Phenotypic Heterogeneity and Instability of Human Ovarian Tumor-Initiating Cells
Session type: Symposia
The cancer stem cell (CSC) model proposes that only some tumor cells indefinitely self-renew, thereby initiating and sustaining tumor growth, and predicts that these tumor-initiating cells (TIC) can be prospectively isolated based on their phenotype. Previous studies have suggested the existence of CSC in serous ovarian cancer (SOC), but these relied on arguably non-fidelitous immortalized cell lines, cultured primary cells or passaged xenografts as the source of tumor cells. We developed a robust assay for TIC from primary SOC. Using this assay, we find that TIC are rare in both NOD/SCID and NOD/SCID/IL2Rg-/- (NSG) mice. The TIC frequency varies substantially between patients, although it is similar in primary ovarian masses and omental metastases, suggesting that TIC frequency is an intrinsic property of ovarian tumors. Of several putative CSC markers tested, only CD133 is expressed in all cases examined and consistently enriches for TIC. However, not all TIC from primary tumors are CD133+, and tumors arising from CD133- cells cannot be explained by contamination with CD133+ cells. The ability of CD133 to enrich for TIC deteriorates upon xenograft passage, such that in many cases, the majority of xenograft-derived TIC are CD133-. Nevertheless, preliminary microarray analyses comparing CD133-positive and negative cells from primary tumors show significant differences in gene expression, with Gene Set Enrichment Analysis demonstrating enrichment of stem cell-related genes in the CD133+ population. Our results show that CD133 expression enriches for TIC in SOC, and suggest that these cells have distinct biological properties. However, our findings also highlight the need for quantitative rigor in the evaluation of TIC and indicate that the heterogeneity of the TIC phenotype may complicate clinical application of the CSC concept.