The first report that the ovaries influenced the development of the normal human female breast, is to be found in a case record from St Bartholomew's Hospital, London, dated 1775. The surgeon Percival (later Sir Percival) Pott had removed the normal ovaries of a twenty-three-year old girl during the repair of bilateral groin hernias, in which they had become entrapped. He reported that the patient had “enjoyed good health ever since” but that … “Her breasts, which were large, are gone, nor has she ever menstruated since the operation, which is now some years” (Pott 1775).

Girls start their puberty including breast development around age 8. Though the best correlation found is that with oestradiol, most girls still have at the onset of breast development oestrogen levels in the normal range for adult males. Therefore, it seems that oestrogen receptor sensitivity plays an important role in breast development. An important insight into the hormonal interplay in breast development was obtained when comparing the effect of exogenous oestrogens in forty-five girls with four different aetiologies of oestrogen deficiency: gonadal dysgenesis (GD). 17-alpha hydroxylase deficiency (17-OHase def), isolated gonadotrophin deficiency (IGnD) and multiple pituitary hormone deficiencies (MPHD). None of the girls had any breast development in the absence of oestrogens. Treatment with oestrogen was given in an identical manner to all. In GD, in whom the hypothalamic-pituitary functions were normal, treatment led to full development of breasts in a relatively short period. In IGnD and MPHD, breast development was incomplete even after years of oestrogen treatment. The conspicuous difference of the hormonal status between GD and IGnD and MPHD is that the latter two groups lack gonadotrophins, while in GD and 17-OHase def. these hormones are pathologically elevated. Thus the gonadotrophins seem to play an important role in mammary gland development. In the girl with 17-OHase def., it was also found that cortisol is necessary for normal breast development. A girl with hyperprolactinaemia due to a microadenoma had delayed puberty including slow breast development. Upon bromocriptine treatment the prolactin fell, the oestrogens rose and puberty including breast development occurred. Patients with Laron-type dwarfism (isolated IGF-I deficiency) have been found to have normal breast development.

The modifications that occur to breast tissue as a consequence of different physiological states of the female are considered with particular respect to oestrogen. Various tissue responses can be noted and the implications of modulation in cell proliferation, as a consequence of menstrual cycles, pregnancy, oophorectomy and oral contraceptive use, are considered. Assessment of oestrogen receptor by immuno-cytochemistry confirms the involvement of the steroid as a cell stimulant. Evidence from a variety of sources indicates the role in normal breast epithelium is more permissive than deterministic.

The epidemiological approach to the investigation of breast cancer is briefly described with special reference to case-control and cohort studies. Risk factors for breast cancer are then reviewed, paying particular attention to menstrual and reproductive factors and to body weight. Some of the hypotheses which attempt to draw the epidemiological observations together are also outlined.

The effects of exposure to exogenous oestrogens are considered in more detail. There is evidence from one study that the administration of large doses of stilboestrol during pregnancy increases breast cancer risk after a latent interval of about twenty years. Conjugated equine oestrogens administered to postmenopausal women may also slightly increase breast cancer risk after prolonged use. Little information is available about the effects of any other type of “hormone replacement therapy”. Combined oral contraceptives do not appear to have any general effect on breast cancer risk, but prolonged early use, especially use before first term pregnancy, may be harmful to the breast.

Oestrogens may be involved in breast cancer development by at least three mechanisms, all of which involve stimulation of growth. Firstly, they may stimulate proliferation of the normal gland and possibly expand the pool of cells susceptible to carcinogens; secondly, they may expand the population of cells initiated by carcinogens by a promoting action to produce clinically detectable tumours; and thirdly, they may stimulate the growth and progression of infiltrating tumours. All of these effects are probably produced by interactions with the oestrogen receptor. There is suggestive evidence from immunocytochemical studies that promotion to the phase of carcinoma in situ in volves expansion of a receptor-positive clone of cells. After progression, there is a gradual loss of endocrine responsiveness either because of loss of the receptor protein or of receptor function. The high proportion of receptor-positive cells in early malignant lesions holds great hope for preventive strategies using anti-oestrogens for suppression of ovarian oestrogen production.

The human breast undergoes profound changes during the menstrual cycle. The development of monocloiial antibodies to oestrogen-receptors (ER) allows study of receptor content and distribution in individual components of breast tissue. Forty-one specimens from premenopausal patients who underwent bilateral mastectomy for benign conditions were evaluated. The patients were categorised into early follicular, late follicular, early luteal, late luteal, and menstrual phases. An HSCORE, a semiquantitative evaluation of the intensity and distribution of receptor antigen localisation, was evaluated for the ductal and lobular epithelia as a tissue mean HSCORE. ER localised to the nuclei of the epithelium and was not observed in stromal elements. Receptor concentration was greater for ER in the follicular phase but this finding was not as striking as differences between lobules in a single breast and even between breasts. This heterogeneous expression between lobules was in contrast to the relatively uniform expression of receptor within a lobule of a given specimen.

A wide variety of methods has been used to investigate the role of the oestrogens in the aetiology and clinical course of breast cancer. They include measurements of urinary metabolites and the concentrations of oestrogens in blood, saliva, nipple aspirates, cyst fluids, lymph and breast tissues. The function of the oestrogen-binding proteins has also been examined. There is no clear evidence that an excessive oestrogenic stimulus is a determinant of risk of breast cancer and it may be sensible to question this widely-held belief. Variation in the oestrogenic stimulus within the normal range might be responsible for the observed differences in tumour doubling times and for clonal selection. Case/control studies which do not include some measure of tumour growth rates are not an efficient way of investigating these possibilities. Endogenous androgen concentrations show such a strong correlation with the natural history of the disease that it would seem unwise to investigate the role of the oestrogens without taking the androgens into account. Geographical variation in the incidence of breast cancer is probably not due to racial differences in endocrine function.

While the role of oestrogens in breast cancer has been recognised for over a century, until recently our understanding of the mechanism by which they act has been limited, despite zealous efforts on the part of many investigators to seek for differences in oestrogen levels or metabolism in cancer patients. Since the initiation of breast cancer occurs many years before the detection of overt disease, it seems reasonable to suggest that these earlier studies may have been carried out at the wrong stage of the disease. Using a time invariant radiometric procedure, we have established that of the three principal sites of oestrogen metabolism at C-2, C-16α, and C-17, only C-16α hydroxylation is elevated in women with breast cancer. Further studies in women at high risk for breast cancer for familial reasons showed that they have an elevated rate of 16α-hydroxylation. Parallel studies in mice with high and low rates of spontaneous mammary tumour formation show a close parallelism between the extent of 16α-hydroxylation and tumour incidence. Studies of 16α-hydroxyoestrone, the product of 16α-hydroxylation, have demonstrated that it is a unique oestrogen, capable of binding irreversibly to the oestrogen receptor and permanently activating it. Studies on the effects of diet and drugs on the extent of C-2 and C-16α hydroxylation have shown that while the former reaction is easily increased or decreased, the latter is refractory to treatment and behaves like a constitutive enzyme. Thus, treatment with cimetidine, obesity, or high fat diets decrease the extent of C-2 hydroxylation, while high protein diets, hyperthyroidism, smoking, or dioxins increase C-2 hydroxylation. In contrast, 16α-hydroxylation is essentially unchanged in any of these conditions.

A link between increased C-2 hydroxylation and decreased levels of breast and endometrial cancer is suggested by epidemiological studies in smokers. Conversely, decreases in this reaction as in obesity are associated with increased risk for these tumours. The aim of these studies is to develop a safe prophylactic regimen which will increase C-2 hydroxylation, resulting in a decrease in active oestrogen and a lower tumour incidence.

We review information highlighting the multiple roles of both steroidal (primarily oestrogen) and polypeptide regulators of mammary epithelial cell growth, emphasising the work of our laboratory. Effects of both classes of hormones are complex and involve multiple interactions with non-tumour, host tissue. Oestrogen may induce growth regulatory polypeptide growth factors and interact with them in hormone dependent breast cancer. Progression of hormone-dependent breast cancer to hormone independence may involve multiple genetic mechanisms of oncogene activation, loss of the oestrogen receptor, or loss of hormone responsivity of other gene products. Initial carcinogenesis and progression of mammary epithelium to cancer probably also requires both proliferative stimuli (oestrogen, polypeptide growth factors) and genetic damage, leading to qualitatively different hormonal responses (hormone responsive cancer). Future therapies should be designed to block hormonal stimulation better and to interfere with necessary activated or induced components of malignant progression such as oncogenes or polypeptide growth factors receptor systems.

In addition to secreted growth factors, acting as autocrine or paracrine mitogens, breast cancer cells secrete other proteins whose function and significance in mammary carcinogenesis may be important. Among them, proteases are particularly interesting since it has been suggested that they play a role in metastatic process, and since at least two of them, the tissue type plasminogen activator and pro-cathepsin D, the precursor of a lysosomal protease, are induced by oestrogens and secreted in excess in some mammary cancer cells.

In oestrogen-receptor-positive human breast cancer cell lines (MCF7, ZR75–1), oestrogens stimulate cell proliferation and specifically increase the secretion into the culture medium of a 52,000-dalton (52-kDa) glycoprotein identified as the secreted precursor of a cathepsin D bearing mannose-6-phosphate signals, which is routed to lysosomes via mannose-6-phosphate-IGF-II receptors. We have determined the structure of this procathepsin D by sequencing its complete cDNA sequence, its chromosomal localisation and its transcriptional regulation by oestrogens and other mitogens. In breast cancer cells, pro-cathepsin D production and secretion is much higher and its processing is altered compared to normal mammary epithelial cells in culture. In vitro, pro-cathepsin D acts as an autocrine mitogen on breast cancer cells and can be activated at acidic pH to degrade extracellular matrix, suggesting a role in mediating the effect of oestrogens on tumour growth and invasion. Retrospective clinical studies indicate a significant correlation between high 52-kDa cathepsin D concentrations in the cytosol of primary breast cancer and poor prognosis (Danish Breast Cancer Group, S. Thorpe, Copenhagen). We propose that among the proteases secreted by cancer cells, 52-kDa cathepsin D is important both as a tissue marker in breast cancer and as a potential factor involved in carcinogenesis.

Oestrogen regulates the growth of human breast cancer cell lines ZR-75–1, T-47-D and MCF-7 (KO and McGrath). Basal cell growth can be reduced (T-47-D) or eliminated (ZR-75–1) by prior growth in the absence of steroid and phenol red for three weeks, demonstrating that oestrogens can have long-lasting effects on cells in culture (termed “steroid memory”). Effects of oestradiol on different cell biological parameters are described and interaction with other steroids and serum growth factors is discussed. Antioestrogen action in these cell lines is affected by at least five parameters: (1) presence of phenol red, (2) time in culture, (3) cell density, (4) antioestrogen concentration, (5) steroid memory.

An in vitro model for loss of oestrogen sensitivity in breast cancer is presented. Both dependent (ZR-75–1) and responsive (T-47-D) cells lose oestrogen sensitivity when deprived of steroid in the long term but show a gradual increase in growth. For ZR-75–1 cells, the effects appear to be clonal but occur at a high frequency (about 1 in 1,000 cells). Parallel alterations in sensitivity to other steroids, antioestrogens and serum growth factors are shown. Molecular markers of this action are described and the results compared with the well-established model for loss of androgen/glucocorticoid sensitivity in SI 15 cells.

Recent advances in our understanding of the receptor components involved in the initial events of oestradiol action are reviewed with an emphasis on aspects of clinical relevance. Two main features of the models have evolved: (1) the domain structure of the ligand-binding unit, with distinct parts of the molecule being concerned with oestrogen-binding and interaction with DNA, and other parts of the molecule having important regulatory functions; (2) the regulatory regions (oestrogen-response elements) of oestrogen-sensitive genes which specifically interact with the receptor protein.

The homologies between different proteins capable of binding ligands as diverse as oestradiol, retinoic acid and thyroid hormone point to possible interactions between such ligands. Oestrogen-response elements can recognise oestradiol receptor proteins complexed with either agonists or antagonists. Interactions with DNA may play an important rôle in determining gene expression additional to those seen at the receptor level.

There is an increasing trend to administer adjuvant therapy to patients with all stages of primary breast cancer. However, since only about half of these patients will eventually relapse, it would be useful to have a set of factors that could predict which patients are at high risk for recurrence and which patients are probably cured by local therapy alone.

Using tumour specimens from our San Antonio Tumour Bank, we have confirmed the prognostic significance of axillary nodal status for predicting disease-free and overall survival. Among patients with node-positive disease, both oestrogen-receptors (ER) and progesterone-receptors (PgR) are predictive of patient outcome, but PgR is the more powerful factor. Using multivariate techniques, we have found that amplification and/or expression of the HER-2/neu oncogene provides additional prognostic information for node-positive patients.

Among patients with node-negative disease, ER and tumour size comprise a powerful set of prognostic factors. But ploidy status and S-phase fraction, as measured by flow cytometry, provide additional independent information for predicting disease-free and overall survival. Using these factors, sub-groups of node-negative patients with low risk of recurrence who will probably not benefit from adjuvant systemic therapy might be identified.

The enzyme complex responsible for the conversion of androgens to oestrogens is called aromatase. This enzyme is expressed in a number of tissue sites, including granulosa cells, placenta, hypothalamus and adipose. Oestrogen formation in adipose tissue has been implicated in the aetiology of endometrial and breast cancer. Aromatase comprises two components: a member of the cytochrome P-450 gene family known as aromatase cytochrome P-450, and a flavoprotein, NADPH-cytochrome P-450 reductase. We have recently obtained a full-length cDNA insert encoding human aromatase cytochrome P-450 and have expressed it in COSI monkey kidney tumour cells. The resulting transcript encodes a protein which is capable of converting androstenedione, testosterone and 16α-hydroxyandrostenedione to the corresponding oestrogens. Furthermore, characterisation of the aromatase cytochrome P-450 gene indicates that in humans there is only one such gene, which spans about 30kb and has at least nine exons. From these results we conclude that there is only one aromatase enzyme in humans. Study of the expression of this enzyme in human adipose stromal cells indicates that it is subject to regulation by a number of factors including cAMP, phorbol esters, and the growth factors, EGF, TGF-α, TGF-β, TNF and 1Lβ. Since several of these growth factors are produced by breast cancer cells in response to oestrogens, the possibility exists for paracrine and autocrine feedback loops within the breast, in which the growth of a tumour is regulated by oestrogen produced locally in the adipose cells of the surrounding tissue, and this in turn is regulated by the growth factors produced by the tumour as a result of oestrogen action.

The levels of endogenous steroids in the target tissues are thought to be more closely related to the biological effects than their concentrations in plasma. Therefore studies on oestrogen levels in malignant and non-malignant breast tissues (expressed per g wet weight) were conducted and the following conclusions were drawn:

(1) malignant tumours contained higher oestradiol levels than normal or benign breast tissues, whereas oestrone levels were more comparable;

(2) in contrast to the large decrease in plasma concentrations after menopause, the levels of oestradiol in tumours and in normal breast tissue did not change with advancing age;

(3) the oestradiol levels in breast tissues were lower than in uterine tissues, particularly in women before menopause; oestrone levels were very similar in all tissues studied;

(4) the mean oestradiol level was higher in oestrogen-receptor-positive tumours, but no correlation between the two parameters was found;

(5) preliminary results indicated lower oestradiol levels in tumours obtained from countries with a lower incidence of breast cancer;

(6) as far as available, oestrone levels were comparable and those of oestradiol were lower in fat tissues than in breast tumours;

(7) neither in vitro studies with breast tumours, nor in vivo results using myometrial tissues support a prominent role of the metabolism of oestrogens at the 16α-position in the development of tumours;

(8) the role of local factors in the production, retention and metabolism of oestradiol in the breast remains to be elucidated.

The breast of the nonpregnant woman secretes, retains or reabsorbs a variety of endogenous and exogenous biochemical substances of potential physiological and clinical significance. Among these substances are steroid hormones, prolactin, growth hormones, and peptide growth factors. Because the concentrations of these hormones secreted into the breast ducts are often significantly higher than the concentrations in serum, the breast epithelium is commonly exposed to very high concentrations of steroid and other hormones which may play a role in the aetiology and pathogenesis of breast disease.

This paper presents a brief review of published data on the cytological and biochemical constituents of breast fluid obtained by nipple aspiration and data on factors affecting oestradiol and oestrone concentrations in breast fluid. Concentrations of oestradiol and oestrone in nipple aspirate fluid are much higher than in serum, not correlated with serum levels, much lower in women who recently have given birth or breast-fed their infants, not related to day in menstrual cycle, and higher in women with biopsied benign breast disease compared with nulliparous women without breast disease.

Further studies of the interrelationships of hormones, other biochemical constituents and the cytopathology of breast fluid may provide new insights into the pathophysiology and aetiology of breast diseases.

Uptake of oestrogens into breast tissue and their subsequent metabolism can be studied by infusing radio-labelled steroids into volunteer patients. Such studies show that oestradiol is preferentially accumulated in breast tumours, oestradiol concentrations exceeding those of oestrone. This contrasts with plasma, in which oestrone concentrations in postmenopausal women are greater than those of the oestradiol. This observation suggests that tissue factors can modulate local oestrogen metabolism, and thus local steroid concentrations.

We have studied the local production of oestrogens from androgen, and also the interconversion of the major oestrogens, oestrone and oestradiol. Using isotopic techniques, it is possible to calculate the proportion of endogenous oestrogen produced from androgen, as opposed to uptake from the circulation. These studies suggest that a very variable proportion of tissue oestrogen derives from endogenous synthesis. After administration of aromatase inhibitors, aromatase activity is substantially inhibited, both in vivo and in vitro.

Relative oestrogen concentrations are determined in part by the activity of oestradiol dehydrogenase. In breast tissue, dehydrogenase activity is present and this is modified by various factors, including androgens. In addition, we have demonstrated that normal, benign and malignant breast tissues produce factors which can modulate both growth and dehydrogenase activity of cancer cells in vitro.

We conclude that breast tissue is a site of synthesis of oestrogens, and that a number of factors can affect their local concentration. Tumour cells produce growth factors which can influence steroid metabolism, and may thus be able to enhance favourably their own endocrine environment.

It is generally accepted that sex hormones play a role in the development of mammary cancer and it is a reasonable assumption that tissue concentration of oestrogens and their androgen precursors, as well as the activity of oestrogen synthetising or metabolising enzymes, may be valuable parameters of hormone dependency. Except for the sulphate conjugates and for testosterone, androgen and oestrogen levels are higher in mammary tissue (ng/g) than in plasma (ng/ml) and as distinct from plasma, tissue oestradiol (the active oestrogen) levels are higher than oestrone levels. Moreover, oestradiol levels are higher in oestrogen-receptor-positive tissues.

Sex hormone levels in mammary tissue find their origin in uptake from plasma and in local synthesis. It is unlikely that local aromatase activity accounts for a large fraction of tissue oestrogens but sulphatase activity may be an important determinant of tissue oestrogen concentration. The major precursors of tissue oestrogens are androstenedione and oestrone sulphate respectively, yielding oestrone which is transformed by the 17β-hydroxysteroid dehydrogenase to oestradiol. However, the reverse reaction, conversion of E2 into E1, is much more active and is inversely correlated to dehydroepiandrosterone (sulphate) concentration, which inhibits non-competitively the conversion of E2 into E1. This E2DH activity is higher in E2R positive tumours, suggesting that E2DH is a good marker of hormone dependency. As to prognostic factors, a follow-up study concerning sixty-two postmenopausal women, with a follow up of at least four years, revealed that in the recurrence group (n = 17 ), besides E2R and PgR concentrations, all hormone concentrations as well as E2DH activity were also lower than in the disease-free group.

Ovarian oestrogens appear to have a central role in the development of the breast. However, peripheral tissues including the breast and its tumours have the potential to synthesise and metabolise oestrogens. When ovarian secretion of hormones is low, for example, after the menopause, such extra-gonadal steroidogenesis may assume biological significance.

While no individual steroid conversion in the breast has been shown to correlate with levels of endogenous oestrogens, certain positive associations have been observed between enzyme activities and the presence, stage and hormone dependence of breast cancer.

Thus biosynthesis of oestrogen (aromatisation) in breast fat is significantly higher in patients with breast cancer as compared to that from women with benign breast disease. Within the breasts of women with cancer, oestrogen biosynthesis is enhanced in quadrants bearing tumour. In contrast, levels of 17β-hydroxysteroid dehydrogenase are not associated with the presence of malignancy, but in women with breast cancer, activity is significantly higher in breast fat of patients with large tumours and those with nodal metastasis. There is also evidence that oestrogen sulphurylation in breast cancer is associated with hormone dependence and among postmenopausal patients with oestrogen-receptor-positive tumours, response to the aromatase inhibitor, aminoglutethimide, is restricted to cancers possessing aromatase activity.