Oncogenes, tumour suppressor genes and the two-hit hypothesis

At a glance

• Cancer comes from accumulated variants in two classes of growth-control gene.
• Proto-oncogenes promote growth. A gain-of-function change in one allele is enough (the accelerator jammed down).
• Tumour suppressor genes restrain growth or repair DNA. Both alleles must be lost (no brakes). This is Knudson's two-hit hypothesis.
• Most hereditary cancer syndromes come from a germline first hit in a tumour suppressor. The family shows autosomal dominant inheritance, even though the gene is recessive in the cell.
• Oncogenes are drug targets (kinase inhibitors, antibodies). Tumour suppressors are lost, so therapy exploits the loss instead (for example, PARP inhibitors in BRCA-deficient cancers).

Oncogenes

Mechanism	Example	Clinical link
Point mutation	KRAS G12C, BRAF V600E, EGFR exon 19 deletion or L858R, JAK2 V617F	Lung, colorectal, melanoma; myeloproliferative neoplasms
Amplification	ERBB2 (HER2) in breast and gastric cancer; MYCN in neuroblastoma	Trastuzumab targets HER2
Translocation forming a fusion	BCR::ABL1, t(9;22), in CML; EML4::ALK in lung adenocarcinoma	Imatinib; alectinib
Translocation to a strong promoter	MYC next to the IgH enhancer, t(8;14), in Burkitt lymphoma; BCL2, t(14;18), in follicular lymphoma	Venetoclax (a BCL2 inhibitor, used in CLL and AML)

• RAS is mutated in about 30% of human cancers. Activating RAS variants make it insensitive to GTPase-activating proteins, so GTP hydrolysis is greatly impaired and RAS stays mostly GTP-bound, signalling through RAF-MEK-ERK.
• Oncogene addiction: tumours driven by one oncogene often respond dramatically to blocking it, then relapse through secondary resistance mutations (for example EGFR T790M, ABL1 T315I).
• Germline oncogene syndromes are uncommon, because constitutive activation is often lethal in development. The main one is MEN2 (RET). See multiple endocrine neoplasia and chronic leukaemia.

Tumour suppressor genes

Gene	Normal role	Germline syndrome
RB1	Holds E2F, blocking G1 to S transition until phosphorylated by cyclin D-CDK4/6	Hereditary retinoblastoma; later osteosarcoma
TP53	"Guardian of the genome": arrests the cycle or triggers apoptosis after DNA damage	Li-Fraumeni (sarcoma, early breast cancer, brain tumours, adrenocortical carcinoma)
APC	Degrades β-catenin (Wnt pathway)	Familial adenomatous polyposis
NF1	GTPase-activating protein that switches RAS off	Neurofibromatosis type 1
VHL	Targets HIF for degradation	Von Hippel-Lindau (clear cell RCC, haemangioblastoma)
BRCA1, BRCA2	Homologous recombination repair	Hereditary breast and ovarian cancer
MLH1, MSH2, MSH6, PMS2	Mismatch repair	Lynch syndrome
PTEN	Antagonises PI3K-AKT	Cowden syndrome
CDKN2A (p16)	Inhibits CDK4/6	Familial melanoma and pancreatic cancer

How the second hit happens

The commonest second hit is loss of heterozygosity (LOH), the loss of the remaining normal allele. In tumour DNA, a marker heterozygous in the patient's blood shows only one allele. LOH mechanisms:

• deletion of the normal allele or of the whole chromosome region
• mitotic recombination, which copies the mutant allele onto the other homologue (copy-neutral LOH)
• chromosome loss, sometimes followed by duplication of the mutant chromosome

Second hits that are not LOH, because the locus stays heterozygous at marker level:

• a second, independent point mutation
• promoter hypermethylation, an epigenetic hit (MLH1 in sporadic MSI-high colorectal cancer, CDKN2A in many tumours)

Exceptions to strict two-hit behaviour: haploinsufficiency, where one copy isn't enough (some PTEN and NF1 effects), and dominant negative TP53 missense variants that disable the p53 tetramer.

Worked example: Knudson's retinoblastoma arithmetic

Knudson (1971) noticed that bilateral retinoblastoma presented earlier, and its age curve fitted one rate-limiting event. Unilateral cases fitted two.

He estimated that a child carrying a germline RB1 hit develops on average about 3 tumours, randomly distributed (Poisson). Then:

• probability of no tumour = e^−3 ≈ 0.05, so penetrance is about 95% (observed: around 90%)
• some carriers get no tumour, which is why an unaffected parent can transmit the variant (incomplete penetrance, not a skipped generation)
• with 3 tumours on average split between two eyes, most affected carriers have bilateral disease

Sporadic retinoblastoma needs two somatic hits in the same retinoblast before it differentiates. That is rare, so sporadic tumours are single and present later.

Pitfall: unilateral doesn't mean non-heritable. About 10 to 15% of children with unilateral, unifocal retinoblastoma carry a germline (or mosaic) RB1 variant, so germline testing is offered to every affected child.

The same logic predicts the clinical clues to a hereditary cancer syndrome: young onset, bilateral or multifocal tumours, multiple primaries and affected relatives.

Hallmarks of cancer

Hanahan and Weinberg's hallmarks (2000, updated 2011 and by Hanahan in 2022) group what these genes do: sustained proliferative signalling, evading growth suppressors, resisting cell death, replicative immortality (telomerase), angiogenesis, invasion and metastasis, reprogrammed metabolism, immune evasion, genome instability and tumour-promoting inflammation. The 2022 update adds phenotypic plasticity, non-mutational epigenetic reprogramming, polymorphic microbiomes and senescent cells. Each targeted drug class maps to a hallmark: VEGF inhibitors to angiogenesis, checkpoint inhibitors to immune evasion, CDK4/6 inhibitors to evading growth suppressors.

How it's tested

• Pedigree reasoning. Why an obligate carrier is unaffected, or why a syndrome caused by recessive cell-level loss is inherited dominantly.
• Matching mechanism to drug. HER2 amplification and trastuzumab; BCR::ABL1 and imatinib; CDK4/6 inhibitors need intact RB1 to work.
• Tumour testing. MSI-high colorectal cancer with MLH1 loss: MLH1 promoter methylation (epigenetic, sporadic) versus germline Lynch. See colorectal cancer.
• Recognising hereditary patterns. Bilateral breast cancer at 32 with a sarcoma in a sibling suggests Li-Fraumeni.

Pitfalls

• An inherited tumour suppressor variant doesn't mean every cell is abnormal in function. It means every cell is one hit away.
• RB1 loss makes a tumour resistant to CDK4/6 inhibitors, because there's no RB protein left for the drug to keep active.
• Promoter methylation silences a gene without any change in sequence, so sequencing alone misses it.
• Not every cancer gene fits neatly. TP53 mutants can be dominant negative or gain new oncogenic functions.
• KRAS status matters in colorectal cancer: activating KRAS or NRAS variants predict no benefit from anti-EGFR antibodies (cetuximab, panitumumab), because the pathway is switched on downstream of the receptor.