METASTATIC PROSTATE CANCER, A JOURNEY OF METABOLIC REWIRING TRIGGERED BY BONE CELLS

ANSELMINO, NICOLÁS; LABANCA, ESTEFANIA; PASCUAL, GASTON; COTIGNOLA, JAVIER; SANCHIS, PABLO; SABATER, AGUSTINA; LAGE VICKERS, SOFIA; NAVONE, NORA; GUERON, GERALDINE; LAVIGNOLLE, ROSARIO; BIZZOTTO, JUAN; SENIUK, ROCIO; VAZQUEZ, ELBA

Congreso; Prostate Cancer Foundation Scientific Retreat 2021; 2021

BACKGROUND: No curative therapy is currently available for metastatic prostate cancer (PCa). Among metastatic PCa, the incidence of bone metastasis is 88%, and represents a severe clinical hurdle to overcome. As a consequence of the metastatic cascade, the arrival of tumor cells to the target homing organ requires a metabolic adaptation which in the case of PCa, could possibly be mediated by bone secreted factors. Here, we sought to identify key metabolic genes that fuel prostate cancer (PCa) bone metastasis.METHODS: By an indirect transwell co-culture system of PCa (PC3) and bone progenitor cells (MC3T3 or Raw264.7) we assessed the transcriptomic profile of PC3 cells modulated by soluble factors released from bone precursors. We performed a GSEA to delineate metabolic pathways dysregulated in co-cultured PC3. We then validated the clinical relevance of key metabolic genes in open access transcriptomic datasets. Finally, we profiled the secretome (ESI MS/MS) of bone cells to establish a communication axis driving the metabolic phenotype of tumoral cells. RESULTS: GSEA analysis showcased an activation of lipid metabolic categories, including PPAR and PI3K-Akt signaling pathways, and fat absorption and digestion, in PC3 cells co-cultured with both, osteoblastic MC3T3 cells or osteoclastic Raw264.7 cells. We then selected the altered metabolic genes for an unsupervised clustering analysis using transcriptomic data from human PCa and bone metastatic samples (GSE74685). Interestingly, those genes could accurately cluster samples by their tissue of origin in two defined groups: primary PCa and bone metastasis, highlighting that the transcriptional metabolic alterations triggered in our co-culture model could discriminate primary tumors from prostate cancer metastasis in the bone. After performing a survival analysis for those genes in the SU2C-PCF dataset, we observed that 4 lipid-associated genes, PPARA, VDR, SLC16A1 and GPX1, were associated with a shorter overall survival time, and that they could be independent risk-predictors of death (HR: 4.96, 2.85, 3.93 and 3.67, respectively; P