Abstract 2234: Detection of minimal residual disease in the bone marrow of breast cancer patients using multiplex gene expression measurements with the Nanostring nCounter Assay

Background: Disseminated tumor cells (DTCs) are detected in up to 40% of breast cancer patients at the time of diagnosis and are an independent prognostic factor for recurrent disease. Presently DTCs are identified by immunohistochemical staining for cytokeratins and/or by molecular detection of single gene expression associated with DTCs. These diagnostic approaches are often subjective, laborious, and insensitive due to the molecular heterogeneity of the DTCs. In this study, we have determined whether a novel, multiplexed gene expression technology platform, Nanostring nCounterTM (NC), could be used for the multi-marker detection of DTCs. Methods: Total RNA was isolated from bone marrow (BM) of breast cancer patients, breast cancer cell lines, and healthy volunteers. RNA was analyzed directly with the NC assay or converted to cDNA and analyzed by qRT-PCR for DTC associated gene-expression. Results: Expression of 9 genes associated with breast cancer and DTC (TWIST1, PITX2, TACSTD1, SCGB2A2, EGFR, SNAI2, S100A3, KRT17 and KRT19) were examined by the NC assay and qRT-PCR. Using cell lines representing 3 main molecular breast tumor subtypes diluted into normal human BM at varying concentrations, we found that expression of KRT19, SNAI2, and EGFR could be simultaneously detected at a sensitivity of 1 cancer cell per 100,000 BM cells, using 0.5ug of input RNA. Expression measurements were quantitative, reproducible, and varied over a 20-fold linear dynamic range, depending on the cell line and its inherent expression of each transcript. Sensitivity was improved approximately 20-fold by Ficoll gradient enrichment of DTCs (a standard methodology for ICC-based detection) and increasing the amount of input RNA to 5 ug. This increased the sensitivity from 1 DTC per 1 million BM cells. The NC assay applied to BM collected prior to treatment from 5 patients with stage II/III breast cancer who had known DTCs detected by conventional ICC. 8 normal BM were used to establish baseline and threshold cutoff values. At least one of the nine genes in the multi-marker panel was detected in four of the five breast cancer patient BM samples. Interestingly, KRT19 itself was detected in only one specimen. The BM from two patients (ER-/Her2-) expressed genes known to be associated with this tumor type (TACSTD1, SNAI2, KRT17). Conclusion: Our data demonstrate the potentiality of the NC platform to detect multigene expression in rare cell populations like DTCs, with sensitivity equa