Metastasis is the predominant cause of mortality in many cancers, including prostate and breast cancers. With a substantial number of men and women being diagnosed with these cancers each year in the United States, there is a need for effective methods to rapidly identify key regulators of the metastatic cascade that can subsequently be employed in diagnostic, prognostic, and therapeutic applications. This has become increasingly important as we come to understand that many metastatic events occur earlier in disease progression than originally thought. Functional genomic screens have arisen as an effective means to rapidly screen for critical genes controlling a specified biology. Loss of function screens, such as RNAi libraries, are particularly useful for the identification of suppressor genes. We utilized genome-wide RNAi screens to identify negative regulators of metastasis, defined as metastasis suppressor genes, with in vivo and in vitro breast and prostate cancer models. We performed multiple screens assessing the individual contributions of genes in Matrigel invasion (T-47D), initial colonization of lungs by prostate cancer cells (LNCaP) in a mouse model, and complete metastasis from orthotopic tumor to distant metastases in vivo (LNCaP).

In the LNCaP screen for enrichment of lung colonizers, we identified a number of gene targets with high confidence due to targeting by multiple shRNA clones in a majority of replicates. These hits were filtered to genes that were downregulated in metastases versus primary tumors across 4 datasets available at Oncomine.org or that exhibited a high percentage of heterozygous or homozygous deletion in The Cancer Genome Atlas (TCGA) provisional prostate cancer dataset. Many of these gene targets were found to be critical nodes regulating adhesion, survival, and apoptosis pathways and included death receptors (TNFRSF10B), map kinases (MAP3K1 and MAPK12), NFkappaB proteins (MALT1), and prion protein ( PRNP).

As a result of the stringent selection regimens, a smaller subset of target genes were enriched in the breast cancer screen for Matrigel invasiveness and the in vivo prostate cancer screen for enhanced metastasis. A shared shRNA clone was strongly enriched in these two distinct screens. The shRNA clone (sh/siAPOL2*) targets apolipoprotein L-II (APOL2) and increases Matrigel invasion and migration in an array of breast and prostate cancer cell lines; however, targeting APOL2 for knockdown with validated, alternative siRNA sequences did not enhance LNCaP chemotaxis. These findings suggest siAPOL2* has off-target, or microRNA-like, effects. Furthermore, gene expression profiling in LNCaP cells attributed 132 differentially downregulated transcripts to the siAPOL2* sequence. Sequence analysis of targeted transcripts and the siAPOL2* sequence revealed a conserved targeting region in the 5' end of siAPOL2* coinciding with nucleotide positions of the seed region known to confer specificity to microRNAs. A comprehensive search of miRNA databases identified 4 human miRNAs (miR-548u, miR-8087, miR-4432, and miR-543) with similar if not identical seed sequences to siAPOL2*, and 3 of these miRNAs recapitulated the enhanced chemotactic phenotype of siAPOL2* in LNCaP cells when introduced as mimics.

MiR-543 is predicted to target G protein-coupled receptor, family C, group 6, member A (GPRC6A) which was identified as a validated off-target of siAPOL2* and found as an enriched shRNA clone target in the LNCaP RNAi screen for metastasis suppressors in vivo. Knockdown of GPRC6A increased LNCaP transwell migration and Matrigel invasion, but consistently diminished migration of androgen-insensitive cell lines, PC3 and DU-145, with no effect on the proliferation of these cells. Transfection with miR-543 mimic phenocopies the pattern of cell-type specific migration effects observed in siGPRC6A treated cells. Interestingly, q-RT-PCR found miR-543 increases GPRC6A transcript either by upregulation or post-transcriptional stabilization; however, luciferase reporter assays incorporating the GPRC6A 3'UTR suggest miR-543 binds the GPRC6A 3'UTR and inhibits protein translation. Further investigation is underway to characterize the mechanism directing cell-specific control of chemotaxis upon GPRC6A silencing with early results indicating that variations in GPRC6A isoform dependency and preferred G protein coupling are contributing factors. This study demonstrates a novel role for GPRC6A in controlling context dependent cell migration and identifies a functionally relevant miRNA capable of repressing GPRC6A.