Discrete microfluidics for the isolation of circulating tumor cell subpopulations targeting fibroblast activation protein alpha and epithelial cell adhesion molecule

Abstract

Circulating tumor cells consist of phenotypically distinct subpopulations that originate from the tumor microenvironment. We report a circulating tumor cell dual selection assay that uses discrete microfluidics to select circulating tumor cell subpopulations from a single blood sample; circulating tumor cells expressing the established marker epithelial cell adhesion molecule and a new marker, fibroblast activation protein alpha, were evaluated. Both circulating tumor cell subpopulations were detected in metastatic ovarian, colorectal, prostate, breast, and pancreatic cancer patients and 90% of the isolated circulating tumor cells did not co-express both antigens. Clinical sensitivities of 100% showed substantial improvement compared to epithelial cell adhesion molecule selection alone. Owing to high purity (>80%) of the selected circulating tumor cells, molecular analysis of both circulating tumor cell subpopulations was carried out in bulk, including next generation sequencing, mutation analysis, and gene expression. Results suggested fibroblast activation protein alpha and epithelial cell adhesion molecule circulating tumor cells are distinct subpopulations and the use of these in concert can provide information needed to navigate through cancer disease management challenges.

Fig. 1

Sinusoidal microfluidic device used in the study and summary of clinical results. a Schematic of the dual selection strategy using mAbs directed against FAPα and EpCAM cell-surface antigens. b SEM of the CTC selection microfluidic device. c Optical micrographs of the CTC selection microchip filled with whole blood, and the chip after rinsing with buffer. d An image (5×) of DAPI-stained Hs578T cells isolated within the channels of the microfluidic device. e Simulation of CTC recovery from blood at different translational velocities as a function of cell rolling distance along the mAb decorated surface. f Box plots for CTCs isolated from the blood of healthy donors, patients with non-cancerous disease, CRPC, M- PDAC, M-CRC, M-BC, and M-EOC. CTC counts were normalized to 1 ml of blood. g Test positivity in cancer patients’ blood using the single EpCAM approach and the dual selection strategy (test positivity based on the CTC^FAPα and/or CTC^EpCAM counts exceeding a level that was 3× SD for counts from non-cancer patients)

Fig. 2

Phenotyping analysis in fluorescence microscopy. Images (40×) of CTC^FAPα and CTC^EpCAM isolated using the sinusoidal microfluidic chips and stained with a panel of markers: DAPI, anti-pan-CK-TR, anti-CD45-FITC, anti-VIM-FITC, anti-EpCAM-Cy5, and anti-FAPα-Cy5

Fig. 3

CTC phenotyping. a Fluorescence micrographs of cells isolated from a patient diagnosed with pancreatitis, and CTC^FAPα and CTC^EpCAM isolated from L/M-PDAC patients. All cells stained negative for CD45. b Immunophenotyping results of CTC^FAPα and CTC^EpCAM. The pie charts show the percent of CTCs with pan-CK and/or VIM expression for L-PDAC patient #66, M-PDAC patient #25, and M-BC patient #5

Fig. 4

Longitudinal tracking of CTC^FAPα and CTC^EpCAM numbers in the blood of PDAC patients. a M-PDAC patient #25. The first CTC analysis was performed during second-line therapy (t = 0). b L-PDAC patient #45. The first CTC analysis in this case was performed pre-operatively on the day of surgery (t = 0). CA19-9 measurements (green stars) are shown when available. CTC^FAPα = red dots, and CTC^EpCAM = blue squares. Points are connected for ease of visualization, but do not represent any type of functional relationship between the individual data points. c A summary of all patients tested in this longitudinal study

Fig. 5

KRAS mutation detection. a Schematic of the polymerase chain reaction/ligase detection reaction (PCR/LDR) assay. b Electropherograms of LDR products for: No gDNA; HT29 wt35 (50 nt); LS180 G35A (44 nt); M-PDAC CTC^FAPα G35A (44 nt), CTC^EpCAM G34C (61 nt); and CTC^FAPα G35T (55 nt). The gray trace shows the DNA markers. The fluorescence intensity values are arbitrary. c Table summarizing PCR/LDR results for HT29 and LS180 cell lines, M-CRC, L-CRC, M-PDAC, and L-PDAC CTCs. d RT-qPCR gene expression profiles for L-PDAC patient #66 and M-PDAC patient #67

Fig. 6

Phenotype, genotype, and CTCs from basal-like breast cancer PDX models. a IHC (400×) of tumor tissue in paraffin sections stained for FAPα, EpCAM, VIM, and pan-CK (scale bar = 20 µm). b Fluorescence microscope images of CTC^FAPα and CTC^EpCAM isolated from the blood via cardiac puncture (scale bar = 15 µm). c CTCs isolated from two PDX models and a healthy NSG control. d Sanger sequencing traces for amplicons generated from exon 6 TP53 DNA isolated from tumor tissue, CTC^FAPα, and CTC^EpCAM with primers designed for human sequence