Androgen receptor functions from reverse genetic models

Abstract

The androgen receptor (AR) is a ligand-dependent transcription factor involved in the regulation of many different physiological processes. AR dysfunction causes a diverse range of clinical conditions, including testicular feminization mutation (Tfm) syndrome, prostate cancer, and motor neuron disease (Kennedy’s disease). However, due to lack of genetic models, the molecular basis of the AR in these disorders remains largely unknown. Using a conditional targeting technique based on the Cre-loxP system, we successfully generated null AR mutant (ARKO) mice. ARKO males exhibited normal healthy growth, but showed typical Tfm abnormalities. Hormonal assay of ARKO males revealed that while serum androgen levels were very low, estrogen levels were normal. Another hallmark of ARKO males was late-onset obesity, with marked accumulation of white adipose tissue. To clarify the role of human AR (hAR) mutants with expanded polyQ stretches as observed in neurodegenerative disease, we also established a Dorsophila model in which either wild-type or polyQ-expanded hAR were ectopically expressed. Although no overt phenotype was detected in adult fly-eye neurons expressing mutant hAR, the ingestion of androgen caused marked neurodegeneraton.

Introduction

The androgen receptor (AR) plays an essential role in a variety of biological processes, not restricted to male reproductive functions such as Wolffian duct development and spermatogenesis [1], [2]. Testosterone and its more potent metabolite, dihydrotestosterone, act as ligands for AR, and liganded AR activates the target gene expression at transcriptional level. Liganded AR form homodimers and bind specific DNA elements referred to as androgen responsive elements (ARE) in target gene promoters [3], [4]. The AR gene comprises eight exons that encode a 110 kDa protein. Members of the steroid/thyroid hormone family share common structural features, with distinct functional domains referred to as domains A to E(F). The highly conserved middle region (C domain) acts as a DNA binding domain, while the ligand binding domain (LBD) is located in the C-terminal E/F domain. During ligand-induced transactivation, the N-terminal domains A/B and the steroid receptor LBD act as interacting regions for the co-activator complexes [5], [6], [7]. The autonomous activation function-1 (AF-1) within the A/B domain is ligand-independent, while AF-2 within the LBD is induced upon ligand binding [8]. While unliganded LBD appears to suppress the function of the A/B domain, ligand binding to the LBD is thought to evoke LDB function and restore A/B domain function through an, as yet undescribed, intramolecular alteration of the entire steroid receptor structure.

In contrast to the other members of the steroid receptor superfamily, a number of clinical disorders of the AR have been reported [9], [10], [11], [12], [13], [14]. Classical AR functional abnormalities cause a spectrum of disorders of androgen insensitivity syndrome (AIS) or testicular feminization mutation (Tfm) [10], [13], [14]. AR mutations underlying these disorders include amino acid substitutions in the DNA or ligand binding domains, point mutations leading to premature stop codons, and deletions of the AR gene. In addition, expansion of a polyQ repeat region within AR has been implicated in the pathogenesis of a motor neuron disease called Kennedy’s disease [9], [12]. AR is a relatively large protein to other steroid receptors, due to the long N-terminal A/B domain that contains this polyQ repeat. However, the molecular basis of AR function underlying these AR-related disorders remains largely unknown due to the lack of stable genetic models. In this article, we present recent results of our studies into genetic models of loss of AR function in mice [15], [16] and gain of AR function in Drosophila [17].