HER2

A Bad Marker That Became a Good Target

When a protein appears in unusual quantities on the surface of cancer cells, it is usually bad news for the patient and of limited practical use for the doctor. HER2 is the exception. An amplified gene that reliably predicted poor outcomes spent roughly a decade as a prognostic marker before it became something more useful: a precise molecular target at which a specific treatment could be aimed. That transition — from information to intervention — is relatively rare in oncology, and the story of how it happened involves a sequence of laboratories, trials and regulatory decisions spread across the final quarter of the twentieth century.

HER2 (Human Epidermal growth factor Receptor 2) is a receptor tyrosine kinase encoded by the ERBB2 gene on chromosome 17. In normal tissue it sits in the cell membrane and participates in the signalling cascades that regulate growth and division. What makes it clinically significant in breast cancer is amplification: when the gene is present in many extra copies, the cell surface becomes crowded with receptor protein, and the growth signals those receptors transmit become, in effect, continuous. The cell behaves as though it is receiving an instruction it cannot switch off.

The discovery belongs primarily to the mid-1980s. Dennis Slamon's laboratory at UCLA, working with material assembled in collaboration with the National Cancer Institute, found in 1987 that HER2 gene amplification was present in approximately 25 to 30 percent of primary breast tumours and correlated strongly with shorter disease-free and overall survival. The finding was published in Science and almost immediately attracted the interest of researchers thinking about targeted treatment, because a receptor crowding the surface of a cancer cell in abnormal numbers is, structurally, an unusually accessible thing to attack. You do not need to penetrate the cell; you need something that binds to the outside.

The idea of using an antibody to block a receptor was not new, but doing it reliably in a living patient was technically very difficult. Axel Ullrich, who had helped characterise the receptor, and Slamon, together with colleagues at the biotechnology company Genentech, set about developing a monoclonal antibody that would bind to HER2's extracellular domain and interrupt its signalling. Early murine antibodies worked in cell lines and in mouse models. The challenge, standard in antibody therapeutics, was making something the human immune system would not simply reject — which led to the engineering of a humanised antibody in which only the small binding region came from mouse sequence and the rest was human. That molecule was trastuzumab.

From Amplification to Assay

The clinical programme required reliable testing. An antibody aimed at HER2 overexpression is only as useful as the method that identifies which tumours overexpress it, and in the early 1990s that method was still being standardised. Two technologies emerged in parallel. Immunohistochemistry — IHC — uses an antibody-based stain applied to a section of tumour tissue on a pathology slide, and scores the intensity and completeness of membrane staining on a scale from 0 to 3+. Fluorescence in situ hybridisation — FISH — goes to the gene itself, counting ERBB2 copies per cell under a fluorescence microscope to detect amplification directly. The two approaches have different sensitivities and occasional discordance, which created real diagnostic complexity that guidelines bodies have worked to resolve through successive rounds of scoring criteria.

FISH, in particular, established the biological logic clearly: it was gene amplification, not simply protein expression caused by other means, that was driving the worst HER2-positive behaviour. A tumour scoring 2+ on IHC — equivocal, not definitively high — is typically reflexed to FISH to determine whether amplification is present before treatment decisions are made. The scoring thresholds have been revised more than once, most significantly in joint guidance issued by the American Society of Clinical Oncology and the College of American Pathologists, because borderline cases have real consequences: a patient miscategorised as HER2-negative loses access to a class of treatment that would benefit her, and one miscategorised as HER2-positive receives treatment that carries toxicity without benefit.

The Trials and the Target

Trastuzumab entered pivotal clinical trials in the late 1990s. In metastatic disease, a randomised trial published in The New England Journal of Medicine in 2001 showed that adding trastuzumab to chemotherapy extended overall survival compared with chemotherapy alone in HER2-positive patients — a finding striking enough that the trial's control arm was closed early so that patients receiving chemotherapy alone could cross over. The magnitude of the effect, roughly five months' additional median survival in a metastatic setting, was clinically meaningful. The US Food and Drug Administration approved trastuzumab for metastatic HER2-positive breast cancer in 1998, before the survival data from the randomised trial were fully mature, on the basis of earlier phase-II response data.

The more consequential question was whether the same benefit existed in early-stage disease, where the goal was cure rather than extension. Four large randomised trials — NSABP B-31, NCCTG N9831, HERA and BCIRG 006 — reported in 2005 and 2006 with remarkable consistency. Adding trastuzumab to adjuvant chemotherapy in HER2-positive early breast cancer approximately halved the rate of disease recurrence and produced significant improvements in overall survival. The absolute magnitude varied by trial and follow-up duration, but the direction was unambiguous. The Early Breast Cancer Trialists' Collaborative Group subsequently incorporated these data into meta-analyses confirming the pattern across a broader evidence base.

The signal from HER2 biology did not stop at trastuzumab. Pertuzumab, which binds a different domain of the same receptor and therefore blocks a complementary signalling pathway, was shown to add further benefit when combined with trastuzumab in the neoadjuvant setting and in metastatic disease. Lapatinib, a small-molecule tyrosine kinase inhibitor that acts intracellularly on both HER2 and the related receptor EGFR, extended the treatment options for patients who had progressed on trastuzumab. Ado-trastuzumab emtansine — T-DM1 — linked the antibody to a cytotoxic payload, turning the receptor's own binding affinity into a delivery mechanism that concentrated chemotherapy inside HER2-positive cells. Each successive agent was a refinement on the same molecular recognition: find what the cancer cell overexpresses and use that against it.

HER2 amplification thus did something unusual. It redefined a poor-prognosis subgroup — patients who in the 1980s had measurably worse outcomes than their HER2-negative counterparts — into a subgroup for whom targeted therapy is now among the most effective in breast oncology. The molecular subtypes that expression profiling later delineated use HER2 status as one of their defining axes precisely because it predicts not just behaviour but response to a specific class of drug. That combination, biological plausibility linking gene amplification to receptor crowding to continuous growth signalling, plus a target sitting accessibly on the cell surface, plus a sequence of trials that confirmed the benefit, is the full architecture of what a therapeutic target actually requires.