CAG repeat number is not inversely associated with androgen receptor activity in vitro

A negative linear association between androgen receptor (AR) function and the CAG repeat numbers is generally assumed. However, in vivo data concerning the association between CAG number and androgenic effects have been conflicting. Since former in vitro studies mostly have been based on extreme CAG lengths and reporter-systems containing viral promoters, the objective of this study was to investigate ARs with CAG lengths within normal range (16, 22 and 28) in a reporter-assay with the human prostate specific antigen promoter as target. We also wished to elucidate whether the interpretation of the results was depending on the methods used for adjustment of transfection efficiency and protein content. With beta-galactosidase as transfection control, 22CAG had the highest activity (set to 100%) compared with 16CAG [mean 78% (range 41-132), P = 0.005] and 28CAG [68% (26-162), P = 0.006], whereas renilla-luciferase resulted in 16CAG behaving similar to 22CAG [104% (56-165), P = 0.7] and 28CAG having lower activity [59% (33-101), P = 0.004]. In these experiments, also the empty vector displayed considerable background activity. When adjusting for AR protein, the 22CAG genotype had the highest activity; 16CAG and 28CAG displaying 20% (10-47, P < 0.0001) and 12% (5-21, P < 0.0001) thereof. Similar results were obtained with adjustment for total protein. Thus, by normalizing for AR-content, contrary to various control vectors, the highest AR activity was confined to the 22CAG and not 16 CAG, which may at least partly explain the discrepancy in data aiming to link physiological conditions to CAG repeat length.