Lentiviral shRNA Screen To Test the Validity of a Gene Signature of Breast Cancer Stem Cells Using High Throughput Mammosphere Assays.

Abstract Background Conventional chemotherapies are effective initially in controlling growth of many tumors, yet, many patients relapse over time. A possible explanation for relapse is that a rare sub-population of cancer stem cells (CSCs) with tumorigenic potential is intrinsically resistant to conventional therapy. We have identified a gene signature which is derived from human breast cancer biopsies that is the overlap between CD44+/CD24-/low vs. all other cell subpopulations, and cancer-derived mammospheres (MS) vs. bulk tumor. Both the CD44+/CD24-/low subpopulation and mammospheres are typically enriched for tumor-initiating “cancer stem cells” and their division-competent descendents. We found that 477 genes were differentially expressed in the combination group of which 185 of these were highly expressed in CD44+/CD24-/low cells and in MS, a highly significant overlap (p<1.0E-9).The rest of the 292 genes demonstrated a reduction in expression in the CD44+/CD24-/low cells and cancer-derived MS vs. all other cells and bulk tumors, respectively (p<5.0E-5, one-sided Fisher's exact). We hypothesized that a subset of these genes may have functional significance in decreasing MS formation efficiency (MSFE).MethodsWe then tested this gene signature by screening the entire list of 477 genes for their ability to increase or decrease MS formation using lentiviral shRNA knockdown vectors (OpenBiosystems, Inc). We created a sub-library comprised of fourteen 96-well plates containing approximately ∼1200 unique shRNA constructs, with an average of 3-4 shRNA constructs per gene, thereby targeting ∼500 unique genes.The library was designed and constructed on a “one gene, one well” structure, such that each sequence validated lentiviral shRNA construct targeting an individual gene is located in a unique well of a 96-well plate. Next, we transduced SUM159 cells with each of the lentiviruses individually, along with a positive control lentivirus targeting Bmi1, a polycomb group gene known to be required for self-renewal of many stem cell types, a non-specific negative control lentivirus, as well as untreated cells. Each transduced cell population was plated into individual wells of a 96-well low-attachment plate and allowed to form MS over 4 days. MS formation was quantified using Gelcount (Oxford Optornix, Inc). ResultsEfficacy of the lentivirus transfection was ∼80%. Of the 477 genes, we identified 151 shRNAs that demonstrated significant change (p≤0.05) in MS formation with respect to median MSFE and controls. ConclusionsWe have developed a high throughput lentiviral shRNA screening method to test the validity of CSC targets using MS assays. These data suggest intricate interactions between MS-initiating cells and downstream progenitors in MS formation and provides a set of ideal candidate genes. Inhibitors to these targets may decrease MS self renewal, as well as eliminate cancer stem cells, and may be important to test as pharmaceutical interventions. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 1158.