We describe an approach to generate patient-derived cancer cells from advanced prostate cancer and apply such cells to discover drugs that could be applied in clinical trials for castration-resistant prostate cancer.

Despite of the progress in the molecular etiology of prostate cancer, the androgen receptor (AR) remains the major druggable target for the advanced disease. In addition to hormonal ligands, AR activity is regulated by posttranslational modifications. Here, we show that androgen induces SUMO-2 and SUMO-3 (SUMO-2/3) modification (SUMOylation) of the endogenous AR in prostate cancer cells, which is also reflected in the chromatin-bound receptor. Although only a small percentage of AR is SUMOylated at the steady state, AR SUMOylation sites have an impact on the receptor's stability, intranuclear mobility, and chromatin interactions and on expression of its target genes. Interestingly, short-term proteotoxic and cell stress, such as hyperthermia, that detaches the AR from the chromatin triggers accumulation of the SUMO-2/3-modified AR pool which concentrates into the nuclear matrix compartment. Alleviation of the stress allows rapid reversal of the SUMO-2/3 modifications and the AR to return to the chromatin. In sum, these results suggest that the androgen-induced SUMOylation is linked to the activity cycles of the holo-AR in the nucleus and chromatin binding, whereas the stress-induced SUMO-2/3 modifications sustain the solubility of the AR and protect it from proteotoxic insults in the nucleus. C ovalent conjugation of proteins by small ubiquitin-related modifiers (SUMOs), SUMOylation, has emerged as a significant regulatory mechanism, especially in nuclear signaling, transport, transcription, and DNA replication/repair (5, 52). The modification pathway has also been implicated in human diseases, including cancer (1, 40). Humans express three ϳ100-aminoacid-long SUMO proteins, SUMO-1, -2, and -3, that can form isopeptide linkages with specific lysine residues of their target proteins. SUMO-2 and SUMO-3 (here called SUMO-2/3) are practically identical, whereas SUMO-1 is only ϳ50% identical with SUMO-2/3. SUMO-2 and -3 can form polymeric chains through an internal lysine residue, whereas SUMO-1 is not thought to be modified to form a polymeric chain, but when linked to the end of a poly-SUMO-2/3 chain, it may terminate the chain growth (47). Moreover, conjugation of SUMO-2/3, but not that of SUMO-1, has been reported to be altered in response to cell stress (39). Different SUMO paralogs may thus have (at least partially) distinct regulatory roles.The SUMOylation pathway requires E1, E2, and E3 activities that are distinct from those in ubiquitylation, and the two modifications have different molecular consequences. The SUMOs are activated by the SAE1 and -2 dimer (E1) and conjugated by UBC9 (E2). PIAS1, -2, -3, and -4 form a major family of SUMO E3 ligases (35). SUMO modifications are thought to be highly dynamic and have been shown to be reversed (deSUMOylated) by a family of SUMO-specific proteases (SENP1, -2, -3, -5, -6, and -7) (15).An increasing number of proteins, especially transcription factors, have been identified as putative SUMO targets (7, 48). However, most of the previous studies addressing SUMOylatio...

The identification of fluorescently stained cell nuclei is the basis of cell detection, segmentation, and feature extraction in high content microscopy experiments. The nuclear morphology of single cells is also one of the essential indicators of phenotypic variation. However, the cells used in experiments can lose their contact inhibition, and can therefore pile up on top of each other, making the detection of single cells extremely challenging using current segmentation methods. The model we present here can detect cell nuclei and their morphology even in high-confluency cell cultures with many overlapping cell nuclei. We combine the “gas of near circles” active contour model, which favors circular shapes but allows slight variations around them, with a new data model. This captures a common property of many microscopic imaging techniques: the intensities from superposed nuclei are additive, so that two overlapping nuclei, for example, have a total intensity that is approximately double the intensity of a single nucleus. We demonstrate the power of our method on microscopic images of cells, comparing the results with those obtained from a widely used approach, and with manual image segmentations by experts.