Circulating Tumour Cells Indicate the Presence of Residual Disease Post-Castration in Prostate Cancer Patient-Derived Xenograft Models.

Castrate-resistant prostate cancer (CRPC) is the lethal form of prostate cancer. Epithelial mesenchymal plasticity (EMP) has been associated with disease progression to CRPC, and prostate cancer therapies targeting the androgen signalling axis, including androgen deprivation therapy (ADT), promote EMP. We explored effects of castration on EMP in the tumours and circulating tumour cells (CTCs) of patient-derived xenograft (PDX)-bearing castrated mice using human-specific RT-qPCR assays and immunocytochemistry. Expression of prostate epithelial cell marker KLK3 was below detection in most tumours from castrated mice (62%, 23/37 mice), consistent with its known up-regulation by androgens. Endpoint tumour size after castration varied significantly in a PDX model-specific pattern; while most tumours were castration-sensitive (BM18, LuCaP70), the majority of LuCaP105 tumours continued to grow following castration. By contrast, LuCaP96 PDX showed a mixed response to castration. CTCs were detected in 33% of LuCaP105, 43% of BM18, 47% of LuCaP70, and 54% of LuCaP96 castrated mice using mRNA measurement in plasma. When present, CTC numbers estimated using human expression ranged from 1 to 458 CTCs per ml blood, similar to our previous observations in non-castrated mice. In contrast to their non-castrated counterparts, there was no relationship between tumour size and CTC burden in castrated mice. Unsupervised hierarchical clustering of the gene expression profiles of CTCs collected from castrated and non-castrated mice revealed distinct CTC sub-groups within the pooled population that were classified as having mesenchymal, epithelial, or EMP hybrid gene expression profiles. The epithelial signature was only found in CTCs from non-castrated mice. Hybrid and mesenchymal signatures were detected in CTCs from both castrated and non-castrated mice, with an emphasis towards mesenchymal phenotypes in castrated mice. Post-castration serum PSA levels were either below detection or very low for all the CTC positive samples highlighting the potential usefulness of CTCs for disease monitoring after androgen ablation therapy. In summary, our study of castration effects on prostate cancer PDX CTCs showed that CTCs were often detected in the castrate setting, even in mice with no palpable tumours, and demonstrated the superior ability of CTCs to reveal residual disease over the conventional clinical biomarker serum PSA.