Abstract

The action of androgens is essential for the development of benign prostatic hyperplasia and carcinoma of the prostate. The androgen receptor is a ligand-dependent nuclear transcription factor. The transcriptional activation domain of the androgen receptor gene contains a polymorphic CAG repeat sequence. A shorter CAG repeat sequence within the normal range has been reported to be associated with increased risk of prostate cancer and symptomatic benign prostatic hyperplasia. Here, we examine the in vitro transcriptional activity of the androgen receptor (AR) with different numbers of CAG repeats within the normal range in a number of different cell lines of prostatic (LNCaP, PC3) and non-prostatic (COS-1, MCF7) origin. We utilize a luciferase reporter driven by the rat probasin promoter (-286/+28) containing two androgen receptor binding sites. Transcriptional activation of the androgen responsive reporter was observed to be greater with the AR containing 15 vs 31 CAG repeats in COS-1 cells (123.2+/-16.6 vs 78.2+/-10.9, P value 0.01) and the well differentiated prostate cancer cell line LNCaP (103.4+/-17.7 vs 81.4+/-7.7, P value 0.045). No difference was observed in the poorly differentiated prostate cancer cell line, PC3 (106.9+/-21.9 vs 109. 6+/-21.4, P>0.5) or the breast cancer cell line MCF7 (120.4+/-39.4 vs 103.1+/-23.1, P value >0.5). Dose-response experiments with varying quantities of ligand (0.01, 0.1, 1 and 10 nM dihydrotestosterone) or AR cDNA did not demonstrate significant differences in transactivation of the androgen responsive reporter in PC3 cells by the different AR constructs. This suggests that the lack of influence of CAG number in this prostatic cell line is not related to dose of ligand or quantity of androgen receptor. Western immunoblot analysis of androgen receptor protein in transiently transfected COS-1 cells did not demonstrate a difference in the expression of the androgen receptor protein with different numbers of CAG repeats following incubation in the presence or absence of androgen. Gel shift assay did not demonstrate increased DNA binding by androgen receptor with a shorter CAG repeat sequence. These experiments using a relatively androgen- and prostate-specific reporter provide evidence for an inverse relationship between androgen receptor transcriptional activity and the number of CAG repeats in the transcriptional activation domain. The effect of CAG repeat number was cell specific suggesting the involvement of accessory factors expressed differentially between different cell lines.