Perinatal imprinting by estrogen and adult prostate disease

One of the major conceptual novelties in human medicine emerging during the last two decades is the fetal-origin hypothesis, challenging our traditional view on the pathogenesis of common chronic disorders. This hypothesis, originally put forward by Barker et al. (1), states that prenatal events, mainly of nutritional and metabolic nature, are memorized (“imprinted”) by the developing organism and, if inappropriate for the postnatal situation, may result in the development of diseases during the adult age. Although intriguing and supported by an increasing number of experimental and epidemiological studies, the “Barker hypothesis” must still be validated with respect to its underlying biology. In this issue of PNAS, Omoto et al. (2) reveal that a well tuned temporal and spatial expression of estrogen receptor α (ERα) and ERβ in the epithelial and stromal cells of the ventral prostate, concomitant with an estrogen signal, is required for normal morphological development of the prostate in mice. By employing ERα and ERβ gene-deleted mice, it was shown that ERα, but not ERβ, is necessary for the neonatal imprinting by estrogen during the first two postnatal weeks and branching morphogenesis of the developing ventral prostate later on. Furthermore, the study supports the yin-yang concept that endogenous estrogenic ligands mediate proliferative signals through ERα and differentiation and antiproliferative signals through ERβ during prostate ontogeny (2). The proliferative signals are thought to be mediated by stromal ERα, resulting in production of paracrine growth factors serving as epithelial mitogens (3). In addition, Omoto et al (2) show that 2- and 3-week-old mice express ERα in the prostatic epithelium, indicating that, during this particular time of development, ERα-mediated cellular proliferation may occur directly in the epithelium.

Imprinting by Estrogenic Compounds

The present results highlight the role of estrogenic compounds in prostate development and have potentially important implications for the understanding of the etiology and pathogenesis of prostate neoplasia. Benign hyperplasia of the prostate is widespread in elderly males, and prostate cancer remains the most common male cancer that, despite extensive previous and ongoing research efforts, is still enigmatic with respect to its causes. The present work by Omoto et al. (2) may also provide mechanistic clues to the observation that a transient exposure of neonatal mice to the nonsteroidal estrogen diethylstilbestrol (DES) causes profound changes in prostate development and predisposes the animals to a state of epithelial dysplasia in adult life, similar to the intraepithelial neoplasia found in humans (4-6). An analogous “imprinting” phenomenon in humans is the well known effect in daughters of women exposed to DES during early pregnancy. All of these “DES daughters” show an epithelial dysplasia in the upper vagina and a greatly increased risk of developing adenocarcinoma of the cervix and vagina, with a peak incidence before the age of 20 (7).

Temporal and spatial expression of estrogen receptors is required for normal morphological development of the prostate.

Estrogen Exposure to Perinatal Boys

The present observations by Omoto et al. (2) widen the perspectives of the fetal-origin hypothesis discussed above and provide a mechanistic framework that is open to challenge in future studies to come. Although results from rodents must be interpreted with caution, it is tempting to speculate on a translation of the current findings into human medicine, assuming that the discussed mechanisms of estrogen imprinting may be operative also in perinatal boys. Accordingly, there is a need to demonstrate possible ways of exposure to estrogenic compounds in boys of this age group. Interestingly, several such ways of exposure seem to exist. First, infant boys show physiologically high levels of testosterone because of a postnatal reproductive hormone surge at the age of 1-3 months, sometimes referred to as “baby puberty” (8, 9). Because testosterone is the natural substrate in estrogen biosynthesis and the responsible enzyme (CYP19/aromatase) is expressed by the prostate (10), it is plausible that estrogen can be produced locally within the prostatic tissue in the newborn period. Second, in late prenatal and early postnatal boys, circulating levels of estrogen are high, reaching equal or higher levels than in girls (8). Third, in the present and previous work (2, 11), it was demonstrated that the steroid 5α-androstane-3β, 17β-diol (3β A-diol), a metabolite of dihydrotestosterone (DHT) with estrogenic properties, could serve as a mitogen in the prostate in the neonatal period acting through ERα. Because 3β A-diol is produced by the human testis in high amounts (12), this steroid may act in conjunction with estradiol to expose newborn boys to estrogenic signals. In addition to the possible means of estrogen exposure summarized here, the growth factor hypothesis mentioned above also has support from clinical and epidemiological data in humans (13).

In conclusion, the results discussed here demonstrate clearly a role for estrogen in the developing prostate. The finely tuned temporal and spatial action of estrogen employing both ERα and ERβ renders the system potentially sensitive to interfering endogenous and exogenous ligands at critical perinatal time windows. Unexpected fluctuations in the levels of endogenous estrogen and exposure to phytoestrogens and other endocrine-disrupting chemicals from the environment, including the diet, may perturb prostate development. Further studies will show whether and how such perturbations (with disturbed imprinting?) may relate to adult prostate disease and cancer.