Androgen deprivation-mediated activation of AKT is enhanced in prostate cancer with TMPRSS2:ERG fusion [RNA-Seq]

The TMPRSS2:ERG gene fusion (T:E fusion) in prostate adenocarcinoma (PCa) puts ERG under the androgen receptor (AR) regulated expression of TMPRSS2. The T:E fusion is frequently associated with PTEN loss, and is highly correlated with decreased expression of INPP4B, both of which may compensate for ERG-mediated suppression of PI3K/AKT signaling. We confirmed in PCa cells and a mouse PCa model that ERG suppresses AKT activation, and that one potential mechanism is through downregulation of IRS2. In contrast, ERG knockdown did not increase INPP4B, suggesting its decreased expression is indirect and reflects selective pressure to suppress INPP4B function. Notably, INPP4B expression is similarly decreased in PTEN-intact and PTEN-deficient T:E fusion tumors, suggesting selection for a function distinct from regulation of PI3K activity. As ERG expression in T:E fusion tumors is AR regulated, we further assessed the extent to which AR inhibition increased AKT activity in T:E fusion tumors. T:E fusion positive versus negative PDXs had greater increases in AKT activity after castration. Moreover, in a neoadjuvant trial of AR inhibition prior to radical prostatectomy we similarly found greater increases in AKT activation in the T:E fusion tumors. Together these findings indicate that AKT activation may mitigate the efficacy of AR targeted therapy in T:E fusion PCa, and that these patients may most benefit from combination therapy targeting AR and AKT. Overall design: A model with doxycycline (DOX) regulated expression of ERG in mouse prostate, driven by a tetracycline operon regulated ERG (pTET-ERG) combined with a probasin promoter regulated reverse tetracycline transactivator (Pb-rTta), was generated to assess the functions of ERG in driving prostate adenocarcinoma (PRAD) in vivo. When crossed onto a Pten-/+ background and treated with doxycycline for ~1 year, these mice developed PRAD. To identify ERG-regulated genes that may be contributing to PRAD development in this model, we carried out time course experiments with both short-term ERG induction and long term induction plus short term DOX withdrawal. For the short-term induction, previously untreated mice were treated with DOX or vehicle for 3 or 6 days. Ventral prostate epithelium was then laser capture microdissected and transcriptome profiles were obtained using Affymetrix Mouse Gene 1.0 ST arrays. Another cohort of mice with DOX regulated ERG on the Pten-/+ background were treated with DOX for ~1 year to generate PRAD. DOX was then discontinued in one set of mice and prostates were harvested after 1-2 weeks and analyzed by RNA-seq using the Illumina TruSeq stranded mRNA kit. To identify ERG regulated genes in VCaP cells (expressing the TMPRSS2:ERG fusion), cells were treated with ERG or nontargeted siRNA and assessed by RNA-seq.