[{"id":"aml-signalling","kind":"pathway","name":"Acute myeloid leukaemia (KEGG map)","aka":["KEGG hsa05221","Acute myeloid leukemia"],"tldr":"KEGG's AML map shows the two hits that turn a normal blood stem cell into a leukaemia: a growth signal jammed on (FLT3, KIT or RAS) plus a broken maturation switch (fusion proteins such as PML-RARA or AML1-ETO, or mutated CEBPA and RUNX1). Drugs now exist for both halves.","summary":"The KEGG acute myeloid leukaemia map (hsa05221) draws the classic two-hit model of AML. The first hit is disordered growth and survival signalling: internal tandem duplications or point mutations in the receptor tyrosine kinase FLT3, mutations in KIT, and mutations in NRAS or KRAS. These feed three downstream routes, RAS to RAF to MEK to ERK, PI3K to AKT to mTOR, and STAT5, which together drive proliferation and switch off apoptosis. The second hit is a block in myeloid differentiation caused by transcription factor fusions created by chromosomal translocations, AML1-ETO (RUNX1-RUNX1T1) from t(8;21), PML-RARA from t(15;17) and PLZF-RARA, which recruit co-repressors and silence the genes that CEBPA, PU.1 and RUNX1 would normally switch on to make mature granulocytes and monocytes. In other cases CEBPA or RUNX1 themselves are mutated. The review by Döhner et al., N Engl J Med, 2015 (doi:10.1056/NEJMra1406184) frames the same biology in terms of recurrently mutated gene classes (signalling genes, myeloid transcription factors, NPM1, DNA methylation and chromatin genes, and the spliceosome) and notes that FLT3-ITD predicts shorter survival while CEBPA biallelic mutation and NPM1 mutation without FLT3-ITD are favourable.\n\nWhat drugs do about it: the FLT3 inhibitors midostaurin (with induction chemotherapy), gilteritinib and quizartinib switch off the first hit in FLT3-mutant AML. The PML-RARA differentiation block is reversed by all-trans retinoic acid plus arsenic trioxide, which degrade the fusion protein and cure most acute promyelocytic leukaemia without cytotoxic chemotherapy. Venetoclax with azacitidine targets the BCL2 survival dependency that sits downstream of these signals in older or unfit patients.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05221","url":"https://www.kegg.jp/pathway/hsa05221"},{"label":"Review: Acute Myeloid Leukemia (NEJM 2015)","url":"https://doi.org/10.1056/NEJMra1406184"}],"tags":[],"related":["ras-mapk","pi3k-akt-mtor","menin-kmt2a"],"cancers":["aml"],"sections":[],"technologies":[],"targets":["flt3","kit","kras","mek","akt","rxr"],"drugs":["midostaurin","gilteritinib","quizartinib","arsenic-trioxide","venetoclax","azacitidine","revumenib"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","rtk-activation","apoptosis-bcl2","menin-kmt2a","transcription-addiction"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-dohner-n-engl-j-med"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Think of a factory line for blood cells. One fault jams the accelerator (FLT3, KIT or RAS stuck on) so cells keep being made; a second fault breaks the finishing station (PML-RARA or AML1-ETO) so the cells never mature. FLT3 inhibitors release the accelerator; retinoic acid and arsenic repair the finishing station.","nodes":[{"id":"flt3","label":"FLT3 (ITD / TKD)","x":20,"y":6,"targetId":"flt3"},{"id":"kit","label":"KIT","x":50,"y":6,"targetId":"kit"},{"id":"ras","label":"NRAS / KRAS","x":35,"y":24,"targetId":"kras"},{"id":"mek","label":"RAF / MEK / ERK","x":15,"y":44,"targetId":"mek"},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":40,"y":44,"targetId":"akt"},{"id":"stat5","label":"STAT5","x":62,"y":44},{"id":"prolif","label":"Proliferation, survival","x":38,"y":66},{"id":"fusion","label":"AML1-ETO, PML-RARA fusions","x":84,"y":20,"targetId":"rxr"},{"id":"cebpa","label":"CEBPA, PU.1, RUNX1","x":84,"y":48},{"id":"diff","label":"Myeloid differentiation","x":84,"y":74},{"id":"aml","label":"Leukaemic blasts","x":60,"y":92}],"edges":[{"from":"flt3","to":"ras","type":"activates"},{"from":"kit","to":"ras","type":"activates"},{"from":"flt3","to":"stat5","type":"activates"},{"from":"flt3","to":"pi3k","type":"activates"},{"from":"ras","to":"mek","type":"activates"},{"from":"ras","to":"pi3k","type":"activates"},{"from":"mek","to":"prolif","type":"activates"},{"from":"pi3k","to":"prolif","type":"activates"},{"from":"stat5","to":"prolif","type":"activates"},{"from":"fusion","to":"cebpa","type":"inhibits"},{"from":"cebpa","to":"diff","type":"activates"},{"from":"prolif","to":"aml","type":"activates"},{"from":"diff","to":"aml","type":"inhibits"}],"interventions":["FLT3 inhibitors: midostaurin with induction chemotherapy, gilteritinib for relapsed or refractory FLT3-mutant AML, quizartinib for FLT3-ITD AML","Differentiation therapy for PML-RARA acute promyelocytic leukaemia: all-trans retinoic acid plus arsenic trioxide","BCL2 inhibitor venetoclax with azacitidine for patients unfit for intensive chemotherapy","Menin inhibitors (revumenib) for NPM1-mutant and KMT2A-rearranged AML, which reopen the differentiation programme","Intensive chemotherapy (cytarabine and daunorubicin) followed by allogeneic stem cell transplant in fit, higher-risk patients"]},{"id":"ar-signaling","kind":"pathway","name":"Androgen receptor signalling","aka":[],"tldr":"Androgen receptor signalling is prostate cancer's engine. Testosterone becomes DHT, binds the androgen receptor, and drives growth genes. Castration removes the fuel; newer pills block the receptor or the enzyme that makes fuel inside the tumour.","summary":"Testicular testosterone (and adrenal precursors converted intratumourally via CYP17A1) is reduced to DHT by 5α-reductase and binds AR, which translocates to the nucleus and drives PSA, TMPRSS2-ERG, and proliferation genes. GnRH agonists/antagonists shut testicular production; abiraterone blocks CYP17A1; enzalutamide, apalutamide, darolutamide block AR. Castration resistance arises via AR amplification, ligand-binding mutations, splice variants (AR-V7 lacks the ligand domain), glucocorticoid receptor substitution, and lineage plasticity to neuroendocrine phenotype.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Androgen_receptor","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Androgen_receptor"}],"tags":[],"related":[],"cancers":["prostate","tnbc"],"sections":[],"technologies":["androgen-deprivation"],"targets":["androgen-receptor","parp","akt","psma"],"drugs":["capivasertib","olaparib","talazoparib","niraparib"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"Same key-and-lock idea as oestrogen: castration stops making keys, abiraterone shuts the tumour's own key factory, enzalutamide blocks the lock. AR-V7 is a lock with no keyhole that is permanently open.","nodes":[{"id":"gnrh","label":"GnRH → LH → testis","x":20,"y":8},{"id":"testo","label":"Testosterone","x":20,"y":28},{"id":"cyp17","label":"CYP17A1 (adrenal/intratumoural)","x":65,"y":15},{"id":"dht","label":"DHT","x":40,"y":45},{"id":"ar","label":"Androgen receptor","x":40,"y":62,"targetId":"androgen-receptor"},{"id":"arv7","label":"AR-V7 / amplification","x":80,"y":55},{"id":"genes","label":"PSA, TMPRSS2-ERG, growth genes","x":40,"y":80},{"id":"prol","label":"Proliferation","x":40,"y":96}],"edges":[{"from":"gnrh","to":"testo","type":"activates"},{"from":"cyp17","to":"testo","type":"activates"},{"from":"testo","to":"dht","type":"activates"},{"from":"dht","to":"ar","type":"activates"},{"from":"arv7","to":"ar","type":"activates"},{"from":"ar","to":"genes","type":"activates"},{"from":"genes","to":"prol","type":"activates"}],"interventions":["GnRH agonists/antagonists (leuprolide, relugolix)","CYP17A1 inhibitor abiraterone","AR antagonists enzalutamide, apalutamide, darolutamide","PARP inhibitors + ARPI in HRR-mutant disease; capivasertib + abiraterone in PTEN-deficient","AR degraders, N-terminal domain inhibitors (trials)"]},{"id":"antigen-presentation-immunoediting","kind":"pathway","name":"Antigen presentation & immune editing","aka":[],"tldr":"How 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.","summary":"Proteasome-processed peptides are loaded on MHC class I (HLA-A/B/C, B2M) and presented to CD8 T cells; dendritic cells cross-present tumour antigens in lymph nodes. Immunoediting proceeds through elimination, equilibrium, and escape: loss of B2M or HLA, JAK1/2 or IFNGR mutations (blunting interferon-γ response), antigen loss, PD-L1 induction, and immunosuppressive niches. Neoantigen quality, not just quantity, predicts response; HLA genotype shapes which mutations are visible. Escape variants explain acquired resistance to checkpoint blockade and CAR-T (CD19 loss).","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Immunoediting","links":[{"label":"Schreiber, Old & Smyth, Cancer immunoediting (Science 2011)","url":"https://doi.org/10.1126/science.1203486"},{"label":"Zaretsky et al., Mutations associated with acquired resistance to PD-1 blockade (NEJM 2016)","url":"https://doi.org/10.1056/NEJMoa1604958"}],"tags":["mechanism"],"related":[],"cancers":["tnbc","non-hodgkin-lymphoma"],"sections":[],"technologies":["neoantigen-mrna-vaccine","tcr-t","t-cell-engager","car-t","car-nk-macrophage","checkpoint-inhibitor"],"targets":["pd1","pdl1","cd3","gp100","mage-a4","prame","hla-a"],"drugs":[],"companies":[],"institutions":["wustl-siteman","mskcc","nci","johns-hopkins"],"pathways":["pd1-checkpoint","cgas-sting","jak-stat"],"terms":["neoantigen","hla-a02-restriction","tmb","cold-vs-hot","avoiding-immune-destruction"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-zaretsky-n-engl-j-med"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Schreiber (WashU) coined immunoediting; Chan and Ribas on neoantigen quality and escape (MSK, UCLA); NCI Surgery Branch on neoantigen-reactive T cells; Balachandran (MSK) on long-term vaccine responders."],"analogy":"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).","nodes":[{"id":"prot","label":"Proteasome → peptides","x":15,"y":30},{"id":"tap","label":"TAP transport","x":35,"y":30},{"id":"mhc","label":"MHC-I / B2M loading","x":55,"y":30},{"id":"surf","label":"Peptide-MHC on surface","x":75,"y":30},{"id":"tcr","label":"CD8 T cell (TCR)","x":75,"y":65,"targetId":"cd3"},{"id":"ifn","label":"IFN-γ → JAK1/2 → STAT1","x":45,"y":65},{"id":"pdl1","label":"PD-L1 induction","x":25,"y":85,"targetId":"pdl1"},{"id":"dc","label":"Dendritic cell cross-presentation","x":92,"y":50},{"id":"escape","label":"Escape: B2M/HLA loss, JAK mutation","x":15,"y":60}],"edges":[{"from":"prot","to":"tap","type":"activates"},{"from":"tap","to":"mhc","type":"activates"},{"from":"mhc","to":"surf","type":"activates"},{"from":"surf","to":"tcr","type":"activates"},{"from":"dc","to":"tcr","type":"activates"},{"from":"tcr","to":"ifn","type":"activates"},{"from":"ifn","to":"mhc","type":"activates"},{"from":"ifn","to":"pdl1","type":"activates"},{"from":"pdl1","to":"tcr","type":"inhibits"},{"from":"escape","to":"mhc","type":"inhibits"},{"from":"escape","to":"ifn","type":"inhibits"}],"interventions":["Checkpoint inhibitors; personalised neoantigen vaccines (intismeran) supply antigen; T-cell engagers and CAR-T bypass MHC entirely","TCR-T and ImmTACs (tebentafusp) target intracellular antigens via peptide-HLA","Epigenetic drugs and interferon can re-express MHC; MHC-independent NK-cell therapies address MHC-loss escape"]},{"id":"autophagy","kind":"pathway","name":"Autophagy","aka":[],"tldr":"Autophagy 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.","summary":"Macroautophagy engulfs organelles and proteins into autophagosomes for lysosomal degradation, supplying amino acids and nucleotides under stress. RAS-mutant and pancreatic cancers are autophagy-addicted; KRAS or MEK inhibition further increases autophagy, so hydroxychloroquine or ULK1 inhibitors are combined with trametinib (phase 1/2, PDAC) and with KRAS inhibitors. Autophagy also degrades MHC-I in PDAC (immune evasion) and is context-dependent: tumour-suppressive early, pro-survival late. No selective autophagy drug is approved; hydroxychloroquine achieves inconsistent lysosomal inhibition at tolerated doses.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Autophagy","links":[{"label":"Yamamoto et al., Autophagy promotes immune evasion of pancreatic cancer by degrading MHC-I (Nature 2020)","url":"https://doi.org/10.1038/s41586-020-2229-5"}],"tags":["mechanism"],"related":[],"cancers":["pancreatic"],"sections":[],"technologies":["kras-inhibitors"],"targets":["kras"],"drugs":[],"companies":[],"institutions":["penn-abramson","cold-spring-harbor"],"pathways":["ras-mapk","pi3k-akt-mtor","cancer-metabolism","antigen-presentation-immunoediting"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-yamamoto-nature"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Amaravadi (Penn) on autophagy inhibition trials; Kimmelman (NYU) on PDAC autophagy and MHC-I; Der (UNC) on KRAS/autophagy."],"analogy":"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.","nodes":[{"id":"stress","label":"Nutrient stress, KRAS/MEK inhibition","x":15,"y":30,"targetId":"kras"},{"id":"ulk","label":"ULK1 / AMPK","x":40,"y":30},{"id":"mtor","label":"mTORC1 (inhibits)","x":40,"y":70},{"id":"phag","label":"Autophagosome","x":65,"y":30},{"id":"lys","label":"Lysosome (HCQ blocks)","x":85,"y":30},{"id":"fuel","label":"Recycled fuel → survival","x":85,"y":70},{"id":"mhc","label":"MHC-I degradation (PDAC)","x":65,"y":90}],"edges":[{"from":"stress","to":"ulk","type":"activates"},{"from":"mtor","to":"ulk","type":"inhibits"},{"from":"ulk","to":"phag","type":"activates"},{"from":"phag","to":"lys","type":"activates"},{"from":"lys","to":"fuel","type":"activates"},{"from":"lys","to":"mhc","type":"activates"}],"interventions":["Hydroxychloroquine + MEK inhibitor or + KRAS inhibitor in PDAC (phase 1/2)","ULK1 inhibitors (DCC-3116) in RAS-driven cancers","Autophagy inhibition to restore MHC-I and immunotherapy response (preclinical)"]},{"id":"bcr-signalling","kind":"pathway","name":"B-cell receptor / BTK signalling (to NF-κB)","aka":[],"tldr":"The B-cell receptor is the survival switch of B cells. Signals from it pass through BTK to free NF-kappa-B, which keeps the cell alive. B-cell cancers hold it on; BTK inhibitors, proteasome inhibitors and lenalidomide each cut the line at a different point.","summary":"Antigen binding to the B-cell receptor activates SYK and BTK, then PLCγ2 and PKCβ, which assemble the CARD11-BCL10-MALT1 complex and activate the IKK kinases. IKK phosphorylates IκB, the protein that holds NF-κB (p65/p50) in the cytoplasm; IκB is ubiquitinated and destroyed by the proteasome, and NF-κB enters the nucleus to switch on BCL2, BCL-XL, IL-6, IL-10 and cyclin D. Toll-like receptors feed the same hub through MYD88 (MYD88 L265P in Waldenström macroglobulinaemia and ABC-type DLBCL), and BAFF and CD40 signals activate the alternative (NIK-dependent) branch, which matters in multiple myeloma. Activated B-cell DLBCL, chronic lymphocytic leukaemia, mantle cell lymphoma and Waldenström macroglobulinaemia depend on this circuit. Ibrutinib and the later BTK inhibitors (acalabrutinib, zanubrutinib, pirtobrutinib) block the receptor arm; bortezomib and carfilzomib stop the proteasome from destroying IκB; lenalidomide and its successors degrade IKZF1/3 and cut IRF4-driven NF-κB output in myeloma and ABC-DLBCL. Resistance comes from BTK C481S mutations, PLCγ2 mutations and CARD11 or MYD88 lesions downstream of BTK.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/B-cell_receptor","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/B-cell_receptor"},{"label":"Davis et al., Nature 2010: chronic active B-cell receptor signalling in diffuse large B-cell lymphoma","url":"https://doi.org/10.1038/nature08638"},{"label":"Ngo et al., Nature 2011: oncogenically active MYD88 mutations in human lymphoma","url":"https://doi.org/10.1038/nature09671"},{"label":"Treon et al., N Engl J Med 2012: MYD88 L265P somatic mutation in Waldenstrom macroglobulinaemia","url":"https://doi.org/10.1056/NEJMoa1200710"}],"tags":[],"related":[],"cancers":["dlbcl","cll","mantle-cell-lymphoma","waldenstrom","multiple-myeloma","non-hodgkin-lymphoma","primary-cns-lymphoma"],"sections":[],"technologies":[],"targets":["bcl2"],"drugs":["ibrutinib","bortezomib","lenalidomide","venetoclax"],"companies":[],"institutions":[],"pathways":["inflammation-nfkb","apoptosis-bcl2"],"terms":["lymphoma-bio-lymphgen"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":["Lymphoma: this is the one pathway in lymphoma where the biology picks the drug. Chronic active signalling in activated B-cell-like disease was shown functionally, by knocking down IgM, Ig-kappa, CD79A, CD79B or BTK and killing only those cells, and structurally, by the slow-diffusing receptor clusters that resemble an antigen-stimulated normal B cell; ITAM mutations of CD79B raise surface receptor and blunt the LYN feedback brake, and were present in 18% of activated B-cell-like cases (Davis 2010). The toll-like receptor arm feeds the same hub through MYD88 L265P, present in 29% of activated B-cell-like cases and in 91% of lymphoplasmacytic lymphoma (Ngo 2011, Treon 2012)."],"analogy":"A guard (IκB) holds a prisoner (NF-kappa-B) who, once free, orders the cell to survive. The B-cell receptor sends a runner (BTK) to hand the guard to the shredder (proteasome). BTK inhibitors stop the runner, proteasome inhibitors jam the shredder, and lenalidomide removes the clerks (IKZF1/3) who file the survival orders.","nodes":[{"id":"bcr","label":"B-cell receptor / antigen","x":30,"y":5},{"id":"tlr","label":"TLR → MYD88 (L265P)","x":70,"y":5},{"id":"btk","label":"SYK → BTK → PLCγ2","x":30,"y":24},{"id":"cbm","label":"CARD11 / BCL10 / MALT1","x":50,"y":42},{"id":"ikk","label":"IKK complex","x":50,"y":58},{"id":"ikb","label":"IκB (destroyed by proteasome)","x":18,"y":74},{"id":"nfkb","label":"NF-κB (p65 / p50)","x":50,"y":74},{"id":"irf4","label":"IRF4 / IKZF1-3 (lenalidomide)","x":82,"y":74},{"id":"out","label":"BCL2, IL-6, IL-10, cyclin D → survival","x":50,"y":95,"targetId":"bcl2"}],"edges":[{"from":"bcr","to":"btk","type":"activates"},{"from":"btk","to":"cbm","type":"activates"},{"from":"tlr","to":"cbm","type":"activates"},{"from":"cbm","to":"ikk","type":"activates"},{"from":"ikk","to":"ikb","type":"inhibits"},{"from":"ikb","to":"nfkb","type":"inhibits"},{"from":"nfkb","to":"out","type":"activates"},{"from":"irf4","to":"nfkb","type":"activates"},{"from":"nfkb","to":"irf4","type":"activates"}],"interventions":["Covalent BTK inhibitors ibrutinib, acalabrutinib, zanubrutinib; non-covalent pirtobrutinib after BTK C481S resistance; BTK degraders in trials","Proteasome inhibitors bortezomib, carfilzomib and ixazomib block IκB degradation in multiple myeloma and mantle cell lymphoma","Lenalidomide and the CELMoDs degrade IKZF1/3, cutting IRF4 and NF-κB output","BCL2 inhibition (venetoclax) removes the main survival gene NF-κB switches on"]},{"id":"basal-cell-carcinoma-signalling","kind":"pathway","name":"Basal cell carcinoma (KEGG map)","aka":["KEGG hsa05217","Basal cell carcinoma"],"tldr":"KEGG's basal cell carcinoma map is the Hedgehog pathway: loss of the brake PTCH1 or activation of SMO leaves GLI transcription factors permanently on. Hedgehog inhibitors (vismodegib, sonidegib) shut this down in advanced disease.","summary":"The KEGG basal cell carcinoma map (hsa05217) shows almost every BCC as a Hedgehog pathway tumour. Normally the ligand Sonic hedgehog (SHH) binds the receptor PTCH1, which relieves PTCH1's inhibition of the seven-pass receptor SMO; SMO then frees the GLI1, GLI2 and GLI3 transcription factors from SUFU and the KIF7 complex, and GLI drives transcription of proliferation and survival genes (cyclin D, MYC, BCL2 and GLI1 itself). In BCC, inactivating mutations of PTCH1 (the gene behind Gorlin syndrome) or activating mutations of SMO, and less often SHH or SUFU changes, give continuous target gene activation without ligand. Ultraviolet exposure is the main environmental cause, and the map also draws TP53, which carries UV-signature mutations in more than half of sporadic BCCs. Epstein, Nat Rev Cancer, 2008 (doi:10.1038/nrc2503) reviews the genetics, the mouse models showing that Hedgehog activation in basal keratinocytes is sufficient to produce BCC, and the rationale for SMO antagonists. KEGG also lists canonical WNT and BMP components as context for basal keratinocyte fate.\n\nWhat drugs do about it: the SMO antagonists vismodegib and sonidegib are approved for locally advanced or metastatic BCC not suitable for surgery or radiotherapy, and produce responses in most Hedgehog-driven tumours. Resistance usually comes from SMO mutations that block drug binding or from downstream SUFU loss and GLI2 amplification. The PD-1 antibody cemiplimab is approved after Hedgehog inhibitor failure or intolerance.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05217","url":"https://www.kegg.jp/pathway/hsa05217"},{"label":"Review: Basal cell carcinomas, attack of the hedgehog","url":"https://doi.org/10.1038/nrc2503"}],"tags":[],"related":["hedgehog"],"cancers":["basal-cell-carcinoma"],"sections":[],"technologies":[],"targets":["smoothened","bcl2","tp53"],"drugs":["vismodegib","sonidegib","cemiplimab"],"companies":[],"institutions":[],"pathways":["hedgehog","p53-cell-cycle","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-epstein-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"PTCH1 is a handbrake on SMO, and SMO is a lever that lets GLI drive. In BCC the handbrake is cut (PTCH1 loss) or the lever is welded down (SMO mutation), so GLI drives constantly. Vismodegib and sonidegib jam the lever itself.","nodes":[{"id":"shh","label":"SHH ligand","x":50,"y":5},{"id":"ptch1","label":"PTCH1 (lost in BCC)","x":50,"y":22},{"id":"smo","label":"SMO (activated)","x":50,"y":40,"targetId":"smoothened"},{"id":"sufu","label":"SUFU / KIF7","x":20,"y":56},{"id":"gli","label":"GLI1 / GLI2","x":50,"y":62},{"id":"targets","label":"Cyclin D, MYC, BCL2, GLI1","x":50,"y":80,"targetId":"bcl2"},{"id":"p53","label":"TP53 (UV mutations)","x":85,"y":62,"targetId":"tp53"},{"id":"bcc","label":"Basal cell proliferation","x":60,"y":95}],"edges":[{"from":"shh","to":"ptch1","type":"inhibits"},{"from":"ptch1","to":"smo","type":"inhibits"},{"from":"smo","to":"sufu","type":"inhibits"},{"from":"sufu","to":"gli","type":"inhibits"},{"from":"gli","to":"targets","type":"activates"},{"from":"targets","to":"bcc","type":"activates"},{"from":"p53","to":"bcc","type":"inhibits"}],"interventions":["SMO antagonists vismodegib and sonidegib for locally advanced or metastatic basal cell carcinoma","PD-1 antibody cemiplimab after Hedgehog inhibitor failure or intolerance","Surgery (including Mohs micrographic surgery) or radiotherapy for localised tumours, which remain curative for most BCCs","Sun protection and skin surveillance, especially in Gorlin syndrome carriers of germline PTCH1 mutations"]},{"id":"base-excision-repair-parp","kind":"pathway","name":"Base excision repair, PARP & alkylation damage","aka":[],"tldr":"Tens of thousands of times a day a single DNA letter is oxidised or chemically scarred. A small crew snips it out and PARP marks the nick so it gets sealed. PARP inhibitors do not just switch PARP off; they trap it on the DNA, turning a harmless nick into a lethal break when the cell copies its DNA.","summary":"Damaged bases (8-oxoguanine, uracil, alkylated bases from temozolomide or endogenous methylation) are removed by glycosylases (OGG1, UNG, MPG), APE1 cuts the backbone, PARP1 binds the single-strand break and PARylates itself and histones to recruit XRCC1, Pol β fills the gap and LIG3 seals it (short-patch) or FEN1/LIG1 (long-patch). PARP inhibitors compete with NAD+ and, crucially, trap PARP1 on DNA (talazoparib >> olaparib > veliparib in trapping potency), so replication forks collide with the complex and collapse into double-strand breaks that only HR can fix, the mechanistic basis of BRCA synthetic lethality; PARP1-selective saruparib spares PARP2 and bone marrow. O6-methylguanine, the lethal TMZ lesion, is reversed directly by MGMT; MGMT promoter methylation predicts TMZ benefit in glioblastoma, and MMR is needed for TMZ toxicity. PARP inhibition also traps PARP at oxidative lesions from radiation and generates cytosolic DNA that fires cGAS-STING.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Base_excision_repair","links":[{"label":"Caldecott, Single-strand break repair and genetic disease (Nat Rev Genet 2008)","url":"https://doi.org/10.1038/nrg2380"},{"label":"Lord & Ashworth, PARP inhibitors: synthetic lethality in the clinic (Science 2017)","url":"https://doi.org/10.1126/science.aam7344"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["glioblastoma","ovarian","prostate","tnbc"],"sections":[],"technologies":["parp-inhibitor","parp-pet","radionuclide-parp-combination","cytotoxic-chemotherapy"],"targets":["parp","brca","atr"],"drugs":["olaparib","niraparib","rucaparib","talazoparib","temozolomide","lomustine"],"companies":[],"institutions":[],"pathways":["ddr","homologous-recombination-repair","cgas-sting","replication-stress"],"terms":["mgmt","synthetic-lethality","hrd","dna-alkylator-payloads"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-lord-science","paper-caldecott-nat-rev-genet"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Potholes on a busy road. Normally a small crew fills them overnight and PARP is the foreman who cones them off. A PARP inhibitor glues the foreman to the pothole; in the morning the traffic (replication) hits him and the road collapses, and only the bridge-building crew (BRCA) could rebuild it.","nodes":[{"id":"les","label":"Oxidised / alkylated base","x":12,"y":15},{"id":"mgmt","label":"MGMT direct reversal","x":12,"y":45},{"id":"glyc","label":"Glycosylase → APE1","x":40,"y":15},{"id":"ssb","label":"Single-strand break","x":68,"y":15},{"id":"parp","label":"PARP1 → XRCC1","x":68,"y":45,"targetId":"parp"},{"id":"polb","label":"Pol β, LIG3 seal","x":92,"y":60},{"id":"trap","label":"PARP trapped on DNA","x":40,"y":72},{"id":"fork","label":"Fork collapse → DSB","x":68,"y":90},{"id":"hr","label":"HR (BRCA) rescue","x":92,"y":90,"targetId":"brca"},{"id":"tmz","label":"Temozolomide, radiation","x":12,"y":78}],"edges":[{"from":"les","to":"glyc","type":"activates"},{"from":"mgmt","to":"les","type":"inhibits"},{"from":"glyc","to":"ssb","type":"activates"},{"from":"ssb","to":"parp","type":"activates"},{"from":"parp","to":"polb","type":"activates"},{"from":"tmz","to":"les","type":"activates"},{"from":"parp","to":"trap","type":"activates"},{"from":"trap","to":"fork","type":"activates"},{"from":"hr","to":"fork","type":"inhibits"}],"interventions":["PARP inhibitors: talazoparib (strongest trapper), olaparib, niraparib, rucaparib; PARP1-selective saruparib","Temozolomide in MGMT-methylated glioblastoma; lomustine, dacarbazine as alkylators","PARP inhibitor + radiotherapy or + TMZ combinations, limited by marrow toxicity; PARP-radioligand combinations in trials","PARP PET imaging to quantify target"]},{"id":"basement-membrane-tissue-barriers","kind":"pathway","name":"Basement membrane & tissue barriers","aka":[],"tldr":"Every 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.","summary":"Epithelia sit on a basement membrane of laminin, collagen IV, nidogen and perlecan anchored by integrins (α6β4 at hemidesmosomes) and dystroglycan, beneath which lies interstitial stroma. Polarity (apical-basal), E-cadherin junctions, and contact inhibition keep cells in place. Carcinoma in situ (DCIS, CIN3, Barrett's dysplasia) respects the sheet; invasion requires protease-driven breach (MMP2/9/14, uPA) at invadopodia, loss of polarity (PAR3, SCRIB), integrin switching and EMT-like changes, often assisted by CAFs and macrophages that widen the gap. Myoepithelial cells in breast add a second barrier. Staging encodes this (Tis versus T1) and screening programmes aim to find lesions before the breach.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Basement_membrane","links":[{"label":"Kelley et al., Traversing the basement membrane in vivo (J Cell Biol 2014)","url":"https://doi.org/10.1083/jcb.201311112"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["cervical","esophageal","breast-hr-positive","urothelial"],"sections":[],"technologies":["colposcopy-excision","endoscopic-resection","precancer-ablation","hpv-vaccine","histopathology-ihc"],"targets":[],"drugs":[],"companies":[],"institutions":[],"pathways":["emt","invasion-ecm-degradation","field-cancerisation"],"terms":["tnm-staging","cin-hsil","barretts-esophagus","stage-shift"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-kelley-j-cell-biol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"epi","label":"Polarised epithelium","x":15,"y":15},{"id":"junc","label":"E-cadherin junctions","x":45,"y":15},{"id":"bm","label":"Basement membrane","x":45,"y":45},{"id":"int","label":"Integrins / hemidesmosomes","x":15,"y":45},{"id":"cis","label":"Carcinoma in situ","x":78,"y":15},{"id":"mmp","label":"MMPs, uPA, invadopodia","x":78,"y":45},{"id":"stroma","label":"Stroma, vessels","x":78,"y":80},{"id":"inv","label":"Invasive carcinoma","x":45,"y":80},{"id":"myo","label":"Myoepithelial layer","x":15,"y":80}],"edges":[{"from":"epi","to":"junc","type":"activates"},{"from":"int","to":"bm","type":"activates"},{"from":"junc","to":"cis","type":"inhibits"},{"from":"bm","to":"inv","type":"inhibits"},{"from":"myo","to":"inv","type":"inhibits"},{"from":"cis","to":"mmp","type":"activates"},{"from":"mmp","to":"bm","type":"inhibits"},{"from":"mmp","to":"stroma","type":"activates"},{"from":"stroma","to":"inv","type":"activates"},{"from":"cis","to":"inv","type":"activates"}],"interventions":["Screening and excision of in situ disease (colposcopy, DCIS surgery, endoscopic resection) before the breach","HPV vaccination removes the commonest driver of cervical in situ lesions","MMP inhibitors failed clinically in the 1990s; invasion is now approached via FAK, integrin and stromal targets","Staging (Tis vs T1) and margins encode whether the barrier was crossed"]},{"id":"bcr-abl1-signalling","kind":"pathway","name":"BCR::ABL1 (Philadelphia chromosome)","aka":[],"tldr":"Chronic myeloid leukaemia is caused by one broken gene: two chromosomes swap pieces and glue a kinase (ABL1) to a protein that forces it permanently on. Imatinib, the first drug to target it, turned a fatal disease into a manageable one, and later drugs cover the mutations that escape it.","summary":"The t(9;22) translocation fuses BCR to ABL1, producing a constitutively active cytoplasmic tyrosine kinase (p210 in CML, p190 in most Ph-positive ALL). BCR::ABL1 autophosphorylates and recruits GRB2/GAB2 to activate RAS-MAPK and PI3K-AKT, phosphorylates STAT5 for survival and CRKL for adhesion changes, and raises reactive oxygen species that drive further mutations (blast crisis). Imatinib (2001) binds the inactive kinase conformation; dasatinib, nilotinib and bosutinib are more potent second-generation ATP-site inhibitors; ponatinib covers the T315I gatekeeper mutation; asciminib (2021) binds the myristoyl pocket (STAMP) and works with ATP-site drugs against compound mutations. Response is tracked by BCR::ABL1 transcript levels (major molecular response, MR4.5), and about half of patients with sustained deep response can stop treatment in treatment-free remission. Ph-positive ALL is treated with TKIs plus chemotherapy or blinatumomab.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Philadelphia_chromosome","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Philadelphia_chromosome"}],"tags":[],"related":[],"cancers":["cml","all-leukemia"],"sections":[],"technologies":[],"targets":["kras","pik3ca","abl1"],"drugs":["imatinib","dasatinib","nilotinib","bosutinib","ponatinib","asciminib"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","jak-stat"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"A car accelerator pedal welded to the floor (BCR::ABL1). Imatinib wedges a block under the pedal so it cannot be pressed; some engines change the pedal's shape (T315I) so the block no longer fits, and ponatinib or asciminib are blocks cut for the new shape.","nodes":[{"id":"t922","label":"t(9;22) translocation","x":50,"y":5},{"id":"bcrabl","label":"BCR::ABL1 kinase (p210 / p190)","x":50,"y":25},{"id":"grb2","label":"GRB2 / GAB2","x":25,"y":45},{"id":"stat5","label":"STAT5","x":75,"y":45},{"id":"ras","label":"RAS → MAPK","x":15,"y":65,"targetId":"kras"},{"id":"pi3k","label":"PI3K → AKT","x":40,"y":65,"targetId":"pik3ca"},{"id":"ros","label":"Reactive oxygen species → new mutations","x":80,"y":65},{"id":"out","label":"Proliferation, survival, genomic instability → CML, Ph+ ALL","x":50,"y":92}],"edges":[{"from":"t922","to":"bcrabl","type":"activates"},{"from":"bcrabl","to":"grb2","type":"activates"},{"from":"bcrabl","to":"stat5","type":"activates"},{"from":"grb2","to":"ras","type":"activates"},{"from":"grb2","to":"pi3k","type":"activates"},{"from":"bcrabl","to":"ros","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"stat5","to":"out","type":"activates"},{"from":"ros","to":"out","type":"activates"}],"interventions":["Imatinib, the first-generation ATP-site inhibitor; dasatinib, nilotinib, bosutinib as more potent second-generation options","Ponatinib for the T315I gatekeeper mutation","Asciminib, an allosteric STAMP inhibitor, alone or with an ATP-site inhibitor against compound mutations","Molecular monitoring (BCR::ABL1 transcripts) to guide treatment-free remission attempts","In Ph-positive ALL: TKI with chemotherapy or with blinatumomab, and transplant for high-risk disease"]},{"id":"bladder-cancer-signalling","kind":"pathway","name":"Bladder cancer (KEGG map)","aka":["KEGG hsa05219","Bladder cancer","Urothelial carcinoma (KEGG)"],"tldr":"KEGG's bladder cancer map shows two routes: low-grade papillary tumours driven by FGFR3 or HRAS activating the MAPK relay, and high-grade invasive tumours that lose TP53 and RB1. Erdafitinib targets the first route; antibody-drug conjugates and PD-1 antibodies now anchor treatment of the second.","summary":"The KEGG bladder cancer map (hsa05219) draws the divergent pathways of urothelial carcinoma. Low-grade, non-invasive papillary tumours arise from urothelial hyperplasia and show constitutive receptor tyrosine kinase to RAS signalling: activating mutations in FGFR3 or HRAS, or overexpression of EGFR and ERBB2, feed RAF to MEK to ERK, MYC and cyclin D1 to drive proliferation. Flat carcinoma in situ and invasive tumours instead show loss of the TP53 and RB pathways: TP53 mutation or MDM2 amplification, RB1 loss or mutation, and deletion of the CDKN2A locus (p16 and p14ARF), which together remove the G1/S checkpoint. Invasion and metastasis are promoted by loss of E-cadherin (CDH1), matrix metalloproteinases (MMP1, MMP2, MMP9), and angiogenic factors such as VEGFA and IL-8 (CXCL8), with thrombospondin-1 as the anti-angiogenic counterweight. Knowles and Hurst, Nat Rev Cancer, 2015 (doi:10.1038/nrc3817) review this two-pathway model in the light of genome sequencing, adding the very high rate of chromatin regulator mutations (KDM6A, ARID1A, KMT2D), TERT promoter mutations, and the luminal and basal expression subtypes of muscle-invasive disease.\n\nWhat drugs do about it: the pan-FGFR inhibitor erdafitinib is approved for advanced urothelial carcinoma with FGFR3 alterations. For the high-grade, p53-pathway-deficient majority, the Nectin-4 antibody-drug conjugate enfortumab vedotin with pembrolizumab has replaced platinum chemotherapy as first-line treatment, PD-1 and PD-L1 antibodies (pembrolizumab, nivolumab, atezolizumab, avelumab, durvalumab) are used as maintenance and in the perioperative setting, and the HER2 antibody-drug conjugate disitamab vedotin targets ERBB2-overexpressing tumours.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05219","url":"https://www.kegg.jp/pathway/hsa05219"},{"label":"Review: Molecular biology of bladder cancer","url":"https://doi.org/10.1038/nrc3817"}],"tags":[],"related":["fgfr-signalling","p53-cell-cycle"],"cancers":["urothelial"],"sections":[],"technologies":[],"targets":["her2","mek","cdk4-6","mdm2","vegf"],"drugs":["erdafitinib","enfortumab-vedotin","pembrolizumab","nivolumab","atezolizumab","avelumab","durvalumab","disitamab-vedotin"],"companies":[],"institutions":[],"pathways":["fgfr-signalling","ras-mapk","p53-cell-cycle","cell-cycle-engine-cdks","emt","vegf-angiogenesis","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-knowles-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Two roads to the same city. The slow road (papillary tumours) is a stuck accelerator, FGFR3 or HRAS. The fast road (invasive tumours) is brake failure, TP53 and RB1 lost. Erdafitinib fixes the accelerator on the slow road; antibody-drug conjugates and immunotherapy deal with cars that have already reached the city.","nodes":[{"id":"fgfr3","label":"FGFR3 (mutated)","x":15,"y":6},{"id":"egfr","label":"EGFR / ERBB2","x":45,"y":6,"targetId":"her2"},{"id":"hras","label":"HRAS","x":30,"y":26},{"id":"mek","label":"RAF / MEK / ERK","x":30,"y":46,"targetId":"mek"},{"id":"cyc","label":"MYC, cyclin D1 / CDK4","x":30,"y":66,"targetId":"cdk4-6"},{"id":"cdkn2a","label":"CDKN2A (p16, p14ARF) deleted","x":82,"y":6},{"id":"mdm2","label":"MDM2 to TP53","x":82,"y":30,"targetId":"mdm2"},{"id":"rb1","label":"RB1 (lost)","x":82,"y":54},{"id":"invade","label":"E-cadherin loss, MMPs, VEGF","x":82,"y":76,"targetId":"vegf"},{"id":"out","label":"Papillary vs invasive carcinoma","x":50,"y":93}],"edges":[{"from":"fgfr3","to":"hras","type":"activates"},{"from":"egfr","to":"hras","type":"activates"},{"from":"hras","to":"mek","type":"activates"},{"from":"mek","to":"cyc","type":"activates"},{"from":"cyc","to":"out","type":"activates"},{"from":"cdkn2a","to":"mdm2","type":"inhibits"},{"from":"cdkn2a","to":"cyc","type":"inhibits"},{"from":"cyc","to":"rb1","type":"inhibits"},{"from":"rb1","to":"out","type":"inhibits"},{"from":"mdm2","to":"out","type":"activates"},{"from":"invade","to":"out","type":"activates"}],"interventions":["Enfortumab vedotin (Nectin-4 antibody-drug conjugate) plus pembrolizumab as first-line treatment for advanced urothelial carcinoma","FGFR inhibitor erdafitinib for FGFR3-altered advanced urothelial carcinoma","PD-1 and PD-L1 antibodies: pembrolizumab, nivolumab, atezolizumab, avelumab (maintenance), durvalumab (perioperative)","HER2 antibody-drug conjugate disitamab vedotin for ERBB2-overexpressing tumours","Intravesical BCG for high-risk non-muscle-invasive disease; cisplatin-based chemotherapy and cystectomy for muscle-invasive disease"]},{"id":"breast-cancer-signalling","kind":"pathway","name":"Breast cancer (KEGG map)","aka":["KEGG hsa05224","Breast cancer"],"tldr":"KEGG's breast cancer map lays out the three clinical subtypes as signalling routes: oestrogen receptor driving cyclin D and CDK4/6 in hormone-receptor-positive disease, HER2 driving PI3K/AKT and MAPK in HER2-positive disease, and EGFR, Notch, Wnt and BRCA defects in triple-negative disease. Each route has its own drug class.","summary":"The KEGG breast cancer map (hsa05224) is organised by molecular subtype. Hormone-receptor-positive tumours (luminal A and B) are driven by oestrogen binding ESR1, which with co-activators (NCOA1, NCOA3) switches on cyclin D1, MYC and the progesterone receptor; cyclin D1 partners CDK4/6 to phosphorylate RB and release E2F, and CCND1 amplification is a recurrent event. HER2-positive tumours overexpress ERBB2, which dimerises with EGFR or HER3 and activates both PI3K to AKT to mTOR and RAS to RAF to MEK to ERK for growth, survival and differentiation, with PIK3CA mutation or amplification and PTEN loss adding to the PI3K arm. Triple-negative and basal-like tumours show EGFR overexpression, deregulated Notch (NOTCH1 to 4, JAG and DLL ligands driving HES and HEY genes) and Wnt/beta-catenin signalling (FZD7 and LRP6 overexpression), and about 8% of all breast cancers are hereditary through germline BRCA1 or BRCA2 mutations that disable homologous recombination repair. TP53, PIK3CA and GATA3 are the only genes mutated in more than 10% of all breast cancers. The map also includes FGF/FGFR1 amplification, IGF1R, KIT and RANKL (TNFSF11). Harbeck et al., Nat Rev Dis Primers, 2019 (doi:10.1038/s41572-019-0111-2) review how these subtypes translate into treatment, with endocrine therapy plus CDK4/6 inhibition, HER2-directed antibodies and small molecules, and chemotherapy plus immunotherapy or PARP inhibition for triple-negative disease.\n\nWhat drugs do about it: for ER-positive disease, aromatase inhibitors (letrozole), tamoxifen, fulvestrant and oral SERDs such as elacestrant, imlunestrant and camizestrant remove or degrade the receptor while palbociclib, ribociclib and abemaciclib block CDK4/6; alpelisib, inavolisib and capivasertib target PIK3CA, PI3K and AKT alterations. For HER2-positive disease, trastuzumab and pertuzumab, the antibody-drug conjugates trastuzumab emtansine and trastuzumab deruxtecan, and the kinase inhibitors tucatinib, neratinib and lapatinib shut down HER2. For triple-negative disease, pembrolizumab with chemotherapy, the TROP2 antibody-drug conjugates sacituzumab govitecan and datopotamab deruxtecan, and the PARP inhibitors olaparib and talazoparib for germline BRCA carriers are the mainstays.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05224","url":"https://www.kegg.jp/pathway/hsa05224"},{"label":"Review: Breast cancer (Nat Rev Dis Primers 2019)","url":"https://doi.org/10.1038/s41572-019-0111-2"}],"tags":[],"related":["er-signaling","pi3k-akt-mtor","cell-cycle-engine-cdks"],"cancers":["breast-hr-positive","breast-her2-positive","tnbc"],"sections":[],"technologies":[],"targets":["estrogen-receptor","cdk4-6","her2","pik3ca","mek","egfr","brca","tp53"],"drugs":["letrozole","tamoxifen","fulvestrant","elacestrant","imlunestrant","camizestrant","palbociclib","ribociclib","abemaciclib","alpelisib","inavolisib","capivasertib","trastuzumab","pertuzumab","trastuzumab-emtansine","trastuzumab-deruxtecan","tucatinib","neratinib","lapatinib","pembrolizumab","sacituzumab-govitecan","datopotamab-deruxtecan","olaparib","talazoparib"],"companies":[],"institutions":[],"pathways":["er-signaling","cell-cycle-engine-cdks","pi3k-akt-mtor","ras-mapk","rtk-activation","ddr","wnt","p53-cell-cycle","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-harbeck-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Three engines that can power a breast cancer. Hormone-receptor-positive tumours run on oestrogen, so endocrine drugs cut the fuel and CDK4/6 inhibitors put a chock under the wheel. HER2-positive tumours have an oversized growth-signal antenna, so anti-HER2 antibodies cap it. Triple-negative tumours have neither, so treatment goes after their broken DNA repair (PARP inhibitors), their surface markers (TROP2 conjugates) and their visibility to the immune system.","nodes":[{"id":"er","label":"Oestrogen / ESR1 (HR-positive)","x":15,"y":6,"targetId":"estrogen-receptor"},{"id":"ccnd1","label":"Cyclin D1 / CDK4/6","x":15,"y":32,"targetId":"cdk4-6"},{"id":"rb","label":"RB to E2F","x":15,"y":56},{"id":"her2","label":"HER2 (ERBB2) amplified","x":50,"y":6,"targetId":"her2"},{"id":"pi3k","label":"PI3K (PIK3CA) / AKT / mTOR","x":40,"y":32,"targetId":"pik3ca"},{"id":"mapk","label":"RAS / RAF / MEK / ERK","x":62,"y":32,"targetId":"mek"},{"id":"egfr","label":"EGFR (TNBC)","x":85,"y":6,"targetId":"egfr"},{"id":"notch","label":"Notch, Wnt/beta-catenin (TNBC)","x":88,"y":32},{"id":"brca","label":"BRCA1/2 loss (HR repair)","x":88,"y":56,"targetId":"brca"},{"id":"p53","label":"TP53 mutation","x":55,"y":62,"targetId":"tp53"},{"id":"out","label":"Proliferation, survival","x":50,"y":90}],"edges":[{"from":"er","to":"ccnd1","type":"activates"},{"from":"ccnd1","to":"rb","type":"inhibits"},{"from":"rb","to":"out","type":"inhibits"},{"from":"her2","to":"pi3k","type":"activates"},{"from":"her2","to":"mapk","type":"activates"},{"from":"egfr","to":"mapk","type":"activates"},{"from":"egfr","to":"pi3k","type":"activates"},{"from":"mapk","to":"ccnd1","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"notch","to":"out","type":"activates"},{"from":"brca","to":"out","type":"activates"},{"from":"p53","to":"out","type":"inhibits"}],"interventions":["HR-positive disease: endocrine therapy (letrozole, tamoxifen, fulvestrant, oral SERDs elacestrant, imlunestrant, camizestrant) with a CDK4/6 inhibitor (palbociclib, ribociclib, abemaciclib), then alpelisib, inavolisib or capivasertib for PIK3CA, AKT1 or PTEN alterations","HER2-positive disease: trastuzumab plus pertuzumab, trastuzumab emtansine, trastuzumab deruxtecan, and the HER2 kinase inhibitors tucatinib, neratinib and lapatinib","Triple-negative disease: pembrolizumab with chemotherapy, TROP2 antibody-drug conjugates sacituzumab govitecan and datopotamab deruxtecan","PARP inhibitors olaparib and talazoparib for germline BRCA1/2-mutated breast cancer","Trastuzumab deruxtecan for HER2-low tumours across subtypes"]},{"id":"cachexia-biology","kind":"pathway","name":"Cancer cachexia","aka":[],"tldr":"The 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.","summary":"Tumour- and host-derived GDF-15 acts on the brainstem GFRAL receptor to suppress appetite; IL-6/STAT3, TNF, activin/myostatin, and glucocorticoids drive muscle proteolysis (ubiquitin-proteasome, autophagy) and adipose lipolysis/browning. Cachexia worsens treatment tolerance and is largely irreversible late. Ponsegromab (anti-GDF-15, Pfizer) increased weight and activity in a phase 2 trial (NEJM 2024) and is in phase 3; anamorelin (ghrelin agonist) is approved in Japan; nutrition and exercise remain foundational. Cachexia is now treated as a target in its own right rather than an inevitability.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Cachexia","links":[{"label":"Groarke et al., Ponsegromab for the treatment of cancer cachexia (NEJM 2024)","url":"https://doi.org/10.1056/NEJMoa2409515"}],"tags":["mechanism"],"related":[],"cancers":["pancreatic","nsclc","gastric"],"sections":[],"technologies":["exercise-oncology","geriatric-assessment"],"targets":[],"drugs":[],"companies":[],"institutions":["fred-hutch","cold-spring-harbor","md-anderson"],"pathways":["jak-stat","inflammation-nfkb","cancer-metabolism"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":["Leading programmes: Janowitz (CSHL) on cachexia mechanisms; Fearon legacy (Edinburgh) on definitions; Roeland (Oregon) and Baracos (Alberta) on clinical cachexia; Pfizer ponsegromab programme; Cancer Grand Challenges CANCAN team."],"analogy":"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.","nodes":[{"id":"tum","label":"Tumour + inflammation","x":15,"y":50},{"id":"gdf","label":"GDF-15 → GFRAL (brainstem)","x":45,"y":25},{"id":"il6","label":"IL-6 / TNF / activin","x":45,"y":75},{"id":"app","label":"Anorexia","x":75,"y":25},{"id":"muscle","label":"Muscle proteolysis, fat lipolysis","x":75,"y":75},{"id":"out","label":"Weight loss, frailty, death","x":92,"y":50}],"edges":[{"from":"tum","to":"gdf","type":"activates"},{"from":"tum","to":"il6","type":"activates"},{"from":"gdf","to":"app","type":"activates"},{"from":"il6","to":"muscle","type":"activates"},{"from":"app","to":"out","type":"activates"},{"from":"muscle","to":"out","type":"activates"}],"interventions":["Ponsegromab (anti-GDF-15) phase 3","Anamorelin (approved Japan), olanzapine for appetite, corticosteroids short term","Exercise and nutrition support (ESPEN/ASCO guidelines)","Anti-IL-6 and activin/myostatin agents in trials"]},{"id":"cancer-metabolism","kind":"pathway","name":"Cancer metabolism","aka":[],"tldr":"Cancer 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.","summary":"Oncogenic signalling (PI3K/AKT/mTOR, MYC, HIF) drives aerobic glycolysis and lactate export, glutaminolysis for TCA anaplerosis, de novo lipogenesis (FASN, SCD), and one-carbon metabolism (SHMT, MTHFD2) for nucleotides and methylation. Mutant IDH1/2 produce the oncometabolite 2-HG. Metabolic plasticity lets tumours switch fuels, which is why single-target metabolic drugs (glutaminase inhibitor telaglenastat, negative in RCC) disappoint. Approved metabolic drugs are antimetabolite chemotherapies and IDH inhibitors; arginine deprivation (ADI-PEG20) is in phase 3 in mesothelioma. Diet interventions (fasting-mimicking, ketogenic) are in trials as adjuncts.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Warburg_effect_(oncology)","links":[{"label":"Pavlova, Zhu & Thompson, The hallmarks of cancer metabolism (Cell Metabolism 2022)","url":"https://doi.org/10.1016/j.cmet.2022.01.007"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":["fdg-pet","cytotoxic-chemotherapy"],"targets":["idh","pik3ca","hif2a"],"drugs":["ivosidenib","vorasidenib"],"companies":[],"institutions":["mskcc","md-anderson","penn-abramson","francis-crick"],"pathways":["pi3k-akt-mtor","myc","hif-vhl"],"terms":["warburg-effect","deregulating-cellular-energetics","suv"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-pavlova-cell-metab-2022"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Thompson (MSK) and Vander Heiden (MIT/Koch) on metabolic dependencies; DeBerardinis (UTSW) on in vivo metabolism; Rabinowitz (Princeton/Ludwig) on nutrient flux."],"analogy":"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.","nodes":[{"id":"glc","label":"Glucose (GLUT1)","x":10,"y":25},{"id":"glyc","label":"Aerobic glycolysis (Warburg)","x":30,"y":25},{"id":"lac","label":"Lactate export (MCT4)","x":50,"y":10},{"id":"tca","label":"TCA cycle","x":50,"y":50},{"id":"gln","label":"Glutamine → glutaminase","x":10,"y":65},{"id":"lipid","label":"De novo lipogenesis (FASN)","x":75,"y":30},{"id":"onec","label":"One-carbon (SHMT2, MTHFD2) → nucleotides","x":75,"y":70},{"id":"idh","label":"Mutant IDH → 2-HG","x":30,"y":85,"targetId":"idh"},{"id":"sig","label":"PI3K/AKT/mTOR, MYC, HIF","x":92,"y":50}],"edges":[{"from":"glc","to":"glyc","type":"activates"},{"from":"glyc","to":"lac","type":"activates"},{"from":"glyc","to":"tca","type":"activates"},{"from":"gln","to":"tca","type":"activates"},{"from":"tca","to":"lipid","type":"activates"},{"from":"glyc","to":"onec","type":"activates"},{"from":"tca","to":"idh","type":"activates"},{"from":"sig","to":"glyc","type":"activates"},{"from":"sig","to":"lipid","type":"activates"},{"from":"sig","to":"onec","type":"activates"}],"interventions":["Antimetabolite chemotherapy (5-FU, gemcitabine, methotrexate) exploits nucleotide demand","IDH inhibitors (ivosidenib, vorasidenib) block 2-HG","Glutaminase, MCT1, FASN, and arginine-deprivation agents in trials","FDG PET images the Warburg effect","Diet and metformin trials as adjuncts"]},{"id":"cancer-neuroscience","kind":"pathway","name":"Cancer neuroscience (nerve-tumour signalling)","aka":[],"tldr":"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.","summary":"Perineural invasion predicts poor outcome in pancreatic, prostate, and head and neck cancer; sympathetic (β-adrenergic) and parasympathetic (muscarinic) signalling promotes growth and metastasis; tumours secrete NGF and axon-guidance factors to recruit nerves. Gliomas form functional synapses with neurons (AMPA receptors) and hijack activity-regulated neuroligin-3; neuronal activity drives glioma growth. Interventions: β-blockers (propranolol trials in melanoma, breast, angiosarcoma), botulinum toxin denervation (gastric cancer trial), NGF/TrkA blockade, AMPA antagonists (perampanel) in glioma, and gabapentinoids. A young field with strong preclinical evidence and early clinical signals.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Neuro-oncology","links":[{"label":"Venkatesh et al., Electrical and synaptic integration of glioma into neural circuits (Nature 2019)","url":"https://doi.org/10.1038/s41586-019-1563-y"},{"label":"Monje et al., Roadmap for the emerging field of cancer neuroscience (Cell 2020)","url":"https://doi.org/10.1016/j.cell.2020.03.034"}],"tags":["mechanism"],"related":[],"cancers":["pancreatic","prostate","glioblastoma","head-and-neck","melanoma"],"sections":[],"technologies":[],"targets":["ntrk"],"drugs":[],"companies":[],"institutions":["stanford","mskcc","md-anderson"],"pathways":["tumor-microenvironment","metastatic-cascade"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-venkatesh-nature","paper-monje-cell"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Monje (Stanford) on neuron-glioma synapses; Winkler (Heidelberg/DKFZ) on tumour microtubes; Wong (MSK) and Amit (MD Anderson) on nerves in pancreatic and head and neck cancer; Cancer Grand Challenges NexTGen/ neuroscience teams."],"analogy":"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.","nodes":[{"id":"tum","label":"Tumour","x":50,"y":50},{"id":"ngf","label":"NGF, axon guidance cues","x":25,"y":25},{"id":"nerve","label":"Nerve ingrowth (PNI)","x":25,"y":75},{"id":"adr","label":"β-adrenergic / cholinergic signals","x":75,"y":75},{"id":"syn","label":"Neuron-glioma synapses (AMPA)","x":75,"y":25},{"id":"growth","label":"Growth, invasion, immunosuppression","x":92,"y":50}],"edges":[{"from":"tum","to":"ngf","type":"activates"},{"from":"ngf","to":"nerve","type":"activates"},{"from":"nerve","to":"adr","type":"activates"},{"from":"adr","to":"growth","type":"activates"},{"from":"syn","to":"growth","type":"activates"},{"from":"tum","to":"syn","type":"activates"}],"interventions":["Propranolol and other β-blockers in trials (melanoma, breast, angiosarcoma)","Botulinum toxin denervation trials (gastric)","NGF/TrkA and AMPA-receptor (perampanel) blockade; gabapentin in glioma trials","Perineural invasion as a staging biomarker"]},{"id":"cancer-stem-cells-plasticity","kind":"pathway","name":"Cancer stem cells & phenotypic plasticity","aka":[],"tldr":"Some 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.","summary":"Cancer stem cells (CSCs; first shown in AML by Dick, 1994-97) are functionally defined by tumour-initiating capacity; in solid tumours stemness is usually a reversible state rather than a fixed population. Lineage plasticity under therapy produces neuroendocrine transformation (EGFR-mutant NSCLC to SCLC; prostate adenocarcinoma to NEPC), basal/mesenchymal switching in breast cancer, and dedifferentiation. Drivers: EMT programmes, Wnt/Notch/Hedgehog, epigenetic remodelling (EZH2, SWI/SNF), TP53/RB1 loss. 'Unlocking phenotypic plasticity' is a 2022 hallmark. Therapeutic routes: differentiation therapy (ATRA in APL), EZH2/LSD1 inhibitors, targeting CSC markers (CD44, LGR5), and MRD-directed therapy.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Cancer_stem_cell","links":[{"label":"Rubin, Bristow, Thienger et al., Impact of lineage plasticity to and from a neuroendocrine phenotype (Molecular Cell 2020)","url":"https://doi.org/10.1016/j.molcel.2020.10.033"},{"label":"Hanahan, Hallmarks of Cancer: New Dimensions (Cancer Discovery 2022)","url":"https://doi.org/10.1158/2159-8290.CD-21-1059"}],"tags":["mechanism"],"related":["theories-of-cancer","cancer-stem-cell-theory"],"cancers":[],"sections":[],"technologies":[],"targets":["ezh2","dll3","menin","tp53"],"drugs":["tarlatamab","revumenib"],"companies":[],"institutions":["princess-margaret","mskcc","stanford","francis-crick"],"pathways":["emt","wnt","notch","hedgehog","swi-snf-chromatin","clonal-evolution"],"terms":["unlocking-phenotypic-plasticity","histologic-transformation","lgr5"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-singh-brain-tumour-initiating-cells-nature-2004","paper-rubin-mol-cell"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Dick (Princess Margaret, leukaemia stem cells); Sawyers and Rudin (MSK, lineage plasticity in prostate and lung); Clevers (Hubrecht, organoids and LGR5); Weissman (Stanford)."],"analogy":"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.","nodes":[{"id":"csc","label":"Stem-like state","x":50,"y":30},{"id":"diff","label":"Differentiated bulk","x":20,"y":60},{"id":"mes","label":"Mesenchymal / drug-tolerant persister","x":80,"y":60},{"id":"ne","label":"Lineage switch (NE transformation)","x":50,"y":88,"targetId":"dll3"},{"id":"epi","label":"EZH2, SWI/SNF, TP53/RB1 loss","x":88,"y":20,"targetId":"ezh2"},{"id":"niche","label":"Wnt / Notch / Hedgehog niche","x":12,"y":20}],"edges":[{"from":"csc","to":"diff","type":"activates"},{"from":"diff","to":"csc","type":"activates"},{"from":"csc","to":"mes","type":"activates"},{"from":"mes","to":"csc","type":"activates"},{"from":"epi","to":"ne","type":"activates"},{"from":"csc","to":"ne","type":"activates"},{"from":"niche","to":"csc","type":"activates"},{"from":"epi","to":"csc","type":"activates"}],"interventions":["Differentiation therapy: ATRA/arsenic in APL (curative), menin inhibitors differentiate KMT2A/NPM1 leukaemias","EZH2, LSD1, and BET inhibitors to block plasticity (trials)","DLL3-directed tarlatamab for neuroendocrine-transformed tumours","MRD-directed therapy to catch persisters before regrowth"]},{"id":"cd47-sirpa","kind":"pathway","name":"CD47 / SIRPα (the 'don't eat me' signal)","aka":[],"tldr":"Macrophages eat cells that look wrong unless the cell shows CD47, a 'don't eat me' badge that binds SIRP-alpha. Leukaemias, myelodysplastic syndromes, lymphomas and carcinomas overproduce CD47 to escape being eaten; antibodies that cover the badge work in the lab, but the leading drug, magrolimab, failed in blood cancers because red cells wear the same badge and lose it too.","summary":"CD47 on a cell binds SIRPα on macrophages and dendritic cells; SIRPα's ITIM motifs recruit SHP-1 and SHP-2, which stop the myosin-driven engulfment programme. Healthy cells, including red cells and platelets, rely on it; tumours (AML, MDS, lymphoma, many carcinomas) express high CD47 to defeat the 'eat me' signals (calreticulin, phosphatidylserine, antibody Fc bound to the cell) that would otherwise trigger phagocytosis. Blocking CD47 or SIRPα therefore synergises with opsonising antibodies (rituximab) and with agents that expose calreticulin (azacitidine). Magrolimab, the first anti-CD47 antibody, showed early promise with azacitidine in MDS and AML but the phase 3 ENHANCE trials were stopped in 2023-2024 for futility and excess deaths, and the programme was discontinued; on-target anaemia (red cells express CD47) limits every agent in the class. Newer designs try to reduce red-cell binding (low-affinity or Fc-silent antibodies, SIRPα-Fc fusion proteins such as evorpacept, CD47 × tumour-antigen bispecifics) and the checkpoint is being explored in solid tumours with anti-EGFR and anti-HER2 antibodies.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/CD47","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/CD47"}],"tags":[],"related":[],"cancers":["aml","mds","dlbcl","non-hodgkin-lymphoma"],"sections":[],"technologies":[],"targets":["cd47"],"drugs":["magrolimab"],"companies":[],"institutions":[],"pathways":["antigen-presentation-immunoediting","tumor-microenvironment"],"terms":[],"trials":["magrolimab"],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"Macrophages are security guards who remove anyone without a staff badge. Cancer cells forge the badge (CD47) so the guards wave them through. Anti-CD47 drugs cover the forged badges, but every red blood cell carries a genuine badge too, so covering them all makes the guards remove red cells as well: anaemia.","nodes":[{"id":"cd47","label":"CD47 on tumour (and red) cells","x":50,"y":5,"targetId":"cd47"},{"id":"sirpa","label":"SIRPα on macrophage","x":50,"y":26},{"id":"shp","label":"SHP-1 / SHP-2","x":50,"y":46},{"id":"eat","label":"Pro-phagocytic signals: calreticulin, antibody Fc (rituximab)","x":16,"y":46},{"id":"phago","label":"Phagocytosis of the tumour cell","x":50,"y":68},{"id":"present","label":"Antigen presentation → T-cell priming","x":50,"y":92},{"id":"rbc","label":"Red cells (on-target anaemia)","x":84,"y":26}],"edges":[{"from":"cd47","to":"sirpa","type":"activates"},{"from":"sirpa","to":"shp","type":"activates"},{"from":"shp","to":"phago","type":"inhibits"},{"from":"eat","to":"phago","type":"activates"},{"from":"phago","to":"present","type":"activates"},{"from":"rbc","to":"sirpa","type":"activates"}],"interventions":["Anti-CD47 magrolimab with azacitidine: phase 3 ENHANCE, ENHANCE-2 and ENHANCE-3 stopped for futility or harm; programme discontinued (2024)","SIRPα-Fc fusions (evorpacept) and Fc-silent or low-affinity anti-CD47 antibodies designed to spare red cells","CD47 × CD19 or CD20 bispecifics to confine blockade to B-cell tumours","Combination with opsonising antibodies (rituximab, cetuximab, trastuzumab) to supply the 'eat me' signal"]},{"id":"senescence","kind":"pathway","name":"Cellular senescence","aka":[],"tldr":"Damaged 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.","summary":"Oncogene-induced and therapy-induced senescence engage p53/p21 and p16/RB to arrest the cycle; the senescence-associated secretory phenotype (SASP: IL-6, IL-8, MMPs) remodels the microenvironment, promotes inflammation, immune recruitment, and paradoxically tumour progression and treatment resistance. 'One-two punch' therapy induces senescence (e.g., CDK4/6 inhibitors, chemotherapy) then clears it with senolytics (BCL-XL/BCL-2 inhibitors, navitoclax; dasatinib+quercetin; uPAR CAR-T in mice). Senescent cells are a 2022 hallmark. Clinical proof in oncology is early; toxicity of BCL-XL inhibition (platelets) drives PROTAC approaches.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Cellular_senescence","links":[{"label":"Schmitt, Wang & Demaria, Senescence and cancer: role and therapeutic opportunities (Nature Reviews Clinical Oncology 2022)","url":"https://doi.org/10.1038/s41571-022-00668-4"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":["cdk46-inhibitor"],"targets":["tp53","cdk4-6","bcl2"],"drugs":[],"companies":[],"institutions":["mskcc","cold-spring-harbor","mayo-clinic"],"pathways":["p53-cell-cycle","apoptosis-bcl2","inflammation-nfkb","tumor-microenvironment"],"terms":["senescent-cells"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-schmitt-nat-rev-clin-oncol"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Lowe (MSK) on senescence and immune clearance; Campisi legacy (Buck); Kirkland (Mayo) on senolytics; Demaria (ERIBA Groningen); Peeper (NKI) on one-two punch."],"analogy":"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.","nodes":[{"id":"dmg","label":"Oncogene / therapy stress","x":15,"y":30},{"id":"p53","label":"p53 → p21","x":40,"y":20,"targetId":"tp53"},{"id":"p16","label":"p16 → RB","x":40,"y":45,"targetId":"cdk4-6"},{"id":"arrest","label":"Stable arrest","x":65,"y":30},{"id":"sasp","label":"SASP (IL-6, IL-8, MMPs)","x":65,"y":70},{"id":"tme","label":"Inflammation, relapse, resistance","x":90,"y":70},{"id":"seno","label":"Senolytics (BCL-XL, uPAR CAR-T)","x":90,"y":30}],"edges":[{"from":"dmg","to":"p53","type":"activates"},{"from":"dmg","to":"p16","type":"activates"},{"from":"p53","to":"arrest","type":"activates"},{"from":"p16","to":"arrest","type":"activates"},{"from":"arrest","to":"sasp","type":"activates"},{"from":"sasp","to":"tme","type":"activates"},{"from":"seno","to":"arrest","type":"inhibits"}],"interventions":["Senescence-inducing therapy (CDK4/6 inhibitors, chemotherapy, radiation) followed by senolytics (one-two punch)","Navitoclax and BCL-XL PROTACs; uPAR-targeted CAR-T (preclinical)","SASP modulation with JAK inhibitors or IL-6 blockade"]},{"id":"cgas-sting","kind":"pathway","name":"cGAS-STING innate sensing","aka":[],"tldr":"cGAS-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.","summary":"Cytosolic double-stranded DNA (from micronuclei after DNA damage, or from dying tumour cells taken up by dendritic cells) is sensed by cGAS, producing cGAMP, which activates STING on the ER, recruiting TBK1 to phosphorylate IRF3 and NF-κB, driving type I interferon and chemokines (CXCL10) that recruit and prime T cells. Explains the immunogenic effects of radiation (abscopal effect), PARP inhibitors, and TOP1-payload ADCs. Tumours silence STING or express ENPP1 (degrades cGAMP). STING agonists as drugs have disappointed; ENPP1 inhibitors and STING-agonist ADCs are being tested.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Stimulator_of_interferon_genes","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Stimulator_of_interferon_genes"}],"tags":[],"related":[],"cancers":[],"sections":[],"technologies":["sbrt","sting-agonist","immune-stimulating-adc","parp-inhibitor","adc"],"targets":["parp","pd1"],"drugs":[],"companies":[],"institutions":[],"pathways":[],"terms":["abscopal-effect","immunogenic-cell-death"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"dmg","label":"Radiation, chemo, ADC payload","x":50,"y":5},{"id":"dna","label":"Cytosolic dsDNA / micronuclei","x":50,"y":22},{"id":"cgas","label":"cGAS","x":50,"y":38},{"id":"cgamp","label":"cGAMP","x":50,"y":52},{"id":"enpp1","label":"ENPP1","x":85,"y":52},{"id":"sting","label":"STING","x":50,"y":66},{"id":"tbk1","label":"TBK1 → IRF3 / NF-κB","x":50,"y":80},{"id":"ifn","label":"Type I IFN, CXCL10 → T-cell recruitment","x":50,"y":95}],"edges":[{"from":"dmg","to":"dna","type":"activates"},{"from":"dna","to":"cgas","type":"activates"},{"from":"cgas","to":"cgamp","type":"activates"},{"from":"enpp1","to":"cgamp","type":"inhibits"},{"from":"cgamp","to":"sting","type":"activates"},{"from":"sting","to":"tbk1","type":"activates"},{"from":"tbk1","to":"ifn","type":"activates"}],"interventions":["Radiotherapy (especially hypofractionated) + checkpoint inhibitors","PARP inhibitor + PD-1 combinations","STING agonists (intratumoural, systemic, antibody-conjugated)","ENPP1 inhibitors","TOP1-payload ADCs + IO (ASCENT-04, EV-302 analogues)"]},{"id":"chemical-carcinogenesis-receptor-activation","kind":"pathway","name":"Chemical carcinogenesis - receptor activation","aka":["KEGG hsa05207","Chemical carcinogenesis - receptor activation"],"tldr":"This KEGG map shows how chemicals in tobacco smoke, industrial pollutants, plastics and hormones cause cancer without directly damaging DNA: they switch on receptors that drive growth signalling. It matters because these routes explain part of the cancer burden from smoking, dioxins and hormone exposure, and several of the receptors are druggable.","summary":"KEGG map hsa05207 covers the non-genotoxic arm of chemical carcinogenesis. Carcinogenesis proceeds through initiation, promotion and progression, and chemicals can act at any stage. Genotoxic agents damage DNA directly; non-genotoxic agents instead activate receptors. The map draws two receptor classes. Cell-surface receptors and some intracellular receptors trigger signal transduction: nicotine and the tobacco nitrosamines NNK and NNN act on nicotinic acetylcholine receptors and beta-adrenergic receptors to drive RAS-ERK, PI3K-AKT, JAK-STAT and cAMP signalling; bisphenol A and oestradiol act through membrane oestrogen receptors to RAS-ERK; arsenic feeds PI3K signalling; progesterone and medroxyprogesterone act through membrane and nuclear progesterone receptors. Nuclear receptors that translocate and act as transcription factors form the second class: oestradiol on the oestrogen receptor, and the aryl hydrocarbon receptor (AhR) activated by dioxin (TCDD), benzo[a]pyrene, polychlorinated biphenyls and hexachlorobenzene, which induces CYP1A1 and CYP1B1 and alters proliferation and immune genes.\n\nMurray, Patterson and Perdew, Nature Reviews Cancer, 2014 (doi:10.1038/nrc3846) review the AhR arm: AhR ligands can promote tumours by inducing metabolic enzymes that activate pro-carcinogens, by crosstalk with oestrogen receptor and growth factor signalling, and by suppressing anti-tumour immunity, while in other contexts AhR agonists suppress tumour growth, hence the title's friend and foe.\n\nWhat can be done: avoidance is the main lever (tobacco control, dioxin and PCB limits). Where the receptor is an established cancer driver the drugs already exist: endocrine therapy against the oestrogen receptor (tamoxifen, fulvestrant, aromatase inhibitors) and the androgen receptor (enzalutamide, abiraterone), and inhibitors of the downstream RAS-ERK and PI3K-AKT routes. AhR antagonists have been tested in early trials as immunotherapy sensitisers but none is approved.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05207","url":"https://www.kegg.jp/pathway/hsa05207"},{"label":"Review: Aryl hydrocarbon receptor ligands in cancer, friend and foe","url":"https://doi.org/10.1038/nrc3846"}],"tags":[],"related":["ras-mapk","pi3k-akt-mtor","jak-stat"],"cancers":[],"sections":[],"technologies":[],"targets":["estrogen-receptor","kras","pik3ca","jak2","androgen-receptor"],"drugs":["tamoxifen","fulvestrant","elacestrant","letrozole","enzalutamide","abiraterone","alpelisib","capivasertib"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","jak-stat","p53-cell-cycle"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-murray-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A doorbell that rings the growth alarm inside the cell. DNA-damaging carcinogens smash the wiring; the chemicals on this map instead stand outside pressing the bell over and over. Nicotine, dioxin, plastic additives and hormones each press a different bell, but the alarms all wire into the same growth circuits.","nodes":[{"id":"tobacco","label":"Nicotine, NNK, NNN","x":12,"y":6},{"id":"nachr","label":"Nicotinic ACh / beta-adrenergic receptors","x":12,"y":28},{"id":"e2","label":"Oestradiol, bisphenol A","x":50,"y":6},{"id":"er","label":"Oestrogen receptor (ER)","x":50,"y":28,"targetId":"estrogen-receptor"},{"id":"p4","label":"Progestins (P4, MPA)","x":72,"y":6},{"id":"dioxin","label":"TCDD, benzo[a]pyrene, PCBs","x":90,"y":6},{"id":"ahr","label":"Aryl hydrocarbon receptor (AhR)","x":90,"y":28},{"id":"ras","label":"RAS / ERK","x":25,"y":55,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT","x":50,"y":55,"targetId":"pik3ca"},{"id":"jak","label":"JAK / STAT","x":70,"y":55,"targetId":"jak2"},{"id":"cyp","label":"CYP1A1 / CYP1B1","x":90,"y":55},{"id":"out","label":"Proliferation, survival, tumour promotion","x":50,"y":92}],"edges":[{"from":"tobacco","to":"nachr","type":"activates"},{"from":"nachr","to":"ras","type":"activates"},{"from":"nachr","to":"pi3k","type":"activates"},{"from":"nachr","to":"jak","type":"activates"},{"from":"e2","to":"er","type":"activates"},{"from":"er","to":"ras","type":"activates"},{"from":"er","to":"out","type":"activates"},{"from":"p4","to":"pi3k","type":"activates"},{"from":"dioxin","to":"ahr","type":"activates"},{"from":"ahr","to":"cyp","type":"activates"},{"from":"ahr","to":"out","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"jak","to":"out","type":"activates"}],"interventions":["Prevention first: tobacco control and limits on dioxin, PCB and bisphenol exposure remove the ligands","Endocrine therapy against hormone receptors that these chemicals activate: tamoxifen, fulvestrant, elacestrant and aromatase inhibitors (letrozole) for ER; enzalutamide and abiraterone for AR","Inhibitors of the downstream growth routes: MEK inhibitors, PI3K/AKT inhibitors (alpelisib, capivasertib)","Aryl hydrocarbon receptor antagonists as immunotherapy sensitisers (early clinical trials, none approved)"]},{"id":"choline-metabolism-in-cancer","kind":"pathway","name":"Choline metabolism in cancer","aka":["KEGG hsa05231","Choline metabolism in cancer"],"tldr":"This KEGG map shows how cancer cells rewire the handling of choline, a nutrient used to build cell membranes, so that growth signals and membrane building feed each other. It matters because the resulting build-up of phosphocholine is visible on MR spectroscopy and PET scans and is one of the metabolic hallmarks of cancer.","summary":"Abnormal choline metabolism is a metabolic hallmark associated with oncogenesis and tumour progression. KEGG map hsa05231 draws how oncogenic signalling through RAS-ERK and PI3K-AKT, together with transcription factors such as hypoxia-inducible factor 1 (HIF1), raises the expression and activity of choline cycle enzymes. Choline enters through transporters (CHT1, CTL1, OCT2), is phosphorylated by choline kinase alpha (CHKA) to phosphocholine, and is built into phosphatidylcholine by the Kennedy pathway. Phospholipases C and D and the phosphatidylcholine-specific enzymes then break membrane phosphatidylcholine back down to phosphocholine, diacylglycerol (DAG) and phosphatidic acid. These products act as second messengers: DAG activates protein kinase C and phosphatidic acid supports the RAS-RAF1-MAPK cascade and mTOR, so membrane turnover and growth signalling reinforce each other.\n\nGlunde, Bhujwalla and Ronen, Nature Reviews Cancer, 2011 (doi:10.1038/nrc3162) review the field: the increase in total choline-containing compounds, and in particular the shift from glycerophosphocholine to phosphocholine, is seen across breast, prostate, brain and other cancers, is driven by CHKA over-expression and increased transporter activity, and can be imaged non-invasively by magnetic resonance spectroscopy and by choline PET. Response to targeted drugs (for example PI3K or MAPK inhibitors) lowers phosphocholine, making it a pharmacodynamic marker.\n\nWhat can be done: choline kinase inhibitors have been developed and one (TCD-717) reached a phase 1 trial, but none is approved. Today the choline pathway is used as a read-out rather than a target: choline imaging helps diagnose and monitor tumours, and drugs against the upstream drivers (PI3K/AKT/mTOR inhibitors such as alpelisib, capivasertib and everolimus; MEK inhibitors such as trametinib) lower choline metabolite levels as they work.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05231","url":"https://www.kegg.jp/pathway/hsa05231"},{"label":"Review: Choline metabolism in malignant transformation","url":"https://doi.org/10.1038/nrc3162"}],"tags":[],"related":["cancer-metabolism","pi3k-akt-mtor","hif-vhl"],"cancers":[],"sections":[],"technologies":[],"targets":["egfr","kras","mtor","hif2a"],"drugs":["alpelisib","capivasertib","everolimus","trametinib","belzutifan"],"companies":[],"institutions":[],"pathways":["cancer-metabolism","pi3k-akt-mtor","ras-mapk","hif-vhl"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-glunde-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A building site where the bricks (choline) are also the walkie-talkies. The cell orders far more bricks than it needs for walls, keeps knocking finished walls down and rebuilding them, and every knocked-down brick sends a message telling the foreman to build faster. The pile of loose bricks is what the scanner sees.","nodes":[{"id":"rtk","label":"Growth factor receptors (EGFR, HER2)","x":50,"y":5,"targetId":"egfr"},{"id":"ras","label":"RAS / RAF / MAPK","x":25,"y":22,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":75,"y":22,"targetId":"mtor"},{"id":"hif","label":"HIF1","x":90,"y":42},{"id":"transport","label":"Choline transporters (CHT1, CTL1)","x":10,"y":45},{"id":"chka","label":"Choline kinase alpha (CHKA)","x":40,"y":45},{"id":"pcho","label":"Phosphocholine (PCho)","x":40,"y":65},{"id":"ptdcho","label":"Phosphatidylcholine (membrane)","x":70,"y":65},{"id":"plc","label":"PLC / PLD","x":90,"y":80},{"id":"dag","label":"DAG / phosphatidic acid","x":55,"y":85},{"id":"pkc","label":"Protein kinase C","x":25,"y":85},{"id":"out","label":"Proliferation, survival","x":50,"y":98}],"edges":[{"from":"rtk","to":"ras","type":"activates"},{"from":"rtk","to":"pi3k","type":"activates"},{"from":"ras","to":"chka","type":"activates"},{"from":"pi3k","to":"chka","type":"activates"},{"from":"pi3k","to":"hif","type":"activates"},{"from":"hif","to":"chka","type":"activates"},{"from":"hif","to":"transport","type":"activates"},{"from":"transport","to":"chka","type":"activates"},{"from":"chka","to":"pcho","type":"activates"},{"from":"pcho","to":"ptdcho","type":"activates"},{"from":"plc","to":"ptdcho","type":"inhibits"},{"from":"plc","to":"dag","type":"activates"},{"from":"dag","to":"pkc","type":"activates"},{"from":"dag","to":"ras","type":"activates"},{"from":"pkc","to":"out","type":"activates"},{"from":"pcho","to":"out","type":"activates"}],"interventions":["Choline imaging (MR spectroscopy, choline PET) to diagnose tumours and read out response to therapy","Cut the upstream drivers: PI3K/AKT/mTOR inhibitors (alpelisib, capivasertib, everolimus) and MEK inhibitors (trametinib) lower phosphocholine as they act","Choline kinase alpha inhibitors (TCD-717 reached phase 1), no approved agent","HIF2 inhibition (belzutifan) where hypoxia signalling drives metabolic rewiring, approved in VHL-related and renal cancers"]},{"id":"chromosomal-instability","kind":"pathway","name":"Chromosomal instability & aneuploidy","aka":[],"tldr":"Most 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.","summary":"Chromosomal instability (CIN) arises from mitotic errors, whole-genome doubling, replication stress, and centrosome amplification; it produces aneuploidy, micronuclei, chromothripsis, and ecDNA (extrachromosomal oncogene amplicons). Consequences: intratumour heterogeneity and rapid adaptation (TRACERx), cytosolic DNA that activates cGAS-STING (immunogenic in bursts, tolerated chronically via non-canonical NF-κB), and proteotoxic and metabolic stress that creates dependencies (KIF18A, spindle assembly checkpoint, BCL-XL). CIN is a poor-prognosis marker across cancers; ecDNA drives resistance to targeted therapy in glioblastoma and others.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Chromosome_instability","links":[{"label":"Bakhoum et al., Chromosomal instability drives metastasis through a cytosolic DNA response (Nature 2018)","url":"https://doi.org/10.1038/nature25432"},{"label":"Cancer Grand Challenges eDyNAmiC team","url":"https://cancergrandchallenges.org/teams/edynamic"}],"tags":["mechanism"],"related":[],"cancers":["colorectal"],"sections":[],"technologies":["wes-wgs","single-cell-spatial"],"targets":["tp53","egfr"],"drugs":[],"companies":[],"institutions":["francis-crick","mskcc","stanford","cruk"],"pathways":["cgas-sting","p53-cell-cycle","replication-stress","clonal-evolution"],"terms":["genome-instability-mutation","whole-genome-doubling","mutational-signature"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bakhoum-nature"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Swanton (Crick/UCL, TRACERx, CIN and immune evasion); Bakhoum (MSK, CIN-STING); Mischel (Stanford) and the eDyNAmiC Cancer Grand Challenge on ecDNA; Sheltzer (Yale) on aneuploidy dependencies."],"analogy":"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.","nodes":[{"id":"mitosis","label":"Mitotic errors, WGD","x":15,"y":25},{"id":"cin","label":"Chromosomal instability","x":45,"y":25},{"id":"aneu","label":"Aneuploidy / karyotype heterogeneity","x":75,"y":25},{"id":"micro","label":"Micronuclei → cytosolic DNA","x":30,"y":60},{"id":"sting","label":"cGAS-STING","x":30,"y":90},{"id":"ecdna","label":"ecDNA oncogene amplification","x":75,"y":60,"targetId":"egfr"},{"id":"evol","label":"Clonal evolution, resistance","x":92,"y":45},{"id":"dep","label":"Dependencies: KIF18A, SAC, BCL-XL","x":60,"y":90},{"id":"p53","label":"TP53 loss permits","x":15,"y":60,"targetId":"tp53"}],"edges":[{"from":"mitosis","to":"cin","type":"activates"},{"from":"cin","to":"aneu","type":"activates"},{"from":"aneu","to":"evol","type":"activates"},{"from":"cin","to":"micro","type":"activates"},{"from":"micro","to":"sting","type":"activates"},{"from":"cin","to":"ecdna","type":"activates"},{"from":"ecdna","to":"evol","type":"activates"},{"from":"aneu","to":"dep","type":"activates"},{"from":"p53","to":"cin","type":"activates"}],"interventions":["KIF18A inhibitors (sovilnesib) selectively kill CIN-high cells; phase 1/2 in ovarian and TNBC","ecDNA-directed strategies (CHK1 inhibition, transcription-replication conflict) from the Cancer Grand Challenges eDyNAmiC team","STING pathway modulation; radiation exploits CIN","Aneuploidy scores (TRACERx) as prognostic biomarkers"]},{"id":"cml-signalling","kind":"pathway","name":"Chronic myeloid leukaemia (KEGG map)","aka":["KEGG hsa05220","Chronic myeloid leukemia"],"tldr":"KEGG's CML map is built around one fusion protein, BCR-ABL1, a kinase that never switches off and drives RAS, PI3K and STAT5 signalling. Because a single enzyme causes the disease, a single class of pill (imatinib and its successors) controls it in most patients.","summary":"The KEGG chronic myeloid leukaemia map (hsa05220) starts with the t(9;22) translocation that forms the Philadelphia chromosome and fuses BCR on chromosome 22 to ABL1 on chromosome 9. The p210 BCR-ABL1 protein has constitutive tyrosine kinase activity and sits in the cytoplasm, where it phosphorylates adaptor proteins (GRB2, GAB2, CRKL, CBL) and switches on three main effector routes: GRB2/SOS to RAS to RAF to MEK to ERK for proliferation, PI3K to AKT to mTOR (with inhibition of BAD and FOXO) for survival, and STAT5 for BCL-XL expression and cytokine-independent growth. Ren, Nat Rev Cancer, 2005 (doi:10.1038/nrc1567) reviews how these signals, plus reduced adhesion to marrow stroma and genomic instability, together explain the expansion of functionally normal myeloid cells in chronic phase. KEGG also draws the secondary lesions that mark progression to accelerated and blast phase: loss of TP53, RB1 and the CDKN2A locus (p16 and p14ARF), and overexpression of EVI1 or the AML1-EVI1 fusion, which disables TGF-beta growth control.\n\nWhat drugs do about it: the ATP-competitive tyrosine kinase inhibitors imatinib, dasatinib, nilotinib and bosutinib block the BCR-ABL1 kinase and turn a fatal disease into a manageable chronic one, with many patients able to stop treatment after sustained deep molecular response. Ponatinib covers the T315I gatekeeper mutation, and asciminib binds a separate myristoyl pocket (STAMP inhibitor), giving an option when the ATP site has mutated.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05220","url":"https://www.kegg.jp/pathway/hsa05220"},{"label":"Review: Mechanisms of BCR-ABL in the pathogenesis of CML","url":"https://doi.org/10.1038/nrc1567"}],"tags":[],"related":["bcr-abl1-signalling"],"cancers":["cml"],"sections":[],"technologies":[],"targets":["bcr-abl","kras","mek","akt","tp53"],"drugs":["imatinib","dasatinib","nilotinib","bosutinib","ponatinib","asciminib"],"companies":[],"institutions":[],"pathways":["bcr-abl1-signalling","ras-mapk","pi3k-akt-mtor","p53-cell-cycle"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-ren-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"BCR-ABL1 is a light switch glued in the on position, feeding power to three circuits (RAS, PI3K, STAT5) that tell blood cells to multiply. Imatinib and its successors are a cover fitted over the switch; asciminib works from the back of the switch plate, so it still fits when the front is damaged.","nodes":[{"id":"bcrabl","label":"BCR-ABL1 (t(9;22))","x":50,"y":6,"targetId":"bcr-abl"},{"id":"adapt","label":"GRB2 / GAB2 / CRKL","x":50,"y":24},{"id":"ras","label":"RAS / RAF","x":18,"y":42,"targetId":"kras"},{"id":"mek","label":"MEK / ERK","x":18,"y":60,"targetId":"mek"},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":50,"y":46,"targetId":"akt"},{"id":"stat5","label":"STAT5 to BCL-XL","x":82,"y":42},{"id":"prolif","label":"Proliferation, survival","x":50,"y":74},{"id":"p53","label":"TP53 / RB1 / p16 loss (blast phase)","x":84,"y":74,"targetId":"tp53"},{"id":"cml","label":"Chronic phase to blast crisis","x":60,"y":93}],"edges":[{"from":"bcrabl","to":"adapt","type":"activates"},{"from":"adapt","to":"ras","type":"activates"},{"from":"ras","to":"mek","type":"activates"},{"from":"adapt","to":"pi3k","type":"activates"},{"from":"bcrabl","to":"stat5","type":"activates"},{"from":"mek","to":"prolif","type":"activates"},{"from":"pi3k","to":"prolif","type":"activates"},{"from":"stat5","to":"prolif","type":"activates"},{"from":"prolif","to":"cml","type":"activates"},{"from":"p53","to":"cml","type":"activates"}],"interventions":["First and second generation BCR-ABL1 tyrosine kinase inhibitors: imatinib, dasatinib, nilotinib, bosutinib","Ponatinib for T315I-mutant or multi-resistant CML","Asciminib, an allosteric STAMP inhibitor of the ABL1 myristoyl pocket, for previously treated CML","Treatment-free remission after sustained deep molecular response on a TKI","Allogeneic stem cell transplant for blast phase or TKI-refractory disease"]},{"id":"circadian-control","kind":"pathway","name":"Circadian control","aka":[],"tldr":"Cells run on a 24-hour clock that gates cell division, DNA repair, and drug metabolism. Cancers often break their clocks, and the time of day a drug or immunotherapy is given can change how well it works.","summary":"The CLOCK/BMAL1-PER/CRY loop times metabolism, cell-cycle checkpoints, and DNA repair; shift work is a probable carcinogen (IARC 2A) and clock-gene disruption accelerates tumorigenesis in mice. Chronomodulated chemotherapy (Lévi, oxaliplatin/5-FU) improved tolerability in colorectal cancer; retrospective and prospective data (MEMOIR, 2024-25) suggest checkpoint inhibitors given earlier in the day yield longer survival, likely through T-cell trafficking rhythms. Clock-targeting drugs (REV-ERB agonists, CRY stabilisers) are preclinical. Implementation is cheap but trial evidence is still limited.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Chronotherapy_(treatment_scheduling)","links":[{"label":"Qian et al., Effect of immunotherapy time-of-day infusion on overall survival among patients with advanced melanoma (Lancet Oncology 2021)","url":"https://doi.org/10.1016/S1470-2045(21)00546-5"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":["checkpoint-inhibitor","cytotoxic-chemotherapy"],"targets":["wee1"],"drugs":[],"companies":[],"institutions":["mskcc","penn-abramson"],"pathways":["p53-cell-cycle","myc","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-qian-lancet-oncol"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Lévi (Warwick/Paris) on chronotherapy; Sassone-Corsi legacy (UC Irvine); Scheiermann (Geneva) on immune rhythms; MEMOIR trial (Winship/Emory)."],"analogy":"A city that lowers its bridges only at certain hours. Send the army (drug, T cells) when the bridges are down and it gets in; send it at midnight and it waits outside.","nodes":[{"id":"clock","label":"CLOCK/BMAL1","x":25,"y":35},{"id":"per","label":"PER/CRY (repress)","x":25,"y":70},{"id":"cycle","label":"Cell-cycle gating (WEE1, MYC)","x":60,"y":20,"targetId":"wee1"},{"id":"repair","label":"DNA repair timing","x":60,"y":50},{"id":"immune","label":"T-cell trafficking rhythm","x":60,"y":80},{"id":"tx","label":"Time-of-day drug response","x":90,"y":50}],"edges":[{"from":"clock","to":"per","type":"activates"},{"from":"per","to":"clock","type":"inhibits"},{"from":"clock","to":"cycle","type":"activates"},{"from":"clock","to":"repair","type":"activates"},{"from":"clock","to":"immune","type":"activates"},{"from":"cycle","to":"tx","type":"activates"},{"from":"repair","to":"tx","type":"activates"},{"from":"immune","to":"tx","type":"activates"}],"interventions":["Morning versus afternoon immunotherapy dosing (prospective trials ongoing)","Chronomodulated chemotherapy infusion","Clock-modulating compounds (preclinical)"]},{"id":"clonal-evolution","kind":"pathway","name":"Clonal evolution & minimal residual disease","aka":[],"tldr":"A 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.","summary":"Truncal mutations are shared by all cells; branched subclones carry private alterations (TRACERx, PCAWG). Therapy imposes selection: pre-existing resistant clones (EGFR T790M, ESR1) expand, or drug-tolerant persisters acquire mutations later. Neutral and punctuated evolution, whole-genome doubling, and CIN modulate the tempo. ctDNA lets clonal dynamics be followed in real time; MRD detection after curative therapy identifies who will relapse. Adaptive therapy (dose modulation to maintain sensitive competitors, Moffitt prostate pilot) and combination strategies aim to steer rather than merely suppress evolution.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Somatic_evolution_in_cancer","links":[{"label":"Jamal-Hanjani et al., Tracking the evolution of non-small-cell lung cancer (NEJM 2017)","url":"https://doi.org/10.1056/NEJMoa1616288"},{"label":"Zhang et al., Integrating evolutionary dynamics into treatment of mCRPC (Nature Communications 2017)","url":"https://doi.org/10.1038/s41467-017-01968-5"}],"tags":["mechanism"],"related":["theories-of-cancer","clonal-evolution-theory"],"cancers":[],"sections":[],"technologies":["mrd-testing","liquid-biopsy","wes-wgs"],"targets":[],"drugs":[],"companies":[],"institutions":["francis-crick","moffitt","broad-institute","mskcc"],"pathways":["chromosomal-instability","metastatic-cascade","tumor-dormancy"],"terms":["mrd","ctdna","resistance","vaf","mutational-signature"],"trials":["imvigor011","dynamic"],"people":[],"bottlenecks":[],"keyPapers":["paper-zhang-nat-commun"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Swanton (Crick, TRACERx); Gatenby and Brown (Moffitt, adaptive therapy); Getz (Broad) on clonal reconstruction; Landau (Weill Cornell) on single-cell evolution."],"analogy":"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.","nodes":[{"id":"trunk","label":"Truncal driver clone","x":12,"y":50,"targetId":"tp53"},{"id":"b1","label":"Subclone A","x":38,"y":25},{"id":"b2","label":"Subclone B (resistant)","x":38,"y":75},{"id":"tx","label":"Therapy (selection)","x":60,"y":50},{"id":"mrd","label":"MRD (ctDNA)","x":78,"y":50},{"id":"relapse","label":"Relapse dominated by B","x":92,"y":75},{"id":"adapt","label":"Adaptive / combination therapy","x":92,"y":20}],"edges":[{"from":"trunk","to":"b1","type":"activates"},{"from":"trunk","to":"b2","type":"activates"},{"from":"tx","to":"b1","type":"inhibits"},{"from":"b2","to":"mrd","type":"activates"},{"from":"mrd","to":"relapse","type":"activates"},{"from":"adapt","to":"relapse","type":"inhibits"}],"interventions":["ctDNA MRD to escalate or de-escalate (IMvigor011, DYNAMIC)","Upfront combinations to pre-empt resistant clones (osimertinib + chemotherapy, BRAF + MEK)","Adaptive therapy trials (Moffitt)","Serial liquid biopsy to switch therapy at molecular progression (SERENA-6)"]},{"id":"clonal-haematopoiesis","kind":"pathway","name":"Clonal haematopoiesis (CHIP)","aka":[],"tldr":"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.","summary":"Clonal haematopoiesis of indeterminate potential (CHIP: DNMT3A, TET2, ASXL1, and after chemotherapy PPM1D, TP53, CHEK2 mutations at VAF ≥2%) is present in >10% of people over 70. It confers ~0.5-1%/year progression to myeloid neoplasm, doubles cardiovascular risk via inflammasome-primed macrophages, and predicts therapy-related MDS/AML after PARP inhibitors, platinum, and radioligand therapy. For liquid biopsy, CHIP variants are the main source of false-positive ctDNA calls, so tumour-informed or paired white-cell sequencing is needed. Interventions are preventive (avoid unnecessary genotoxic exposure, monitor) with IL-1β/IL-6 blockade under study.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Clonal_hematopoiesis","links":[{"label":"Jaiswal et al., Age-related clonal hematopoiesis associated with adverse outcomes (NEJM 2014)","url":"https://doi.org/10.1056/NEJMoa1408617"},{"label":"Bolton et al., Cancer therapy shapes the fitness landscape of clonal hematopoiesis (Nature Genetics 2020)","url":"https://doi.org/10.1038/s41588-020-00710-0"}],"tags":["mechanism"],"related":["ageing-tissue-field-theory"],"cancers":["aml","prostate","prostate-mcrpc"],"sections":[],"technologies":["liquid-biopsy","mrd-testing","parp-inhibitor"],"targets":["tp53","parp","tet2"],"drugs":[],"companies":[],"institutions":["dana-farber","broad-institute","mskcc"],"pathways":["clonal-evolution","inflammation-nfkb","epigenetic-reprogramming"],"terms":["vaf","ctdna"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-jensen-clonal-haematopoiesis-cfdna-interference-prostate-jama-oncol-2021","paper-tukachinsky-ctdna-3334-advanced-prostate-ccr-2021","paper-bolton-nat-genet"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Ebert (Dana-Farber/Broad) and Jaiswal (Stanford) who defined CHIP; Levine and Bolton (MSK) on therapy-related clonal haematopoiesis.","Prostate cancer: this is where clonal haematopoiesis does concrete harm. In 69 men with advanced prostate cancer, clonal haematopoiesis variants at 2% or more were present in plasma from 13, 19%, and 7, 10%, had one in a gene used to decide PARP inhibitor candidacy, most often ATM, accounting for almost half of all somatic DNA repair variants detected. A paired whole-blood control distinguishes them (Jensen 2021)."],"analogy":"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.","nodes":[{"id":"hsc","label":"Ageing HSCs","x":15,"y":50},{"id":"mut","label":"DNMT3A, TET2, ASXL1, PPM1D, TP53","x":40,"y":50,"targetId":"tp53"},{"id":"chip","label":"CHIP clone (VAF ≥2%)","x":65,"y":50},{"id":"mds","label":"Therapy-related MDS/AML","x":90,"y":25},{"id":"cvd","label":"Inflammation → cardiovascular disease","x":90,"y":60},{"id":"fp","label":"False-positive ctDNA","x":90,"y":90},{"id":"chemo","label":"Chemo, PARPi, radioligands select","x":40,"y":85}],"edges":[{"from":"hsc","to":"mut","type":"activates"},{"from":"mut","to":"chip","type":"activates"},{"from":"chip","to":"mds","type":"activates"},{"from":"chip","to":"cvd","type":"activates"},{"from":"chip","to":"fp","type":"activates"},{"from":"chemo","to":"chip","type":"activates"}],"interventions":["Paired white-blood-cell sequencing in liquid biopsy pipelines","Monitoring after PARP inhibitors and radioligand therapy","IL-1β/IL-6 blockade trials for CHIP-associated cardiovascular risk"]},{"id":"immune-desert-exclusion","kind":"pathway","name":"Cold tumours: immune deserts and exclusion","aka":[],"tldr":"Tumours 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.","summary":"Deserts arise from low antigenicity (low TMB, MHC loss), failed priming (few BATF3+ cDC1, low CCL4 because of tumour-intrinsic WNT/β-catenin signalling; PTEN loss; MYC-driven CD47/PD-L1), and absent chemokines (CXCL9/10 silenced by EZH2 and DNMT1). Exclusion arises from stroma: TGF-β-activated CAFs and dense collagen (Mariathasan 2018), CXCL12 from FAP+ fibroblasts, abnormal VEGF-driven vessels lacking adhesion molecules, and myeloid barriers. Inflamed tumours still fail through PD-L1, exhaustion and Tregs. Converters: radiotherapy and chemotherapy (immunogenic death, STING), oncolytic viruses and in situ vaccines, STING agonists (systemic versions disappointed), anti-VEGF and TGF-β blockade (bintrafusp alfa failed), FAP/CXCR4 targeting, epigenetic priming to restore chemokines, and antigen-independent killers (engagers, CAR-T) that do not need a hot tumour. Gene signatures (T-cell inflamed GEP, TIS) and spatial pathology grade the weather.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Tumor-infiltrating_lymphocytes","links":[{"label":"Chen & Mellman, Elements of cancer immunity and the cancer-immune set point (Nature 2017)","url":"https://doi.org/10.1038/nature21349"},{"label":"Spranger, Bao & Gajewski, Melanoma-intrinsic β-catenin signalling prevents anti-tumour immunity (Nature 2015)","url":"https://doi.org/10.1038/nature14404"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["pancreatic"],"sections":[],"technologies":["checkpoint-inhibitor","oncolytic-virus","in-situ-vaccination","sting-agonist","sbrt","epigenetic-drugs","t-cell-engager","single-cell-spatial","il12-electroporation"],"targets":["pd1","pdl1","vegf","fap","ezh2","csf1r","cxcr4","pik3ca"],"drugs":["talimogene-laherparepvec","vusolimogene-oderparepvec","ivonescimab","adu-s100","tazemetostat"],"companies":[],"institutions":[],"pathways":["tgf-beta","wnt","cgas-sting","tumor-microenvironment","cancer-immunity-cycle","myeloid-suppression-axis"],"terms":["cold-vs-hot","immune-exclusion","tils","tmb","cps","desmoplasia"],"trials":[],"people":[],"bottlenecks":["b-immunotherapy-response","b-tme-immunosuppression"],"keyPapers":["paper-chen-nature","paper-spranger-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"ag","label":"Low TMB, MHC loss","x":12,"y":12},{"id":"wnt","label":"β-catenin, PTEN loss → no cDC1","x":12,"y":40,"targetId":"pik3ca"},{"id":"chemo","label":"CXCL9/10 silenced (EZH2)","x":12,"y":68,"targetId":"ezh2"},{"id":"desert","label":"Immune desert","x":42,"y":40},{"id":"tgf","label":"TGF-β CAFs, collagen","x":72,"y":12,"targetId":"fap"},{"id":"vess","label":"Abnormal vessels (VEGF)","x":72,"y":40,"targetId":"vegf"},{"id":"excl","label":"Immune exclusion","x":92,"y":60},{"id":"hot","label":"Inflamed → PD-1 response","x":42,"y":90,"targetId":"pd1"},{"id":"conv","label":"RT, STING, viruses, vaccines","x":12,"y":92},{"id":"myeloid","label":"Myeloid barrier","x":72,"y":68,"targetId":"csf1r"}],"edges":[{"from":"ag","to":"desert","type":"activates"},{"from":"wnt","to":"desert","type":"activates"},{"from":"chemo","to":"desert","type":"activates"},{"from":"tgf","to":"excl","type":"activates"},{"from":"vess","to":"excl","type":"activates"},{"from":"myeloid","to":"excl","type":"activates"},{"from":"desert","to":"hot","type":"inhibits"},{"from":"excl","to":"hot","type":"inhibits"},{"from":"conv","to":"hot","type":"activates"},{"from":"conv","to":"desert","type":"inhibits"}],"interventions":["Radiotherapy, immunogenic chemotherapy and TOP1 ADCs to seed antigen and STING signalling; oncolytic viruses (T-VEC, RP1) and in situ vaccination","Anti-VEGF and PD-1×VEGF bispecifics open the vessel gate; FAP-, CXCR4- and TGF-β-directed agents (mostly modest so far)","Epigenetic priming (EZH2, DNMT inhibitors) to restore chemokines; STING agonists intratumourally","Bypass the weather: T-cell engagers, CAR-T, TCR-T and vaccines that bring or make their own T cells"]},{"id":"colorectal-cancer-signalling","kind":"pathway","name":"Colorectal cancer (KEGG map)","aka":["KEGG hsa05210","Colorectal cancer"],"tldr":"This KEGG map traces the step-by-step genetic route from normal bowel lining to colorectal cancer: APC loss unleashes Wnt, KRAS mutation drives growth, then TP53 and TGF-beta/SMAD4 loss remove the last brakes, or alternatively mismatch repair fails and mutations pile up. Knowing which route a tumour took decides which drugs work.","summary":"Colorectal cancer arises from the colorectal epithelium through accumulated alterations in defined oncogenes and tumour suppressors. KEGG map hsa05210 draws two mechanisms of genomic instability. The chromosomal instability (CIN) route follows the classical adenoma to carcinoma sequence: inactivation of APC (germline in familial adenomatous polyposis) or activating mutation of CTNNB1 stabilises beta-catenin and switches on Wnt target genes (MYC, cyclin D1); activating KRAS mutation, or over-expression of EGFR and its ligands (EGF, TGF-alpha, amphiregulin, epiregulin), drives RAS-ERK and PI3K-AKT signalling; then loss of TP53, of SMAD4 and DCC on chromosome 18q, and of TGF-beta receptor II removes growth arrest and apoptosis. The microsatellite instability (MSI) route follows inactivation of the mismatch repair genes MLH1 (usually by promoter hypermethylation) or MSH2 (germline in Lynch syndrome, HNPCC), with secondary frameshift mutations in genes carrying coding microsatellites such as TGFBR2 and BAX.\n\nDekker and colleagues, The Lancet, 2019 (doi:10.1016/S0140-6736(19)32319-0) place these routes in the clinic: KRAS, NRAS and BRAF status decides whether anti-EGFR antibodies can be used, MSI status decides whether immune checkpoint inhibitors are appropriate, and the consensus molecular subtypes separate MSI-immune, canonical Wnt, metabolic and mesenchymal TGF-beta-driven tumours.\n\nWhat drugs do about it: anti-EGFR antibodies (cetuximab, panitumumab) block the receptor layer in RAS wild-type, left-sided tumours; encorafenib plus cetuximab treats BRAF V600E disease; KRAS G12C inhibitors (sotorasib, adagrasib) combined with anti-EGFR treat the G12C subset; pembrolizumab, nivolumab and ipilimumab treat MSI-high tumours; bevacizumab, ramucirumab, regorafenib and fruquintinib block VEGF-driven angiogenesis; HER2-directed therapy (trastuzumab with tucatinib, trastuzumab deruxtecan) treats HER2-amplified tumours. No approved drug yet targets the APC/Wnt step itself.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05210","url":"https://www.kegg.jp/pathway/hsa05210"},{"label":"Review: Colorectal cancer (Lancet seminar)","url":"https://doi.org/10.1016/S0140-6736(19)32319-0"}],"tags":[],"related":["wnt","ras-mapk","mismatch-repair-msi","tgf-beta"],"cancers":["colorectal"],"sections":[],"technologies":[],"targets":["egfr","kras","braf","pik3ca","tp53","her2","pd1","vegf"],"drugs":["cetuximab","panitumumab","encorafenib","sotorasib","adagrasib","pembrolizumab","nivolumab","ipilimumab","bevacizumab","ramucirumab","regorafenib","fruquintinib","tucatinib","trastuzumab-deruxtecan","trifluridine-tipiracil"],"companies":[],"institutions":[],"pathways":["wnt","ras-mapk","pi3k-akt-mtor","p53-cell-cycle","tgf-beta","mismatch-repair-msi","apoptosis-bcl2","pd1-checkpoint","vegf-angiogenesis"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-dekker-lancet"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A car with several independent brakes and an accelerator. APC loss removes the handbrake (Wnt runs), KRAS jams the accelerator, TP53 and SMAD4 loss cut the foot brakes. In the MSI route the mechanic who proofreads the repair manual (mismatch repair) is sacked, so faults accumulate everywhere at once, which oddly makes the car more visible to the immune system's traffic police.","nodes":[{"id":"apc","label":"APC (lost)","x":12,"y":8},{"id":"bcat","label":"beta-catenin / TCF","x":12,"y":30},{"id":"egfr","label":"EGFR + ligands","x":45,"y":8,"targetId":"egfr"},{"id":"kras","label":"KRAS","x":45,"y":30,"targetId":"kras"},{"id":"braf","label":"BRAF / MEK / ERK","x":32,"y":52,"targetId":"braf"},{"id":"pi3k","label":"PI3K / AKT","x":58,"y":52,"targetId":"pik3ca"},{"id":"tgfb","label":"TGF-beta receptor II / SMAD4","x":85,"y":8},{"id":"p53","label":"TP53","x":85,"y":40,"targetId":"tp53"},{"id":"mmr","label":"MLH1 / MSH2 (mismatch repair)","x":85,"y":68},{"id":"bax","label":"BAX / DCC apoptosis","x":62,"y":80},{"id":"out","label":"Adenoma to carcinoma","x":30,"y":95}],"edges":[{"from":"apc","to":"bcat","type":"inhibits"},{"from":"bcat","to":"out","type":"activates"},{"from":"egfr","to":"kras","type":"activates"},{"from":"kras","to":"braf","type":"activates"},{"from":"kras","to":"pi3k","type":"activates"},{"from":"braf","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"tgfb","to":"out","type":"inhibits"},{"from":"p53","to":"out","type":"inhibits"},{"from":"p53","to":"bax","type":"activates"},{"from":"bax","to":"out","type":"inhibits"},{"from":"mmr","to":"tgfb","type":"activates"},{"from":"mmr","to":"bax","type":"activates"}],"interventions":["Anti-EGFR antibodies (cetuximab, panitumumab) for RAS wild-type, BRAF wild-type, left-sided tumours","Encorafenib plus cetuximab for BRAF V600E colorectal cancer","KRAS G12C inhibitors (sotorasib, adagrasib) combined with anti-EGFR antibodies for the G12C subset","Immune checkpoint inhibitors (pembrolizumab, nivolumab plus ipilimumab) for mismatch repair deficient, MSI-high tumours","Anti-angiogenics (bevacizumab, ramucirumab, regorafenib, fruquintinib) and HER2-directed therapy (trastuzumab with tucatinib, trastuzumab deruxtecan) for HER2-amplified disease","Wnt/APC step: no approved drug yet"]},{"id":"complement-in-cancer","kind":"pathway","name":"Complement in cancer","aka":[],"tldr":"Complement 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.","summary":"Three routes converge on C3 convertase: classical (C1q binding IgG1/IgG3 Fc clusters, the basis of complement-dependent cytotoxicity by rituximab, obinutuzumab less so, daratumumab), lectin, and alternative (spontaneous C3 tick-over). C3b opsonises for phagocytosis via CR3; C5 convertase releases C5a and assembles the membrane attack complex (C5b-9). Tumours over-express membrane regulators CD46, CD55 (DAF) and CD59 (protectin) and secrete factor H, limiting CDC; CD20 shaving and antigen loss add to rituximab resistance. Conversely, intratumoural complement activation is often pro-tumour: C5a-C5aR1 and C3a-C3aR recruit and activate MDSCs and neutrophils, suppress CD8 T cells, and promote angiogenesis; C1q from macrophages supports invasion. Blocking C5aR1 (avdoralimab) with anti-PD-1 was tested; complement is also a mediator of infusion reactions and of CAR-T-associated inflammation. Fc engineering (afucosylation for ADCC versus hexamerisation for CDC) tunes which effector dominates.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Complement_system","links":[{"label":"Roumenina et al., Context-dependent roles of complement in cancer (Nat Rev Cancer 2019)","url":"https://doi.org/10.1038/s41568-019-0210-0"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["dlbcl","follicular-lymphoma","cll","multiple-myeloma"],"sections":[],"technologies":["monoclonal-antibody","checkpoint-inhibitor"],"targets":["cd20","cd38","cd19","pd1","csf1r"],"drugs":["rituximab","obinutuzumab","daratumumab","isatuximab"],"companies":[],"institutions":[],"pathways":["myeloid-suppression-axis","inflammation-nfkb","nk-cell-recognition","extrinsic-apoptosis-death-receptors"],"terms":["adcc","fc-effector","crs","myeloid-derived-suppressor-cells"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-roumenina-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"ab","label":"IgG1 antibody (rituximab)","x":12,"y":15,"targetId":"cd20"},{"id":"c1q","label":"C1q classical route","x":40,"y":15},{"id":"alt","label":"Alternative / lectin","x":40,"y":45},{"id":"c3","label":"C3 convertase → C3b","x":68,"y":30},{"id":"c5","label":"C5 → C5a + MAC (C5b-9)","x":92,"y":45},{"id":"reg","label":"CD46, CD55, CD59 shields","x":68,"y":65},{"id":"lysis","label":"CDC lysis, opsonisation","x":68,"y":92},{"id":"c5a","label":"C5a → MDSC, neutrophils","x":40,"y":78,"targetId":"csf1r"},{"id":"tcell","label":"CD8 T cells suppressed","x":12,"y":78,"targetId":"pd1"},{"id":"adcc","label":"ADCC (NK, CD16)","x":12,"y":45}],"edges":[{"from":"ab","to":"c1q","type":"activates"},{"from":"ab","to":"adcc","type":"activates"},{"from":"c1q","to":"c3","type":"activates"},{"from":"alt","to":"c3","type":"activates"},{"from":"c3","to":"c5","type":"activates"},{"from":"reg","to":"c3","type":"inhibits"},{"from":"reg","to":"c5","type":"inhibits"},{"from":"c5","to":"lysis","type":"activates"},{"from":"c5","to":"c5a","type":"activates"},{"from":"c5a","to":"tcell","type":"inhibits"}],"interventions":["CD20 and CD38 antibodies (rituximab, obinutuzumab, daratumumab, isatuximab) kill partly through complement; Fc engineering tunes CDC versus ADCC","Blocking complement regulators (CD55/CD59) to restore CDC (preclinical)","C5aR1 antagonists with checkpoint inhibitors to remove myeloid recruitment (early trials)","Complement inhibition to manage infusion reactions and CAR-T inflammation"]},{"id":"ddr","kind":"pathway","name":"DNA damage response & homologous recombination","aka":[],"tldr":"The 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.","summary":"Single-strand breaks recruit PARP1 for base-excision repair. Unrepaired breaks collapse replication forks into double-strand breaks, repaired by homologous recombination (BRCA1, BRCA2, PALB2, RAD51) in S/G2 or by error-prone NHEJ/POLQ-mediated end joining. ATR senses replication stress and signals via CHK1; ATM senses double-strand breaks. HR-deficient tumours (BRCA mutation, HRD) depend on PARP, POLQ, and ATR: the basis of synthetic lethality. Resistance: BRCA reversion, 53BP1/Shieldin loss restoring HR, drug efflux.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/DNA_repair","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/DNA_repair"}],"tags":[],"related":[],"cancers":["ovarian","tnbc","prostate","pancreatic"],"sections":[],"technologies":["parp-inhibitor","hrd-testing","synthetic-lethality-approaches","platinum","parp-pet"],"targets":["parp","brca","atr","wee1","tp53"],"drugs":["olaparib","niraparib","talazoparib","carboplatin"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"ssb","label":"Single-strand break","x":20,"y":8},{"id":"parp","label":"PARP1","x":20,"y":28,"targetId":"parp"},{"id":"fork","label":"Replication fork collapse","x":50,"y":45},{"id":"atr","label":"ATR / CHK1","x":80,"y":30,"targetId":"atr"},{"id":"dsb","label":"Double-strand break","x":50,"y":62},{"id":"brca","label":"BRCA1/2-RAD51 (HR)","x":25,"y":82,"targetId":"brca"},{"id":"nhej","label":"NHEJ / POLQ (error-prone)","x":75,"y":82},{"id":"fix","label":"Accurate repair","x":25,"y":97},{"id":"death","label":"Genomic collapse / death","x":75,"y":97}],"edges":[{"from":"ssb","to":"parp","type":"activates"},{"from":"parp","to":"fork","type":"inhibits"},{"from":"ssb","to":"fork","type":"activates"},{"from":"fork","to":"dsb","type":"activates"},{"from":"atr","to":"fork","type":"inhibits"},{"from":"dsb","to":"brca","type":"activates"},{"from":"dsb","to":"nhej","type":"activates"},{"from":"brca","to":"fix","type":"activates"},{"from":"nhej","to":"death","type":"activates"}],"interventions":["PARP inhibitors in BRCA/HRD ovarian, breast, prostate, pancreatic cancer","Platinum chemotherapy (crosslinks) in HRD tumours","ATR inhibitors (ceralasertib) in ATM-deficient or PARP-resistant tumours","POLQ inhibitors (novobiocin analogues) in HRD","PARP1-selective saruparib to widen therapeutic window"]},{"id":"dna-replication-licensing","kind":"pathway","name":"DNA replication & origin licensing","aka":[],"tldr":"Before a cell divides it must copy three billion letters of DNA exactly once, 'licensing' thousands of start points in advance and firing them in waves. Cancers driven by MYC, cyclin E or RAS fire excess start points too fast, and antimetabolites such as 5-FU, topoisomerase poisons such as irinotecan and platinum drugs all jam this copying machinery.","summary":"In G1, ORC binds origins and with CDC6 and CDT1 loads double MCM2-7 hexamers (licensing). At S-phase entry CDK2 and DDK (CDC7) convert MCM into the active CMG helicase (with CDC45, GINS), recruiting polymerases: Pol α-primase starts, Pol ε (leading) and Pol δ (lagging) extend, with PCNA as clamp, RPA coating single strands, and Okazaki fragments joined by FEN1/LIG1. Re-licensing within one cycle is prevented by CDT1 degradation (CRL4-CDT2), geminin, and CDK activity. Oncogenes (MYC, cyclin E, RAS) shorten G1, fire excess and ectopic origins, and cause origin-fork collisions and dNTP exhaustion, the root of replication stress. Chemotherapy exploits this: antimetabolites deplete dNTPs (5-FU, methotrexate, pemetrexed, hydroxyurea inhibits RNR) or terminate chains (gemcitabine, cytarabine); topoisomerase I poisons (irinotecan, topotecan, ADC payloads SN-38, DXd, exatecan) trap the enzyme ahead of the fork; platinum crosslinks block polymerases. POLQ, POLA1 and CDC7 inhibitors are experimental.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Eukaryotic_DNA_replication","links":[{"label":"Bell & Dutta, DNA replication in eukaryotic cells (Annu Rev Biochem 2002)","url":"https://doi.org/10.1146/annurev.biochem.71.110601.135425"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["cytotoxic-chemotherapy","topoisomerase-inhibitors","platinum","adc"],"targets":["kras","cdk4-6","atr","trop2","her2"],"drugs":["gemcitabine","fluorouracil","methotrexate","pemetrexed","hydroxyurea","irinotecan","topotecan","trastuzumab-deruxtecan","sacituzumab-govitecan"],"companies":[],"institutions":[],"pathways":["replication-stress","cell-cycle-engine-cdks","ddr","myc"],"terms":["topoisomerase-i-payloads","sn-38","dxd","exatecan","payload"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bell-annu-rev-biochem"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"orc","label":"ORC, CDC6, CDT1","x":12,"y":15},{"id":"mcm","label":"MCM2-7 loaded (licence)","x":40,"y":15},{"id":"gem","label":"Geminin, CRL4-CDT2","x":40,"y":45},{"id":"cdk","label":"CDK2 / DDK firing","x":68,"y":15,"targetId":"cdk4-6"},{"id":"cmg","label":"CMG helicase + Pol ε/δ","x":68,"y":45},{"id":"dntp","label":"dNTP supply (RNR)","x":90,"y":75},{"id":"fork","label":"Replication fork","x":68,"y":75},{"id":"top1","label":"TOP1 (chemo, ADC payloads)","x":40,"y":75,"targetId":"trop2"},{"id":"onc","label":"MYC, cyclin E: excess origins","x":12,"y":45,"targetId":"kras"},{"id":"stress","label":"Replication stress","x":12,"y":80}],"edges":[{"from":"orc","to":"mcm","type":"activates"},{"from":"gem","to":"mcm","type":"inhibits"},{"from":"mcm","to":"cdk","type":"activates"},{"from":"cdk","to":"cmg","type":"activates"},{"from":"cmg","to":"fork","type":"activates"},{"from":"dntp","to":"fork","type":"activates"},{"from":"top1","to":"fork","type":"activates"},{"from":"onc","to":"cdk","type":"activates"},{"from":"onc","to":"stress","type":"activates"},{"from":"fork","to":"stress","type":"inhibits"}],"interventions":["Antimetabolites (5-FU, gemcitabine, methotrexate, pemetrexed, hydroxyurea, cytarabine) starve or terminate synthesis","Topoisomerase I poisons (irinotecan, topotecan) and their ADC payloads (SN-38, DXd, exatecan) trap the enzyme ahead of forks","Platinum crosslinks and alkylators block polymerases","ATR, CHK1, WEE1 inhibitors exploit the stress cancers create (see replication stress); CDC7 and POLQ inhibitors in trials"]},{"id":"replication-stress","kind":"pathway","name":"DNA replication stress","aka":[],"tldr":"Cancers 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.","summary":"Oncogene activation (MYC, cyclin E, RAS) shortens G1, increases origin firing, and causes fork stalling, ssDNA gaps, and transcription-replication conflicts. ATR senses stalled forks and signals via CHK1 to slow origin firing and stabilise forks; WEE1 restrains CDK1/2. TP53-mutant and CCNE1-amplified cells depend on this axis (G1 checkpoint gone, G2/M checkpoint essential), the rationale for ATR (ceralasertib, camonsertib), CHK1, WEE1 (azenosertib), and PKMYT1 (lunresertib) inhibitors, often with PARP inhibitors or chemotherapy. Toxicity is the recurring limitation because normal proliferating tissue also uses these checkpoints.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Replication_stress","links":[{"label":"Saxena & Zou, Hallmarks of DNA replication stress (Molecular Cell 2022)","url":"https://doi.org/10.1016/j.molcel.2022.05.004"}],"tags":["mechanism"],"related":[],"cancers":["colorectal"],"sections":[],"technologies":["synthetic-lethality-approaches","parp-inhibitor"],"targets":["atr","wee1","tp53","parp"],"drugs":[],"companies":[],"institutions":["icr-london","dana-farber","nki","francis-crick"],"pathways":["ddr","p53-cell-cycle","myc"],"terms":["synthetic-lethality","genome-instability-mutation"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-saxena-mol-cell"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Zou (MGH/Duke) and Cimprich (Stanford) on replication-stress signalling; ICR and Dana-Farber early-phase ATR/WEE1 programmes; NKI on CCNE1 dependencies."],"analogy":"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.","nodes":[{"id":"onc","label":"Oncogenes (MYC, cyclin E, RAS)","x":15,"y":20,"targetId":"kras"},{"id":"origin","label":"Excess origin firing, short G1","x":45,"y":20},{"id":"fork","label":"Stalled forks, ssDNA gaps","x":75,"y":20},{"id":"atr","label":"ATR → CHK1","x":75,"y":55,"targetId":"atr"},{"id":"wee1","label":"WEE1 / PKMYT1 restrain CDK1","x":45,"y":55,"targetId":"wee1"},{"id":"g2m","label":"G2/M checkpoint","x":15,"y":55},{"id":"collapse","label":"Fork collapse → DSBs","x":75,"y":88},{"id":"mitosis","label":"Mitotic catastrophe","x":30,"y":88},{"id":"p53","label":"TP53 (lost)","x":92,"y":55,"targetId":"tp53"}],"edges":[{"from":"onc","to":"origin","type":"activates"},{"from":"origin","to":"fork","type":"activates"},{"from":"fork","to":"atr","type":"activates"},{"from":"atr","to":"wee1","type":"activates"},{"from":"wee1","to":"g2m","type":"activates"},{"from":"atr","to":"origin","type":"inhibits"},{"from":"fork","to":"collapse","type":"activates"},{"from":"atr","to":"collapse","type":"inhibits"},{"from":"g2m","to":"mitosis","type":"inhibits"},{"from":"p53","to":"g2m","type":"activates"}],"interventions":["ATR inhibitors (ceralasertib, camonsertib) alone and with PARP inhibitors or IO","WEE1 (azenosertib) and PKMYT1 (lunresertib) inhibitors in CCNE1-amplified and TP53-mutant tumours","PARP inhibitors trap forks in HRD tumours (see DDR)","Gemcitabine and other antimetabolites are classical replication-stress inducers"]},{"id":"homologous-recombination-repair","kind":"pathway","name":"Double-strand break repair: HR versus end joining","aka":[],"tldr":"A 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.","summary":"The MRN complex (MRE11-RAD50-NBS1) senses the break and activates ATM, which phosphorylates H2AX, CHK2 and p53. Pathway choice is set by resection: 53BP1-RIF1-Shieldin protect ends and favour classical NHEJ (Ku70/80, DNA-PKcs, Artemis, XRCC4-LIG4), dominant in G1 and used by radiation-damaged cells; BRCA1-CtIP promote resection in S/G2, RPA coats the single strand, and PALB2-BRCA2 load RAD51 to form a filament that invades the sister chromatid for accurate copying. Polymerase-theta-mediated end joining (TMEJ) is the backup that leaves microhomology-flanked deletions, the scar of HR deficiency. BRCA1/2, PALB2, RAD51C/D and ATM loss define HRD; such tumours cannot repair replication-associated breaks caused by PARP trapping, platinum crosslinks or topoisomerase poisons, and depend on POLQ and ATR. Reversion mutations, 53BP1/Shieldin loss (restoring resection in BRCA1-null cells), and fork protection restore HR under PARP-inhibitor pressure. DNA-PK inhibitors radiosensitise; ATM loss sensitises to ATR inhibition.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Homologous_recombination","links":[{"label":"Chang, Pannunzio, Adachi & Lieber, Non-homologous DNA end joining and alternative pathways to double-strand break repair (Nat Rev Mol Cell Biol 2017)","url":"https://doi.org/10.1038/nrm.2017.48"},{"label":"Lord & Ashworth, PARP inhibitors: synthetic lethality in the clinic (Science 2017)","url":"https://doi.org/10.1126/science.aam7344"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["ovarian","tnbc","prostate","pancreatic","gallbladder"],"sections":[],"technologies":["parp-inhibitor","hrd-testing","platinum","imrt-igrt","sbrt","germline-testing","synthetic-lethality-approaches"],"targets":["brca","parp","atr","wee1"],"drugs":["olaparib","niraparib","rucaparib","talazoparib","carboplatin","cisplatin"],"companies":[],"institutions":[],"pathways":["ddr","replication-stress","base-excision-repair-parp","synthetic-lethality-map","cgas-sting"],"terms":["hrd","synthetic-lethality","mutational-signature","germline-vs-somatic","hereditary-cancer-syndromes"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-lord-science","paper-chang-nat-rev-mol-cell-biol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"dsb","label":"Double-strand break","x":45,"y":8},{"id":"mrn","label":"MRN → ATM → CHK2","x":45,"y":28,"targetId":"atr"},{"id":"53bp1","label":"53BP1-Shieldin (protect)","x":78,"y":28},{"id":"nhej","label":"NHEJ: Ku, DNA-PKcs, LIG4","x":78,"y":55},{"id":"res","label":"Resection: BRCA1-CtIP","x":15,"y":40,"targetId":"brca"},{"id":"rad51","label":"PALB2-BRCA2 → RAD51","x":15,"y":65,"targetId":"brca"},{"id":"hr","label":"Accurate HR (sister copy)","x":15,"y":90},{"id":"polq","label":"POLQ end joining (backup)","x":45,"y":90},{"id":"join","label":"Error-prone joining","x":78,"y":82},{"id":"parp","label":"PARP trapping → breaks","x":12,"y":8,"targetId":"parp"}],"edges":[{"from":"parp","to":"dsb","type":"activates"},{"from":"dsb","to":"mrn","type":"activates"},{"from":"mrn","to":"res","type":"activates"},{"from":"53bp1","to":"res","type":"inhibits"},{"from":"dsb","to":"nhej","type":"activates"},{"from":"53bp1","to":"nhej","type":"activates"},{"from":"res","to":"rad51","type":"activates"},{"from":"rad51","to":"hr","type":"activates"},{"from":"res","to":"polq","type":"activates"},{"from":"nhej","to":"join","type":"activates"},{"from":"polq","to":"join","type":"activates"}],"interventions":["PARP inhibitors (olaparib, niraparib, rucaparib, talazoparib) in BRCA/HRD ovarian, breast, prostate, pancreatic cancer","Platinum chemotherapy and radiation add breaks HR-deficient cells cannot fix","POLQ inhibitors (novobiocin analogues), ATR inhibitors after ATM loss, DNA-PK inhibitors with radiotherapy (trials)","HRD testing (genomic scars, BRCA sequencing) and RAD51 foci assays select patients; reversion mutations detected in ctDNA flag resistance"]},{"id":"oncogene-activation-two-hit","kind":"pathway","name":"Drivers, passengers & the two-hit model","aka":[],"tldr":"Of 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.","summary":"Vogelstein's landscape: ~140 driver genes, each tumour carrying 2-8 driver events in ~12 pathways, on a background of tens to thousands of passengers. Oncogene activation is dominant and recurrent at hotspots: point mutation (KRAS G12, BRAF V600E, PIK3CA H1047R), amplification (HER2, MYC, MDM2), fusion (BCR-ABL, ALK, RET, NTRK, EWSR1-FLI1), promoter mutation (TERT) or ecDNA. Tumour suppressor loss follows Knudson's two hits, one inherited in hereditary syndromes (RB1, BRCA1/2, APC, MLH1, TP53), the second by deletion, LOH, mutation or methylation; some are haploinsufficient (PTEN). Gatekeepers (APC, RB1) control proliferation directly; caretakers (BRCA, MMR) guard the genome so their loss accelerates all other hits. Oncogene addiction, the dependence of a tumour on its driver, is why kinase inhibitors work; passenger load creates neoantigens and, occasionally, collateral vulnerabilities (MTAP deletion next to CDKN2A → PRMT5 dependence).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Knudson_hypothesis","links":[{"label":"Vogelstein et al., Cancer genome landscapes (Science 2013)","url":"https://doi.org/10.1126/science.1235122"},{"label":"Kinzler & Vogelstein, Gatekeepers and caretakers (Nature 1997)","url":"https://doi.org/10.1038/386761a0"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["retinoblastoma","colorectal","cml"],"sections":[],"technologies":["cgp","germline-testing","kinase-inhibitors","synthetic-lethality-approaches","crispr-screens"],"targets":["kras","braf","her2","egfr","alk","tp53","brca","prmt5-mtap","ewsr1-fli1","bcr-abl"],"drugs":["imatinib","osimertinib","sotorasib","trastuzumab","olaparib"],"companies":[],"institutions":[],"pathways":["ras-mapk","p53-cell-cycle","clonal-evolution","mutagenesis-signatures"],"terms":["oncogene-addiction","gene-fusion","hereditary-cancer-syndromes","germline-vs-somatic","synthetic-lethality","vus"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-kinzler-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"proto","label":"Proto-oncogene","x":15,"y":15},{"id":"act","label":"Mutation, amp, fusion","x":15,"y":45,"targetId":"kras"},{"id":"onc","label":"Oncogene ON (1 hit)","x":15,"y":78},{"id":"ts","label":"Tumour suppressor","x":55,"y":15,"targetId":"tp53"},{"id":"hit1","label":"1st hit (germline/somatic)","x":55,"y":45},{"id":"hit2","label":"2nd hit: LOH, methylation","x":85,"y":45},{"id":"loss","label":"Brake lost (2 hits)","x":55,"y":78},{"id":"pass","label":"Passengers, neoantigens","x":85,"y":15},{"id":"growth","label":"Clonal expansion","x":55,"y":96}],"edges":[{"from":"proto","to":"act","type":"activates"},{"from":"act","to":"onc","type":"activates"},{"from":"onc","to":"growth","type":"activates"},{"from":"ts","to":"growth","type":"inhibits"},{"from":"hit1","to":"ts","type":"inhibits"},{"from":"hit2","to":"ts","type":"inhibits"},{"from":"hit1","to":"loss","type":"activates"},{"from":"hit2","to":"loss","type":"activates"},{"from":"loss","to":"growth","type":"activates"},{"from":"pass","to":"growth","type":"activates"}],"interventions":["Oncogene addiction is the basis of every targeted kinase inhibitor and of HER2 antibodies","Suppressor loss cannot be 'inhibited', so it is exploited indirectly: synthetic lethality (BRCA-PARP, MTAP-PRMT5), CDK4/6 for RB-intact, MDM2 for TP53-wild-type","Germline first hits drive surveillance and risk-reducing surgery in hereditary syndromes","Comprehensive genomic profiling separates drivers from passengers at diagnosis"]},{"id":"drug-efflux-pumps","kind":"pathway","name":"Drug efflux pumps (ABC transporters)","aka":[],"tldr":"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.","summary":"ATP-binding cassette transporters, chiefly ABCB1 (P-glycoprotein/MDR1), ABCG2 (BCRP) and ABCC1 (MRP1), hydrolyse ATP to export hydrophobic drugs: taxanes, vinca alkaloids, anthracyclines, MMAE, DM1, several TKIs (ABCB1); SN-38, topotecan, mitoxantrone, methotrexate (ABCG2). They are constitutive in intestine, liver, kidney, the blood-brain barrier and stem cells (including leukaemia and cancer stem cells, the 'side population'), and are induced by EMT transcription factors, hypoxia, NRF2 (KEAP1-mutant tumours), and by drug exposure itself (ABCB1 amplification and promoter fusions in ovarian cancer after taxanes). Three generations of P-gp inhibitors (verapamil, valspodar, tariquidar) failed in phase 3 through pharmacokinetic interactions and lack of benefit. Modern responses: choose payloads that are poor substrates (DXd and exatecan are less affected than SN-38 and MMAE; PBD dimers; radionuclides are pump-independent), use ADCs to raise intracellular delivery, and exploit collateral sensitivity (pump-expressing cells are more sensitive to some agents). Efflux at the BBB limits CNS activity of most drugs; brain-penetrant TKIs are engineered to evade P-gp.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/P-glycoprotein","links":[{"label":"Robey et al., Revisiting the role of ABC transporters in multidrug-resistant cancer (Nat Rev Cancer 2018)","url":"https://doi.org/10.1038/s41568-018-0005-8"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["adc","radioligand-therapy","cytotoxic-chemotherapy","topoisomerase-inhibitors"],"targets":["trop2","her2","kras","alk"],"drugs":["trastuzumab-deruxtecan","datopotamab-deruxtecan","sacituzumab-tirumotecan","sacituzumab-govitecan","enfortumab-vedotin","paclitaxel","doxorubicin","lorlatinib"],"companies":[],"institutions":[],"pathways":["emt","keap1-nrf2","blood-brain-barrier-metastasis","resistance-routes-map","cancer-stem-cells-plasticity"],"terms":["efflux-pump","payload","dxd","exatecan","sn-38","mmae","bystander-effect","adc-sequencing","resistance"],"trials":[],"people":[],"bottlenecks":["b-resistance"],"keyPapers":["paper-robey-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"drug","label":"Chemo / payload enters","x":12,"y":30},{"id":"abcb1","label":"ABCB1 (P-gp)","x":45,"y":15},{"id":"abcg2","label":"ABCG2 (BCRP), ABCC1","x":45,"y":48},{"id":"out","label":"Drug exported (ATP)","x":78,"y":30},{"id":"conc","label":"Sub-lethal intracellular dose","x":78,"y":62},{"id":"ind","label":"EMT, hypoxia, NRF2 induce","x":12,"y":62,"targetId":"kras"},{"id":"bbb","label":"BBB, stem cells, marrow","x":12,"y":90},{"id":"payload","label":"Non-substrate payloads (DXd)","x":45,"y":82,"targetId":"trop2"},{"id":"kill","label":"Cell death","x":78,"y":92}],"edges":[{"from":"drug","to":"abcb1","type":"activates"},{"from":"drug","to":"abcg2","type":"activates"},{"from":"abcb1","to":"out","type":"activates"},{"from":"abcg2","to":"out","type":"activates"},{"from":"out","to":"conc","type":"activates"},{"from":"conc","to":"kill","type":"inhibits"},{"from":"ind","to":"abcb1","type":"activates"},{"from":"ind","to":"abcg2","type":"activates"},{"from":"bbb","to":"abcb1","type":"activates"},{"from":"payload","to":"abcb1","type":"inhibits"},{"from":"payload","to":"kill","type":"activates"}],"interventions":["Payload selection: DXd and exatecan (T-DXd, Dato-DXd, sac-TMT) retain activity where MMAE and SN-38 are pumped out; PBD dimers and radionuclides are pump-independent","Brain-penetrant TKIs engineered to evade P-gp (lorlatinib, tucatinib, osimertinib)","P-gp inhibitors (valspodar, tariquidar, zosuquidar) failed in phase 3; a museum exhibit","Collateral sensitivity and MDR-selective compounds are experimental"]},{"id":"drug-tolerant-persisters","kind":"pathway","name":"Drug-tolerant persister cells","aka":[],"tldr":"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. 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.","summary":"First described by Sharma et al. (2010) in EGFR-mutant lung cancer cells surviving erlotinib, persisters are a reversible, largely non-genetic state resembling bacterial persistence and embryonic diapause: slow-cycling, with chromatin changes (KDM5A-dependent histone demethylation, H3K27me3 gain), activation of IGF1R, YAP/TAZ, NF-κB, Notch and AXL, autophagy, altered metabolism (reliance on fatty acid oxidation, low glutathione), reduced apoptotic priming, upregulated ABC transporters, and a mesenchymal-like, ALDH-high phenotype. Their lipid metabolism creates dependence on GPX4, so ferroptosis inducers kill them selectively in models. During persistence, downregulated repair and APOBEC activity increase mutagenesis, so genuine resistance mutations (EGFR T790M, C797S) often arise from persisters ('bet hedging' then 'commitment'). MRD detected by ctDNA after targeted therapy or in adjuvant settings partly reflects this population; senescence-like persisters share SASP features. Strategies: drug holidays and intermittent dosing to prevent commitment, GPX4/ferroptosis inducers, BCL-XL/MCL-1 inhibitors to remove the survival buffer, KDM5 or EZH2 inhibitors to block the epigenetic switch, and upfront combinations that hit the persister programme (osimertinib + chemotherapy in FLAURA2, + amivantamab in MARIPOSA).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Persister_cell","links":[{"label":"Sharma et al., A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations (Cell 2010)","url":"https://doi.org/10.1016/j.cell.2010.02.027"},{"label":"Shen, Vagner & Robert, Persistent cancer cells: the deadly survivors (Cell 2020)","url":"https://doi.org/10.1016/j.cell.2020.10.027"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["mrd-testing","continuous-ctdna-monitoring","epigenetic-drugs","senescence-targeting","kinase-inhibitors"],"targets":["egfr","ezh2","bcl2","kras","braf"],"drugs":["osimertinib","amivantamab","lazertinib","venetoclax","signatera"],"companies":[],"institutions":[],"pathways":["ferroptosis-cell-death","epigenetic-reprogramming","autophagy","senescence","clonal-evolution","cancer-stem-cells-plasticity","resistance-routes-map"],"terms":["mrd","ctdna","resistance","nonmutational-epigenetic-reprogramming","c797s"],"trials":[],"people":[],"bottlenecks":["b-resistance","b-dormancy-mrd"],"keyPapers":["paper-sharma-cell","paper-shen-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"drug","label":"Targeted drug or chemo","x":12,"y":15,"targetId":"egfr"},{"id":"bulk","label":"Bulk tumour dies","x":45,"y":15},{"id":"pers","label":"Persister: slow-cycling, reversible","x":45,"y":45},{"id":"epi","label":"KDM5A, H3K27me3, YAP, NF-κB","x":12,"y":48,"targetId":"ezh2"},{"id":"meta","label":"FAO, low GSH → GPX4 dependence","x":78,"y":30},{"id":"abc","label":"Efflux, autophagy, BCL-XL","x":78,"y":62,"targetId":"bcl2"},{"id":"mut","label":"APOBEC → resistance mutation","x":45,"y":78},{"id":"relapse","label":"Relapse (MRD → clinical)","x":78,"y":92},{"id":"ferro","label":"Ferroptosis inducers","x":92,"y":8},{"id":"combo","label":"Upfront combinations, holidays","x":12,"y":85}],"edges":[{"from":"drug","to":"bulk","type":"activates"},{"from":"drug","to":"pers","type":"activates"},{"from":"epi","to":"pers","type":"activates"},{"from":"pers","to":"meta","type":"activates"},{"from":"pers","to":"abc","type":"activates"},{"from":"pers","to":"mut","type":"activates"},{"from":"mut","to":"relapse","type":"activates"},{"from":"pers","to":"relapse","type":"activates"},{"from":"ferro","to":"meta","type":"inhibits"},{"from":"combo","to":"pers","type":"inhibits"}],"interventions":["Upfront combinations that pre-empt persisters: osimertinib + chemotherapy (FLAURA2), amivantamab + lazertinib (MARIPOSA), BRAF + MEK + anti-PD-1","GPX4 and ferroptosis inducers, BCL-XL/MCL-1 inhibitors, KDM5 and EZH2 inhibitors (preclinical to phase 1)","MRD-guided treatment: ctDNA clearance to de-escalate, persistence to intensify or switch","Intermittent or adaptive dosing to delay commitment to resistance (trials in melanoma and prostate cancer)"]},{"id":"endometrial-cancer-signalling","kind":"pathway","name":"Endometrial cancer (KEGG map)","aka":["KEGG hsa05213","Endometrial cancer"],"tldr":"KEGG's endometrial cancer map shows oestrogen-related type I tumours with PTEN loss, KRAS and beta-catenin mutations and faulty mismatch repair, and type II tumours with TP53 mutation and HER2 amplification. Immunotherapy for mismatch-repair-deficient tumours and HER2-directed therapy follow directly from this split.","summary":"The KEGG endometrial cancer map (hsa05213) separates the two classical types. Type I (endometrioid) carcinoma follows unopposed oestrogen exposure and endometrial hyperplasia and shows loss of PTEN, the most frequent event, releasing PI3K to AKT signalling (with PIK3CA and PIK3R1 mutations adding to it), activating KRAS mutations feeding RAF to MEK to ERK, stabilising CTNNB1 (beta-catenin) mutations that escape the APC/AXIN/GSK3B destruction complex and drive TCF/LEF and MYC transcription, and defects in DNA mismatch repair (MLH1 promoter methylation, or germline Lynch syndrome mutations) that produce microsatellite instability. Type II (serous and other non-endometrioid) carcinoma arises in atrophic endometrium, lacks hormone receptors and shows TP53 mutation, aneuploidy and ERBB2 (HER2) amplification with downstream EGFR/ERBB2 to RAS and PI3K signalling. Crosbie et al., Lancet, 2022 (doi:10.1016/S0140-6736(22)00323-3) update this into the four molecular classes now used clinically: POLE ultramutated, mismatch repair deficient, p53 abnormal, and no specific molecular profile, which predict outcome and guide adjuvant therapy.\n\nWhat drugs do about it: PD-1 antibodies (dostarlimab, pembrolizumab, durvalumab) added to carboplatin and paclitaxel are now first-line for advanced or recurrent disease, with the largest benefit in mismatch-repair-deficient tumours; pembrolizumab with lenvatinib treats mismatch-repair-proficient tumours after chemotherapy; trastuzumab and trastuzumab deruxtecan target HER2-positive serous tumours; and hormonal therapy (progestins, aromatase inhibitors such as letrozole, sometimes with the mTOR inhibitor everolimus) exploits the oestrogen dependence of low-grade type I disease.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05213","url":"https://www.kegg.jp/pathway/hsa05213"},{"label":"Review: Endometrial cancer (Lancet 2022)","url":"https://doi.org/10.1016/S0140-6736(22)00323-3"}],"tags":[],"related":["mismatch-repair-msi","pi3k-akt-mtor"],"cancers":["endometrial"],"sections":[],"technologies":[],"targets":["estrogen-receptor","pik3ca","kras","her2","tp53"],"drugs":["dostarlimab","pembrolizumab","durvalumab","lenvatinib","trastuzumab","trastuzumab-deruxtecan","letrozole","everolimus"],"companies":[],"institutions":[],"pathways":["pi3k-akt-mtor","ras-mapk","wnt","mismatch-repair-msi","p53-cell-cycle","pd1-checkpoint","er-signaling"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-crosbie-lancet"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Two different break-ins. In type I the alarm system (PTEN, mismatch repair) is disabled and oestrogen keeps the doors open; the intruders leave so many fingerprints (mutations) that the immune police can recognise them once PD-1 antibodies unmask them. In type II a single master key (TP53 loss, HER2 amplification) does the damage, so HER2-targeted drugs are the tool.","nodes":[{"id":"oestrogen","label":"Unopposed oestrogen, hyperplasia","x":20,"y":5,"targetId":"estrogen-receptor"},{"id":"pten","label":"PTEN (lost)","x":40,"y":22},{"id":"pi3k","label":"PI3K (PIK3CA) / AKT","x":20,"y":42,"targetId":"pik3ca"},{"id":"kras","label":"KRAS to RAF / MEK / ERK","x":50,"y":42,"targetId":"kras"},{"id":"wnt","label":"Beta-catenin (CTNNB1) to TCF/LEF","x":78,"y":42},{"id":"mmr","label":"Mismatch repair loss (MLH1), MSI","x":78,"y":12},{"id":"her2","label":"ERBB2 (HER2) amplified (type II)","x":92,"y":65,"targetId":"her2"},{"id":"p53","label":"TP53 mutated (type II)","x":65,"y":78,"targetId":"tp53"},{"id":"out","label":"Proliferation, survival","x":35,"y":90}],"edges":[{"from":"oestrogen","to":"out","type":"activates"},{"from":"pten","to":"pi3k","type":"inhibits"},{"from":"pi3k","to":"out","type":"activates"},{"from":"kras","to":"out","type":"activates"},{"from":"wnt","to":"out","type":"activates"},{"from":"mmr","to":"kras","type":"activates"},{"from":"her2","to":"pi3k","type":"activates"},{"from":"her2","to":"kras","type":"activates"},{"from":"p53","to":"out","type":"inhibits"}],"interventions":["PD-1 antibodies dostarlimab, pembrolizumab or durvalumab with carboplatin and paclitaxel for advanced or recurrent disease, especially mismatch-repair-deficient tumours","Pembrolizumab plus lenvatinib for mismatch-repair-proficient disease after platinum chemotherapy","HER2-directed therapy (trastuzumab with chemotherapy, trastuzumab deruxtecan) for HER2-positive serous carcinoma","Hormonal therapy (progestins, letrozole, with or without everolimus) for low-grade oestrogen-receptor-positive disease","Molecular classification (POLE, mismatch repair, p53) to de-escalate or intensify adjuvant treatment after surgery"]},{"id":"epigenetic-reprogramming","kind":"pathway","name":"Epigenetic reprogramming","aka":[],"tldr":"Cancer 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.","summary":"DNA methylation (DNMT1/3A, TET2, IDH-driven hypermethylation), histone marks (EZH2/H3K27me3, KMT2 family, H3K27M in glioma, NSD2 in myeloma), chromatin readers (BET proteins), and remodelling (SWI/SNF) are all mutated or hijacked. 'Non-mutational epigenetic reprogramming' is a 2022 hallmark: drug-tolerant persister states arise without new mutations. Approved: azacitidine/decitabine (MDS/AML), HDAC inhibitors (T-cell lymphoma), EZH2 (tazemetostat, withdrawn 2026), IDH inhibitors, menin inhibitors (KMT2A/NPM1 leukaemia). Solid tumour activity remains modest; combinations to re-express antigens or hormone receptors are the current bet. Methylation classifiers diagnose brain tumours and underlie cfDNA cancer detection.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Cancer_epigenetics","links":[{"label":"Hanahan, Hallmarks of Cancer: New Dimensions (Cancer Discovery 2022)","url":"https://doi.org/10.1158/2159-8290.CD-21-1059"},{"label":"Baylin & Jones, Epigenetic determinants of cancer (Cold Spring Harbor Perspectives 2016)","url":"https://doi.org/10.1101/cshperspect.a019505"}],"tags":["mechanism"],"related":["epigenetic-progenitor-theory"],"cancers":["colorectal","non-hodgkin-lymphoma"],"sections":[],"technologies":["epigenetic-drugs","methylation-profiling","mced"],"targets":["idh","ezh2","menin","kmt2a","npm1","tet2","brd4"],"drugs":["azacitidine","revumenib","ziftomenib","vorasidenib"],"companies":[],"institutions":["johns-hopkins","dkfz","dana-farber","mskcc"],"pathways":["swi-snf-chromatin","cancer-stem-cells-plasticity","er-signaling"],"terms":["nonmutational-epigenetic-reprogramming","mgmt","h3k27m"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-baylin-cold-spring-harb-perspect-biol"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Baylin (Johns Hopkins, DNA methylation); Pfister (DKFZ, methylation classification); Armstrong (Dana-Farber, menin/KMT2A); Allis legacy (Rockefeller, histone code)."],"analogy":"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.","nodes":[{"id":"dnmt","label":"DNA methylation (DNMT, TET2, IDH/2-HG)","x":15,"y":25,"targetId":"idh"},{"id":"hist","label":"Histone marks (EZH2, KMT2A, H3K27M)","x":15,"y":55,"targetId":"ezh2"},{"id":"reader","label":"Readers (BET) & remodellers (SWI/SNF)","x":15,"y":85},{"id":"chrom","label":"Chromatin state","x":50,"y":55},{"id":"sil","label":"Silenced tumour suppressors, antigens","x":82,"y":30},{"id":"act","label":"Active oncogenic programmes, persister states","x":82,"y":75},{"id":"menin","label":"Menin-KMT2A scaffold","x":50,"y":20,"targetId":"menin"}],"edges":[{"from":"dnmt","to":"chrom","type":"activates"},{"from":"hist","to":"chrom","type":"activates"},{"from":"reader","to":"chrom","type":"activates"},{"from":"chrom","to":"sil","type":"activates"},{"from":"chrom","to":"act","type":"activates"},{"from":"menin","to":"hist","type":"activates"}],"interventions":["Hypomethylating agents (azacitidine, decitabine) with venetoclax in AML","Menin inhibitors (revumenib, ziftomenib) in KMT2A/NPM1 leukaemia","IDH inhibitors reverse 2-HG hypermethylation","HDAC, BET, LSD1 inhibitors mostly in trials; epigenetic priming for immunotherapy"]},{"id":"emt","kind":"pathway","name":"Epithelial-mesenchymal transition & drug efflux","aka":[],"tldr":"How 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.","summary":"TGF-β, Wnt, Notch, hypoxia, and inflammation induce ZEB1/2, SNAIL, SLUG, TWIST, which repress E-cadherin and epithelial genes and induce vimentin, N-cadherin, and matrix proteases. EMT confers invasiveness, stemness, resistance to apoptosis, upregulated ABC efflux transporters (ABCB1/P-gp, ABCG2), and immune exclusion. Partial EMT states dominate in tumours. Clinically: the mesenchymal/claudin-low TNBC subtype, sarcomatoid carcinomas, and TKI resistance (EGFR NSCLC). No approved EMT drug; strategies include TGF-β blockade (bintrafusp failed), efflux-agnostic payloads, and targeting mesenchymal-state dependencies (GPX4, ferroptosis).","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Epithelial–mesenchymal_transition","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Epithelial–mesenchymal_transition"}],"tags":[],"related":["theories-of-cancer","cancer-stem-cell-theory"],"cancers":["tnbc","nsclc"],"sections":[],"technologies":["adc","radioligand-therapy","single-cell-spatial"],"targets":["trop2"],"drugs":[],"companies":[],"institutions":[],"pathways":[],"terms":["efflux-pump","resistance"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-lamouille-emt-molecular-mechanisms-nrmcb-2014"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"tgf","label":"TGF-β, Wnt, Notch, hypoxia","x":50,"y":5},{"id":"tf","label":"ZEB1/2, SNAIL, TWIST","x":50,"y":25},{"id":"ecad","label":"E-cadherin, claudins","x":20,"y":48},{"id":"vim","label":"Vimentin, N-cadherin, MMPs","x":80,"y":48},{"id":"abc","label":"ABCB1 / ABCG2 efflux","x":20,"y":72},{"id":"stem","label":"Stemness, apoptosis resistance","x":50,"y":72},{"id":"imm","label":"Immune exclusion","x":80,"y":72},{"id":"out","label":"Invasion, metastasis, drug resistance","x":50,"y":95}],"edges":[{"from":"tgf","to":"tf","type":"activates"},{"from":"tf","to":"ecad","type":"inhibits"},{"from":"tf","to":"vim","type":"activates"},{"from":"tf","to":"abc","type":"activates"},{"from":"tf","to":"stem","type":"activates"},{"from":"tf","to":"imm","type":"activates"},{"from":"vim","to":"out","type":"activates"},{"from":"abc","to":"out","type":"activates"},{"from":"stem","to":"out","type":"activates"}],"interventions":["No approved direct EMT inhibitor","Payloads with low efflux susceptibility (sac-TMT's belotecan derivative claims this) and radiation (efflux-independent)","TGF-β pathway blockade (mostly failed so far)","Ferroptosis inducers for mesenchymal-state cells (preclinical)","Immune approaches to overcome exclusion (STING, radiation)"]},{"id":"extrinsic-apoptosis-death-receptors","kind":"pathway","name":"Extrinsic apoptosis (death receptors)","aka":[],"tldr":"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.","summary":"FASL, TRAIL and TNF bind death receptors (FAS, DR4/DR5, TNFR1), recruiting FADD and pro-caspase-8 into the DISC; active caspase-8 cleaves caspase-3/7 directly (type I cells) or cleaves BID to tBID to engage the mitochondrial pathway (type II). c-FLIP competes with caspase-8; XIAP restrains caspase-3/9; decoy receptors (DcR1-3) soak up ligand. Cytotoxic T and NK cells also deliver granzyme B through perforin pores, cleaving BID and caspases independently of receptors. Tumours lose FAS or CASP8 (head and neck, HPV-negative), overexpress c-FLIP and XIAP, and express FASL to kill infiltrating lymphocytes ('counterattack'). TRAIL agonists were safe but inactive; SMAC mimetics (IAP antagonists) and second-generation DR5 agonists are in trials; caspase-8 loss also switches death toward necroptosis.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Death_receptor","links":[{"label":"Ashkenazi, Targeting the extrinsic apoptotic pathway in cancer (J Clin Invest 2015)","url":"https://doi.org/10.1172/JCI80420"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["car-t","t-cell-engager","checkpoint-inhibitor","bh3-profiling"],"targets":["bcl2","cd3","pd1"],"drugs":["venetoclax"],"companies":[],"institutions":[],"pathways":["apoptosis-bcl2","pd1-checkpoint","nk-cell-recognition"],"terms":["immunogenic-cell-death","adcc"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-ashkenazi-j-clin-invest"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"ctl","label":"CTL / NK cell","x":15,"y":12,"targetId":"cd3"},{"id":"lig","label":"FASL, TRAIL","x":45,"y":12},{"id":"dr","label":"FAS, DR4/DR5","x":45,"y":38},{"id":"dec","label":"Decoy receptors, c-FLIP","x":82,"y":25},{"id":"disc","label":"DISC: FADD, caspase-8","x":45,"y":62},{"id":"gzm","label":"Perforin / granzyme B","x":15,"y":40},{"id":"bid","label":"tBID → mitochondria","x":82,"y":62,"targetId":"bcl2"},{"id":"iap","label":"XIAP (IAPs)","x":82,"y":88},{"id":"casp","label":"Caspase-3/7 → death","x":45,"y":88}],"edges":[{"from":"ctl","to":"lig","type":"activates"},{"from":"ctl","to":"gzm","type":"activates"},{"from":"lig","to":"dr","type":"activates"},{"from":"dec","to":"dr","type":"inhibits"},{"from":"dr","to":"disc","type":"activates"},{"from":"dec","to":"disc","type":"inhibits"},{"from":"disc","to":"casp","type":"activates"},{"from":"disc","to":"bid","type":"activates"},{"from":"bid","to":"casp","type":"activates"},{"from":"gzm","to":"casp","type":"activates"},{"from":"iap","to":"casp","type":"inhibits"}],"interventions":["Checkpoint inhibitors, engagers and CAR-T all ultimately act through this wire, so caspase-8 or FAS loss confers immune resistance","SMAC mimetics (IAP antagonists) lower the threshold; birinapant, xevinapant tested with chemoradiation","DR5 agonist antibodies and TRAIL-receptor engagers, largely inactive so far","BH3 mimetics engage the mitochondrial arm downstream of tBID"]},{"id":"ferroptosis-cell-death","kind":"pathway","name":"Ferroptosis & regulated cell death","aka":[],"tldr":"Cells 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.","summary":"Ferroptosis is iron-dependent lipid peroxidation restrained by GPX4 (using glutathione from the cystine transporter SLC7A11/xCT) and by FSP1. Therapy-persistent, mesenchymal, and dedifferentiated cancer cells depend on GPX4; radiation and some immunotherapies act partly via ferroptosis. Necroptosis (RIPK3/MLKL) and pyroptosis (gasdermins) are inflammatory death modes that can be immunogenic. No ferroptosis inducer is approved; GPX4 inhibitors lack drug-like properties, so xCT inhibition, cyst(e)inase, and sulfasalazine repurposing are the clinical routes.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Ferroptosis","links":[{"label":"Stockwell, Ferroptosis turns 10 (Cell 2022)","url":"https://doi.org/10.1016/j.cell.2022.06.003"},{"label":"Viswanathan et al., Dependency of a therapy-resistant state on a lipid peroxidase pathway (Nature 2017)","url":"https://doi.org/10.1038/nature23007"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":[],"targets":["bcl2"],"drugs":["venetoclax"],"companies":[],"institutions":["mskcc","stanford","broad-institute"],"pathways":["apoptosis-bcl2","emt","cancer-metabolism"],"terms":["resisting-cell-death","resistance"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-stockwell-cell","paper-viswanathan-nature"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Stockwell (Columbia, coined ferroptosis); Dixon (Stanford); Schreiber (Broad) on persister-cell GPX4 dependence; Jiang (MSK) on ferroptosis in immunotherapy."],"analogy":"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.","nodes":[{"id":"cys","label":"Cystine import (SLC7A11)","x":15,"y":30},{"id":"gsh","label":"Glutathione","x":40,"y":30},{"id":"gpx4","label":"GPX4","x":65,"y":30},{"id":"lpo","label":"Lipid peroxidation","x":65,"y":70},{"id":"iron","label":"Labile iron (Fenton)","x":35,"y":70},{"id":"ferro","label":"Ferroptosis","x":90,"y":70},{"id":"mes","label":"Mesenchymal / persister state","x":90,"y":30},{"id":"apo","label":"Apoptosis (BCL-2 family)","x":15,"y":90,"targetId":"bcl2"}],"edges":[{"from":"cys","to":"gsh","type":"activates"},{"from":"gsh","to":"gpx4","type":"activates"},{"from":"gpx4","to":"lpo","type":"inhibits"},{"from":"iron","to":"lpo","type":"activates"},{"from":"lpo","to":"ferro","type":"activates"},{"from":"mes","to":"gpx4","type":"activates"}],"interventions":["xCT inhibitors, cyst(e)inase, sulfasalazine repurposing (early trials)","Radiotherapy and IFN-γ from T cells induce lipid peroxidation","BH3 mimetics (venetoclax) exploit apoptosis; MCL-1 inhibitors in development","Persister-cell targeting after EGFR/ALK inhibitors (preclinical)"]},{"id":"fgfr-signalling","kind":"pathway","name":"FGF / FGFR signalling","aka":[],"tldr":"Fibroblast growth factor receptors are growth antennas on the cell surface. Bladder cancer mutates FGFR3, bile duct cancer fuses FGFR2 to other genes, and stomach cancer overproduces FGFR2b; each has its own drug, and each brings a tell-tale side effect (high phosphate) because the same receptors control phosphate in the kidney.","summary":"Twenty-two FGF ligands bind four receptor tyrosine kinases (FGFR1-4) with heparan sulphate or, for the endocrine FGFs, with Klotho co-receptors. Dimerised receptors phosphorylate FRS2, which recruits GRB2/SOS to activate RAS-MAPK and GAB1 to activate PI3K-AKT; PLCγ and STAT branches add to the output. Oncogenic alterations: FGFR3 point mutations (S249C, Y373C) and FGFR3-TACC3 fusions in urothelial carcinoma (about 20% of advanced disease, more in upper tract); FGFR2 fusions and rearrangements in 10-15% of intrahepatic cholangiocarcinoma; FGFR2 amplification or FGFR2b overexpression in gastric cancer; FGFR1 amplification in squamous lung and luminal breast cancer. Erdafitinib (pan-FGFR) is approved for FGFR3-altered urothelial cancer after the THOR trial; pemigatinib and futibatinib (covalent, active against gatekeeper mutations) for FGFR2-fusion cholangiocarcinoma; bemarituzumab, an anti-FGFR2b antibody, is in phase 3 in gastric cancer. Hyperphosphataemia (FGF23-FGFR1 in the kidney), nail and skin changes and central serous retinopathy are class effects. Resistance arises through gatekeeper (V564) and molecular-brake mutations and through MAPK or PI3K bypass.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Fibroblast_growth_factor_receptor","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Fibroblast_growth_factor_receptor"}],"tags":[],"related":[],"cancers":["urothelial","cholangiocarcinoma","gastric"],"sections":[],"technologies":[],"targets":["fgfr2","kras","pik3ca"],"drugs":["erdafitinib","pemigatinib","futibatinib","bemarituzumab"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"Four aerials (FGFR1-4) tuned to growth-factor broadcasts. Bladder cancer bends an aerial so it hears a signal that is not there; bile duct cancer welds it to a foreign mast (fusion) that keeps it switched on. The inhibitors mute the aerial, but the same aerials manage phosphate in the kidney, so phosphate rises as the price.","nodes":[{"id":"fgf","label":"FGF ligands (+ heparan sulphate / Klotho)","x":50,"y":5},{"id":"fgfr","label":"FGFR1-4 (FGFR2 fusions, FGFR3 mutations)","x":50,"y":24,"targetId":"fgfr2"},{"id":"frs2","label":"FRS2","x":50,"y":42},{"id":"ras","label":"GRB2 / SOS → RAS → MAPK","x":22,"y":62,"targetId":"kras"},{"id":"pi3k","label":"GAB1 → PI3K → AKT","x":50,"y":62,"targetId":"pik3ca"},{"id":"plc","label":"PLCγ / STAT","x":78,"y":62},{"id":"out","label":"Proliferation, survival, angiogenesis","x":50,"y":88},{"id":"phos","label":"FGF23 → FGFR1 (kidney phosphate; drug side effect)","x":84,"y":24}],"edges":[{"from":"fgf","to":"fgfr","type":"activates"},{"from":"fgfr","to":"frs2","type":"activates"},{"from":"frs2","to":"ras","type":"activates"},{"from":"frs2","to":"pi3k","type":"activates"},{"from":"fgfr","to":"plc","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"plc","to":"out","type":"activates"},{"from":"phos","to":"fgfr","type":"activates"}],"interventions":["Erdafitinib for FGFR3-altered advanced urothelial cancer after platinum and PD-1/PD-L1 therapy (THOR)","Pemigatinib and futibatinib for FGFR2-fusion cholangiocarcinoma; futibatinib's covalent binding keeps activity against gatekeeper mutations","Bemarituzumab (anti-FGFR2b) with chemotherapy in FGFR2b-overexpressing gastric cancer (FORTITUDE-101)","Phosphate binders and diet for hyperphosphataemia; eye examinations for central serous retinopathy","Selective FGFR2 and FGFR3 inhibitors under development to widen the therapeutic window"]},{"id":"caf-activation-desmoplasia","kind":"pathway","name":"Fibroblast activation, desmoplasia & matrix stiffness","aka":[],"tldr":"Tumours 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.","summary":"Resident fibroblasts, pancreatic and hepatic stellate cells, and mesenchymal stromal cells are activated by TGF-β, PDGF, IL-1, Hedgehog ligand from tumour cells and by stiffness itself (YAP/TAZ feed-forward) into myofibroblastic CAFs (myCAF: αSMA, collagen I/III, FAP, LRRC15), inflammatory CAFs (iCAF: IL-6, LIF, CXCL12, driven by IL-1/JAK-STAT) and antigen-presenting CAFs. They deposit and crosslink collagen (LOX/LOXL2), hyaluronan and fibronectin, raising interstitial pressure (collapsing vessels, blocking drug delivery) and stiffness, which signals through integrins → FAK → RHO → YAP/TAZ to drive proliferation, EMT and chemoresistance in tumour cells; aligned fibres guide invasion; TGF-β-CAFs exclude T cells. CAFs also feed tumours (alanine, lipids, exosomes) and shield them. Depleting all fibroblasts (Shh-deleted or αSMA-ablated mice) made tumours more aggressive, and the SMO inhibitor vismodegib and hyaluronidase PEGPH20 failed in PDAC: hence a shift to reprogramming (vitamin D receptor agonists, losartan/angiotensin blockade in trials, IL-1/JAK for iCAF, LRRC15 or FAP targeting with radioligands and CAR-T) and to exploiting FAP for imaging (FAPI PET) and therapy.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Cancer-associated_fibroblast","links":[{"label":"Sahai et al., A framework for advancing our understanding of cancer-associated fibroblasts (Nat Rev Cancer 2020)","url":"https://doi.org/10.1038/s41568-019-0238-1"},{"label":"Kalluri, The biology and function of fibroblasts in cancer (Nat Rev Cancer 2016)","url":"https://doi.org/10.1038/nrc.2016.73"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["pancreatic","cholangiocarcinoma","breast-hr-positive"],"sections":[],"technologies":["fapi-pet","stroma-directed-car","mechanobiology-therapy","radioligand-therapy","single-cell-spatial"],"targets":["fap","smoothened","jak2","pd1"],"drugs":["fap-2286","vismodegib","sonidegib","ruxolitinib","avutometinib-defactinib"],"companies":[],"institutions":[],"pathways":["tumor-microenvironment","tgf-beta","hippo-yap","hedgehog","jak-stat","immune-desert-exclusion","invasion-ecm-degradation"],"terms":["desmoplasia","cancer-associated-fibroblasts","immune-exclusion","cold-vs-hot"],"trials":[],"people":[],"bottlenecks":["b-tme-immunosuppression"],"keyPapers":["paper-kalluri-nat-rev-cancer","paper-sahai-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"tum","label":"Tumour: TGF-β, PDGF, Hh, IL-1","x":12,"y":15},{"id":"fib","label":"Fibroblast / stellate cell","x":45,"y":15},{"id":"my","label":"myCAF (FAP, αSMA, collagen)","x":78,"y":15,"targetId":"fap"},{"id":"icaf","label":"iCAF (IL-6, CXCL12, LIF)","x":78,"y":48,"targetId":"jak2"},{"id":"ecm","label":"Collagen, HA, LOX crosslinks","x":45,"y":48},{"id":"stiff","label":"Stiffness → FAK → YAP/TAZ","x":12,"y":48},{"id":"press","label":"Pressure: vessels collapse","x":45,"y":82},{"id":"excl","label":"T-cell exclusion","x":78,"y":82,"targetId":"pd1"},{"id":"grow","label":"Growth, EMT, chemoresistance","x":12,"y":82},{"id":"smo","label":"Hedgehog (SMO) paradox","x":12,"y":96,"targetId":"smoothened"}],"edges":[{"from":"tum","to":"fib","type":"activates"},{"from":"fib","to":"my","type":"activates"},{"from":"fib","to":"icaf","type":"activates"},{"from":"my","to":"ecm","type":"activates"},{"from":"ecm","to":"stiff","type":"activates"},{"from":"stiff","to":"grow","type":"activates"},{"from":"stiff","to":"fib","type":"activates"},{"from":"ecm","to":"press","type":"activates"},{"from":"my","to":"excl","type":"activates"},{"from":"icaf","to":"excl","type":"activates"},{"from":"smo","to":"fib","type":"activates"}],"interventions":["FAP-targeted imaging (FAPI PET) and radioligands (FAP-2286) and FAP/LRRC15 CAR-T","Reprogramming: vitamin D analogues, losartan with chemoradiation (PDAC trials), IL-1/JAK inhibition for iCAFs","FAK inhibitors soften stroma and improve immunotherapy entry (trials); mechanobiology approaches","Vismodegib and PEGPH20 failed in PDAC: stromal depletion can accelerate disease"]},{"id":"field-cancerisation","kind":"pathway","name":"Field cancerisation","aka":[],"tldr":"Cancer 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.","summary":"Deep sequencing of normal tissue shows that by middle age, sun-exposed skin, oesophagus (NOTCH1, TP53), bronchial epithelium, endometrium, and colon crypts are colonised by clones carrying cancer-driver mutations, most of which never progress. Field effects explain second primary tumours and local recurrence after resection, and they define who benefits from interception (chemoprevention, surveillance). Clonal competition can even be protective (NOTCH1-mutant oesophageal clones outcompete TP53 clones). Understanding the transition from field to tumour is the central question for early detection and prevention.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Field_cancerization","links":[{"label":"Martincorena et al., Somatic mutant clones colonize the human esophagus with age (Science 2018)","url":"https://doi.org/10.1126/science.aau3879"}],"tags":["mechanism"],"related":["ageing-tissue-field-theory","topical-and-destructive-treatment-bcc","skin-cancer-after-organ-transplant"],"cancers":["esophageal","head-and-neck","nsclc","cervical","colorectal","bowens-disease","cutaneous-scc","basal-cell-carcinoma","skin-cancer"],"sections":[],"technologies":["chemoprevention","precancer-ablation","endoscopic-resection","hpv-vaccine"],"targets":["tp53"],"drugs":[],"companies":[],"institutions":["cruk","francis-crick","johns-hopkins"],"pathways":["clonal-evolution","epigenetic-reprogramming"],"terms":["barretts-esophagus","cin-hsil","actinic-keratosis","keratinocyte-cancer-counting"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-martincorena-science"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Jones (Wellcome Sanger) on clones in normal oesophagus and skin; Campbell/Stratton (Sanger) on somatic mutation in normal tissues; Fitzgerald (Cambridge/CRUK) on Barrett's interception; Hopkins on field cancerisation in head and neck.","Sun-damaged skin is the clearest field in medicine, and it is why skin cancer is hard to count. A person with one keratinocyte cancer usually has a field of damaged skin that will produce more, which is why treatment is often aimed at an area rather than a spot, and why a registry rule that records only the first tumour per person misses most of the disease (see `keratinocyte-cancer-counting`)."],"analogy":"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.","nodes":[{"id":"expo","label":"Carcinogen exposure, ageing","x":15,"y":40},{"id":"clones","label":"Mutant clones in normal tissue (NOTCH1, TP53)","x":42,"y":40,"targetId":"tp53"},{"id":"comp","label":"Clonal competition","x":68,"y":20},{"id":"prog","label":"Progression to dysplasia → cancer","x":68,"y":65},{"id":"second","label":"Second primaries, local recurrence","x":92,"y":65},{"id":"interc","label":"Interception: chemoprevention, ablation","x":92,"y":20}],"edges":[{"from":"expo","to":"clones","type":"activates"},{"from":"clones","to":"comp","type":"activates"},{"from":"clones","to":"prog","type":"activates"},{"from":"prog","to":"second","type":"activates"},{"from":"interc","to":"prog","type":"inhibits"}],"interventions":["Surveillance and ablation of precancer (Barrett's RFA, cervical precancer ablation, colon polypectomy)","Chemoprevention (aspirin in Lynch, tamoxifen, HPV vaccination)","Field-directed therapy after resection (e.g., 5-FU cream for actinic keratosis)","Molecular monitoring of fields (cytosponge, sputum, urine)"]},{"id":"gastric-cancer-signalling","kind":"pathway","name":"Gastric cancer (KEGG map)","aka":["KEGG hsa05226","Gastric cancer"],"tldr":"This KEGG map splits stomach cancer into two routes: the intestinal type that accumulates TP53, APC and HER2 changes step by step, and the diffuse type driven by loss of the cell glue E-cadherin plus MET or FGFR2 amplification. It matters because HER2, FGFR2, claudin 18.2 and PD-1 status now decide first-line treatment.","summary":"Gastric cancer is one of the most common cancers worldwide and is divided by the Lauren classification into intestinal and diffuse types. KEGG map hsa05226 draws the intestinal-type sequence: intestinal metaplasia carries TP53 mutation, reduced retinoic acid receptor beta and hTERT expression; gastric adenomas add APC mutation, reduced p27 (CDKN1B) and cyclin E amplification; advanced tumours show ERBB2 (HER2) amplification, reduced TGF-beta receptor I and complete loss of p27. Amplified HER2, FGFR2 and MET all feed RAS-ERK and PI3K-AKT; CDX2 over-expression alters transcription. The diffuse type is defined by loss of E-cadherin through CDH1 mutation (germline in hereditary diffuse gastric cancer), which releases beta-catenin, and by MET and FGFR2 amplification.\n\nSmyth and colleagues, The Lancet, 2020 (doi:10.1016/S0140-6736(20)31288-5) review the disease: Helicobacter pylori is the main cause, The Cancer Genome Atlas separates EBV-positive, MSI, genomically stable (diffuse, CDH1 and RHOA altered) and chromosomally unstable (intestinal, receptor amplified) subtypes, and biomarker testing for HER2, PD-L1 and MSI guides therapy.\n\nWhat drugs do about it: trastuzumab plus chemotherapy, now with pembrolizumab, is standard for HER2-positive disease, with trastuzumab deruxtecan and zanidatamab after progression; nivolumab or pembrolizumab with chemotherapy treats HER2-negative disease with PD-L1 expression; zolbetuximab treats claudin 18.2-positive tumours; ramucirumab blocks VEGFR2 in second line; the FGFR2b antibody bemarituzumab is in phase 3 for FGFR2b over-expressing tumours; MSI-high tumours respond to checkpoint inhibitors. The CDH1/E-cadherin loss that defines diffuse cancer has no direct drug.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05226","url":"https://www.kegg.jp/pathway/hsa05226"},{"label":"Review: Gastric cancer (Lancet seminar)","url":"https://doi.org/10.1016/S0140-6736(20)31288-5"}],"tags":[],"related":["rtk-activation","fgfr-signalling","emt","pd1-checkpoint"],"cancers":["gastric"],"sections":[],"technologies":[],"targets":["tp53","her2","fgfr2","met","kras","pik3ca","cldn18-2","pd1","vegf"],"drugs":["trastuzumab","pembrolizumab","trastuzumab-deruxtecan","zanidatamab","nivolumab","zolbetuximab","bemarituzumab","ramucirumab","paclitaxel"],"companies":[],"institutions":[],"pathways":["rtk-activation","fgfr-signalling","ras-mapk","pi3k-akt-mtor","p53-cell-cycle","wnt","tgf-beta","emt","pd1-checkpoint","mismatch-repair-msi"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-smyth-lancet"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Two different ways a wall fails. In the intestinal type the bricks are replaced one at a time with faulty ones (TP53, APC, HER2) until the wall bows outward. In the diffuse type the mortar (E-cadherin) dissolves, so the bricks scatter and slip between other layers. The drugs mostly aim at loudspeakers bolted to the wall (HER2, FGFR2, claudin 18.2) rather than the bricks or mortar.","nodes":[{"id":"hp","label":"H. pylori, metaplasia","x":50,"y":4},{"id":"p53","label":"TP53 / hTERT","x":15,"y":22,"targetId":"tp53"},{"id":"apc","label":"APC / p27 / cyclin E","x":40,"y":22},{"id":"her2","label":"HER2 amplified","x":65,"y":22,"targetId":"her2"},{"id":"fgfr2","label":"FGFR2 / MET amplified","x":90,"y":22,"targetId":"fgfr2"},{"id":"cdh1","label":"E-cadherin (CDH1) lost","x":15,"y":50},{"id":"bcat","label":"beta-catenin","x":15,"y":72},{"id":"ras","label":"RAS / ERK","x":60,"y":50,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT","x":85,"y":50,"targetId":"pik3ca"},{"id":"tgfb","label":"TGF-beta receptor I (reduced)","x":40,"y":50},{"id":"cldn","label":"Claudin 18.2 / PD-L1 (surface)","x":72,"y":74,"targetId":"cldn18-2"},{"id":"out","label":"Intestinal or diffuse gastric cancer","x":45,"y":94}],"edges":[{"from":"hp","to":"p53","type":"inhibits"},{"from":"hp","to":"apc","type":"activates"},{"from":"hp","to":"her2","type":"activates"},{"from":"hp","to":"fgfr2","type":"activates"},{"from":"hp","to":"cdh1","type":"inhibits"},{"from":"cdh1","to":"bcat","type":"inhibits"},{"from":"bcat","to":"out","type":"activates"},{"from":"her2","to":"ras","type":"activates"},{"from":"her2","to":"pi3k","type":"activates"},{"from":"fgfr2","to":"ras","type":"activates"},{"from":"fgfr2","to":"pi3k","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"p53","to":"out","type":"inhibits"},{"from":"apc","to":"out","type":"activates"},{"from":"tgfb","to":"out","type":"inhibits"},{"from":"cldn","to":"out","type":"activates"}],"interventions":["HER2-positive: trastuzumab plus chemotherapy and pembrolizumab first line; trastuzumab deruxtecan or zanidatamab after progression","HER2-negative, PD-L1 expressing: nivolumab or pembrolizumab with chemotherapy","Claudin 18.2-positive: zolbetuximab with chemotherapy","FGFR2b over-expressing: bemarituzumab (phase 3); MET amplified: MET inhibitors in trials","Second line: ramucirumab (VEGFR2) with paclitaxel; MSI-high: checkpoint inhibitors","E-cadherin/CDH1 loss: no direct drug yet"]},{"id":"glioma-signalling","kind":"pathway","name":"Glioma (KEGG map)","aka":["KEGG hsa05214","Glioma"],"tldr":"This KEGG map shows the two genetic roads to glioblastoma: primary tumours amplify EGFR and lose PTEN and p16, secondary tumours from lower-grade astrocytomas over-express PDGF and CDK4 and lose TP53 and RB. It explains why growth-factor and cell-cycle drugs are the main targeted options in brain tumours, and why paediatric low-grade gliomas with BRAF changes respond to MAPK inhibitors.","summary":"Gliomas are the most common primary brain tumours, more than 40 percent of central nervous system neoplasms, and include astrocytomas, oligodendrogliomas, mixed gliomas and ependymomas. KEGG map hsa05214 draws the routes to glioblastoma, the most malignant infiltrating astrocytoma. Primary glioblastoma arises de novo in older patients and typically shows EGFR amplification or mutation, p16 (CDKN2A) and p14ARF deletion and PTEN mutation or deletion. Secondary glioblastoma progresses from low-grade or anaplastic astrocytoma in younger patients and frequently shows PDGF and PDGFR over-expression, CDK4 or MDM2 amplification, TP53 mutation and RB1 loss. Both routes converge on the same effectors: RAS-ERK, PI3K-AKT (unleashed by PTEN loss), PLC-gamma/calcium signalling from EGFR and PDGFR, and loss of the p16-CDK4/6-RB and p14ARF-MDM2-p53 checkpoints.\n\nWeller and colleagues, Nature Reviews Disease Primers, 2015 (doi:10.1038/nrdp.2015.17) review the field: IDH mutation separates lower-grade and secondary tumours from IDH wild-type primary glioblastoma, MGMT promoter methylation predicts benefit from temozolomide, and BRAF V600E and KIAA1549-BRAF fusions characterise many paediatric low-grade gliomas.\n\nWhat drugs do about it: temozolomide with radiotherapy remains the standard for glioblastoma, with bevacizumab used for symptomatic control; vorasidenib treats IDH-mutant grade 2 glioma; in paediatric low-grade glioma the BRAF-MAPK arm is drugged with tovorafenib (listed by KEGG) and with dabrafenib plus trametinib for BRAF V600E tumours, and the MEK inhibitors selumetinib and mirdametinib treat NF1-related tumours. EGFR and PDGFR inhibitors have so far failed in glioblastoma, partly because of the blood-brain barrier and tumour heterogeneity.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05214","url":"https://www.kegg.jp/pathway/hsa05214"},{"label":"Review: Glioma (Nature Reviews Disease Primers)","url":"https://doi.org/10.1038/nrdp.2015.17"}],"tags":[],"related":["ras-mapk","pi3k-akt-mtor","idh-2hg"],"cancers":["glioblastoma","paediatric-low-grade-glioma"],"sections":[],"technologies":[],"targets":["egfr","pdgfra","braf","pik3ca","cdk4-6","mdm2","idh","vegf"],"drugs":["tovorafenib","dabrafenib","trametinib","selumetinib","mirdametinib","vorasidenib","temozolomide","bevacizumab","lomustine"],"companies":[],"institutions":[],"pathways":["rtk-activation","ras-mapk","pi3k-akt-mtor","p53-cell-cycle","idh-2hg","vegf-angiogenesis"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-weller-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Two roads to the same wreck. On the fast road (primary glioblastoma) the accelerator is welded down (EGFR) and the main brake line is cut (PTEN). On the slow road (secondary glioblastoma) a different accelerator is pressed (PDGF) and the brakes wear out one by one (TP53, RB). In children's low-grade gliomas a single stuck relay (BRAF) is the fault, which is why one drug can help.","nodes":[{"id":"egfr","label":"EGFR amplified (primary)","x":20,"y":6,"targetId":"egfr"},{"id":"pdgfr","label":"PDGF / PDGFR (secondary)","x":80,"y":6,"targetId":"pdgfra"},{"id":"ras","label":"RAS / RAF / ERK","x":20,"y":30,"targetId":"braf"},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":50,"y":30,"targetId":"pik3ca"},{"id":"pten","label":"PTEN (lost)","x":50,"y":52},{"id":"plc","label":"PLC-gamma / calcium","x":80,"y":30},{"id":"p16","label":"p16 / CDK4 / RB","x":20,"y":70,"targetId":"cdk4-6"},{"id":"mdm2","label":"p14ARF / MDM2 / TP53","x":80,"y":70,"targetId":"mdm2"},{"id":"idh","label":"IDH1/2 mutation (lower grade)","x":50,"y":70,"targetId":"idh"},{"id":"out","label":"Astrocytoma to glioblastoma","x":50,"y":94}],"edges":[{"from":"egfr","to":"ras","type":"activates"},{"from":"egfr","to":"pi3k","type":"activates"},{"from":"egfr","to":"plc","type":"activates"},{"from":"pdgfr","to":"ras","type":"activates"},{"from":"pdgfr","to":"pi3k","type":"activates"},{"from":"pdgfr","to":"plc","type":"activates"},{"from":"pten","to":"pi3k","type":"inhibits"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"plc","to":"out","type":"activates"},{"from":"p16","to":"out","type":"inhibits"},{"from":"mdm2","to":"out","type":"inhibits"},{"from":"idh","to":"out","type":"activates"}],"interventions":["Paediatric low-grade glioma: BRAF-MAPK inhibitors tovorafenib (listed by KEGG), dabrafenib plus trametinib for BRAF V600E, selumetinib or mirdametinib for NF1-related tumours","IDH-mutant grade 2 glioma: vorasidenib","Glioblastoma: temozolomide with radiotherapy (MGMT methylated tumours benefit most), bevacizumab for symptom control","EGFR and PDGFR inhibitors and CDK4/6 inhibitors have not yet shown benefit in glioblastoma trials"]},{"id":"glutamine-metabolism","kind":"pathway","name":"Glutamine addiction","aka":[],"tldr":"After 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.","summary":"Glutamine enters via SLC1A5 (ASCT2) and SLC38A2; glutaminase (GLS, MYC-induced via miR-23 suppression; GLS2 is p53-induced) converts it to glutamate, which GLUD1 or transaminases (GOT1/2, GPT2) convert to α-ketoglutarate for TCA anaplerosis, supporting oxidative phosphorylation and, under hypoxia or IDH mutation, reductive carboxylation to citrate for lipids. Glutamine's amide nitrogen feeds purine and pyrimidine synthesis (CAD, PPAT) and hexosamines; glutamate makes glutathione (with cysteine via SLC7A11) and aspartate (KRAS-mutant pancreatic cancer routes glutamine through GOT1 to maintain NADPH). mTORC1 senses glutamine-derived α-KG. Tumour uptake depletes glutamine in the microenvironment, impairing T-cell effector function. Drugs: the glutaminase inhibitor telaglenastat (CB-839) failed in RCC (CANTATA) and was inactive in KEAP1-mutant NSCLC (KEAPSAKE); the broad glutamine antagonist prodrug DRP-104 (sirpiglenastat) and ASCT2 inhibitors are in early trials; asparaginase (which also depletes glutamine) is standard in ALL. Glutamine PET tracers (18F-FGln) image glioma where FDG fails.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Glutaminolysis","links":[{"label":"Altman, Stine & Dang, From Krebs to clinic: glutamine metabolism to cancer therapy (Nat Rev Cancer 2016)","url":"https://doi.org/10.1038/nrc.2016.71"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["all-leukemia","rcc","pancreatic","glioblastoma"],"sections":[],"technologies":["metabolic-therapy","fdg-pet","pet"],"targets":["kras","idh","akt","pd1"],"drugs":["asparaginase","ivosidenib","vorasidenib"],"companies":[],"institutions":[],"pathways":["cancer-metabolism","myc","ras-mapk","keap1-nrf2","nutrient-competition-tme"],"terms":["warburg-effect","deregulating-cellular-energetics"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-altman-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"gln","label":"Glutamine (SLC1A5)","x":12,"y":30},{"id":"gls","label":"Glutaminase (GLS)","x":40,"y":30},{"id":"glu","label":"Glutamate","x":66,"y":30},{"id":"akg","label":"α-KG → TCA anaplerosis","x":66,"y":60},{"id":"gsh","label":"Glutathione, NADPH","x":90,"y":45},{"id":"nuc","label":"Nucleotides (N donor)","x":40,"y":8},{"id":"myc","label":"MYC, KRAS drive uptake","x":12,"y":65,"targetId":"kras"},{"id":"idh","label":"IDH → 2-HG / reductive","x":90,"y":80,"targetId":"idh"},{"id":"tcell","label":"T cells starved","x":40,"y":85,"targetId":"pd1"},{"id":"mtor","label":"mTORC1 sensing","x":12,"y":92,"targetId":"akt"}],"edges":[{"from":"gln","to":"gls","type":"activates"},{"from":"gls","to":"glu","type":"activates"},{"from":"gln","to":"nuc","type":"activates"},{"from":"glu","to":"akg","type":"activates"},{"from":"glu","to":"gsh","type":"activates"},{"from":"akg","to":"idh","type":"activates"},{"from":"myc","to":"gls","type":"activates"},{"from":"myc","to":"gln","type":"activates"},{"from":"gln","to":"tcell","type":"inhibits"},{"from":"akg","to":"mtor","type":"activates"}],"interventions":["Asparaginase (depletes asparagine and glutamine) in ALL","Glutaminase inhibitor telaglenastat: negative in RCC and NSCLC; glutamine antagonist DRP-104 and ASCT2 blockers in early trials","IDH inhibitors (ivosidenib, vorasidenib) block the oncometabolite branch","Glutamine PET (18F-FGln) for glioma; metabolic imaging to pick fuel-dependent tumours"]},{"id":"hedgehog","kind":"pathway","name":"Hedgehog signalling","aka":[],"tldr":"A developmental pathway that shapes embryos and is switched back on in basal cell skin cancer and some brain tumours. Blocking it cures most advanced basal cell carcinomas, but tumours learn to reactivate it downstream.","summary":"Hedgehog ligands bind Patched (PTCH1), relieving inhibition of Smoothened (SMO), which activates GLI transcription factors. PTCH1 loss or SMO mutation drives basal cell carcinoma (Gorlin syndrome) and the SHH subgroup of medulloblastoma. SMO inhibitors vismodegib and sonidegib are approved for advanced BCC; glasdegib is approved in AML with low-dose cytarabine. Resistance arises via SMO mutations or SUFU/GLI2 alterations downstream; stromal Hedgehog signalling in pancreatic cancer had paradoxical tumour-restraining effects, and SMO inhibitors failed there.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Hedgehog_signaling_pathway","links":[{"label":"Sekulic et al., Vismodegib in advanced basal-cell carcinoma (NEJM 2012)","url":"https://doi.org/10.1056/NEJMoa1113713"}],"tags":["mechanism"],"related":[],"cancers":["glioblastoma","aml"],"sections":[],"technologies":["kinase-inhibitors","hedgehog-inhibitors"],"targets":["ptch1"],"drugs":[],"companies":[],"institutions":["stanford","dana-farber"],"pathways":["wnt","notch","cancer-stem-cells-plasticity"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":["Leading programmes: Scott (Stanford) and Oro (Stanford) on Hedgehog in skin; Beachy (Stanford, cyclopamine); Rudin/Pomeroy on SHH medulloblastoma."],"analogy":"A construction crew that is supposed to leave once the building is finished. In these cancers the foreman (SMO) never gets the 'stop' message from the site manager (PTCH1).","nodes":[{"id":"shh","label":"SHH ligand","x":15,"y":30},{"id":"ptch","label":"PTCH1","x":38,"y":30},{"id":"smo","label":"SMO","x":60,"y":30},{"id":"sufu","label":"SUFU","x":60,"y":65},{"id":"gli","label":"GLI1/2","x":82,"y":50},{"id":"out","label":"Proliferation (BCC, SHH-medulloblastoma)","x":82,"y":85}],"edges":[{"from":"shh","to":"ptch","type":"inhibits"},{"from":"ptch","to":"smo","type":"inhibits"},{"from":"smo","to":"gli","type":"activates"},{"from":"sufu","to":"gli","type":"inhibits"},{"from":"gli","to":"out","type":"activates"}],"interventions":["SMO inhibitors vismodegib, sonidegib (advanced BCC), glasdegib (AML)","GLI inhibitors (arsenic trioxide, BET inhibitors) for downstream resistance (investigational)","Surgery and radiation remain first line for most BCC"]},{"id":"hepatocellular-carcinoma-signalling","kind":"pathway","name":"Hepatocellular carcinoma (KEGG map)","aka":["KEGG hsa05225","Hepatocellular carcinoma"],"tldr":"This KEGG map shows how hepatitis viruses, alcohol and aflatoxin leave the liver with mutations in telomerase, TP53, Wnt/beta-catenin, PI3K/AKT/mTOR and the oxidative stress sensor NRF2, which together drive liver cancer. It matters because the map explains why liver cancer is treated mainly with angiogenesis blockers and immunotherapy rather than a single targeted drug.","summary":"Hepatocellular carcinoma (HCC) is the most common primary liver cancer and one of the few human tumours with a viral aetiology. KEGG map hsa05225 draws the genetic and epigenetic changes that follow HBV or HCV infection and alcohol or aflatoxin B1 exposure. Recurrently mutated genes cluster in several driver processes: telomere maintenance (TERT promoter mutation and over-expression), TP53, cell cycle control (CDKN2A deletion, RB1 loss), Wnt/beta-catenin (activating CTNNB1 mutation, AXIN1 inactivation, FZD7 over-expression), PI3K/AKT/mTOR (PIK3CA mutation, PTEN loss), receptor signalling through MET/HGF, IGF2/IGF1R and TGF-alpha/EGFR into RAS-ERK and PLC-gamma/PKC, reduced TGF-beta receptor II, and, from exome sequencing, chromatin remodelling (ARID1A, ARID2) and the oxidative stress pathway (KEAP1 loss or NFE2L2/NRF2 activation).\n\nLlovet and colleagues, Nature Reviews Disease Primers, 2021 (doi:10.1038/s41572-020-00240-3) review the disease: TERT promoter, CTNNB1 and TP53 are the most frequent mutations, none of the main drivers is directly druggable, and CTNNB1-mutant tumours are immune-excluded and respond poorly to checkpoint inhibitors, while MET-high and VEGF-driven tumours underpin the success of multikinase and anti-angiogenic drugs.\n\nWhat drugs do about it: first-line systemic therapy is atezolizumab plus bevacizumab (PD-L1 plus VEGF) or durvalumab plus tremelimumab (PD-L1 plus CTLA-4), with the multikinase inhibitors sorafenib and lenvatinib (both listed by KEGG) as alternatives; regorafenib, cabozantinib (which also hits MET) and ramucirumab (for high AFP) are used after progression. In China camrelizumab plus rivoceranib and donafenib are approved. Glypican-3, a proteoglycan over-expressed on most HCC, is an antigen for antibodies and CAR-T cells in trials. The Wnt/beta-catenin and TERT steps remain undrugged.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05225","url":"https://www.kegg.jp/pathway/hsa05225"},{"label":"Review: Hepatocellular carcinoma (Nature Reviews Disease Primers)","url":"https://doi.org/10.1038/s41572-020-00240-3"}],"tags":[],"related":["wnt","keap1-nrf2","vegf-angiogenesis","pd1-checkpoint","proteoglycans-in-cancer"],"cancers":["hcc"],"sections":[],"technologies":[],"targets":["tp53","met","kras","mtor","vegf","pdl1","ctla4","gpc3"],"drugs":["atezolizumab","bevacizumab","durvalumab","tremelimumab","sorafenib","lenvatinib","regorafenib","cabozantinib","ramucirumab","camrelizumab","rivoceranib","donafenib"],"companies":[],"institutions":[],"pathways":["wnt","p53-cell-cycle","pi3k-akt-mtor","ras-mapk","keap1-nrf2","tgf-beta","vegf-angiogenesis","pd1-checkpoint","rtk-activation"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-llovet-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A factory that has been running in a toxic, inflamed environment for decades. The chronic damage rewrites several control panels at once: the age counter is disabled (TERT), the emergency stop is cut (TP53), the growth dial is stuck (Wnt, PI3K) and the smoke alarm is disconnected (NRF2). No single repair fixes it, so treatment cuts the blood supply and calls in the immune system.","nodes":[{"id":"cause","label":"HBV / HCV, alcohol, aflatoxin","x":50,"y":4},{"id":"tert","label":"TERT (telomerase)","x":10,"y":24},{"id":"wnt","label":"CTNNB1 / AXIN1 (Wnt)","x":32,"y":24},{"id":"p53","label":"TP53 / RB1 / CDKN2A","x":55,"y":24,"targetId":"tp53"},{"id":"nrf2","label":"KEAP1 / NRF2","x":78,"y":24},{"id":"rtk","label":"MET, IGF1R, EGFR","x":95,"y":24,"targetId":"met"},{"id":"ras","label":"RAS / ERK","x":90,"y":50,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT / mTOR (PTEN lost)","x":65,"y":50,"targetId":"mtor"},{"id":"vegf","label":"VEGF angiogenesis","x":20,"y":60,"targetId":"vegf"},{"id":"pdl1","label":"PD-L1 immune escape","x":45,"y":72,"targetId":"pdl1"},{"id":"out","label":"Hepatocellular carcinoma","x":50,"y":94}],"edges":[{"from":"cause","to":"tert","type":"activates"},{"from":"cause","to":"wnt","type":"activates"},{"from":"cause","to":"p53","type":"inhibits"},{"from":"cause","to":"nrf2","type":"activates"},{"from":"cause","to":"rtk","type":"activates"},{"from":"rtk","to":"ras","type":"activates"},{"from":"rtk","to":"pi3k","type":"activates"},{"from":"tert","to":"out","type":"activates"},{"from":"wnt","to":"out","type":"activates"},{"from":"p53","to":"out","type":"inhibits"},{"from":"nrf2","to":"out","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"vegf","to":"out","type":"activates"},{"from":"pdl1","to":"out","type":"activates"}],"interventions":["First line: atezolizumab plus bevacizumab (PD-L1 plus VEGF) or durvalumab plus tremelimumab (PD-L1 plus CTLA-4)","Multikinase inhibitors sorafenib and lenvatinib (listed by KEGG), and camrelizumab plus rivoceranib or donafenib in China","Second line: regorafenib, cabozantinib (also blocks MET), ramucirumab for AFP-high disease","Glypican-3 directed antibodies and CAR-T cells in clinical trials","Wnt/beta-catenin and TERT steps: no approved drug yet"]},{"id":"hippo-yap","kind":"pathway","name":"Hippo-YAP/TAZ","aka":[],"tldr":"The 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.","summary":"Mechanical and contact cues activate the Hippo kinases MST1/2-LATS1/2, which phosphorylate and exclude YAP/TAZ from the nucleus. NF2 (Merlin) loss (mesothelioma, meningioma), LATS loss, and YAP/TAZ fusions (epithelioid haemangioendothelioma) unleash YAP/TAZ-TEAD transcription. TEAD palmitoylation-pocket inhibitors (IK-930, VT3989, IAG933) are in phase 1/2, notably in NF2-mutant mesothelioma and as combinations to overcome KRAS/EGFR-inhibitor resistance, where YAP is a bypass route. YAP also drives stiffness-induced signalling and CAF activation.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Hippo_signaling_pathway","links":[{"label":"Franklin, Wu & Guan, Insights into recent findings and clinical application of YAP and TAZ in cancer (Nature Reviews Cancer 2023)","url":"https://doi.org/10.1038/s41568-023-00579-1"}],"tags":["mechanism"],"related":[],"cancers":["mesothelioma"],"sections":[],"technologies":[],"targets":["kras","egfr"],"drugs":[],"companies":[],"institutions":["mgh","mskcc"],"pathways":["ras-mapk","emt","tumor-microenvironment"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-franklin-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Camargo (Stanford), Guan (UCSD), Pan (Johns Hopkins) on Hippo biology; Vivace Therapeutics, Ikena, Novartis on TEAD inhibitors."],"analogy":"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.","nodes":[{"id":"cue","label":"Contact, stiffness, GPCRs","x":12,"y":40},{"id":"nf2","label":"NF2 (Merlin)","x":32,"y":20},{"id":"mst","label":"MST1/2 → LATS1/2","x":50,"y":40},{"id":"yap","label":"YAP/TAZ","x":70,"y":40},{"id":"tead","label":"TEAD transcription","x":88,"y":40},{"id":"out","label":"Growth, EMT, drug tolerance","x":88,"y":80}],"edges":[{"from":"cue","to":"mst","type":"activates"},{"from":"nf2","to":"mst","type":"activates"},{"from":"mst","to":"yap","type":"inhibits"},{"from":"yap","to":"tead","type":"activates"},{"from":"tead","to":"out","type":"activates"}],"interventions":["TEAD inhibitors (VT3989, IK-930, IAG933) in NF2-mutant mesothelioma and with KRAS/EGFR inhibitors","Verteporfin repurposing (preclinical)","Combination rationale: YAP bypass after MAPK inhibition"]},{"id":"inflammation-nfkb","kind":"pathway","name":"Inflammation & NF-κB","aka":[],"tldr":"Chronic 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.","summary":"Tumour-promoting inflammation (a hallmark enabling characteristic) acts through IL-6/STAT3, TNF/NF-κB, IL-1β, COX-2/PGE2, and inflammasome signalling. NF-κB (canonical IKK/IκB and non-canonical NIK/RelB) drives survival, cytokine production, and the SASP; it is constitutively active in many lymphomas (ABC-DLBCL via MYD88/CD79B, BTK-dependent) and myeloma. Aspirin reduces colorectal cancer in Lynch syndrome (CAPP2) and canakinumab reduced lung cancer incidence in CANTOS (secondary analysis), though as treatment it failed. Anti-inflammatory chemoprevention is one of the few validated mechanism-based preventions.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/NF-κB","links":[{"label":"Greten & Grivennikov, Inflammation and cancer: triggers, mechanisms, and consequences (Immunity 2019)","url":"https://doi.org/10.1016/j.immuni.2019.06.025"}],"tags":["mechanism"],"related":[],"cancers":["gallbladder","non-hodgkin-lymphoma"],"sections":[],"technologies":["chemoprevention","hpv-vaccine"],"targets":["btk"],"drugs":["ibrutinib","zanubrutinib"],"companies":[],"institutions":["ucsf","mskcc","dkfz"],"pathways":["jak-stat","senescence","tumor-microenvironment","cachexia-biology"],"terms":["tumor-promoting-inflammation","crs"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-greten-immunity"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Karin (UCSD) on NF-κB and inflammation-driven cancer; Coussens (OHSU) on inflammation and TME; Greten (Frankfurt); CAPP2 (Burn, Newcastle)."],"analogy":"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.","nodes":[{"id":"trig","label":"Infection, injury, obesity","x":12,"y":50},{"id":"cyto","label":"TNF, IL-1β, IL-6","x":35,"y":25},{"id":"ikk","label":"IKK → IκB degradation","x":58,"y":25},{"id":"nfkb","label":"NF-κB","x":80,"y":25},{"id":"stat3","label":"IL-6 → STAT3","x":58,"y":75},{"id":"out","label":"Survival, proliferation, SASP, immune suppression","x":92,"y":60},{"id":"btk","label":"BCR → BTK (lymphoma)","x":35,"y":75,"targetId":"btk"},{"id":"cox","label":"COX-2 → PGE2","x":12,"y":85}],"edges":[{"from":"trig","to":"cyto","type":"activates"},{"from":"cyto","to":"ikk","type":"activates"},{"from":"ikk","to":"nfkb","type":"activates"},{"from":"nfkb","to":"out","type":"activates"},{"from":"cyto","to":"stat3","type":"activates"},{"from":"stat3","to":"out","type":"activates"},{"from":"btk","to":"ikk","type":"activates"},{"from":"trig","to":"cox","type":"activates"},{"from":"cox","to":"out","type":"activates"}],"interventions":["Aspirin chemoprevention in Lynch syndrome (CAPP2); NSAIDs in FAP","BTK inhibitors (ibrutinib, zanubrutinib) shut NF-κB in CLL/lymphoma","IL-6/STAT3 blockade in cachexia and CRS (tocilizumab)","Anti-H. pylori therapy prevents gastric cancer; HBV/HCV treatment prevents HCC"]},{"id":"intravasation-ctc-survival","kind":"pathway","name":"Intravasation & circulating tumour cells","aka":[],"tldr":"Getting into the bloodstream kills almost every cell that tries: 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.","summary":"Intravasation happens at 'TMEM doorways' (a tumour cell, a perivascular TIE2+ macrophage and an endothelial cell in direct contact) where macrophage-derived VEGF-A transiently opens the vessel; leaky angiogenic vessels and lymphatics (VEGF-C) also admit cells. In the blood, detachment triggers anoikis (integrin loss → BIM/BMF → apoptosis), which metastatic cells resist via TrkB, PI3K, and mesenchymal states; shear stress and oxidative stress (antioxidants paradoxically aid metastasis) kill most; NK cells clear the rest. Survivors use platelet cloaking (P-selectin, tissue-factor-thrombin, platelet TGF-β shielding from NK and inducing EMT), neutrophil escorts and NETs, and travel as clusters held by plakoglobin and CD44, which are up to 50-fold more metastatic than singles and carry hypomethylated stemness genes. CTCs (EpCAM-based CellSearch; ≥5 per 7.5 mL prognostic in breast, prostate, colorectal) and ctDNA are the clinical windows; fragmentomics and methylation extend detection. Aspirin and anticoagulants reduce metastasis in models via the platelet arm; Na+/K+-ATPase inhibitors dissolve clusters preclinically.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Circulating_tumor_cell","links":[{"label":"Aceto et al., Circulating tumor cell clusters are oligoclonal precursors of breast cancer metastasis (Cell 2014)","url":"https://doi.org/10.1016/j.cell.2014.07.013"},{"label":"Massagué & Obenauf, Metastatic colonization by circulating tumour cells (Nature 2016)","url":"https://doi.org/10.1038/nature17038"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["liquid-biopsy","mrd-testing","continuous-ctdna-monitoring","fragmentomics","aspirin-cancer-prevention","cancer-associated-thrombosis"],"targets":["vegf","csf1r","epcam","pik3ca"],"drugs":["signatera","shield","bevacizumab"],"companies":[],"institutions":[],"pathways":["metastatic-cascade","angiogenic-switch","nk-cell-recognition","emt","pre-metastatic-niche"],"terms":["ctdna","mrd","vaf","disseminated-tumor-cells","activating-invasion-metastasis"],"trials":[],"people":[],"bottlenecks":["b-metastasis-biology","b-dormancy-mrd"],"keyPapers":["paper-massague-nature","paper-aceto-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"tmem","label":"TMEM doorway (macrophage)","x":12,"y":15,"targetId":"csf1r"},{"id":"leak","label":"Leaky vessels (VEGF)","x":40,"y":15,"targetId":"vegf"},{"id":"intra","label":"Intravasation","x":68,"y":15},{"id":"anoik","label":"Anoikis (detachment death)","x":12,"y":48},{"id":"shear","label":"Shear, oxidative stress","x":40,"y":48},{"id":"nk","label":"NK-cell clearance","x":68,"y":48},{"id":"ctc","label":"Surviving CTCs / clusters","x":92,"y":48},{"id":"plat","label":"Platelet cloak, NETs","x":92,"y":82},{"id":"clust","label":"Clusters (plakoglobin, CD44)","x":62,"y":82},{"id":"lb","label":"Liquid biopsy detection","x":30,"y":82}],"edges":[{"from":"tmem","to":"intra","type":"activates"},{"from":"leak","to":"intra","type":"activates"},{"from":"intra","to":"ctc","type":"activates"},{"from":"anoik","to":"ctc","type":"inhibits"},{"from":"shear","to":"ctc","type":"inhibits"},{"from":"nk","to":"ctc","type":"inhibits"},{"from":"plat","to":"nk","type":"inhibits"},{"from":"clust","to":"anoik","type":"inhibits"},{"from":"plat","to":"ctc","type":"activates"},{"from":"clust","to":"ctc","type":"activates"},{"from":"ctc","to":"lb","type":"activates"}],"interventions":["Liquid biopsy: CTC enumeration (CellSearch), ctDNA (Signatera, Guardant Reveal), fragmentomics for detection and MRD","Aspirin and low-molecular-weight heparin target platelet cloaking; aspirin reduces recurrence in PIK3CA-mutant colorectal cancer (ALASCCA)","Anti-VEGF and CSF1R inhibition close TMEM doorways in models","Cluster-dissociating agents (digoxin analogues) and NK-boosting therapies are experimental"]},{"id":"apoptosis-bcl2","kind":"pathway","name":"Intrinsic apoptosis (BCL-2 family)","aka":[],"tldr":"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.","summary":"Stress signals induce BH3-only proteins (BIM, PUMA, NOXA, BAD) that either inhibit anti-apoptotic BCL-2, BCL-XL, MCL-1 or directly activate BAX/BAK, which permeabilise the mitochondrial outer membrane, releasing cytochrome c to activate caspase-9 and executioner caspases. Cancers overexpress BCL-2 (t(14;18) in follicular lymphoma; CLL), MCL-1 (myeloma, AML), or BCL-XL (solid tumours, platelets). Venetoclax (BCL-2) transformed CLL and AML; MCL-1 inhibitors have cardiac toxicity; BCL-XL degraders and platelet-sparing PROTACs are in development.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Bcl-2_family","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Bcl-2_family"},{"label":"Tsujimoto et al., Science 1985: the t(14;18) translocation results from a mistake in VDJ joining","url":"https://doi.org/10.1126/science.3929382"},{"label":"Blombery et al., Cancer Discov 2019: the recurrent BCL2 Gly101Val mutation confers resistance to venetoclax","url":"https://doi.org/10.1158/2159-8290.CD-18-1119"}],"tags":[],"related":[],"cancers":["cll","aml","dlbcl","non-hodgkin-lymphoma","follicular-lymphoma","mantle-cell-lymphoma"],"sections":[],"technologies":[],"targets":["bcl2","tp53","menin","bim","bax"],"drugs":["venetoclax"],"companies":[],"institutions":[],"pathways":[],"terms":["lymphoma-bio-germinal-centre"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":["Lymphoma: the founding lesion of follicular lymphoma is a recombination accident. The t(14;18) breakpoints carry N-region nucleotides and sit beside signal-like sequences, which identifies the translocation as a mistake by the VDJ recombinase at the pre-B-cell stage rather than a late event (Tsujimoto 1985). Venetoclax has not repeated its chronic lymphocytic leukaemia result in lymphoma, because these cells also lean on MCL1 and BCL-xL, and the escape mutation when it does work, BCL2 G101V, lowers drug affinity about 180-fold and appears months before clinical progression (Blombery 2019)."],"analogy":"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.","nodes":[{"id":"stress","label":"DNA damage, oncogene stress, p53","x":50,"y":5},{"id":"bh3","label":"BH3-only (BIM, PUMA, NOXA)","x":50,"y":22},{"id":"bcl2","label":"BCL-2 / BCL-XL / MCL-1","x":20,"y":42,"targetId":"bcl2"},{"id":"bax","label":"BAX / BAK","x":65,"y":45},{"id":"momp","label":"Mitochondrial permeabilisation","x":65,"y":62},{"id":"cytc","label":"Cytochrome c → caspase-9","x":65,"y":78},{"id":"casp","label":"Caspase-3/7 → apoptosis","x":65,"y":95}],"edges":[{"from":"stress","to":"bh3","type":"activates"},{"from":"bh3","to":"bcl2","type":"inhibits"},{"from":"bh3","to":"bax","type":"activates"},{"from":"bcl2","to":"bax","type":"inhibits"},{"from":"bax","to":"momp","type":"activates"},{"from":"momp","to":"cytc","type":"activates"},{"from":"cytc","to":"casp","type":"activates"}],"interventions":["Venetoclax (BCL-2) in CLL, AML, mantle cell lymphoma","Next-generation BCL-2 inhibitors sonrotoclax, lisaftoclax","MCL-1 inhibitors (limited by cardiotoxicity)","BCL-XL PROTACs sparing platelets","Combinations with hypomethylating agents, BTK inhibitors, menin inhibitors"]},{"id":"invasion-ecm-degradation","kind":"pathway","name":"Invasion: proteases, adhesion & the invasive front","aka":[],"tldr":"To 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.","summary":"Invasion cycles through protrusion (RAC1-driven lamellipodia, actin-rich invadopodia with cortactin and TKS5), adhesion (integrins α5β1, αvβ3, αvβ6 to fibronectin/collagen, signalling via FAK-SRC and ILK), matrix degradation (membrane-anchored MT1-MMP/MMP14 activating MMP2, secreted MMP9, uPA-uPAR-plasmin, cathepsins), and RHO-ROCK-myosin contraction. Modes: mesenchymal (protease-dependent, elongated), amoeboid (protease-independent squeezing through pores, RHO/ROCK-high), and collective invasion led by leader cells (keratin-14+ in breast cancer) or by CAFs that generate tracks and pull via N-cadherin/E-cadherin heterotypic junctions. Perineural and lymphovascular invasion are histological markers of the process; stiff, aligned collagen (TACS-3) promotes it. Broad-spectrum MMP inhibitors (marimastat, prinomastat) failed in phase 3 with musculoskeletal toxicity and because MMPs also produce anti-angiogenic fragments. Current approaches: FAK inhibitors (defactinib, approved with avutometinib in KRAS-mutant low-grade serous ovarian cancer; also reduce stromal density), integrin antagonists (cilengitide failed in GBM), uPAR-targeted CAR-T for senescent cells, and RHO/ROCK inhibitors.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Invadopodia","links":[{"label":"Friedl & Alexander, Cancer invasion and the microenvironment: plasticity and reciprocity (Cell 2011)","url":"https://doi.org/10.1016/j.cell.2011.11.016"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["lgsoc","pancreatic","head-and-neck"],"sections":[],"technologies":["fluorescence-guided-surgery","robotic-surgery","mechanobiology-therapy","fapi-pet","single-cell-spatial","histopathology-ihc"],"targets":["fap","kras","met","smoothened","fak"],"drugs":["avutometinib-defactinib","fap-2286"],"companies":[],"institutions":[],"pathways":["emt","basement-membrane-tissue-barriers","tgf-beta","hippo-yap","caf-activation-desmoplasia","metastatic-cascade"],"terms":["desmoplasia","cancer-associated-fibroblasts","tnm-staging","activating-invasion-metastasis"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-friedl-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"cue","label":"TGF-β, HGF, hypoxia, stiffness","x":12,"y":12},{"id":"emt","label":"EMT programme (ZEB1, SNAIL)","x":40,"y":12},{"id":"int","label":"Integrins → FAK / SRC","x":68,"y":12},{"id":"inv","label":"Invadopodia, MT1-MMP","x":68,"y":42},{"id":"mmp","label":"MMP2/9, uPA → ECM breach","x":92,"y":55},{"id":"rho","label":"RHO-ROCK contraction","x":40,"y":42},{"id":"amoe","label":"Amoeboid squeezing","x":12,"y":42},{"id":"coll","label":"Collective invasion (leaders)","x":40,"y":75},{"id":"caf","label":"CAF tracks","x":12,"y":75,"targetId":"fap"},{"id":"front","label":"Invasive front → vessels","x":72,"y":82}],"edges":[{"from":"cue","to":"emt","type":"activates"},{"from":"emt","to":"int","type":"activates"},{"from":"int","to":"inv","type":"activates"},{"from":"inv","to":"mmp","type":"activates"},{"from":"cue","to":"rho","type":"activates"},{"from":"rho","to":"amoe","type":"activates"},{"from":"emt","to":"coll","type":"activates"},{"from":"caf","to":"coll","type":"activates"},{"from":"mmp","to":"front","type":"activates"},{"from":"coll","to":"front","type":"activates"},{"from":"amoe","to":"front","type":"activates"}],"interventions":["FAK inhibitor defactinib with avutometinib (approved 2025, KRAS-mutant low-grade serous ovarian cancer); FAK inhibition also softens stroma","Broad MMP inhibitors and the integrin antagonist cilengitide failed in phase 3; lesson retained in the failure museum","Surgery and radiotherapy margins are the practical answer to local invasion; perineural and lymphovascular invasion drive adjuvant decisions","Anti-stromal strategies (FAP theranostics, Hedgehog paradox) reshape the tracks"]},{"id":"jak-stat","kind":"pathway","name":"JAK-STAT signalling","aka":[],"tldr":"The relay that turns cytokine signals into gene changes. Overactive in blood cancers (JAK2 in myelofibrosis), it is also the wire that carries interferon's cancer-killing message, so tumours cut it to escape immunotherapy.","summary":"Cytokine receptors recruit JAK kinases that phosphorylate STAT transcription factors. JAK2 V617F drives myeloproliferative neoplasms (ruxolitinib, fedratinib, momelotinib approved); STAT3 is a hub for IL-6-driven survival, immunosuppression, and cachexia; STAT5 in leukaemias. Interferon-γ signalling via JAK1/2-STAT1 upregulates MHC and PD-L1; JAK1/2 loss-of-function mutations cause acquired resistance to PD-1 blockade. Direct STAT3 inhibitors (degraders, antisense) are in early trials.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/JAK-STAT_signaling_pathway","links":[{"label":"Hu et al., The JAK/STAT signaling pathway: from bench to clinic (Signal Transduction and Targeted Therapy 2021)","url":"https://doi.org/10.1038/s41392-021-00791-1"}],"tags":["mechanism"],"related":[],"cancers":["non-hodgkin-lymphoma"],"sections":[],"technologies":[],"targets":["pdl1","stat5","mpl","jak1"],"drugs":[],"companies":[],"institutions":["mskcc","mayo-clinic"],"pathways":["antigen-presentation-immunoediting","inflammation-nfkb","cachexia-biology","pd1-checkpoint"],"terms":["crs"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-darnell-jak-stat-science-1994","paper-hu-signal-transduct-target-ther"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Levine (MSK) on JAK2 and MPN; Ribas (UCLA) on JAK mutations and PD-1 resistance; Darnell legacy (Rockefeller)."],"analogy":"JAK-STAT is a doorbell wired to the house lights. Cytokines ring, JAK flips the switch, STAT turns on the lights. Cancers either jam the switch on (JAK2 mutation) or cut the wire so interferon cannot turn on the 'I am infected' lights.","nodes":[{"id":"cyto","label":"Cytokine / IFN-γ","x":15,"y":40},{"id":"rec","label":"Receptor","x":35,"y":40},{"id":"jak","label":"JAK1/2","x":55,"y":40},{"id":"stat","label":"STAT1 / STAT3 / STAT5","x":75,"y":40},{"id":"mhc","label":"MHC-I, PD-L1 (STAT1)","x":92,"y":20,"targetId":"pdl1"},{"id":"surv","label":"Survival, SASP, cachexia (STAT3)","x":92,"y":65},{"id":"socs","label":"SOCS feedback","x":55,"y":80}],"edges":[{"from":"cyto","to":"rec","type":"activates"},{"from":"rec","to":"jak","type":"activates"},{"from":"jak","to":"stat","type":"activates"},{"from":"stat","to":"mhc","type":"activates"},{"from":"stat","to":"surv","type":"activates"},{"from":"stat","to":"socs","type":"activates"},{"from":"socs","to":"jak","type":"inhibits"}],"interventions":["JAK inhibitors in myeloproliferative neoplasms and GVHD","STAT3 degraders/antisense (early trials)","Anti-IL-6 (tocilizumab) for CRS and under study for cachexia","JAK-loss tumours: MHC-independent therapies (T-cell engagers, NK cells)"]},{"id":"keap1-nrf2","kind":"pathway","name":"KEAP1-NRF2 antioxidant pathway","aka":[],"tldr":"KEAP1-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.","summary":"KEAP1 targets NRF2 (NFE2L2) for degradation; oxidative stress or KEAP1 mutation (~20% of lung adenocarcinoma, often with KRAS or STK11) stabilises NRF2, inducing glutathione synthesis, drug efflux, and NADPH production. KEAP1/STK11-mutant NSCLC responds poorly to chemotherapy and immunotherapy. Vulnerabilities: glutaminase dependence (telaglenastat, KEAPSAKE trial negative), NRF2-activated prodrugs (TRC102?), and CDK4/6 dependence; NRF2 also drives ferroptosis resistance.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/NFE2L2","links":[{"label":"Romero et al., Keap1 loss promotes Kras-driven lung cancer and results in dependence on glutaminolysis (Nature Medicine 2017)","url":"https://doi.org/10.1038/nm.4407"}],"tags":["mechanism"],"related":[],"cancers":["nsclc"],"sections":[],"technologies":[],"targets":["kras"],"drugs":[],"companies":[],"institutions":["mskcc","md-anderson"],"pathways":["ferroptosis-cell-death","cancer-metabolism","ras-mapk"],"terms":["efflux-pump"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-romero-nat-med"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Papagiannakopoulos (NYU) on KEAP1 in lung cancer; DeNicola (Moffitt); Skoulidis and Heymach (MD Anderson) on STK11/KEAP1 and IO."],"analogy":"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.","nodes":[{"id":"ros","label":"Oxidative stress","x":12,"y":40},{"id":"keap","label":"KEAP1 (mutated)","x":38,"y":40},{"id":"nrf2","label":"NRF2","x":62,"y":40},{"id":"are","label":"ARE genes: GSH, NQO1, efflux","x":86,"y":40},{"id":"res","label":"Chemo/RT/ferroptosis resistance","x":86,"y":80},{"id":"stk11","label":"STK11/LKB1 loss (co-mutation)","x":38,"y":80}],"edges":[{"from":"ros","to":"keap","type":"inhibits"},{"from":"keap","to":"nrf2","type":"inhibits"},{"from":"nrf2","to":"are","type":"activates"},{"from":"are","to":"res","type":"activates"},{"from":"stk11","to":"res","type":"activates"}],"interventions":["Glutaminase inhibition (negative in KEAPSAKE)","NRF2-activated prodrugs and NRF2 inhibitors (preclinical)","Biomarker: KEAP1/STK11 status predicts poor IO benefit in NSCLC","CDK4/6 and mTOR combinations under study"]},{"id":"lineage-plasticity-neuroendocrine","kind":"pathway","name":"Lineage plasticity & neuroendocrine transformation","aka":[],"tldr":"Under 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.","summary":"Lineage plasticity requires loss of the gatekeepers TP53 and RB1 (Ku et al., Mu et al. 2017), which unlocks SOX2, EZH2-mediated repression of lineage genes, and reactivation of neural programmes (ASCL1, NEUROD1, INSM1, BRN2), producing AR-indifferent neuroendocrine prostate cancer in 15-20% of castration-resistant cases after potent AR inhibitors, and small-cell transformation in ~5-15% of EGFR-mutant NSCLC on osimertinib (also after ALK inhibitors and in immunotherapy-treated adenocarcinoma). Related transitions: squamous transdifferentiation of adenocarcinoma, sarcomatoid dedifferentiation in RCC and mesothelioma, MITF-low neural-crest states in melanoma under BRAF inhibitors, and blast/Richter transformation in lymphoid cancers. The new state expresses DLL3, SEZ6, B7-H3, CEACAM5 and loses PSMA or EGFR dependence, is transiently sensitive to platinum-etoposide, and is detected by biopsy at progression (recommended when PSA is low relative to disease burden or ctDNA shows TP53/RB1 loss) and by DLL3 PET. Therapeutics: DLL3 engagers (tarlatamab), EZH2 inhibitors (mevrometostat + enzalutamide, tazemetostat) to block or reverse the switch, Aurora A inhibitors for MYCN/ASCL1 states, and B7-H3 or SEZ6 ADCs.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Neuroendocrine_tumor","links":[{"label":"Ku et al., Rb1 and Trp53 cooperate to suppress prostate cancer lineage plasticity, metastasis, and antiandrogen resistance (Science 2017)","url":"https://doi.org/10.1126/science.aah4199"},{"label":"Beltran et al., Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer (Nat Med 2016)","url":"https://doi.org/10.1038/nm.4045"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["prostate","nsclc","sclc","melanoma"],"sections":[],"technologies":["t-cell-engager","adc","epigenetic-drugs","liquid-biopsy","histopathology-ihc"],"targets":["dll3","b7h3","ezh2","tp53","androgen-receptor","egfr","ceacam5","psma"],"drugs":["tarlatamab","ifinatamab-deruxtecan","mevrometostat","tazemetostat","platinum-etoposide","enzalutamide","osimertinib"],"companies":[],"institutions":[],"pathways":["cancer-stem-cells-plasticity","ar-signaling","epigenetic-reprogramming","notch","p53-cell-cycle","resistance-routes-map"],"terms":["histologic-transformation","castration-resistance","unlocking-phenotypic-plasticity","richter-transformation","epithelioid-vs-sarcomatoid"],"trials":[],"people":[],"bottlenecks":["b-resistance"],"keyPapers":["paper-ku-rb1-trp53-lineage-plasticity-science-2017"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"adeno","label":"Adenocarcinoma (AR / EGFR)","x":12,"y":15,"targetId":"androgen-receptor"},{"id":"drug","label":"ARPI or EGFR TKI pressure","x":12,"y":48,"targetId":"egfr"},{"id":"loss","label":"TP53 + RB1 loss","x":45,"y":15,"targetId":"tp53"},{"id":"sox2","label":"SOX2, EZH2, ASCL1/NEUROD1","x":45,"y":48,"targetId":"ezh2"},{"id":"ne","label":"Neuroendocrine / small-cell","x":78,"y":30},{"id":"dll3","label":"DLL3, B7-H3, SEZ6 surface","x":78,"y":62,"targetId":"dll3"},{"id":"indiff","label":"AR / EGFR indifferent","x":45,"y":82},{"id":"tx","label":"Tarlatamab, platinum-etoposide","x":78,"y":92},{"id":"ezh2i","label":"EZH2 inhibitors block switch","x":12,"y":82}],"edges":[{"from":"adeno","to":"ne","type":"activates"},{"from":"drug","to":"sox2","type":"activates"},{"from":"loss","to":"sox2","type":"activates"},{"from":"sox2","to":"ne","type":"activates"},{"from":"ne","to":"dll3","type":"activates"},{"from":"ne","to":"indiff","type":"activates"},{"from":"drug","to":"adeno","type":"inhibits"},{"from":"tx","to":"ne","type":"inhibits"},{"from":"ezh2i","to":"sox2","type":"inhibits"}],"interventions":["Re-biopsy at progression when the clinical picture and markers diverge; ctDNA TP53/RB1 loss as a warning","DLL3 T-cell engager tarlatamab (SCLC; trials in neuroendocrine prostate cancer); B7-H3 and SEZ6 ADCs","EZH2 inhibitors (mevrometostat with enzalutamide, tazemetostat) to prevent or reverse plasticity; Aurora A inhibitors for MYCN/ASCL1-high states","Platinum-etoposide gives transient responses in transformed disease"]},{"id":"lipid-metabolism-cancer","kind":"pathway","name":"Lipid synthesis, uptake & cholesterol","aka":[],"tldr":"Dividing 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.","summary":"Citrate exported from mitochondria is cleaved by ACLY to acetyl-CoA, carboxylated by ACC to malonyl-CoA, and built into palmitate by FASN, then desaturated by SCD1 and elongated; SREBP1/2, activated by mTORC1 and hypoxia and restrained by AMPK and LKB1, is the master transcriptional switch. The mevalonate pathway (HMGCR) makes cholesterol and the isoprenoids needed to anchor RAS to membranes; mutant p53 activates it. Exogenous uptake via CD36, LDLR and lipoprotein lipase supplies fatty acids from adipocytes (omental metastasis of ovarian cancer, prostate cancer in bone) and fatty acid oxidation (CPT1) fuels dormant, metastatic and leukaemia stem cells. Lipid saturation sets ferroptosis sensitivity (PUFAs are the substrate; SCD1 protects). Epidemiologically obesity raises risk of 13 cancers via insulin/IGF-1, oestrogen, inflammation and lipid supply; GLP-1 agonists and bariatric surgery lower incidence. Drugs: FASN inhibitor denifanstat (TVB-2640) in trials, statins repurposed (mostly negative in RCTs), CD36 antibodies preclinical; SCD1 inhibition to sensitise to ferroptosis.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Fatty_acid_synthesis","links":[{"label":"Röhrig & Schulze, The multifaceted roles of fatty acid synthesis in cancer (Nat Rev Cancer 2016)","url":"https://doi.org/10.1038/nrc.2016.89"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["ovarian","prostate","breast-hr-positive","endometrial","hcc"],"sections":[],"technologies":["metabolic-therapy","glp1-agonists-cancer-risk","bariatric-surgery-cancer-incidence","mediterranean-plant-forward-diet","exercise-oncology","dietitian-led-weight-loss-breast"],"targets":["akt","kras","estrogen-receptor"],"drugs":[],"companies":[],"institutions":[],"pathways":["cancer-metabolism","ferroptosis-cell-death","pi3k-akt-mtor","tumor-dormancy"],"terms":["obesity-related-cancers","energy-balance","metabolic-syndrome","body-composition","deregulating-cellular-energetics"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-rohrig-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"cit","label":"Citrate → ACLY → acetyl-CoA","x":12,"y":20},{"id":"acc","label":"ACC → malonyl-CoA","x":40,"y":20},{"id":"fasn","label":"FASN → palmitate","x":68,"y":20},{"id":"scd","label":"SCD1 → membranes","x":90,"y":40},{"id":"srebp","label":"SREBP (mTORC1, hypoxia)","x":40,"y":50,"targetId":"akt"},{"id":"ampk","label":"AMPK / LKB1","x":12,"y":50},{"id":"mev","label":"Mevalonate → cholesterol","x":68,"y":55,"targetId":"kras"},{"id":"cd36","label":"CD36 uptake from adipocytes","x":12,"y":82},{"id":"fao","label":"Fatty acid oxidation (CPT1)","x":40,"y":82},{"id":"ferro","label":"PUFA → ferroptosis","x":90,"y":80},{"id":"obes","label":"Obesity, insulin, IGF-1","x":68,"y":88}],"edges":[{"from":"cit","to":"acc","type":"activates"},{"from":"acc","to":"fasn","type":"activates"},{"from":"fasn","to":"scd","type":"activates"},{"from":"srebp","to":"fasn","type":"activates"},{"from":"srebp","to":"mev","type":"activates"},{"from":"ampk","to":"srebp","type":"inhibits"},{"from":"cd36","to":"fao","type":"activates"},{"from":"cd36","to":"scd","type":"activates"},{"from":"scd","to":"ferro","type":"inhibits"},{"from":"obes","to":"srebp","type":"activates"},{"from":"obes","to":"cd36","type":"activates"}],"interventions":["FASN inhibitor denifanstat and SCD1 inhibitors in trials; statins repurposed with mostly negative randomised data","Weight management, GLP-1 agonists and bariatric surgery reduce obesity-related cancer incidence","Exercise and dietary pattern interventions during treatment","Ferroptosis inducers exploit the lipid composition of mesenchymal and persister cells"]},{"id":"melanoma-signalling","kind":"pathway","name":"Melanoma (KEGG map)","aka":["KEGG hsa05218","Melanoma"],"tldr":"KEGG's melanoma map shows BRAF or NRAS mutations driving the MAPK growth relay, PTEN loss driving PI3K/AKT, and loss of the CDKN2A brakes (p16 and p14ARF) on CDK4/6 and p53. BRAF plus MEK inhibitors and immune checkpoint antibodies have transformed treatment.","summary":"The KEGG melanoma map (hsa05218) traces how melanocytes become melanoma. Growth factor receptors and, far more often, mutations in BRAF (mostly V600E) or NRAS lock on the RAF to MEK to ERK relay, which drives cyclin D1 and proliferation. NRAS also feeds PI3K, and loss or mutation of PTEN releases AKT signalling for survival. These mutations arise early, in the benign naevus stage, and persist. Progression is then tied to the two tumour suppressor arms drawn on the map: p16INK4a normally holds CDK4/6 away from cyclin D so RB stays bound to E2F, and p14ARF normally blocks MDM2 so that p53 is stable; deletion or mutation of the CDKN2A locus, or activating CDK4 mutation, removes both brakes at once. MITF, the melanocyte master transcription factor, and TP53 mutations are implicated in further progression. Shain and Bastian, Nat Rev Cancer, 2016 (doi:10.1038/nrc.2016.37) review the genetic evolution from melanocyte to melanoma, showing MAPK activation as the initiating event, CDKN2A loss and telomerase promoter mutations at the transition to invasive melanoma, and PTEN and TP53 alterations later, on a background of ultraviolet-signature mutations.\n\nWhat drugs do about it: BRAF inhibitors paired with MEK inhibitors (dabrafenib with trametinib, vemurafenib with cobimetinib, encorafenib with binimetinib) shut the MAPK relay in BRAF V600-mutant melanoma. Immune checkpoint antibodies against PD-1 (nivolumab, pembrolizumab) with or without CTLA-4 (ipilimumab) or LAG-3 (relatlimab) produce durable remissions regardless of mutation and are now first choice for most patients with advanced disease.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05218","url":"https://www.kegg.jp/pathway/hsa05218"},{"label":"Review: From melanocytes to melanomas","url":"https://doi.org/10.1038/nrc.2016.37"}],"tags":[],"related":["ras-mapk","pd1-checkpoint"],"cancers":["melanoma"],"sections":[],"technologies":[],"targets":["braf","mek","akt","cdk4-6","mdm2"],"drugs":["dabrafenib","trametinib","vemurafenib","cobimetinib","encorafenib","binimetinib","nivolumab","pembrolizumab","ipilimumab"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","p53-cell-cycle","cell-cycle-engine-cdks","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-shain-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A car with the accelerator taped down (BRAF or NRAS) and both brake cables cut (p16 and p14ARF). BRAF and MEK inhibitors peel off the tape; immunotherapy calls in the police to stop the car from outside.","nodes":[{"id":"rtk","label":"Growth factor RTKs","x":50,"y":5},{"id":"nras","label":"NRAS","x":50,"y":20},{"id":"braf","label":"BRAF V600E","x":25,"y":36,"targetId":"braf"},{"id":"mek","label":"MEK / ERK","x":25,"y":54,"targetId":"mek"},{"id":"pten","label":"PTEN (lost)","x":90,"y":30},{"id":"pi3k","label":"PI3K / AKT","x":72,"y":46,"targetId":"akt"},{"id":"cdkn2a","label":"CDKN2A: p16 and p14ARF","x":50,"y":62},{"id":"cdk46","label":"Cyclin D / CDK4/6 to RB","x":30,"y":78,"targetId":"cdk4-6"},{"id":"mdm2","label":"MDM2 to p53","x":72,"y":78,"targetId":"mdm2"},{"id":"mitf","label":"MITF","x":90,"y":62},{"id":"out","label":"Proliferation, survival","x":50,"y":95}],"edges":[{"from":"rtk","to":"nras","type":"activates"},{"from":"nras","to":"braf","type":"activates"},{"from":"braf","to":"mek","type":"activates"},{"from":"nras","to":"pi3k","type":"activates"},{"from":"pten","to":"pi3k","type":"inhibits"},{"from":"mek","to":"cdk46","type":"activates"},{"from":"cdkn2a","to":"cdk46","type":"inhibits"},{"from":"cdkn2a","to":"mdm2","type":"inhibits"},{"from":"cdk46","to":"out","type":"activates"},{"from":"mdm2","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"mitf","to":"out","type":"activates"}],"interventions":["BRAF plus MEK inhibitor doublets for BRAF V600-mutant melanoma: dabrafenib with trametinib, vemurafenib with cobimetinib, encorafenib with binimetinib","PD-1 checkpoint antibodies nivolumab or pembrolizumab, alone or combined with ipilimumab (CTLA-4) or relatlimab (LAG-3), for advanced disease and as adjuvant therapy","Adjuvant BRAF plus MEK inhibition or anti-PD-1 after resection of stage III melanoma","MEK inhibitors and CDK4/6 inhibitor combinations under study for NRAS-mutant melanoma"]},{"id":"menin-kmt2a","kind":"pathway","name":"Menin / KMT2A (HOXA9-MEIS1 axis)","aka":[],"tldr":"In some leukaemias a broken chromatin protein (KMT2A, once called MLL) or a mutant NPM1 keeps embryonic growth genes (HOXA9, MEIS1) switched on, so blood cells never mature. Both need a partner called menin to stay on the DNA. Menin inhibitors pull the plug and the cells mature; the first was approved in 2024.","summary":"KMT2A (MLL1) is a histone H3K4 methyltransferase that, fused to one of more than 80 partner genes by chromosomal translocation, becomes a leukaemia driver in infant ALL and in about 10% of adult AML (often therapy-related). The fusion protein is tethered to chromatin through menin (MEN1) and LEDGF, and recruits DOT1L (H3K79 methylation) to keep HOXA9 and MEIS1 transcribed; these homeobox genes hold the cell in a stem-like, undifferentiated state. Mutant NPM1 (about 30% of adult AML), mislocalised to the cytoplasm, drives the same HOX programme and is equally menin-dependent. Small molecules that occupy the KMT2A-binding pocket of menin (revumenib, ziftomenib, bleximenib, enzomenib) release the complex from chromatin, HOXA9 and MEIS1 fall, and the blasts differentiate. Revumenib was approved in 2024 for relapsed or refractory KMT2A-rearranged acute leukaemia (AUGMENT-101) and ziftomenib in 2025 for NPM1-mutant AML (KOMET-001). Differentiation syndrome and QT prolongation are the class toxicities; acquired MEN1 mutations at the drug-binding site are the resistance mechanism.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/KMT2A","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/KMT2A"}],"tags":[],"related":[],"cancers":["aml","all-leukemia"],"sections":[],"technologies":[],"targets":["menin","flt3","hoxa9","meis1"],"drugs":["revumenib","ziftomenib","venetoclax","azacitidine"],"companies":[],"institutions":[],"pathways":["epigenetic-reprogramming"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"Two hands are needed to hold a switch down: KMT2A (or mutant NPM1) is one hand and menin is the other. As long as both hold, the 'stay immature' genes stay on. Menin inhibitors slip a wedge under menin's hand, the switch springs up and the leukaemia cells finish growing up into normal blood cells.","nodes":[{"id":"kmt2a","label":"KMT2A fusion (MLL-r)","x":25,"y":5},{"id":"npm1","label":"Mutant NPM1 (cytoplasmic)","x":75,"y":5},{"id":"menin","label":"Menin (MEN1) + LEDGF","x":50,"y":28,"targetId":"menin"},{"id":"dot1l","label":"DOT1L (H3K79me)","x":82,"y":46},{"id":"chrom","label":"Complex bound at chromatin","x":50,"y":48},{"id":"hox","label":"HOXA9 / MEIS1 transcription","x":50,"y":70},{"id":"flt3","label":"FLT3 (co-mutated; co-target)","x":18,"y":70,"targetId":"flt3"},{"id":"out","label":"Differentiation block → acute leukaemia","x":50,"y":93}],"edges":[{"from":"kmt2a","to":"menin","type":"activates"},{"from":"npm1","to":"menin","type":"activates"},{"from":"menin","to":"chrom","type":"activates"},{"from":"dot1l","to":"chrom","type":"activates"},{"from":"chrom","to":"hox","type":"activates"},{"from":"hox","to":"out","type":"activates"},{"from":"flt3","to":"out","type":"activates"}],"interventions":["Revumenib for relapsed or refractory KMT2A-rearranged acute leukaemia (approved 2024) and NPM1-mutant AML","Ziftomenib for relapsed or refractory NPM1-mutant AML (approved 2025)","Combinations with venetoclax plus azacitidine and with FLT3 inhibitors in front-line trials","Differentiation syndrome and QT prolongation need monitoring; MEN1 pocket mutations cause resistance","DOT1L inhibition (pinometostat) showed limited single-agent activity"]},{"id":"microbiome-tumour","kind":"pathway","name":"Microbiome-tumour interactions","aka":[],"tldr":"The 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.","summary":"Gut microbiota modulate systemic immunity (Bifidobacterium, Akkermansia, Faecalibacterium associated with PD-1 response; antibiotics associated with worse outcomes); faecal microbiota transplant from responders converted a fraction of refractory melanoma patients to responders (2021, and 2023-25 follow-ups). Intratumoural bacteria (Fusobacterium nucleatum in colorectal cancer, Gammaproteobacteria degrading gemcitabine in pancreatic cancer) affect chemoresistance and inflammation; colibactin-producing pks+ E. coli leaves a mutational signature in CRC. 'Polymorphic microbiomes' is a 2022 hallmark. Defined consortia (VE800, SER-155) and diet interventions are in trials; causality beyond melanoma is not settled.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Microbiome","links":[{"label":"Davar et al., Fecal microbiota transplant overcomes resistance to anti-PD-1 therapy in melanoma (Science 2021)","url":"https://doi.org/10.1126/science.abf3363"},{"label":"Hanahan, Hallmarks of Cancer: New Dimensions (Cancer Discovery 2022)","url":"https://doi.org/10.1158/2159-8290.CD-21-1059"}],"tags":["mechanism"],"related":[],"cancers":["melanoma","colorectal","pancreatic"],"sections":[],"technologies":[],"targets":["pd1"],"drugs":[],"companies":[],"institutions":["md-anderson","gustave-roussy","penn-abramson","johns-hopkins"],"pathways":["pd1-checkpoint","inflammation-nfkb","tumor-microenvironment"],"terms":["polymorphic-microbiomes","mutational-signature"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-davar-science"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Wargo (MD Anderson); Zitvogel (Gustave Roussy); Gajewski (Chicago); Sears (Johns Hopkins) on colibactin; Straussman (Weizmann) on intratumoural bacteria."],"analogy":"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.","nodes":[{"id":"gut","label":"Gut microbiota","x":15,"y":40},{"id":"meta","label":"Metabolites (SCFA, inosine, bile acids)","x":40,"y":20},{"id":"dc","label":"Dendritic / T-cell priming","x":65,"y":20},{"id":"io","label":"Immunotherapy response","x":90,"y":40,"targetId":"pd1"},{"id":"intra","label":"Intratumoural bacteria","x":40,"y":70},{"id":"chemo","label":"Chemotherapy degradation, inflammation","x":65,"y":70},{"id":"abx","label":"Antibiotics","x":15,"y":80},{"id":"colib","label":"Colibactin → mutational signature","x":90,"y":80}],"edges":[{"from":"gut","to":"meta","type":"activates"},{"from":"meta","to":"dc","type":"activates"},{"from":"dc","to":"io","type":"activates"},{"from":"abx","to":"gut","type":"inhibits"},{"from":"intra","to":"chemo","type":"activates"},{"from":"intra","to":"colib","type":"activates"},{"from":"gut","to":"intra","type":"activates"}],"interventions":["FMT from responders with PD-1 blockade (phase 2)","Defined bacterial consortia (VE800, SER-155) and diet (fibre) trials","Antibiotic stewardship around immunotherapy","Fusobacterium-targeted strategies (research)"]},{"id":"micrornas-in-cancer","kind":"pathway","name":"MicroRNAs in cancer","aka":["KEGG hsa05206","MicroRNAs in cancer"],"tldr":"This KEGG map collects the small RNA molecules (microRNAs) that are switched up or down in nine common cancers and shows which oncogenes and tumour suppressors they silence. It matters because a single microRNA can dial down dozens of genes at once, so losing or gaining one reshapes whole signalling routes.","summary":"MicroRNAs (miRNAs) are non-coding RNAs of about 21 to 23 nucleotides that control gene expression after transcription, either by degrading target messenger RNAs or by blocking their translation. KEGG map hsa05206 summarises profiling studies across nine cancers and draws two kinds of event: over-expressed miRNAs that repress tumour suppressors, and under-expressed miRNAs that release oncogenes. Both directions push proliferation, block differentiation and blunt apoptosis. KEGG notes that reported miRNA signatures differ between studies because of differences in populations and methods, so the map is a summary rather than a consensus.\n\nCalin and Croce, Nature Reviews Cancer, 2006 (doi:10.1038/nrc1997) set out the founding evidence: miRNA genes sit at fragile sites and cancer-associated regions of the genome, the miR-15a/16-1 cluster is deleted in chronic lymphocytic leukaemia and represses BCL2, the let-7 family restrains RAS, and miR-21 and the miR-17-92 cluster are over-expressed oncogenic miRNAs. The map connects these to familiar nodes: BCL2, KRAS, PTEN (a target of miR-21), MYC (which drives miR-17-92), TP53 (which induces the miR-34 family) and E2F. Epigenetic silencing of miRNA promoters by DNA methylation and by EZH2-containing polycomb complexes is one route to miRNA loss, which is why KEGG lists the demethylating agent decitabine and the EZH2 inhibitors tazemetostat and valemetostat against this map.\n\nWhat can be done: there is not yet an approved drug that directly replaces or blocks a miRNA in cancer. The practical levers today are indirect. Epigenetic drugs (decitabine, azacitidine, EZH2 inhibitors) can re-express silenced miRNAs, while drugs against the protein nodes the miRNAs control (BCL2 inhibitors, KRAS inhibitors, PI3K/AKT inhibitors) act downstream of the miRNA change. Antisense oligonucleotides against oncogenic miRNAs, such as the miR-155 inhibitor cobomarsen listed by KEGG, have reached early clinical trials.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05206","url":"https://www.kegg.jp/pathway/hsa05206"},{"label":"Review: MicroRNA signatures in human cancers","url":"https://doi.org/10.1038/nrc1997"}],"tags":[],"related":["ras-mapk","pi3k-akt-mtor","p53-cell-cycle","apoptosis-bcl2"],"cancers":[],"sections":[],"technologies":[],"targets":["ezh2","tp53","kras","bcl2"],"drugs":["decitabine","azacitidine","tazemetostat","sotorasib","adagrasib","alpelisib","capivasertib"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","p53-cell-cycle","apoptosis-bcl2"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-calin-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Think of each gene as a light in a house and microRNAs as dimmer switches wired to many lights at once. Cancer either jams a dimmer to full (silencing a whole bank of safety lights) or rips one out (so oncogene lights blaze). Because each dimmer controls many lights, fixing one switch can change the whole room.","nodes":[{"id":"methyl","label":"Promoter methylation / EZH2","x":15,"y":8,"targetId":"ezh2"},{"id":"myc","label":"MYC","x":60,"y":8},{"id":"mir17","label":"miR-17-92 cluster","x":60,"y":28},{"id":"mir21","label":"miR-21","x":85,"y":28},{"id":"let7","label":"let-7 family","x":15,"y":45},{"id":"mir15","label":"miR-15a / miR-16-1","x":40,"y":45},{"id":"p53","label":"TP53","x":60,"y":62,"targetId":"tp53"},{"id":"mir34","label":"miR-34 family","x":60,"y":80},{"id":"kras","label":"KRAS","x":15,"y":70,"targetId":"kras"},{"id":"bcl2","label":"BCL2","x":40,"y":70,"targetId":"bcl2"},{"id":"pten","label":"PTEN","x":85,"y":55},{"id":"out","label":"Proliferation, survival","x":50,"y":96}],"edges":[{"from":"methyl","to":"let7","type":"inhibits"},{"from":"methyl","to":"mir34","type":"inhibits"},{"from":"myc","to":"mir17","type":"activates"},{"from":"mir17","to":"out","type":"activates"},{"from":"mir21","to":"pten","type":"inhibits"},{"from":"let7","to":"kras","type":"inhibits"},{"from":"mir15","to":"bcl2","type":"inhibits"},{"from":"p53","to":"mir34","type":"activates"},{"from":"mir34","to":"bcl2","type":"inhibits"},{"from":"mir34","to":"out","type":"inhibits"},{"from":"kras","to":"out","type":"activates"},{"from":"bcl2","to":"out","type":"activates"},{"from":"pten","to":"out","type":"inhibits"}],"interventions":["Epigenetic drugs that re-express silenced microRNAs: hypomethylating agents (decitabine, azacitidine) and EZH2 inhibitors (tazemetostat, valemetostat)","Drugs against the protein nodes microRNAs control: BCL2 inhibitors, KRAS inhibitors (sotorasib, adagrasib), PI3K/AKT inhibitors (alpelisib, capivasertib)","Antisense oligonucleotides against oncogenic microRNAs (miR-155 inhibitor cobomarsen), early clinical trials only","MicroRNA signatures as diagnostic and prognostic biomarkers rather than drug targets"]},{"id":"mismatch-repair-msi","kind":"pathway","name":"Mismatch repair & microsatellite instability","aka":[],"tldr":"After 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.","summary":"MutSα (MSH2-MSH6) recognises base mismatches and small insertion-deletion loops, MutSβ (MSH2-MSH3) larger loops; MutLα (MLH1-PMS2) is recruited and nicks the new strand, EXO1 excises, Pol δ resynthesises, LIG1 seals. Loss of MLH1 (usually by promoter hypermethylation in sporadic colorectal and endometrial cancer, often with BRAF V600E), MSH2, MSH6 or PMS2 (germline in Lynch syndrome, or EPCAM deletion silencing MSH2) produces microsatellite instability (MSI-H), a hypermutator phenotype with 10-100x more mutations, frameshift neoantigens, and the SBS6/15/26 signatures. Clinically MSI-H/dMMR is tumour-agnostic for pembrolizumab and dostarlimab; dMMR rectal cancer can be cured with dostarlimab alone (complete responses in >90% of patients in the MSK study); neoadjuvant nivolumab-ipilimumab gives near-universal pathological responses in dMMR colon cancer. MMR loss also confers tolerance to temozolomide and thiopurines (the lesions are no longer recognised). Synthetic lethality: MSI-H cells depend on the WRN helicase; WRN inhibitors are in trials.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/DNA_mismatch_repair","links":[{"label":"Le et al., Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade (Science 2017)","url":"https://doi.org/10.1126/science.aan6733"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["colorectal","endometrial","gastric","tnbc","pancreatic","gallbladder"],"sections":[],"technologies":["checkpoint-inhibitor","cgp","germline-testing","histopathology-ihc","colorectal-screening","aspirin-cancer-prevention","synthetic-lethality-approaches"],"targets":["pd1","ctla4","wrn","pdl1"],"drugs":["pembrolizumab","dostarlimab","nivolumab","ipilimumab"],"companies":[],"institutions":[],"pathways":["mutagenesis-signatures","antigen-presentation-immunoediting","pd1-checkpoint","synthetic-lethality-map"],"terms":["msi","lynch-syndrome","tmb","neoantigen","tumour-agnostic","mutational-signature","endometrial-molecular-classes","cms-subtypes"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"err","label":"Replication mismatch","x":12,"y":15},{"id":"muts","label":"MutSα (MSH2-MSH6)","x":40,"y":15},{"id":"mutl","label":"MutLα (MLH1-PMS2)","x":68,"y":15},{"id":"exo","label":"EXO1, Pol δ resynthesis","x":92,"y":30},{"id":"fixed","label":"Corrected DNA","x":92,"y":60},{"id":"loss","label":"MMR loss (Lynch, MLH1 methylation)","x":40,"y":45},{"id":"msi","label":"MSI-H, hypermutation","x":40,"y":72},{"id":"neo","label":"Frameshift neoantigens","x":68,"y":72},{"id":"io","label":"Checkpoint-inhibitor response","x":68,"y":94,"targetId":"pd1"},{"id":"wrn","label":"WRN dependence","x":12,"y":72,"targetId":"wrn"}],"edges":[{"from":"err","to":"muts","type":"activates"},{"from":"muts","to":"mutl","type":"activates"},{"from":"mutl","to":"exo","type":"activates"},{"from":"exo","to":"fixed","type":"activates"},{"from":"loss","to":"muts","type":"inhibits"},{"from":"loss","to":"mutl","type":"inhibits"},{"from":"loss","to":"msi","type":"activates"},{"from":"msi","to":"neo","type":"activates"},{"from":"neo","to":"io","type":"activates"},{"from":"msi","to":"wrn","type":"activates"}],"interventions":["Tumour-agnostic pembrolizumab and dostarlimab for MSI-H/dMMR; nivolumab ± ipilimumab in colorectal cancer","Organ-sparing: dostarlimab alone cures most dMMR rectal cancers; neoadjuvant nivolumab-ipilimumab in dMMR colon cancer","Universal MMR/MSI testing of colorectal and endometrial cancer finds Lynch syndrome; colonoscopic surveillance and aspirin for carriers","WRN helicase inhibitors as synthetic-lethal therapy for MSI-H tumours (trials)"]},{"id":"mitotic-spindle-checkpoint","kind":"pathway","name":"Mitosis & the spindle assembly checkpoint","aka":[],"tldr":"When 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.","summary":"In prometaphase, unattached kinetochores recruit MPS1, MAD1/MAD2, BUB1/BUBR1 to assemble the mitotic checkpoint complex, which inhibits APC/C-CDC20, keeping securin and cyclin B intact so anaphase cannot start. Once all kinetochores are attached under tension (Aurora B corrects mis-attachments), the checkpoint silences, separase cleaves cohesin, and chromosomes segregate; PLK1 and Aurora A drive spindle assembly and centrosome maturation, KIF18A tames chromosome oscillation. Microtubule drugs (taxanes, epothilones, eribulin stabilise or sequester; vinca alkaloids and ADC payloads MMAE/DM1 destabilise) cause persistent checkpoint activation; cells either die in mitosis (BCL-XL degradation, MCL-1 loss) or 'slip' into a tetraploid G1 where p53 arrests them, which is why TP53 loss tolerates slippage and drives aneuploidy. Whole-genome doubling and chromosomal instability arise here; CIN-high tumours depend on KIF18A, the checkpoint and BCL-XL. Aurora, PLK1 and MPS1 inhibitors have narrow windows; alisertib is being retested in SCLC and neuroendocrine prostate cancer.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Spindle_checkpoint","links":[{"label":"Musacchio & Salmon, The spindle-assembly checkpoint in space and time (Nat Rev Mol Cell Biol 2007)","url":"https://doi.org/10.1038/nrm2163"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["cytotoxic-chemotherapy","adc"],"targets":["her2","nectin4","cd30","folr1","bcl2","tp53"],"drugs":["paclitaxel","docetaxel","cabazitaxel","eribulin","vincristine","vinblastine","vinorelbine","trastuzumab-emtansine","enfortumab-vedotin","brentuximab-vedotin","mirvetuximab-soravtansine"],"companies":[],"institutions":[],"pathways":["chromosomal-instability","cell-cycle-engine-cdks","p53-cell-cycle","apoptosis-bcl2"],"terms":["tubulin-inhibitor-payloads","mmae","dm1","dm4","whole-genome-doubling"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-musacchio-nat-rev-mol-cell-biol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"cent","label":"Centrosomes (Aurora A, PLK1)","x":12,"y":15},{"id":"spin","label":"Spindle microtubules","x":40,"y":15},{"id":"kin","label":"Kinetochore attachment","x":68,"y":15},{"id":"tax","label":"Taxanes, vincas, MMAE, DM1","x":40,"y":45,"targetId":"her2"},{"id":"sac","label":"SAC: MAD2, BUBR1, MPS1","x":68,"y":45},{"id":"apc","label":"APC/C-CDC20","x":90,"y":45},{"id":"ana","label":"Anaphase (separase)","x":90,"y":78},{"id":"death","label":"Mitotic death (BCL-XL↓)","x":60,"y":78,"targetId":"bcl2"},{"id":"slip","label":"Slippage → tetraploid","x":30,"y":78},{"id":"p53","label":"p53 arrest","x":8,"y":78,"targetId":"tp53"}],"edges":[{"from":"cent","to":"spin","type":"activates"},{"from":"spin","to":"kin","type":"activates"},{"from":"tax","to":"spin","type":"inhibits"},{"from":"kin","to":"sac","type":"inhibits"},{"from":"sac","to":"apc","type":"inhibits"},{"from":"apc","to":"ana","type":"activates"},{"from":"sac","to":"death","type":"activates"},{"from":"sac","to":"slip","type":"activates"},{"from":"p53","to":"slip","type":"inhibits"}],"interventions":["Taxanes (paclitaxel, docetaxel, cabazitaxel), eribulin, vinca alkaloids (vincristine, vinblastine, vinorelbine)","Tubulin-payload ADCs: MMAE (enfortumab, brentuximab, polatuzumab vedotin), DM1 (T-DM1), DM4 (mirvetuximab)","Aurora A (alisertib), PLK1 and MPS1 inhibitors; KIF18A inhibitors for CIN-high tumours (trials)","BCL-XL degraders to tip arrested cells into death"]},{"id":"mrna-translation-eif4f","kind":"pathway","name":"mRNA translation (eIF4F / mTOR)","aka":[],"tldr":"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.","summary":"mTORC1 phosphorylates 4E-BPs, releasing eIF4E to form eIF4F (eIF4E/4G/4A) and translate cap-dependent, oncogene-rich mRNAs (MYC, cyclin D1, BCL-2, MCL-1); MNK kinases phosphorylate eIF4E. Ribosomal protein and rRNA synthesis (POLR1, RNA pol I inhibitors such as CX-5461) are collateral dependencies. Drugs: eIF4A inhibitors (zotatifin, phase 1/2 in ER+ breast and KRAS-mutant cancers), MNK inhibitors (tomivosertib), mTOR inhibitors (everolimus approved), and RNA pol I inhibitors in BRCA/HRD (CX-5461 with PARP). Translation control is a shared node downstream of PI3K, RAS, and MYC.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/EIF4E","links":[{"label":"Fabbri, Chakraborty, Robert & Vagner, The plasticity of mRNA translation during cancer progression and therapy resistance (Nature Reviews Cancer 2021)","url":"https://doi.org/10.1038/s41568-021-00380-y"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":[],"targets":["akt","pik3ca"],"drugs":["everolimus"],"companies":[],"institutions":["mgh","mskcc","ucsf"],"pathways":["pi3k-akt-mtor","myc","ras-mapk"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-fabbri-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Sonenberg (McGill); Ruggero (UCSF) on translation control in cancer; Pelletier (McGill) on eIF4A; eFFECTOR Therapeutics."],"analogy":"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.","nodes":[{"id":"mtor","label":"mTORC1","x":15,"y":30,"targetId":"akt"},{"id":"4ebp","label":"4E-BP","x":40,"y":30},{"id":"eif4e","label":"eIF4E","x":62,"y":30},{"id":"eif4f","label":"eIF4F (4E/4G/4A)","x":82,"y":30},{"id":"mnk","label":"MNK1/2","x":62,"y":65},{"id":"onc","label":"MYC, cyclin D1, MCL-1 translation","x":82,"y":70},{"id":"ribo","label":"Ribosome biogenesis (Pol I)","x":15,"y":70}],"edges":[{"from":"mtor","to":"4ebp","type":"inhibits"},{"from":"4ebp","to":"eif4e","type":"inhibits"},{"from":"eif4e","to":"eif4f","type":"activates"},{"from":"eif4f","to":"onc","type":"activates"},{"from":"mnk","to":"eif4e","type":"activates"},{"from":"mtor","to":"ribo","type":"activates"},{"from":"ribo","to":"onc","type":"activates"}],"interventions":["mTOR inhibitors (everolimus) approved in breast, RCC, NET","eIF4A inhibitor zotatifin (eFT226) phase 1/2 with fulvestrant/abemaciclib","MNK inhibitors (tomivosertib) in NSCLC (mixed)","RNA Pol I inhibitors (CX-5461, pidnarulex) in HRD cancers"]},{"id":"mutagenesis-signatures","kind":"pathway","name":"Mutagenesis & mutational signatures","aka":[],"tldr":"Every 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.","summary":"Mutations arise when damage (exogenous: UV, tobacco polycyclics, aflatoxin, alkylators, platinum; endogenous: deamination, APOBEC3A/B cytidine deaminases, ROS, replication errors, colibactin) meets replication before repair, or when repair itself is defective (MMR loss → SBS6/15/26 and MSI; HRD → SBS3 and indel/rearrangement patterns; POLE exonuclease mutation → ultramutation). COSMIC catalogues >60 single-base substitution signatures, plus doublet, indel and copy-number signatures. Clock-like SBS1/SBS5 accumulate with age; APOBEC (SBS2/13) is episodic and therapy-associated; SBS31/35 record prior platinum; temozolomide leaves SBS11. Signatures are now clinical: HRD scores (SBS3, LOH, TAI, LST) select PARP inhibitors; MSI and TMB select immunotherapy; APOBEC activity predicts resistance evolution.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Mutational_signatures","links":[{"label":"Alexandrov et al., The repertoire of mutational signatures in human cancer (Nature 2020)","url":"https://doi.org/10.1038/s41586-020-1943-3"},{"label":"COSMIC mutational signatures","url":"https://cancer.sanger.ac.uk/signatures/"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["tnbc"],"sections":[],"technologies":["wes-wgs","hrd-testing","cgp","smoking-cessation-after-diagnosis","hpv-vaccine"],"targets":["brca","parp","pd1"],"drugs":["olaparib","pembrolizumab","dostarlimab","temozolomide"],"companies":[],"institutions":[],"pathways":["ddr","mismatch-repair-msi","homologous-recombination-repair","clonal-evolution"],"terms":["mutational-signature","tmb","msi","hrd","mgmt","neoantigen","genome-instability-mutation"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-alexandrov-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"exo","label":"Exogenous: UV, tobacco","x":12,"y":12},{"id":"endo","label":"Endogenous: APOBEC, ROS","x":40,"y":12},{"id":"rep","label":"Replication errors","x":70,"y":12},{"id":"dmg","label":"DNA lesions","x":40,"y":42},{"id":"repair","label":"Repair: MMR, HR, BER, NER","x":78,"y":42,"targetId":"brca"},{"id":"fix","label":"Fixed mutations","x":40,"y":70},{"id":"sig","label":"Signature (SBS, ID, CN)","x":12,"y":70},{"id":"bio","label":"HRD, MSI, TMB biomarkers","x":78,"y":70},{"id":"drv","label":"Drivers, neoantigens","x":40,"y":94}],"edges":[{"from":"exo","to":"dmg","type":"activates"},{"from":"endo","to":"dmg","type":"activates"},{"from":"rep","to":"dmg","type":"activates"},{"from":"repair","to":"dmg","type":"inhibits"},{"from":"dmg","to":"fix","type":"activates"},{"from":"fix","to":"sig","type":"activates"},{"from":"repair","to":"bio","type":"inhibits"},{"from":"fix","to":"bio","type":"activates"},{"from":"fix","to":"drv","type":"activates"}],"interventions":["Prevention removes the mutagen: smoking cessation, UV protection, HPV/HBV vaccination, aflatoxin control","HRD signatures select PARP inhibitors and platinum; MSI/TMB select checkpoint inhibitors","Signature-aware design: avoid TMZ in MGMT-unmethylated tumours, expect APOBEC-driven resistance","Whole-genome sequencing and methylation profiling read the fingerprints"]},{"id":"idh-2hg","kind":"pathway","name":"Mutant IDH / 2-hydroxyglutarate","aka":[],"tldr":"A single mutation in a metabolic enzyme (IDH1 or IDH2) makes cells pour out a molecule they should never make, 2-hydroxyglutarate. It jams the machinery that erases chemical marks on DNA and histones, so blood and brain cells get stuck before they mature. Pills that block the mutant enzyme let them mature again.","summary":"Wild-type IDH1 (cytoplasm) and IDH2 (mitochondria) convert isocitrate to α-ketoglutarate. Hotspot mutations (IDH1 R132, IDH2 R140/R172) give the enzyme a new activity: reducing α-ketoglutarate to the oncometabolite (R)-2-hydroxyglutarate, which reaches millimolar levels. 2-HG competitively inhibits α-ketoglutarate-dependent dioxygenases: the TET family of DNA demethylases (producing the CpG island hypermethylator phenotype, G-CIMP, in glioma and a TET2-like phenotype in AML), the JmjC histone demethylases and prolyl hydroxylases. The result is a block in differentiation rather than fast proliferation, which is why IDH-mutant gliomas are slow-growing and why the drugs work by differentiation. Mutant-selective inhibitors ivosidenib (IDH1), olutasidenib (IDH1), enasidenib (IDH2) and the brain-penetrant dual inhibitor vorasidenib are approved: ivosidenib in AML, cholangiocarcinoma and MDS; vorasidenib for grade 2 IDH-mutant glioma after the INDIGO trial. Differentiation syndrome is the class toxicity; resistance arises through second-site IDH mutations and switching from IDH1 to IDH2 (or back).","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Isocitrate_dehydrogenase","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Isocitrate_dehydrogenase"}],"tags":[],"related":[],"cancers":["aml","glioblastoma","cholangiocarcinoma","mds"],"sections":[],"technologies":[],"targets":["idh"],"drugs":["ivosidenib","olutasidenib","enasidenib","vorasidenib"],"companies":[],"institutions":[],"pathways":["epigenetic-reprogramming"],"terms":[],"trials":["indigo"],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"A factory machine (IDH) that is supposed to make one part is bent so it makes a rogue part (2-HG) that gums up the erasers (TET, histone demethylases) which normally wipe old instructions off the blueprint. Cells cannot read the 'grow up' instructions and stay immature. The inhibitors straighten the machine; the erasers work again and the cells mature.","nodes":[{"id":"idh","label":"Mutant IDH1 / IDH2 (R132, R140, R172)","x":50,"y":5,"targetId":"idh"},{"id":"akg","label":"α-ketoglutarate","x":22,"y":24},{"id":"hg","label":"2-hydroxyglutarate (oncometabolite)","x":50,"y":30},{"id":"tet","label":"TET2 DNA demethylase","x":22,"y":54},{"id":"kdm","label":"JmjC histone demethylases","x":78,"y":54},{"id":"meth","label":"DNA and histone hypermethylation (G-CIMP)","x":50,"y":74},{"id":"block","label":"Differentiation block → leukaemia, glioma","x":50,"y":95}],"edges":[{"from":"idh","to":"hg","type":"activates"},{"from":"akg","to":"hg","type":"activates"},{"from":"hg","to":"tet","type":"inhibits"},{"from":"hg","to":"kdm","type":"inhibits"},{"from":"tet","to":"meth","type":"inhibits"},{"from":"kdm","to":"meth","type":"inhibits"},{"from":"meth","to":"block","type":"activates"}],"interventions":["Ivosidenib (IDH1) for relapsed AML, first-line AML with azacitidine (AGILE), IDH1-mutant cholangiocarcinoma and MDS","Olutasidenib (IDH1) and enasidenib (IDH2) for relapsed or refractory AML","Vorasidenib (dual IDH1/2, brain-penetrant) for residual or recurrent grade 2 IDH-mutant astrocytoma and oligodendroglioma (INDIGO)","Differentiation syndrome is the class toxicity; steroids and hydroxyurea manage it","Combination with venetoclax and azacitidine, and with PARP inhibitors (2-HG induces a homologous-recombination defect), under study"]},{"id":"myc","kind":"pathway","name":"MYC","aka":[],"tldr":"MYC 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.","summary":"MYC family transcription factors (MYC, MYCN, MYCL) dimerise with MAX to drive ribosome biogenesis, metabolism, and proliferation; deregulated in >50% of cancers via amplification, translocation (Burkitt), or upstream signalling (Wnt, RAS, Notch). MYC also suppresses immune recognition (CD47, PD-L1). Direct approaches: OMO-103 (Omomyc mini-protein, phase 1/2 in PDAC), MYC degraders, and MAX-stabilising molecules; indirect: CDK9 and BET inhibitors that reduce MYC transcription, PLK1 and AURKA inhibitors that destabilise MYC/MYCN protein (neuroblastoma), and synthetic-lethal dependencies (mTOR, splicing).","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Myc","links":[{"label":"Dhanasekaran et al., The MYC oncogene: the grand orchestrator of cancer growth and immune evasion (Nature Reviews Clinical Oncology 2022)","url":"https://doi.org/10.1038/s41571-021-00549-2"},{"label":"OMO-103 first-in-human (Nature Medicine 2024)","url":"https://doi.org/10.1038/s41591-024-02875-1"}],"tags":["mechanism"],"related":[],"cancers":["neuroblastoma","dlbcl","pancreatic","tnbc","non-hodgkin-lymphoma"],"sections":[],"technologies":[],"targets":["cd47","kras"],"drugs":[],"companies":[],"institutions":["vall-dhebron","ucsf","dana-farber","cold-spring-harbor"],"pathways":["wnt","ras-mapk","notch","cancer-metabolism","mrna-translation-eif4f","rna-splicing"],"terms":["mycn-amplification","sustaining-proliferative-signaling"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-dhanasekaran-nat-rev-clin-oncol","paper-brown-nat-med"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Soucek (VHIO, Omomyc/Peptomyc); Felsher (Stanford, MYC addiction); Eilers (Würzburg); Dang (Ludwig); Bradner/Vakoc (CSHL, BET inhibitors)."],"analogy":"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.","nodes":[{"id":"up","label":"Wnt, RAS, Notch, amplification","x":12,"y":40},{"id":"myc","label":"MYC/MAX","x":40,"y":40},{"id":"ribo","label":"Ribosome biogenesis, metabolism","x":70,"y":20},{"id":"prol","label":"Proliferation","x":90,"y":40},{"id":"imm","label":"CD47, PD-L1 (immune evasion)","x":70,"y":70,"targetId":"cd47"},{"id":"cdk9","label":"CDK9 / BET (transcription)","x":40,"y":80},{"id":"aurka","label":"AURKA / PLK1 (stability)","x":12,"y":80}],"edges":[{"from":"up","to":"myc","type":"activates"},{"from":"myc","to":"ribo","type":"activates"},{"from":"ribo","to":"prol","type":"activates"},{"from":"myc","to":"imm","type":"activates"},{"from":"cdk9","to":"myc","type":"activates"},{"from":"aurka","to":"myc","type":"activates"}],"interventions":["OMO-103 (Omomyc) phase 1/2 in pancreatic cancer","BET and CDK9 inhibitors reduce MYC transcription (haematologic trials)","AURKA inhibitors in MYCN-amplified neuroblastoma","MYC-driven dependencies: mTOR, spliceosome, glutamine"]},{"id":"myeloid-suppression-axis","kind":"pathway","name":"Myeloid suppression: TAMs, MDSCs & don't-eat-me signals","aka":[],"tldr":"Tumours 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.","summary":"Tumour and stromal CSF1, CCL2, CXCL1/2/8, VEGF, G-CSF and IL-6 recruit monocytes and neutrophils and expand myeloid-derived suppressor cells (PMN- and M-MDSC) from the bone marrow. In the tumour, hypoxia, lactate, IL-4/IL-13 and IL-10 polarise macrophages toward an immunosuppressive TAM state (TREM2+, SPP1+, MRC1+) that secretes IL-10, TGF-β, arginase, PGE2, expresses PD-L1, SIRPα and VISTA, and promotes angiogenesis and metastasis (TMEM doorways). MDSCs suppress via arginase, iNOS, ROS and PD-L1. CD47 on tumour cells binds SIRPα to block phagocytosis; magrolimab (anti-CD47) failed in AML/MDS after promising phase 1 data, but CD47 remains pursued with bispecifics. CSF1R inhibition (pexidartinib, vimseltinib in tenosynovial giant cell tumour) depletes TAMs but compensatory MDSC influx blunts it in cancer; CD40 agonists and TLR agonists reprogram; CXCR2 and CCR2 blockade limit recruitment; STAT3 and PI3Kγ inhibitors reprogram. Trained immunity (BCG) and engineered CAR-macrophages try to conscript the same cells for the attack.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Tumor-associated_macrophage","links":[{"label":"DeNardo & Ruffell, Macrophages as regulators of tumour immunity and immunotherapy (Nat Rev Immunol 2019)","url":"https://doi.org/10.1038/s41577-019-0127-6"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["car-nk-macrophage","trained-innate-immunity","bcg-and-intravesical-therapy","checkpoint-inhibitor","monoclonal-antibody"],"targets":["csf1r","cd47","pdl1","vegf","vista","cxcr4"],"drugs":["pexidartinib","vimseltinib","magrolimab","bcg-intravesical"],"companies":[],"institutions":[],"pathways":["tumor-microenvironment","nutrient-competition-tme","angiogenic-switch","myc","jak-stat"],"terms":["tumor-associated-macrophages","myeloid-derived-suppressor-cells","fc-effector","immune-exclusion"],"trials":[],"people":[],"bottlenecks":["b-tme-immunosuppression"],"keyPapers":["paper-denardo-nat-rev-immunol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"rec","label":"CSF1, CCL2, G-CSF, VEGF","x":12,"y":15,"targetId":"vegf"},{"id":"mono","label":"Monocytes, neutrophils","x":40,"y":15},{"id":"mdsc","label":"MDSCs (arginase, ROS)","x":40,"y":45},{"id":"tam","label":"TAMs (CSF1R, TREM2)","x":68,"y":15,"targetId":"csf1r"},{"id":"sup","label":"IL-10, TGF-β, PD-L1","x":68,"y":45,"targetId":"pdl1"},{"id":"tcell","label":"T cells suppressed","x":92,"y":62},{"id":"cd47","label":"CD47 → SIRPα 'don't eat'","x":40,"y":78,"targetId":"cd47"},{"id":"phag","label":"Phagocytosis (ADCP)","x":68,"y":82},{"id":"ang","label":"Angiogenesis, metastasis","x":92,"y":20},{"id":"cd40","label":"CD40 agonists, TLRs reprogram","x":12,"y":78}],"edges":[{"from":"rec","to":"mono","type":"activates"},{"from":"mono","to":"mdsc","type":"activates"},{"from":"mono","to":"tam","type":"activates"},{"from":"tam","to":"sup","type":"activates"},{"from":"mdsc","to":"sup","type":"activates"},{"from":"sup","to":"tcell","type":"inhibits"},{"from":"tam","to":"ang","type":"activates"},{"from":"cd47","to":"phag","type":"inhibits"},{"from":"tam","to":"phag","type":"activates"},{"from":"cd40","to":"sup","type":"inhibits"},{"from":"cd40","to":"phag","type":"activates"}],"interventions":["CSF1R inhibitors pexidartinib and vimseltinib (approved in tenosynovial giant cell tumour; disappointing in cancer)","CD47/SIRPα: magrolimab failed in AML/MDS; next-generation bispecifics and SIRPα-Fc in trials","CD40 agonists, TLR7/9 agonists, CXCR2/CCR2 blockade, PI3Kγ inhibitors as reprogrammers","CAR-macrophages, trained innate immunity (BCG) and antibody engineering for ADCP"]},{"id":"nk-cell-recognition","kind":"pathway","name":"NK-cell recognition: missing self & stress ligands","aka":[],"tldr":"Natural 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.","summary":"NK activation integrates inhibitory signals from KIRs and NKG2A (binding classical HLA-A/B/C and HLA-E respectively; 'missing self' when MHC-I is lost) against activating signals from NKG2D (MICA/B, ULBP1-6, induced by DNA damage and oncogenes), DNAM-1 (CD155, CD112), NKp30 (B7-H6), NKp46, and CD16 (FcγRIIIa, mediating ADCC by IgG1 antibodies such as trastuzumab and cetuximab). Killing is by perforin/granzyme and death ligands; IFN-γ recruits and licenses the adaptive response. Tumour escape: proteolytic shedding of MICA/B (ADAM10/17), HLA-E upregulation engaging NKG2A (monalizumab, mixed results), TIGIT and PVRIG competing with DNAM-1 for CD155, TGF-β and adenosine downregulating NKG2D, platelet cloaking of CTCs. NK cells are central to clearing circulating tumour cells and dormant cells and to controlling MHC-I-negative escape variants after checkpoint or CAR-T therapy. Therapeutics: CAR-NK (cord blood, iPSC-derived, off the shelf, low CRS), NK engagers (CD16×target), IL-15 superagonists (nogapendekin alfa approved in NMIBC), anti-NKG2A, antibody afucosylation to boost ADCC.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Natural_killer_cell","links":[{"label":"Morvan & Lanier, NK cells and cancer: you can teach innate cells new tricks (Nat Rev Cancer 2016)","url":"https://doi.org/10.1038/nrc.2015.5"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["car-nk-macrophage","monoclonal-antibody","cytokine-therapy","allogeneic-cell-therapy","bispecific-antibody"],"targets":["tigit","her2","egfr","cd20","cd38","pd1"],"drugs":["trastuzumab","cetuximab","rituximab","obinutuzumab","margetuximab","daratumumab","nogapendekin-alfa","tiragolumab"],"companies":[],"institutions":[],"pathways":["antigen-presentation-immunoediting","extrinsic-apoptosis-death-receptors","tgf-beta","intravasation-ctc-survival","tumor-dormancy"],"terms":["adcc","fc-effector","bcg-unresponsive","avoiding-immune-destruction"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-morvan-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"mhc","label":"MHC-I → KIR (inhibit)","x":12,"y":15},{"id":"hlae","label":"HLA-E → NKG2A (inhibit)","x":12,"y":45},{"id":"stress","label":"MICA/B, ULBP → NKG2D","x":12,"y":75},{"id":"nk","label":"NK cell decision","x":45,"y":45},{"id":"cd16","label":"CD16 ← IgG1 antibody (ADCC)","x":45,"y":12,"targetId":"her2"},{"id":"tigit","label":"TIGIT vs DNAM-1 (CD155)","x":45,"y":80,"targetId":"tigit"},{"id":"kill","label":"Perforin, granzyme, IFN-γ","x":78,"y":30},{"id":"shed","label":"MICA shedding, TGF-β","x":78,"y":65},{"id":"carnk","label":"CAR-NK, IL-15, NK engagers","x":92,"y":90}],"edges":[{"from":"mhc","to":"nk","type":"inhibits"},{"from":"hlae","to":"nk","type":"inhibits"},{"from":"stress","to":"nk","type":"activates"},{"from":"cd16","to":"nk","type":"activates"},{"from":"tigit","to":"nk","type":"inhibits"},{"from":"nk","to":"kill","type":"activates"},{"from":"shed","to":"stress","type":"inhibits"},{"from":"shed","to":"nk","type":"inhibits"},{"from":"carnk","to":"nk","type":"activates"}],"interventions":["IgG1 antibodies (trastuzumab, cetuximab, rituximab) recruit NK ADCC; afucosylated antibodies (obinutuzumab, margetuximab) bind CD16 harder","IL-15 superagonist nogapendekin alfa (BCG-unresponsive NMIBC); CAR-NK and NK engagers in trials","Anti-NKG2A (monalizumab) and anti-TIGIT (tiragolumab) release inhibitory checks, with mixed phase 3 results","NK-based therapies address MHC-I-loss escape from T-cell therapies"]},{"id":"nsclc-signalling","kind":"pathway","name":"Non-small cell lung cancer (KEGG map)","aka":["KEGG hsa05223","Non-small cell lung cancer"],"tldr":"KEGG's non-small cell lung cancer map shows a set of alternative on-switches (EGFR mutation, KRAS mutation, EML4-ALK, RET and MET alterations) that all feed the same RAS/ERK, PI3K/AKT and STAT relays, plus loss of the p16 and p53 brakes. Each on-switch now has its own targeted pill, which is why molecular testing comes before treatment.","summary":"The KEGG non-small cell lung cancer map (hsa05223) is a map of mutually exclusive oncogenic drivers. Activating EGFR mutations (exon 19 deletions and L858R) or EGFR overexpression signal through GRB2/SOS to KRAS, through PLC-gamma and PKC to ERK, and through PI3K to AKT. KRAS point mutations (G12C among them) disable GTP hydrolysis so the protein continuously transmits growth signals to RAF, MEK and ERK. The EML4-ALK fusion, created by an inversion on chromosome 2p, gives constitutive ALK kinase activity feeding PLC-gamma to ERK, PI3K to AKT and JAK to STAT3/STAT5. RET fusions and MET mutations or amplification are further drivers feeding RAS to ERK and PI3K, and ERBB2 overexpression adds to the receptor input. On the suppressor side, loss of p16INK4a (CDKN2A) frees CDK4/6 and cyclin D to phosphorylate RB, TP53 mutation removes apoptosis and arrest, RASSF1A loss shifts RAS output toward growth, and RAR-beta silencing removes retinoid growth control. Herbst, Morgensztern and Boshoff, Nature, 2018 (doi:10.1038/nature25183) review how this driver map, together with PD-L1 expression and tumour mutational burden, now dictates first-line treatment, and how resistance emerges through on-target mutations (EGFR T790M, C797S; ALK G1202R), bypass through MET amplification, and histological transformation to small cell carcinoma.\n\nWhat drugs do about it: EGFR mutations are treated with osimertinib (alone or with chemotherapy) or amivantamab with lazertinib, and exon 20 insertions with amivantamab or sunvozertinib; KRAS G12C with sotorasib, adagrasib and the newer divarasib and olomorasib; ALK fusions with alectinib, brigatinib or lorlatinib; RET fusions with selpercatinib or pralsetinib; MET exon 14 skipping with capmatinib, tepotinib or savolitinib, and MET overexpression with telisotuzumab vedotin; BRAF V600E with dabrafenib plus trametinib; HER2 mutations with trastuzumab deruxtecan or zongertinib; and driver-negative tumours with pembrolizumab or other PD-1/PD-L1 antibodies, with or without chemotherapy.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05223","url":"https://www.kegg.jp/pathway/hsa05223"},{"label":"Review: The biology and management of non-small cell lung cancer","url":"https://doi.org/10.1038/nature25183"}],"tags":[],"related":["rtk-activation","ras-mapk"],"cancers":["nsclc"],"sections":[],"technologies":[],"targets":["egfr","alk","ret","met","kras","mek","akt","cdk4-6","tp53"],"drugs":["osimertinib","amivantamab","lazertinib","sunvozertinib","sotorasib","adagrasib","divarasib","olomorasib","alectinib","brigatinib","lorlatinib","selpercatinib","pralsetinib","savolitinib","telisotuzumab-vedotin","dabrafenib","trametinib","trastuzumab-deruxtecan","zongertinib","pembrolizumab"],"companies":[],"institutions":[],"pathways":["rtk-activation","ras-mapk","pi3k-akt-mtor","p53-cell-cycle","cell-cycle-engine-cdks","pd1-checkpoint","lineage-plasticity-neuroendocrine"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-herbst-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A room full of light switches wired to the same bulb. In each patient's tumour just one switch is jammed on (EGFR, KRAS, ALK, RET or MET), so the first job is to find which one, then fit a cover that matches that switch. If no switch is jammed, the immune system is unleashed on the bulb instead.","nodes":[{"id":"egfr","label":"EGFR (mutated) / ERBB2","x":12,"y":6,"targetId":"egfr"},{"id":"alk","label":"EML4-ALK fusion","x":37,"y":6,"targetId":"alk"},{"id":"ret","label":"RET fusion","x":62,"y":6,"targetId":"ret"},{"id":"met","label":"MET (mutated, amplified)","x":87,"y":6,"targetId":"met"},{"id":"kras","label":"KRAS (G12C and others)","x":25,"y":30,"targetId":"kras"},{"id":"mek","label":"RAF / MEK / ERK","x":25,"y":52,"targetId":"mek"},{"id":"pi3k","label":"PI3K / AKT","x":62,"y":30,"targetId":"akt"},{"id":"stat","label":"JAK / STAT3, STAT5","x":87,"y":30},{"id":"p16","label":"p16 loss to cyclin D / CDK4/6","x":62,"y":60,"targetId":"cdk4-6"},{"id":"p53","label":"TP53 (mutated)","x":87,"y":60,"targetId":"tp53"},{"id":"out","label":"Proliferation, survival, invasion","x":50,"y":92}],"edges":[{"from":"egfr","to":"kras","type":"activates"},{"from":"egfr","to":"pi3k","type":"activates"},{"from":"alk","to":"kras","type":"activates"},{"from":"alk","to":"pi3k","type":"activates"},{"from":"alk","to":"stat","type":"activates"},{"from":"ret","to":"kras","type":"activates"},{"from":"met","to":"kras","type":"activates"},{"from":"met","to":"pi3k","type":"activates"},{"from":"kras","to":"mek","type":"activates"},{"from":"mek","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"stat","to":"out","type":"activates"},{"from":"p16","to":"out","type":"activates"},{"from":"p53","to":"out","type":"inhibits"}],"interventions":["EGFR-mutant disease: osimertinib (alone or with chemotherapy), amivantamab plus lazertinib; amivantamab or sunvozertinib for exon 20 insertions","KRAS G12C: sotorasib, adagrasib, divarasib, olomorasib","ALK fusions: alectinib, brigatinib, lorlatinib; RET fusions: selpercatinib, pralsetinib; MET exon 14 or overexpression: savolitinib, telisotuzumab vedotin","BRAF V600E: dabrafenib plus trametinib; HER2 mutations: trastuzumab deruxtecan, zongertinib","Driver-negative tumours: pembrolizumab or other PD-1/PD-L1 antibodies with or without platinum chemotherapy"]},{"id":"notch","kind":"pathway","name":"Notch signalling","aka":[],"tldr":"A cell-to-cell contact signal that decides cell fate. It drives T-cell leukaemia when mutated on, acts as a tumour suppressor in some squamous cancers when lost, and its ligand DLL3 became a drug target in small-cell lung cancer.","summary":"Notch receptors are cleaved by γ-secretase on ligand binding (Delta-like, Jagged), releasing NICD to activate HES/HEY genes. Activating NOTCH1 mutations occur in >50% of T-ALL and in CLL (shorter survival); inactivating mutations in head and neck, oesophageal, and cutaneous squamous carcinomas mark a tumour-suppressor role. γ-secretase inhibitors failed as anticancer drugs (gut toxicity) but nirogacestat is approved in desmoid tumours (2023). DLL3, an inhibitory Notch ligand aberrantly on SCLC cell surfaces, is targeted by tarlatamab. Context dependence is the central lesson.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Notch_signaling_pathway","links":[{"label":"Aster, Pear & Blacklow, The varied roles of Notch in cancer (Annual Review of Pathology 2017)","url":"https://doi.org/10.1146/annurev-pathol-052016-100127"}],"tags":["mechanism"],"related":[],"cancers":["sclc","all-leukemia","cll","head-and-neck","tnbc"],"sections":[],"technologies":[],"targets":["dll3"],"drugs":["tarlatamab","nirogacestat"],"companies":[],"institutions":["dana-farber","penn-abramson"],"pathways":["cancer-stem-cells-plasticity","myc"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-aster-annu-rev-pathol"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Aster (Brigham/Harvard) on Notch in leukaemia; Pear (Penn); Radtke (EPFL)."],"analogy":"A doorbell that only works when a neighbour presses it. In some cancers the bell rings constantly (T-ALL); in others it has been ripped out so cells never hear 'stop and mature'.","nodes":[{"id":"lig","label":"DLL/JAG ligand (neighbour)","x":15,"y":40},{"id":"notch","label":"NOTCH receptor","x":40,"y":40},{"id":"gs","label":"γ-secretase cleavage","x":60,"y":40},{"id":"nicd","label":"NICD → RBPJ","x":80,"y":40},{"id":"hes","label":"HES1/HEY, MYC","x":80,"y":78},{"id":"dll3","label":"DLL3 (inhibitory; SCLC surface)","x":15,"y":80,"targetId":"dll3"}],"edges":[{"from":"lig","to":"notch","type":"activates"},{"from":"notch","to":"gs","type":"activates"},{"from":"gs","to":"nicd","type":"activates"},{"from":"nicd","to":"hes","type":"activates"},{"from":"dll3","to":"notch","type":"inhibits"}],"interventions":["Nirogacestat (γ-secretase inhibitor) in desmoid tumours","DLL3-directed tarlatamab in SCLC; DLL3 ADCs and trispecifics","NOTCH1-mutant CLL: reduced benefit from anti-CD20, informs regimen choice"]},{"id":"nutrient-competition-tme","kind":"pathway","name":"Nutrient competition & metabolic immunosuppression","aka":[],"tldr":"Tumours 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.","summary":"Glycolytic tumour cells deplete glucose, leaving T cells unable to sustain aerobic glycolysis needed for IFN-γ production (Chang et al. 2015); lactate export (MCT4) acidifies the niche, blocks T-cell lactate export and polarises macrophages to M2 (via GPR81, HIF-1α); hypoxia induces adenosine via CD39 → CD73, which signals through A2A receptors to suppress T and NK cells. IDO1 and TDO deplete tryptophan and produce kynurenine, activating AHR in T cells and Tregs (the IDO1 inhibitor epacadostat failed in ECHO-301 with pembrolizumab); arginase-1 from MDSCs and TAMs depletes arginine, which T cells need for proliferation; glutamine depletion and cysteine competition add to the exhaustion. Lipid-laden dendritic cells present antigen poorly. Countermeasures: A2A antagonists and CD73 antibodies (modest), arginase inhibitors, glutamine antagonists that spare T cells (DRP-104), metabolic engineering of CAR-T, and diet interventions in early trials.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Tumor_microenvironment","links":[{"label":"Chang et al., Metabolic competition in the tumor microenvironment is a driver of cancer progression (Cell 2015)","url":"https://doi.org/10.1016/j.cell.2015.08.016"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["checkpoint-inhibitor","metabolic-therapy","armored-car","hypoxia-activated-therapy","fasting-mimicking-diet"],"targets":["cd73-adenosine","pd1","csf1r","hif2a","vegf"],"drugs":["epacadostat","pembrolizumab"],"companies":[],"institutions":[],"pathways":["cancer-metabolism","glutamine-metabolism","tumor-microenvironment","hif-vhl","myeloid-suppression-axis"],"terms":["warburg-effect","myeloid-derived-suppressor-cells","tumor-associated-macrophages","cold-vs-hot"],"trials":[],"people":[],"bottlenecks":["b-tme-immunosuppression"],"keyPapers":["paper-chang-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"tum","label":"Glycolytic tumour cell","x":45,"y":12},{"id":"glc","label":"Glucose, glutamine depleted","x":12,"y":35},{"id":"lac","label":"Lactate, acidity (MCT4)","x":45,"y":40},{"id":"ado","label":"CD39 → CD73 → adenosine","x":78,"y":35,"targetId":"cd73-adenosine"},{"id":"ido","label":"IDO1 → kynurenine","x":12,"y":65},{"id":"arg","label":"Arginase (MDSC, TAM)","x":45,"y":68,"targetId":"csf1r"},{"id":"tcell","label":"T-cell / NK dysfunction","x":78,"y":68,"targetId":"pd1"},{"id":"m2","label":"M2 macrophage polarisation","x":78,"y":92},{"id":"hyp","label":"Hypoxia","x":12,"y":92,"targetId":"hif2a"}],"edges":[{"from":"tum","to":"glc","type":"activates"},{"from":"tum","to":"lac","type":"activates"},{"from":"tum","to":"ido","type":"activates"},{"from":"hyp","to":"ado","type":"activates"},{"from":"hyp","to":"tum","type":"activates"},{"from":"glc","to":"tcell","type":"inhibits"},{"from":"lac","to":"tcell","type":"inhibits"},{"from":"lac","to":"m2","type":"activates"},{"from":"ado","to":"tcell","type":"inhibits"},{"from":"ido","to":"tcell","type":"inhibits"},{"from":"arg","to":"tcell","type":"inhibits"}],"interventions":["IDO1 inhibitor epacadostat failed with pembrolizumab in melanoma (ECHO-301); lesson on phase 2 mirages","Adenosine axis: CD73 antibodies (oleclumab) and A2A antagonists, modest activity so far","Arginase inhibitors and glutamine antagonists (DRP-104) designed to spare T cells","Vessel normalisation and hypoxia relief; metabolically armoured CAR-T; diet trials"]},{"id":"er-signaling","kind":"pathway","name":"Oestrogen receptor signalling","aka":[],"tldr":"In hormone-positive breast cancer, oestrogen binds its receptor, which switches on genes that make the cell divide. Every endocrine therapy cuts this chain somewhere.","summary":"Aromatase converts androgens to oestradiol (in ovaries pre-menopause, fat and tumour post-menopause). Oestradiol binds ERα (ESR1), which dimerises, recruits co-activators, and drives transcription of cyclin D1, MYC, PGR, and growth factors, activating CDK4/6 and cross-talking with PI3K and HER2. Tamoxifen (SERM) competes; aromatase inhibitors deplete ligand; fulvestrant and oral SERDs degrade ER; vepdegestrant (PROTAC) degrades it via cereblon. Resistance: ESR1 ligand-binding-domain mutations (ligand-independent, ~30% after AI), cyclin D1/CDK4 amplification, PI3K activation, lineage switch.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Estrogen_receptor","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Estrogen_receptor"}],"tags":[],"related":[],"cancers":["breast-hr-positive"],"sections":[],"technologies":["endocrine-therapy","protac-degrader","cdk46-inhibitor"],"targets":["estrogen-receptor","cdk4-6","pik3ca"],"drugs":["vepdegestrant","elacestrant","palbociclib","ribociclib","abemaciclib"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"A key (oestrogen) turns a lock (ER) that opens a gate to the division machinery. Aromatase inhibitors stop making keys, tamoxifen jams a fake key in the lock, SERDs and PROTACs remove the lock from the door, and ESR1 mutations are a lock that opens without any key.","nodes":[{"id":"andro","label":"Androgens","x":20,"y":8},{"id":"arom","label":"Aromatase","x":20,"y":25},{"id":"e2","label":"Oestradiol","x":20,"y":42},{"id":"er","label":"ERα (ESR1)","x":50,"y":42,"targetId":"estrogen-receptor"},{"id":"esr1mut","label":"ESR1 Y537S / D538G","x":80,"y":25},{"id":"coact","label":"Co-activators, FOXA1","x":80,"y":55},{"id":"ccnd1","label":"Cyclin D1, MYC, PGR","x":50,"y":65},{"id":"cdk","label":"CDK4/6","x":50,"y":82,"targetId":"cdk4-6"},{"id":"prol","label":"Proliferation","x":50,"y":97}],"edges":[{"from":"andro","to":"arom","type":"activates"},{"from":"arom","to":"e2","type":"activates"},{"from":"e2","to":"er","type":"activates"},{"from":"esr1mut","to":"er","type":"activates"},{"from":"coact","to":"er","type":"activates"},{"from":"er","to":"ccnd1","type":"activates"},{"from":"ccnd1","to":"cdk","type":"activates"},{"from":"cdk","to":"prol","type":"activates"}],"interventions":["Aromatase inhibitors (letrozole, anastrozole, exemestane) ± ovarian suppression","SERMs (tamoxifen)","SERDs: fulvestrant, elacestrant, imlunestrant, camizestrant","PROTAC degrader vepdegestrant (2026)","CDK4/6 inhibitors downstream; PI3K/AKT inhibitors for cross-talk"]},{"id":"oncogenic-viruses","kind":"pathway","name":"Oncogenic viruses","aka":[],"tldr":"About 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.","summary":"Direct carcinogens express viral oncoproteins: HPV E6 degrades p53 and E7 inactivates RB (cervical, anal, oropharyngeal; p16 overexpression is the surrogate); EBV LMP1 mimics CD40 to drive NF-κB and EBNA proteins immortalise B cells (nasopharyngeal, Burkitt, Hodgkin, post-transplant lymphoma; plasma EBV DNA is a biomarker); HTLV-1 Tax/HBZ (adult T-cell leukaemia); KSHV vFLIP/vCyclin (Kaposi sarcoma); MCPyV truncated large T antigen (Merkel cell). Indirect carcinogens act via chronic inflammation and regeneration: HBV (also integrates and expresses HBx) and HCV cause cirrhosis and HCC; H. pylori (a bacterium) drives gastric cancer via CagA and inflammation. Virus-driven tumours carry foreign antigens, explaining high immunotherapy response in Merkel cell carcinoma and HPV+ head and neck cancer, and enabling viral-antigen TCR-T and vaccines. Prevention: HPV (Gardasil 9) and HBV vaccination, HCV cure with direct-acting antivirals, H. pylori eradication.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Oncovirus","links":[{"label":"Mesri, Feitelson & Munger, Human viral oncogenesis: a cancer hallmarks analysis (Cell Host Microbe 2014)","url":"https://doi.org/10.1016/j.chom.2014.02.011"},{"label":"IARC Monographs, biological agents (volume 100B)","url":"https://publications.iarc.fr/119"},{"label":"Young and Rickinson, Nat Rev Cancer 2004: Epstein-Barr virus, 40 years on","url":"https://doi.org/10.1038/nrc1452"},{"label":"Kataoka et al., Nat Genet 2015: integrated molecular analysis of 426 adult T-cell leukaemia/lymphoma cases","url":"https://doi.org/10.1038/ng.3415"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["cervical","head-and-neck","nasopharyngeal","hcc","merkel-cell-carcinoma","kaposi-sarcoma","anal","hodgkin-lymphoma","non-hodgkin-lymphoma","peripheral-t-cell-lymphoma","burkitt-lymphoma"],"sections":[],"technologies":["hpv-vaccine","hpv-testing","colposcopy-excision","cthpv-dna","hcc-surveillance","tcr-t"],"targets":["tp53","cdk4-6","pd1","pdl1"],"drugs":["gardasil-9","avelumab","retifanlimab","pembrolizumab","cemiplimab"],"companies":[],"institutions":[],"pathways":["inflammation-nfkb","p53-mdm2-axis","p53-cell-cycle","antigen-presentation-immunoediting"],"terms":["hpv-p16","hbv-hcv","plasma-ebv-dna","cin-hsil","tumor-promoting-inflammation","lymphoma-bio-ebv-latency","lymphoma-bio-htlv1"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-mesri-cell-host-microbe"],"journals":[],"dependsOn":[],"notes":["Lymphoma: two viruses, two mechanisms. Epstein-Barr virus causes different diseases depending on which latency programme the infected cell runs, from EBNA1 alone in Burkitt lymphoma to the full set in post-transplant disease, and the programme tracks how closely T cells are watching (Young and Rickinson 2004). HTLV-1 works by starting a process the host genome then finishes along the lines the virus drew: across 426 adult T-cell leukaemia/lymphoma cases, the acquired alterations overlapped significantly with the Tax interactome and concentrated in T-cell receptor and NF-kB signalling, trafficking and immune surveillance (Kataoka 2015)."],"analogy":"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.","nodes":[{"id":"hpv","label":"HPV E6 / E7","x":12,"y":15},{"id":"ebv","label":"EBV LMP1, EBNA","x":12,"y":45},{"id":"hbv","label":"HBV / HCV","x":12,"y":78},{"id":"p53","label":"p53 degraded","x":45,"y":8,"targetId":"tp53"},{"id":"rb","label":"RB inactivated","x":45,"y":30,"targetId":"cdk4-6"},{"id":"nfkb","label":"NF-κB, immortalisation","x":45,"y":52},{"id":"infl","label":"Chronic inflammation, cirrhosis","x":45,"y":78},{"id":"ca","label":"Virus-driven cancer","x":80,"y":45},{"id":"ag","label":"Viral antigens → IO response","x":80,"y":80,"targetId":"pd1"},{"id":"vax","label":"Vaccination, antivirals","x":80,"y":12}],"edges":[{"from":"hpv","to":"p53","type":"activates"},{"from":"hpv","to":"rb","type":"activates"},{"from":"ebv","to":"nfkb","type":"activates"},{"from":"hbv","to":"infl","type":"activates"},{"from":"p53","to":"ca","type":"activates"},{"from":"rb","to":"ca","type":"activates"},{"from":"nfkb","to":"ca","type":"activates"},{"from":"infl","to":"ca","type":"activates"},{"from":"ca","to":"ag","type":"activates"},{"from":"vax","to":"hpv","type":"inhibits"},{"from":"vax","to":"hbv","type":"inhibits"}],"interventions":["HPV vaccination (Gardasil 9) and HBV vaccination prevent the infection; HCV direct-acting antivirals and H. pylori eradication remove the driver","HPV testing and colposcopy find and excise precursor lesions","Checkpoint inhibitors work well in Merkel cell (avelumab, retifanlimab) and HPV+ cancers; EBV- and HPV-specific TCR-T and vaccines are in trials","Plasma EBV DNA screens for nasopharyngeal carcinoma and tracks response"]},{"id":"organ-tropism-seed-soil","kind":"pathway","name":"Organ tropism: seed and soil","aka":[],"tldr":"Breast 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.","summary":"Blood flow anatomy sets first-pass organs (portal vein → liver for gut cancers; lung for most), but survival there is selective. Bone: tumour PTHrP, IL-11 and IL-6 induce osteoblast RANKL, activating osteoclasts that release matrix-bound TGF-β and IGF-1, which feed the tumour (the vicious cycle broken by bisphosphonates and denosumab; radium-223 targets the bone-forming surface); CXCL12-CXCR4 and osteomimicry attract seeds. Lung: tenascin C, periostin and VCAM1 support stem-like cells; SPARC and ID1 programmes. Liver: Kupffer-cell and stellate-cell activation, CREB-driven metabolic adaptation, low immune surveillance; PDAC and uveal melanoma (BAP1 loss) show strong hepatic tropism. Brain: astrocyte-mediated support via gap junctions and plasmin inhibition (see blood-brain barrier). Peritoneum: ascites, omental fat, mesothelial attachment (ovarian, gastric, appendiceal). Seed programmes are partly encoded by subtype (basal breast → lung and brain; luminal → bone; HER2 → brain; SCLC → brain). Clinically, tropism informs surveillance (brain MRI in ALK/HER2 disease), bone-modifying agents, metastasis-directed SBRT, and regional therapies (HIPEC, liver-directed radioembolisation).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Seed_and_soil_hypothesis","links":[{"label":"Obenauf & Massagué, Surviving at a distance: organ-specific metastasis (Trends Cancer 2015)","url":"https://doi.org/10.1016/j.trecan.2015.07.009"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["prostate","breast-hr-positive","colorectal","uveal-melanoma","ovarian"],"sections":[],"technologies":["bone-modifying-agents","sbrt","hipec","radioembolisation-tare","liver-transplant-oncology","whole-body-mri","psma-pet"],"targets":["cxcr4","psma","her2","alk"],"drugs":["radium-223","pluvicto","tucatinib","lorlatinib"],"companies":[],"institutions":[],"pathways":["metastatic-cascade","pre-metastatic-niche","blood-brain-barrier-metastasis","tgf-beta","tumor-dormancy"],"terms":["oligometastatic","peritoneal-metastasis","her2-brain-metastases","pam50","blood-brain-barrier"],"trials":[],"people":[],"bottlenecks":["b-metastasis-biology"],"keyPapers":["paper-obenauf-trends-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"ctc","label":"CTC (seed programme)","x":12,"y":45},{"id":"flow","label":"Blood-flow anatomy","x":12,"y":12},{"id":"bone","label":"Bone: RANKL vicious cycle","x":45,"y":12},{"id":"lung","label":"Lung: tenascin C, periostin","x":45,"y":38},{"id":"liver","label":"Liver: Kupffer, stellate cells","x":45,"y":64},{"id":"brain","label":"Brain: astrocytes, BBB","x":45,"y":90},{"id":"tgf","label":"TGF-β, IGF-1 released","x":78,"y":12},{"id":"growth","label":"Organ-specific colonisation","x":88,"y":55},{"id":"bma","label":"Denosumab, radium-223","x":78,"y":32},{"id":"regional","label":"SBRT, HIPEC, TARE","x":88,"y":88}],"edges":[{"from":"flow","to":"ctc","type":"activates"},{"from":"ctc","to":"bone","type":"activates"},{"from":"ctc","to":"lung","type":"activates"},{"from":"ctc","to":"liver","type":"activates"},{"from":"ctc","to":"brain","type":"activates"},{"from":"bone","to":"tgf","type":"activates"},{"from":"tgf","to":"bone","type":"activates"},{"from":"tgf","to":"growth","type":"activates"},{"from":"lung","to":"growth","type":"activates"},{"from":"liver","to":"growth","type":"activates"},{"from":"brain","to":"growth","type":"activates"},{"from":"bma","to":"bone","type":"inhibits"},{"from":"regional","to":"growth","type":"inhibits"}],"interventions":["Bone: denosumab and zoledronic acid break the vicious cycle; radium-223 and 177Lu-PSMA in bone-predominant prostate cancer","Brain-penetrant TKIs and surveillance MRI in HER2+ breast, ALK+ lung and SCLC","Regional therapy for organ-confined spread: HIPEC (peritoneum), radioembolisation and liver transplant (liver), SBRT for oligometastases","Tropism-aware adjuvant trials and exosome profiling to predict relapse site"]},{"id":"p53-cell-cycle","kind":"pathway","name":"p53 / RB / cell-cycle checkpoint","aka":[],"tldr":"The 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.","summary":"DNA damage activates ATM/ATR → CHK2/CHK1 → p53 stabilisation (MDM2 degrades p53 normally). p53 induces p21, which inhibits CDK4/6-cyclin D and CDK2-cyclin E, keeping RB bound to E2F and the cell in G1. Cyclin D1 amplification, CDK4 amplification, CDKN2A (p16) loss, RB loss, and TP53 mutation (50% of cancers) each release the brake. CDK4/6 inhibitors re-impose it in RB-intact HR+ breast cancer; WEE1 and PLK1 inhibitors exploit G2/M dependence in TP53-mutant cells; MDM2 inhibitors reactivate wild-type p53.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Cell_cycle_checkpoint","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Cell_cycle_checkpoint"}],"tags":[],"related":[],"cancers":["breast-hr-positive","tnbc","ovarian","sarcoma","gallbladder"],"sections":[],"technologies":[],"targets":["tp53","cdk4-6","wee1","atr"],"drugs":["palbociclib","ribociclib","abemaciclib"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"dmg","label":"DNA damage","x":20,"y":5},{"id":"atm","label":"ATM / ATR","x":20,"y":20,"targetId":"atr"},{"id":"p53","label":"p53","x":20,"y":40,"targetId":"tp53"},{"id":"mdm2","label":"MDM2","x":50,"y":30},{"id":"p21","label":"p21","x":20,"y":58},{"id":"cdk46","label":"CDK4/6-cyclin D","x":65,"y":58,"targetId":"cdk4-6"},{"id":"rb","label":"RB","x":65,"y":75},{"id":"e2f","label":"E2F","x":65,"y":90},{"id":"s","label":"S phase (DNA replication)","x":90,"y":97},{"id":"wee1","label":"WEE1 (G2/M)","x":90,"y":40,"targetId":"wee1"}],"edges":[{"from":"dmg","to":"atm","type":"activates"},{"from":"atm","to":"p53","type":"activates"},{"from":"mdm2","to":"p53","type":"inhibits"},{"from":"p53","to":"p21","type":"activates"},{"from":"p21","to":"cdk46","type":"inhibits"},{"from":"cdk46","to":"rb","type":"inhibits"},{"from":"rb","to":"e2f","type":"inhibits"},{"from":"e2f","to":"s","type":"activates"},{"from":"atm","to":"wee1","type":"activates"}],"interventions":["CDK4/6 inhibitors (palbociclib, ribociclib, abemaciclib)","WEE1 inhibitors (azenosertib) and PLK1 inhibitors in TP53-mutant tumours","MDM2 inhibitors (brigimadlin) in TP53-wild-type tumours","p53 Y220C reactivator rezatapopt","CDK2 inhibitors for CCNE1-amplified and CDK4/6-resistant disease"]},{"id":"pancreatic-cancer-signalling","kind":"pathway","name":"Pancreatic cancer (KEGG map)","aka":["KEGG hsa05212","Pancreatic cancer"],"tldr":"This KEGG map shows the order of genetic hits that turn normal pancreatic duct cells into ductal adenocarcinoma: KRAS mutation first, then loss of the p16 brake, then loss of TP53, SMAD4 and BRCA2. It matters because nearly every pancreatic cancer is driven by KRAS, which until recently had no drug.","summary":"Infiltrating ductal adenocarcinoma is the most common pancreatic malignancy and progresses from normal duct epithelium through histologically defined precursors (PanINs). KEGG map hsa05212 draws the sequence: HER2 (ERBB2) over-expression and activating KRAS point mutation occur early, inactivation of p16 (CDKN2A) at an intermediate stage, and inactivation of TP53, SMAD4 and BRCA2 relatively late. Activated KRAS engages several effector arms at once: RAF-MEK-ERK, PI3K-AKT and RalGDS. EGFR family receptors sit both upstream of RAS and downstream of it, because RAS induces autocrine EGFR ligands; HER2 also feeds JAK-STAT. Loss of p16 releases CDK4/6 and the G1/S checkpoint; loss of TP53 removes damage-induced arrest; loss of SMAD4 removes the growth-inhibitory arm of TGF-beta signalling. The tumours show extensive genomic instability and aneuploidy, to which telomere attrition and TP53 and BRCA2 mutation contribute.\n\nKleeff and colleagues, Nature Reviews Disease Primers, 2016 (doi:10.1038/nrdp.2016.22) review the disease: KRAS mutation is found in more than 90 percent of tumours, CDKN2A, TP53 and SMAD4 are the other frequently altered drivers, a dense stroma limits drug delivery, and germline BRCA1/2 and other homologous recombination defects define a subset that responds to platinum and PARP inhibition.\n\nWhat drugs do about it: direct KRAS inhibition has arrived, with KRAS G12C inhibitors (sotorasib, adagrasib) for the small G12C subset and the pan-RAS(ON) inhibitor daraxonrasib in phase 3 for the common G12D and G12V mutations; olaparib maintenance treats germline BRCA-mutant disease after platinum; erlotinib with gemcitabine (listed by KEGG) gives a small EGFR-directed benefit; larotrectinib and entrectinib treat rare NTRK fusion tumours; pembrolizumab treats the rare MSI-high tumours. Cytotoxic chemotherapy (FOLFIRINOX, gemcitabine with nab-paclitaxel) remains the backbone.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05212","url":"https://www.kegg.jp/pathway/hsa05212"},{"label":"Review: Pancreatic cancer (Nature Reviews Disease Primers)","url":"https://doi.org/10.1038/nrdp.2016.22"}],"tags":[],"related":["ras-mapk","ddr","tgf-beta"],"cancers":["pancreatic"],"sections":[],"technologies":[],"targets":["her2","kras","braf","pik3ca","jak2","cdk4-6","tp53","brca","egfr","ntrk","parp"],"drugs":["sotorasib","adagrasib","daraxonrasib","olaparib","erlotinib","gemcitabine","folfirinox","larotrectinib","entrectinib","pembrolizumab"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","p53-cell-cycle","tgf-beta","ddr","jak-stat","rtk-activation"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-kleeff-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A locked room escape in reverse. KRAS mutation is the first lock picked, and it opens three doors at once (ERK, PI3K, RalGDS). p16, TP53 and SMAD4 are guards posted at later doors, removed one by one. The room is also surrounded by a thick wall of scar-like stroma that keeps rescuers (drugs) out.","nodes":[{"id":"her2","label":"HER2 / EGFR (early)","x":50,"y":5,"targetId":"her2"},{"id":"kras","label":"KRAS mutation (early)","x":50,"y":24,"targetId":"kras"},{"id":"erk","label":"RAF / MEK / ERK","x":20,"y":44,"targetId":"braf"},{"id":"pi3k","label":"PI3K / AKT","x":50,"y":44,"targetId":"pik3ca"},{"id":"ral","label":"RalGDS","x":80,"y":44},{"id":"jak","label":"JAK / STAT","x":92,"y":24,"targetId":"jak2"},{"id":"p16","label":"p16 / CDK4-6 (intermediate)","x":15,"y":66,"targetId":"cdk4-6"},{"id":"p53","label":"TP53 (late)","x":45,"y":66,"targetId":"tp53"},{"id":"smad4","label":"SMAD4 / TGF-beta (late)","x":70,"y":66},{"id":"brca2","label":"BRCA2 (late)","x":92,"y":66,"targetId":"brca"},{"id":"out","label":"PanIN to ductal adenocarcinoma","x":50,"y":94}],"edges":[{"from":"her2","to":"kras","type":"activates"},{"from":"kras","to":"her2","type":"activates"},{"from":"her2","to":"jak","type":"activates"},{"from":"kras","to":"erk","type":"activates"},{"from":"kras","to":"pi3k","type":"activates"},{"from":"kras","to":"ral","type":"activates"},{"from":"erk","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"ral","to":"out","type":"activates"},{"from":"jak","to":"out","type":"activates"},{"from":"p16","to":"out","type":"inhibits"},{"from":"p53","to":"out","type":"inhibits"},{"from":"smad4","to":"out","type":"inhibits"},{"from":"brca2","to":"out","type":"inhibits"}],"interventions":["Direct KRAS inhibition: sotorasib or adagrasib for KRAS G12C; pan-RAS(ON) inhibitor daraxonrasib in phase 3 for G12D/G12V disease","PARP inhibitor olaparib as maintenance for germline BRCA-mutant tumours after platinum chemotherapy","EGFR inhibitor erlotinib with gemcitabine (small benefit, listed by KEGG)","NTRK fusion tumours: larotrectinib, entrectinib; MSI-high tumours: pembrolizumab","Chemotherapy backbone: FOLFIRINOX or gemcitabine with nab-paclitaxel"]},{"id":"pd1-checkpoint","kind":"pathway","name":"PD-1 / PD-L1 immune checkpoint & T-cell activation","aka":[],"tldr":"How 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.","summary":"Dendritic cells prime T cells in lymph nodes via peptide-MHC to TCR (signal 1) plus B7 to CD28 (signal 2); CTLA-4 competes for B7 and terminates priming. In the tumour, interferon-gamma from activated T cells induces PD-L1 on tumour and myeloid cells, which engages PD-1 on T cells and inhibits their effector function (adaptive resistance). LAG-3, TIM-3, TIGIT are further exhaustion checkpoints. Anti-CTLA-4 acts at priming; anti-PD-1/PD-L1 in the tumour. Predictors: PD-L1, TMB, MSI, TILs, interferon signatures. Resistance: loss of MHC-I/B2M, JAK1/2, antigen loss, immunosuppressive myeloid cells, TGF-β exclusion.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Immune_checkpoint","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Immune_checkpoint"},{"label":"Green et al., Blood 2010: selective 9p24.1 amplification and PD-1 ligand induction through JAK2 in Hodgkin lymphoma and mediastinal large B-cell lymphoma","url":"https://doi.org/10.1182/blood-2010-05-282780"},{"label":"Roemer et al., J Clin Oncol 2016: PD-L1 and PD-L2 genetic alterations in 108 classical Hodgkin lymphomas","url":"https://doi.org/10.1200/JCO.2016.66.4482"}],"tags":[],"related":[],"cancers":["melanoma","nsclc","tnbc","rcc","urothelial","non-hodgkin-lymphoma","hodgkin-lymphoma","primary-mediastinal-b-cell-lymphoma"],"sections":[],"technologies":["checkpoint-inhibitor","t-cell-engager","neoantigen-mrna-vaccine"],"targets":["pd1","pdl1","ctla4","lag3","tigit","cd3","cd80","cd28","b7h4"],"drugs":["pembrolizumab","nivolumab","ipilimumab","atezolizumab","durvalumab","relatlimab-nivolumab","ivonescimab"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":["Lymphoma: Hodgkin lymphoma and primary mediastinal B-cell lymphoma are the two diseases in which the checkpoint is switched on by a copy-number event rather than by the microenvironment. The 9p24.1 amplicon carries CD274, PDCD1LG2 and JAK2, so the ligands rise by gene dose and by transcription at once (Green 2010), and 97% of 108 classical Hodgkin lymphomas carried concordant alterations of the two ligand loci (Roemer 2016)."],"analogy":"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.","nodes":[{"id":"dc","label":"Dendritic cell (antigen + B7)","x":20,"y":8},{"id":"tcr","label":"TCR-peptide/MHC","x":20,"y":28},{"id":"cd28","label":"CD28-B7 (go)","x":50,"y":28},{"id":"ctla4","label":"CTLA-4 (stop)","x":80,"y":28,"targetId":"ctla4"},{"id":"tcell","label":"Activated CD8 T cell","x":50,"y":50},{"id":"ifn","label":"IFN-γ","x":80,"y":62},{"id":"pdl1","label":"PD-L1 on tumour","x":80,"y":80,"targetId":"pdl1"},{"id":"pd1","label":"PD-1 on T cell","x":50,"y":70,"targetId":"pd1"},{"id":"lag3","label":"LAG-3 / TIGIT","x":20,"y":70,"targetId":"lag3"},{"id":"kill","label":"Tumour cell killing","x":50,"y":95}],"edges":[{"from":"dc","to":"tcr","type":"activates"},{"from":"dc","to":"cd28","type":"activates"},{"from":"tcr","to":"tcell","type":"activates"},{"from":"cd28","to":"tcell","type":"activates"},{"from":"ctla4","to":"cd28","type":"inhibits"},{"from":"tcell","to":"ifn","type":"activates"},{"from":"ifn","to":"pdl1","type":"activates"},{"from":"pdl1","to":"pd1","type":"activates"},{"from":"pd1","to":"tcell","type":"inhibits"},{"from":"lag3","to":"tcell","type":"inhibits"},{"from":"tcell","to":"kill","type":"activates"}],"interventions":["Anti-PD-1 (pembrolizumab, nivolumab), anti-PD-L1 (atezolizumab, durvalumab)","Anti-CTLA-4 (ipilimumab) ± anti-PD-1","Anti-LAG-3 (relatlimab) + nivolumab","PD-1×VEGF bispecifics (ivonescimab)","Vaccines, T-cell engagers, and CAR-T supply signal 1 by other means"]},{"id":"pi3k-akt-mtor","kind":"pathway","name":"PI3K / AKT / mTOR","aka":[],"tldr":"The 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.","summary":"Receptor tyrosine kinases (HER2, EGFR, IGF1R) activate PI3K (p110α, encoded by PIK3CA), producing PIP3, which recruits AKT. PTEN reverses this step and is a tumour suppressor. AKT phosphorylates many targets including TSC2, releasing mTORC1 to drive protein synthesis, and inhibits FOXO and BAD. The most frequently altered pathway in cancer: PIK3CA mutation (~40% HR+ breast, endometrial, head and neck), PTEN loss (prostate, endometrial, glioblastoma), AKT1 E17K. Drugs: alpelisib, inavolisib (PI3Kα), capivasertib (AKT), everolimus (mTOR), gedatolisib (PI3K/mTOR). Feedback: mTOR inhibition releases AKT, so combinations and endocrine partners are needed.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/PI3K/AKT/mTOR_pathway","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/PI3K/AKT/mTOR_pathway"}],"tags":[],"related":[],"cancers":["breast-hr-positive","endometrial","prostate","tnbc","gallbladder"],"sections":[],"technologies":[],"targets":["pik3ca","akt","her2","egfr"],"drugs":["capivasertib","inavolisib","gedatolisib"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"rtk","label":"RTK (HER2, EGFR)","x":50,"y":5,"targetId":"her2"},{"id":"pi3k","label":"PI3K (PIK3CA)","x":50,"y":25,"targetId":"pik3ca"},{"id":"pten","label":"PTEN","x":15,"y":40},{"id":"pip3","label":"PIP3","x":50,"y":42},{"id":"akt","label":"AKT","x":50,"y":58,"targetId":"akt"},{"id":"tsc","label":"TSC1/2","x":25,"y":74},{"id":"mtor","label":"mTORC1","x":50,"y":82},{"id":"foxo","label":"FOXO / BAD (apoptosis)","x":85,"y":74},{"id":"growth","label":"Protein synthesis, growth","x":50,"y":97}],"edges":[{"from":"rtk","to":"pi3k","type":"activates"},{"from":"pi3k","to":"pip3","type":"activates"},{"from":"pten","to":"pip3","type":"inhibits"},{"from":"pip3","to":"akt","type":"activates"},{"from":"akt","to":"tsc","type":"inhibits"},{"from":"tsc","to":"mtor","type":"inhibits"},{"from":"akt","to":"foxo","type":"inhibits"},{"from":"mtor","to":"growth","type":"activates"}],"interventions":["PI3Kα inhibitors (alpelisib, inavolisib) for PIK3CA-mutant HR+ breast cancer","AKT inhibitor capivasertib for PIK3CA/AKT1/PTEN-altered breast and PTEN-deficient prostate cancer","mTOR inhibitor everolimus","Dual PI3K/mTOR gedatolisib (2026)","Upstream: anti-HER2, anti-EGFR"]},{"id":"prostate-cancer-signalling","kind":"pathway","name":"Prostate cancer (KEGG map)","aka":["KEGG hsa05215","Prostate cancer"],"tldr":"KEGG's prostate cancer map centres on the androgen receptor, the hormone switch that prostate cells depend on, plus loss of PTEN and NKX3.1 that lets PI3K/AKT growth signalling run free. Hormone therapy, AR antagonists and now AKT inhibitors act on these two arms.","summary":"The KEGG prostate cancer map (hsa05215) draws the androgen receptor (AR) as the central node. Testosterone is converted by SRD5A2 to dihydrotestosterone, which releases AR from HSP90 chaperones so it enters the nucleus and, with co-activators (CREBBP, EP300, NCOA family), switches on target genes such as KLK3 (PSA), TMPRSS2 and, through the TMPRSS2-ERG fusion, the ERG oncogene. During androgen deprivation the map shows how AR signalling is re-established: AR gene amplification, AR mutations that respond to other ligands, altered co-activator activity and growth-factor cross-talk. The second arm is growth-factor signalling: EGFR, ERBB2, IGF1R, FGFR and PDGFR feed PI3K to AKT to mTOR, and AKT both stabilises AR output and inhibits FOXO and BAD. PTEN loss, which is very common, and NKX3.1 loss lower p27 (CDKN1B) and remove restraint on the cell cycle, while MDM2 activation by AKT lowers p53. KEGG also draws GSTP1 silencing, which removes carcinogen detoxification in prostatic intraepithelial neoplasia, and FOLH1 (PSMA), the surface protein now used for imaging and radioligand therapy. Watson, Arora and Sawyers, Nat Rev Cancer, 2015 (doi:10.1038/nrc4016) review how castration-resistant tumours restore AR signalling through amplification, ligand-binding-domain mutations (such as F877L under enzalutamide), constitutively active splice variants like AR-V7, intratumoural androgen synthesis and glucocorticoid receptor substitution, and how PTEN loss and lineage plasticity provide AR-independent escape.\n\nWhat drugs do about it: androgen deprivation (GnRH agonists or antagonists) removes the ligand; abiraterone blocks CYP17-dependent androgen synthesis; the AR antagonists enzalutamide, apalutamide and darolutamide stop AR binding DNA even when amplified; the AKT inhibitor capivasertib with abiraterone is approved for PTEN-deficient metastatic castration-resistant disease; and the PSMA radioligand lutetium-177 vipivotide tetraxetan (Pluvicto) delivers radiation to FOLH1-expressing cells after AR pathway inhibitors.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05215","url":"https://www.kegg.jp/pathway/hsa05215"},{"label":"Review: Emerging mechanisms of resistance to AR inhibitors in prostate cancer","url":"https://doi.org/10.1038/nrc4016"}],"tags":[],"related":["ar-signaling","pi3k-akt-mtor"],"cancers":["prostate"],"sections":[],"technologies":[],"targets":["androgen-receptor","egfr","akt","mdm2","psma"],"drugs":["enzalutamide","abiraterone","apalutamide","darolutamide","capivasertib","pluvicto","olaparib","niraparib","talazoparib"],"companies":[],"institutions":[],"pathways":["ar-signaling","pi3k-akt-mtor","rtk-activation","p53-cell-cycle","lineage-plasticity-neuroendocrine"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"The androgen receptor is an engine that runs on testosterone. Cutting the fuel (androgen deprivation) works until the tumour fits a bigger tank (AR amplification) or an engine that runs on anything (AR mutations and splice variants). Enzalutamide and its cousins clamp the engine itself; capivasertib deals with the separate PI3K/AKT motor that PTEN loss switches on.","nodes":[{"id":"androgen","label":"Testosterone to DHT (SRD5A2)","x":20,"y":6},{"id":"ar","label":"Androgen receptor (amplified, mutated)","x":20,"y":30,"targetId":"androgen-receptor"},{"id":"coact","label":"Co-activators (CBP/p300, NCOA)","x":50,"y":20},{"id":"argenes","label":"KLK3 (PSA), TMPRSS2-ERG","x":20,"y":56},{"id":"rtk","label":"EGFR, IGF1R, FGFR, PDGFR","x":80,"y":6,"targetId":"egfr"},{"id":"pten","label":"PTEN, NKX3.1 (lost)","x":95,"y":30},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":75,"y":40,"targetId":"akt"},{"id":"p27","label":"p27 (CDKN1B)","x":75,"y":62},{"id":"mdm2","label":"MDM2 to p53","x":95,"y":62,"targetId":"mdm2"},{"id":"psma","label":"FOLH1 (PSMA)","x":50,"y":78,"targetId":"psma"},{"id":"out","label":"Growth, survival, castration resistance","x":50,"y":95}],"edges":[{"from":"androgen","to":"ar","type":"activates"},{"from":"coact","to":"ar","type":"activates"},{"from":"ar","to":"argenes","type":"activates"},{"from":"argenes","to":"out","type":"activates"},{"from":"rtk","to":"pi3k","type":"activates"},{"from":"pten","to":"pi3k","type":"inhibits"},{"from":"pi3k","to":"ar","type":"activates"},{"from":"pi3k","to":"p27","type":"inhibits"},{"from":"pi3k","to":"mdm2","type":"activates"},{"from":"p27","to":"out","type":"inhibits"},{"from":"mdm2","to":"out","type":"activates"},{"from":"psma","to":"out","type":"activates"}],"interventions":["Androgen deprivation (GnRH agonists or antagonists) combined with an AR pathway inhibitor: enzalutamide, apalutamide, darolutamide or abiraterone (CYP17 inhibitor)","AKT inhibitor capivasertib plus abiraterone for PTEN-deficient metastatic castration-resistant prostate cancer","PSMA radioligand therapy (lutetium-177 vipivotide tetraxetan, Pluvicto) for PSMA-positive disease after AR pathway inhibitors","PARP inhibitors (olaparib, niraparib, talazoparib) with an AR pathway inhibitor for BRCA-altered disease","Docetaxel or cabazitaxel chemotherapy; radiotherapy or prostatectomy for localised disease"]},{"id":"proteoglycans-in-cancer","kind":"pathway","name":"Proteoglycans in cancer","aka":["KEGG hsa05205","Proteoglycans in cancer"],"tldr":"This KEGG map shows how sugar-coated proteins on the cell surface and in the surrounding matrix (proteoglycans such as syndecans, glypicans, CD44 and decorin) catch growth factors and hand signals to receptors. It matters because these molecules set how loudly growth signals reach the tumour cell and how easily it invades.","summary":"Proteoglycans are core proteins carrying long sulphated sugar chains (heparan sulphate, chondroitin sulphate, dermatan sulphate or keratan sulphate); hyaluronan is the related free sugar polymer. KEGG map hsa05205 draws how these molecules in the tumour microenvironment feed proliferation, adhesion, angiogenesis and metastasis. Hyaluronan binding to CD44 promotes growth and migration. Cell-surface heparan sulphate proteoglycans, the syndecans (1 to 4) and glypicans (1 and 3), bind growth factors, cytokines and morphogens through their heparan sulphate chains and present them to receptors such as FGFR, EGFR, IGF1R, MET and VEGFR, driving RAS-ERK, PI3K-AKT, Wnt and Hedgehog signalling. Perlecan in the basement membrane does the same for angiogenic factors. In contrast the small leucine-rich proteoglycans decorin and lumican act as tumour repressors: decorin binds EGFR and MET and dampens their signalling.\n\nIozzo and Sanderson, Journal of Cellular and Molecular Medicine, 2011 (doi:10.1111/j.1582-4934.2010.01236.x) review this biology and stress two enzymes that remodel the system: heparanase, which clips heparan sulphate and releases bound growth factors, and the sheddases that release syndecan-1 from the cell surface, both linked to invasive and angiogenic behaviour in myeloma and solid tumours.\n\nWhat can be done: the proteoglycans themselves are mostly targeted indirectly. Glypican-3, which is expressed on most hepatocellular carcinomas and almost no normal adult tissue, is an antibody and CAR-T antigen. The receptor tyrosine kinases that proteoglycans co-activate are druggable with FGFR inhibitors (pemigatinib, futibatinib, erdafitinib), EGFR inhibitors and MET inhibitors, and the downstream Hedgehog route with vismodegib, which KEGG lists against this map. Heparan sulphate mimetics such as necuparanib, also listed by KEGG, have been tested in early trials.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05205","url":"https://www.kegg.jp/pathway/hsa05205"},{"label":"Review: Proteoglycans in cancer biology, tumour microenvironment and angiogenesis","url":"https://doi.org/10.1111/j.1582-4934.2010.01236.x"}],"tags":[],"related":["rtk-activation","fgfr-signalling","hedgehog","vegf-angiogenesis"],"cancers":[],"sections":[],"technologies":[],"targets":["gpc3","fgfr2","kras","pik3ca","smoothened"],"drugs":["pemigatinib","futibatinib","erdafitinib","savolitinib","vismodegib","bevacizumab","ramucirumab"],"companies":[],"institutions":[],"pathways":["rtk-activation","fgfr-signalling","ras-mapk","pi3k-akt-mtor","hedgehog","wnt","vegf-angiogenesis","emt"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-iozzo-j-cell-mol-med"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Picture the space around a cell as a sticky sugar lawn. Growth-factor messages land on the lawn and cling to it, and the proteoglycans act like ushers that pick messages up and walk them to the receptor doors. Some ushers (syndecans, glypicans) shout the message louder; a few (decorin) quietly bin it. Cancer hires more of the loud ushers and sends in gardeners (heparanase) that cut the lawn so trapped messages fly free.","nodes":[{"id":"ha","label":"Hyaluronan","x":10,"y":8},{"id":"cd44","label":"CD44","x":10,"y":30},{"id":"gf","label":"Growth factors (FGF, HGF, VEGF, Wnt)","x":50,"y":8},{"id":"hspg","label":"Syndecans / glypicans (HS chains)","x":50,"y":28,"targetId":"gpc3"},{"id":"heparanase","label":"Heparanase","x":80,"y":8},{"id":"decorin","label":"Decorin / lumican","x":85,"y":40},{"id":"rtk","label":"RTKs (FGFR, MET, EGFR, VEGFR)","x":50,"y":50,"targetId":"fgfr2"},{"id":"ras","label":"RAS / ERK","x":30,"y":70,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT","x":55,"y":70,"targetId":"pik3ca"},{"id":"hh","label":"Hedgehog / Wnt","x":80,"y":70,"targetId":"smoothened"},{"id":"out","label":"Proliferation, angiogenesis, invasion","x":50,"y":95}],"edges":[{"from":"ha","to":"cd44","type":"activates"},{"from":"cd44","to":"out","type":"activates"},{"from":"gf","to":"hspg","type":"activates"},{"from":"hspg","to":"rtk","type":"activates"},{"from":"heparanase","to":"hspg","type":"inhibits"},{"from":"heparanase","to":"gf","type":"activates"},{"from":"decorin","to":"rtk","type":"inhibits"},{"from":"rtk","to":"ras","type":"activates"},{"from":"rtk","to":"pi3k","type":"activates"},{"from":"hspg","to":"hh","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"hh","to":"out","type":"activates"}],"interventions":["Glypican-3 directed antibodies and CAR-T cells in hepatocellular carcinoma (clinical trials)","Inhibit the co-activated receptors: FGFR inhibitors (pemigatinib, futibatinib, erdafitinib), MET inhibitors (savolitinib), EGFR inhibitors","Hedgehog pathway inhibitor vismodegib (listed by KEGG on this map) where SMO signalling drives the tumour","Anti-angiogenic drugs against VEGF signalling that perlecan and heparan sulphate present (bevacizumab, ramucirumab)","Heparan sulphate mimetics and heparanase inhibitors (necuparanib), early trials only"]},{"id":"ras-mapk","kind":"pathway","name":"RAS / RAF / MEK / ERK (MAPK)","aka":[],"tldr":"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.","summary":"Growth-factor receptors recruit GRB2/SOS to load GTP onto RAS (KRAS, NRAS, HRAS). RAS-GTP recruits RAF (BRAF, CRAF) dimers, which phosphorylate MEK1/2, which phosphorylate ERK1/2. ERK drives transcription of cyclin D1, MYC, and negative feedback (DUSP, SPRY). Alterations: KRAS (pancreatic 90%, CRC 45%, lung 30%), BRAF V600E (melanoma 50%), NF1 loss, receptor fusions (ALK, RET, NTRK). Drugs: KRAS G12C inhibitors, pan-RAS(ON) inhibitors, BRAF+MEK doublets, upstream RTK inhibitors. Feedback reactivation (loss of ERK-mediated inhibition of RTKs) is why single agents fail and vertical combinations (BRAF+MEK, KRAS+EGFR in CRC) work.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/MAPK/ERK_pathway","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/MAPK/ERK_pathway"}],"tags":[],"related":[],"cancers":["pancreatic","colorectal","nsclc","melanoma","gallbladder"],"sections":[],"technologies":[],"targets":["kras","braf","egfr","alk","ret","met","ntrk","fgfr2","kit","nras","shp2","nf1"],"drugs":["sotorasib","adagrasib","daraxonrasib","encorafenib","osimertinib","lorlatinib"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"rtk","label":"RTK (EGFR, ALK, RET, MET)","x":50,"y":5,"targetId":"egfr"},{"id":"sos","label":"GRB2 / SOS1","x":50,"y":20},{"id":"ras","label":"RAS (KRAS)","x":50,"y":36,"targetId":"kras"},{"id":"nf1","label":"NF1 (GAP)","x":15,"y":36},{"id":"raf","label":"RAF (BRAF)","x":50,"y":52,"targetId":"braf"},{"id":"mek","label":"MEK1/2","x":50,"y":68},{"id":"erk","label":"ERK1/2","x":50,"y":82},{"id":"out","label":"Cyclin D1, MYC → proliferation","x":50,"y":97},{"id":"fb","label":"DUSP / SPRY feedback","x":85,"y":60}],"edges":[{"from":"rtk","to":"sos","type":"activates"},{"from":"sos","to":"ras","type":"activates"},{"from":"nf1","to":"ras","type":"inhibits"},{"from":"ras","to":"raf","type":"activates"},{"from":"raf","to":"mek","type":"activates"},{"from":"mek","to":"erk","type":"activates"},{"from":"erk","to":"out","type":"activates"},{"from":"erk","to":"fb","type":"activates"},{"from":"fb","to":"rtk","type":"inhibits"}],"interventions":["KRAS G12C inhibitors (sotorasib, adagrasib) ± anti-EGFR in colorectal cancer","Pan-RAS(ON) inhibitor daraxonrasib (phase 3, pancreatic)","BRAF + MEK inhibitors (dabrafenib/trametinib, encorafenib/binimetinib)","Encorafenib + cetuximab (+ chemo) in BRAF V600E CRC","Upstream: EGFR, ALK, RET, MET, NTRK inhibitors and bispecifics"]},{"id":"rtk-activation","kind":"pathway","name":"Receptor tyrosine kinase activation","aka":[],"tldr":"Growth-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.","summary":"Ligand binding (EGF, HGF, FGF, NRG, SCF, PDGF, VEGF) dimerises RTKs, trans-autophosphorylating tyrosines that recruit SH2/PTB adaptors: GRB2-SOS to RAS-MAPK, p85 to PI3K-AKT, PLCγ, SRC, STAT3/5, and CBL for ubiquitin-mediated downregulation. Oncogenic activation: kinase-domain mutations (EGFR L858R/exon 19, exon 20 insertions; FLT3-ITD; KIT exon 11; HER2), extracellular-domain truncation (EGFRvIII), amplification (HER2, MET, FGFR1/2), fusions (ALK, ROS1, RET, NTRK, FGFR2/3, with dimerisation domains that pair the kinase constitutively), autocrine loops, and PTPN11 or CBL loss. Drug classes: reversible and covalent TKIs (with resistance via gatekeeper/solvent-front mutations such as EGFR C797S, ALK G1202R), antibodies that block ligand or receptor and recruit Fc effectors (trastuzumab, cetuximab), bispecifics (amivantamab EGFR×MET, zanidatamab HER2×HER2), and ADCs that use the receptor as a delivery address regardless of signalling (T-DXd on HER2-low). Bypass via a parallel RTK (MET amplification under EGFR blockade, HER3 upregulation) is the classic escape.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Receptor_tyrosine_kinase","links":[{"label":"Lemmon & Schlessinger, Cell signaling by receptor tyrosine kinases (Cell 2010)","url":"https://doi.org/10.1016/j.cell.2010.06.011"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["nsclc","breast-her2-positive","gastric","gist","thyroid","cholangiocarcinoma","tnbc","gallbladder"],"sections":[],"technologies":["kinase-inhibitors","monoclonal-antibody","bispecific-antibody","adc","companion-diagnostic"],"targets":["egfr","her2","her3","met","alk","ret","ros1","ntrk","fgfr2","kit","flt3","pdgfra","kras","pik3ca"],"drugs":["osimertinib","lorlatinib","alectinib","selpercatinib","larotrectinib","entrectinib","capmatinib-tepotinib","imatinib","zongertinib","trastuzumab","pertuzumab","cetuximab","panitumumab","amivantamab","zanidatamab","zenocutuzumab","trastuzumab-deruxtecan","patritumab-deruxtecan","telisotuzumab-vedotin"],"companies":[],"institutions":[],"pathways":["ras-mapk","pi3k-akt-mtor","jak-stat","resistance-routes-map"],"terms":["oncogene-addiction","gene-fusion","egfr-exon19-l858r","egfr-exon20-insertion","c797s","met-amplification","egfrviii","her2-low","adcc"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"lig","label":"Ligand (EGF, HGF, NRG)","x":15,"y":10},{"id":"rtk","label":"RTK dimer (EGFR, HER2, MET)","x":45,"y":25,"targetId":"egfr"},{"id":"alt","label":"Mutation, amp, fusion","x":80,"y":10,"targetId":"alk"},{"id":"cbl","label":"CBL → degradation","x":82,"y":40},{"id":"ras","label":"GRB2/SOS → RAS","x":20,"y":55,"targetId":"kras"},{"id":"pi3k","label":"PI3K → AKT","x":50,"y":55,"targetId":"pik3ca"},{"id":"stat","label":"SRC, STAT3, PLCγ","x":80,"y":68},{"id":"out","label":"Proliferation, survival","x":50,"y":88},{"id":"bypass","label":"Bypass RTK (MET, HER3)","x":15,"y":88,"targetId":"met"}],"edges":[{"from":"lig","to":"rtk","type":"activates"},{"from":"alt","to":"rtk","type":"activates"},{"from":"cbl","to":"rtk","type":"inhibits"},{"from":"rtk","to":"ras","type":"activates"},{"from":"rtk","to":"pi3k","type":"activates"},{"from":"rtk","to":"stat","type":"activates"},{"from":"ras","to":"out","type":"activates"},{"from":"pi3k","to":"out","type":"activates"},{"from":"stat","to":"out","type":"activates"},{"from":"bypass","to":"pi3k","type":"activates"},{"from":"bypass","to":"ras","type":"activates"}],"interventions":["TKIs by driver: osimertinib (EGFR), lorlatinib/alectinib (ALK), selpercatinib (RET), larotrectinib/entrectinib (NTRK), capmatinib/tepotinib (MET), imatinib (KIT/PDGFRA/BCR-ABL), zongertinib (HER2)","Antibodies: trastuzumab/pertuzumab (HER2), cetuximab/panitumumab (EGFR); bispecifics amivantamab (EGFR×MET), zanidatamab, zenocutuzumab (HER2×HER3)","ADCs use the receptor as an address: T-DXd, T-DM1, patritumab deruxtecan (HER3), telisotuzumab vedotin (MET)","Combining with MET or downstream inhibitors closes bypass routes"]},{"id":"renal-cell-carcinoma-signalling","kind":"pathway","name":"Renal cell carcinoma (KEGG map)","aka":["KEGG hsa05211","Renal cell carcinoma"],"tldr":"KEGG's kidney cancer map shows how losing VHL lets the oxygen sensor HIF pile up and order new blood vessels (VEGF, PDGF), while MET and PI3K drive growth in other subtypes. Anti-VEGF drugs, HIF-2a inhibitors and immunotherapy all act on this circuit.","summary":"The KEGG renal cell carcinoma map (hsa05211) organises RCC by subtype and driver gene. In clear cell RCC, the majority, the VHL tumour suppressor is lost. VHL is the substrate-recognition part of an E3 ubiquitin ligase (with elongin B and C, CUL2 and RBX1) that degrades the alpha subunits of hypoxia-inducible factor once prolyl hydroxylases (EGLN1 to EGLN3) have marked them in the presence of oxygen. Without VHL, HIF-1a and especially HIF-2a (EPAS1) accumulate, pair with ARNT and switch on VEGFA, PDGFB, TGFA, GLUT1 (SLC2A1) and other hypoxia genes, producing a highly vascular tumour even in normal oxygen. In hereditary and sporadic papillary RCC, activating MET mutations (with its ligand HGF) drive GAB1, RAS to ERK, PI3K to AKT and RAC/PAK signalling for motility and invasion. Loss of fumarate hydratase (FH) accumulates fumarate, which inhibits the prolyl hydroxylases and again stabilises HIF; folliculin (FLCN) loss underlies Birt-Hogg-Dube syndrome; and PRCC-TFE3 fusions define translocation RCC. Hsieh et al., Nat Rev Dis Primers, 2017 (doi:10.1038/nrdp.2017.9) review this genetics, adding the chromatin genes PBRM1, SETD2, BAP1 and KDM5C that are co-deleted or mutated on chromosome 3p alongside VHL.\n\nWhat drugs do about it: VEGF receptor tyrosine kinase inhibitors (sunitinib, pazopanib, axitinib, cabozantinib which also hits MET, and lenvatinib) and the VEGF antibody bevacizumab starve the HIF-driven vasculature; belzutifan blocks HIF-2a directly and is approved for VHL disease and previously treated clear cell RCC; mTOR inhibitors everolimus and temsirolimus act on the PI3K axis; and PD-1 antibodies (nivolumab, pembrolizumab) combined with a VEGFR TKI or with ipilimumab are the standard first-line regimens.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05211","url":"https://www.kegg.jp/pathway/hsa05211"},{"label":"Review: Renal cell carcinoma (Nat Rev Dis Primers 2017)","url":"https://doi.org/10.1038/nrdp.2017.9"}],"tags":[],"related":["hif-vhl","vegf-angiogenesis"],"cancers":["rcc"],"sections":[],"technologies":[],"targets":["hif2a","vegf","met","kras","mtor"],"drugs":["belzutifan","sunitinib","pazopanib","axitinib","cabozantinib","lenvatinib","bevacizumab","everolimus","temsirolimus","nivolumab","pembrolizumab","ipilimumab","savolitinib"],"companies":[],"institutions":[],"pathways":["hif-vhl","vegf-angiogenesis","pi3k-akt-mtor","ras-mapk","rtk-activation","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-hsieh-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"VHL is the bin lorry that removes HIF every day. When it stops coming, HIF piles up and sends out non-stop orders for new blood vessels. Anti-VEGF drugs intercept the orders; belzutifan gags HIF-2a so the orders are never written.","nodes":[{"id":"o2","label":"Oxygen, prolyl hydroxylases (EGLN)","x":20,"y":6},{"id":"fh","label":"FH loss, fumarate","x":55,"y":6},{"id":"vhl","label":"VHL E3 ligase (lost)","x":20,"y":26},{"id":"hif","label":"HIF-1a / HIF-2a","x":20,"y":48,"targetId":"hif2a"},{"id":"vegf","label":"VEGF, PDGF, TGFA, GLUT1","x":20,"y":68,"targetId":"vegf"},{"id":"angio","label":"Angiogenesis, growth","x":35,"y":92},{"id":"met","label":"HGF / MET (papillary RCC)","x":78,"y":26,"targetId":"met"},{"id":"ras","label":"RAS / ERK","x":62,"y":50,"targetId":"kras"},{"id":"pi3k","label":"PI3K / AKT / mTOR","x":90,"y":50,"targetId":"mtor"},{"id":"invade","label":"Motility, invasion, survival","x":78,"y":74}],"edges":[{"from":"o2","to":"vhl","type":"activates"},{"from":"fh","to":"o2","type":"inhibits"},{"from":"vhl","to":"hif","type":"inhibits"},{"from":"hif","to":"vegf","type":"activates"},{"from":"vegf","to":"angio","type":"activates"},{"from":"met","to":"ras","type":"activates"},{"from":"met","to":"pi3k","type":"activates"},{"from":"ras","to":"invade","type":"activates"},{"from":"pi3k","to":"invade","type":"activates"}],"interventions":["PD-1 antibody (nivolumab or pembrolizumab) combined with a VEGFR TKI (cabozantinib, axitinib, lenvatinib) or with ipilimumab as first-line therapy for advanced clear cell RCC","VEGF pathway inhibitors: sunitinib, pazopanib, axitinib, cabozantinib, lenvatinib, bevacizumab","HIF-2a inhibitor belzutifan for VHL disease tumours and previously treated advanced RCC","mTOR inhibitors everolimus (with lenvatinib) and temsirolimus","MET-directed TKIs (cabozantinib, savolitinib) for MET-driven papillary RCC"]},{"id":"resistance-routes-map","kind":"pathway","name":"Resistance routes: how a blocked pathway comes back","aka":[],"tldr":"When 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.","summary":"Vasan, Baselga and Hyman's framework. (1) On-target: secondary mutations that block binding (EGFR T790M and C797S, ALK G1202R, BTK C481S, KRAS Y96D, ESR1 ligand-domain, AR F877L, BCR-ABL T315I) or amplification of the target (AR, BCR-ABL, MET after MET inhibition, BRAF splice variants). (2) Bypass: a parallel input restores the downstream signal (MET or HER3 amplification under EGFR blockade, RTK upregulation via loss of ERK feedback after BRAF/MEK inhibition, NRG1 fusions, IGF1R). (3) Downstream: mutation or amplification below the block (KRAS/NRAS, PIK3CA, MAP2K1, PTEN loss, CDK4/cyclin E, RB1 loss under CDK4/6 inhibition). (4) Phenotypic: lineage plasticity (neuroendocrine transformation, EMT, squamous transdifferentiation) or entry into a drug-tolerant persister state that no longer depends on the target. (5) Pharmacological/antigenic: efflux pumps, sanctuary sites (brain), drug metabolism, and for immune therapies antigen loss (CD19, BCMA, B2M) or payload-related mechanisms for ADCs (SLFN11 loss, TOP1 mutation, antigen downregulation). Pre-existing resistant subclones are selected (clonal evolution) and new mutations arise under APOBEC-driven mutagenesis. Countermeasures: next-generation inhibitors, vertical combinations (BRAF+MEK, KRAS+EGFR), parallel combinations, ctDNA-guided switching, and non-cross-resistant modalities (ADCs, radioligands, cell therapy).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Drug_resistance","links":[{"label":"Vasan, Baselga & Hyman, A view on drug resistance in cancer (Nature 2019)","url":"https://doi.org/10.1038/s41586-019-1730-1"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["nsclc","non-hodgkin-lymphoma"],"sections":[],"technologies":["liquid-biopsy","cgp","kinase-inhibitors","adc","radioligand-therapy","car-t"],"targets":["egfr","alk","kras","met","her3","pik3ca","estrogen-receptor","androgen-receptor","btk","bcr-abl","cd19","bcma","dll3"],"drugs":["osimertinib","lorlatinib","amivantamab","pirtobrutinib","asciminib","sotorasib","cetuximab","elacestrant","tarlatamab"],"companies":[],"institutions":[],"pathways":["rtk-activation","ras-mapk","clonal-evolution","drug-tolerant-persisters","drug-efflux-pumps","lineage-plasticity-neuroendocrine","antigen-presentation-immunoediting"],"terms":["resistance","c797s","met-amplification","esr1-mutation","ar-v7","efflux-pump","histologic-transformation","oligoprogression"],"trials":[],"people":[],"bottlenecks":["b-resistance","b-tumor-heterogeneity"],"keyPapers":["paper-vasan-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"drug","label":"Drug blocks target","x":45,"y":8,"targetId":"egfr"},{"id":"target","label":"Target → signal → growth","x":45,"y":40},{"id":"ontarget","label":"1 Target mutation / amp","x":12,"y":25,"targetId":"kras"},{"id":"bypass","label":"2 Bypass RTK (MET, HER3)","x":12,"y":55,"targetId":"met"},{"id":"down","label":"3 Downstream (PIK3CA, RB1)","x":12,"y":85,"targetId":"pik3ca"},{"id":"pheno","label":"4 Lineage switch, persisters","x":78,"y":25,"targetId":"dll3"},{"id":"pharm","label":"5 Efflux, sanctuary, antigen loss","x":78,"y":55,"targetId":"cd19"},{"id":"growth","label":"Regrowth under therapy","x":45,"y":72},{"id":"counter","label":"Next-gen, vertical combos, switch","x":78,"y":88}],"edges":[{"from":"drug","to":"target","type":"inhibits"},{"from":"target","to":"growth","type":"activates"},{"from":"ontarget","to":"drug","type":"inhibits"},{"from":"bypass","to":"growth","type":"activates"},{"from":"down","to":"growth","type":"activates"},{"from":"pheno","to":"growth","type":"activates"},{"from":"pharm","to":"drug","type":"inhibits"},{"from":"counter","to":"growth","type":"inhibits"}],"interventions":["Next-generation inhibitors for on-target mutations (osimertinib for T790M, lorlatinib for ALK G1202R, pirtobrutinib for BTK C481S, asciminib for T315I)","Vertical and parallel combinations: BRAF+MEK, KRAS G12C + EGFR in CRC, CDK4/6 + endocrine + PI3K/AKT","Switching modality on progression: ADCs, radioligands, engagers and CAR-T are not cross-resistant with small molecules","Serial ctDNA to detect the route (C797S, MET amp, ESR1) and adaptive dosing; see the resistance atlas for every class"]},{"id":"rna-splicing","kind":"pathway","name":"RNA splicing","aka":[],"tldr":"Genes are cut and pasted into messages before they are used. Blood cancers often carry mutations in the splicing machinery, and the errors create abnormal proteins that could serve as targets or immune flags.","summary":"Spliceosome mutations (SF3B1, SRSF2, U2AF1, ZRSR2) occur in ~50% of MDS and in CLL and uveal melanoma; they cause mis-splicing, R-loops, and dependence on the remaining wild-type spliceosome (H3B-8800, negative), and on PRMT5 and NMD. Splicing produces drug-resistance isoforms (AR-V7 in prostate cancer, BCL2L1 isoforms) and neoantigens (splice-junction-derived) that may be immunogenic across patients. RNA-targeting small molecules (risdiplam-like) and antisense oligonucleotides are the therapeutic tools; MYC-driven tumours are spliceosome-dependent.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/RNA_splicing","links":[{"label":"Bradley & Anczuków, RNA splicing dysregulation and the hallmarks of cancer (Nature Reviews Cancer 2023)","url":"https://doi.org/10.1038/s41568-022-00541-7"}],"tags":["mechanism"],"related":[],"cancers":["aml","cll","melanoma","prostate"],"sections":[],"technologies":["antisense-sirna","rna-seq"],"targets":["androgen-receptor"],"drugs":[],"companies":[],"institutions":["mskcc","dana-farber","fred-hutch"],"pathways":["myc","replication-stress","ar-signaling","antigen-presentation-immunoediting"],"terms":["ar-v7","neoantigen"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bradley-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Abdel-Wahab (MSK) on spliceosome mutations; Bradley (Fred Hutch) on splicing and neoantigens; Ebert (Dana-Farber) on SF3B1 in MDS."],"analogy":"Splicing is film editing. The raw footage (pre-mRNA) is cut into a final movie. Cancer's editor makes odd cuts: some create villains (AR-V7), some create scenes no one has seen before (neoantigens), and the editing room itself becomes a place where the cancer can be attacked.","nodes":[{"id":"pre","label":"Pre-mRNA","x":12,"y":40},{"id":"sf3b1","label":"SF3B1 / SRSF2 / U2AF1 (mutant)","x":38,"y":20},{"id":"spl","label":"Spliceosome","x":38,"y":60},{"id":"iso","label":"Aberrant isoforms (AR-V7)","x":66,"y":25,"targetId":"androgen-receptor"},{"id":"neo","label":"Splice neoantigens","x":66,"y":60},{"id":"rloop","label":"R-loops, replication stress","x":66,"y":90},{"id":"prmt5","label":"PRMT5 dependence","x":90,"y":60}],"edges":[{"from":"pre","to":"spl","type":"activates"},{"from":"sf3b1","to":"spl","type":"activates"},{"from":"spl","to":"iso","type":"activates"},{"from":"spl","to":"neo","type":"activates"},{"from":"sf3b1","to":"rloop","type":"activates"},{"from":"spl","to":"prmt5","type":"activates"}],"interventions":["Spliceosome modulators (H3B-8800 negative; next generation in development)","PRMT5 inhibitors in MTAP-deleted and splicing-mutant cancers","Antisense oligonucleotides to redirect splicing (AR-V7, BCL2L1)","Splice-derived neoantigen vaccines (research)"]},{"id":"sclc-signalling","kind":"pathway","name":"Small cell lung cancer (KEGG map)","aka":["KEGG hsa05222","Small cell lung cancer"],"tldr":"KEGG's small cell lung cancer map shows a tumour with both master brakes removed, RB1 and TP53, plus amplified MYC pushing the cell cycle and PTEN loss and BCL2 keeping cells alive. Chemotherapy with PD-L1 antibodies and the DLL3 T-cell engager tarlatamab are the current answers.","summary":"The KEGG small cell lung cancer map (hsa05222) describes a neuroendocrine tumour defined by loss of tumour suppressors rather than by a druggable activated kinase. RB1 loss or mutation removes the G1/S gate so E2F transcription runs freely; TP53 mutation removes DNA-damage-induced arrest and apoptosis (loss of p21, GADD45, BAX and BAK induction); PTEN loss releases PI3K to AKT survival signalling; and FHIT deletion on chromosome 3p removes a pro-apoptotic control. On the oncogene side, MYC family amplification (MYC, MYCL, MYCN) drives cyclin D and CDK4/6 and CDK2 while repressing the CDK inhibitors p15 (CDKN2B) and p27 (CDKN1B) through MAX and MIZ1. KEGG also draws BCL2 and BCL-XL overexpression that blocks the intrinsic apoptotic pathway, laminin and integrin signalling through FAK (PTK2) to PI3K and NF-kB, and retinoic acid signalling through RAR-beta and RXR, which is frequently silenced. Rudin et al., Nat Rev Dis Primers, 2021 (doi:10.1038/s41572-020-00235-0) review the near-universal RB1 and TP53 inactivation, the transcription-factor-defined subtypes (ASCL1, NEUROD1, POU2F3 and an inflamed subtype), and the therapeutic targets that follow, including DLL3, which is displayed on the surface of ASCL1-driven cells.\n\nWhat drugs do about it: platinum and etoposide chemotherapy combined with a PD-L1 antibody (atezolizumab, durvalumab, or serplulimab against PD-1) is standard first-line treatment, with durvalumab also used after chemoradiotherapy in limited-stage disease. Tarlatamab, a DLL3 and CD3 bispecific T-cell engager, is approved for relapsed disease, and lurbinectedin or topotecan are second-line chemotherapy options. The B7-H3 antibody-drug conjugate ifinatamab deruxtecan is in late-stage trials.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05222","url":"https://www.kegg.jp/pathway/hsa05222"},{"label":"Review: Small-cell lung cancer (Nat Rev Dis Primers 2021)","url":"https://doi.org/10.1038/s41572-020-00235-0"}],"tags":[],"related":["lineage-plasticity-neuroendocrine","p53-cell-cycle"],"cancers":["sclc","nsclc"],"sections":[],"technologies":[],"targets":["cdk4-6","tp53","akt","bcl2","rxr"],"drugs":["atezolizumab","durvalumab","serplulimab","tarlatamab","lurbinectedin","topotecan","ifinatamab-deruxtecan"],"companies":[],"institutions":[],"pathways":["p53-cell-cycle","cell-cycle-engine-cdks","apoptosis-bcl2","pi3k-akt-mtor","lineage-plasticity-neuroendocrine","transcription-addiction","pd1-checkpoint"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-rudin-nat-rev-dis-primers"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A train with both braking systems removed (RB1 and TP53) and a driver (MYC) shovelling coal. Nothing on the train can be switched off, so treatment works from outside: chemotherapy derails the fastest cars, and immunotherapy and tarlatamab summon T cells to attack a flag (DLL3) that the train happens to fly.","nodes":[{"id":"myc","label":"MYC amplified","x":20,"y":6},{"id":"p27","label":"p15 / p27 (repressed)","x":50,"y":20},{"id":"cdk","label":"Cyclin D / CDK4/6, cyclin E / CDK2","x":20,"y":36,"targetId":"cdk4-6"},{"id":"rb1","label":"RB1 (lost)","x":20,"y":56},{"id":"e2f","label":"E2F, proliferation","x":20,"y":78},{"id":"p53","label":"TP53 (mutated)","x":55,"y":50,"targetId":"tp53"},{"id":"pten","label":"PTEN loss to PI3K / AKT","x":82,"y":20,"targetId":"akt"},{"id":"bcl2","label":"BCL2 / BCL-XL","x":82,"y":50,"targetId":"bcl2"},{"id":"apop","label":"Apoptosis","x":82,"y":78},{"id":"rxr","label":"RAR-beta / RXR (silenced)","x":55,"y":78,"targetId":"rxr"},{"id":"out","label":"Small cell lung cancer","x":50,"y":95}],"edges":[{"from":"myc","to":"p27","type":"inhibits"},{"from":"p27","to":"cdk","type":"inhibits"},{"from":"myc","to":"cdk","type":"activates"},{"from":"cdk","to":"rb1","type":"inhibits"},{"from":"rb1","to":"e2f","type":"inhibits"},{"from":"p53","to":"apop","type":"activates"},{"from":"p53","to":"cdk","type":"inhibits"},{"from":"pten","to":"apop","type":"inhibits"},{"from":"bcl2","to":"apop","type":"inhibits"},{"from":"e2f","to":"out","type":"activates"},{"from":"apop","to":"out","type":"inhibits"},{"from":"rxr","to":"out","type":"inhibits"}],"interventions":["Platinum plus etoposide chemotherapy with a PD-L1 antibody (atezolizumab or durvalumab) or PD-1 antibody (serplulimab) for extensive-stage disease; durvalumab consolidation after chemoradiotherapy in limited-stage disease","DLL3 x CD3 bispecific T-cell engager tarlatamab for relapsed small cell lung cancer","Second-line chemotherapy: lurbinectedin or topotecan","B7-H3 antibody-drug conjugate ifinatamab deruxtecan in late-stage trials","Prophylactic or therapeutic cranial radiotherapy and thoracic radiotherapy in selected patients"]},{"id":"swi-snf-chromatin","kind":"pathway","name":"SWI/SNF chromatin remodelling","aka":[],"tldr":"A 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.","summary":"SWI/SNF (BAF, PBAF, ncBAF) complexes remodel nucleosomes; subunits are mutated in ~20% of cancers: ARID1A (ovarian clear cell, endometrial, bladder), SMARCA4 (lung, SCCOHT), PBRM1 (clear-cell RCC), SMARCB1 (rhabdoid tumours, epithelioid sarcoma). Dependencies: SMARCA4-loss → SMARCA2 (degraders in trials), ARID1A-loss → EZH2/ATR/PARP, SMARCB1-loss → EZH2 (tazemetostat, withdrawn 2026 for safety). PBRM1 loss associates with IO response in RCC (debated). FHD-286 (BRM/BRG1 inhibitor) is in AML trials.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/SWI/SNF","links":[{"label":"Centore, Sandoval, Soares, Kadoch & Chan, Mammalian SWI/SNF chromatin remodeling complexes (Trends in Genetics 2020)","url":"https://doi.org/10.1016/j.tig.2020.07.011"}],"tags":["mechanism"],"related":[],"cancers":["rcc","ovarian","nsclc","sarcoma"],"sections":[],"technologies":["synthetic-lethality-approaches","protac-degrader"],"targets":["ezh2","atr"],"drugs":[],"companies":[],"institutions":["dana-farber","broad-institute","stanford"],"pathways":["epigenetic-reprogramming","cancer-stem-cells-plasticity","replication-stress"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-centore-trends-genet"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Kadoch (Dana-Farber/Broad) on SWI/SNF structure and Foghorn Therapeutics; Crabtree (Stanford); Roberts (St. Jude) on SMARCB1."],"analogy":"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.","nodes":[{"id":"baf","label":"BAF / PBAF / ncBAF","x":40,"y":30},{"id":"nuc","label":"Nucleosome repositioning","x":70,"y":30},{"id":"tx","label":"Enhancer access, differentiation genes","x":90,"y":55},{"id":"arid","label":"ARID1A loss","x":12,"y":20},{"id":"smarca4","label":"SMARCA4 loss → SMARCA2 dependence","x":12,"y":50},{"id":"smarcb1","label":"SMARCB1 loss → EZH2","x":12,"y":80,"targetId":"ezh2"},{"id":"prc2","label":"PRC2 (EZH2) antagonism","x":60,"y":75,"targetId":"ezh2"}],"edges":[{"from":"baf","to":"nuc","type":"activates"},{"from":"nuc","to":"tx","type":"activates"},{"from":"arid","to":"baf","type":"inhibits"},{"from":"smarca4","to":"baf","type":"inhibits"},{"from":"smarcb1","to":"baf","type":"inhibits"},{"from":"prc2","to":"tx","type":"inhibits"},{"from":"baf","to":"prc2","type":"inhibits"}],"interventions":["SMARCA2 degraders (PRT3789, PRT7732) in SMARCA4-mutant cancers (phase 1/2)","EZH2 inhibition in SMARCB1-deficient tumours (tazemetostat withdrawn 2026)","ATR/PARP inhibitors in ARID1A-mutant tumours (trials)","FHD-286 in AML"]},{"id":"synthetic-lethality-map","kind":"pathway","name":"Synthetic lethality: paired dependencies","aka":[],"tldr":"Two genes are synthetically lethal when losing either alone is fine but losing both kills the cell. Cancers that have already lost one (a tumour suppressor you cannot put back) become uniquely dependent on the other, which you can drug. BRCA and PARP was the first proof; a dozen more pairs are now in trials.","summary":"Genome-wide CRISPR screens (DepMap, Project Score) and clinical experience have mapped paired dependencies: BRCA1/2 or HRD → PARP1 and POLQ; ATM loss → ATR; TP53 loss or CCNE1 amplification → WEE1 and PKMYT1; MSI-H → WRN helicase; MTAP deletion (co-deleted with CDKN2A in ~15% of cancers) → PRMT5 and MAT2A (MTA-cooperative PRMT5 inhibitors AMG 193, MRTX1719); SMARCA4 loss → SMARCA2 (degraders); ARID1A loss → EZH2 and ATR; RB1 loss → Aurora kinase; VHL loss → HIF-2α (belzutifan is an oncogene-addiction-style version); KRAS mutation → SHP2, SOS1 (collateral dependencies). The concept also covers 'collateral lethality' (passenger deletions removing a paralogue) and 'induced essentiality' under therapy. Clinical success requires a clean biomarker, a selective inhibitor with a therapeutic window (PARP2 and marrow; PRMT5 in normal cells), and tolerance for resistance via restoration of the lost pathway (BRCA reversion) or loss of the dependency (53BP1).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Synthetic_lethality","links":[{"label":"Lord & Ashworth, PARP inhibitors: synthetic lethality in the clinic (Science 2017)","url":"https://doi.org/10.1126/science.aam7344"},{"label":"DepMap portal (Broad Institute)","url":"https://depmap.org/portal/"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["synthetic-lethality-approaches","crispr-screens","parp-inhibitor","hrd-testing","cgp","functional-drug-testing"],"targets":["parp","brca","wee1","atr","wrn","prmt5-mtap","tp53","hif2a","ezh2"],"drugs":["olaparib","niraparib","talazoparib","belzutifan"],"companies":[],"institutions":[],"pathways":["ddr","homologous-recombination-repair","mismatch-repair-msi","replication-stress","p53-mdm2-axis","hif-vhl"],"terms":["synthetic-lethality","hrd","msi","oncogene-addiction"],"trials":[],"people":[],"bottlenecks":["b-undruggable-targets"],"keyPapers":["paper-lord-science"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A building held up by two pillars. Knock one out and it still stands; nobody notices. But a tumour has already lost one pillar to get where it is, so the second pillar, harmless to attack in every normal cell, brings the whole tumour down when it goes.","nodes":[{"id":"brca","label":"BRCA / HRD loss","x":12,"y":10,"targetId":"brca"},{"id":"parp","label":"→ PARP1, POLQ","x":40,"y":10,"targetId":"parp"},{"id":"p53","label":"TP53 loss, CCNE1 amp","x":12,"y":32,"targetId":"tp53"},{"id":"wee1","label":"→ WEE1, PKMYT1, ATR","x":40,"y":32,"targetId":"wee1"},{"id":"msi","label":"MSI-H (MMR loss)","x":12,"y":54},{"id":"wrn","label":"→ WRN helicase","x":40,"y":54,"targetId":"wrn"},{"id":"mtap","label":"MTAP deletion","x":12,"y":76},{"id":"prmt5","label":"→ PRMT5, MAT2A","x":40,"y":76,"targetId":"prmt5-mtap"},{"id":"atm","label":"ATM loss","x":12,"y":96},{"id":"atr","label":"→ ATR","x":40,"y":96,"targetId":"atr"},{"id":"screen","label":"CRISPR screens (DepMap)","x":78,"y":20},{"id":"drug","label":"Selective inhibitor","x":78,"y":52},{"id":"res","label":"Resistance: restore lost path","x":78,"y":84}],"edges":[{"from":"brca","to":"parp","type":"activates"},{"from":"p53","to":"wee1","type":"activates"},{"from":"msi","to":"wrn","type":"activates"},{"from":"mtap","to":"prmt5","type":"activates"},{"from":"atm","to":"atr","type":"activates"},{"from":"screen","to":"drug","type":"activates"},{"from":"drug","to":"parp","type":"inhibits"},{"from":"drug","to":"wee1","type":"inhibits"},{"from":"drug","to":"wrn","type":"inhibits"},{"from":"drug","to":"prmt5","type":"inhibits"},{"from":"res","to":"drug","type":"inhibits"}],"interventions":["PARP inhibitors for BRCA/HRD (approved in four cancers)","MTA-cooperative PRMT5 inhibitors for MTAP-deleted tumours (mesothelioma, NSCLC, pancreatic; phase 1-2)","WEE1 (azenosertib), PKMYT1 (lunresertib) and ATR (ceralasertib, camonsertib) inhibitors in TP53-mutant, CCNE1-amplified and ATM-deficient tumours","WRN inhibitors for MSI-H; SMARCA2 degraders for SMARCA4-deficient cancers"]},{"id":"t-cell-exhaustion","kind":"pathway","name":"T-cell exhaustion","aka":[],"tldr":"T 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.","summary":"Chronic TCR stimulation with insufficient help drives a stepwise programme: TCF1+ progenitor-exhausted cells (stem-like, lymph-node and tertiary-lymphoid-structure resident, PD-1 intermediate) self-renew and give rise to transitory effector-like cells and then terminally exhausted cells (PD-1 high, TIM-3, CD39, CD101) with poor IL-2/TNF/IFN-γ production, driven by NFAT without AP-1, TOX and NR4A, and fixed by DNMT3A-dependent de novo methylation ('epigenetic scarring'). PD-1 blockade expands the progenitor pool and its transitory progeny rather than reviving terminal cells; response correlates with the size of the TCF1+ reservoir and with CD28 co-stimulation (PD-1 dephosphorylates CD28). Co-inhibitory receptors partition: LAG-3 (MHC-II, FGL1) blocked by relatlimab; TIGIT (CD155) by tiragolumab (mixed phase 3 results); TIM-3 (galectin-9, CEACAM1) by sabatolimab. CAR-T cells exhaust the same way (CD8 CAR-T with c-Jun overexpression or TET2 loss resist it); interval rests and PD-1 knockout are engineering countermeasures.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/T_cell_exhaustion","links":[{"label":"Wherry & Kurachi, Molecular and cellular insights into T cell exhaustion (Nat Rev Immunol 2015)","url":"https://doi.org/10.1038/nri3862"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["checkpoint-inhibitor","car-t","til-therapy","cytokine-therapy","armored-car"],"targets":["pd1","pdl1","lag3","tigit","tim3","ctla4","cd19"],"drugs":["pembrolizumab","nivolumab","relatlimab-nivolumab","fianlimab","tiragolumab","nogapendekin-alfa","bempegaldesleukin"],"companies":[],"institutions":[],"pathways":["pd1-checkpoint","antigen-presentation-immunoediting","cancer-immunity-cycle","epigenetic-reprogramming"],"terms":["tils","cold-vs-hot","irae","avoiding-immune-destruction"],"trials":[],"people":[],"bottlenecks":["b-immunotherapy-response"],"keyPapers":["paper-wherry-nat-rev-immunol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"chronic","label":"Chronic antigen, no help","x":12,"y":15},{"id":"tcf1","label":"TCF1+ progenitor (stem-like)","x":40,"y":15},{"id":"trans","label":"Transitory effector","x":68,"y":15},{"id":"term","label":"Terminally exhausted","x":90,"y":40},{"id":"tox","label":"TOX, NR4A, NFAT","x":40,"y":45},{"id":"epi","label":"DNMT3A epigenetic scar","x":68,"y":45},{"id":"brakes","label":"PD-1, LAG-3, TIM-3, TIGIT","x":12,"y":50,"targetId":"pd1"},{"id":"kill","label":"Cytokines, killing","x":40,"y":82},{"id":"block","label":"Checkpoint blockade","x":12,"y":82,"targetId":"lag3"},{"id":"cart","label":"CAR-T exhaustion","x":78,"y":82,"targetId":"cd19"}],"edges":[{"from":"chronic","to":"tcf1","type":"activates"},{"from":"tcf1","to":"trans","type":"activates"},{"from":"trans","to":"term","type":"activates"},{"from":"chronic","to":"tox","type":"activates"},{"from":"tox","to":"brakes","type":"activates"},{"from":"tox","to":"epi","type":"activates"},{"from":"epi","to":"term","type":"activates"},{"from":"brakes","to":"kill","type":"inhibits"},{"from":"trans","to":"kill","type":"activates"},{"from":"block","to":"brakes","type":"inhibits"},{"from":"block","to":"tcf1","type":"activates"},{"from":"chronic","to":"cart","type":"activates"}],"interventions":["Anti-PD-1/PD-L1; combinations with anti-LAG-3 (relatlimab-nivolumab, fianlimab) and anti-TIGIT (tiragolumab, mixed results)","Earlier use (neoadjuvant) when the TCF1+ reservoir is larger; IL-2 variants and IL-15 superagonists to expand progenitors","CAR-T engineering: c-Jun overexpression, TET2/DNMT3A editing, transient rest, PD-1 knockout","Epigenetic drugs to reverse scarring (preclinical)"]},{"id":"telomere-maintenance","kind":"pathway","name":"Telomere maintenance & replicative immortality","aka":[],"tldr":"Normal cells can divide only a limited number of times because the protective caps on their chromosomes, telomeres, wear down. About 90% of cancers switch the cap-rebuilding enzyme telomerase back on, often through TERT promoter mutations, and roughly 10% use an alternative lengthening route (ALT), so they divide indefinitely; imetelstat is the first approved telomerase inhibitor.","summary":"About 90% of cancers reactivate telomerase (TERT), often via TERT promoter mutations (glioblastoma, melanoma, bladder, thyroid) or amplification; ~10% use alternative lengthening of telomeres (ALT) through homologous recombination, associated with ATRX/DAXX loss (sarcomas, pancreatic NETs, gliomas). Telomerase inhibition (imetelstat, an oligonucleotide) reached approval in lower-risk MDS (2024) and is in phase 3 in myelofibrosis; ALT cells are sensitive to ATR inhibition. TERT promoter mutation is a diagnostic and prognostic marker and a candidate for ctDNA detection in bladder cancer and glioma.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Telomerase","links":[{"label":"Shay & Wright, Telomeres and telomerase: three decades of progress (Nature Reviews Genetics 2019)","url":"https://doi.org/10.1038/s41576-019-0099-1"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":[],"targets":["atr"],"drugs":[],"companies":[],"institutions":["johns-hopkins","dana-farber","ucsf"],"pathways":["replication-stress","senescence"],"terms":["tert-promoter","enabling-replicative-immortality"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-shay-nat-rev-genet"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Greider (Johns Hopkins, Nobel 2009); Meyerson (Dana-Farber) on TERT promoter genetics; Costello (UCSF) on TERT in glioma; Reddel (CMRI Sydney) on ALT."],"analogy":"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.","nodes":[{"id":"div","label":"Each division shortens telomeres","x":15,"y":30},{"id":"crisis","label":"Senescence / crisis","x":45,"y":30},{"id":"tert","label":"TERT reactivation (promoter mutation)","x":45,"y":70},{"id":"alt","label":"ALT (ATRX/DAXX loss)","x":80,"y":70},{"id":"immortal","label":"Replicative immortality","x":80,"y":30},{"id":"atr","label":"ATR dependence (ALT)","x":92,"y":90,"targetId":"atr"}],"edges":[{"from":"div","to":"crisis","type":"activates"},{"from":"tert","to":"crisis","type":"inhibits"},{"from":"alt","to":"crisis","type":"inhibits"},{"from":"tert","to":"immortal","type":"activates"},{"from":"alt","to":"immortal","type":"activates"},{"from":"alt","to":"atr","type":"activates"}],"interventions":["Imetelstat (telomerase inhibitor) approved in MDS (2024); solid-tumour use unproven","ATR inhibitors in ALT-positive tumours","TERT promoter mutation as a urine/plasma biomarker (bladder, glioma)"]},{"id":"tgf-beta","kind":"pathway","name":"TGF-β signalling","aka":[],"tldr":"A 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.","summary":"TGF-β binds TGFBR2/TGFBR1 (ALK5) to phosphorylate SMAD2/3, which with SMAD4 regulates transcription. Early tumour suppressor (cytostatic; SMAD4 and TGFBR2 loss in pancreatic and MSI colorectal cancer) and later promoter of EMT, CAF activation, immune exclusion (T-cell exclusion in bladder and CRC), and metastasis. Bintrafusp alfa (PD-L1/TGF-β trap) failed in phase 3 (NSCLC, biliary); galunisertib was discontinued; SRK-181 (latent TGF-β1) and dalutrafusp continue. Blocking TGF-β to re-sensitise cold tumours remains an active, unproven idea.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/TGF_beta_signaling_pathway","links":[{"label":"Tauriello et al., TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis (Nature 2018)","url":"https://doi.org/10.1038/nature25492"}],"tags":["mechanism"],"related":[],"cancers":["pancreatic"],"sections":[],"technologies":[],"targets":["acvr1"],"drugs":[],"companies":[],"institutions":["mskcc","vall-dhebron","cold-spring-harbor"],"pathways":["emt","tumor-microenvironment","metastatic-cascade","pd1-checkpoint"],"terms":["immune-exclusion","cold-vs-hot","desmoplasia"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-tauriello-tgfbeta-immune-evasion-colorectal-nature-2018"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Massagué (MSK, TGF-β biology and metastasis); Tauriello & Batlle (IRB Barcelona, TGF-β and immune exclusion in CRC); Sheppard (UCSF, integrin activation)."],"analogy":"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).","nodes":[{"id":"lig","label":"Latent TGF-β (activated by integrins)","x":12,"y":40},{"id":"rec","label":"TGFBR2 / ALK5","x":35,"y":40},{"id":"smad","label":"SMAD2/3-SMAD4","x":58,"y":40},{"id":"cyto","label":"Cytostasis (early)","x":80,"y":15},{"id":"emt","label":"EMT, CAF activation","x":80,"y":45},{"id":"excl","label":"T-cell exclusion","x":80,"y":75},{"id":"smad4","label":"SMAD4 loss (PDAC)","x":35,"y":80}],"edges":[{"from":"lig","to":"rec","type":"activates"},{"from":"rec","to":"smad","type":"activates"},{"from":"smad","to":"cyto","type":"activates"},{"from":"smad","to":"emt","type":"activates"},{"from":"smad","to":"excl","type":"activates"},{"from":"smad4","to":"smad","type":"inhibits"}],"interventions":["TGF-β traps and antibodies (mostly failed: bintrafusp alfa); latent-TGF-β1-selective agents in trials","ALK5 inhibitors (vactosertib) with IO in trials","Integrin αvβ6/αvβ8 blockade to prevent activation (investigational)"]},{"id":"angiogenic-switch","kind":"pathway","name":"The angiogenic switch & tumour vessels","aka":[],"tldr":"A 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.","summary":"Hypoxia (HIF), oncogenes (RAS, MYC), p53 loss (which normally induces thrombospondin-1) and inflammatory cells tip the balance of pro-angiogenic (VEGF-A/C, FGF2, PDGF-B, ANG2, PlGF) over anti-angiogenic (TSP-1, endostatin, angiostatin) factors. Endothelial tip cells (VEGFR2, DLL4-Notch selection) sprout, stalk cells proliferate, pericytes (PDGFRβ) stabilise; ANG2 destabilises. Alternatives: vessel co-option (liver, lung, brain metastases; a mechanism of bevacizumab resistance), intussusception, vasculogenic mimicry, and bone-marrow-derived endothelial progenitors. Tumour vessels are tortuous, leaky, poorly pericyte-covered and lack hierarchy, producing heterogeneous hypoxia, acidity, high interstitial pressure (which collapses vessels further and blocks drug penetration), and easy intravasation. Anti-VEGF therapy at moderate doses 'normalises' vessels transiently, improving perfusion, drug delivery and T-cell entry, the basis of IO-VEGF combinations. Resistance: alternative ligands (FGF, ANG2, PlGF), myeloid cells, co-option, and increased invasiveness.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Angiogenesis","links":[{"label":"Bergers & Benjamin, Tumorigenesis and the angiogenic switch (Nat Rev Cancer 2003)","url":"https://doi.org/10.1038/nrc1093"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["rcc","hcc","colorectal","glioblastoma","ovarian"],"sections":[],"technologies":["antiangiogenic","hypoxia-activated-therapy","kinase-inhibitors"],"targets":["vegf","hif2a","pd1","csf1r","pdgfra"],"drugs":["bevacizumab","ramucirumab","axitinib","cabozantinib","lenvatinib","sunitinib","pazopanib","tivozanib","fruquintinib","regorafenib","sorafenib","nintedanib","ivonescimab","belzutifan"],"companies":[],"institutions":[],"pathways":["vegf-angiogenesis","hif-vhl","tumor-microenvironment","intravasation-ctc-survival"],"terms":["inducing-angiogenesis"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bergers-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"hyp","label":"Hypoxia (HIF), RAS, p53 loss","x":12,"y":15,"targetId":"hif2a"},{"id":"pro","label":"VEGF, FGF2, ANG2, PDGF","x":40,"y":15,"targetId":"vegf"},{"id":"anti","label":"TSP-1, endostatin","x":40,"y":45},{"id":"switch","label":"Angiogenic switch","x":68,"y":30},{"id":"tip","label":"Tip / stalk sprouting (VEGFR2)","x":90,"y":50},{"id":"peri","label":"Pericytes (PDGFRβ)","x":68,"y":65},{"id":"vess","label":"Leaky, chaotic vessels","x":90,"y":82},{"id":"coopt","label":"Vessel co-option","x":40,"y":82},{"id":"out","label":"Hypoxia, poor delivery, exit","x":12,"y":82},{"id":"myeloid","label":"TAMs, MDSCs (bypass)","x":12,"y":48,"targetId":"csf1r"}],"edges":[{"from":"hyp","to":"pro","type":"activates"},{"from":"pro","to":"switch","type":"activates"},{"from":"anti","to":"switch","type":"inhibits"},{"from":"switch","to":"tip","type":"activates"},{"from":"tip","to":"vess","type":"activates"},{"from":"peri","to":"vess","type":"inhibits"},{"from":"vess","to":"out","type":"activates"},{"from":"coopt","to":"out","type":"activates"},{"from":"myeloid","to":"pro","type":"activates"},{"from":"out","to":"hyp","type":"activates"}],"interventions":["Anti-VEGF antibodies (bevacizumab, ramucirumab) and VEGFR TKIs (axitinib, cabozantinib, lenvatinib, sunitinib, pazopanib, tivozanib, fruquintinib, regorafenib, sorafenib)","Vascular normalisation windows for IO-VEGF combinations in RCC, HCC, endometrial cancer; PD-1×VEGF bispecific ivonescimab","HIF-2α inhibition upstream (belzutifan); multikinase inhibitors hit FGFR/PDGFR escape ligands","Vessel co-option limits anti-angiogenics in liver and brain metastases"]},{"id":"blood-brain-barrier-metastasis","kind":"pathway","name":"The blood-brain barrier & brain metastasis","aka":[],"tldr":"The 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.","summary":"The BBB is endothelium with continuous tight junctions (claudin-5, occludin), pericyte coverage, astrocyte end-feet, and ABC efflux transporters (P-gp/ABCB1, BCRP/ABCG2) that limit drug entry; large molecules such as antibodies achieve ~0.1% of plasma levels. Metastasising cells arrest in capillary branch points, extravasate over days using cathepsin S, COX-2, HBEGF, ST6GALNAC5, and survive by hugging vessels (vascular co-option via L1CAM and β1 integrin). Reactive astrocytes first attack (plasminogen activator → FasL), which tumour cells block with serpins (neuroserpin); then astrocytes are co-opted through connexin-43 gap junctions that transfer cGAMP, activating astrocytic STING and IFN-α/TNF that support tumour growth; glioma-like neuron-tumour synapses add to it. The blood-tumour barrier in established metastases is leakier but heterogeneous, so drug levels vary lesion to lesion. Brain-penetrant small molecules (osimertinib, lorlatinib, alectinib, tucatinib, capivasertib partially, temozolomide, lomustine) changed outcomes; T-DXd shows intracranial activity despite size; focused ultrasound, intrathecal delivery and LITT are engineering solutions. Meningeal disease and leptomeningeal spread remain the hardest compartment.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Blood–brain_barrier","links":[{"label":"Chen et al., Carcinoma-astrocyte gap junctions promote brain metastasis by cGAMP transfer (Nature 2016)","url":"https://doi.org/10.1038/nature18268"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["nsclc","breast-her2-positive","melanoma","sclc","glioblastoma"],"sections":[],"technologies":["bbb-focused-ultrasound","litt","sbrt","prophylactic-cranial-irradiation","mri","glioma-car-t"],"targets":["egfr","alk","her2","akt"],"drugs":["osimertinib","lorlatinib","alectinib","tucatinib","trastuzumab-deruxtecan","temozolomide","lomustine","capivasertib"],"companies":[],"institutions":[],"pathways":["organ-tropism-seed-soil","cgas-sting","cancer-neuroscience","drug-efflux-pumps","metastatic-cascade"],"terms":["blood-brain-barrier","her2-brain-metastases","efflux-pump","mgmt"],"trials":[],"people":[],"bottlenecks":["b-brain-delivery"],"keyPapers":["paper-chen-nature-2016"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"ctc","label":"CTC arrest at capillary","x":12,"y":15},{"id":"bbb","label":"BBB: tight junctions, pericytes","x":45,"y":15},{"id":"pgp","label":"P-gp / BCRP efflux","x":78,"y":15},{"id":"extra","label":"Extravasation (cathepsin S)","x":12,"y":48},{"id":"coopt","label":"Vascular co-option (L1CAM)","x":45,"y":48},{"id":"astro","label":"Astrocytes: Cx43, cGAMP","x":78,"y":48},{"id":"met","label":"Brain metastasis","x":45,"y":82},{"id":"drug","label":"Drugs excluded","x":92,"y":82},{"id":"tki","label":"Brain-penetrant TKIs, FUS","x":12,"y":82,"targetId":"her2"}],"edges":[{"from":"ctc","to":"extra","type":"activates"},{"from":"bbb","to":"extra","type":"inhibits"},{"from":"extra","to":"coopt","type":"activates"},{"from":"coopt","to":"met","type":"activates"},{"from":"astro","to":"met","type":"activates"},{"from":"met","to":"astro","type":"activates"},{"from":"pgp","to":"drug","type":"activates"},{"from":"bbb","to":"drug","type":"activates"},{"from":"drug","to":"met","type":"activates"},{"from":"tki","to":"met","type":"inhibits"},{"from":"tki","to":"pgp","type":"inhibits"}],"interventions":["Brain-penetrant TKIs: osimertinib (EGFR), lorlatinib/alectinib (ALK), tucatinib (HER2, HER2CLIMB), plus T-DXd intracranial activity","Radiosurgery and hippocampal-sparing whole-brain RT; prophylactic cranial irradiation vs MRI surveillance in SCLC","Focused ultrasound BBB opening, LITT, intrathecal and intraventricular delivery for leptomeningeal disease","Astrocyte gap-junction (meclofenamate, tonabersat) and STING-axis blockade in trials"]},{"id":"cancer-immunity-cycle","kind":"pathway","name":"The cancer-immunity cycle","aka":[],"tldr":"Seven 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.","summary":"Chen and Mellman's cycle: (1) dying tumour cells release neoantigens; (2) dendritic cells (BATF3+ cDC1) capture and cross-present them; (3) in lymph nodes, T cells are primed via TCR-MHC plus CD28-B7, with CTLA-4 as the brake; (4) effector T cells traffic via CXCL9/10-CXCR3; (5) infiltrate through vasculature and stroma; (6) recognise peptide-MHC-I; (7) kill via perforin/granzyme and IFN-γ, releasing more antigen. Each step has failure modes (low antigenicity, poor DC function, Treg-dominated priming, abnormal vessels, TGF-β stroma, MHC loss, PD-1 exhaustion) and matching drugs (radiation, STING agonists and vaccines for 1-2; anti-CTLA-4 for 3; anti-VEGF for 4-5; engagers and CAR-T bypass 6; anti-PD-1 for 7).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Cancer_immunology","links":[{"label":"Chen & Mellman, Oncology meets immunology: the cancer-immunity cycle (Immunity 2013)","url":"https://doi.org/10.1016/j.immuni.2013.07.012"},{"label":"Chen & Mellman, Elements of cancer immunity and the cancer-immune set point (Nature 2017)","url":"https://doi.org/10.1038/nature21349"}],"tags":["mechanism","mechanics-atlas"],"related":["theories-of-cancer","immune-surveillance-immunoediting"],"cancers":[],"sections":[],"technologies":["checkpoint-inhibitor","neoantigen-mrna-vaccine","sting-agonist","t-cell-engager","car-t","oncolytic-virus","sbrt"],"targets":["pd1","pdl1","ctla4","vegf","cd3","cd137","ido1"],"drugs":["pembrolizumab","nivolumab","ipilimumab","ivonescimab","intismeran-autogene"],"companies":[],"institutions":[],"pathways":["pd1-checkpoint","antigen-presentation-immunoediting","cgas-sting","vegf-angiogenesis"],"terms":["neoantigen","immunogenic-cell-death","cold-vs-hot","tils"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-chen-nature"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"rel","label":"1 Antigen release","x":15,"y":15},{"id":"dc","label":"2 DC capture (cDC1)","x":45,"y":15},{"id":"prime","label":"3 Priming (CD28 / CTLA-4)","x":78,"y":15,"targetId":"ctla4"},{"id":"traf","label":"4 Trafficking (CXCL9/10)","x":88,"y":50},{"id":"inf","label":"5 Infiltration","x":62,"y":50,"targetId":"vegf"},{"id":"rec","label":"6 Recognition (MHC-I)","x":45,"y":82},{"id":"kill","label":"7 Killing (PD-1 brake)","x":15,"y":82,"targetId":"pd1"},{"id":"block","label":"Escape at any step","x":15,"y":50}],"edges":[{"from":"rel","to":"dc","type":"activates"},{"from":"dc","to":"prime","type":"activates"},{"from":"prime","to":"traf","type":"activates"},{"from":"traf","to":"inf","type":"activates"},{"from":"inf","to":"rec","type":"activates"},{"from":"rec","to":"kill","type":"activates"},{"from":"kill","to":"rel","type":"activates"},{"from":"block","to":"kill","type":"inhibits"},{"from":"block","to":"rel","type":"inhibits"}],"interventions":["Radiation, chemotherapy, oncolytic viruses and ADC payloads feed step 1 (immunogenic cell death)","Vaccines and STING agonists load step 2; anti-CTLA-4 acts at step 3","Anti-VEGF and stromal agents open steps 4-5","Engagers, CAR-T and TCR-T replace step 6; anti-PD-1/PD-L1 releases step 7"]},{"id":"cell-cycle-engine-cdks","kind":"pathway","name":"The cell-cycle engine (cyclins & CDKs)","aka":[],"tldr":"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.","summary":"Mitogens induce cyclin D, which with CDK4/6 mono-phosphorylates RB; cyclin E-CDK2 completes RB hyperphosphorylation, releasing E2F and committing the cell at the restriction point (the point of no return). Cyclin A-CDK2 drives S phase, cyclin A/B-CDK1 drive G2 and mitosis, with CDK1 activation controlled by WEE1/PKMYT1 (inhibitory phosphorylation) and CDC25 phosphatases. CDK7 (CAK) activates all CDKs and also drives transcription. INK4 proteins (p16 from CDKN2A) inhibit CDK4/6; CIP/KIP (p21, p27) inhibit CDK2; APC/C and SCF ubiquitin ligases destroy cyclins in order. Cancer alterations: CCND1 amplification, CDK4 amplification (liposarcoma), CDKN2A deletion (very common, also removes ARF), CCNE1 amplification (ovarian, gastric, endocrine-resistant breast), RB1 loss (SCLC, TNBC, resistant prostate). CDK4/6 inhibitors need intact RB; CCNE1 amplification and RB1 loss are the escape routes and the rationale for CDK2 inhibitors and CDK4-selective atirmociclib.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Cyclin-dependent_kinase","links":[{"label":"Malumbres & Barbacid, Cell cycle, CDKs and cancer: a changing paradigm (Nat Rev Cancer 2009)","url":"https://doi.org/10.1038/nrc2602"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["breast-hr-positive","sarcoma","ovarian","sclc","gallbladder"],"sections":[],"technologies":["cdk46-inhibitor","endocrine-therapy"],"targets":["cdk4-6","wee1","estrogen-receptor","tp53"],"drugs":["palbociclib","ribociclib","abemaciclib","atirmociclib"],"companies":[],"institutions":[],"pathways":["p53-cell-cycle","er-signaling","replication-stress","mitotic-spindle-checkpoint"],"terms":["endocrine-resistance","evading-growth-suppressors"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-malumbres-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"mit","label":"Mitogens (ER, RTK, RAS)","x":12,"y":12,"targetId":"estrogen-receptor"},{"id":"d46","label":"Cyclin D-CDK4/6","x":12,"y":42,"targetId":"cdk4-6"},{"id":"p16","label":"p16 (CDKN2A)","x":12,"y":75},{"id":"rb","label":"RB → E2F released","x":40,"y":42},{"id":"e2","label":"Cyclin E-CDK2","x":40,"y":72},{"id":"p21","label":"p21 / p27","x":40,"y":12,"targetId":"tp53"},{"id":"s","label":"S phase (cyclin A)","x":68,"y":72},{"id":"b1","label":"Cyclin B-CDK1","x":68,"y":42},{"id":"wee1","label":"WEE1 / PKMYT1","x":90,"y":20,"targetId":"wee1"},{"id":"m","label":"Mitosis","x":90,"y":72}],"edges":[{"from":"mit","to":"d46","type":"activates"},{"from":"p16","to":"d46","type":"inhibits"},{"from":"d46","to":"rb","type":"activates"},{"from":"p21","to":"e2","type":"inhibits"},{"from":"rb","to":"e2","type":"activates"},{"from":"e2","to":"rb","type":"activates"},{"from":"e2","to":"s","type":"activates"},{"from":"s","to":"b1","type":"activates"},{"from":"wee1","to":"b1","type":"inhibits"},{"from":"b1","to":"m","type":"activates"}],"interventions":["CDK4/6 inhibitors (palbociclib, ribociclib, abemaciclib) in HR+ breast cancer; CDK4-selective atirmociclib to spare neutrophils","CDK2 inhibitors for CCNE1-amplified and CDK4/6-resistant disease (trials)","WEE1 (azenosertib) and PKMYT1 (lunresertib) inhibitors force premature mitosis in CCNE1-amplified or TP53-mutant cells","CDK7 and CDK9 inhibitors hit transcription as well as the cycle"]},{"id":"germinal-centre-reaction","kind":"pathway","name":"The germinal centre reaction","aka":["Germinal centre","Germinal center reaction","Somatic hypermutation and class switching","Affinity maturation"],"tldr":"To make a good antibody, a B cell has to deliberately damage its own DNA and keep dividing while it does. The germinal centre is where that happens, under strict time limits. Most B-cell lymphomas are cells that went through it and did not come out.","summary":"When a B cell meets an antigen it has not seen before, it enters a lymph node follicle and starts a germinal centre. There it does three dangerous things at once. It switches on activation-induced cytidine deaminase, which deliberately mutates the variable region of its own immunoglobulin genes so that a better-binding version can be selected. It cuts and rejoins the constant region to change antibody class. And it divides faster than almost any cell in the body. To survive this, BCL6 holds down the DNA damage response and the differentiation programme, polycomb repressive complex 2 keeps the exit genes methylated shut, and CREBBP and EP300 provide the acetylation that turns the programme back off at the end.\n\nAlmost every lesion in B-cell lymphoma is a failure of one of those safeguards. Translocations of BCL2 and MYC to an immunoglobulin locus are mistakes made by the recombinases that cut these genes on purpose. EZH2 gain-of-function mutations hold the polycomb mark on, so the exit never happens. CREBBP and EP300 loss leaves BCL6 acetylated less and therefore active more. The cell of origin classification of diffuse large B-cell lymphoma is a description of where in this reaction the cancer got stuck: germinal-centre-like cells are still inside it, activated B-cell-like cells are at the exit and cannot complete it.\n\nThe reaction is also why these cancers keep their normal lineage's weaknesses. A germinal-centre B cell still carries CD19, CD20, CD79b and CD22, which is why antibodies work and why they strip out the healthy B-cell compartment at the same time.","asOf":"2026-09-30","links":[{"label":"Tsujimoto et al., Science 1985: the t(14;18) translocation results from a mistake in VDJ joining","url":"https://doi.org/10.1126/science.3929382"},{"label":"Morin et al., Nat Genet 2010: somatic EZH2 Tyr641 mutations in follicular and germinal-centre diffuse large B-cell lymphoma","url":"https://doi.org/10.1038/ng.518"},{"label":"Pasqualucci et al., Nature 2011: inactivating mutations of the acetyltransferase genes CREBBP and EP300 in B-cell lymphoma","url":"https://doi.org/10.1038/nature09730"},{"label":"Alizadeh et al., Nature 2000: distinct types of diffuse large B-cell lymphoma identified by gene expression profiling","url":"https://doi.org/10.1038/35000501"}],"tags":["pathway","lymphoma"],"related":[],"cancers":["dlbcl","follicular-lymphoma","burkitt-lymphoma","non-hodgkin-lymphoma","primary-mediastinal-b-cell-lymphoma"],"sections":[],"technologies":[],"targets":["bcl6","ezh2","crebbp","ep300","bcl2","myc-gene","kmt2d"],"drugs":[],"companies":[],"institutions":[],"pathways":["epigenetic-reprogramming","apoptosis-bcl2","bcr-signalling","myc"],"terms":["lymphoma-bio-germinal-centre","cell-of-origin"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"A proof-reading workshop where the only way to improve a design is to make random changes to the blueprint and test them. While the workshop is running, the fire alarm (p53 and the damage response) is switched off and the exit door (differentiation) is locked, because nobody must leave mid-experiment. Lymphoma is what happens when a worker jams the lock or cuts the alarm wire permanently.","nodes":[{"id":"antigen","label":"Antigen and T-cell help (CD40, IL-21)","x":50,"y":5},{"id":"bcl6","label":"BCL6 switches off the damage response","x":25,"y":26,"targetId":"bcl6"},{"id":"aid","label":"AID mutates the immunoglobulin genes","x":72,"y":26},{"id":"prc2","label":"PRC2 / EZH2 keeps the exit genes shut","x":25,"y":48,"targetId":"ezh2"},{"id":"select","label":"Selection in the light zone: better binders survive","x":72,"y":48},{"id":"hat","label":"CREBBP / EP300 acetylation releases the brake","x":50,"y":68,"targetId":"crebbp"},{"id":"exit","label":"Exit as a plasma cell or memory B cell","x":25,"y":88},{"id":"lymphoma","label":"Stuck: BCL2 and MYC translocation, EZH2 gain, CREBBP loss","x":75,"y":88,"targetId":"bcl2"}],"edges":[{"from":"antigen","to":"bcl6","type":"activates"},{"from":"antigen","to":"aid","type":"activates"},{"from":"bcl6","to":"prc2","type":"activates"},{"from":"aid","to":"select","type":"activates"},{"from":"prc2","to":"hat","type":"activates"},{"from":"select","to":"hat","type":"activates"},{"from":"hat","to":"exit","type":"activates"},{"from":"hat","to":"lymphoma","type":"inhibits"},{"from":"prc2","to":"lymphoma","type":"activates"},{"from":"bcl6","to":"exit","type":"inhibits"}],"interventions":["EZH2 inhibition (tazemetostat) in follicular lymphoma, which is the only licensed drug that acts on this reaction directly and the only one selected by a germinal-centre genotype","BCL-2 inhibition (venetoclax) against the anti-apoptotic protein the t(14;18) translocation put there","HDAC inhibition as the proposed answer to CREBBP loss, on the argument that a cell short of acetyltransferase is dependent on keeping the deacetylase in check; in trials, not approved for this indication","BCL6 degraders and inhibitors, in early trials","Every anti-CD19, anti-CD20, anti-CD79b and anti-CD22 medicine, which work because the lymphoma kept the surface of the normal cell it came from"]},{"id":"metastatic-cascade","kind":"pathway","name":"The metastatic cascade","aka":[],"tldr":"How 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.","summary":"Invasion (EMT, matrix proteases), intravasation, survival in circulation as CTCs or clusters (platelet cloaking, anoikis resistance), arrest and extravasation, then colonisation, which is the rate-limiting step: most disseminated tumour cells (DTCs) die or stay dormant. Organ tropism follows Paget's seed-and-soil: pre-metastatic niches are prepared by tumour-derived exosomes and myeloid cells. Clinically, ctDNA and MRD assays detect the cascade before imaging; adjuvant therapy targets it blind; no drug specifically blocks colonisation yet.","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Metastasis","links":[{"label":"Massagué & Obenauf, Metastatic colonization by circulating tumour cells (Nature 2016)","url":"https://doi.org/10.1038/nature17038"},{"label":"Gerstberger, Jiang & Ganesh, Metastasis (Cell 2023)","url":"https://doi.org/10.1016/j.cell.2023.03.003"}],"tags":["mechanism"],"related":[],"cancers":[],"sections":[],"technologies":["mrd-testing","liquid-biopsy","sbrt","single-cell-spatial"],"targets":[],"drugs":[],"companies":[],"institutions":["mskcc","francis-crick","cold-spring-harbor","dkfz"],"pathways":["emt","tumor-dormancy","tumor-microenvironment","clonal-evolution"],"terms":["mrd","ctdna","oligometastatic","disseminated-tumor-cells"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-massague-nature","paper-gerstberger-cell"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Massagué lab (MSK) on metastasis-initiating cells and dormancy; Swanton/TRACERx (Crick) on evolution and dissemination timing; Egeblad (CSHL) on neutrophils and dormancy awakening; DKFZ Metastasis Research.","Massagué & Ganesh, Cancer Cell 2023 review: https://doi.org/10.1016/j.ccell.2023.02.020"],"analogy":"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.","nodes":[{"id":"primary","label":"Primary tumour","x":8,"y":50},{"id":"emt","label":"Invasion (EMT, MMPs)","x":22,"y":50},{"id":"intra","label":"Intravasation","x":36,"y":50},{"id":"ctc","label":"CTCs in circulation","x":50,"y":50},{"id":"extra","label":"Extravasation","x":64,"y":50},{"id":"dtc","label":"DTC dormancy","x":78,"y":35},{"id":"colon","label":"Colonisation / macrometastasis","x":92,"y":60},{"id":"niche","label":"Pre-metastatic niche (exosomes, myeloid cells)","x":64,"y":82},{"id":"immune","label":"NK / T-cell clearance","x":50,"y":18}],"edges":[{"from":"primary","to":"emt","type":"activates"},{"from":"emt","to":"intra","type":"activates"},{"from":"intra","to":"ctc","type":"activates"},{"from":"ctc","to":"extra","type":"activates"},{"from":"extra","to":"dtc","type":"activates"},{"from":"dtc","to":"colon","type":"activates"},{"from":"extra","to":"colon","type":"activates"},{"from":"primary","to":"niche","type":"activates"},{"from":"niche","to":"colon","type":"activates"},{"from":"immune","to":"ctc","type":"inhibits"},{"from":"immune","to":"dtc","type":"inhibits"}],"interventions":["Adjuvant systemic therapy and ctDNA-guided escalation aim at DTCs before colonisation","Anti-EMT and anti-MMP drugs failed historically; TGF-β blockade is being retried in combinations","Dormancy-maintaining strategies (see tumour dormancy) are the newest idea","Metastasis-directed SBRT for oligometastatic disease"]},{"id":"p53-mdm2-axis","kind":"pathway","name":"The p53 network (guardian of the genome)","aka":[],"tldr":"p53 is the cell's emergency coordinator: DNA damage, oncogene stress or lack of oxygen switch it on, and it then pauses division, orders repairs, or triggers suicide or permanent retirement, while MDM2 keeps it off in healthy cells. About half of cancers mutate p53 outright, and sarcomas, gliomas, melanomas and retinoblastomas silence it instead by amplifying MDM2 or MDM4.","summary":"Stress inputs converge on p53 stabilisation: DNA damage via ATM/ATR-CHK2/CHK1 phosphorylation, oncogene activation via p14ARF (CDKN2A) sequestering MDM2, ribosomal stress via RPL5/RPL11, hypoxia. MDM2 (with MDMX/MDM4) ubiquitinates p53 for proteasomal degradation and is itself a p53 target, forming a negative feedback loop. Active p53 tetramers transactivate CDKN1A (p21, arrest), PUMA/NOXA/BAX (apoptosis), GADD45/DDB2 (repair), TIGAR/SCO2 (metabolism), and senescence programmes; outcome depends on stress intensity and cofactors. TP53 mutations (~50% of cancers) are mostly missense DNA-binding-domain hotspots (R175, R248, R273) with dominant-negative and gain-of-function effects; MDM2 amplification (sarcoma, glioma) and MDM4 amplification (melanoma, retinoblastoma) silence wild-type p53. Drugs: MDM2 inhibitors (milademetan, brigimadlin, navtemadlin) in TP53-wild-type disease, limited by thrombocytopenia; Y220C reactivator rezatapopt; eprenetapopt (APR-246) failed in MDS; TP53 status guides WEE1/ATR synthetic lethality and predicts chemoresistance in CLL (del17p).","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/P53","links":[{"label":"Kastenhuber & Lowe, Putting p53 in context (Cell 2017)","url":"https://doi.org/10.1016/j.cell.2017.08.028"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["sarcoma","cll","mds","ovarian","gallbladder"],"sections":[],"technologies":["germline-testing","cgp","synthetic-lethality-approaches"],"targets":["tp53","mdm2","atr","wee1","bcl2"],"drugs":["eprenetapopt","venetoclax"],"companies":[],"institutions":[],"pathways":["p53-cell-cycle","apoptosis-bcl2","senescence","replication-stress"],"terms":["li-fraumeni","del17p-tp53","evading-growth-suppressors","germline-vs-somatic"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-kastenhuber-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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).","nodes":[{"id":"dmg","label":"DNA damage (ATM/ATR)","x":12,"y":12,"targetId":"atr"},{"id":"onc","label":"Oncogene stress → ARF","x":40,"y":12},{"id":"hyp","label":"Hypoxia, ribosome stress","x":70,"y":12},{"id":"mdm2","label":"MDM2 / MDMX","x":82,"y":42,"targetId":"mdm2"},{"id":"p53","label":"p53","x":45,"y":45,"targetId":"tp53"},{"id":"mut","label":"TP53 mutation (~50%)","x":12,"y":45},{"id":"arr","label":"p21 → arrest","x":12,"y":82},{"id":"apo","label":"PUMA, NOXA → apoptosis","x":40,"y":82,"targetId":"bcl2"},{"id":"sen","label":"Senescence, repair","x":70,"y":82}],"edges":[{"from":"dmg","to":"p53","type":"activates"},{"from":"onc","to":"mdm2","type":"inhibits"},{"from":"hyp","to":"p53","type":"activates"},{"from":"mdm2","to":"p53","type":"inhibits"},{"from":"p53","to":"mdm2","type":"activates"},{"from":"mut","to":"p53","type":"inhibits"},{"from":"p53","to":"arr","type":"activates"},{"from":"p53","to":"apo","type":"activates"},{"from":"p53","to":"sen","type":"activates"}],"interventions":["MDM2 inhibitors (brigimadlin, milademetan, navtemadlin) for TP53-wild-type, MDM2-amplified tumours","Mutant p53 reactivators: rezatapopt (Y220C); eprenetapopt failed in phase 3 MDS","TP53-mutant tumours are approached via WEE1, ATR, PLK1 dependence and via p53-independent chemotherapy","TP53 status as biomarker: del17p CLL, MDS/AML risk, Li-Fraumeni surveillance"]},{"id":"pre-metastatic-niche","kind":"pathway","name":"The pre-metastatic niche","aka":[],"tldr":"Before 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.","summary":"Tumour-secreted factors (VEGF-A, PlGF, TNF, TGF-β, G-CSF, LOX) and exosomes prime distant sites: exosomal integrins address organs (α6β4/α6β1 to lung fibroblasts and epithelium, αvβ5 to liver Kupffer cells), exosomal MIF induces hepatic stellate cells to secrete fibronectin and TGF-β, and PDAC exosomes prepare liver niches; tumour-derived LOX crosslinks collagen; S100A8/A9 and SAA3 recruit VEGFR1+ haematopoietic progenitors, neutrophils and monocytes, which lay fibronectin, secrete MMP9, and increase vascular permeability. Hypoxia in the primary tumour amplifies the programme. The niche provides adhesion (fibronectin, periostin, tenascin C), survival signals, immune suppression (MDSCs, Treg), and later awakening cues (NETs, inflammation). Lymph nodes are similarly pre-conditioned (lymphangiogenesis via VEGF-C). Clinically the niche explains organ tropism and the benefit of adjuvant therapy; exosome profiling (integrin patterns, protein cargo) is being tested as a predictor of the site of relapse; LOX inhibitors and CXCR2/CCR2 blockade are experimental.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Metastasis","links":[{"label":"Peinado et al., Pre-metastatic niches: organ-specific homes for metastases (Nat Rev Cancer 2017)","url":"https://doi.org/10.1038/nrc.2017.6"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":[],"sections":[],"technologies":["exosome-therapeutics","mrd-testing","liquid-biopsy","single-cell-spatial"],"targets":["vegf","hif2a","csf1r","cxcr4"],"drugs":["bevacizumab","signatera"],"companies":[],"institutions":[],"pathways":["metastatic-cascade","intravasation-ctc-survival","organ-tropism-seed-soil","tumor-dormancy","hif-vhl"],"terms":["disseminated-tumor-cells","myeloid-derived-suppressor-cells","mrd","late-recurrence"],"trials":[],"people":[],"bottlenecks":["b-metastasis-biology"],"keyPapers":["paper-peinado-nat-rev-cancer"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"prim","label":"Primary tumour (hypoxic)","x":12,"y":40,"targetId":"hif2a"},{"id":"exo","label":"Exosomes (integrins, MIF)","x":40,"y":15},{"id":"fact","label":"VEGF, G-CSF, LOX, S100A8/9","x":40,"y":65,"targetId":"vegf"},{"id":"bmdc","label":"Bone-marrow cells (VEGFR1+)","x":68,"y":65},{"id":"resid","label":"Resident cells: Kupffer, fibroblasts","x":68,"y":15},{"id":"niche","label":"Niche: fibronectin, MMP9, leaky","x":92,"y":40},{"id":"ctc","label":"Arriving CTCs","x":92,"y":82},{"id":"col","label":"Colonisation","x":62,"y":90},{"id":"mdsc","label":"MDSCs, immune suppression","x":30,"y":90,"targetId":"csf1r"}],"edges":[{"from":"prim","to":"exo","type":"activates"},{"from":"prim","to":"fact","type":"activates"},{"from":"exo","to":"resid","type":"activates"},{"from":"fact","to":"bmdc","type":"activates"},{"from":"resid","to":"niche","type":"activates"},{"from":"bmdc","to":"niche","type":"activates"},{"from":"bmdc","to":"mdsc","type":"activates"},{"from":"niche","to":"col","type":"activates"},{"from":"ctc","to":"col","type":"activates"},{"from":"mdsc","to":"col","type":"activates"}],"interventions":["Adjuvant systemic therapy and ctDNA-guided escalation act during niche formation and early seeding","Exosome and integrin profiling to predict organ of relapse (research); exosome-based therapeutics in early development","LOX, CXCR2 and CCR2 inhibitors, and G-CSF neutralisation in models","Anti-VEGF has not prevented metastasis in adjuvant trials (bevacizumab in colon and breast), a cautionary result"]},{"id":"theories-of-cancer","kind":"pathway","name":"Theories of cancer: how the ideas connect","aka":["theories of carcinogenesis","what is cancer","origin of cancer theories","cancer theory map"],"tldr":"Scientists have given more than a dozen answers to what cancer is: mutated genes, runaway evolution, misbehaving stem cells, disordered tissue, ancient cell programmes, broken chromosomes, metabolism, epigenetic switches, unhealed wounds, a failed immune system, ageing tissue, force or electricity. This map shows each theory, who proposed it, what held up, and how the ideas connect.","summary":"The mainstream trunk runs left to right along the top of the diagram. The somatic mutation theory (cancer starts when one cell accumulates mutations in growth-controlling genes) was refined into the driver and passenger model (only a handful of those mutations matter) and then set in motion by clonal evolution (Nowell 1976: the mutant cell's descendants compete, diversify and are selected, including by treatment). The hallmarks of cancer (Hanahan and Weinberg 2000, 2011, 2022) sit at the end of the trunk as a synthesis: whatever the origin, cancers converge on the same acquired capabilities.\n\nBelow the trunk are the theories that challenge or extend it. The aneuploidy theory (Boveri, Duesberg) says broken chromosomes, not point mutations, are the engine; modern work on chromosomal instability and chromothripsis has absorbed it as an accelerator rather than the sole cause. The epigenetic progenitor theory (Feinberg) says the first step is a reversible change in gene regulation, not a mutation, and it feeds the cancer stem cell and plasticity view (Dick, Weissman, Clevers), in which a tumour is a caricature of a tissue with a stem-like compartment that treatment misses. The metabolic theory (Warburg, Seyfried) puts damaged respiration first; its modern reading (Vander Heiden, Thompson) treats metabolic rewiring as a consequence of oncogenic signalling that is nonetheless druggable. The atavistic theory (Davies, Lineweaver, Vincent) frames all of this as a reversion to an ancient unicellular survival programme.\n\nThe tissue-level theories occupy the bottom row. The tissue organisation field theory (Sonnenschein and Soto) says cancer is a disease of tissue architecture in which mutations are secondary; the microenvironment and inflammation view (Virchow, Dvorak, Coussens, Bissell) says stroma, vessels and immune cells co-create the tumour; Paget's seed and soil hypothesis applies the same logic to where metastases grow; the immune surveillance and immunoediting theory (Burnet, Thomas, Schreiber) says the immune system continuously removes nascent tumours and sculpts the ones that survive. The ageing tissue view (DeGregori, Martincorena, Ebert) adds time: mutant clones fill normal tissue throughout life and cancer arrives when old tissue changes what is selected. The newest proposals are physical: the mechanical theory (Bissell, Weaver, Jain) treats stiffness and pressure as instructive signals, and the bioelectric theory (Levin) treats membrane voltage patterns as a tissue-level control layer that can override mutations in animal models.\n\nThe arrows record how the schools relate. Solid arrows mean one theory feeds another (driver mutations feed clonal evolution; inflammation and immune escape became hallmarks). Inhibiting arrows mean one theory was proposed against another (tissue organisation, aneuploidy, epigenetics, metabolism and bioelectricity were each framed as alternatives to the somatic mutation theory). In practice the field has stopped treating them as rivals: the 2022 hallmarks paper explicitly imports non-mutational epigenetic reprogramming, phenotypic plasticity, the microbiome and senescence, and the somatic mutation theory is now read as necessary but not sufficient.\n\nStatus: a scorecard rather than one verdict. Established: somatic mutation theory, driver and passenger model, clonal evolution, immune surveillance and immunoediting, seed and soil, hallmarks as a framework. Partly confirmed: cancer stem cells (real hierarchy in leukaemias, rigid hierarchy replaced by plasticity), epigenetic progenitor theory, aneuploidy and chromosomal instability, metabolic reprogramming, ageing tissue as a field, mechanical theory. Contested: tissue organisation field theory, atavistic theory, bioelectric theory. Superseded: the strong forms of the aneuploidy theory (chromosomes instead of genes) and of the metabolic theory (respiration damage as the origin), and Ewing's purely mechanical account of metastatic spread.","asOf":"2026-09-17","wikipedia":"https://en.wikipedia.org/wiki/Carcinogenesis","links":[{"label":"Sonnenschein and Soto, Theories of carcinogenesis: an emerging perspective (Seminars in Cancer Biology 2008)","url":"https://doi.org/10.1016/j.semcancer.2008.03.012"},{"label":"Hanahan, Hallmarks of Cancer: New Dimensions (Cancer Discovery 2022)","url":"https://doi.org/10.1158/2159-8290.CD-21-1059"},{"label":"Greaves and Maley, Clonal evolution in cancer (Nature 2012)","url":"https://doi.org/10.1038/nature10762"},{"label":"Wikipedia: Carcinogenesis","url":"https://en.wikipedia.org/wiki/Carcinogenesis"}],"tags":["mechanism","mechanics-atlas","theory"],"related":["somatic-mutation-theory","driver-passenger-model","clonal-evolution-theory","hallmarks-synthesis","cancer-stem-cell-theory","epigenetic-progenitor-theory","aneuploidy-theory-of-cancer","metabolic-theory-of-cancer","tissue-organisation-field-theory","microenvironment-inflammation-theory","seed-and-soil-hypothesis","immune-surveillance-immunoediting","ageing-tissue-field-theory","atavistic-theory-of-cancer","mechanical-theory-of-cancer","bioelectric-theory-of-cancer","hallmarks-of-cancer","mathematical-oncology"],"cancers":[],"sections":[],"technologies":[],"targets":[],"drugs":[],"companies":[],"institutions":[],"pathways":["oncogene-activation-two-hit","clonal-evolution","cancer-stem-cells-plasticity","epigenetic-reprogramming","chromosomal-instability","cancer-metabolism","tumor-microenvironment","cancer-immunity-cycle","organ-tropism-seed-soil","clonal-haematopoiesis","field-cancerisation","caf-activation-desmoplasia"],"terms":["hallmarks-of-cancer","driver-mutation","warburg-effect","inflammation","metastasis"],"trials":[],"people":["bert-vogelstein","mel-greaves","robert-gatenby","robert-schreiber","john-dick","matthew-vander-heiden","benjamin-ebert"],"bottlenecks":[],"keyPapers":["paper-hallmarks-of-cancer-cell-2000","paper-hallmarks-new-dimensions-cancer-discov-2022","paper-vogelstein-cancer-genome-landscapes-science-2013","paper-schreiber-cancer-immunoediting-science-2011","paper-greaves-nature","paper-sonnenschein-semin-cancer-biol"],"journals":[],"dependsOn":[],"notes":[],"analogy":"Sixteen detectives arguing over one crime scene. The geneticist blames the suspect's record (mutations), the ecologist blames the neighbourhood that let a small-time offender flourish (evolution and microenvironment), the architect blames the building (tissue organisation), the historian says the suspect is behaving as its ancestors did (atavism), the nutritionist blames what it ate (metabolism), the immunologist asks where the police were (surveillance), and the physicist checks the wiring (mechanics and bioelectricity). The hallmarks paper is the chief inspector's summary: whoever is right about the motive, here is what the culprit always does.","nodes":[{"id":"smt","label":"Somatic mutation theory","x":12,"y":10},{"id":"driver","label":"Drivers and passengers","x":38,"y":10},{"id":"clonal","label":"Clonal evolution and ecology","x":64,"y":10},{"id":"hall","label":"Hallmarks (synthesis)","x":90,"y":28},{"id":"aneu","label":"Aneuploidy and chromothripsis","x":12,"y":34},{"id":"epi","label":"Epigenetic progenitor","x":38,"y":34},{"id":"csc","label":"Cancer stem cells and plasticity","x":64,"y":34},{"id":"atav","label":"Atavism (ancient programme)","x":12,"y":58},{"id":"metab","label":"Metabolic (Warburg)","x":38,"y":58},{"id":"age","label":"Ageing tissue and clonal fields","x":64,"y":58},{"id":"imm","label":"Immune surveillance and editing","x":90,"y":58},{"id":"toft","label":"Tissue organisation field theory","x":12,"y":82},{"id":"tme","label":"Microenvironment and inflammation","x":38,"y":82},{"id":"soil","label":"Seed and soil","x":64,"y":82},{"id":"bio","label":"Bioelectric (Levin)","x":12,"y":97},{"id":"mech","label":"Mechanical (stiffness, pressure)","x":38,"y":97}],"edges":[{"from":"smt","to":"driver","type":"activates"},{"from":"driver","to":"clonal","type":"activates"},{"from":"clonal","to":"hall","type":"activates"},{"from":"clonal","to":"imm","type":"activates"},{"from":"aneu","to":"smt","type":"inhibits"},{"from":"aneu","to":"clonal","type":"activates"},{"from":"epi","to":"smt","type":"inhibits"},{"from":"epi","to":"csc","type":"activates"},{"from":"csc","to":"hall","type":"activates"},{"from":"atav","to":"metab","type":"activates"},{"from":"metab","to":"smt","type":"inhibits"},{"from":"metab","to":"hall","type":"activates"},{"from":"age","to":"clonal","type":"activates"},{"from":"age","to":"tme","type":"activates"},{"from":"imm","to":"hall","type":"activates"},{"from":"toft","to":"smt","type":"inhibits"},{"from":"toft","to":"tme","type":"activates"},{"from":"tme","to":"soil","type":"activates"},{"from":"tme","to":"hall","type":"activates"},{"from":"tme","to":"imm","type":"activates"},{"from":"bio","to":"smt","type":"inhibits"},{"from":"bio","to":"toft","type":"activates"},{"from":"mech","to":"tme","type":"activates"},{"from":"mech","to":"toft","type":"activates"}],"interventions":["Somatic mutation theory and the driver model gave targeted therapy, genomic profiling, hereditary testing and synthetic lethality (imatinib, osimertinib, olaparib)","Clonal evolution gave combination therapy, residual-disease monitoring, rechallenge and adaptive dosing","Immune surveillance and immunoediting gave checkpoint inhibitors, CAR-T, TIL therapy and neoantigen vaccines","The microenvironment and inflammation view gave anti-angiogenics, aspirin chemoprevention and the vaccines against HPV and hepatitis B; seed and soil gave adjuvant bisphosphonates","The metabolic view gave FDG PET and IDH inhibitors; the epigenetic view gave azacitidine, decitabine, EZH2 and menin inhibitors and methylation-based tests","Cancer stem cells and plasticity gave the case for hitting the persister compartment (DLL3 engagers after neuroendocrine transformation); the mechanical view gave stromal decompression and vascular normalisation in trials; the bioelectric and atavistic views have not yet produced a therapy"]},{"id":"thyroid-cancer-signalling","kind":"pathway","name":"Thyroid cancer (KEGG map)","aka":["KEGG hsa05216","Thyroid cancer"],"tldr":"This KEGG map shows thyroid cancers driven by one relay, the MAPK pathway: RET or NTRK fusions and BRAF mutations in papillary tumours, RAS mutations or PAX8-PPARG fusion in follicular tumours, and TP53 loss marking anaplastic cancer. It matters because RET, NTRK and BRAF alterations each have their own drug, and MAPK blockade can restore iodine uptake.","summary":"Thyroid cancer is the most common endocrine malignancy and more than 95 percent of cases arise from follicular cells. KEGG map hsa05216 draws the spectrum from indolent well-differentiated papillary (PTC) and follicular (FTC) carcinoma to aggressive undifferentiated (anaplastic) carcinoma. Somatic RET and TRK (NTRK) rearrangements are found almost exclusively in PTC and appear early; BRAF V600E is the other dominant PTC driver, and all three feed RAS-RAF-MEK-ERK. FTC is marked by aneuploidy, RAS (KRAS, NRAS) mutations and PAX8-PPARG fusions, which alter PPAR-gamma-mediated transcription. Reduced E-cadherin releases beta-catenin signalling, and TP53 inactivation is crucial in dedifferentiation to anaplastic carcinoma.\n\nCabanillas, McFadden and Durante, The Lancet, 2016 (doi:10.1016/S0140-6736(16)30172-6) review the disease: mutually exclusive MAPK drivers (BRAF, RAS, RET/PTC, NTRK) define most differentiated thyroid cancers, MAPK activation lowers expression of the sodium-iodide symporter and so causes radioiodine resistance, and RET point mutations drive medullary thyroid cancer.\n\nWhat drugs do about it: radioactive iodine treats differentiated tumours that still take up iodine; the multikinase inhibitors lenvatinib and sorafenib (both listed by KEGG) treat radioiodine-refractory disease; selpercatinib and pralsetinib treat RET-altered tumours including medullary cancer, replacing vandetanib and cabozantinib for most patients; larotrectinib, entrectinib and repotrectinib treat NTRK fusion tumours; dabrafenib plus trametinib is approved for BRAF V600E anaplastic thyroid cancer; and MAPK inhibition (selumetinib, dabrafenib) can restore iodine uptake in refractory tumours, a strategy called redifferentiation.","asOf":"2026-09-10","links":[{"label":"KEGG map hsa05216","url":"https://www.kegg.jp/pathway/hsa05216"},{"label":"Review: Thyroid cancer (Lancet seminar)","url":"https://doi.org/10.1016/S0140-6736(16)30172-6"}],"tags":[],"related":["ras-mapk","rtk-activation"],"cancers":["thyroid"],"sections":[],"technologies":[],"targets":["ret","ntrk","kras","braf","mek","tp53"],"drugs":["selpercatinib","pralsetinib","vandetanib","cabozantinib","larotrectinib","entrectinib","repotrectinib","dabrafenib","trametinib","lenvatinib","sorafenib","selumetinib","radioactive-iodine"],"companies":[],"institutions":[],"pathways":["ras-mapk","rtk-activation","p53-cell-cycle","wnt","emt"],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-cabanillas-lancet"],"journals":[],"dependsOn":[],"notes":[],"analogy":"A thermostat where the heating relay (MAPK) is stuck on. Different faults jam it: a mis-wired fuse (RET or NTRK fusion), a broken switch (BRAF), or a stuck sensor (RAS). Because it is the same relay each time, an engineer can usually find the exact broken part and swap it, and fixing it even lets the room's iodine meter work again.","nodes":[{"id":"ret","label":"RET fusion / mutation","x":12,"y":6,"targetId":"ret"},{"id":"ntrk","label":"NTRK fusion","x":38,"y":6,"targetId":"ntrk"},{"id":"ras","label":"KRAS / NRAS (follicular)","x":62,"y":6,"targetId":"kras"},{"id":"pax8","label":"PAX8-PPARG fusion","x":88,"y":6},{"id":"braf","label":"BRAF V600E (papillary)","x":38,"y":30,"targetId":"braf"},{"id":"mek","label":"MEK / ERK","x":38,"y":52,"targetId":"mek"},{"id":"nis","label":"Sodium-iodide symporter (NIS)","x":10,"y":72},{"id":"cdh1","label":"E-cadherin / beta-catenin","x":70,"y":52},{"id":"p53","label":"TP53 (dedifferentiation)","x":88,"y":72,"targetId":"tp53"},{"id":"out","label":"Papillary / follicular to anaplastic","x":50,"y":94}],"edges":[{"from":"ret","to":"braf","type":"activates"},{"from":"ntrk","to":"braf","type":"activates"},{"from":"ras","to":"braf","type":"activates"},{"from":"braf","to":"mek","type":"activates"},{"from":"mek","to":"out","type":"activates"},{"from":"mek","to":"nis","type":"inhibits"},{"from":"pax8","to":"out","type":"activates"},{"from":"cdh1","to":"out","type":"activates"},{"from":"p53","to":"out","type":"inhibits"}],"interventions":["RET-altered tumours (including medullary cancer): selpercatinib, pralsetinib; older options vandetanib, cabozantinib","NTRK fusion tumours: larotrectinib, entrectinib, repotrectinib","BRAF V600E anaplastic thyroid cancer: dabrafenib plus trametinib","Radioiodine-refractory differentiated cancer: lenvatinib, sorafenib (listed by KEGG)","Redifferentiation: MAPK inhibition (selumetinib, dabrafenib) to restore iodine uptake before radioactive iodine","Differentiated cancer that still takes up iodine: radioactive iodine after surgery"]},{"id":"transcription-addiction","kind":"pathway","name":"Transcriptional machinery & addiction","aka":[],"tldr":"Cancer 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.","summary":"Lineage transcription factors (ER, AR, MITF, SOX2, ASCL1) and oncogenic fusions (EWSR1-FLI1, TMPRSS2-ERG, PAX3-FOXO1, NUT-BRD4) nucleate super-enhancers: clusters of enhancers densely loaded with Mediator, BRD4 (reading acetyl-lysine), p300/CBP, and cohesin loops to promoters. RNA Pol II is released from promoter-proximal pausing by P-TEFb (CDK9) and initiated by TFIIH (CDK7); CDK12/13 couple elongation to DNA-repair gene expression. MYC, itself super-enhancer-driven and amplified or ecDNA-borne, is an amplifier of all active genes. Because oncogene transcripts and proteins (MYC, MCL-1) are short-lived, transient inhibition of BET, CDK7 or CDK9 collapses them first ('transcriptional addiction'). Menin-KMT2A is a lineage-specific version (revumenib, ziftomenib in NPM1/KMT2A AML); CDK12 loss creates a tandem-duplicator phenotype in prostate cancer. Hormone receptors are the oldest transcription drugs; BET inhibitors were limited by thrombocytopenia, CDK9 inhibitors and BET/CBP degraders are in trials.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Super-enhancer","links":[{"label":"Bradner, Hnisz & Young, Transcriptional addiction in cancer (Cell 2017)","url":"https://doi.org/10.1016/j.cell.2016.12.013"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["ewing-sarcoma","aml","prostate","breast-hr-positive"],"sections":[],"technologies":["endocrine-therapy","androgen-deprivation","protac-degrader","epigenetic-drugs"],"targets":["androgen-receptor","estrogen-receptor","ewsr1-fli1","menin","kmt2a","npm1","bcl2","ezh2","irf4","ikzf1","ikzf3"],"drugs":["revumenib","ziftomenib","vepdegestrant","enzalutamide","elacestrant"],"companies":[],"institutions":[],"pathways":["myc","er-signaling","ar-signaling","epigenetic-reprogramming","swi-snf-chromatin","mrna-translation-eif4f"],"terms":["differentiation-syndrome","esr1-mutation","ar-v7"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bradner-cell"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"tf","label":"Lineage TFs (ER, AR, ASCL1)","x":12,"y":15,"targetId":"androgen-receptor"},{"id":"fus","label":"Fusion TFs (EWSR1-FLI1)","x":12,"y":48,"targetId":"ewsr1-fli1"},{"id":"se","label":"Super-enhancer","x":42,"y":30},{"id":"brd4","label":"BRD4, Mediator, p300","x":42,"y":60},{"id":"cdk7","label":"CDK7 (TFIIH) initiation","x":72,"y":15},{"id":"cdk9","label":"CDK9 (P-TEFb) elongation","x":72,"y":48},{"id":"pol","label":"RNA Pol II","x":72,"y":78},{"id":"myc","label":"MYC, MCL-1 (short-lived)","x":42,"y":90,"targetId":"bcl2"},{"id":"menin","label":"Menin-KMT2A (AML)","x":12,"y":82,"targetId":"menin"}],"edges":[{"from":"tf","to":"se","type":"activates"},{"from":"fus","to":"se","type":"activates"},{"from":"se","to":"brd4","type":"activates"},{"from":"brd4","to":"cdk7","type":"activates"},{"from":"brd4","to":"cdk9","type":"activates"},{"from":"cdk7","to":"pol","type":"activates"},{"from":"cdk9","to":"pol","type":"activates"},{"from":"pol","to":"myc","type":"activates"},{"from":"menin","to":"pol","type":"activates"}],"interventions":["Nuclear receptor drugs (endocrine therapy, ARPIs, SERDs, PROTAC vepdegestrant) are transcription drugs","Menin inhibitors revumenib and ziftomenib in KMT2A-rearranged and NPM1-mutant AML","BET inhibitors and BET/CBP degraders; CDK7 (samuraciclib) and CDK9 inhibitors in trials","Fusion-TF cancers (Ewing, NUT carcinoma) are the proving ground for transcriptional drugs"]},{"id":"tumor-dormancy","kind":"pathway","name":"Tumour dormancy","aka":[],"tldr":"Cancer 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.","summary":"Disseminated tumour cells enter quiescence (p38-high/ERK-low, NR2F1, DYRK1A), are held by niche signals (TGF-β2, BMP7, endothelial thrombospondin) and immune surveillance (NK cells, T cells), and evade cytotoxics by not dividing. Awakening triggers include inflammation (neutrophil extracellular traps), ageing, surgery-induced wound healing, and loss of immune control. Extended endocrine therapy in ER+ breast cancer is de facto dormancy maintenance. Strategies: keep cells asleep (NR2F1 agonists, 5-azacytidine + retinoic acid pilot in prostate), wake and kill (with cell-cycle-dependent drugs), or clear them immunologically. MRD assays make dormancy measurable for the first time.","asOf":"2026-09-08","links":[{"label":"Phan & Croucher, The dormant cancer cell life cycle (Nature Reviews Cancer 2020)","url":"https://doi.org/10.1038/s41568-020-0263-0"},{"label":"Albrengues et al., Neutrophil extracellular traps produced during inflammation awaken dormant cancer cells (Science 2018)","url":"https://doi.org/10.1126/science.aao4227"}],"tags":["mechanism"],"related":[],"cancers":["breast-hr-positive","prostate","melanoma"],"sections":[],"technologies":["mrd-testing","endocrine-therapy","cdk46-inhibitor"],"targets":[],"drugs":[],"companies":[],"institutions":["mount-sinai","cold-spring-harbor","mskcc"],"pathways":["metastatic-cascade","clonal-evolution","tgf-beta","er-signaling"],"terms":["late-recurrence","disseminated-tumor-cells","mrd"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-phan-nat-rev-cancer","paper-albrengues-science"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Aguirre-Ghiso (Einstein/Mount Sinai) on dormancy signalling and NR2F1; Egeblad (CSHL) on NETs awakening dormant cells; Massagué (MSK) on latency competent cells; Ghajar (Fred Hutch) on perivascular niches."],"analogy":"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).","nodes":[{"id":"dtc","label":"Disseminated tumour cell","x":15,"y":50},{"id":"quies","label":"Quiescence (p38↑, ERK↓, NR2F1)","x":42,"y":30},{"id":"niche","label":"Niche: TGF-β2, BMP7, endothelium","x":42,"y":70},{"id":"immune","label":"NK / T-cell surveillance","x":15,"y":15},{"id":"wake","label":"Awakening: NETs, inflammation, ageing","x":70,"y":50},{"id":"relapse","label":"Late relapse","x":92,"y":50}],"edges":[{"from":"dtc","to":"quies","type":"activates"},{"from":"niche","to":"quies","type":"activates"},{"from":"immune","to":"dtc","type":"inhibits"},{"from":"wake","to":"quies","type":"inhibits"},{"from":"wake","to":"relapse","type":"activates"},{"from":"quies","to":"relapse","type":"inhibits"}],"interventions":["Extended adjuvant endocrine therapy and CDK4/6 inhibitors (de facto dormancy maintenance)","5-azacytidine + all-trans retinoic acid to enforce dormancy (pilot, prostate)","MRD-guided intervention at molecular relapse","NET/inflammation inhibitors to prevent awakening (preclinical)"]},{"id":"tumor-microenvironment","kind":"pathway","name":"Tumour microenvironment (TME)","aka":[],"tldr":"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.","summary":"Cancer-associated fibroblasts (CAFs; myCAF/iCAF/apCAF subtypes) build desmoplastic stroma and secrete TGF-β, CXCL12, and IL-6; tumour-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) suppress T cells; regulatory T cells and exhausted CD8 T cells define immune phenotypes (inflamed, excluded, desert); abnormal vasculature creates hypoxia and blocks drug delivery; extracellular matrix stiffness signals through integrins and YAP; nerves and adipocytes add fuel and signals. Single-cell and spatial profiling has turned the TME from a concept into a map with druggable niches (FAP, CSF1R, CXCR4, TGF-β, adenosine).","asOf":"2026-09-08","wikipedia":"https://en.wikipedia.org/wiki/Tumor_microenvironment","links":[{"label":"de Visser & Joyce, The evolving tumor microenvironment (Cancer Cell 2023)","url":"https://doi.org/10.1016/j.ccell.2023.02.016"},{"label":"Human Tumor Atlas Network","url":"https://humantumoratlas.org"}],"tags":["mechanism"],"related":["theories-of-cancer","microenvironment-inflammation-theory","tissue-organisation-field-theory"],"cancers":[],"sections":[],"technologies":["single-cell-spatial","fapi-pet","checkpoint-inhibitor","antiangiogenic"],"targets":["fap","pd1","ctla4","vegf","cd47"],"drugs":[],"companies":[],"institutions":["mskcc","md-anderson","cold-spring-harbor","dana-farber","institut-curie"],"pathways":["pd1-checkpoint","vegf-angiogenesis","cgas-sting","emt","cancer-neuroscience","inflammation-nfkb"],"terms":["cold-vs-hot","tils","cancer-associated-fibroblasts","tumor-associated-macrophages","myeloid-derived-suppressor-cells","immune-exclusion"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-quail-joyce-microenvironment-metastasis-natmed-2013","paper-de-visser-cancer-cell"],"journals":[],"dependsOn":[],"notes":["Leading programmes: Tuveson (CSHL) on CAF subtypes; Sahai (Crick) and Mechta-Grigoriou (Institut Curie) on fibroblast heterogeneity; Joyce (Ludwig Lausanne) on TAMs; MD Anderson and MSK spatial atlases; Human Tumor Atlas Network (NCI)."],"analogy":"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.","nodes":[{"id":"tumour","label":"Tumour cells","x":50,"y":50},{"id":"caf","label":"CAFs (FAP+)","x":20,"y":30,"targetId":"fap"},{"id":"tam","label":"TAMs (CSF1R, CD47 axis)","x":80,"y":30,"targetId":"cd47"},{"id":"mdsc","label":"MDSCs","x":80,"y":70},{"id":"treg","label":"Tregs (CTLA-4)","x":62,"y":85,"targetId":"ctla4"},{"id":"cd8","label":"Exhausted CD8 T cells (PD-1)","x":38,"y":85,"targetId":"pd1"},{"id":"vessel","label":"Abnormal vessels (VEGF)","x":20,"y":70,"targetId":"vegf"},{"id":"ecm","label":"Stiff ECM / desmoplasia","x":8,"y":50},{"id":"nerve","label":"Nerves","x":92,"y":50}],"edges":[{"from":"tumour","to":"caf","type":"activates"},{"from":"caf","to":"ecm","type":"activates"},{"from":"ecm","to":"cd8","type":"inhibits"},{"from":"tumour","to":"vessel","type":"activates"},{"from":"vessel","to":"cd8","type":"inhibits"},{"from":"tumour","to":"tam","type":"activates"},{"from":"tam","to":"cd8","type":"inhibits"},{"from":"mdsc","to":"cd8","type":"inhibits"},{"from":"treg","to":"cd8","type":"inhibits"},{"from":"cd8","to":"tumour","type":"inhibits"},{"from":"nerve","to":"tumour","type":"activates"},{"from":"caf","to":"tumour","type":"activates"}],"interventions":["Checkpoint blockade (PD-1, CTLA-4, LAG-3) releases exhausted T cells","Anti-VEGF normalises vessels and improves infiltration","FAP-targeted imaging and radioligands attack CAFs","CSF1R, CD47/SIRPα, CXCR4, TGF-β, adenosine (A2A/CD73) agents target myeloid and stromal suppression; most have been modest so far","Radiation and oncolytic viruses convert excluded tumours"]},{"id":"ubiquitin-proteasome-system","kind":"pathway","name":"Ubiquitin-proteasome system & protein homeostasis","aka":[],"tldr":"Cells 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.","summary":"E1 activates ubiquitin, E2 carries it, and one of ~600 E3 ligases (CRL4-CRBN, VHL, MDM2, SCF-FBXW7, APC/C) attaches it to a substrate lysine; K48 chains send substrates to the 26S proteasome, whose β5 subunit is the target of bortezomib, carfilzomib and ixazomib. Deubiquitinases (USP7, USP14) reverse tagging. Plasma cells and myeloma depend on proteasome capacity to clear misfolded immunoglobulin; inhibition triggers the unfolded protein response (PERK, IRE1, ATF6) and death, and stabilises IκB to shut NF-κB. Cereblon modulators (thalidomide, lenalidomide, pomalidomide; CELMoDs iberdomide, mezigdomide, golcadomide) are molecular glues that redirect CRL4-CRBN to degrade IKZF1/3. PROTACs (vepdegestrant for ER, ARV-766 for AR, BGB-16673 for BTK) link a target ligand to an E3 ligand. Oncogenic lesions in the system: FBXW7 loss stabilises MYC, cyclin E and NOTCH; SPOP mutations in prostate; VHL loss stabilises HIF; MDM2 amplification degrades p53. HSP90 and chaperones buffer mutant kinases; HSP90 inhibitors mostly failed on toxicity.","asOf":"2026-09-09","wikipedia":"https://en.wikipedia.org/wiki/Proteasome","links":[{"label":"Békés, Langley & Crews, PROTAC targeted protein degraders: the past is prologue (Nat Rev Drug Discov 2022)","url":"https://doi.org/10.1038/s41573-021-00371-6"}],"tags":["mechanism","mechanics-atlas"],"related":[],"cancers":["multiple-myeloma","mantle-cell-lymphoma","prostate","breast-hr-positive"],"sections":[],"technologies":["protac-degrader","molecular-glue-platforms","degrader-antibody-conjugate"],"targets":["mdm2","hif2a","estrogen-receptor","btk","bcma","androgen-receptor"],"drugs":["bortezomib","carfilzomib","ixazomib","lenalidomide","pomalidomide","iberdomide","mezigdomide","golcadomide","vepdegestrant","bgb-16673"],"companies":[],"institutions":[],"pathways":["hif-vhl","p53-mdm2-axis","inflammation-nfkb","myc"],"terms":["r-iss","high-risk-myeloma"],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-bekes-nat-rev-drug-discov"],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"e1","label":"E1 → E2 ubiquitin","x":12,"y":15},{"id":"e3","label":"E3 ligase (CRBN, VHL, MDM2)","x":40,"y":15,"targetId":"mdm2"},{"id":"sub","label":"Substrate (IKZF1/3, p53, HIF)","x":70,"y":15,"targetId":"hif2a"},{"id":"chain","label":"K48 ubiquitin chain","x":70,"y":45},{"id":"dub","label":"DUBs (USP7)","x":92,"y":30},{"id":"prot","label":"26S proteasome (β5)","x":70,"y":75},{"id":"upr","label":"UPR, IκB → NF-κB","x":40,"y":75},{"id":"glue","label":"Glues / PROTACs hijack E3","x":40,"y":45,"targetId":"estrogen-receptor"},{"id":"deg","label":"Degradation","x":92,"y":90},{"id":"chap","label":"HSP90 chaperones","x":12,"y":75}],"edges":[{"from":"e1","to":"e3","type":"activates"},{"from":"e3","to":"sub","type":"activates"},{"from":"sub","to":"chain","type":"activates"},{"from":"dub","to":"chain","type":"inhibits"},{"from":"chain","to":"prot","type":"activates"},{"from":"prot","to":"deg","type":"activates"},{"from":"prot","to":"upr","type":"inhibits"},{"from":"glue","to":"e3","type":"activates"},{"from":"glue","to":"sub","type":"activates"},{"from":"chap","to":"sub","type":"inhibits"}],"interventions":["Proteasome inhibitors bortezomib, carfilzomib, ixazomib in multiple myeloma and mantle cell lymphoma","Cereblon glues: lenalidomide, pomalidomide; CELMoDs iberdomide, mezigdomide, golcadomide","PROTACs: vepdegestrant (ER), BGB-16673 (BTK), AR degraders; degrader-antibody conjugates deliver them by antibody","Reactivating degradation of oncoproteins (MDM2 inhibition for p53) and blocking DUBs are in trials"]},{"id":"vegf-angiogenesis","kind":"pathway","name":"VEGF angiogenesis","aka":[],"tldr":"How tumours grow their own blood supply. Low oxygen makes cells release VEGF, which tells blood-vessel cells to sprout toward the tumour.","summary":"Hypoxia stabilises HIF-1α/HIF-2α, which induce VEGF-A. VEGF-A binds VEGFR2 on endothelial cells → PLCγ/PKC/MAPK and PI3K/AKT → proliferation, migration, and permeability. Tumour vessels are chaotic and leaky, causing hypoxia and impaired drug and immune-cell delivery. Bevacizumab neutralises VEGF-A; VEGFR TKIs (axitinib, cabozantinib, lenvatinib) block the receptor; 'vascular normalisation' at moderate doses improves immune infiltration, the rationale for IO-VEGF combinations and PD-1×VEGF bispecifics. VEGF also suppresses dendritic cell maturation and expands Tregs.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Angiogenesis","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Angiogenesis"}],"tags":[],"related":[],"cancers":["rcc","hcc","colorectal","nsclc"],"sections":[],"technologies":["antiangiogenic"],"targets":["vegf","hif2a","pd1"],"drugs":["ivonescimab","belzutifan","atezolizumab"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"hyp","label":"Hypoxia","x":50,"y":5},{"id":"hif","label":"HIF-1α / HIF-2α","x":50,"y":22,"targetId":"hif2a"},{"id":"vegf","label":"VEGF-A","x":50,"y":40,"targetId":"vegf"},{"id":"vegfr","label":"VEGFR2 (endothelium)","x":50,"y":58},{"id":"sig","label":"PLCγ / MAPK / PI3K","x":50,"y":74},{"id":"ang","label":"Angiogenesis, permeability","x":25,"y":92},{"id":"imm","label":"Immune suppression (DC, Treg)","x":75,"y":92}],"edges":[{"from":"hyp","to":"hif","type":"activates"},{"from":"hif","to":"vegf","type":"activates"},{"from":"vegf","to":"vegfr","type":"activates"},{"from":"vegfr","to":"sig","type":"activates"},{"from":"sig","to":"ang","type":"activates"},{"from":"vegf","to":"imm","type":"activates"}],"interventions":["Bevacizumab, ramucirumab (antibodies)","VEGFR TKIs: axitinib, cabozantinib, lenvatinib, sunitinib","IO + VEGF combinations in RCC, HCC, endometrial cancer","PD-1×VEGF bispecifics: ivonescimab and successors","HIF-2α inhibitor belzutifan upstream in VHL-deficient RCC"]},{"id":"hif-vhl","kind":"pathway","name":"VHL / HIF oxygen sensing","aka":[],"tldr":"The 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.","summary":"In normoxia, prolyl hydroxylases (PHD) hydroxylate HIF-α, allowing the VHL E3 ligase to ubiquitinate it for proteasomal degradation. In hypoxia (or with VHL loss in ~90% of clear-cell RCC), HIF-α accumulates, dimerises with HIF-1β (ARNT), and transcribes VEGF, PDGF, GLUT1, CAIX, cyclin D1, and EPO. HIF-2α is the oncogenic paralogue in RCC; belzutifan blocks its dimerisation. CAIX is a PET and radioligand target (89Zr-girentuximab).","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Hypoxia-inducible_factor","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Hypoxia-inducible_factor"}],"tags":[],"related":[],"cancers":["rcc"],"sections":[],"technologies":[],"targets":["hif2a","vegf","hif1a"],"drugs":["belzutifan"],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"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.","nodes":[{"id":"o2","label":"Oxygen","x":20,"y":8},{"id":"phd","label":"PHD hydroxylases","x":20,"y":28},{"id":"vhl","label":"VHL E3 ligase","x":20,"y":48},{"id":"hif","label":"HIF-2α","x":55,"y":48,"targetId":"hif2a"},{"id":"arnt","label":"HIF-1β (ARNT)","x":85,"y":35},{"id":"genes","label":"VEGF, CAIX, GLUT1, cyclin D1","x":55,"y":72,"targetId":"vegf"},{"id":"out","label":"Angiogenesis, glycolysis, growth","x":55,"y":92}],"edges":[{"from":"o2","to":"phd","type":"activates"},{"from":"phd","to":"vhl","type":"activates"},{"from":"vhl","to":"hif","type":"inhibits"},{"from":"arnt","to":"hif","type":"activates"},{"from":"hif","to":"genes","type":"activates"},{"from":"genes","to":"out","type":"activates"}],"interventions":["Belzutifan (HIF-2α) in VHL disease and RCC, adjuvant with pembrolizumab (2026)","VEGF-directed therapy downstream","CAIX-targeted imaging (89Zr-girentuximab) and radioligands in development"]},{"id":"wnt","kind":"pathway","name":"Wnt / β-catenin","aka":[],"tldr":"Wnt/β-catenin is a developmental pathway hijacked by colorectal cancer. Normally a destruction complex keeps β-catenin low; losing APC lets it flood the nucleus and drive growth genes.","summary":"Without Wnt, β-catenin is phosphorylated by the destruction complex (APC, AXIN, GSK3β, CK1) and degraded. Wnt binding to Frizzled/LRP5/6 disables the complex; β-catenin accumulates, enters the nucleus, and with TCF/LEF drives MYC, cyclin D1, LGR5, AXIN2. APC loss initiates ~80% of colorectal cancers; CTNNB1 mutations occur in HCC, endometrial, and desmoid tumours; RNF43/RSPO alterations define a ligand-dependent subset. Wnt is also immunosuppressive (excludes dendritic cells). Drugs have been hard: porcupine inhibitors for RSPO/RNF43 tumours, tankyrase inhibitors, and nirogacestat (gamma-secretase, desmoid) are the closest.","asOf":"2026-09-04","wikipedia":"https://en.wikipedia.org/wiki/Wnt_signaling_pathway","links":[{"label":"Wikipedia","url":"https://en.wikipedia.org/wiki/Wnt_signaling_pathway"}],"tags":[],"related":[],"cancers":["colorectal","hcc","endometrial","gallbladder"],"sections":[],"technologies":["chemoprevention"],"targets":["gpc3"],"drugs":[],"companies":[],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"analogy":"β-catenin is a messenger constantly being shredded by a committee (APC and friends). A Wnt signal tells the committee to stand down. Colorectal cancer fires the committee (APC loss), so the messenger runs unchecked into the nucleus.","nodes":[{"id":"wnt","label":"Wnt ligand","x":20,"y":8},{"id":"fzd","label":"Frizzled / LRP5/6","x":20,"y":26},{"id":"rnf43","label":"RNF43 / RSPO","x":60,"y":12},{"id":"dc","label":"Destruction complex (APC, AXIN, GSK3β)","x":55,"y":45},{"id":"bcat","label":"β-catenin","x":55,"y":65},{"id":"tcf","label":"TCF/LEF","x":55,"y":82},{"id":"genes","label":"MYC, cyclin D1, LGR5","x":55,"y":97}],"edges":[{"from":"wnt","to":"fzd","type":"activates"},{"from":"rnf43","to":"fzd","type":"inhibits"},{"from":"fzd","to":"dc","type":"inhibits"},{"from":"dc","to":"bcat","type":"inhibits"},{"from":"bcat","to":"tcf","type":"activates"},{"from":"tcf","to":"genes","type":"activates"}],"interventions":["Porcupine inhibitors (RSPO-fusion / RNF43-mutant tumours, trials)","Gamma-secretase inhibitor nirogacestat in desmoid tumours (approved 2023)","Tankyrase inhibitors (preclinical/early)","Indirect: chemoprevention with aspirin/COX-2 in Lynch and FAP"]}]