Journalist(s):

Janet Fricker

Introduction

A review entitled Past, present and future challenges in breast cancer treatment, published in 2014 to mark the 50th anniversary of the American Society of Clinical Oncology (ASCO), reported that anti-oestrogen treatments have had a “greater global impact than any other treatment intervention in cancer medicine”. The flagship in hormonal, or endocrine, therapy has been tamoxifen, one of the World Health Organisation’s essential drugs.

This was echoed by Harold Burstein, a medical oncologist at the Dana-Farber Cancer Institute, who said in 2004: “Tamoxifen is the most important drug in the history of medical oncology.”

Tamoxifen came of age in 1992 with the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) meta-analysis of data from adjuvant tamoxifen trials. The overview data, considered ‘hypothesis supporting’ as opposed to ‘hypothesis generating’ (as the individual trials drew the same conclusion), found that four and six additional women per 100 will be alive at 5 and 10 years respectively after taking tamoxifen as an adjuvant treatment in early breast cancer than women not taking the drug.

Over the years, tamoxifen and subsequent endocrine treatments have saved and extended the lives of hundreds of thousands of women with early breast cancer and tamoxifen has had an impact on families all over the world. “Mothers see their children grow up, children experience the affection of a grandmother,” wrote Craig Jordan, a British–American breast cancer pharmacologist and one of the leading figures in hormonal therapy, the cancer world and indeed in other spheres – he served in the Special Air Service (SAS), an elite arm of the British army.

The story of one of the most important advances in cancer prevention and treatment has extensive European input and collaboration with researchers in the US, and also shows differences in approaches across the Atlantic.

This story is described as follows:

• A briefing on what comprises endocrine therapy
• The impact of tamoxifen
• The history of approaches to hormone therapy before tamoxifen
• How tamoxifen was developed as a cancer treatment and prevention therapy (and how it nearly did not make it)
• The launch of a different class of hormonal agent, aromatase inhibitors.

Endocrine therapy for breast cancer - types and mechanism

Endocrine therapy is used to treat women in all breast cancer stages and for prevention. It is given to:

• Prevent the development of breast cancer in high-risk women
• Shrink tumours prior to surgery in women with inoperable or large operable tumours (neoadjuvant treatment)
• Reduce the risk of breast cancer recurrence following surgery (adjuvant therapy)
• Relieve symptoms and delay progression in women with metastatic or advanced breast cancer.

The hormone oestrogen is the major stimulus that drives growth of hormone-dependent breast cancers, and hormone therapy is designed to obstruct this process. It has effect in about 60% to 70% of breast cancers that have oestrogen receptors and require oestrogen to grow. Note that oestrogen receptor is abbreviated to ER after the American spelling (estrogen) and that HR (hormone receptor) is also an abbreviated term although HR can be oestrogen or progesterone, or both.

Hormone therapy in breast cancer has progressed from the irreversible destruction of endocrine glands, achieved through surgery or radiation (both with high comorbidity), to the development of targeted drugs.

Drugs work by interfering with the supply of oestrogen to the breast tumour, which can be achieved by either blocking the action of the hormone on the oestrogen receptor or reducing production of the hormone.

Tamoxifen, originally developed as an antifertility agent and later discovered to have beneficial effects in breast cancer, was the first drug used. It was found to bind directly to the oestrogen receptor, a protein that was discovered in 1958, whose presence in breast tissue predicts the likely benefit of hormone therapy. It is the first of a class of agents known as selective (o)estrogen receptor modulators (SERMs).

Despite tamoxifen having anti-oestrogen effects in breast tumours, it was later found to have oestrogen-like actions in the uterus and other tissues that had unfavourable side-effects (e.g. risk of endometrial cancer and thromboembolic events). There was a need to find alternative agents that avoided this undesirable ‘agonist’ activity, and also because people developed resistance to tamoxifen.

Among other agents are:

• Aromatase inhibitors (AIs), which have a different approach by blocking synthesis of oestrogen. Oestrogen represents the end-product of a sequence of steroid transformation with more specific suppression resulting from the inhibition of the final step, which involves a single enzyme – aromatase.
• Other SERMs designed to be oestrogen receptor antagonists in the breast and uterus, and ER agonists in bone (thereby having the additional benefit of preventing bone loss).
• Selective oestrogen receptor downregulators/degraders (SERDs) such as fulvestrant, which work instead by binding to the receptor and stopping the action of oestrogen.

An important concept guiding the type of endocrine therapy for women with breast cancer is that primary oestrogen production sites differ between premenopausal and postmenopausal women. In premenopausal women, the ovaries are the major source of oestrogen; in postmenopausal women, peripheral tissues (fat, muscle and the tumour itself) produce oestrogen.

Consequently, AIs have come to play a greater role in the treatment of postmenopausal women where they can completely block oestrogen production, and tamoxifen in premenopausal women.

AIs can be used in premenopausal women in combination with ovarian suppression, most often with luteinising hormone-releasing hormone (LHRH) agonists, which produce temporary suppression of ovarian oestrogen as opposed to the permanent interruption that occurs with oophorectomy (surgical ovary removal) or radiotherapy.

Impact of tamoxifen - and Europe - US differences

Tamoxifen introduced into oncology the concept of personalising treatment according to biological markers, in this case expression of oestrogen receptors on breast tissue. That the presence of the oestrogen receptor was necessary for response to endocrine therapy provided a predictive factor, allowing substantial numbers of women who were hormone-receptor negative to avoid ineffective therapy. (Male breast cancers are almost all hormone-receptor positive, so the much smaller number of men are also part of the story.)

The clinical success of tamoxifen encouraged evaluation of more agents, especially aromatase inhibitors and further selective oestrogen receptor modulators.

Identifying the oestrogen receptor, and the finding that tamoxifen attached to it, opened the way for a new field of nuclear receptors and their intracellular signalling pathways. The researcher who discovered the oestrogen receptor (making a first presentation in 1958) and set in train this field was Elwood Jensen at the University of Chicago, joined by others such as Jack Gorski in the many studies of this complex field that followed.

Identifying the key role of the enzyme aromatase also provided one of the first molecular targets for rational drug development – knowledge of the structure, function and mechanism of the target – thereby avoiding testing of thousands of molecules. The result was the development of aromatase inhibitors that have exceptional potency and specificity.

Tamoxifen’s potential to prevent contralateral breast cancer in women at high risk (who already had cancer in one breast) made tamoxifen the first preventive treatment for any cancer, and helped to establish the broader principle of chemoprevention, although drugs such as aspirin and metformin are still confined to research.

Recognition of the need for large numbers of patients to detect moderate treatment differences led to the development of the EBCTCG Oxford Overview process, where all trials addressing the same questions are combined to provide clear-cut answers.

Tamoxifen and other endocrine treatments have influenced patient empowerment. The widespread publicity surrounding the endocrine trials led to the public becoming more familiar with breast cancer treatments, and ultimately encouraged more patients to become involved in the design of trials. Tamoxifen trials resulted in the development of the first instrument for measuring quality of life in patients undergoing treatment for cancer, the linear analogue self-assessment (LASA) tool.

From the outset, there was a cultural divide between Europe and the US, with Europe having a stronger tradition for hormone therapy, and cytotoxic agents in the US.

“The discipline of medical oncology took off much earlier in the US than Europe. As a result, the US had medical oncologists treating breast cancer who were more gung ho in using aggressive chemotherapy. In Europe, treatment of breast cancer was much more in the hands of surgeons who felt more comfortable giving less toxic endocrine therapy,” said Jack Cuzick, an epidemiologist at Queen Mary University of London who led chemoprevention trials and was a statistician on other breast cancer studies.

Another factor making chemotherapy attractive to American health providers, added Cuzick, was that it requires clinicians to have more contact with patients, generating larger fees for service. In the case of tamoxifen and aromatase inhibitors, clinicians write prescriptions and review the patient every few months, leading to smaller workloads.

Endocrine therapies: the early years

Beatson: the origin of hormone therapy in breast cancer

Scottish surgeon George Thomas Beatson (after whom the Beatson West of Scotland Cancer Centre is named) provided the foundations for hormone therapy in breast cancer. In 1874, soon after graduating from the University of Edinburgh, Beatson worked on an estate in western Scotland caring for a man whose ‘mind was affected’. The job afforded plenty of leisure time, which Beatson spent observing lactation practices in cows and sheep on a nearby farm. It was well known among farmers, he discovered, that cows could give milk indefinitely if their ovaries were removed after calving.

George Beatson also served as a volunteer in medical corps and received many honours. Source: Wikipedia

These observations inspired Beatson’s thesis, Functions of the cortex cerebri, with investigations, in which he confirmed in rabbits that cell proliferation seen in the gestational breast proceeds to fatty degeneration if the ovaries are removed. He compared the histologic changes in the lactating breasts after pregnancy with those seen in cancer.

From these studies, Beatson concluded that the breasts were controlled by the ovaries (rather than nerves as previously thought) and that cancer results from failure of the natural involution (shrinking) of the lining of breast ducts. Over the next 20 years, he experimented by removing ovaries from rabbits, sheep and cows and studying resulting tissue changes. It was this animal work that gave him the idea that removing the ovaries might lead to regression of advanced breast cancer.

“That we must look in the female to the ovaries as the seat of the exciting cause of carcinoma, certainly of the mamma, in all probability of the female generative organs generally, and possibly of the rest of the body,” Beatson wrote.

In 1896 Beatson, then at the Glasgow Cancer Hospital, published a seminal paper detailing treatment of three women with advanced breast cancer with bilateral oophorectomy (surgical removal of the ovaries).

Beatson’s first case was a 33 year old woman who had noticed a small lump on her left breast when feeding her first child, which got bigger after the birth of her second baby. The cancer was already advanced when a 12 cm tumour was removed, leading to her being referred to Beatson who removed both her ovaries. The woman survived for nearly 4 years before dying of recurrent disease.

Encouraged by these results, Beatson went on to perform bilateral oophorectomy on two other women, who also showed marked postoperative improvements. (It appears Beatson would not have been the first if German surgeon Alfred Schinzinger had followed up his proposal in 1889 for oophorectomy as a treatment for breast cancer, but it remained only an idea. A British surgeon, Thomas William Nunn, who wrote a book on breast cancer, is also said to be the first to see a relationship between ovarian function and cancer in a patient case in 1882. See this history of oophorectomy for breast cancer.)

Gaining access to Beatson’s operative records, David Smith, a Glasgow breast surgeon, found Beatson had treated nine breast cancer patients with oophorectomy, but only reported on his success with three.

“That is not surprising as the 30% response rate to endocrine therapy has been an obstinate fixed number until the discovery of the oestrogen and progesterone receptors in breast cancer cells,” wrote British breast surgeon, Michael Baum, in his book The History and Mystery of Breast Cancer. By selecting what he was going to publish and leaving out the denominator, added Baum, Beatson missed something of critical importance that took many years for scientists to recognise. “It demonstrates how negative results are as important as positive results,” he said.

Reading Beatson’s paper, Baum commented on how impressed he was by the description of informed consent. “It was very moving and incidentally demonstrated a degree of compassion and ethical probity which would be in keeping with modern attitudes,” he said.

Commenting on the contribution of Beatson, Charles MacMahon and John Cahill wrote: “He intuitively developed a thesis which he then proceeded to explore in the laboratory. Guided by his clinical observations and knowledge of the disease process, he proceeded to apply a mode of treatment which has become a standard therapy in advanced breast cancer.”

Focus on the ovaries

English surgeon Stanley Boyd, at Charing Cross Hospital, London, followed Beatson’s lead and championed oophorectomy as an effective (although not curative) treatment for his breast cancer patients. In 1900, Boyd published a summary of his cases in which he reported the significant finding that one third of breast cancer patients benefited from this approach and that the response only lasted 1 or 2 years.

Another early proponent was Hugh Lett, a surgeon at the London Hospital, who performed oophorectomy on a series of 99 women with inoperable breast cancer. Lett observed a marked improvement in about 23% of patients, with the best results obtained in women aged between 45 and 50 years. An observation of little benefit in women over 50 provided an early signal that hormonal therapy (targeting the ovaries) delivered the greatest effects in premenopausal women.

Irradiation of the ovaries, as an alternative means of stopping ovarian function, was first attempted in 1904 by French radiologist Francois Victor Foveau de Courmelles. The arrival of radiotherapy brought surgical oophorectomy into disfavour.

“Surgery involved anaesthesia and risk of infections, making non-invasive irradiation castration oophorectomy a big advance,” said Monica Castiglione, emeritus professor at the University of Geneva and an oncologist member of international breast cancer trial groups. (Castration, a term most often used in suppressing/removing testosterone in prostate cancer, also applies to ovaries in women.) “However, radiation therapy wasn’t without issues including damaging organs located near the ovaries, like the bladder and the intestines.” Another drawback, she added, was that surgery takes immediate effect, whereas radiation takes 3 to 6 months for hormonal levels to fall.

The first randomised trial of ovarian ablation in breast cancer was one of two trials initiated in 1948 by Ralston Paterson, a radiologist and director of radiotherapy at the Christie Hospital in Manchester. In the ovarian study, 598 pre- and postmenopausal women were randomised to ovarian irradiation or a control procedure. Results at 10 and 15 years showed the survival for irradiated patients was superior to the control but did not reach statistical significance. (The other trial concerned radiation of the breast after mastectomy.)

But the randomisation protocol, which used the women’s months of birth to allocate treatment, was a major issue as it gave investigators prior knowledge of treatment, creating potential for bias.

Nevertheless, said Michael Baum, the achievements of these early trials should be more widely acknowledged. “It’s remarkable that these very large studies at the Christie Hospital were begun at almost the same time as a much smaller study of tuberculosis treatment that’s often cited as the first randomised trial performed in the clinic.” Furthermore, the ovarian study demonstrated for the first time that survival in premenopausal women could be prolonged by adjuvant ablation.

See Paterson’s Breast cancer. A report of two clinical trials and also the summary of the paper in the book, Classic Papers in Breast Disease, by Michael Baum and Craig Henderson.

Roar Nissen-Meyer, a radiation oncologist at Norwegian Radium Hospital, Oslo, undertook a large trial investigating prophylactic (now known as adjuvant treatment) radiation castration versus therapeutic castration.

From 1957 to 1963, 932 patients were admitted for treatment of operable breast cancer (stage I and II), with oophorectomy performed as part of primary treatment in 74 women (classified as adjuvant treatment), ovarian irradiation as part of the primary treatment in 590 women (classified as adjuvant treatment), while the remaining 268 only had castration performed after cancer recurrence (classified as therapeutic castration).

When he analysed the survival curves, Nissen-Meyer found that the first 4 years showed a divergence favouring the castrated arm, but after 5 years the curves converged. From this, he concluded that there were no grounds for expecting that castration could prevent recurrences, but that it could delay them and possibly increase survival of patients who could be cured by primary surgery.

An advance in trial design was that Nissen-Meyer minimised the possibility of foreknowledge of treatment by ensuring allocation was performed by a telephone call to an independent person.

“Roar was a remarkable one-man band who used a punch card system to keep track of the study,” said Baum. “He was also a decorated war hero. Working as a young doctor in northern Norway he sent messages to the British Army about the German occupation. When the Nazis discovered that he was working as an agent he escaped by sledge skiing over the mountains with his wife and baby. In Britain he retrained as an agent and was parachuted back into Norway to spy on the Germans.”

In the early 1960s, five similar ovarian ablation trials were started in the US and Canada, including the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-03 trial, undertaken in the US by Bernard Fisher and colleagues. This trial, which compared ovarian ablation with control procedures, was of particular significance as it represented the first NSABP formal collaborative effort to use randomised trials to assess breast cancer treatment.

Finally, in 1996 the benefits of adjuvant oophorectomy were conclusively established by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), led by epidemiologist Richard Peto at the University of Oxford (see more later), who conducted a meta-analysis of data from clinical trials of adjuvant oophorectomy (both radiation and surgery) undertaken in the 1960s and 1970s.

The analysis, including 12 randomised trials involving 2102 women aged under 50, found survival was significantly improved with 5 year survival of 52.4% for those receiving ovarian ablation vs 46.1% for those receiving control treatment. Overall, results showed there were 6.3 fewer deaths per 100 women among those receiving oophorectomy.

Other surgical approaches used in breast cancer to influence hormones included bilateral adrenalectomy (surgical removal of the adrenal glands) and <a name="_hlk82189513"></a>hypophysectomy (surgical removal of the pituitary gland). Adrenalectomy as a treatment for breast cancer was first described in 1952 by Charles Huggins (who won a Nobel prize for work on hormones in prostate cancer) and Delbert Bergenstal at the University of Chicago. They were only able to perform the operation following the introduction of cortisone acetate as substitution therapy for adrenal insufficiency.

At the time, the rationale for the treatment was obscure and patients had to be maintained on cortisone for life. “We now understand why the procedure worked. Androgens synthesised in the adrenal glands are metabolised in fatty tissue to produce oestrogens by the enzyme aromatase,” explained Baum.

A review of 455 adrenalectomised patients with breast cancer (many of whom had been operated on by Huggins) found 5 year survival was 12.5% in the surgical group versus 3.5% in an historical control group treated prior to 1951.

Hypophysectomy was first described in humans by the endocrinologist Rolf Luft and neurosurgeon Herbert Olivecrona, both leading figures in Swedish medicine, in the early 1950s.

The rationale for the procedure was based on animal work by Remmert Korteweg and Frédéric Thomas at the Netherlands Cancer Institute, who in 1939 showed breast cancer in mice grew more slowly after hypophysectomy. “The mortality and morbidity of this procedure was awful, with diabetes insipidus being added to adrenocortical insufficiency and occasionally blindness from damage to the optic chiasma,” said Baum.

For a while adrenalectomy and hypophysectomy were widely used, but thankfully became obsolete after aromatase inhibitors (AIs) were introduced to chemically block the aromatase enzyme (see later).

Endocrine ablation, surgical or radiation induced oophorectomy, adrenalectomy or hypophysectomy remained palliative options for breast cancer patients for the first 60 years of the 20th century. In all cases, a uniform result was obtained where only about one in three patients derived any benefit and the duration of response was about a year.

The rationale for the persistent efficacy result of one in three patients only became evident with the discovery of the oestrogen receptor. But one of the long-term consequences of endocrine surgery was that the field of hormone therapy in breast cancer was dominated by surgeons, who continued to coordinate the treatment of patients even after the introduction of oral tamoxifen.

Hormone therapy and Tamoxifen – but a battle for its recognition

Hormone therapy in the form of tamoxifen came to be at about the same time as the introduction of chemotherapy for breast cancer (chemotherapy was pioneered by Gianni Bonadonna in Italy and Bernard Fisher in the US). But a paradoxical hormonal therapy had already been tried in breast cancer, namely oestrogen itself in high-dose form, and also high-dose androgen.

The paradox is that it became known that oestrogen can act as both a promoter and inhibitor of breast cancer. The finding that synthetic oestrogens could produce a 30% response rate in advanced breast cancer was reported in 1944 by Alexander Haddow and colleagues at the Royal Cancer Hospital, London and the Christie in Manchester, but it was not known then what mechanisms were in play. This study involved other cancer types and prostate also responded.

Craig Jordan has said this was the first chemical therapy to treat any cancer successfully. Synthetic oestrogens were discovered and produced in the 1930s and high-dose oestrogen therapy remained a standard of care until the introduction of tamoxifen, he wrote. See also this article on the history of oestrogen.

It was hospital records of oestrogen and also androgen treatments that were compared with tamoxifen in its first trial. (Oestrogen and androgen have also undergone more recent and continuing investigation in cancer.)

Tamoxifen is an anti-oestrogen and its discovery and development in the UK result in part from research in the US that met a dead end. The first nonsteroidal anti-oestrogen, MER-25 (ethamoxytriphetol), was discovered by Leonard Lerner, an endocrinologist at Merrell, a US pharmaceutical firm, in the 1950s. Another anti-oestrogen, clomiphene, was also produced at Merrell.

There was a long list of possible applications, with particular interest in reproductive indications for both fertility and anti-fertility (the ‘morning after’ pill was mooted as they were found to be antifertility agents in rats); breast cancer was also one of the possibilities. However, while nonsteroidal anti-oestrogens were found to be good contraceptives in rats they induced ovulation in women with poor fertility. MER-25 though had side-effects that curtailed its cancer investigation, and attention switched to the ovulation-inducing properties of clomiphene, for which it was approved in 1967 in the US.

The story switches across the Atlantic to the UK for the breakthrough in cancer. Tamoxifen was also put forward as a potential ‘morning after pill’ but was a failure – and its eventual status as one of the most important cancer drugs can largely be attributed to the insight and tenacity of British reproductive endocrinologist Arthur Walpole. He was head of the fertility regulation programme at the pharmaceuticals division of chemical company ICI in Alderley Park, and had experience of oncology research. ICI has played a big role in drug development (the pharmaceutical division became Zeneca, and then the major company AstraZeneca following a merger in the 1990s).

Tamoxifen was synthesised by Dora Richardson, an organic chemist who had been given the task of synthesising all the isomers of triphenylethylene derivatives to test as antifertility agents. By a process of fractional crystallisation, Richardson, working with spectroscopist G.R. Bedford, succeeded in separating the cis and trans isomers, revealing that activity resided with the trans isomer (ICI 46474) rather than its minor cis isomer (ICI 47699). Michael Harper, an endocrinologist, was also involved with animal testing work.

The comparative structures of early non-steroidal anti-oestrogens. Triparanol, a cholesterol-lowering medicine, marketed in the early 1960s, is illustrated to demonstrate the similarity of structure with the first non-steroidal anti-oestrogen MER-25. The structure of the anti-oestrogen tamoxifen and its oestrogenic-cis isomer are compared with the geometric isomers of clomiphene. Source: 50th anniversary of the first clinical trial with ICI 46474 (tamoxifen): then what happened?**

By 1964, ICI 46474 had been shown to be the most potent and least toxic of the compounds. It was identified as an effective post-coital contraceptive in rats, raising the possibility that it might be used as a morning-after pill in women. However, the basic pharmacology and physiology of ovulation and implantation turned out to be entirely different in women, where in preliminary clinical studies tamoxifen was found to induce ovulation rather than reduce fertility, suggesting that ICI 46474 might instead be used as a fertility agent, in a similar way to the US research.

But Walpole insisted the drug be trialled in breast cancer. Earlier in his career he had worked as a cancer researcher using ‘nitrogen mustard’ to inhibit blood and lymph tumours in rats, and had also worked with clinicians at the Christie Hospital in Manchester on high-dose oestrogen in breast cancer. This collaboration provided Walpole with clinical contacts and made him conscious that, in comparison to other similar compounds, ICI 46474 had unique characteristics in lacking androgenic properties.

ICI’s Arthur Walpole, Dora Richardson, Michael Harper. Source: 50th anniversary of the first clinical trial with ICI 46474 (tamoxifen): then what happened?**

In the first breast cancer trial of what was to become tamoxifen, conducted by Moira Cole and colleagues at the Christie Hospital, 46 postmenopausal women with metastatic breast cancer received either 10 mg or 20 mg daily of ICI 46474 for longer than 3 months. Results showed that 10 patients (22%) had a clear response, and a further 37% a partial response. Notably, pulmonary metastases in two patients resolved and lytic bone metastasis in one patient reosified. A total of 19 different side-effects were noted, with hot flushes and gastrointestinal intolerance being the most common.

The response rate of 22% was compared with historical hospital records showing a response rate of 25% for 64 patients treated with high-dose oestrogen and 16% for 60 patients treated with high-dose androgen. But where ICI 46474 distinguished itself from oestrogen and androgen was in the discontinuation rate, which occurred in 4% of patient receiving ICI 46474 versus 18% receiving oestrogen and 8% androgen.

In the conclusion the authors commented how impressed they had been by ICI 46474’s absence of toxicity and the low incidence and trivial nature of the side-effects. “The lower toxicity of anti-oestrogen seems to constitute its main advantage over other hormones,” they wrote.

The Christie findings were subsequently confirmed by a dose-response study comparing 10 mg or 20 mg ICI 46474 (by this time known as tamoxifen) in 64 patients with advanced, recurrent or metastatic breast cancer. Results of the study, by Harold Ward, an oncologist at Queen Elizabeth Hospital, Birmingham, UK showed 26 patients had reductions in tumour size to half or less, five had been in remission for over a year, and another 10 for over 6 months. “The particular advantages of this drug were the absence of virilisation or fluid retention and the fact that injections were not required,” wrote Ward.

However, at Alderley Park the studies were not received with universal enthusiasm by ICI. At a senior management meeting in 1972 the decision was made to abandon further development. The reasons were that there was estimated not to be a significant market for a palliative drug that would only be effective for a year in one in three cancer patients (at the time the incidence of metastatic breast cancer in the UK was only a few thousand a year), and that the compound had no patent protection in the US (this would only be granted in 1985).

At the time, major hopes for advances in cancer were pinned on new cytotoxic chemotherapeutic agents.

Walpole and colleagues were told they were supposed to be looking for contraceptive agents, not anti-cancer drugs. ICI only relented on their decision to pull the plug on ICI 46474 when Walpole tendered his resignation. The condition of Walpole staying in post was that ICI 46474 would be advanced for clinical use as an orphan drug for treatment of metastatic breast cancer and induction of ovulation.

In 1973, ICI 46474 was approved in the UK under the name Nolvadex (tamoxifen) for palliative treatment of advanced breast cancer in postmenopausal women and also for anovulatory infertility. Due to patent complications it was not approved by the US FDA until 1977. Tamoxifen became established as a non-toxic endocrine agent for treatment of postmenopausal women with metastatic breast cancer.

However, although about half of women with advanced disease responded to tamoxifen it had little impact on survival.

A note on oestrogen receptor assays

A crucial step was developing oestrogen receptor assays that could be used to identify this critical biomarker in breast cancer. It was the work of Elwood Jensen, Jack Gorski and others in the 1960s/1970s on the oestrogen receptor that led to the assays and the answer to the question about why only a third of breast cancer patients benefited from surgical ablative ovarian treatment.

As Craig Jordan noted: “Simply stated, the OER [oestrogen receptor] assay for breast cancer is based on the presence or absence of receptor protein extracted from a tumour biopsy. If the receptor is present, there is a high probability that the tumour will respond to ablative surgery, but if the receptor is absent, there is an extremely low probability of a tumour response.”

The ER assay was also effective at predicting the outcome of breast cancer treatment with high-dose synthetic oestrogen or androgen therapy, he added.

Mitch Dowsett, a breast cancer researcher at the Royal Marsden, writing on the evolution of ER testing, said that in 1974 overwhelming evidence was presented that breast cancers with a high content of the receptor protein would respond to endocrine treatment. This was at a meeting held by the National Cancer Institute in the US, attended by many of the oncologists destined to become major authorities in breast cancer.

“Consistent data showing clinically important separation of responders from nonresponders according to the detection of ER in their primary tumours was presented by several groups and summarised by McGuire et al.” (William McGuire was an important figure in breast cancer with multiple achievements including receptor and assay work. He co-founded the San Antonio Breast Cancer Symposium.)

“The following response rates were observed for ER+ versus ER- tumours: surgical therapies (e.g. adrenalectomy), 55% vs 8%; additive medical therapies (e.g. androgens), 60% vs 8%; and anti-oestrogens, 40% vs 19%. In most countries, analyses of ER rapidly became an important component of a breast cancer patient’s work-up, particularly if she was under consideration for surgical endocrine therapy.”

See also this article by Jensen and Jordan that includes details on steps in oestrogen receptor and assay history.

Much subsequent work was carried out on new developments in ER assays and other hormonal targets (progesterone, PR, in particular) and tests became routine in treatment and research. Research continues on aspects such as the prognostic value of the percentage of tumour cells that express ER and the balance between ER and PR.

Moving Tamoxifen to early breast cancer and adjuvant treatment

The second chapter in the history of tamoxifen is adjuvant therapy, where the goal is to destroy micrometastases that spread around the body at the time of surgery to prevent recurrence. Recognition of the value of tamoxifen in the adjuvant setting can be largely credited to the parallel work of Michael Baum and Craig Jordan.

Baum first started using tamoxifen in advanced breast cancer in the mid-1970s when at University Hospital, Cardiff, after becoming familiar with the Manchester and Birmingham studies. Working with Terry Priestman (an oncologist who also wrote books on breast cancer for patients) at the then Welsh National School of Medicine, and others, Baum compared chemotherapy to endocrine treatment (including tamoxifen) in 92 women with advanced breast cancer.

The trial was notable for being the first to incorporate a linear analogue self-assessment (LASA) tool to evaluate quality of life in treatment of cancer patients. The concept was borrowed from psychologists, allowing patients to assess their own appreciation of their illness by marking how they felt on a scale of 1 to 10, where at one end they could not feel worse and at the other could not feel better. “We found that the length of survival was the same for tamoxifen and chemotherapy, but that quality of life was much better with tamoxifen,” said Baum.

What was novel about the approach was that they were randomising patients between two treatment arms rather than giving everyone the same treatment and comparing outcomes with historical controls.

Baum was well versed in the importance of conducting randomised trials from his time in the early 1970s working as a research fellow for Bernard Fisher at the University of Pittsburgh. Fisher had been appointed chair of the National Surgical Adjuvant Breast and Bowel Project (NSABP) and was conducting large scale randomised trials evaluating less extensive surgery and postoperative chemotherapy in breast cancer patients. “The big lessons I learnt from Bernie were to define the question, define the outcome measure, and bring in a statistician at the beginning rather than the end of the study,” said Baum.

Baum also learnt about the concept of micrometastases. Working with his pathologist brother Edwin, Fisher had tracked labelled tumour cells and demonstrated that most cells forming metastases gained access to organs through the bloodstream rather than the prevailing concept at the time of metastases spreading in an orderly fashion as extensions of the tumour.

Baum was also influenced by the work of Diana Brinkley and John Haybittle, who had followed up the long-term outcomes for breast cancer patients. In their analysis of subsequent cases of death of breast cancer patients seen in the Cambridge area between 1947 and 1950 they found that after 20 years those who had initially recovered from breast cancer had 16 times more deaths from breast cancer than would have been expected in normal populations.

“The Cambridge study demonstrated breast cancer was likely to recur,” said Baum, who at the time was becoming disillusioned with working in advanced breast cancer as although tamoxifen slowed progression, it did not improve survival. “As a surgeon I wanted to influence the natural history of breast cancer for the better. I felt there was greater opportunity to make a difference in early breast cancer. My theory was by the time women show their first symptoms they already have micrometastases in a state of dynamic equilibrium. So I put my money on adjuvant systemic therapy to stop spread.”

Baum, who by this time had moved to King’s College, London, approached John Paterson, the tamoxifen product manager at ICI, who agreed to provide the drug for an investigator-led trial exploring tamoxifen as an adjuvant treatment in early breast cancer.

NATO study on adjuvant tamoxifen and other trials

The study acronym NATO – Novaldex Adjuvant Trial Organization – was chosen in a team brain storming session, said Baum, to make the Americans think it was a US trial and to read the results.

NATO recruited 1285 women with stage II premenopausal and stage I and II postmenopausal breast cancer who underwent total mastectomy with axillary node clearance or sampling followed by randomisation to tamoxifen therapy for 2 years or no further treatment.

Results published in 1983 showed that 2 years of adjuvant tamoxifen was associated with a reduction in recurrence and death (recurrence or death was 14.2% in women receiving tamoxifen vs 20.5% for controls). This was equivalent to a prolongation of the disease-free survival interval from 21 months to 30 months at the mean follow-up time of 21 months. Tamoxifen was well-tolerated and treatment was discontinued in only 14 (2.2%) of patients as a direct result of side-effects.

“The magnitude of the effect is comparable with that associated with adjuvant cytotoxic chemotherapy at a similar follow-up time, but with minimal toxicity and excellent compliance,” noted the authors.

The Scottish adjuvant tamoxifen trial, which started in 1978 and was led by surgeon Patrick Forrest and Helen Stewart and sponsored by the UK Medical Research Council, randomised 1312 patients to receive adjuvant tamoxifen for 5 years, or tamoxifen for treating relapse. Again, results showed a highly significant delay in relapse in the arm receiving tamoxifen.

Like NATO, the Scottish trial failed to find any outcome difference according to subgroup analyses exploring whether patients were hormone receptor positive or negative (later found in subsequent trials). “If you’ve just over 1000 patients it’s unlikely that you would have the statistical power to see a significant difference in terms of 30%,” said Baum.

Critics have subsequently suggested that the poor signal observed in the early British adjuvant studies were due to delays in testing that resulted in a deterioration of receptors. “At the time it was common practice in the UK for pathologists to wait a day after surgery before testing,” said Monica Castiglione.

The first trial to show the effect of hormone receptor status was a NSABP trial by Fisher and colleagues that randomised 1892 women with primary operable breast cancer to receive L-phenylalanine mustard (L-PAM) and 5-fluorouracil (5-FU) with or without tamoxifen, which showed the greatest benefits were obtained in women with the highest levels of oestrogen and progesterone receptors (see also this paper by Fisher and others).

However, it was not all good news for tamoxifen. A Swedish trial led by Tommy Fornander, at the Karolinska, Stockholm, was the first to recognise the long-term side-effect of endometrial cancer. The study showed that although a second breast cancer occurred less often in patients taking tamoxifen than controls, endometrial cancers occurred more frequently.

EBCTCG meta-analysis of tamoxifen trials

The tamoxifen adjuvant trials paved the way for the Oxford Overview analysis that ultimately established the supremacy of tamoxifen. The overview process, led by Richard Peto at the University of Oxford, involved a meta-analysis of relatively small and underpowered clinical trials to reveal robust information that would otherwise not have been available.

For the adjuvant tamoxifen trials, which informed the first meta-analysis of the Oxford series (and which went on to review many other aspects of breast cancer), Peto established a framework where international collaborators contributed data on an ongoing basis to a central group who produced a combined analysis. The methodology collected individual patient data, including a wide variety of items such as dates of randomisation and treatment allocation, allowing reliable exploration of the effects of treatment in subgroups.

“The overview provided an answer that was the best estimate from all the different studies. It represented an enormous contribution to advancement of tamoxifen because rather than having investigators squabble about their slightly different results, they could all agree on the overall result, which ultimately resulted in more rapid implementation of the findings,” said Jack Cuzick. “The approach succeeded largely because of Richard Peto’s huge ability to pester people until they gave him their data because they didn’t want to hear from him again.”

“By bringing all the studies together and using new statistical tools, it was possible to show a clear signal against the background noise,” said Baum.

The first results, presented at a hotel in Heathrow Airport in 1984, showed a clearly significant reduction in short-term mortality among women receiving tamoxifen. The EBCTCG, which developed the infrastructure to update the meta-analysis every 5 years with new data, went on to irrefutably demonstrate the benefits of tamoxifen in the adjuvant breast cancer setting (see here and here).

The latest meta-analysis of 20 trials (involving over 21,000 patients) showed that 5 years of tamoxifen in early breast cancer versus no adjuvant tamoxifen substantially reduced recurrence by 39% and breast cancer mortality by 30% in treated patients.

Craig Jordan’s work

In parallel to the clinical studies, Craig Jordan was using mouse models to investigate tamoxifen as a breast cancer adjuvant treatment. Jordan benefited greatly from working in the UK and US. As a pharmacology student at the University of Leeds, UK, Jordan worked as an intern at ICI, and later had his PhD (on oestrogenic and anti-oestrogenic activities of triphenylethylene) examined by Arthur Walpole.

Jordan did his ‘been to America stint’ at the Worcester Foundation for Experimental Biology (WFEB), Shrewsbury, Massachusetts, famous for the development of the combined oral contraceptive pill by Gregory Pincus and Min Chueh Chang. Here, a chance meeting with Elwood Jensen, the chemist and cancer researcher then at the University of Chicago, which led to Jensen, then best known for the discovery of the oestrogen receptor, teaching him how to induce mammary tumours in rats using dimethylbenzanthracene (DMBA) and techniques to measure oestrogen receptors.

Back in the UK and equipped with technical skills from Jensen and invaluable ICI contacts, Jordan was able to evaluate anti-tumour actions of tamoxifen in the laboratory. His contacts at ICI provided access to tamoxifen and arranged for thousands of rats to be sent from Alderley Park to Leeds.

In his DMBA rat model, Jordan demonstrated that long-term therapy with tamoxifen was more effective than short-term therapy, reasoning that daily treatment with tamoxifen for a month in rats was equivalent to about a year in women. The findings led to the idea that ‘longer’ was going to be a better strategy for adjuvant tamoxifen, with Jordan coining the catchphrase ‘Tamoxifen for ever’ when he presented his findings at an ICI symposium, held at King’s College, Cambridge, in 1977.

In the audience were Michael Baum and Helen Stewart, who at the time were finalising plans to use tamoxifen as an adjuvant treatment in the NATO trial and Scottish trials for 2 and 5 years respectively (see earlier).

ICI symposium, 1977, King’s College, Cambridge. Highlighted are Michael Baum, Craig Jordan, Helen Stewart. Source: 50th anniversary of the first clinical trial with ICI 46474 (tamoxifen): then what happened?**

Jordan’s Cambridge presentation did not meet with universal approval. Some delegates argued that adjuvant treatment encouraged premature drug resistance and clinicians would waste valuable palliative agents by using them too soon.

It was not until the publication of the Adjuvant Tamoxifen Longer Against Shorter (ATLAS) and adjuvant Tamoxifen – To offer more? (aTTom) studies in 2013 that the full extent of Jordan’s vision for the benefits of longer tamoxifen treatment strategies became apparent.

ATLAS study: 5 years vs 10 years tamoxifen

The international ATLAS study, started by Christina Davies, a medical researcher at the University of Oxford in 1995, randomised 6846 women with ER+ breast cancer who had been taking tamoxifen for 5 years to continue treatment for another 5 years or immediately stop treatment.

Results showed 617 recurrences in 3428 women allocated to continue versus 711 among 3418 controls, and reduced breast cancer mortality (331 vs 397). But treatment allocation had no effect on breast cancer outcomes for the 1248 women with oestrogen receptor negative disease (ER-). Additionally, results showed women taking tamoxifen had an absolute mortality increase for endometrial cancer of 0.2%.

aTTom study: 5 years vs 10 years tamoxifen

In the aTTom trial, the UK counterpart of ATLAS, 6953 women with breast cancer (2755 with ER+ cancer and 4198 untested) who had taken tamoxifen for 5 years were randomised to continue tamoxifen for another 5 years or stop treatment. With follow-up exceeding 10 years, 580 recurrences were recorded for women taking tamoxifen for 10 years vs 672 for those stopping after 5 years.

Between years 5 and 6 there was no difference between the groups, but starting in year 7 patients who continued on tamoxifen had a significant decrease in recurrence compared with those stopping treatment.

Like ATLAS, aTTom showed a small increase in the risk of endometrial cancer among patients in the tamoxifen group.

From ATLAS and aTTom longer duration of tamoxifen emerged as a new standard of care, leading to an updated ASCO clinical practice guideline in 2014, recommending treatment with adjuvant tamoxifen for 10 years in women with stage I-III HR+ breast cancer.

Tamoxifen in prevention

Craig Jordan had set the agenda for a possible role for tamoxifen in prevention with laboratory studies on animals, while a multidisciplinary project on breast cancer, which met in 1980 in the UK, considered that tentative steps towards prevention seemed possible from examining a range of factors including endocrine status.

Then, in a letter to The Lancet in 1985, Michael Baum and Jack Cuzick highlighted an analysis of the NATO trial showing adjuvant tamoxifen significantly reduced the incidence of contralateral breast cancer (i.e. new cancers in the other breast). The observation suggested that prevention of breast cancer using tamoxifen might be possible in high-risk women.

Writing in The Lancet in 1986, Cuzick (the statistician on NATO) and colleagues set out reasons for wanting to treat high-risk women with tamoxifen, arguing the approach was a logical extension of screening programmes and was scientifically based on the effectiveness of tamoxifen in established disease, the lack of side-effects and relationship of free oestradiol to risk of breast cancer.

Cuzick was influenced by the Guernsey study, initiated by Richard Bulbrook, a biochemist at the Imperial Cancer Research Fund, and John Hayward, a breast cancer surgeon at Guy’s Hospital, London, and colleagues, who had collected blood specimens from 6127 women aged 35 and over living on the island of Guernsey (selected for its stable and well-defined population). Analysis showed mean oestradiol concentrations were 12% higher among subjects who went on to develop breast cancer.

“Reading the Guernsey study, it really struck me that if development of breast cancer was related to hormones, then maybe we should be changing the hormonal background,” remembered Cuzick.

Royal Marsden prevention pilot becomes a landmark trial

Before several international tamoxifen prevention trials eventually got underway, one led by Cuzick, Trevor Powles, a medical oncologist at the Royal Marsden and colleagues embarked on a feasibility study as early as 1986. Known as Tamoplac, it expanded into a pilot and then a larger trial that may still be the largest single-centre preventive trial carried out in the UK. Powles, who led the trial, eventually enrolled nearly 2500 women at high risk of developing breast cancer to take tamoxifen or a placebo for 8 years. 

The Royal Marsden pilot paved the way for international trials by evaluating potential problems with recruitment, toxicity and safety, and compliance. It was successful: by 1994, Powles and colleagues reported: “Accrual remained high in spite of extensive informed consent regarding potential risk. Acute symptomatic toxicity was low for participants on tamoxifen or placebo and compliance remained correspondingly high with a predicted 77% of women on tamoxifen and 82% of women on placebo continuing medication at 5 years.”

They also said that safety monitoring indicated no adverse anti-oestrogenic effects of tamoxifen. The known link with endometrial cancer was noted, and other concerns, but the expected reduction in breast cancers would far outweigh all risks, they said.

The women were followed up after the 8 year treatment period and by 20 years “a highly statistically significant risk reduction was found that could be attributed principally to a reduction in the risk of ER+ breast cancers”.

But Powles had to contend with much scepticism about conducting a trial on healthy women and also a safety alarm about a possible relationship with liver cancer after a report by ICI of liver tumours in rats given high doses of tamoxifen. The feasibility study was stopped for 7 months before it could recruit all the first tranche of 200 women but recommenced after clearance by the ethics committee.

Mitch Dowsett, who was also involved with the trial, said Powles was instrumental in advocating for a prevention trial and overcoming objections. “Trevor was the only person prepared to drive through his study several years before the multicentre trials. His data and drive showed it could be done.” Added: Ian Smith, a medical oncologist at the Royal Marsden who also worked on the trial, said: “This ground-breaking trial led the world’s attempts to prevent breast cancer developing in  healthy women who were deemed at risk.”

Mitch Dowsett, Ian Smith, Trevor Powles at the Royal Marsden. Source: Tamoxifen: a landmark trial**

Cuzick encountered similar objections and delays as he prepared the International Breast Cancer Intervention Study I (IBIS-I) trial, one of three subsequent large-scale trials that followed the Royal Marsden lead. The possible connection with liver cancer was one, which went on to be reported in a study led by Philip Carthew at the University of Leicester, UK

“It was a ridiculous claim because by this time millions of women had already taken tamoxifen for breast cancer with no evidence of increased liver cancer,” said Cuzick. The dose used in the rat studies, he added, was proportionally 10 times greater than that used in women.

Despite his best arguments, Cuzick also encountered a reluctance among clinicians to embrace the concept of prescribing preventive medicines to apparently healthy women. “The IBIS-I study was truly a trial that not even Kafka could have imagined, with scrutiny from dozens of review committees questioning the value of preventive therapy for breast cancer,” he said.

Three international trials start

But in 1992 three large international trials did begin to recruit high-risk women: IBIS-I (see also the IBIS site), the P-1 study (coordinated by the NSABP in the US), and a trial led by Umberto Veronesi at the European Institute of Oncology [https://www.oncopedia.wiki/research-centres/european-institute-of-oncology], Milan, Italy.

By 2003, all three trials, and the Royal Marsden trial, had shown a reduction in development of breast cancer of 30% to 40% among high-risk women given tamoxifen vs those receiving placebo.

In the race to publish, due to better funding and the extensive NSABP network (which enabled quicker patient recruitment) the P-1 trial got there first.

“But prevention is a long-term game. The Americans declared victory too soon and didn’t do the long-term follow-up. With IBIS-I, we followed patients out to 20 years and found that the really big story was that the benefits of tamoxifen carried on for another 15 years after stopping treatment,” said Cuzick.

Long-term analysis of IBIS-I showed that protection provided by tamoxifen over placebo was similar from years 0–10.

“The most likely explanation is tamoxifen is reversing early changes in the carcinogenic process, which are known to take at least 20 years, so that – at least in some cases – the process must start again for new tumours to develop,” said Cuzick.

The Americans, explained Dowsett, had already “shot themselves in the foot” with regard to a long-term follow-up by allowing some women in the control arm to cross over to tamoxifen. “The result was they had no possibility of doing a long-term comparison of placebo versus tamoxifen,” he said.

In 1998, following the positive P-1 study, the FDA approved tamoxifen to prevent primary breast cancer in women at high risk. Approval was especially significant because it represented the first authorisation by the FDA of any agent for reducing risk in cancer.

But tamoxifen as a preventive agent never gained majority support. A study in 2001 suggested only 5% to 30% of high-risk women in the US who were recommended tamoxifen took it.

And tamoxifen uptake does not seem to be higher since. Systematic reviews and meta-analyses suggest that less than 10% of high-risk women offered an anti-oestrogen for primary prevention agree to take it.

One of the reasons for reluctance to take preventive agents, suggested Monica Castiglione, was that tamoxifen became a victim of its own success, changing the public’s view of breast cancer. “Today, as a result of drugs like tamoxifen breast cancer patients have a much better prognosis. The effectiveness of treatment alters the patient’s risk-benefit analysis. Healthy women don’t want to take a drug for 5 years that gives them hot flushes and a dry mucosa when their perception is that the disease itself is treatable.”

Summary of tamoxifen’s actions

Selective action of tamoxifen in target tissues. Tamoxifen is a SERM and has anti-oestrogenic action in the breast, but oestrogenic properties in the bone, endometrium. Source: Tamoxifen: Pioneering Medicine in Breast Cancer*

Aromatase inhibitors

As the disadvantages of tamoxifen became more apparent (including associations with endometrial cancer, increased risk of thromboembolic events, and treatment resistance) alternative agents were sought to suppress oestrogen.

Unlike tamoxifen, which interferes with the binding of oestrogen to oestrogen receptors on breast tissue, aromatase inhibitors (AIs) work by interfering with production of oestrogen. Before menopause, the enzyme aromatase in the ovaries is responsible for producing the majority of circulating oestrogen, but after menopause fat and muscle still produce aromatase, which is able to convert testosterone to oestrogen, albeit in much smaller quantities. Blocking the aromatase enzyme has proved of particular value in postmenopausal women.

A number of investigators provided critical insights into the aromatisation process, ultimately leading to the development of AIs. In 1955, Andre Meyer, a Swiss scientist working at the Worcester Foundation for Experimental Biology (WFEB) in the US, at that time the world centre for steroid endocrinology and reproductive biology, discovered that androstenedione was hydroxylated by the bovine adrenal to yield 19-hydroxy-A (19-OH-A).

Meyer recognised the enzymatic nature of 19-hydroxylation and suggested this was likely a first step in the biosynthesis of oestrogens from nonaromatic steroids, which he termed the aromatisation process.

Ralph Dorfman and Kenneth Savard, also at WFEB, provided further insights into aromatisation by demonstrating that conversion of radiolabelled testosterone to oestrogen occurred in the human ovary.

Then, in the 1970s, researchers at Southwestern Medical School, Dallas, discovered extra-glandular aromatase activity in men and women. They demonstrated that peripheral tissue provides the major source of oestrogen synthesis in men and postmenopausal women and that the reaction involves aromatisation of adrenal androstenedione. Prior to this, the consensus had been that steroids were only produced in endocrine glands, such as the ovary, testis and adrenal.

The quest then was to find agents that blocked aromatase, the enzyme catalysing the final step of oestrogen biosynthesis, with the aim of blocking oestrogen levels in postmenopausal women with breast cancer.

Mechanism of action of aromatase inhibitors and tamoxifen. Source: Aromatase inhibitors in breast cancer**

Aminoglutethimide (known as Elipten), a drug originally introduced in 1960 as an anticonvulsant, was the first AI. Aminoglutethimide was withdrawn for use in epilepsy in 1966 after it was found to cause adrenal suppression by inhibiting multiple cytochrome P-450 enzymes.

Since adrenalectomy was being used to treat breast cancer (see earlier), Richard Santen, at the University of Virginia, reasoned that aminoglutethimide might be repurposed for ‘medical’ adrenalectomy.

When aminoglutethimide was given in sufficient doses to inhibit the production of adrenal steroids, it was found that replacement corticosteroids were needed to avoid potential problems of adrenal insufficiency. Santen demonstrated that the aminoglutethimide-corticoid regimen blocked peripheral conversion of androgens to oestrogen and suppressed circulating oestrogens in postmenopausal women with breast cancer.

In subsequent work it was found that aminoglutethimide actually increased adrenal androgens and it was the co-administered hydrocortisone that led to the androgens being suppressed

In 1981 at the Royal Marsden, UK, aminoglutethimide was compared with tamoxifen in a randomised trial of 117 women with advanced breast cancer with both drugs found to be equally effective.

But it was clear that aminoglutethimide was far from an ideal agent, with several side-effects (largely resulting from its actions on other cytochrome P-450 systems, including lethargy, somnolence, nausea and vomiting). Another drawback of inhibition of P-450 enzymes was that patients given aminoglutethimide, which is now considered a first-generation AI, needed to take cortisol replacement therapy.

To overcome the non-specificity and side-effects of aminoglutethimide, there was a need to find more selective AIs that drugs would be better tolerated.

Second- and third-generation AIs

Formestane was the first specific AI to be developed, and here there is a story of a husband and wife team.

Harry Brodie, an organic chemist working at WFEB, first began developing AIs as potential contraceptive agents. However, under the influence of his pharmacologist wife Angela Hartley Brodie, he transferred his AI interest to breast cancer.

By serendipity, Angela, who was originally from the UK, had studied for her PhD at the University of Manchester, and was familiar with the Christie tamoxifen work.

“At the time, tamoxifen and other antioestrogens were known to be partial oestrogen agonists as well as antagonists, raising concerns that they may not be optimally effective against breast cancer and may have adverse oestrogenic effects. We reasoned that by using a different approach, compounds that blocked the production of oestrogen without having significant oestrogenic activity themselves might be identified,” wrote Angela.

The Brodies undertook systematic structure/function studies examining nearly 100 steroidal AIs, ultimately identifying 4-hydroxy-androstenedione (4-OHA) as their most promising candidate.

“Angela and Harry published a paper in 1973 listing their AIs, but didn’t mention 4-OHA. Later, when I asked them why they omitted 4-OHA, they confessed that they knew it was the most interesting one and wanted the opportunity to work on it before anyone else,” said Mitch Dowsett.

Focusing on 4-OHA as their lead compound, the Brodies demonstrated that it reduced oestrogen levels in rats. With the help of Craig Jordan, who was proficient with the DMBA rat model (see earlier), they showed that after 4 weeks of treatment with 4-OHA over 90% of tumours regressed to less than half their size.

Significantly, in contrast to tamoxifen, 4-OHA was not oestrogenic on other tissues, such as the rat uterus.

But when the Brodies tried to pursue drug development for 4-OHA, they faced scepticism from both academics and the pharmaceutical industry. At that time, pharma in the US was intensely focused on developing chemotherapy agents.

In the autumn of 1981, Angela presented her 4-OHA results at a meeting in Rome. After the talk, Charles Coombes, an oncologist then at the Royal Marsden, who had worked on the aminoglutethimide trial, approached her to express interest in conducting a small trial with 4-OHA in London.

In the resulting transatlantic collaboration, Angela was able to produce a kilogram of 4-OHA at her laboratory (now at the University of Maryland), which was then flown to London where it was formulated for injection and given to 11 women with metastatic breast cancer. Results showed 4-OHA successfully lowered oestrogen and induced a clinical response in four patients who responded for periods up to 4 months with healing of bone metastases and reduction in soft-tissue metastases.

This small study was enough to provide the proof of principle that a selective AI had efficacy as an anti-tumour compound. Heartened by the trial success, Brodie and Coombes approached Ciba-Geigy (now Novartis), which had produced aminoglutethimide, proposing it take on 4-OHA and expand trials.

In clinical trials, such as this conducted in Spain, 4-OHA (renamed formestane) was tested against tamoxifen as first-line treatment and found to have equivalent efficacy.

Formestane became the first selective AI to be available for treatment of breast cancer in the UK. However, with further studies it became apparent that formestane (which was never approved in the US) did not block aromatase sufficiently to improve efficacy over aminoglutethimide.

To find more potent and selective inhibitors, the pharmaceutical companies then used structure/function analysis to identify lead compounds. To enable this quest, Mitch Dowsett developed an ultra-sensitive oestrogen assay that was at least 10-fold more sensitive than other oestrogen assays at the time. “The technology allowed us to see how well oestrogen levels in postmenopausal women, barely measurable by other assays at the time, could be suppressed by different AI candidates,” explained Dowsett.

The first agent to emerge from this rationale drug development process was the second generation inhibitor CGS-1694 (later named fadrozole), an imidazole derivative of aminoglutethimide which inhibited the aromatase system in human placenta and rat ovary 400 to 1000 fold.

But investigators unexpectedly discovered fadrozole had aldosterone-blocking properties limiting the doses that could be given to block aromatase more effectively than with aminoglutethimide or formestane.

The field eventually narrowed to three companies that had developed highly potent AIs with higher specificity for the aromatase enzyme and fewer adverse events. AstraZeneca had anastrozole (Arimidex); Novartis, letrozole (Femara); and Pfizer, exemestane (Aromasin). Anastrozole and letrozole are non-steroidal AIs binding reversibly to the aromatase enzyme, while exemestane is a steroidal AI binding irreversibly. An additional advance was that all three agents were formulated to be taken orally rather than by injection.

Notably, development of all three agents took place in Europe.

Structures of the main aromatase inhibitors and the natural substrate androstenedione. Source: Aromatase inhibitors in breast cancer**

ATAC study: comparing anastrozole with tamoxifen

The ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial was the landmark study started in 1996 that challenged the long-held supremacy of tamoxifen in adjuvant treatment of breast cancer. ATAC was the first study to report a third-generation AI (anastrozole) against tamoxifen in the setting of early breast cancer. According to Michael Baum, who independently initiated the study, ATAC (with 9366 patients from 380 centres in 21 countries) is the biggest international trial ever undertaken.

After engaging with several companies, Baum chose anastrozole from Zeneca (soon to become AstraZeneca) for his first trial of an AI, partly due to his previous experience of working with the in the NATO study. Anastrozole was patented by ICI in 1987 and approved by the FDA in 1995 for postmenopausal women with advanced breast cancer after progression on tamoxifen, and then in the first-line for this group.

From the outset, ATAC was unusual in involving statistical input (by Jack Cuzick) into the trial design. “I’ve always believed that the statistician should play a role in the obstetrics team devising a study and not just be brought in at autopsy,” said Baum, who by this time had moved to the Royal Marsden.

The initial concept for a two-arm study was scuppered after the ethics committee dismissed the idea of using anastrozole in a treatment arm alone, although patient advocates supported the anastrozole-only approach. Tamoxifen was so effective, the committee argued, that patient’s lives would be put at risk by omitting the drug in any arm. By way of a compromise Baum and Cuzick persuaded the committee to run a three-arm trial.

ATAC was notable for being one of the first trials involving patient representatives in study design. Unlike other adjuvant breast cancer studies, at the time of the start of ATAC there were no data demonstrating anastrozole to be better than tamoxifen in advanced breast cancer. “The patient advocates argued to the regulatory authorities that breast cancer was such a dreadful disease that they would be prepared to take small risks to enable big leaps forward,” said Mitch Dowsett.

In the three-arm design, postmenopausal women with operable breast cancer who had completed primary treatment (including chemotherapy in about 20% of high-risk cases) were randomised to one of anastrozole daily plus a placebo for tamoxifen; tamoxifen daily plus a placebo for anastrozole; or a combination of the two active agents. Of the 9366 women in the study, 7839 (84%) were known to be HR+.

Results at a median follow-up of 33.3 months showed disease free survival (DFS) was 89.4% for anastrozole vs 87.4% for tamoxifen and 87.2% for the combination. Notably, DFS for combination treatment was not significantly different from tamoxifen alone. The DFS improvement with anastrozole was seen in HR+ but not in HR- patients. Additionally, results showed anastrozole was associated with fewer side-effects than tamoxifen, especially gynaecological problems and vascular events, but that arthralgia and fractures were increased.

Undoubtedly, the most surprising finding from ATAC was that women given combination treatment (anastrozole and tamoxifen) fared no better than those receiving tamoxifen alone. “One possible explanation is that when an oestrogen-deprived environment is created by an AI, tamoxifen is ‘seen’ as an agonist; whereas in a normal oestrogen-rich environment, tamoxifen can exert its classical anti-oestrogen effects,” said Baum.

For those receiving anastrozole, 5 year follow-up results of ATAC led by Tony Howell, an oncologist at the University of Manchester, UK, showed a 23% reduction in breast cancer relapse and a 42% reduction in the development of contralateral breast cancers compared with tamoxifen.

For the first time in nearly 20 years, agents had been found that were better than tamoxifen for adjuvant treatment of early breast cancer. “Publication resulted in a sea change in the adjuvant treatment of breast cancer and the drug was called a ‘blockbuster’ by the company because of its international widespread use (and sales!),” wrote Howell.

The ultimate success of anastrozole, added Howell, was undoubtedly due to the investigators’ determination to test anastrozole alone (versus in combination with tamoxifen). “If we had not persuaded the ethics committee to change its mind, we would have a null result and anastrozole would not be the widespread preferred treatment for early breast cancer that it is today.”

One missed opportunity from ATAC was that AstraZeneca did not support the investigators in collecting patient samples. However, samples were collected from UK patients, allowing Mitch Dowsett in the late 1990s to initiate TransATAC, a translational research initiative with access to 2000 tumour blocks from the trial. He used suitable blocks to evaluate the prognostic value of the 21 gene Oncotype DX score that predicts distant recurrence in postmenopausal HR+ women treated with tamoxifen.

In 2002, the FDA granted accelerated approval for anastrozole in early breast cancer based on ATAC, and in 2005 it gained regular approval.

BIG 1-98: comparing letrozole and tamoxifen

The BIG 1-98 trial, coordinated by the European Breast International Group (BIG), was the first study to compare tamoxifen with the AI letrozole in the adjuvant setting. The idea for BIG came in 1995 over a dinner that Aron Goldhirsch, a breast cancer oncologist who worked at the University of Bern and the European Institute of Oncology, had with Martine Piccart, a breast cancer specialist at Jules Bordet Institute, Brussels.

Piccart voiced her frustration with how European research on breast cancer had become “terribly fragmented”, with academic groups running many similar trials and consequently duplicating research efforts. “I immediately saw the eyes of my friend Aron getting suddenly very excited,” Piccart is quoted in an obituary of Goldhirsch.

In the early days, BIG had to contend with how research group leaders felt threatened. Piccart recalls how they told them, “BIG is not there to kill national groups; it is there to solve the question a lot more efficiently.”

BIG 1-98, (the first in the BIG series) was built on finding that women with metastatic breast cancer given letrozole as first-line treatment had significant improvement in 1 and 2 year survival rates compared with those given tamoxifen.

The initial proposal had been for a simple comparison between tamoxifen and one of three AIs, but Goldhirsch felt the study could also address more complex questions about sequential use of tamoxifen and letrozole.

The eventual BIG 1-98 study had four arms comparing 5 years of monotherapy with tamoxifen or with letrozole with sequences of 2 years of one of these agents followed by 3 years of the other in women with early stage, HR+ breast cancer. The intention-to-treat population included 8010 patients from 247 centres in 27 countries. The first analysis showed that compared with tamoxifen, letrozole significantly reduced risk of an event ending a period of disease-free survival (DFS).

Side-effect profiles were markedly different, with thromboembolism, endometrial cancer and vaginal bleeding occurring more frequently in the tamoxifen arm and skeletal fractures and cardiac events more common in the letrozole arm.

Analysis of overall survival proved problematic because women in the tamoxifen alone arm were given the option to cross over to letrozole once initial results became available. “Different analytical tools were developed to overcome this, including inverse probability of censoring weighted (IPCW) analysis, which achieves better estimates of relative treatment effects in the presence of selective crossover,” wrote Gaia Schiavon and Ian Smith, at the Royal Marsden.

At a median follow-up of 8.7 years, letrozole monotherapy was confirmed to be significantly better than tamoxifen for DFS and overall survival by both intention-to-treat and IPCW analysis.

Advantages for letrozole in terms of DFS and overall survival continued to 12.6 years, but contralateral breast cancer prevention proved more complex. Analysis showed letrozole was significantly more effective at reducing contralateral breast cancer in the first 10 years but that the advantage was reversed beyond this. The BIG 1-98 study team commented that this long carryover effect for tamoxifen was also seen in the 16 year follow-up of the IBIS-1 study.

BIG 1-98 also showed that sequential treatment offered no advantages. At a median follow-up of 71 months, DFS was not significantly improved with either sequential treatment approach compared with letrozole alone.

“In summary, there is no therapeutic gain in starting with tamoxifen and switching to an AI rather than starting with an AI upfront…a woman who wishes to switch from an AI to tamoxifen after 2 to 3 years because of side-effects or for whatever reason can do so without adversely affecting outcome, at least for up to 5 years of treatment,” wrote Schiavon and Smith.

Intergroup Exemestane Study (IES): switching to exemestane after tamoxifen

AIs were also studied as sequential therapy after 2 to 3 years of tamoxifen. The Intergroup Exemestane Study (IES), led by Charles Coombes, who had moved to Imperial College London, evaluated exemestane for sequential therapy after tamoxifen in the adjuvant setting. For the study, 4742 postmenopausal women with ER+ or ER-unknown breast cancer who had no evidence of disease after 2 to 3 years adjuvant tamoxifen were randomised to either switching to exemestane or continuing treatment with tamoxifen for 5 years.

Results after a median follow-up of 30.6 months found 183 first events (local or metastatic recurrence, contralateral breast cancer or death) occurred in the exemestane group vs 266 in the tamoxifen group.

Updated analyses at 91 months and 120 months demonstrated that the benefit associated with a switch to exemestane observed early in follow-up was maintained and that modest benefits could still be seen in overall survival.

EBCTCG meta-analysis: comparing AIs with tamoxifen in early breast cancer

The EBCTCG undertook a meta-analysis of randomised trials comparing AIs to tamoxifen in early breast cancer. For the analysis the group divided 31,920 postmenopausal women with ER+ early breast cancer into four subgroups: group 1 had received 5 years of an AI; group 2, 5 years of tamoxifen; group 3, 3 years of tamoxifen followed by 2 years of an AI; and group 4, 2 to 3 years of an AI followed by tamoxifen.

The meta-analysis showed AIs reduced recurrence rates by nearly 30% compared with tamoxifen and that 5 years of an AI reduced 10 year cancer mortality rates by about 15% compared with 5 years of tamoxifen.

“In summary, we have estimated that while tamoxifen reduces the risk of dying from breast cancer by about 30%, AIs do so by 40%,” said Dowsett.

Extended adjuvant endocrine therapy

Women with HR+ operable breast cancer remain at risk of relapse for many years beyond the completion of adjuvant endocrine therapy.

Recognition of this risk led to a number of studies to explore the role of extended adjuvant endocrine therapy beyond 5 years. Studies investigating the use of AIs after 5 years of tamoxifen (including ABCSG 6a with anastrozole, MA.17 with letrozole and NSABP B-33 with exemestane) all showed benefit for an additional 5 years of AI treatment after an initial five years of tamoxifen.

The EBCTCG undertook an analysis of 5 years of AIs versus 10 years. The meta-analysis, involving 24,912 patients from 12 randomised trials, found risk of recurrence was reduced by 33% when the AI was given for 5 additional years after tamoxifen but that the effect was a more modest 19% when it followed 5 years treatment with an AI.

As a result of these studies, ASCO guidelines recommended women with node-positive breast cancer receive extended therapy, including an AI, for up to 10 years of adjuvant endocrine treatment. Additionally, the guidelines stated that women with node-negative breast cancer should consider extended therapy for up to 10 years of adjuvant endocrine treatment based on consideration of recurrence risk using established prognostic factors.

In a 2021 interview, oncologist George Sledge at Stanford University Medical Center, US, gave an overview of AIs and tamoxifen in breast cancer. “Aromatase inhibitors have been shown to provide a somewhat higher disease-free survival rate than tamoxifen, so these are typically our initial go-to drugs in postmenopausal women,” he explained.

One of the tougher decisions to make, he added, was what to recommend for a postmenopausal woman who has been on an AI for 5 years. The outstanding question was, “Does this woman need further endocrine therapy?”

For women who cannot tolerate AIs, Sledge said, tamoxifen still has an important role to play since although it is not quite as good as an AI from the standpoint of DFS it could still be considered quite good. “I would rather have a patient on adjuvant tamoxifen than on adjuvant nothing because tamoxifen certainly does lower the risk for a distant recurrence. I usually try a second AI when a patient is having problems with the first one she tries, but if someone continues to object to the use of AIs, I will use tamoxifen.”

Monica Castiglione takes a more cautious view on long-term endocrine therapy “While longer appears to be better, with benefits of endocrine therapy lasting out to 15 or 20 years, questions remain about whether long-term treatment is really better for all patients.” The reality, she added, was that after 5 years women who had recovered from breast cancer want to put the experience behind them, “It’s not so easy for women to have to take a pill every morning that reminds them of their cancer.”

Both the optimal duration of endocrine treatment and the best type of endocrine therapy for individual women remain unanswered questions so this story will continue, with European researchers playing a central role.

SERMs, SERDs, LHRH and oestrogen therapy – brief notes

Tamoxifen is not the only SERM and others have been developed to take advantage of the oestrogenic agonist/antagonist effects depending on the target tissue, and also for better safety (such as for the uterus). Among the most widely used SERMs apart from tamoxifen is raloxifene, a second-generation SERM, which has a particular application in preventing osteoporosis (for which it was approved in 1997 by the FDA) while also lowering breast cancer risk in postmenopausal women (FDA approval in 2007).

Much of the work on establishing SERMs as a novel drug group was conducted by Craig Jordan and a tamoxifen team at the University of Wisconsin Clinical Cancer Center, and in studies such as the NSABP Study of Tamoxifen and Raloxifene (STAR) trial. Jordan has described the evolution of SERMs and the emergence of multifunction use; see also a paper on the SERM saga.

----------

Also an active field, although with only one approved drug (fulvestrant), is the selective oestrogen receptor degrader, disrupter or downregulator (SERD), which blocks oestrogen receptors and which is designed to be a ‘pure’ anti-oestrogen agent with no agonist activity. Fulvestrant is given to postmenopausal women whose disease becomes resistant to other anti-oestrogen therapy through the ESR1 mutation, and may also be given in combination with the inhibitor palbociclib. It has also been approved as a monotherapy for HR+/HER2- breast cancer.

Fulvestrant is, like tamoxifen, a result of research at ICI in the UK, under the code name ICI 182780, and was approved for its first indications in the US in 2002 and in Europe in 2004. Alan Wakeling and Jean Bowler were the scientists at ICI behind the creation of the SERD field, and trials of fulvestrant were carried out in North America and Europe. A limitation of fulvestrant is that it requires injection into muscle, and oral SERDs for early and advanced breast cancer are under investigation.

----------

Luteinising hormone-releasing hormone (LHRH) agonists are alternatives to ovarian ablation with surgery and radiotherapy in premenopausal women. The most widely used is goserelin (brand name Zoladex), which is yet another product of ICI’s pharmaceutical research in the UK (code name ICI 118630). Barry Furr was the endocrinologist who started a project on LHRH agents in 1973, working with chemist Anand Dutta, and they identified goserelin as a candidate in 1978.

Further work on a clinically viable product with colleague Frank Hutchinson led to the launch of goserelin as Zoladex in 1987, and later to an LHRH for prostate cancer. Furr was also instrumental in the development of other agents including fulvestrant and anastrozole. In 1991, Furr, Dutta and Hutchinson received a Society for Medicines Research award for their work on Zoladex. An important trial involving goserelin and tamoxifen was Stockholm STO-5, which ran from 1990 to 1997.

------------

A recent paper co-authored by Craig Jordan on the oestrogen paradox and how it is used to treat and prevent breast cancer refers to the contribution of Alexander Haddow and how oestrogen derivatives are being currently tested as candidates for women after anti-oestrogen therapies have failed. Another article by Jordan also covers the history and current clinical landscape.

------------

The complex treatment landscape for HR+ breast cancer and research on new drugs is set out in New steps on an old path: novel oestrogen receptor inhibitors in breast cancer.

------------

Further reading

The book Tamoxifen: Pioneering Medicine in Breast Cancer, by Philipp Maximov, Russell McDaniel and Craig Jordan, is a comprehensive volume of referenced articles on all aspects of the tamoxifen story.

Craig Jordan has written many papers on the development of hormonal therapy and tamoxifen in particular, and on the personalities involved. See:

• 50th anniversary of the first clinical trial with ICI 46474 (tamoxifen): then what happened?
• Tamoxifen as the first targeted long-term adjuvant therapy for breast cancer
• *Tamoxifen (ICI 46474) as a targeted therapy to treat and prevent breast cancer*
• Tamoxifen, a most unlikely pioneering medicine
• *Turning scientific serendipity into discoveries in breast cancer research and treatment: a tale of PhD students and a 50-year roaming tamoxifen team.

Viviane Quirke, an academic at Oxford Brookes University specialising in the history of medicine, has written two articles on tamoxifen: Imperial Chemical Industries and Craig Jordan, “the first tamoxifen consultant”, 1960s–1990s, and Tamoxifen – from failed contraceptive pill to best-selling breast cancer medicine: a case-study in pharmaceutical innovation. They are of particular note because Quirke had access to ICI documents and also to an unpublished history of tamoxifen by ICI’s Dora Richardson.

See also Case histories of significant medical advances: tamoxifen, by authors at Harvard Business School.