NEW YORK(GenomeWeb) – A team from the Foundation IRCCS National Cancer Institute in Milan has published a proof of concept of a method for capturing and measuring gene expression in circulating tumor cells, which they hope to use in the clinical care of advanced tumors.
In their approach, the researchers coupled magnetic bead-based capture of CTCs, using antibodies against both EpCAM and MUC1, with gene-expression analysis using Illumina's BeadChip-based Whole Genome-DASL HT assay.
Applying the array-based method to a series of blood samples spiked with different concentrations of cells from two breast cancer cell lines, the group found that in samples containing 25 or more cells, gene-expression results correlated well with directly analyzed cell-line RNA and distinguished cleanly from control blood samples.
The approach was also able to distinguish between the two different breast cancer cell lines, the authors reported in this month's Clinical Chemistry.
Maria Daidone, one of the study's authors, told GenomeWeb in an email, that the report serves as a technical proof of concept for the approach, which the institute researchers have since continued to validate and advance as a method for clinical evaluation and care of advanced cancer patients.
"We started with [this] approach since we were (and actually still are) not interested in counting CTCs, but in characterizing them. Now we are testing the feasibility of using filters … to isolate selected cells [for example, perhaps using Silicon Biosystems' DEPArray] approach and then processing them by genomic assays," Daidone wrote.
While the team's initial study was focused on the DASL array and demonstrated that this type of broad genomic profiling can be done on CTCs — in contrast to efforts using more limited PCR assays — Daidone said that in the future, she and her colleagues are also potentially interested in exploring beyond arrays into next-gen sequencing.
In their initial POC study, Daidone and her colleagues tested the gene-expression results they gleaned using their array analysis approach with those obtained from RNA isolated from either the raw cultured cell lines or from unspiked blood samples.
According to the study authors, the WG-DASL HT assay, designed for low-abundance RNA in FFPE samples, allows measurement of the expression of 29,000 genes. To evaluate the reliability of the assay in obtaining gene-expression profiles from both a low-RNA sample input and from samples with varying concentrations of CTCs, the researchers performed several spiking experiments, using breast cancer cells spiked into blood from a single healthy donor at concentrations of 400, 200, and 100 cells per 5 mL. This showed that, as expected, gene expression in samples of unspiked blood was well differentiated from that of cancer cell-spiked blood, the researchers reported.
The team then assessed the biological reliability of DASL assay-derived gene expression by comparing the profiles from captured cancer cells to those of isolated RNA from the two breast cancer cell lines, MDA-MB-468 and MCF7.
According to the authors, more than 90 percent of genes were shared between the isolated cancer cells and the cell line RNA. Moreover, when they compared the genes expressed in isolated cancer cells to those of control blood samples, the two sets were clearly differentiated.
In addition, the researchers tested the method to see if the captured blood-spiked cancer cells from each individual cell line would correlate in their gene expression to measurements made directly from the relevant cell line.
According to the authors, 73 percent of the MDA-MB-468-spiked samples and 64 percent of the MCF7-spiked samples correlated to the proper cell line.
The researchers also observed that at concentrations below 25 cells per 5 mL of blood, the rate of failure to obtain a reliable gene-expression profile began to exceed the success rate. In fact, the majority of the lowest-input samples showed gene-expression patterns more close to unspiked blood than to the relevant cancer cell line. The success rate for 25 cells was about 80 percent, the group reported.
Overall, Daidone and her coauthors wrote that the results support the "possibility of obtaining technically reliable gene-expression profiles from as few as 25 cells in 5ml [of] blood."
According to the group, the data also show that gene-expression profiles obtained using their method "convey biologically useful information that may allow differences among samples to be distinguished.
The authors acknowledged that other groups have also demonstrated gene expression measurement in isolated CTCs. But most, being PCR-based, have been limited to the analysis of far fewer genes than allowed by the DASL array.
According to the group, their method's requirement of 25 cells per 5 mL of blood may be high enough to somewhat limit the method's clinical potential.
However, the team is actively looking for ways to improve the sensitivity of DASL-based gene-expression profiling of CTCs either by using larger blood samples or improved CTC capture methods.
Considering a clinically significant CTC threshold to be 5 CTCs per 7.5 mL of blood, a 38 mL draw should be enough to guarantee the necessary 25 CTCs for gene expression analysis, the authors wrote.
The lab has also collected some preliminary data from a study of 30 breast cancer patients, which indicated that using an improved antibody cocktail for CTC enrichment also improves the ability to obtain reliable gene-expression measurements.
Daidone told GenomeWeb that the team also is hoping to apply their approach to characterizing CTCs in terms of both "transcriptome and DNA mutations, compared to the corresponding primary and/or metastatic lesions to investigate tumor heterogeneity and its implications on tumor biology and patient outcome."