Abstract 646: Feasibility of assessing circulating tumor cells in patient-derived xenograft tumor models

Abstract

Abstract Background. Patient-derived xenograft (PDX) tumor models more accurately reflect human tumor biology than traditional cancer cell lines or their derived xenografts. Pre-clinical studies in cohorts of PDX lines emulate clinical trials, with each line representing a unique “patient”. Stemcentrx has developed more than 600 PDX lines from various cancer types to use in development of novel therapies directed against the cancer stem cells, which can initiate tumors and drive recurrent disease. In order to better understand the biology of metastasis - the primary cause of cancer mortality - a newly developed RareCyte method for detection of human circulating tumor cells (CTCs) in the blood of tumor bearing mice was independently validated. Methods. For platform validation studies, cells from 5 distinct PDX tumor models were counted and spiked into human blood at approximately 10, 50, 200 or 1000 per 7.5 mL. Blood was processed using AccuCyte cell separation and collection (RareCyte) and the nucleated cells spread onto 8 microscope slides. Slides were stained with anti-human cytokeratin, EpCAM and CD45 antibodies and DAPI (nuclear stain) using an automated instrument. A single slide from each spike-in tube was imaged by the CyteFinder digital fluorescence scanning microscope (RareCyte). Cells with morphology and phenotype (cytokeratin+/EpCAM+, CD45-) consistent with human cancer cells were tallied after review of images and compared to expected counts. For mouse PDX feasibility studies, 0.5-1 mL of blood from PDX tumor-bearing mice (∼150-2000 mm3) was collected by cardiac puncture. After red cell lysis and washing, nucleated blood cells were spread onto one slide. Cells were stained, imaged and identified as above substituting anti-mouse CD45. Results. CTCs were clearly identified by human epithelial cell marker staining in several different PDX lines. Moreover, there was a high correlation between the expected and experimental counts in the validation studies (R = 0.99) that was independent of spike-in number. Regression analysis gave a line of fit with slope 0.95, suggesting 95% recovery efficiency. In the PDX studies, CTCs were identified in 9 of 9 models. The number of CTCs was dependent on PDX line and tumor volume and ranged from 1 to 282 [per 800uL]. Clusters of 4 to 100+ CTCs were seen in blood samples from PDX tumor bearing mice. There was no correlation between number of CTCs and the duration of tumor in mice. However, there was a correlation between total number of CTCs and CTC clusters with the frequency and size of lung metastasis in mice. Conclusions. Spike-in studies demonstrated linear, high-recovery identification of CTCs from PDX tumor models, using a small volume mouse blood protocol. Captured CTC appear to be present in a high proportion of models across several indications. These results demonstrate the feasibility of evaluating CTCs as a parameter of drug response in pre-clinical studies. Citation Format: Somdutta Roy, Kevin Martinez, Arturo Ramirez, Daniel Campton, Joshua Nordberg, Eric Kaldjian, Scott J. Dylla, Holger Karsunky. Feasibility of assessing circulating tumor cells in patient-derived xenograft tumor models. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 646.