Although hints and reports of a role for hormones and the causation of human cancer had ap- peared over the past several centuries, the first practical application of hormones in oncology was their use in the therapy of specific human neoplasms. The first practical application to the human of what little knowledge existed at the time was the demonstration by Huggins and Hodges of the partial androgen dependence of many human prostatic cancers, as evidenced by the beneficial effects of orchiectomy or synthetic estrogens in many patients even with meta- static disease (Huggins and Hodges, 1941). Since then, significant evidence has developed to indicate a role for hormones in the development of several common human neoplasms. Miller (1978) pointed out such a relationship with breast, ovarian, endometrial, and prostate cancer in the human. More recently Henderson and colleagues (Henderson et al., 1988) have presented reasonable evidence to argue for a major role of estrogens in cancers of the breast and genital tissues in both males and females. However, as Kodama and Kodama (1987) have pointed out along with others, a variety of additional factors such as diet and other lifestyle components (Chapter 11) are involved in the hormonal interactions with tissues that lead ultimately to neoplasia.
To date there is little evidence for the carcinogenicity of polypeptide hormones in the hu- man, such as with the animal systems indicated above. The best evidence is for a role of thy- rotropin in the genesis of human thyroid cancer, as referred to above. In Williams’ review (1989), several epidemiological studies indicated that risk factors such as iodide deficiency, “dyshormonogenesis,” and ionizing radiation were all associated with elevated thyrotropin lev- els in the serum of these individuals. Furthermore, patients with neoplasms of the pituitary that secrete high levels of growth hormone (Chapter 18) do have an increased risk of several types of malignant neoplasms (Barzilay et al., 1991). Thus, based on animal experimentation together with these few epidemiological studies, the data are suggestive of a carcinogenic effect of en- dogenous polypeptide pituitary hormones in the human analogous to experimental investigations in the animal.
That estrogens are carcinogenic in the human, either by direct or indirect mechanisms, was first exemplified by the report of the development of primary vaginal adenocarcinoma in young women whose mothers had taken diethylstilbestrol (DES) during the first trimester of the rele- vant pregnancy (Herbst and Scully, 1970). A registry for such exposed women was developed, and in 1987 Melnick and associates (1987) reviewed 519 cases of clear-cell adenocarcinoma of the vagina and cervix in which the majority of the patients’ mothers had received DES or a re- lated hormone during pregnancy. The risk that such a neoplasm will develop in an exposed fe- male from birth through age 34 was found to be approximately 1 in 1000. On the other hand, the incidence of a nonneoplastic condition resulting from the retention of embryonic glands in the vagina and cervix, termed vaginal adenosis, was much higher, approaching 70% of females ex- posed to DES in utero (Johnson et al., 1979). As further support for the causative relationship of DES and this vaginal pathology, Newbold and McLachlan (1982) reported that mice exposed prenatally or neonatally to DES developed both vaginal adenosis and adenocarcinoma.
For more than three decades, preparations of synthetic estrogens and progestogens, singly or combined, have been used as oral contraceptives by women in the United States (Huggins and Zucker, 1987). After the first decade of their widespread use, reports (Edmondson et al., 1976; Goldfarb, 1976) of a significant increase in the incidence of liver adenomas in women using such oral contraceptive preparations appeared. The increased risk of such lesions directly paral-leled the duration of use; after 8 years, there was a 500-fold increase in the risk of the develop- ment of hepatocellular adenomas in women using oral contraceptives continuously for this period (cf. Huggins and Zucker, 1987). The use of sequential oral contraceptives, in which estro- gen was given for 2 weeks followed by an estrogen-progestogen combination for 1 week, re- sulted in the occurrence of endometrial carcinoma in young women on this regimen (cf. Huggins and Zucker, 1987). Following the first reports of such occurrences (Silverberg and Makowski,
1975; Lyon, 1975), manufacturers voluntarily removed such preparations from the market. As further documentation of the carcinogenicity of estrogens in the human, the administration of synthetic or natural estrogens as hormone replacement therapy to postmenopausal women has revealed a strong association between estrogen use and cancer risk. The risk is in proportion to dose and duration of use, with a three- to sixfold increase in risk after 3 to 10 years of use and a more than tenfold increment in risk after more than 10 years of using unopposed estrogen (Bar- rett-Connor, 1992). Strikingly, when exogenous estrogens given to postmenopausal women are supplemented with progestogens, the risk of endometrial cancer is dramatically reduced (cf. Voigt et al., 1991). Thus, these findings with the use of exogenous estrogens further support the concepts of Henderson, Miller, and others of the importance of endogenous hormones, espe- cially estrogens, in the development of human cancers of a variety of tissues, most of which are recognized as end organs for estrogen effects.