|[Frontiers in Bioscience 1, d59-71, March 1, 1996]|
THE ANDROGEN RECEPTOR: A MEDIATOR OF DIVERSE RESPONSES
Evan T. Keller, William B. Ershler, and Chawnshang Chang
The Institute on Aging and the Department of Human Oncology, University of Wisconsin, Madison, WI 53706, USA.
Received 01/16/96; Accepted 02/22/96; On-line 03/01/96
When initially diagnosed, the majority of prostate cancers respond to androgen deprivation; however, over time they lose their dependency on androgens and the tumor recurs. The molecular events responsible for development of androgen independence are currently unknown. Altered AR expression or function, loss of tumor suppressor genes, and overexpression of or altered response to growth factors may all be implicated in this phenomenon.
AR is an important mediator of early prostate cancer growth. This is underscored by the marked tumor regression associated with androgen deprivation. However, the importance of the AR during the late stages of prostate cancer is not very clear. Initial studies demonstrated that AR mRNA is present in androgen-sensitive prostate cancer cell lines but is absent or is expressed at low levels in androgen-independent lines (102). Sato et al. demonstrated that as prostate cancer progresses to a less differentiated state AR expression was decreased (103). Recent immunohistochemical studies have revealed that the total AR content of prostate cancer cannot be used for assessment of prognosis; rather the heterogeneity of AR expression throughout the tumor was significant as a predictor of response of tumor to hormonal therapy (104-106). Chodak et al. also demonstrated that AR was uniformly present in the nuclei of epithelial cells in normal prostate, whereas in prostatic adenocarcinoma only some of the epithelial nuclei stained for AR (104). Sadi and Barrack demonstrated that cancer cells in patients with advanced prostate cancer who responded poorly to therapy exhibited a significant heterogeneity in the staining intensity of the AR (105). These observations suggest that AR expression may be modified during prostate cancer progression and may somehow be associated with the development of androgen independence. One must bear in mind that these studies do not eliminate the possibility that a mutated form of the AR may be expressed that no longer reacts with the antibodies to AR.
AR mutation may confer androgen independence on prostate cancer cells. Although mutations of the AR in prostate cancer are not often identified, they have been reported in up to 25% of the patients in one study (107). A variety of unique mutations, typically occurring in the ligand-binding domain, have been found in prostate cancer (103, 108-111). The LNCaP AR has a single point mutation in codon 877 (ACT->GCT) changing Thr to Ala in the ligand binding domain (112). This mutation decreased dependency of AR activation on androgen because progestagens, estradiol and anti-androgens could compete with androgens for binding to the androgen receptor and could stimulate both cell growth and secretion of prostate specific acid phosphatase (113, 114). This mutation may have a biologic relevance in vivo since Gaddipati et al. reported that 6 of 24 prostatic tissue specimens derived from transurethral resections of patients with metastatic prostate cancer had AR mutations identical to those found in the LNCaP cell line (107). In addition to activation of mutated AR by a greater number of ligands, the AR can be mutated in such a fashion that it is constitutively active (6). In vitro induced deletion of the AR's ligand-binding domain resulted in an AR that in the absence of steroid possessed up to 90% of the activity a steroid-induced wildtype AR (6). Though not observed in prostate cancer, a constitutively activated AR would circumvent the need for androgen, resulting in a phenotype of androgen independence.
A variety of growth factors have been implicated in either autocrine or paracrine prostate cancer cell growth (reviewed in (115)). Culig et al. provided compelling evidence for the ability of various growth factors to activate androgen responsive genes via AR in prostate cancer cells (116). Insulin-like growth factor I (IGF-I), epidermal growth factor (EGF), or keratinocyte growth factor (KGF), stimulated an AR-responsive reporter gene used to cotransfect DU-145, an AR negative prostate cancer cell line, with an AR expression vector. In the absence of the AR expression vector, the reporter gene was not activated. Furthermore, exposure of LNCaP cells to IGF-1 resulted in expression of prostate-specific antigen, which is encoded by an androgen responsive gene (93, 117). These actions were blocked by casodex, a pure AR antagonist. These experiments suggest that androgen independence could arise as a result of overexpression of growth factors in the local environment of the prostate.
In summary, the molecular events which facilitate development of androgen independence are currently unknown. However, loss or mutation of androgen receptor may play a role in selection of androgen-independent tumors. Such mutations in the ligand binding domain of the AR may result in a promiscuous response to non-androgen ligands allowing prostate cancer proliferation in response to non-androgenic growth factors.