, survivalBypass RTK (MET, HER3)activatesinhibitsdruggable target (click)hit by selected productescape route
Receptor tyrosine kinase activationGrowth-factor receptors are antennas on the cell surface that pair up when a signal lands and switch on the growth relays inside. Cancers mutate, multiply, or fuse these antennas so they broadcast 'grow' with no signal at all. Most targeted drugs, antibodies and ADCs start here.

Two halves of a walkie-talkie that only transmit when clipped together by a signal from outside. Cancer glues them together (fusions, mutations) or installs hundreds of extra sets (amplification), so the room is full of shouted 'grow' orders. TKIs pull the battery, antibodies tape over the microphone, ADCs use the aerial as a mailing address for poison.

Light up a product:+20 more via the Connected tab
RTK (EGFR, ALK, RET, MET)GRB2 / SOS1RAS (KRAS)NF1 (GAP)RAF (BRAF)MEK1/2ERK1/2Cyclin D1, MYC → prolifer…DUSP / SPRY feedbackactivatesinhibitsdruggable target (click)hit by selected productescape route
RAS / RAF / MEK / ERK (MAPK)The RAS-MAPK pathway is the cell's 'divide' relay. A signal at the surface flips RAS on, which passes to RAF, MEK, and ERK, which tell the nucleus to make the cell divide. KRAS and BRAF mutations jam it in the on position.

A relay race: receptor hands the baton to RAS, RAS to RAF, RAF to MEK, MEK to ERK, ERK runs into the nucleus and shouts 'divide'. A KRAS mutation is a runner who never stops running whether or not anyone handed them the baton.

Stage 3.5

PI3K/AKT/mTOR

The 'grow and survive' circuit: PI3K to AKT to mTOR, with PTEN as the off switch. The most frequently altered pathway in cancer, and one of the hardest to drug without hitting blood sugar and skin.

Light up a product:+11 more via the Connected tab
RTK (HER2, EGFR)PI3K (PIK3CA)PTENPIP3AKTTSC1/2mTORC1FOXO / BAD (apoptosis)Protein synthesis, growthactivatesinhibitsdruggable target (click)hit by selected productescape route
PI3K / AKT / mTORThe cell's 'grow and survive' circuit. Growth signals from the surface switch on PI3K, which switches on AKT, which switches on mTOR, which builds proteins and blocks self-destruction.

Think of a factory: the receptor is the order desk, PI3K and AKT are the managers relaying the order, PTEN is the accountant cancelling orders, and mTOR is the assembly line. Cancer forges orders (PIK3CA mutation) or fires the accountant (PTEN loss).

Open questions
  • Can mutant-selective PI3Kα inhibitors avoid hyperglycaemia and widen the window?
  • Why do PTEN-loss tumours respond less than PIK3CA-mutant ones?
Stage 3.6

Transcription and MYC

Cancers run a few genes at deafening volume from super-enhancers, and MYC is the master amplifier. The machinery is shared with normal cells, but tumours depend on it more, and hormone receptors are the oldest transcription drugs.

Light up a product:
Wnt, RAS, Notch, amplific…MYC/MAXRibosome biogenesis, meta…ProliferationCD47, PD-L1 (immune evasi…CDK9 / BET (transcription)AURKA / PLK1 (stability)activatesinhibitsdruggable target (click)hit by selected productescape route
MYCMYC is the most commonly amplified cancer gene, a master switch that turns on thousands of growth genes. It has no pocket for a conventional drug, so it remained 'undruggable' for 40 years; the first direct MYC drugs finally entered trials in the 2020s.

A conductor who can make every section of the orchestra play louder at once. You cannot take away the baton, so drugs try to silence the score (transcription), tire the conductor (degradation), or exploit the fact that a full-volume orchestra cannot afford a single missing player.

Light up a product:
Lineage TFs (ER, AR, ASCL…Fusion TFs (EWSR1-FLI1)Super-enhancerBRD4, Mediator, p300CDK7 (TFIIH) initiationCDK9 (P-TEFb) elongationRNA Pol IIMYC, MCL-1 (short-lived)Menin–KMT2A (AML)activatesinhibitsdruggable target (click)hit by selected productescape route
Transcriptional machinery & addictionCancer cells run a few genes (MYC, their lineage factors, their fusion oncogenes) at extreme volume from giant control regions called super-enhancers. The amplifiers, BRD4, CDK7, CDK9 and Mediator, are the same in every cell, but cancers are unusually dependent on them, and that dependence is druggable.

A concert where a few songs are played at deafening volume through rented amplifiers. Cutting the mains for a moment (BET, CDK7/9 inhibitors) silences the loudest songs first because their sound decays fastest, while the quieter household appliances keep humming.

Also drawnRNA splicing
Open questions
  • Can MYC be drugged directly (degraders, OMOMYC) and will patients tolerate it?
  • Which fusion transcription factors (EWSR1-FLI1) are reachable with glues or degraders?
Stage 3.7

Translation

Genes are read into RNA, then RNA into protein. mTOR controls the rate, and the short-lived oncoproteins MYC, cyclin D1 and MCL-1 are the first casualties when translation slows.

Light up a product:
mTORC14E-BPeIF4EeIF4F (4E/4G/4A)MNK1/2MYC, cyclin D1, MCL-1 tra…Ribosome biogenesis (Pol …activatesinhibitsdruggable target (click)hit by selected productescape route
mRNA translation (eIF4F / mTOR)Cancer cells must make protein at furious speed. The eIF4F complex that starts protein synthesis is the funnel where growth signals converge, and drugs that pinch the funnel starve the tumour of the proteins it needs most.

A factory's single loading dock (eIF4F). No matter how many orders the managers (RAS, PI3K, MYC) shout, everything must pass through the dock. Narrow the dock and the most urgent, oversized orders (oncogene proteins) are the first to fail.

Open questions
  • Are eIF4A/eIF4E inhibitors (zotatifin) selectively toxic to tumours in patients?
  • Can ribosome biogenesis be targeted with an acceptable window?
Stage 3.8

Protein homeostasis and the proteasome

Unwanted proteins are tagged with ubiquitin and fed into a shredder. Myeloma dies when the shredder jams; and the newest drugs forge the tags so that a cancer destroys its own oncoproteins.

Light up a product:
E1 → E2 ubiquitinE3 ligase (CRBN, VHL, MDM…Substrate (IKZF1/3, p53, …K48 ubiquitin chainDUBs (USP7)26S proteasome (β5)UPR, IκB → NF-κBGlues / PROTACs hijack E3DegradationHSP90 chaperonesactivatesinhibitsdruggable target (click)hit by selected productescape route
Ubiquitin–proteasome system & protein homeostasisCells tag unwanted proteins with a small marker called ubiquitin and feed them into a shredder, the proteasome. Myeloma cells, which make antibody in bulk, die if the shredder jams; and the newest drugs hijack the tagging machinery to make a cancer destroy its own oncoproteins.

A recycling plant with barcode stickers (ubiquitin) and a shredder (proteasome). Myeloma is a paper mill that produces so much waste it dies when the shredder stops (bortezomib). PROTACs and glues are forged stickers that get the plant to shred the cancer's own machinery.

Chapter 4 · 4 stages

Evading death and repair

To survive the damage they generate and the treatments thrown at them, cancer cells rewire death and repair. They block apoptosis, lean on whichever repair crew they have left, eat themselves to survive famine, and stiffen their membranes against oxidative death. Each rewiring is a dependence.

  1. 4.1 Apoptosis and the BCL-2 family
  2. 4.2 DNA repair pathways and synthetic lethality
  3. 4.3 Autophagy
  4. 4.4 Ferroptosis
Stage 4.1

Apoptosis and the BCL-2 family

A dam (the mitochondrial membrane) held by guards (BCL-2, MCL-1, BCL-XL) against demolition crews (BAX, BAK). Cancer hires extra guards; venetoclax fires the BCL-2 guards. Immune killing uses the same dam from the outside.

Open questions
  • Why does venetoclax work in CLL and AML but barely in solid tumours?
  • Can platelet-sparing BCL-XL degraders open solid-tumour use?
Stage 4.2

DNA repair pathways and synthetic lethality

Double-strand breaks (HR versus end joining), copying errors (mismatch repair), and single damaged letters (base excision, PARP) each have their crew. Tumours that lost one crew survive on the others, and that dependence is the first widely successful way to drug a lost gene.

Light up a product:
Double-strand breakMRN → ATM → CHK253BP1–Shieldin (protect)NHEJ: Ku, DNA-PKcs, LIG4Resection: BRCA1–CtIPPALB2–BRCA2 → RAD51Accurate HR (sister copy)POLQ end joining (backup)Error-prone joiningPARP trapping → breaksactivatesinhibitsdruggable target (click)hit by selected productescape route
Double-strand break repair: HR versus end joiningA break through both strands of DNA is the most dangerous lesion a cell faces. Two crews compete to fix it: homologous recombination copies the answer from the sister chromosome (accurate, needs BRCA), while end joining simply glues the ends (fast, sloppy). Which crew wins decides whether PARP inhibitors and radiation kill the cell.

A torn page. The careful archivist (HR) fetches the twin copy from the shelf and transcribes it letter for letter; the hurried clerk (NHEJ) tapes the two halves together, losing a few words. Tumours missing the archivist survive on the clerk, so anything that adds more torn pages (PARP inhibitors, platinum, radiation) buries them.

Light up a product:
Replication mismatchMutSα (MSH2–MSH6)MutLα (MLH1–PMS2)EXO1, Pol δ resynthesisCorrected DNAMMR loss (Lynch, MLH1 met…MSI-H, hypermutationFrameshift neoantigensCheckpoint-inhibitor resp…WRN dependenceactivatesinhibitsdruggable target (click)hit by selected productescape route
Mismatch repair & microsatellite instabilityAfter DNA is copied, a proofreading crew fixes the letters the polymerase got wrong. Lose it and the genome fills with thousands of small errors, especially in repetitive stretches (microsatellites). Those errors make abnormal proteins that the immune system can see, which is why immunotherapy works so well in these cancers.

A spell-checker that runs after every page is typed. Without it, typos pile up, especially in words like 'banana' where it is easy to lose count of the repeats. The garbled words in the resulting proteins read as foreign, so the immune system, once its brakes are released, attacks with unusual vigour.

Stage 4.3

Autophagy

Under famine or drug stress a cell eats its own components to survive. KRAS-driven cancers depend on it, and it also destroys the molecules that display antigens to T cells.

Light up a product:
Nutrient stress, KRAS/MEK…ULK1 / AMPKmTORC1 (inhibits)AutophagosomeLysosome (HCQ blocks)Recycled fuel → survivalMHC-I degradation (PDAC)activatesinhibitsdruggable target (click)hit by selected productescape route
AutophagyAutophagy is the cell's recycling programme. Cancer cells, especially pancreatic and RAS-driven tumours, use it to survive starvation and drug stress, which is why hydroxychloroquine, an old malaria drug that blocks it, keeps appearing in trials.

A besieged city that starts recycling furniture into firewood. It keeps the lights on through the siege, and burning the 'wanted posters' (MHC) hides its criminals too.

Acts here
Open questions
  • Does hydroxychloroquine with MEK/KRAS inhibition help patients (trials in pancreatic cancer)?
  • Is autophagy inhibition safe long-term for neurons and muscle?
Stage 4.4

Ferroptosis

A form of death by rusting: iron-driven oxidation of membrane lipids. Cancer cells in mesenchymal or drug-tolerant states depend on the GPX4 enzyme to prevent it, and the NRF2 antioxidant programme protects them further.

Light up a product:
Cystine import (SLC7A11)GlutathioneGPX4Lipid peroxidationLabile iron (Fenton)FerroptosisMesenchymal / persister s…Apoptosis (BCL-2 family)activatesinhibitsdruggable target (click)hit by selected productescape route
Ferroptosis & regulated cell deathCells can die in several programmed ways. Beyond the classic apoptosis, ferroptosis kills through iron-driven fat oxidation, and drug-resistant, mesenchymal cancer cells turn out to be unusually prone to it.

Ferroptosis is rust. Iron plus oxygen eats through the cell's membranes unless an antioxidant crew (GPX4) keeps repainting them. Cells that changed shape to dodge chemotherapy have thinner paint.

Oxidative stressKEAP1 (mutated)NRF2ARE genes: GSH, NQO1, eff…Chemo/RT/ferroptosis resi…STK11/LKB1 loss (co-mutat…activatesinhibitsdruggable target (click)hit by selected productescape route
KEAP1–NRF2 antioxidant pathwayKEAP1–NRF2 is the cell's antioxidant defence switch. Lung cancers often break the off-switch (KEAP1), leaving NRF2 permanently on, which detoxifies chemotherapy and radiation and makes these tumours resistant to almost everything.

KEAP1–NRF2 is a smoke detector wired to a sprinkler system. Cancers jam the detector on, so the sprinklers run constantly and wash away every poison you throw at them.

Acts here
Open questions
  • Can GPX4 be inhibited in patients without killing kidneys and neurons?
  • Do dietary fats change ferroptosis sensitivity in tumours?
Chapter 5 · 6 stages

Feeding the tumour

A tumour is a construction site that never stops. It burns glucose fast and dirty, gorges on glutamine and fat, calls for new blood vessels when it runs short of oxygen, and starves the immune cells that share its table. Metabolism is how we image cancer (FDG PET) and increasingly how we starve it.

  1. 5.1 Warburg metabolism
  2. 5.2 Glutamine
  3. 5.3 Lipids
  4. 5.4 Hypoxia and HIF
  5. 5.5 Angiogenesis and VEGF
  6. 5.6 Nutrient competition
Stage 5.1

Warburg metabolism

Cancer cells burn glucose into lactate even with oxygen around: inefficient but fast, and it supplies building blocks. This is why an FDG PET scan lights up tumours.

Light up a product:
Glucose (GLUT1)Aerobic glycolysis (Warbu…Lactate export (MCT4)TCA cycleGlutamine → glutaminaseDe novo lipogenesis (FASN)One-carbon (SHMT2, MTHFD2…Mutant IDH → 2-HGPI3K/AKT/mTOR, MYC, HIFactivatesinhibitsdruggable target (click)hit by selected productescape route
Cancer metabolismCancer cells rewire how they eat. They burn glucose inefficiently but fast (the Warburg effect), gorge on glutamine and fats, and build the nucleotides and lipids needed to divide. This is why the FDG PET scan works, and why metabolism is a drug target.

A factory that switches from a clean, efficient power plant to burning everything it can find, fast and dirty, because speed matters more than efficiency when you are building a new factory every day.

Stage 5.2

Glutamine

The tumour's second favourite food: it feeds the energy cycle, donates nitrogen for DNA letters, and makes antioxidants. MYC- and KRAS-driven cancers eat so much that the T cells next door go hungry.

Light up a product:+13 more via the Connected tab
Glutamine (SLC1A5)Glutaminase (GLS)Glutamateα-KG → TCA anaplerosisGlutathione, NADPHNucleotides (N donor)MYC, KRAS drive uptakeIDH → 2-HG / reductiveT cells starvedmTORC1 sensingactivatesinhibitsdruggable target (click)hit by selected productescape route
Glutamine addictionAfter glucose, glutamine is the tumour's favourite food. It feeds the energy cycle, donates nitrogen for making DNA letters, and makes the antioxidant glutathione. MYC- and KRAS-driven cancers eat so much of it that they starve the T cells next door.

A construction site that runs on two deliveries: sand (glucose) for bulk and steel (glutamine) for the frame and the rebar. MYC doubles the steel order. Cutting one delivery rarely stops the build because the site switches suppliers; that is why single metabolic drugs have disappointed.

Open questions
  • Can glutamine be blocked in tumour cells while sparing T cells (DRP-104 concept)?
  • Which tumours are truly glutamine-addicted in patients rather than in dishes?
Stage 5.3

Lipids

Dividing cells need membranes, and membranes are fat. Cancers switch on fat-building enzymes adults normally keep off and steal lipids from nearby fat cells, which links obesity to thirteen cancers.

Light up a product:
Citrate → ACLY → acetyl-C…ACC → malonyl-CoAFASN → palmitateSCD1 → membranesSREBP (mTORC1, hypoxia)AMPK / LKB1Mevalonate → cholesterolCD36 uptake from adipocyt…Fatty acid oxidation (CPT…PUFA → ferroptosisObesity, insulin, IGF-1activatesinhibitsdruggable target (click)hit by selected productescape route
Lipid synthesis, uptake & cholesterolDividing cells need membranes, and membranes are fat. Cancers switch on the fat-building enzymes most adult tissues keep off, and in fatty environments (breast, omentum, bone marrow) they also steal lipids from neighbouring fat cells. This links obesity to cancer and offers new drug targets.

A factory that must build its own walls. Normal adult tissue buys wall panels from a supplier (dietary fat); cancer reopens its own panel plant (FASN) and, when it lands in a fat depot, simply strips panels off the buildings next door.

Stage 5.4

Hypoxia and HIF

When oxygen runs low, HIF proteins switch on a survival programme: new vessels, more glucose uptake, escape. VHL normally destroys HIF; kidney cancer loses VHL and runs the programme permanently. Belzutifan blocks HIF-2α directly.

Light up a product:
OxygenPHD hydroxylasesVHL E3 ligaseHIF-2αHIF-1β (ARNT)VEGF, CAIX, GLUT1, cyclin…Angiogenesis, glycolysis,…activatesinhibitsdruggable target (click)hit by selected productescape route
VHL / HIF oxygen sensingThe VHL/HIF pathway is how cells sense oxygen (the 2019 Nobel Prize). VHL destroys HIF when oxygen is present. Kidney cancers lose VHL, so HIF-2α is permanently on and drives blood vessel growth and proliferation.

VHL is the shredder that destroys the 'we are suffocating' memo whenever there is oxygen around. Kidney cancer breaks the shredder, so the memo piles up and the cell keeps ordering new blood vessels and sugar.

Open questions
  • Why does HIF-2α inhibition work in RCC and not in other hypoxic tumours?
  • Can hypoxia imaging select patients for hypoxia-activated prodrugs or dose-painted radiotherapy?
Stage 5.5

Angiogenesis and VEGF

A tumour cannot grow past a couple of millimetres without new blood vessels. VEGF calls them in; the result is a leaky, chaotic network that starves the tumour, blocks drugs and lets cells escape. Anti-VEGF drugs cancel the order and, at the right dose, straighten the roads that remain.

Light up a product:
Hypoxia (HIF), RAS, p53 l…VEGF, FGF2, ANG2, PDGFTSP-1, endostatinAngiogenic switchTip / stalk sprouting (VE…Pericytes (PDGFRβ)Leaky, chaotic vesselsVessel co-optionHypoxia, poor delivery, e…TAMs, MDSCs (bypass)activatesinhibitsdruggable target (click)hit by selected productescape route
The angiogenic switch & tumour vesselsA tumour cannot grow beyond a couple of millimetres without its own blood supply. The 'switch' flips when the signals calling for new vessels (VEGF, FGF, angiopoietin) outweigh the ones holding them back (thrombospondin). The vessels that result are leaky and chaotic, which starves the tumour of oxygen, blocks drugs, and gives cancer cells a way out.

A new housing estate demanding roads. Once the developers (VEGF) outvote the planners (thrombospondin), roads are laid overnight: badly, with dead ends and potholes, so deliveries (oxygen, drugs) fail and the estate's residents can slip out onto the motorway. Anti-VEGF drugs do not close the roads; used well they make the few that remain drivable.

Light up a product:
HypoxiaHIF-1α / HIF-2αVEGF-AVEGFR2 (endothelium)PLCγ / MAPK / PI3KAngiogenesis, permeabilityImmune suppression (DC, T…activatesinhibitsdruggable target (click)hit by selected productescape route
VEGF angiogenesisHow tumours grow their own blood supply. Low oxygen makes cells release VEGF, which tells blood-vessel cells to sprout toward the tumour.

A growing town (tumour) that keeps sending out road-building orders (VEGF). Anti-angiogenic drugs cancel the orders; at the right dose the roads that remain are straighter and better, so police (immune cells) and supplies (drugs) get in.

Open questions
  • Is the benefit of PD-1×VEGF bispecifics due to the VEGF arm, the dose, or the format?
  • Can vascular normalisation be timed and measured in patients?
Stage 5.6

Nutrient competition

Tumour and immune cells eat from the same plate. Cancer hoards glucose, dumps lactate and acid, and burns tryptophan and arginine into by-products that paralyse T cells. Fixing the food fight is part of making immunotherapy work.

Light up a product:
Glycolytic tumour cellGlucose, glutamine deplet…Lactate, acidity (MCT4)CD39 → CD73 → adenosineIDO1 → kynurenineArginase (MDSC, TAM)T-cell / NK dysfunctionM2 macrophage polarisationHypoxiaactivatesinhibitsdruggable target (click)hit by selected productescape route
Nutrient competition & metabolic immunosuppressionTumours and immune cells eat from the same plate. Cancer cells hoard glucose and glutamine, dump lactate and acid, and burn tryptophan and arginine into by-products that paralyse T cells. The tumour wins the food fight, and the immune system loses before it has fired a shot.

A buffet where the tumour arrives first, eats the protein, and leaves the table sticky with lactate. The immune guests arrive hungry and find nothing but by-products that make them drowsy.

Open questions
  • After IDO1 failed, which metabolic checkpoint (adenosine, arginase, lactate) is worth a phase 3?
  • Can CAR-T cells be engineered to thrive in a nutrient-poor, acidic tumour?
Chapter 6 · 7 stages

Escaping the immune system

Every tumour that exists has already beaten the immune system once. It hides its antigens, raises checkpoint brakes, keeps T cells out, recruits myeloid bodyguards, soaks the neighbourhood in TGF-β, shields itself from complement and dodges NK cells. Immunotherapy works when it undoes the particular trick a tumour used.

  1. 6.1 Antigen presentation
  2. 6.2 Checkpoints: PD-1, CTLA-4, LAG-3
  3. 6.3 Cold tumours: deserts and exclusion
  4. 6.4 Myeloid suppression
  5. 6.5 TGF-β
  6. 6.6 Complement
  7. 6.7 NK-cell evasion
Stage 6.1

Antigen presentation

Cells pin fragments of their proteins on MHC molecules like wanted posters. T cells read them. Tumours take the posters down (MHC or B2M loss) or lose the mutant proteins that made them visible. Engagers and CAR-T bypass the posters entirely.

Open questions
  • What makes a neoantigen truly immunogenic, so vaccines can be designed rather than guessed?
  • Can MHC-I be re-expressed pharmacologically in tumours that lost it?
Stage 6.2

Checkpoints: PD-1, CTLA-4, LAG-3

T cells carry brakes so they do not attack the body. Tumours lean on them: PD-L1 on their surface, Tregs with CTLA-4, LAG-3 and TIGIT on exhausted T cells. Checkpoint inhibitors release the brakes; exhausted cells that are only tired recover, the broken do not.

Light up a product:+4 more via the Connected tab
Dendritic cell (antigen +…TCR – peptide/MHCCD28 – B7 (go)CTLA-4 (stop)Activated CD8 T cellIFN-γPD-L1 on tumourPD-1 on T cellLAG-3 / TIGITTumour cell killingactivatesinhibitsdruggable target (click)hit by selected productescape route
PD-1 / PD-L1 immune checkpoint & T-cell activationHow T cells decide to attack. A T cell needs to see the target (TCR-MHC) and get a 'go' signal (CD28). PD-1 and CTLA-4 are 'stop' signals; tumours exploit them. Checkpoint inhibitors remove the stop.

A soldier needs a target in the sights (TCR) and an order to fire (CD28). CTLA-4 is a commander revoking orders during training; PD-1 is a white flag the enemy waves on the battlefield that makes the soldier lower their weapon. Checkpoint inhibitors tear up the white flag.

Light up a product:+5 more via the Connected tab
Chronic antigen, no helpTCF1+ progenitor (stem-li…Transitory effectorTerminally exhaustedTOX, NR4A, NFATDNMT3A epigenetic scarPD-1, LAG-3, TIM-3, TIGITCytokines, killingCheckpoint blockadeCAR-T exhaustionactivatesinhibitsdruggable target (click)hit by selected productescape route
T-cell exhaustionT cells that see their target for weeks on end without winning gradually shut down: they raise a set of brakes (PD-1, LAG-3, TIM-3, TIGIT), lose their ability to kill, and eventually lock this state into their DNA. Checkpoint drugs rescue the ones that are only partly exhausted; the terminally exhausted are beyond reach.

A soldier posted at a wall for months with no relief. First tired, then unwilling to fire, finally unable to, and the last stage is written into their habits so deeply that no order can undo it. Checkpoint inhibitors work on the tired, not the broken.

Stage 6.3

Cold tumours: deserts and exclusion

Three immune weathers: inflamed (T cells inside), excluded (stuck at the edge), desert (none). Most common cancers are cold. Radiation, viruses, STING agonists and vessel-opening drugs try to warm them; engagers and CAR-T bring their own T cells.

Light up a product:+23 more via the Connected tab
Low TMB, MHC lossβ-catenin, PTEN loss → no…CXCL9/10 silenced (EZH2)Immune desertTGF-β CAFs, collagenAbnormal vessels (VEGF)Immune exclusionInflamed → PD-1 responseRT, STING, viruses, vacci…Myeloid barrieractivatesinhibitsdruggable target (click)hit by selected productescape route
Cold tumours: immune deserts and exclusionTumours come in three immune weathers: inflamed (T cells inside, checkpoint drugs work), excluded (T cells stuck at the edge), and desert (no T cells at all). Most common cancers are excluded or desert, and turning them 'hot' is the central problem of immunotherapy.

Three kinds of town: one where the police already patrol the streets (inflamed), one where they mill about outside a wall (excluded), and one with no police station at all (desert). Removing the officers' handcuffs (PD-1 blockade) only helps in the first; the second needs a gate, the third needs recruitment.

Radiation, chemo, ADC pay…Cytosolic dsDNA / micronu…cGAScGAMPENPP1STINGTBK1 → IRF3 / NF-κBType I IFN, CXCL10 → T-ce…activatesinhibitsdruggable target (click)hit by selected productescape route
cGAS–STING innate sensingcGAS-STING is the cell's alarm for DNA in the wrong place. Radiation, chemotherapy, and ADCs spill DNA into the cytoplasm; cGAS detects it, STING sounds the alarm, and interferon calls in the immune system.

A smoke detector wired to the fire brigade. DNA in the cytoplasm is smoke; cGAS is the detector; STING is the alarm bell; interferon is the 999 call that brings the immune system. Many tumours have quietly removed the batteries.

Stage 6.4

Myeloid suppression

Tumours recruit macrophages and immature myeloid cells and re-train them as bodyguards that switch off T cells, build vessels, and obey 'don't eat me' signals such as CD47.

Light up a product:+5 more via the Connected tab
CSF1, CCL2, G-CSF, VEGFMonocytes, neutrophilsMDSCs (arginase, ROS)TAMs (CSF1R, TREM2)IL-10, TGF-β, PD-L1T cells suppressedCD47 → SIRPα 'don't eat'Phagocytosis (ADCP)Angiogenesis, metastasisCD40 agonists, TLRs repro…activatesinhibitsdruggable target (click)hit by selected productescape route
Myeloid suppression: TAMs, MDSCs & don't-eat-me signalsTumours recruit the body's clean-up cells (macrophages and immature myeloid cells) and re-train them as bodyguards. They switch off T cells, build vessels, and, when a therapeutic antibody flags a cancer cell for eating, are told 'don't eat me' by CD47 on its surface.

A landlord who hires the neighbourhood's own bouncers, pays them in sugar and lactate, and hangs a sign on every door that reads 'don't touch, friend' (CD47). The police (T cells) are stopped at the door by the bouncers, and the cleaners (macrophages) read the sign and leave.

Light up a product:+14 more via the Connected tab
Tumour cellsCAFs (FAP+)TAMs (CSF1R, CD47 axis)MDSCsTregs (CTLA-4)Exhausted CD8 T cells (PD…Abnormal vessels (VEGF)Stiff ECM / desmoplasiaNervesactivatesinhibitsdruggable target (click)hit by selected productescape route
Tumour microenvironment (TME)A tumour is not just cancer cells. It is a neighbourhood of fibroblasts, immune cells, blood vessels, nerves, and scaffolding that the cancer recruits and corrupts, and that decides whether drugs and immune cells can get in.

A castle under siege from the inside: the cancer conscripts the town's builders (fibroblasts) to raise walls, bribes the guards (macrophages) to look away, and diverts the water supply (vessels) so that reinforcements (T cells, drugs) never arrive.

Open questions
  • After magrolimab, is CD47 still a target with the right format?
  • Can macrophages be re-educated rather than depleted?
Stage 6.5

TGF-β

A growth factor that starts as a brake and becomes an accelerator: late in cancer it drives invasion, activates fibroblasts, and walls T cells out of the tumour.

Latent TGF-β (activated b…TGFBR2 / ALK5SMAD2/3–SMAD4Cytostasis (early)EMT, CAF activationT-cell exclusionSMAD4 loss (PDAC)activatesinhibitsdruggable target (click)hit by selected productescape route
TGF-β signallingA signal that stops normal cells from dividing but, once a cancer is established, switches sides: it builds scar-like stroma, walls out immune cells, and pushes cells into a migratory state.

A town planner who first refuses all new building (tumour suppressor) and then, corrupted, builds walls and moats around the tumour that keep the police out (immune exclusion).

Acts here
Open questions
  • Is TGF-β blockade salvageable with the right selectivity (TGF-β1 only) or local delivery?
  • Which patients have TGF-β-driven exclusion versus other causes?
Stage 6.6

Complement

A cascade of blood proteins that punches holes in cells flagged by antibodies and calls in inflammatory cells. Rituximab uses it; tumours shield themselves with CD55 and CD59, and the cascade's own by-products recruit suppressive myeloid cells.

Light up a product:+3 more via the Connected tab
IgG1 antibody (rituximab)C1q classical routeAlternative / lectinC3 convertase → C3bC5 → C5a + MAC (C5b-9)CD46, CD55, CD59 shieldsCDC lysis, opsonisationC5a → MDSC, neutrophilsCD8 T cells suppressedADCC (NK, CD16)activatesinhibitsdruggable target (click)hit by selected productescape route
Complement in cancerComplement is a cascade of blood proteins that punches holes in things marked by antibodies and calls in inflammatory cells. Therapeutic antibodies such as rituximab use it to kill cancer cells; tumours defend themselves with shields (CD46, CD55, CD59), and the cascade's own by-products (C5a) can recruit the myeloid cells that protect the tumour.

A demolition crew that follows the flags an antibody plants on a building: they blow holes in the walls (membrane attack complex) and call in the bulldozers (phagocytes). Tumours paint over the flags (CD55, CD59), and the noise of the demolition attracts the wrong kind of crowd (C5a-recruited suppressor cells).

Open questions
  • Should therapeutic antibodies be engineered for complement or against it?
  • Does C5a blockade add to checkpoint inhibitors in patients?
Stage 6.7

NK-cell evasion

Natural killer cells hunt cells that lost their identity papers (MHC-I) or show stress flags. Tumours that hide from T cells by dropping MHC-I become visible to NK cells unless they also shed the flags or borrow a second badge (HLA-E).

Light up a product:+3 more via the Connected tab
MHC-I → KIR (inhibit)HLA-E → NKG2A (inhibit)MICA/B, ULBP → NKG2DNK cell decisionCD16 ← IgG1 antibody (ADC…TIGIT vs DNAM-1 (CD155)Perforin, granzyme, IFN-γMICA shedding, TGF-βCAR-NK, IL-15, NK engagersactivatesinhibitsdruggable target (click)hit by selected productescape route
NK-cell recognition: missing self & stress ligandsNatural killer cells patrol for cells that have lost their identity papers (MHC-I) or that display stress flags. Cancers that hide from T cells by dropping MHC-I become visible to NK cells, unless they also shed the stress flags, wrap themselves in a second inhibitory badge (HLA-E), or soak the neighbourhood in TGF-β.

Guards who stop anyone not wearing a staff badge (MHC-I) or anyone visibly panicking (stress ligands). Cancer's trick against T cells (throwing away the badge) makes it conspicuous to these guards, so successful tumours also learn to stop panicking, borrow a visitor badge (HLA-E) and bribe the guards with TGF-β.

Open questions
  • Can off-the-shelf CAR-NK cells persist long enough to matter in solid tumours?
  • Does NKG2A blockade work outside head and neck cancer?
Chapter 7 · 7 stages

Invasion and metastasis

Metastasis causes about nine in ten cancer deaths, and no approved drug targets it directly. Cells loosen their grip, cut a path, enter the blood, survive a brutal journey, land where the soil was prepared for them, often sleep for years, and grow in the organ whose welcome matches their programme. Each step is a bottleneck we could learn to hold.

  1. 7.1 Epithelial-mesenchymal transition
  2. 7.2 Invasion
  3. 7.3 Intravasation and circulating tumour cells
  4. 7.4 The pre-metastatic niche
  5. 7.5 Dormancy
  6. 7.6 Organ tropism
  7. 7.7 The brain barrier
Stage 7.1

Epithelial-mesenchymal transition

A brick in a wall becomes a nomad: it lets go of its neighbours, packs pumps to spit out drugs, and puts on camouflage. Most tumour cells stop halfway, in a partial state that is the most dangerous of all.

TGF-β, Wnt, Notch, hypoxiaZEB1/2, SNAIL, TWISTE-cadherin, claudinsVimentin, N-cadherin, MMPsABCB1 / ABCG2 effluxStemness, apoptosis resis…Immune exclusionInvasion, metastasis, dru…activatesinhibitsdruggable target (click)hit by selected productescape route
Epithelial–mesenchymal transition & drug effluxHow a cancer cell changes shape to migrate and to shrug off drugs. Transcription factors like ZEB1 and SNAIL loosen the cell, switch on pumps that eject chemotherapy, and hide it from the immune system.

A brick in a wall (epithelial cell) turning into a nomad: it lets go of its neighbours, packs pumps to spit out poison, and puts on camouflage. The wall-brick was easy to hit; the nomad is not.

Open questions
  • Is EMT required for metastasis in patients, or mainly for drug resistance?
  • Can partial-EMT states be targeted through their ferroptosis sensitivity?
Stage 7.2

Invasion

Grip the scaffolding (integrins), dissolve a path (MMPs), haul forward (myosin), often along tracks that fibroblasts cut first. The protease blockers of the 1990s failed; the grip (FAK) is the modern target.

Light up a product:
TGF-β, HGF, hypoxia, stif…EMT programme (ZEB1, SNAI…Integrins → FAK / SRCInvadopodia, MT1-MMPMMP2/9, uPA → ECM breachRHO–ROCK contractionAmoeboid squeezingCollective invasion (lead…CAF tracksInvasive front → vesselsactivatesinhibitsdruggable target (click)hit by selected productescape route
Invasion: proteases, adhesion & the invasive frontTo invade, a cancer cell must grip the scaffolding around it, dissolve a path with enzymes, and pull itself forward, alone or in a chain led by a scout cell. Fibroblasts often cut the trail first. The enzyme blockers of the 1990s failed; today's targets are the grip (integrins, FAK) and the trail-makers.

A climber in a collapsing tunnel: grip the wall (integrins), chip away the rock ahead (MMPs), and haul forward (myosin). Some climbers squeeze through cracks without chipping (amoeboid). Often a guide (a fibroblast) has already carved the passage.

Open questions
  • Does FAK inhibition reduce metastasis, or only soften stroma for other drugs?
  • Can perineural invasion be blocked pharmacologically?
Stage 7.3

Intravasation and circulating tumour cells

Getting into the blood and surviving there kills all but one cell in a thousand. Survivors travel in clusters, under a cloak of platelets, or with a neutrophil escort. Liquid biopsies catch what is left.

Light up a product:
TMEM doorway (macrophage)Leaky vessels (VEGF)IntravasationAnoikis (detachment death)Shear, oxidative stressNK-cell clearanceSurviving CTCs / clustersPlatelet cloak, NETsClusters (plakoglobin, CD…Liquid biopsy detectionactivatesinhibitsdruggable target (click)hit by selected productescape route
Intravasation & circulating tumour cellsGetting into the bloodstream and surviving there is brutal: cells are ripped from their neighbours, battered by flow, and hunted by NK cells. Fewer than one in a thousand survive. The ones that do travel in clusters, wear a cloak of platelets, or ride with neutrophils. Liquid biopsies catch what is left.

Leaving a fortress through the drains and swimming a river in flood while archers (NK cells) fire from the bank. Survivors go in rafts (clusters), under wet blankets (platelets), and with a friendly escort (neutrophils).

Stage 7.4

The pre-metastatic niche

Before a single cancer cell arrives, the primary tumour sends parcels ahead: vesicles and hormones that recruit bone-marrow cells to a distant organ and turn it into fertile soil.

Light up a product:
Primary tumour (hypoxic)Exosomes (integrins, MIF)VEGF, G-CSF, LOX, S100A8/9Bone-marrow cells (VEGFR1…Resident cells: Kupffer, …Niche: fibronectin, MMP9,…Arriving CTCsColonisationMDSCs, immune suppressionactivatesinhibitsdruggable target (click)hit by selected productescape route
The pre-metastatic nicheBefore a single cancer cell arrives, the primary tumour sends parcels ahead: tiny vesicles (exosomes) and hormones that recruit bone-marrow cells to a distant organ and remodel it into fertile soil. By the time the seed lands, the bed is already made.

A colonising power that sends engineers, seed and fertiliser to a distant shore before the settlers sail. The settlers (CTCs) that land where the soil has been prepared survive; the ones that land elsewhere starve.

Primary tumourInvasion (EMT, MMPs)IntravasationCTCs in circulationExtravasationDTC dormancyColonisation / macrometas…Pre-metastatic niche (exo…NK / T-cell clearanceactivatesinhibitsdruggable target (click)hit by selected productescape route
The metastatic cascadeHow cancer spreads: cells leave the tumour, squeeze into blood or lymph vessels, survive the journey, exit into a new organ, often sleep there for years, and finally grow. Metastasis causes about 90% of cancer deaths.

A seed leaving a plant: it must detach, ride the wind, land somewhere with the right soil, survive the winter, and only then sprout. Almost every seed fails; the few that grow are the metastases.

Open questions
  • Can exosome profiling predict the organ of relapse in patients?
  • Is there a drug that stops niche formation without the harms of adjuvant anti-VEGF?
Stage 7.5

Dormancy

Disseminated cells can sleep for years, held quiet by their niche and watched by immune cells, then wake after inflammation, injury or ageing. Late relapse in breast and prostate cancer is dormancy ending.

Disseminated tumour cellQuiescence (p38↑, ERK↓, N…Niche: TGF-β2, BMP7, endo…NK / T-cell surveillanceAwakening: NETs, inflamma…Late relapseactivatesinhibitsdruggable target (click)hit by selected productescape route
Tumour dormancyCancer cells can hide in bone marrow, lung, or brain for years or decades, asleep and invisible to scans and chemotherapy, then wake up. Late relapse in breast and prostate cancer is dormancy ending.

Seeds that stay in the soil for years waiting for the right spring. You can keep the ground cold (maintenance therapy), force them to sprout and mow them (wake-and-kill), or dig them out (immune clearance).

Open questions
  • Should we wake dormant cells to kill them, or keep them asleep for life?
  • What wakes them: surgery, infection, inflammation, ageing?
Stage 7.6

Organ tropism

Breast cancer goes to bone, lung, liver and brain; prostate to bone; colon to liver. Seed and soil: the cell's programme and the organ's welcome. Bone's vicious cycle of tumour and osteoclast is the one we can already drug.

CTC (seed programme)Blood-flow anatomyBone: RANKL vicious cycleLung: tenascin C, periost…Liver: Kupffer, stellate …Brain: astrocytes, BBBTGF-β, IGF-1 releasedOrgan-specific colonisati…Denosumab, radium-223SBRT, HIPEC, TAREactivatesinhibitsdruggable target (click)hit by selected productescape route
Organ tropism: seed and soilBreast cancer goes to bone, lung, liver and brain; prostate cancer to bone; colon cancer to liver; uveal melanoma almost only to liver. Paget's 1889 idea still holds: where a cancer spreads depends on both the seed (the cell's programme) and the soil (the organ's welcome). Each soil has its own vicious cycle, and some are druggable.

Dandelion seeds fall everywhere, but only take root where the soil suits them. Bone is a greenhouse for breast and prostate seeds because the seeds can trick the gardeners (osteoclasts) into digging up food for them.

Open questions
  • Does treating oligometastases with SBRT change survival across cancers, or only in selected ones?
  • Can tropism be predicted from the primary tumour to target surveillance?
Stage 7.7

The brain barrier

The brain's vessels are sealed tight and fitted with pumps that eject most drugs. Cancer cells that squeeze through recruit astrocytes to feed and shield them. Brain-penetrant pills, focused ultrasound and radiosurgery are the answers so far.

Light up a product:+2 more via the Connected tab
CTC arrest at capillaryBBB: tight junctions, per…P-gp / BCRP effluxExtravasation (cathepsin …Vascular co-option (L1CAM)Astrocytes: Cx43, cGAMPBrain metastasisDrugs excludedBrain-penetrant TKIs, FUSactivatesinhibitsdruggable target (click)hit by selected productescape route
The blood–brain barrier & brain metastasisThe brain's blood vessels are sealed tight and fitted with pumps that eject most drugs. That protects the brain from poisons but also from chemotherapy and antibodies. Cancer cells that do squeeze through recruit the brain's own support cells, astrocytes, to feed and shield them.

A gated city with customs officers who throw out most goods (efflux pumps). Smugglers who get in bribe the local guards (astrocytes) to feed them. Drugs that work here are either small enough to slip past customs or arrive by a special convoy (focused ultrasound, intrathecal delivery).

Open questions
  • Why do some large molecules (T-DXd) work in the brain when the barrier should exclude them?
  • Can focused ultrasound deliver antibodies and cell therapies to brain tumours routinely?
Chapter 8 · 5 stages

The tumour ecosystem

A tumour is a corrupted organ: cancer cells plus the fibroblasts, matrix, vessels, nerves, microbes and immune cells they recruit, and the signals they send to the rest of the body. The ecosystem decides whether drugs and immune cells get in, and it is why the same mutation behaves differently in different tissues.

  1. 8.1 Fibroblasts and the extracellular matrix
  2. 8.2 Vasculature
  3. 8.3 Nerves
  4. 8.4 Microbiome
  5. 8.5 Cachexia signals
Stage 8.1

Fibroblasts and the extracellular matrix

Tumours keep the body's repair cells in wound-healing mode forever. The scar they lay down squeezes vessels shut, walls out immune cells, and its very stiffness tells cancer cells to grow. Demolishing it made pancreatic cancer worse; retraining it is the new plan.

Light up a product:+5 more via the Connected tab
Tumour: TGF-β, PDGF, Hh, …Fibroblast / stellate cellmyCAF (FAP, αSMA, collage…iCAF (IL-6, CXCL12, LIF)Collagen, HA, LOX crossli…Stiffness → FAK → YAP/TAZPressure: vessels collapseT-cell exclusionGrowth, EMT, chemoresista…Hedgehog (SMO) paradoxactivatesinhibitsdruggable target (click)hit by selected productescape route
Fibroblast activation, desmoplasia & matrix stiffnessTumours recruit the body's repair cells, fibroblasts, and keep them in wound-healing mode forever. The scar tissue they lay down (desmoplasia) squeezes blood vessels shut, walls out immune cells, stiffens the tissue in a way that itself tells cancer cells to grow, and is why pancreatic cancer is so hard to treat.

Builders hired to repair a wall who never stop: they pour concrete around the tumour until the roads are blocked (vessels), the police cannot get in (T cells), and the very hardness of the concrete tells the tenants to multiply. Demolishing the builders' work made things worse; the newer plan is to retrain them.

Stage 8.2

Vasculature

Chaotic, leaky vessels create pockets of hypoxia, high pressure that collapses capillaries, and easy exits for cancer cells. Normalising rather than destroying them improves drug and immune delivery.

Open questions
  • How do we measure vessel normalisation in patients to time combination therapy?
  • Why do anti-angiogenics fail in adjuvant settings?
Stage 8.3

Nerves

Tumours grow their own nerve supply and use it: adrenaline and acetylcholine signal growth, gliomas form synapses with neurons, and nerve invasion causes pain. Old drugs (beta blockers) are being retested.

TumourNGF, axon guidance cuesNerve ingrowth (PNI)β-adrenergic / cholinergi…Neuron–glioma synapses (A…Growth, invasion, immunos…activatesinhibitsdruggable target (click)hit by selected productescape route
Cancer neuroscience (nerve–tumour signalling)Cancer neuroscience is the study of how tumours talk to nerves. Nerves grow into tumours and feed them signals; brain tumours even wire themselves into neural circuits. Cutting the conversation with common drugs such as beta-blockers is now being tested.

A tumour that taps into the city's telephone lines: it receives growth orders from the nervous system, and brain tumours go further, plugging themselves into the switchboard so that ordinary brain chatter feeds them.

Open questions
  • Do beta blockers improve survival in randomised trials?
  • Can neuron-glioma synapses be blocked with epilepsy drugs?
Stage 8.4

Microbiome

Gut bacteria shape whether immunotherapy works, bacteria inside tumours degrade chemotherapy and inflame tissue, and one strain's toxin leaves a mutational fingerprint. Diet, antibiotics and stool transplants all move the needle.

Light up a product:
Gut microbiotaMetabolites (SCFA, inosin…Dendritic / T-cell primingImmunotherapy responseIntratumoural bacteriaChemotherapy degradation,…AntibioticsColibactin → mutational s…activatesinhibitsdruggable target (click)hit by selected productescape route
Microbiome–tumour interactionsThe bacteria in the gut, and even inside tumours, influence whether cancer starts and whether immunotherapy works. Transplanting stool from responders has made some non-responders respond.

Soil bacteria decide whether a garden thrives. Some feed the plants' defenders, some produce poisons, and some even eat the pesticide before it reaches the weeds.

Stage 8.5

Cachexia signals

Tumours send hormones (GDF-15, IL-6) that switch off appetite and melt muscle and fat. Cachexia kills a fifth of patients directly and stops treatment in many more; the first drugs against it are now in phase 3.

Tumour + inflammationGDF-15 → GFRAL (brainstem)IL-6 / TNF / activinAnorexiaMuscle proteolysis, fat l…Weight loss, frailty, dea…activatesinhibitsdruggable target (click)hit by selected productescape route
Cancer cachexiaThe wasting syndrome that kills up to a third of cancer patients: tumours send hormonal signals (GDF-15, IL-6) that switch off appetite and burn muscle and fat. The first drug to reverse it, ponsegromab, showed weight gain in 2024.

A thermostat hijacked by the tumour: it tells the body it is full when it is starving and orders the furnace to burn muscle for fuel. Ponsegromab cuts the wire to the thermostat.

Chapter 9 · 5 stages

Why treatments fail

Every cancer drug eventually meets resistance. Tumours change the part the drug binds, take side roads, switch engines, hide in sanctuaries, or pump the drug out. Some cells simply go quiet and wait. Understanding the route a tumour took decides the next move; the resistance atlas lists them class by class.

  1. 9.1 Resistance mechanics
  2. 9.2 Heterogeneity
  3. 9.3 Persisters
  4. 9.4 Efflux pumps
  5. 9.5 Lineage plasticity
Stage 9.1

Resistance mechanics

Five routes back when a pathway is blocked: mutate the target, make more of it, bypass, mutate downstream, or change identity. Plus the pharmacological escapes: pumps, sanctuaries, lost antigens.

Light up a product:+37 more via the Connected tab
Drug blocks targetTarget → signal → growth1 Target mutation / amp2 Bypass RTK (MET, HER3)3 Downstream (PIK3CA, RB1)4 Lineage switch, persist…5 Efflux, sanctuary, anti…Regrowth under therapyNext-gen, vertical combos…activatesinhibitsdruggable target (click)hit by selected productescape route
Resistance routes: how a blocked pathway comes backWhen a drug blocks a cancer's engine, the cancer has five ways back: change the part the drug binds, make more of it, take a side road, switch to a different engine altogether, or stop letting the drug in. Knowing which route a tumour took decides the next drug.

Blocking a motorway. Traffic re-routes through a changed junction (target mutation), an extra lane (amplification), a parallel A-road (bypass), a road further along (downstream), a different form of transport (lineage switch), or simply avoids the roadblock's jurisdiction (efflux, sanctuary sites).

Stage 9.2

Heterogeneity

A biopsy samples one place at one time; the tumour is many places changing over time. Subclones with different drivers coexist, and the one that survives treatment was often a minority nobody sequenced.

Open questions
  • How many regions or how much blood must be sampled to see the clones that matter?
  • Can heterogeneity itself be reduced (anti-CIN drugs) to slow evolution?
Stage 9.3

Persisters

Even when a drug wipes out 99% of a tumour, a few cells survive without any resistance mutation. They go quiet, stop dividing, and wait; true resistance often grows out of them. They are hard to kill because they do so little, but they have weaknesses of their own.

Light up a product:+8 more via the Connected tab
Targeted drug or chemoBulk tumour diesPersister: slow-cycling, …KDM5A, H3K27me3, YAP, NF-…FAO, low GSH → GPX4 depen…Efflux, autophagy, BCL-XLAPOBEC → resistance mutat…Relapse (MRD → clinical)Ferroptosis inducersUpfront combinations, hol…activatesinhibitsdruggable target (click)hit by selected productescape route
Drug-tolerant persister cellsEven when a drug wipes out 99% of a tumour, a few cells survive without any resistance mutation: they go quiet, stop dividing, and wait. These persisters are the seed of relapse. They are hard to kill precisely because they are not doing much, but they have their own weaknesses.

Bears in hibernation while the forest burns. Poison meant for grazing animals does nothing to a sleeping bear; but a hibernating bear cannot run, so a hunter who knows where the den is (GPX4, BCL-XL) can strike.

Open questions
  • Do upfront combinations prevent persisters or just delay the same evolution?
  • Which persister dependency (GPX4, BCL-XL, KDM5) is the first to reach a phase 3?
Stage 9.4

Efflux pumps

Cancer cells can install pumps that throw chemotherapy back out. The same pumps guard gut, brain and marrow, which is why blocking them failed and why ADC designers now pick payloads the pumps cannot grip.

Light up a product:
Chemo / payload entersABCB1 (P-gp)ABCG2 (BCRP), ABCC1Drug exported (ATP)Sub-lethal intracellular …EMT, hypoxia, NRF2 induceBBB, stem cells, marrowNon-substrate payloads (D…Cell deathactivatesinhibitsdruggable target (click)hit by selected productescape route
Drug efflux pumps (ABC transporters)Cancer cells can install pumps in their outer membrane that throw chemotherapy back out as fast as it comes in. The same pumps guard the gut, brain and bone marrow in healthy tissue, which is why blocking them failed as a strategy and why drug designers now choose payloads the pumps cannot grip.

A nightclub bouncer who throws out anyone in a particular jacket. Sacking the bouncer (P-gp inhibitors) also emptied the club of the staff who kept the place safe (gut, brain, marrow). The fix was to change jackets: payloads the bouncer does not recognise.

Open questions
  • Which payload should follow which after an ADC fails: is efflux the reason for cross-resistance?
  • Can pump expression be measured on a biopsy to guide payload choice?
Stage 9.5

Lineage plasticity

Under a drug that blocks its identity, a tumour can become a different kind of cell, most dramatically a small-cell neuroendocrine cancer that no longer needs the blocked signal. Not a new mutation in the engine: a new engine.

Light up a product:+18 more via the Connected tab
Adenocarcinoma (AR / EGFR)ARPI or EGFR TKI pressureTP53 + RB1 lossSOX2, EZH2, ASCL1/NEUROD1Neuroendocrine / small-ce…DLL3, B7-H3, SEZ6 surfaceAR / EGFR indifferentTarlatamab, platinum-etop…EZH2 inhibitors block swi…activatesinhibitsdruggable target (click)hit by selected productescape route
Lineage plasticity & neuroendocrine transformationUnder pressure from a drug that blocks its identity (the androgen receptor in prostate cancer, EGFR in lung cancer), a tumour can change what kind of cell it is, becoming a small-cell neuroendocrine cancer that no longer needs the blocked signal. It is the ultimate escape: not a new mutation in the engine, but a new engine.

A shop that sells hats is fined every time it sells a hat (AR blockade). One day it reopens as a bakery. The fine no longer applies, the old inspectors (PSA, PSMA scans) see nothing, and only a new set of tools works against the new business.

Light up a product:
Stem-like stateDifferentiated bulkMesenchymal / drug-tolera…Lineage switch (NE transf…EZH2, SWI/SNF, TP53/RB1 l…Wnt / Notch / Hedgehog ni…activatesinhibitsdruggable target (click)hit by selected productescape route
Cancer stem cells & phenotypic plasticitySome cancer cells behave like stem cells: they can regrow the whole tumour, resist treatment, and switch identities. This plasticity explains why tumours come back and why some lung and prostate cancers transform into a different cancer type under therapy.

A weed that can turn from leaf to root to seed depending on what you spray on it. Kill the leaves and the roots wait; kill the roots and a seed reawakens.

How this page is built: the chapters and stages are a curated atlas (src/data/mechanics-atlas.ts); every diagram, drug, target, technology and term is an object in the knowledge graph with its own page and sources. Each diagram is drawn once, at the first stage that needs it, and later stages link back to it. Nothing here is medical advice; see about and methodology.