We thank Setlur et al. for their comments and agree with them that despite essential differences between their study and ours there are striking overlaps in the conclusions. Both studies identify a link between the gene expression signatures of prostate cancers containing the TMPRSS2:ERG fusions and β-oestradiol signalling. Also, Setlur et al. completed in vitro functional studies showing a role for both of the ERs in regulating the expression of TMPRSS2:ERG. Each study concluded that a better understanding of the role of the β-oestradiol signalling pathway in prostate cancer development could have important clinical implications. This idea is interesting because in animal models the synergy between oestrogens and androgens in prostate cancer development is well established, and in clinical trials treatment with the ERα agonist toremifene is known to delay cancer formation (reviewed in [1]).

Although the signatures defining TMPRSS2:ERG cancers defined by Setlur et al. (87 genes) and Jhavar et al. (142 genes) shared only 14 genes in common, both studies also highlighted the potential importance of Ca2+- and K+-ion channel overexpression in defining cancers harbouring the TMPRSS2:ERG fusion. In all, our study identified six interlinked networks defining TMPRSS2:ERG cancers, but differed from the work of Setlur et al. in that it also implicated pathways modulated by retinoic acid and by progesterone. Retinoic acid is known to be involved in the development of the normal prostate [2], and progesterone has a proposed role in the development of prostate cancer [1]. Taken together, our two studies identify many new avenues of investigations that might assist in therapeutically targeting cancer with ERG-gene alterations.