Part IV · Molecular biology and therapeutic vulnerabilities · Chapter 13

Hallmarks and pathway logic in breast cancer

A framework that lists capabilities, applied to a disease whose question is which capability can be removed.

1 · Mapping the hallmarks onto breast cancer subtypes

The hallmark framework names the capabilities a tumour acquires on the way to becoming one¹. It does not rank those capabilities. Nor does it say which of them is load-bearing in any particular disease. Mapping the framework onto breast cancer is therefore an exercise in weighting rather than in ticking boxes.

Three weightings are worth stating plainly, because the clinical categories follow from them.

In hormone receptor-positive disease, sustained proliferative signalling is supplied by a physiological ligand acting through an unmutated receptor. The capability is present at diagnosis without any alteration that produces it, which is a situation the framework was not designed to describe. The oncogenic input is the host's own endocrine system.

In HER2-amplified disease the same capability is supplied by gene dosage rather than by sequence. The receptor is normal in sequence and abnormal in amount, a distinction that organises the whole of HER family signaling.

In triple-negative disease, genome instability dominates as the enabling characteristic, and proliferative signalling has no single named source. This is why the first two categories are named for a target while the third is named for what it lacks Triple-negative and other biologically diverse phenotypes.

The 2022 revision matters more here than in most tumour types. Phenotypic plasticity and disrupted differentiation were proposed as a discrete hallmark capability, with non-mutational epigenetic reprogramming as an enabling characteristic². Breast cancer's principal drug targets are lineage factors. A framework that treats differentiation state as an outcome rather than a variable describes this disease badly.

2 · Subtype-specific dependencies and why one framework does not fit all

A hallmark is a capability the tumour has. A dependency is something that can be removed to kill it. Those are different lists, which is why conflating them is the commonest way this framework is misused at the bedside.

Every breast cancer sustains proliferative signalling, though only some of them depend on a removable source of it. In hormone receptor-positive disease the dependency is not on oestrogen receptor protein as such. It is on the transcriptional programme that the receptor runs, which is why receptor degradation and ligand deprivation count as different strategies against the same pathway Endocrine therapy.

In HER2-amplified disease the dependency is on a dimer rather than on a single receptor. HER2 requires HER3 to drive proliferation in amplified cells, and the pair behaves as one oncogenic unit³. Removing HER2 kinase activity and removing HER2 dimerisation are therefore not interchangeable manoeuvres.

In triple-negative disease no comparable single dependency has been established, a gap that is visible in the shape of the treatment landscape. Where a dependency exists, therapy is defined by the target. Where it does not, therapy is defined by delivery and by DNA damage DNA repair, replication stress, and genomic instability Antibody-drug conjugates and targeted delivery.

The general point is that a capability shared across subtypes can have entirely different actionable content in each. Shared hallmarks predict shared descriptions. They do not predict shared treatment.

3 · Pathway redundancy, feedback, and why single-node inhibition fails

The pathways in question are not linear chains but feedback systems. Inhibiting one node relieves the negative feedback that the node was exerting. The drug's own mechanism therefore generates the escape route that ends its usefulness.

Three worked examples make the pattern concrete.

Inhibition of mTOR raises insulin receptor substrate 1 expression and abrogates feedback inhibition of the pathway, which activates Akt in cell lines and in treated patient tumours⁴.

Inhibition of PI3K increases oestrogen receptor transcriptional activity, ESR1 messenger RNA and receptor protein, and the effect was confirmed in tumours from patients receiving a PI3K alpha inhibitor⁵. The mechanism runs partly through the epigenetic regulator KMT2D⁶. Blocking the kinase makes the tumour more dependent on the receptor.

Inhibition of HER-family kinases is buffered by HER3. A compensatory shift in the HER3 phosphorylation and dephosphorylation equilibrium, driven by Akt-mediated negative feedback, preserves PI3K and Akt signalling despite effective inhibition of HER2⁷.

Each case predicts its own combination partner, which is the useful part of reasoning this way. The partner for an mTOR inhibitor is suggested by which receptor rebounds, not by screening. The clinical payoff of this reasoning is taken up in PI3K, AKT, mTOR, and related networks and Targeted therapy.

4 · Where the hallmark framework misleads

Four failures are specific to this disease rather than general complaints about the framework.

The first is that it is a list of capabilities rather than a list of rate-limiting steps. Every advanced breast cancer satisfies every hallmark, and satisfying them says nothing about which one can be removed. Removability is the only property of a tumour that changes a prescription.

The second is that the framework was built around acquired oncogenic alteration. In hormone receptor-positive disease the dominant proliferative input at diagnosis is a normal receptor responding to a normal ligand. Nothing about that is an acquired capability in the intended sense, and yet the most effective systemic treatment in oncology acts on exactly it.

The third is latency. The framework describes acquisition, while this disease's defining clinical problem is a recurrence appearing fifteen years after apparently curative treatment. A dormant cell has not been acquiring capabilities during that interval. It has been declining to exercise them, which is a different state governed by different biology Dormancy, residual disease, and late recurrence.

The fourth is that hallmarks are assigned to tumours and exercised by cells. A capability present in a minority compartment is scored as present for the whole tumour. No weight attaches to how large that compartment was. That averaging is precisely the error the heterogeneity argument in A framework for heterogeneity and HER2 heterogeneity exists to correct.

References

1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011 144:646-674. PMID 21376230
2. Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov 2022 12:31-46. PMID 35022204
3. Holbro T, Beerli RR, Maurer F, et al. The ErbB2/ErbB3 heterodimer functions as an oncogenic unit: ErbB2 requires ErbB3 to drive breast tumor cell proliferation. Proc Natl Acad Sci U S A 2003 100:8933-8938. PMID 12853564
4. O'Reilly KE, Rojo F, She QB, et al. mTOR inhibition induces upstream receptor tyrosine kinase signaling and activates Akt. Cancer Res 2006 66:1500-1508. PMID 16452206
5. Bosch A, Li Z, Bergamaschi A, et al. PI3K inhibition results in enhanced estrogen receptor function and dependence in hormone receptor-positive breast cancer. Sci Transl Med 2015 7:283ra51. PMID 25877889
6. Toska E, et al. PI3K pathway regulates ER-dependent transcription in breast cancer through the epigenetic regulator KMT2D. Science 2017. PMID 28336670
7. Sergina NV, Rausch M, Wang D, et al. Escape from HER-family tyrosine kinase inhibitor therapy by the kinase-inactive HER3. Nature 2007 445:437-441. PMID 17206155