The androgen receptor CAG repeat: a modifier of carcinogenesis?

The first exon of the human androgen receptor (AR) contains a translated CAG (poly-glutamine) repeat. The repeat length is polymorphic in the normal population ranging from 8 to 35 repeats. Expansions to over 40 repeats lead to spinal bulbar muscular atrophy (SBMA), a late onset neurodegenerative disease. The repeat is located between the two parts of a bipartite amino-terminal transactivation function and the repeat length, also within in the normal range, is inversely correlated to the transactivation power of the receptor. P160 type co-activators bind more strongly to shorter repeats. A correlation between AR CAG repeat length and total risk, age at diagnosis, recurrence after surgery and aggressive growth has been reported for tumors of classical androgen target tissues. In the prostate, where androgens exert a mitogenic effect, the cancer risk increases with decreasing AR-CAG repeat length. In contrast, in the breast, where the hormone probably acts as anti-mitogen, a higher risk and earlier onset of breast cancer has been reported for carriers of BRCA1 mutations who also have long CAG repeats in the receptor gene. Somatic alterations during carcinogenesis appear to be frequent in endometrial and in colon cancer. In the endometrium the AR CAG repeat prevalently undergoes expansions consistent with the putative protective function of androgens in this tissue. Frequent repeat reductions during colon carcinogenesis would be consistent with a mitogenic effect of androgens. Analysis of AR protein expression by Western blot reveals expression of the AR in healthy and neoplastic colon tissues. Normal mucosa of the colon expresses both AR-isoforms of 110 and 87 kDa, while the tumor samples have lost the expression of the 110-kDa isoform. The 87-kDa isoform is devoid of the amino-terminal portion of the receptor molecule that also contains the poly-glutamine tract. The temporal and causal relation between isoform switch and somatic repeat reductions during colon carcinogenesis is as yet unclear, but the two events could both enhance p160 mediated androgen signaling. The recent finding that smad3 interacts with the AR in a way similar to p160 links the AR to TGFbeta signaling. Interruption of this signaling pathway is a frequent event in colon carcinogenesis.