Description
From Beatson to Aromatase Inhibitors
The use of endocrine deprivation therapy to treat breast cancer was the first effective systemic therapy in breast cancer. Beginning with the work of Beatson in Glasgow in 1896 [1], oophorectomy became a management standard in premenopausal women, in part because of the important series of patients reported by Boyd in 1900 [2]. This approach had no efficacy in post-menopausal women, though at the time this was not fully understood. Further effective endocrine manipulations were achieved by hypophysectomy and adrenalectomy, which were active in post-menopausal women since they inhibited or removed (respectively) the source of the androgen precursors of oestrogens in post-menopausal women. This led to pharmacological approaches replacing surgery and in steps to today’s use of potent, well-tolerated third generation aromatase inhibitors (AIs).
Such surgical approaches, pioneered by Huggins and Bergenstal in 1951 [3], were of course not without their surgical and endocrine morbidity. Operative mortalities of the order of a few percent were accepted, though it was noted that the presence of pulmonary and/or pleural disease increased this risk. This is a stark comparison to the mortality risks of today’s endocrine therapies! Responses were seen in around 40 % of patients, though fewer in younger women, and of reason-able duration—one series reported that almost a third of women were still alive 5 years later [4]. These data are not substantially different from today’s endocrine therapies, and in some ways the major developments since, such as the pharmacological approaches which began in the late 1960s, were more about tolerability and increasing the range of therapeutic options than about major improvements in efficacy.
Aminoglutethimide was developed as an inhibitor of the final enzymatic step in the synthesis of oestrogens from androgens, but it also inhibits the earlier conversion of pregnenolone from cholesterol, thus inhibiting the synthesis of all hormonally active steroids, so that when used in the full effective dose of 1000 mg daily, even with the required supplementary hydrocortisone, there were endocrine tox-icities such as hypothyroidism and ACTH stimulation. Importantly, in a 96-patient randomised trial, it was demonstrated to be at least as effective as surgical adrenolectomy [5]. Subsequently it was also found that the lower dose of 250 mg daily was essentially both biochemically and clinically as effective as the standard 1000 mg daily [6, 7] but with less toxicity.
Tamoxifen had of course been developed somewhat in parallel, initially as a contraceptive, and had evidence of its efficacy established back in the 1960s [8], so that once medical approaches had supplanted surgical adrenalectomy, the question arose as to which of the drugs, aminoglutethimide or tamoxifen was the superior therapy for women with advanced breast cancer. One very informative study comparing these drugs was a phase II randomised cross-over trial conducted in the UK [9]. No significant differences in efficacy were found, though what was informative was that there was a differential cross-resistance. More patients responded to aminoglutethimide after tamoxifen than with the opposite sequence of drugs, and unsurprisingly, tamoxifen was generally better tolerated. Thus tamoxifen remained first-line therapy: though the study was too small to exclude small differences in efficacy. This, and other data, established that when more potent aromatase inhibitors were developed, they started in the later lines of therapy and only once on the market were compared directly with tamoxifen. This phase II study is also noteworthy for the observation, confirmed with a later follow-up [10], that there was no benefit in combining tamoxifen and aminoglu-tethimide: a lesson perhaps forgotten in the design of the ATAC trial which again confirmed the lack of any benefit in the combination of tamoxifen and a later generation non-steroidal aromatase inhibitor [11].
One of the major developments in the endocrine therapy of women with advanced breast cancer was the ability to identify those tumours very unlikely to respond, namely those without nuclear expression of the oestrogen recep-tor (ER) and/or progesterone (PgR) receptor. Older studies did not select for ER and/or PgR positive breast cancers, but gradually it has become the norm to test for expression in the histopathological diagnosis of breast cancer, preferably in laboratories that have some form of certification (e.g. CLIA in North America, NEQAS in the UK) that their assays are robust and reproducible. There remains debate as to just how frequently woman have ER-negative PgR-positive
(ER-ve/PgR+ve) breast cancer, but that notwithstanding, it is fair to say that the only indication for the use of an endocrine therapy is in patients with breast cancers that are ER and/or PgR positive. Those cancers that express both receptors are on the whole more likely to be endocrine sensitive, and have a lower rate of proliferation and as such are mostly Luminal A; those which only express the ER tend to have higher rates of proliferation, earlier relapses and fit into the category of Luminal B. The importance of proliferation in breast cancer biology led to extensive stud-ies of Ki67, which is now reaching the status of an ancillary biomarker to separate Luminal A and B subtypes within ER+ve tumours [12].
Further research has been conducted to explore the drivers of endocrine resist-ance, both as a class of agents and specifically to aromatase inhibitors. There is a very extensive literature on this, with many markers all having some evidence that they are relevant, but none yet has been able to enter clinical practice as hav-ing sufficient analytical validity and strong enough positive and negative predictive values to allow identification of ER+ve patients whose cancer will not respond to endocrine therapy, over and above the presence of the oestrogen and/or progester one receptor. One reason why this extensive research has yet to identify a better marker is perhaps illustrated in a study conducted using mRNA from ER+breast cancers treated in the neo-adjuvant setting [13] in which a heat map of similarity of gene expression shows that responders are more similar than non-responders, such that accurate prediction of non-response before therapy is challenging. More recent data suggest that the use of the primary tumour to predict sensitivity of met-astatic disease may not be ideal: many years of work looked for mutations in the ESR1 gene (which codes for the ERα protein) in the historic primary tumour as a cause of resistance and failed to confirm its relevance; more recent studies have looked at biopsies taken from the metastases and reported that up to 20 % of cases have mutations in the ESR gene [14] and these are associated with resistance to endocrine therapy including aromatase inhibitors [15].
Thus whilst there remain controversies over the exact optimal selection tools for the endocrine sensitivity, the absence of nuclear steroid receptor expression means that endocrine therapy is not appropriate, and older studies were often una-ble to exclude such patients and so often reported slightly lower overall beneficial effects. The study by Murray and Pitt of 53 women treated with aminoglutethimide is important as it laid down three seminal observations which have not really been disproved in the subsequent 30+ years [16]. Firstly, the median duration of benefit in advanced breast cancer was about 1 year but with some patients having much longer disease control; and secondly, that sensitivity to prior endocrine therapy increased the chance of response to an AI with approximately twice as many women who had responded to prior tamoxifen responded to the AI (69 % vs. 35 %). It also confirmed that once a breast cancer has developed overt clini-cal metastases, resistance to sequential lines of therapy almost inevitably develops, though this can happen after months or years of any particular therapeutic approach to oestrogen deprivation.
Current Situation
Today there is widespread use of the three clinically available aromatase inhibi-tors. Letrozole and anastrazole, both non-steroidal compounds are the two more dominant in the market, and the third agent, exemestane, is often referred to as an aromatase inactivator as being a steroidal compound it irreversibly inhibits the enzyme. A series of trials have confirmed modest superiority of AIs over the previous standards of care, whether megestrol acetate in the second line ther-apy of advanced disease after tamoxifen, or against tamoxifen itself, both in the adjuvant setting and also in advanced breast cancer. In the adjuvant setting, the meta- analysis of 18,871 women randomised across several large studies confirms an overall small, but significant survival advantage for the sequence of an AI after a few years’ tamoxifen [17]. Such confirmed survival advantages have not been robustly seen in advanced breast cancer, but either non-inferiority and/or some clinical superiority have been shown in large phase III trials, and on the basis of those data and their good tolerability, they have become the standard first line
endocrine therapy in ER+ breast cancer for post-menopausal women.