|[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
The AR is a member of the steroid hormone receptor family. These receptors regulate gene transcription by interacting with a specific DNA sequence in a ligand-dependent manner. The AR is a 110 kD nuclear protein which consists of approximately 918 amino acid residues (1, 3, 5). Similar to other steroid hormone receptors, the AR consists of transactivation (6-10), DNA binding (11-16), nuclear localization (17-19), dimerization (20-22), and ligand binding domains (6, 23-27) (Fig. 1). Of these domains, the N-terminus is the most variable while the other regions are highly conserved. Within the N-terminus, the region of amino acids from 141 to 338 consist of polyglutamine and polyproline residues (3) which appear important in transcriptional activation (18, 28). The DNA-binding domain consists of 68 amino acids which fold into two zinc fingers capable of binding DNA. Four cysteine residues non-covalently bind the zinc ion within each finger. The remaining 295 amino acids form the C-terminus and encode the hinge region/nuclear localization signal and the ligand-binding domain.
Figure 1. Schematic illustration of the functional domains of the human androgen receptor. The exons are labeled A-H. The numbers above the box indicate the amino acid number at the junction between the exon and the intron. The domains responsible for their specific function are also indicated by lines with the numbers of amino acid residues.
As demonstrated by a variety of methods, AR is present in most tissues. 3H-androgen binding assays were the initial approach used to determine tissue distribution of AR. By injecting (3H)-testosterone into rats followed by quantitation of (3H)-androgen uptake in various tissues, Gustafsson and Pousette (29) were able to identify target organs of androgens. The development of anti-AR antibodies has complemented the autoradiographic methods of AR detection (30-33). Using anti-AR antibodies, Takeda et al. (32) evaluated AR distribution in various rat tissues. They were able to demonstrate that all male sexual organs in the rat showed a strong positive nuclear staining for AR, whereas several other tissues, including hepatic, renal, neuronal, muscular and female reproductive organs had weak, albeit positive, nuclear staining. In fact, the only tissue which did not demonstrate staining for AR was the spleen. The use of microwave-based antigen retrieval for AR (34) enhanced the immunohistochemical detection of AR in paraffin sections allowing for evaluation of archival tissue sections.
In addition to identification of AR protein, detection of AR mRNA in tissues has been accomplished by several methods. Cloning of human AR (hAR) and rat AR (rAR) cDNA (1, 35) allowed for development of probes which were used to detect in various tissues the AR mRNA by Northern blotting or by in situ hybridization. In addition to confirming previous immunohistochemical staining data, the analysis of AR mRNA by Northern blot resulted in identification of two isoforms in the A/B domain of AR mRNA in the vocal organ of Xenopus (36) and the brain of rodents (37). The sensitivity of detection of AR mRNA was greatly improved by competitive RT-PCR (38). Table 1. summarizes the relative expression (compared to prostate) of AR mRNA in various tissues of male and female Sprague-Dawley rats based on competitive RT-PCR (38). These results demonstrate agreement between mRNA levels and immunohistochemical staining intensity (39). One notable exception was the ability to demonstrate AR mRNA in the spleen using RT-PCR (38) as opposed to the inability to detect AR protein by immunohistochemical staining (32). This result is most likely attributable to the extreme sensitivity of RT-PCR as compared to immunohistochemistry.
AR expression is modified during fetal development, sexual development, aging, and malignant transformation. Regulation of AR levels may occur anywhere along the path from AR gene transcription to post-translational modification. A variety of factors, including androgens, have been implicated in modulating the AR protein and mRNA expression.
In the case of mouse fetal development, AR mRNA, based on in situ hybridization, was not found in the urogenital sinus at 13.5 days of gestation, whereas at 15.5 days of gestation both AR mRNA and protein levels were detectable (39).
Table1: The relative abundance of androgen receptor (AR) mRNA in rat tissues (38). Total RNA from the indicated tissues was subjected to competitive reverse transcription-polymerase chain reaction for quantitation of AR mRNA levels. The data are reported as percentage of AR mRNA relative to the prostate AR mRNA levels (averaged from 5 male rats). The data represent mean values from 3-5 rats. ND = not determined.
In the rat neonate, castration at 3 days of age did not result in altered AR expression in the rat prostate (40). In contrast, castration in the adult rat altered AR mRNA and protein levels (41-43). These findings suggest that one or more developmentally regulated factor(s) influence the AR expression.
An age-dependent decline in hepatic rAR expression has been shown to be associated with the expression of an Age-dependent factor (ADF) which is ubiquitously expressed in tissues (44). ADF binds to a rAR fragment between -310 to -330. Rat hepatic ADF binding activity on the rAR promoter in vitro was shown to decrease with age, When the ADF binding site was mutated in a reporter construct, a decreased rAR promoter activity was observed.
As discussed below, in prostate cancer, development of an androgen independent state is associated with a heterogeneous AR expression. In addition to prostate cancer, the AR protein or mRNA has been detected in other forms of cancer such as hepatomas (45), germ-cell neoplasias (46), and ovarian tumors (47). These findings suggest that dedifferentiation may result in expression of factors which modulate the AR expression.
Several factors have been reported to modify the AR expression. These factors include androgens which were shown to decrease the AR mRNA expression in the rat ventral prostate (42, 48), a human androgen-responsive prostate carcinoma cell line (LNCaP) (42, 49), and a hepatoma cell line (HepG2) (48). However, these observations are controversial because up-regulation of AR by androgens has been demonstrated in the rat and mouse prostate (43), genital skin fibroblasts (50), rat penile smooth muscle (51), and male rat fat-pad adipose precursor cells (52). Some of these conflicting observations may be due to different methodologies used such as moderately sensitive Northern blots versus highly sensitive in situ hybridization. However, the presence of tissue-specific transcription factors may also account for the opposite effects androgens play on AR expression.
The regulation of the AR also takes place at the protein level. For example, in the androgen-dependent prostate cancer cell line, PC-82, although the AR mRNA level was stable AR protein was up-regulated by androgen (53). On the other hand, it was reported that the down-regulation of AR mRNA by androgen in the LNCaP line was associated with an increased AR protein expression (42, 54, 55). Mizokami et al. showed that the AR protein up-regulation by androgen resulted from enhanced stability of AR protein (55) providing a clue for these discrepant observations.
In addition to androgen, several other hormones and growth factors can regulate the AR expression. Follicle-stimulating hormone increases the level of mRNA of AR in the Sertoli cells (56). Growth hormone, prolactin (57) and epidermal growth factor (55, 58) increase the AR mRNA levels in prostatic cells. Interestingly, trans-retinoic acid down-regulated AR mRNA in T47D breast cancer cells but up-regulated the expression of AR mRNA in MDA-MB-453 breast cancer cells (59) demonstrating that the effect of the signal depends on the cell type.
In conclusion, AR expression can be modified by a variety of factors which appear to act in a tissue and cell specific fashion. Close examination of the AR gene control elements may provide clues as to how this regulation is achieved.
Transcriptional control of the AR gene has been examined in a variety of conditions. The AR promoter lacks TATA and CCAAT boxes, but appears to have a GC box (a Sp1 binding site) which may play an important role in transcriptional initiation (60, 61). Additionally, it appears that there are two transcription initiation sites located 1127 base pairs (bp) and 1116 bp upstream of the translation initiation site (61, 62). A variety of putative cis-acting elements are present on the AR promoter (63). A cAMP response element (CRE) has been shown to stimulate AR transcription in both the human and murine AR (63-65). Though not reported for the human AR, the rAR promoter contains a NF-kB site which down-regulates transcription of the AR (66). In addition to the above studies, a cis-acting region between +57 and +575 of the 5'-untranslated region (5'-UTR) of the AR gene was identified which up-regulated an AR reporter gene's protein expression (62). However, this occurred without increasing the steady state mRNA levels suggesting that this region of the 5'-UTR functions by inducing translation (67).