During the 1960s, the treatment of breast cancer was basic. Following self-detection or detection during a routine medical examination by a physician (there was no mammography), the tumour would be identified pathologically and the disease staged. Metastatic breast cancer would be treated with new combinations of cytotoxic chemotherapy and some early improvements were noted. Chemotherapy was the most widely used treatment and, based on successes in defeating childhood leukaemia, there was every reason for the medical community to believe that, if the right combination was discovered, cures would result. Regrettably, this was not the case as the non-specific nature of cytotoxic chemicals resulted in major life-threatening side-effects such as the destruction of normal tissue in the colon, bone marrow and reproductive organs.
Hormone therapy, despite being the standard of care for more than half a century,1–3 produced only transient responses for 1–2 years in one out of three patients; it was palliative and none of the patients survived. Those with earlier stages of breast cancer were treated with modified radical mastectomy and radiation of the surgical field. By today’s standards, the operations were barbaric and, despite the hope that all diseased tissue would be removed, often the disease had already spread microscopically. Over the next few decades, surgical techniques would improve to embrace lumpectomy and irradiation, but survival rates would remain the same as those observed with modified radical mastectomy. Adjuvant therapy for postsurgical destruction of micrometastases had not yet been considered, but, when the concept was advanced in the 1970s, there was optimism that cures would be developed4,5 if only enough of the cytotoxic chemotherapeutic combination could be given to the patient and she could tolerate the severe life-threatening side-effects. However, the correct combination of cytotoxic chemotherapies was elusive and a strategy of using bone marrow transplantation to aid patient survival was unsuccessful. Therefore, another approach was necessary and the tumour needed to be specifically targeted and destroyed to avoid killing the patient.
The solution that would introduce tamoxifen, the first successful targeted therapy to treat breast cancer, would revolutionize the approach to patient care and result in a cheap and effective way to enhance and prolong the lives of millions of women.6,7
Non-steroidal antioestrogens to regulate fertility
The discovery of the first non-steroidal antioestrogen was reported in 1958,8 but low potency and toxic side-effects prohibited clinical development. However, the observation that antioestrogens were potent and effective postcoital antifertility agents in laboratory animals started a search by the pharmaceutical industry for an effective ‘morning after pill’ to compete with the widely used oral contraceptive. An occasional contraceptive would reduce side-effects and improve safety, but this was not to be. The non-steroidal antioestrogens did exactly the opposite in women by inducing ovulation. One compound, clomiphene, remains the standard of care to this day,9 and was the first drug discovered to enhance fertility. However, the market was small in the 1960s and new agents were unlikely to succeed economically. The non-steroidal antioestrogens were medicines that required an application in a bigger market.
The link between oestrogen and the growth of some breast cancers has been recognized for most of the first half of the twentieth century.10 Oophorectomy was effective in causing tumour regression in premenopausal women with some breast cancers10 but, paradoxically, high-dose oestrogen therapy in postmenopausal women was similarly effective; one in three tumours responded for about 2 years but side-effects were severe and life-threatening with high doses of synthetic oestrogen.11 Early small clinical trials of non-steroidal antioestrogens for the treatment of breast cancer were all disappointing because of a poor therapeutic index, except for ICI 46,474 (Figure 1), a failed contraceptive.12–14 Side-effects were much lower than observed using high-dose oestrogen therapy, but development was terminated in May 1972 as there was no perceived financial incentive.15 Only the fact that Dr Arthur Walpole, the retiring head of the fertility control programme at ICI Pharmaceuticals Division, had been the external examiner for my PhD16 would change the path to progress in medicine. The meeting for my examination in September 1972 became the key that indirectly opened the door to advance ICI 46,474 to become tamoxifen.15 He recommended investment in my work to create a strategy for the use of ICI 46,474 in clinical trials. Regrettably, Arthur Walpole died suddenly on 2 July 1977 and never saw the success of tamoxifen.
Birth of a strategy to treat breast cancer successfully
The discovery that a high-affinity oestrogen-binding protein was located in oestrogen target tissues17 and human breast tumours18 provided the basis for a plausible explanation for the variable responsiveness of breast cancer to oophorectomy or adrenalectomy. The presence of the oestrogen receptor (ER) in the tumour would herald a high likelihood of responsiveness to oestrogen withdrawal, but the absence of ER would predict that the tumour would not respond to endocrine ablation and, therefore, would not benefit from surgical ablation of the ovaries or adrenals. At this time, the ER assay was seen as a test for whether or not to carry out ablative surgery.18 However, to a pharmacologist, the ER was a potential target for drugs.19,20 Dr Walpole and Dr Roy Cotton at ICI Pharmaceuticals Division agreed to my request for support (which continued for 8 years at the Worcester Foundation for Experimental Biology, USA, and the University of Leeds, UK). The Research Director at ICI Pharmaceuticals Division was persuaded to apply for approval from the Committee for the Safety of Medicines to place tamoxifen on the market (see Figure 1) in the UK for the palliative treatment of late-stage breast cancers, but funds would be made available to me both in the USA and subsequently at Leeds University (over the period 1974–9) to devise a clinical strategy to develop tamoxifen to its full potential. As a pharmacologist, I wanted to know how tamoxifen worked at the ER target. Thus, over a period of two decades, tamoxifen went from being an orphan drug with no prospects of development to the ‘gold standard’ for the adjuvant treatment of breast cancer and being successfully evaluated as the first chemopreventive to prevent any cancer.
The advance in the laboratory at the University of Leeds, UK, funded by the Yorkshire Cancer Campaign and an unrestricted research grant from ICI Pharmaceuticals Division (a wise investment in my young staff and students and unlimited shipment of laboratory rats from Alderley Park, Cheshire, UK, to the Department of Pharmacology at the University of Leeds), would provide the scientific basis for all future clinical work. The work at Leeds University focused on the molecular mechanism of the action of tamoxifen as an antioestrogen and antitumour agent in animal models, the metabolic activation of tamoxifen to hydroxylated metabolites with a high binding affinity for the ER, and the development of an animal model system to devise clinical strategies for the appropriate use of tamoxifen as a long-term adjuvant therapy in patients with ER-positive primary breast cancer and as a preventative for breast cancer.
The early work on the molecular mechanism of action of tamoxifen as an antitumour agent has been reviewed.19–21 This provided the basis for the idea that tamoxifen, or a hydroxylated metabolite,22–27 bound and blocked the binding of oestrogens to the human tumour ER as this was the target to treat breast cancer. Although clinical triallists in America embraced the concept,28,29 it was not until the 1990s that the British clinicians followed suit. Nevertheless, even the idea of targeting the ER in metastatic breast cancer was not going to save lives and a new approach was necessary that could provide a principle to be used by clinicians to move forward with scientific support to execute adjuvant clinical trials. In the mid-1970s, the clinical community was planning a logical but conservative year of adjuvant tamoxifen in unselected breast cancer patients with node-positive breast cancer. It was reasoned that patients respond for only 1–2 years during the treatment of metastatic breast cancer so the same must be true for the treatment of micrometastatic breast cancer following surgery.
The laboratory advance with animal models in the 1970s at Leeds University was that tamoxifen was effective as an antitumour agent only in ER-positive disease30 and long-term therapy using a carcinogen-induced rat mammary carcinoma model of micrometastatic disease demonstrated that ‘longer was better than shorter’ when it came to duration of adjuvant tamoxifen therapy. The body of experimental work came to two major conclusions:31–35
The tumour ER is the selective target for long-term adjuvant therapy; a month in the rat (equivalent to a year in a patient) is ineffective, but 6 months (equivalent to 6 years in a patient) is the correct strategy for complete tumour suppression and prevention of recurrence.34 Indeed, sequential antitumour agents may be more effective (this strategy is used clinically today, with an aromatase inhibitor following tamoxifen).
Any new short-acting but more potent antioestrogen will be less effective as a long-term antitumour agent than the long-acting but metabolically activated tamoxifen.25,34 The principle of targeted therapy proved to be true for the subsequent applications of raloxifene (Evista®, Eli Lilly, Indianapolis, IN, USA) as a chemopreventive agent some 25 years later.36 Raloxifene, a rapidly excreted drug, needs to be given indefinitely.36
Early indications from individual clinical trials suggested that, although 1 year of adjuvant tamoxifen did not enhance survival,37 2 years of adjuvant tamoxifen did,38,39 and 5 years of adjuvant tamoxifen enhanced survival considerably.40 However, the link with the ER target was not confirmed in the British trials.39,40
The innovation provided by the Oxford overview analysis, conducted by Sir Richard Peto, was to take the individual results of small randomized clinical trials of adjuvant tamoxifen and combine them to make a gold mine of statistically reliable data.41,42 The principles were clear in the analysis of the adjuvant clinical trials data: tamoxifen is effective only in ER-positive breast cancer, a longer duration of adjuvant therapy is better than shorter and there is a 50% decrease in recurrence and a 30% decrease in mortality even after stopping tamoxifen. The value of the first effective targeted therapy for cancer was confirmed.
The idea of preventing breast cancer
The idea of preventing breast cancer is not new. As long ago as 1936, Professor Antoine Lacassagne set out a vision for the prevention of breast cancer at the annual meeting of the American Association for Cancer Research in Boston.43 His vision, based on animal work, was to stop oestrogen accumulating in breast tissue.
This vision had no chance of becoming a reality until the discovery of non-steroidal antioestrogens and the commitment to address the clinical applications of tamoxifen. The idea of using tamoxifen as a chemopreventive was based on the idea that the drug would block oestrogen-stimulated growth promotion of nascent disease via the blockade of the ER and the drug would be safe enough to administer to healthy women at high risk of breast cancer for prolonged periods. The first studies using tamoxifen to prevent the early stages of rat mammary carcinogenesis44,45 took place long before the initiation of clinical trials, but this was because there were legitimate toxicological concerns. Nevertheless, multiple animal model systems were explored to ensure the scientific veracity of the prevention strategy46,47 before major randomized clinical trials were initiated in the early 1990s. Tamoxifen was selected as the only plausible medicine to test the worth of chemoprevention in women. Three pieces of evidence were important to advance the strategy.
Laboratory data supported the idea that an antioestrogen could prevent mammary carcinogenesis in animals.
The use of adjuvant tamoxifen to treat breast cancer reduced the risk of contralateral breast cancer.48
The extensive clinical use of tamoxifen had created a feeling of security, and the side-effects of the medicine and how to manage them were understood, or so it was believed.
In the 1980s, many aspects of the toxicology of tamoxifen were unknown. The toxicological and safety scrutiny to which the drug had been exposed was a fraction of that needed for a medicine intended to be administered to a population without a life-threatening disease.
The translational research story that was to emerge throughout the late 1980s and early 1990s was to create controversy, but did not prevent tamoxifen being tested successfully in clinical trials49,50 and becoming the first drug to be approved for the prevention of breast cancer. Perhaps most importantly, the intense scrutiny of tamoxifen toxicology would open the door to the discovery of a new group of medicines referred to as the selective oestrogen receptor modulators (SERMs). Tamoxifen had originally been expected to block the development of tumours via the ER signal transduction system: a targeted therapy. The reality, after two decades of molecular pharmacology (Figure 2), was that SERMs could play a much more versatile role in promoting women’s health.
Selective oestrogen receptor modulators
The predominant question to be addressed in the 1980s was whether or not a non-steroidal antioestrogen, tamoxifen, could be prescribed to women at elevated risk of breast cancer on a long-term basis. If oestrogen was essential to maintain bone density in women, would an antioestrogen prevent breast cancer but expose women to osteoporosis? A few women in every thousand would be saved from breast cancer, but perhaps hundreds would develop osteoporosis, which would be a disaster. Unexpectedly, all of the initial animal experiments consistently demonstrated that tamoxifen maintained bone density52–54 of ovariectomized rats, and these findings subsequently translated to clinical trials.55,56 Thus, it was possible that tamoxifen would be beneficial to postmenopausal women and prevent both bone loss and breast cancer. However, the tissue specificity of tamoxifen as an oestrogen in bone but an antioestrogen in the breast also involved target specificity in other organs and end points. Tamoxifen lowered circulating low-density lipoprotein cholesterol, but high-density lipoprotein cholesterol was unaffected.57,58 There was a possibility that tamoxifen could lower the risk of coronary heart disease. The results of some early studies were positive59–61 but the Oxford overview analysis and their enormous database was negative for this end point. Nevertheless, there was potential for SERM success.
The major concern with the target site specificity of tamoxifen was oestrogen-like action in the uterus and carcinogenesis. The interesting target tissue selectivity of tamoxifen’s pharmacology was first noted in mouse models of human tumour xenografts. It was known that tamoxifen exhibited species specificity: it was antioestrogenic in the rat uterus with weak oestrogen-like actions, but a full oestrogen in both the mouse vagina and uterus.13 Oestrogen is essential for the growth of ER-positive breast cancers implanted into athymic mice.62 Paradoxically, the administration of tamoxifen to athymic mice with a human ER-positive breast tumour xenograft did not cause oestrogen-dependent tumour growth. Tamoxifen was an antioestrogen in the human breast tumour but an oestrogen on mouse uterine growth.63 More importantly for patient care, a bitransplanted athymic mouse with an ER-positive breast tumour implanted in one axilla and an endometrial carcinoma on the other side produced a differential growth effect with tamoxifen. Oestradiol stimulated growth of both tumours, but tamoxifen blocked oestradiol-stimulated breast tumour growth but simultaneously encouraged the growth of the endometrial tumour.64 Clinical translation was rapid and clinical trial data demonstrated the same target site specificity of tamoxifen to reduce breast cancer incidence, but to increase the incidence of endometrial cancer.65 Almost simultaneously, tamoxifen was found to increase liver carcinogenesis in rats through DNA adduct formation during long-term therapy.66–69 Clinical epidemiology never subsequently showed an increased risk of liver cancer in women who had taken tamoxifen. Tamoxifen went forward to be tested successfully as the first chemopreventive agent in cancer; however, a new strategy was clearly needed to maintain momentum in the tissue targeting of the new group of drugs: the SERMs.
Raloxifene: the first of several selective oestrogen receptor modulators
The discovery of the target site oestrogenic and antioestrogenic actions of tamoxifen created an unanticipated opportunity to improve women’s health. The road map for the development of SERMs was articulated on two occasions at international meetings and the new strategy to treat multiple diseases was published in the literature.20,70 In 1988, at the first International Conference on the Chemoprevention of Breast Cancer, the idea was raised:
an extensive clinical investigation of available antioestrogens. Could analogues be developed to treat osteoporosis or even retard the development of atherosclerosis? Should the agent also retain anti-breast tumour actions then it might be expected to act as a chemosuppressive on all developing breast cancers.70
. . . a bold commitment to drug discovery and clinical pharmacology will potentially place us in a key position to prevent the development of breast cancer by the end of this century.70
Reproduced with permission from Jordan VC. Chemosuppression of breast cancer with tamoxifen: laboratory evidence and future clinical investigations. Cancer Invest 1988; 6:589–9570
This concept was refined and defined further at the 1989 AACR meeting in San Francisco with the presentation of the B. F. Cain Award:
We have obtained valuable clinical information about this group of drugs that can be applied in other disease states. Research does not travel in straight lines, and observations in one field of science often become major discoveries in another. Important clues have been garnered about the effects of tamoxifen on bone and lipids; it is possible that derivatives could find targeted applications to retard osteoporosis or atherosclerosis. The ubiquitous application of novel compounds to prevent diseases associated with the progressive changes after menopause may, as a side effect, significantly retard the development of breast cancer. The target population would be postmenopausal women in general, thereby avoiding the requirement to select a high-risk group to prevent breast cancer.20
Reproduced with permission from Lerner LJ, Jordan VC. Development of antiestrogens and their use in breast cancer: eighth Cain memorial award lecture. Cancer Res 1990; 50:4177–8920
But the proposal was based on the known pharmacology of the failed breast cancer drug called LY 156758. or keoxifene (see Figure 1), clinical development of which had been terminated in 1988.
The finding that tamoxifen, with low binding affinity for the ER tumour target, was metabolically activated to 4-hydroxytamoxifen,23,25 and the proposal to develop tamoxifen clinically, started a search for new agents with improved pharmacological properties.19 Keoxifene is a polyhydroxylated benzothiophene with high binding affinity for the ER71 but, unfortunately, drugs of this group are rapidly excreted, which makes keoxifene suboptimal as an anticancer agent.72,73 The programme to develop keoxifene as an antitumour agent to treat breast cancer was terminated once it was found that it has no anticancer activity clinically.74
However, keoxifene had already been tested in the laboratory against tamoxifen and, interestingly, it was observed that keoxifene was less oestrogenic in the rodent uterus than tamoxifen.71 However, the compound was comparable to tamoxifen in preventing bone loss in the ovariectomized rat;52 it could partially block tamoxifen-stimulated endometrial cancer growth,75 and it was effective at preventing rat mammary carcinogenesis, although tamoxifen was superior.46 It was suggested that the poor antitumour activity of keoxifene was due to its rapid excretion, leading to poor bioavailability,46 and as a result it would need to be given indefinitely. And, in fact, this is case for the drug, reinvented as raloxifene, for the treatment of osteoporosis76 and breast cancer prevention36 some 20 years later.
All the pieces were in place in 1990 (prevention of bone loss,52 prevention of mammary cancer46 and reduced oestrogenicity in the uterus and endometrial cancer71,75) to exploit a new strategy of using keoxifene to prevent osteoporosis and breast cancer without the concern about endometrial cancer.20 The repurposing of keoxifene as raloxifene created a new dimension in women’s health when it was approved for the prevention and treatment osteoporosis as it was associated with a reduction in breast cancer but no increase in endometrial cancer.76,77 Subsequently, it provided an alternative for postmenopausal women and for the prevention of breast cancer in women at high risk.36,78 However, despite the fact that both tamoxifen and raloxifene lower the levels of circulating cholesterol, the dimension of reducing the risk of coronary heart disease remained elusive.79
Nevertheless, the commercial success of repurposing raloxifene stimulated a huge increase in medicinal chemistry in the pharmaceutical industry and academia to exploit the ER target.80–82 Some fascinating new molecules have been tested clinically and show promise for creating a broad landscape of SERMs. These new molecules have recently been reviewed,83 but it is perhaps pertinent to state that ospemifene is available to treat vaginal dryness, a bazedoxifene-conjugated equine oestrogen combination is approved to treat hot flushes and menopausal symptoms and lasofoxifene is used to prevent osteoporosis, but with documented reductions in coronary heart disease, strokes and breast cancer, and no increases in endometrial cancer. The next advance is to exploit the progress made in the pharmacology of SERMs in medicine with new classes of compounds called selective nuclear receptor modulators to treat diseases never before considered possible.84
Forty years of progress in women’s health targeting the oestrogen receptor
The strategy of repurposing failed drugs (ICI 46,474 and keoxifene) to create medicines that selectively target ERs in different parts of the body is a significant advance. The principles developed in the 1970s to treat patients with ER-positive tumours with long-term adjuvant tamoxifen or use the ER to prevent the disease7 saved lives and reduced mortality and morbidity from breast cancer. The result is that the principles have now been exploited dramatically with the recent announcements that the ‘Adjuvant Tamoxifen: Longer Against Shorter’ (ATLAS) trial and the ‘adjuvant Tamoxifen – To offer more?’ (aTTom) trial, using 10 years of adjuvant tamoxifen, saves additional lives.85,86 However, there is a twist. Mortality decreases significantly after tamoxifen treatment stops, i.e. in the decade following therapy. This novel clinical finding had already been anticipated with the discovery of the new biology of oestrogen-induced apoptosis.87–91 Additionally, the National Institute for Health and Care Excellence has recommended that both tamoxifen and raloxifene be made available through the NHS in the UK to reduce the incidence of breast cancer. None of this was possible in 1970, but it has now happened. The advance in women’s health, targeting multiple sites around a woman’s body, is having profound effects on the prospects for progress. In the past 20 years, since the first description of SERMs,20,92 medical practice has witnessed unprecedented progress with the potential, through targeting all members of the nuclear receptor superfamily, to treat and prevent specific diseases with fewer side-effects.