The generation of the TMPRSS2:ERG recombinant bacterial artificial chromosome (BAC) using recombineering has been completed. Transgenic animals were generated in the FVB and C57/BL6 backgrounds, and one line from each background, A5 and H7, respectively, was selected for further investigation. RT-PCR analysis of prostate-derived RNA utilizing primer pairs that span TMPRSS2 and ERG exons showed specific expression of the fusion transcripts in transgenic animals. Sequencing revealed appropriately spliced transcripts, and transient transfection produced the expected-sized translation products. Transcripts with and without the 72 bp exon 11 were observed, similar to endogenous transcripts described in human cell lines and clinical samples. Western blot analysis of protein lysates from prostate epithelial organoids using monoclonal antibody 2805 (Epitomics, Burlingame, CA) directed to the COOH-terminus of ERG revealed a protein migrating slightly slower and therefore distinct from endogenous ERG. ERG fusion protein expression was 2-4 fold greater than endogenous ERG. The regulation of the TMPRSS2-ERG fusion by androgen in primary prostate epithelial cells has not been investigated. To establish the regulatory potential of the integrated TMPRSS2-ERG BAC, we analyzed the androgen receptor (AR) binding sites that previously have been identified for the TMPRSS2 promoter in LnCaP. Using ChIP analysis performed on total prostates, we confirmed that AR bound to sites in the transgenic human TMPRSS2 promoter. Importantly, these data suggest that appropriate chromatin organization leading to AR association was established in the BAC transgenic prostates. Regulation of the fusion transcript by androgen was analyzed in vivo and in vitro, and no induction of fusion RNA or protein was observed despite induction of the AR-regulated FKBP5 RNA. In addition, castration of transgenic animals confirmed partial androgen dependence for the fusion construct and endogenous mouse Tmprss2. The TMPRSS2 promoter has been shown to be androgen inducible in human cell lines derived from advanced prostate cancers, and it has been commonly assumed that androgen regulation of TMPRSS2-ERG is an important feature of transformation. Our data challenge this concept and suggest that TMPRSS2 regulation be investigated in additional cellular and genetic contexts to define the full spectrum of TMPRSS2 control by androgen. To establish the pattern of ERG expression in the prostate, tissue sections from wild type and transgenic fusion mice were stained with monoclonal antibodies which recognize mouse and human ERG. ERG antibody staining verified epithelial-specific increased expression of the fusion construct and demonstrated expression in the majority of luminal cells and in a fraction of basal cells. Interestingly, it has been shown that the majority of prostate progenitor activity fractionates with basal cells expressing high Sca-1, and we asked whether the expression of full-length ERG in this fraction was associated with changes in progenitor activity. Sphere-forming activity was assayed in prostate cells fractionated using Lin, Sca-1, and EpCam markers, and virtually all of the sphere-forming units (SFU) from wt and fusion transgenics were found in the Sca-1+/EpCam+ fraction. A comparison of sphere-forming activity in dissociated whole prostates revealed on average a 2-3 fold increase in the relative number of spheres in fusion mice preparations, showing the existence of an increased pool of primary prostate cells capable of self-renewal. Furthermore, the ability to form spheres upon serial passage in vitro, a measure of intrinsic self-renewal, was 2-3 fold higher in fusion compared to wt as measured by sphere or colony formation. TMPRSS2-ERG translocations often occur in combination with decreased PTEN expression in human prostate cancer. The effect of Pten loss upon TMPRSS2-driven ERG expression has not been evaluated directly. Histologically, fusion A5 mice containing a heterozygous deletion of Pten showed at 28 weeks increased total numbers and severity of mPIN lesions compared to Pten+/-. Further studies using models such as the one described here in conjunction with clinical samples will aid in defining the mechanism and prognostic value of mutations occurring in conjunction with TMPRSS2-ERG translocations. We have used a technique developed by Tan Ince to grow luminal cell lines from primary prostate samples of TMPRSS2-ERG transgenic mice, some in combination with other relevant alterations. These cell lines have allowed the direct analysis of ERG-dependent properties, which we have shown include invasion and clonogenic growth propagation. We have been able to show using TMPRSS2-ERG and wild type lines that mutant Ras synergizes with ERG to produce adenocarcinoma, which does not occur with mutant Ras or transgenic ERG alone. We have observed that ERG expressing lines display reduced senescence in response to Ras induction. One question concerns the pathways that are effected by ERG expression to counteract a senescence response. TP53 and RB mediated Ras induced senescence have been well-described. We have determined that ERG action is independent of TP53. Because ERG translocations occur early in prostate cancer tumorigenesis, these data suggest that one role for ERG may be the inhibition of senescence that accompanies oncogene activation. We have developed a variety of genetically-modified cell lines and high content imaging protocols as well as genetically engineered mouse models to analyze the context of ERG dependent inhibition of senescence. Further mechanistic analyses are on-going.
| Casey, Orla M; Fang, Lei; Hynes, Paul G et al. (2012) TMPRSS2- driven ERG expression in vivo increases self-renewal and maintains expression in a castration resistant subpopulation. PLoS One 7:e41668 |
