Prostate cancer is a leading cause of cancer-related death in males worldwide. Indeed, advanced and metastatic disease characterized by androgen resistance and often associated with neuroendocrine (NE) differentiation remains incurable. Using the spontaneous prostate cancer TRAMP model, we have shown that mast cells (MCs) support in vivo the growth of prostate adenocarcinoma, whereas their genetic or pharmacologic targeting favors prostate NE cancer arousal. Aiming at simultaneously targeting prostate NE tumor cells and MCs, both expressing the cKit tyrosine kinase receptor, we have tested the therapeutic effect of imatinib in TRAMP mice. Imatinib-treated TRAMP mice experience a partial benefit against prostate adenocarcinoma, because of inhibition of supportive MCs. However, they show an unexpected outgrowth of prostate NE tumors, likely because of defective signaling pathway downstream of cKit receptor. Also unexpected but very effective was the inhibition of epithelial-stromal tumors of the seminal vesicles achieved by imatinib treatment. These tumors normally arise in the seminal vesicles of TRAMP mice, independently of the degree of prostatic glandular lesions, and resemble phyllodes tumors found in human prostate and seminal vesicles, and in breast. In both mice and in patients, these tumors are negative for cKit but express PDGFR-β, another tyrosine kinase receptor specifically inhibited by imatinib. Our results imply a possible detrimental effect of imatinib in prostate cancer patients but suggest a promising therapeutic application of imatinib in the treatment of recurrent or metastatic phyllodes tumors.