Mutational Analysis of Circulating Tumor Cells Using a Novel Microfluidic Collection Device and qPCR Assay

Abstract

Circulating tumor cells (CTCs) provide a readily accessible source of tumor material from patients with cancer. Molecular profiling of these rare cells can lead to insight on disease progression and therapeutic strategies. A critical need exists to isolate CTCs with sufficient quantity and sample integrity to adapt to conventional analytical techniques. We present a microfluidic platform (IsoFlux) that uses flow control and immunomagnetic capture to enhance CTC isolation. A novel cell retrieval mechanism ensures complete transfer of CTCs into the molecular assay. Improved sensitivity to the capture antigen was demonstrated by spike-in experiments for three cell lines of varying levels of antigen expression. We obtained spike-in recovery rates of 74%, 75%, and 85% for MDA-MB-231 (low), PC3 (middle), and SKBR3 (high) cell lines. Recovery using matched enumeration protocols and matched samples (PC3) yielded 90% and 40% recovery for the IsoFlux and CellSearch systems, respectively. In matched prostate cancer samples (N = 22), patients presenting more than four CTCs per blood draw were 95% and 36% using IsoFlux and CellSearch, respectively. An assay for detecting KRAS mutations was described along with data from patients with colorectal cancer, of which 87% presented CTCs above the assay's limit of detection (four CTCs). The CTC KRAS mutant rate was 50%, with 46% of patients displaying a CTC KRAS mutational status that differed from the previously acquired tissue biopsy data. The microfluidic system and mutation assay presented here provide a complete workflow to track oncogene mutational changes longitudinally with high success rates.

Figure 1

Microfluidic design and control instrumentation. (A) The microfluidic device for CTC isolation contains a sample inlet reservoir that feeds into a microfluidic channel. (B) A schematic depiction of the flow path consisting of two inputs (a sample and a wash buffer well), isolation region, and waste well. (C) The sample passes through an isolation zone where magnetically labeled cells are captured on the top of the channel in the presence of a magnet. After processing, the cells are retrieved inside the instrument by lifting the cap off the channel and securing it to either microfuge tube or holder to remove the target cells through pipetting (not shown). (D) Up to four microfluidic cartridges can be processed in parallel inside the instrument, which uses pneumatic-driven flow to process the samples.

Figure 2

CTC capture efficiency and linearity using tumor cell lines. Model CTC samples were prepared by spiking PC3 (A) and MDA-MB-231 (B) cells into tubes of healthy donor blood in concentrations ranging from 20 to 300 total target cells per tube. Average CTC capture efficiency was 73% and 81% with an overall SD of 16% and 27%, respectively. Standard errors are shown for each spiked concentration. (B) CTC capture levels varied linearly with spike-in concentration (C and D). R² values were 0.9 and 0.5 with slopes of 0.75 (75% of cells captured) and 0.74.

Figure 3

Recovery dependence on EpCAM expression. A comparison of recovery rates for three different tumor cell lines is shown. Mean results using the IsoFlux System are plotted with standard error bars (solid squares) for three cell lines: MDA-MB-231 (n = 7), PC3 (n = 38), and SKBR3 (n = 8). Solid triangles represent matched sample results (n = 4 samples each) that were run on either CellSearch (blue) or IsoFlux (red) using exactly the same cell counting protocols. Open circles represent literature data for Cell-Search spike-in recovery from two of the cell lines studied: MDA-MB-231 [10] and SKBR3 [9].

Figure 4

CTC recovery from matched clinical samples. Matched samples from 22 patients with prostate cancer were used to compare CTC recovery using the IsoFlux System compared to the CellSearch platform that is also based on immunomagnetic separation. For concomitantly drawn 7.5 ml of blood samples, counts are presented using the two platforms; in both cases, immunostaining was used to identify CTCs as cells that are CK+, CD45-, and DAPI+ (nucleated).

Figure 5

KRAS mutation assay LODs using purified gDNA. Varying levels of G13D mutant gDNA were mixed with high wild-type background (10,000 wild-type cells) and analyzed with a KRAS G13D mutant assay and KRAS reference assay. The C_t difference between KRAS mutant and wild-type assays (dC_t) was calculated. KRAS mutation is detected when the dC_t is below the cutoff dC_t value generated with pure wild-type gDNA (green line). Samples containing four cell equivalents of mutant gDNA in a background equivalent of 10,000 wild-type cells were consistently detected.

Figure 6

Assay LODs using model CTC samples. Varying levels of KRAS G13D mutation-positive tumor cells (MDA-MB-231) were spiked into healthy donor blood and processed on the IsoFlux System. gDNA was amplified, purified, and tested for KRAS mutations. The C_t difference between KRAS mutant and wild-type assays (dC_t) was calculated. Samples having as few as nine recovered tumor cells were detectable using the qPCR assay. Healthy control samples remained above the mutation detection cutoff.

Figure 7

Representative images of CTCs and background WBCs from patients with CRC. CTCs were enumerated using immunofluorescence staining for CK (fluorescein isothiocyanate, green), CD45 (Cy3, red), and nucleus (Hoescht 33342, blue). A CTC is defined as CK+, CD45-, DAPI+, and morphologically intact.

Figure 8

Summary of colorectal cancer test group.