Abstract
Background: Circulating tumor cells (CTCs) likely derive from clones in the primary tumor, suggesting that they can be used for all biological tests applying to the primary cells. Materials and Methods: The ScreenCell® devices are single-use and low-cost innovative devices that use a filter to isolate and sort tumor cells by size. Results: The ScreenCell® Cyto device is able to isolate rare, fixed, tumor cells, with a high recovery rate. Cells are well preserved morphologically. Immunocytochemistry and FISH assays can be performed directly on the filter. The ScreenCell® CC device allows isolation of live cells able to grow in culture. High quality genetic materials can be obtained directly from tumor cells isolated on the ScreenCell® MB device filter. Conclusion: Due to their reduced size, versatility, and capacity to isolate CTCs within minutes, the ScreenCell® devices may be able to simplify and improve non-invasive access to tumor cells.
- Circulating tumor cells
- CTCs
- microfiltration by size
- real-time quantitative reverse transcriptase-polymerase chain reaction
- qRT-PCR
- molecular biology
- cytomorphology
- FISH
- cell culture
- lung cancer
Circulating tumor cells (CTCs) can be used as a surrogate for primary tumor cells when obtaining material non-invasively for the monitoring of tumor phenotypes. Indeed, CTCs are likely derived from clones in the primary tumor (1), suggesting that they can be used for all biological tests applying to the primary cells they represent. Several methods have been developed to detect and isolate CTCs in the peripheral blood of patients with cancer (2-13). Such methods rely mainly on cytometric/immunological characteristics (14, 15), although CTCs may also be isolated by size (16). Any useful method for isolation of CTCs must allow: (i) their identification and enumeration and (ii) their characterization through immunocytochemistry, fluorescence in situ hybridization (FISH) assays and all relevant molecular techniques using optimal quality DNA/RNA. However, it is as yet unknown to what extent different methods compare in terms of sensitivity, specificity and reproducibility.
Quantification of CTCs can be performed through the use of magnetic bead-conjugated antibodies against epithelial-cell adhesion molecule (EpCAM). However, its use in affecting treatment decisions remains a point of discussion (17-19). Indeed, isolation of an adequate number of CTCs in a reproducible manner and their use for molecular studies has been limited due to their extreme rarity (around 1 per 109 cells in peripheral blood of patients with metastatic cancer) and methodological limitations (20). It is also important to note that the concept of EpCAM-dependent assays is based upon the assumption that the presence of epithelial cells in peripheral blood indicates the presence of tumor cells. However, epithelial cells may be found in healthy donors (21), with their apparent prevalence varying according to: (i) the detection procedure used (9, 22); (ii) the ‘cut-off’ of reference (9), and (iii) morphological assessment of circulating cells by different investigators, which is a major source of error (23). Furthermore, EpCAM-based assays are not able to detect normal-like breast tumor cells (24). This, combined with the fact that certain tumor types such as melanoma are not of epithelial origin, suggests that EpCAM-based assays may be of limited use. Moreover, a precise cytological analysis of isolated cells is not possible due to technological limitations, therefore limiting the capacity to assess the true tumoral nature of individual cells and microclusters. Indeed, through different operators, a number of CK+ DAPI+ cells were not classified as CTCs (9). Finally, some data suggest that epithelial antigen may be lost on CTCs due to the epithelial–mesenchymal transition (EMT), which is considered to be a crucial event in the metastatic process (24-26).
Altogether, these results strongly suggest that morphological analysis including nucleo-cytoplasmic ratio, nuclear details and size of nucleoli, must be correlated with enumeration and immunocytochemical methods for identifying the tumoral origin of circulating cells. Furthermore, CTCs must be isolated alive for testing their potential capacity to initiate tumor formation in animal models and must become easily accessible to a large range of molecular biological analysis.
This study describes a newly developed filtration mini-device, the ScreenCell®, which can isolate, quantify, and analyze circulating tumor cells from a blood sample. In the ScreenCell® device, blood flow passes through a microporous membrane filter allowing size-selective isolation of rare tumor cells under fully reproducible and standardized conditions. The ScreenCell® device has been designed as a low-cost innovative technology with the aim of achieving isolation of tumor cells without the requirement for large and expensive apparatus. This device is fully accessible so that tumor cells isolated onto the filter can be analyzed using all relevant cellular and molecular biological techniques pertinent to the identification and characterization of CTCs and their potential genetic abnormalities. The device can isolate living cells, allowing further tissue culture experiments. Furthermore, tumor cells can be isolated without using an antibody-based assay, suggesting that the ScreenCell® device can be used for the isolation of a large spectrum of tumor cells, including cells of non-epithelial origin.
Materials and Methods
Cell culture and cell spiking. The NCI-H2030 and -H1975 cell lines (derived from non-small cell lung cancer), as well as the HT29 cell line (derived from colorectal adenocarcinoma), were cultured according to the suppliers instructions (CRL-5914, CRL-5908, and HTB-38; American Type Culture Collection (ATCC), Manassas, VA, USA). Following culture, cells were harvested using 0.05% trypsin. Cell suspensions were only used when their viability exceeded 90% as assessed by trypan blue exclusion. For accuracy, linearity, and sensitivity experiments, the spiked cell numbers were estimated to be two and five in 1 ml of peripheral blood from healthy donors. Blood samples from patients with melanoma and colorectal carcinoma were also filtered and analyzed for cytomorphology and immunocytochemistry.
Devices and buffers. The ScreenCell® filtration devices were developed in order to isolate CTCs by size on a microporous membrane filter. These devices are 19 cm long and designed for isolation of: (i) fixed cells for cytological studies (ScreenCell® Cyto); (ii) live cells for culture (ScreenCell® CC) and (iii) molecular biology (ScreenCell® MB). The filtration devices comprise a filtration tank, a filter capped by a removable nozzle/holder (Figure 1Aa and Ba) which, after removal of a protective membrane (Figure 1Ab and Bb), allows insertion and guidance of a collection tube (Figures 1Ac and Bc). The circular track-etched filter, which is composed of polycarbonate (it4ip, Belgium) material, is 18 μm thick with a smooth flat and hydrophilic surface. Circular pores are calibrated (7.5±0.36 μm or 6.5±0.33 μm for isolation of fixed or live cells, respectively) and randomly distributed throughout the filter (1×105 pores/cm2). Before filtration and in order to lyse red blood cells (RBCs), 1 ml blood samples must be diluted in 7 or 8 ml of a specific dilution buffer for fixed or live cells, respectively. Following filtration of fixed cells and for better cytological studies, an additional 1 ml of PBS is filtered for removing RBC debris from the filter. Filtration is usually completed within approximately 50 s (a sample must be considered as micro-coagulated when filtration exceeds 60 s). At the end of filtration, the nozzle/holder of the ScreenCell® device is unclipped and removed from the filtration tank (Figures 1A and 1B).
ScreenCell® Cyto and CC devices (Figure 1A) are devoted to cytological studies and cell culture, respectively. The filter allows a fast and regular filtration, preserving the CTC morphology and microcluster structures. Blood samples are diluted with the ScreenCell® FC or ScreenCell® LC/CC dilution buffers for fixed or live cells, respectively. At the end of filtration, the ScreenCell® CC filter is released into a well of a 24-well tissue culture plate, by pushing down a rod located at the bottom part of the filtration device (Figure 1Ae). Adequate tissue culture medium and growth factors are added into the well. The multiwell plate is then closed and incubated under defined conditions. Similarly, the ScreenCell® Cyto filter is released onto a standard microscopy glass slide by pushing down a rod located at the bottom part of the filtration device; a 7 mm circular coverslip (Cat# CB00070RA1; Menzel-Gläser, Braunschweig, Germany) can be laid down on the filter (Figure 1Af) with the appropriate mounting medium. Cytological studies including staining, cell enumeration, immunocytochemistry, FISH assays, can then be conducted directly on the filter.
For both the ScreenCell® Cyto and CC device filters, the filtration area is delimited by an O ring made of surgical inox with a bar code to insure traceability of the filtered samples.
ScreenCell® MB device is nuclease-free and devoted to molecular biological studies before or after the cell culture (Figure 1B). CTCs are isolated on a circular filter clipped at the bottom part of a capsule (Figure 1Ba); a bar code ensures traceability of the filtered sample.
Before filtration, the blood sample is diluted in the ScreenCell® LC dilution buffer. At the end of filtration, the capsule-filter is then ejected and either inserted inside the upper inner part of an nuclease-free Eppendorf® tube (Figure 1Be) or into a well of a 24-well tissue culture plate (Figure 1Bf).The ScreenCell® MB filtration unit allows extraction of either DNA or RNA directly from cells isolated on the filter before or after cell culture (in this case, the capsule-filter containing the cultured cells is transferred with nuclease-free forceps into a nuclease-free Eppendorf® tube) (Figure 1Be and f). An adequate volume of lysis buffer is added into the capsule-filter which is then closed with the Eppendorf® tube cap (Figure 1Be). Following incubation at the appropriate temperature, the Eppendorf® tube containing the capsule-filter is centrifuged for 1 min at 12,000 ×g (Figure 1Be), and the capsule-filter removed and discarded. The flow-through is either stored in the closed Eppendorf® tube or used immediately to conduct further molecular biological procedures.
Sensitivity tests. Two and five cultured NCI-H2030 cells were collected by micropippeting under a microscope and spiked in 1 ml of peripheral blood from a healthy donor. The whole blood was collected in EDTA and filtrated within three hours of collection. Blood spiked with the tumor cells was mixed with a dilution buffer for fixed cells and transferred into the ScreenCell® Cyto device for filtration. Cells isolated on the filter of the device were stained with hematoxylin and eosin. The filters were then mounted on a glass slide with Faramount mounting Medium (S3025, Dakocytomation, Glostrup, Denmark) and a circular glass coverslip. Cell enumeration was performed using a NIKON eclipse 80i fluorescence microscope integrated with cooled CCD camera system and NIS-Elements BR2.30 imaging software (NIKON, France).
qRT-PCR analysis. Lysate obtained from cells isolated on the capsule-filter was recovered into an Eppendorf tube by a one minute centrifugation at 12,000 ×g (10,500 rpm). Total RNA was then extracted and purified using the RNeasy FFPE kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions with minor modifications. Purified RNA was eluted with 14 μl nuclease-free water (Ambion, NY, USA). RNA (10 μl) was reverse transcribed for 60 min at 37°C followed by 5 min of RT inactivation at 95°C using 2.5 μl of 20X RT Enzyme Mix and 25 μl of 2× RT Buffer in 50μl of TaqMan® PreAmp Cells-to-Ct™ Kit (Ambion) according to the manufacturer's instructions. Sequences of interest in the cDNA were then pre-amplified through 14 cycles using 12.5 μl of 1:200 diluted pool of primers and probes (20× Taqman Gene Expression Assays EGFR -Hs00193306_m1- and TaqMan Genotyping Assays; Applied Biosystems, Foster City, USA) and 25 μl of TaqMan PreAmp Master Mix (Ambion) in a final volume of 50 μl. Cycling conditions were as follows: 10 min of denaturing at 95°C and 14 cycles of 95°C for 15 s and annealing at 60°C for 4 min.
For EGFR gene expression assay, 10 μl of the 1:10 diluted pre-amplification product was amplified at a final volume of 40 μl by qRT-PCR with 20 μl of 2× Taqman Gene expression Master Mix, 2 μl of 20× Taqman Gene Expression Assays EGFR primers and probes and 8 μl of 2.5 μg/μl BSA (Ambion). The qRT-PCR reactions were run using the ABI PRISM 7300 apparatus (Applied Biosystems). The cycling conditions were as follow: 2 min of UDG incubation at 50°C, 10 min of enzyme activation at 95°C and 40 cycles of 95°C for 15 s and annealing at 60°C for 1 min.
For detection of the EGFR exon 19 deletion, 1 ml of whole peripheral blood spiked with PC9 cells (20 and 100 cells), which harbor an exon 19 deletion (E746-A750), was lysed with 7 ml of ScreenCell® MB dilution buffer at room temperature for 2 min and then passed through ScreenCell® MB filters. Genomic DNA was extracted according to manufacturer's suggestion using QIAamp DNA Micro Kit (Qiagen Cat# 56304), and then amplified in a PCR reaction for exon 19 of EGFR (29). The PCR primers for amplification of the EGFR exon 19 were designed to hybridize to intron sequences flanking the exon. The sequences are as follows: EGFR-EX19F 5’-GTGGCACCATCTCACAATTGCC-3’, EGFR-EX19R 5’-GGGCC TGAGGTTCAGAGCCAT-3’. The amplicon generated was 203 bp. The PCR reactions were as follows: 200 μM dNTP mix, 300 nM forward primer, 300 nM reverse primer, 1 μl genomic DNA, 0.25 μl JumpStart Taq (Sigma Cat#D4184), 1× JumpStart Buffer, 2.5 mM MgCl2, and water up to 25 μl. Touchdown PCR conditions were as follows: 95°C for 5 min, followed by 14 cycles of 95°C for 20 s, 69°C-62°C for 20 s, and 72°C for 40 s, where temperature was reduced by 0.5°C in each successive round of hybridization. Then 30 more cycles of the following conditions: 95°C for 20 s, 62°C for 20 s and 72°C for 40 s, followed by a final extension of 72°C for 5 min. Samples were then denatured and slowly renatured in order to form heteroduplexes using the following conditions: 95°C for 2 min followed by a step decrease in temperature of 0.5°C for 15 s per step, until the temperature reached 45°C. The PCR products were analyzed on an Agilent 2100 BioAnalyzer. Genomic DNA extracted from cell lines A549 and PC9 was included as wild-type and deletion mutation-positive cells respectively. Two independent filtrations and PCR assays were performed.
For EGFR 21 L858R exon point mutation, detection was assessed by allelic discrimination using two sets of primers-probes designed to hybridize to the mutated and the wild-type sequence as follows: EGFR-EX21F 5’-GCAGCATGTCAAGATCACAGATTT-3’, EGFR-EX21R 5’-CCTCCTTCTGCATGGTATTCTTTCT-3’, EGFR-EX21 VIC 5’–CAGTTTGGCCAGCCCA-3’, and EGFR-EX21 FAM 5’-CAGTTTGGCCCGCCCA-3’. The targets of interest (L858R mutation and wild-type EGFR) in the cDNA were pre-amplified for 14 cycles. Amplification was performed on an ABI PRISM 7300 apparatus (Applied Biosystems) in a 25 μl volume composed of 5 μl of a 1:20 dilution of the pre-amplified product, 12.5 μl of 2xTaqMan® GTXpress™ Master Mix (Applied Biosystems), 0.625 μl of 40x Assay Mix, 5.6 μl of 2.5 μg/μl BSA (Ambion), and 1.275 μl nuclease-free water. The reaction was performed with a preliminary step for 20 s at 95°C, followed by 60 cycles of two steps at 95°C for 1 s for denaturing, and 60°C for 27 s for annealing. The results were analyzed using the allelic discrimination assay program as provided by the constructor.
Immunocytochemistry. Prior to the immunocytochemical analysis, cells isolated on the circular filter of the ScreenCell® Cyto device were dried overnight at room temperature and then hydrated with tris-buffered saline (TBS). When needed, the antigens were retrieved with target retrieval solution at 95-99°C for 20 min and rinsed with TBS. For detection of intracellular proteins, isolated cells were treated 5 min at room temperature with a permeabilizing buffer, and incubated 30 min at room temperature with a peroxidase-blocking solution. Filters were incubated with a monoclonal antibody against human cytokeratins (clone KL1; Abcam, Cambridge, UK or clone AE1/AE3; Dakocytomation, Glostrup, Denmark), or against human CD45 (clone 2B11 + PD7/26) (Dakocytomation). A standard HRP-peroxidase complex method (EnVision + Dual link system-HRP, DAB+, kit; Dakocytomation) was used for immunocytochemical detection directly on the ScreenCell® Cyto filters which were then incubated with 3,3’-diaminobenzidine chromogen solution (Dakocytomation). After a final wash with distilled water, the filter was counterstained with hematoxylin for 5 min at room temperature. The filter was then dried at room temperature and mounted on a glass slide with Faramount mounting medium (Dakocytomation) and a glass circular coverslip.
Fluorescence in situ hybridization. FISH was applied directly on the filter with the EGFR/CEN-7 FISH Probe Mix and Histology FISH Accessory kit (Dako Denmark, Germany), according to the manufacturer's protocol with minor modifications. In brief, the filter was hydrated with the kit washing buffer and 250 μl of cold (2-8°C) pepsin were added for 15 min at 37°C on each area of the filter inside the O ring. The filter was soaked twice for 2 min at room temperature in the kit washing buffer. It was then dehydrated through a graded ethanol series (70%, 90% and 100%) for 2 min at room temperature and dried at room temperature. Following a 5 min denaturation step at 82 (±2)°C, the EGFR and chromosome 7 centromere locus were directly labeled overnight at 45°C with 10 μl EGFR/CEN-7 PNA probe mix in a Hybridizer (Dako Denmark). The filter was washed for 10 min at 65°C with the stringent washing buffer and twice for 3 min at room temperature with the washing buffer. Dehydration was then conducted through an ethanol series and the filter was dried at room temperature and mounted on a slide with fluorescence mounting medium containing DAPI and a glass coverslip. The FISH analysis was performed using a NIKON eclipse 80i fluorescence microscope integrated with a cooled CCD camera system and NIS-Elements BR 2.30 imaging software (NIKON).
Statistics. P-values of unpaired unilateral/bilateral Student's t-test and Fisher's exact test were calculated to assess the sensitivity of the ScreenCell® Cyto device. It is usually considered that a p-value above 0.05 implies that there is no significant difference between compared groups.
Results
Sensitivity of the ScreenCell® Cyto device for CTC detection. Twenty-five independent experiments were conducted with two and five fixed H2030 cells (Tables IA to ID). These cells were micropipetted after being released from the tissue culture flask following standard trypsinisation. After fixation cells were spiked into whole peripheral blood drawn from a healthy donor, and filtered through the ScreenCell® Cyto device. The average filtration time was 50 s. The expected numbers of H2030 cells spiked into the blood sample plotted against the actual number of H2030 cells observed in the sample are shown in Tables IA and IB. The average percentage of H2030 cells recovered was 91.2% and 74% with five and two spiked cells respectively. In the samples spiked with five cells, no fewer than three cells were detected in all 25 samples. For two and five spiked H2030 cells, the average numbers of cells recovered were 1.48 (standard deviation: 0.71) and 4.56 (standard deviation: 0.71), respectively.
To verify whether the percentage of cell loss was related to the filtration device, two and five H2030 cells were harvested as indicated above. Once fixed, cells were micropipetted and collected directly into an Eppendorf tube. Under these conditions, the mean percentage of recovery was 82% and 88% for two and five cells, respectively (Tables IC and ID). For two and five cells, the average numbers of cells recovered were 1.64 (standard deviation 0.57) and 4.40 (standard deviation 0.71), respectively. The relative sensitivity of the ScreenCell® Cyto device versus direct cell collection was assessed through P-values calculated for unpaired unilateral Student's t-test (0.19 and 0.20 for two and five spiked cells respectively), unpaired bilateral Student's t-test (0.39 and 0.41 for two and five cells, respectively) and Fisher's exact test (0.14 and 0.34 for two and five cells, respectively). These tests showed that collection of two and five spiked tumor cells through the ScreenCell® Cyto device or by direct collection of the micropipetted cells directly into an Eppendorf tube resulted in similar sensitivities. Through the different series of tests using the ScreenCell® Cyto device and direct collection, similar numbers of cells were lost after 25 independent collections of two and five spiked tumor cells. Indeed, the percentage of cells lost through the ScreenCell® Cyto device was 26% (standard deviation 0.71, with an average of 0.52 cells lost), and 9% (standard deviation 0.65, with an average of 0.44 cells lost) for two and five spiked H2030 cells, respectively, while it was 18% (standard deviation 0.57, with an average of 0.36 cells lost) and 12% (standard deviation 0.71, with an average of 0.60 cells lost) through direct collection for two and five cells, respectively. The P-value for unpaired unilateral Student's t-test indicated similar numbers of lost cells using the ScreenCell® Cyto device or by direct collection with two or five tumor cells.
No significant differences were found when using the P-value for unpaired unilateral Student's t-test to compare the results obtained with two versus five spiked tumor cells through the ScreenCell® Cyto device or by direct collection (0.19 versus 0.20 for two and five spiked tumor cells, respectively). Furthermore, no significant differences were found when using the P-value for unpaired unilateral Student's t-test to compare the results obtained with two versus five spiked tumor cells through the ScreenCell® Cyto device (0.34) or by direct collection (0.10). Altogether, these results strongly suggest that cells were lost essentially through the micropipetting procedure and that the recovery rate of the ScreenCell® Cyto device was close to 100%.
Viability and culture of live H2030 cells following filtration through a ScreenCell® CC device. Five independent experiments were conducted to assess the viability of tumor cells after filtration through the ScreenCell® CC device. In each experiment, live H2030 cells were trypsinised and 50 cells were filtrated through the device. Viable cells were counted immediately after filtration using a standard trypan blue exclusion test; the mean was 0.85±0.09.
The capacity of isolated H2030 cells to grow in tissue culture was further tested through eight independent experiments. In each case, isolated cells were able to grow and expand directly onto the filter under adequate tissue culture conditions (Figures 2A and B).
Morphological and immunocytochemical analysis of fixed tumor cells after filtration through a ScreenCell® Cyto device. Cultured H2030 cells isolated onto ScreenCell® filters after spiking into blood exhibited a well-preserved morphology, with intact nuclear and cytoplasmic contents and membranes (Figure 3A). Isolated H2030 cells which were immunostained with an anti-pankeratin (KL1) antibody directly onto the filter on which they were isolated exhibited a dense brown granular cytoplasmic staining, while lymphocytes or monocytes were negative (Figures 3A and 3B). CD45-positive cells were also identified on the filter (~50 CD45-positive cells, with a large variation among donors) together with CD45-negative spiked tumor cells, with a minimal presence of RBCs (Figure 3C). Circulating tumor cells from patients with solid tumors were also analyzed morphologically. Figure 4 shows CTCs from patients with melanoma and colorectal carcinoma. Cell clusters with anisocaryosis and high nucleo-cytoplasmic ratio and hyperbasophilic nucleus were observed. After immunolabeling, a fine and granular cytoplasmic expression of the AE1/AE3 cytokeratins was detected on CTCs from the patient with colorectal cancer.
Detection of the epidermal growth factor receptor (EGFR) gene by FISH assay in isolated HT29 cells. Diluted HT29 cells were filtrated through the ScreenCell® Cyto device and a FISH assay was conducted directly on the filter using an EGFR/CEN-7 FISH probe mix (EGFR with a red fluorophore specific to the Red Texas spectrum and CEN7 with a green fluorophore specific to the FITC spectrum). After hybridization, the filters were counterstained with DAPI and the EGFR and CEN7 copy numbers were evaluated by FISH using fluorescence microscopy at ×60 magnification (Figure 5). EGFR gene amplification was attested to by the detection in the nuclei of more than two EGFR signals, as well as a high level of incidence of polysomy 7.
Detection of EGFR expression in isolated H2030 cells. Quantitative real-time RT-PCR was used for the detection of gene copy number changes of EGFR. Two, five and ten H2030 cells were spiked into whole peripheral blood from a healthy donor and filtered through the ScreenCell® MB device. Five independent filtrations were conducted. Isolated tumor cells were lysed directly on the capsule filter and mRNA was extracted. Reverse transcription was performed followed by a pre-amplification step, and PCR. The number of PCR cycles (Ct) required for the 6-carboxyfluorescein (FAM) intensities to exceed a threshold just above background was calculated using 7300 Real-Time PCR System software (Applied Biosystems). EGFR expression was detected in all samples of two, five and ten cells (Figure 6A). No significant differences were observed when two, five and ten H2030 cells were micropipetted and directly collected (without filtration) into an Eppendorf tube, suggesting that most spiked tumor cells were recovered through the ScreenCell® MB device (Figure 6Aa vs. B). For ten filtered H2030 cells, EGFR was detected after 26.5 cycles, while 30.8 cycles were necessary to amplify only two cells.
For material accumulation, it was possible to conduct a second filtration through the same ScreenCell® MB filter. Figure 7 shows EGFR detection for a single ten-cell filtration versus two five-cell filtrations through the same filter. Similar Cts were observed.
Reproducibility of RNA extraction and amplification was verified using five H2030 cells spiked into whole blood from a healthy donor and isolated through the ScreenCell® MB device (Figure 8).
Detection of EGFR exon 19 deletion and exon 21 L858R EGFR mutation in cells spiked into healthy donor blood. Full length PCR product (204 bp) was amplified from the DNA from A549 cells. A mixture of full-length product and deletion product (188 bp) was amplified from the DNA of PC9 cells, as well as DNA extracted from PC9-spiked blood samples that went through the filters (Figure 9A). Sequencing analysis confirmed that the ScreenCell® MB filters successfully captured PC9 cells which carry the EGFR exon 19 deletion mutation (Figure 9B). EGFR exon 21 L858R EGFR point mutation was detected in two, five, ten and thirty H1975 cells containing the heterozygous mutation, while, as expected, wild-type EGFR was detected in H2030 cells (Figure 9C).
Discussion
As a proof of principle, this study describes a new filtration device that allows the isolation of CTCs. While very easy to use and not requiring any large equipment, the ScreenCell® device allows CTC isolation with high sensitivity and specificity, with the capability of identifying tumor cells and assessing their specific genotypes. It is not limited to the isolation of CTCs of epithelial origin (28) and the device allows isolation of live cells which may be used for further tissue culture experiments and potentially for testing their capacity to grow in animal models.
A number of devices have been described for isolating CTCs. They use either direct methods, relying essentially on combining negative and positive immunological selection of non-epithelial and epithelial cells, respectively, or indirect methods, mainly the detection of epithelial-specific mRNA transcripts by RT-PCR. The development of the CellSearch™ system (Veridex, Warren, NJ, USA), has been instrumental in opening the field of CTCs to clinical trials. This semi-automated device, based on isolation of potential CTCs by immunomagnetic beads coated with antibodies against EpCAM, aims at identifying cytokeratin-positive cells with positive nuclear staining and CD45 negativity (10, 18, 29). The CellSearch system has been approved by the U.S. Food and Drug Administration as an aid in the monitoring of metastatic breast, colorectal cancer, and metastatic prostate cancer.
This immunological approach to detecting CTCs has given rise to various devices such as the commercially available AdnaTest BreastCancer™ device (AdnaGen AG, Langenhagen, Germany), which allows the isolation of potential CTCs by immunomagnetic beads coated with anti-MUC1 and -EpCAM antibodies. After isolation of the mRNA, transcripts of epithelial-specific markers such as GA 73.3, EpCAM and human epidermal growth factor receptor 2 (HER2) were amplified by a multiplex PCR (31-33).
A relevant marker must be useful to the monitoring of treatment and particularly of targeted therapy. Therefore, a main issue is to obtain cells and genetic material under the best conditions in order to assess whether the monitoring of tumor genotypes can be accomplished using CTCs as a surrogate for primary tumor cells. To this extent, Maheswaran et al. (34) have used a microfluidic device, a CTC chip, mediated by the interaction between CTCs and microposts coated with antibody against EpCAM under controlled laminar-flow conditions (35). As a proof of principle, these authors used such a device to isolate viable CTCs from 27 patients with metastatic non-small cell lung cancer (34) and to identify EGFR-activating mutations in 92% of patients with advanced stage disease. Furthermore, the T790M mutation, which confers drug resistance, was found in 55% of patients. That the authors claimed a sensitivity of 92%, together with the fact that in some patients lower CTC numbers correlated with a radiographic response to gefitinib, while increased CTC numbers correlated with an onset of additional EGFR mutations and clinical progression, suggested that a survey of the genotypic CTC status is relevant to monitoring tumor cells to assess whether the treatment is likely to reach its target efficiently.
Although this type of approach may allow real-time tumor genotyping, Maheswaran et al. indicated that optimization and automation of their device for high-throughput processing will be required to allow large-scale clinical trials. In a recent development of the CTC chip device, Stott et al. (36) described a high-throughput microfluidic mixing device, the herringbone-chip, or ‘HB-Chip’, which provides an enhanced platform for CTC isolation, allowing detection of microclusters of CTCs, previously unappreciated tumor cell aggregates which may contribute to the hematogenous dissemination of cancer.
Using DNA sequencing of tumor tissue as well as a predefined gene expression arrays to trace the presence of CTCs in peripheral blood of patients with metastatic colorectal cancer treated with cetuximab plus FOLFOX or FOLFIRI, Yen et al. (37) observed a strong correlation between the detection of KRAS mutations between circulating tumor cells and the primary malignant tissue. Furthermore, the majority of responders to the cetuximab-containing regimen had wild-type KRAS and experienced superior progression-free and overall survival than patients with mutant KRAS. This method suggests that CTCs harbor KRAS mutations representative of the primary tumor status. However, this methodology does not per se allow CTC isolation, and the necessity for a predefined expression array to trace the presence of CTCs appears of limited interest for monitoring a targeted therapy, since detection of a KRAS mutation indicates likely resistance to cetuximab (38).
Whether the technique described by Nagrath et al. (35) and Yen et al. (37) can be adapted to allow consistent detection and isolation of viable CTCs for molecular analysis of pathways other than that of EGFR and KRAS in different tumor types remains to be determined.
Although recent meeting reports disclosed concordance rates of 68% and 88% for data obtained with the CellSearch™ system and with the AdnaTest BreastCancer™ kit, it is difficult (39, 40) to know to what extent these methods may differ in terms of sensitivity, specificity and reproducibility.
Due to its reduced size, large versatility, and capacity to isolate and characterize large circulating cells within minutes, the ScreenCell® device is likely to simplify and amplify clinical access to CTCs considerably. This study conducted spiking experiments using tumor cell lines for assessing the sensitivity, specificity and reproducibility of the system in isolating and characterizing CTCs. The ScreenCell® filtration device is a technology able to capture isolated tumor cells and clusters from patient samples. Further investigations are needed to analyze whether such clusters may provide insight into the metastatic process. Moreover, the fact that the ScreenCell® CC device allows isolation of live cells may open the possibility of testing the presence of tumor-initiating cells among CTCs using highly permissive xenotransplantation conditions. Finally, the use of flat circular filters allows analysis of full microscopy laser scan images of CTCs (data not shown).
Through different devices, the clinical validation of using CTCs as surrogates of primary tumor cells for detecting mutations relevant to the monitoring of targeted therapies is urgently required. Indeed, once this critical point is clarified, a direct comparison between CTC enrichment techniques will be required which will be best conducted through prospective clinical trials.
Acknowledgements
The Authors thank Professor M.F. Avril, Head, Department of Dermatology, Hôpital Cochin-Tarnier, Paris and Dr. P Langlois, Clinique Geoffroy Saint-Hilaire, Paris for providing patients' blood samples, and Stanislas Lechevalier for statistical calculations.
Footnotes
-
↵* These Authors contributed equally to this work.
-
Author Contributions
YEC conceived and designed the study. NBF, BG, ID, YK, MY, LW, and JAB, contributed to designing protocols and performed experiments. JW contributed to cytological studies. PAJ and RJD contributed to designing the EGFR exon 19 deletion experiment. NBF, BG, ID, YEC, and RJD contributed to data analysis. YEC wrote the paper with contributions by NBF, BG, ID, JW, and RJD.
-
This article is freely accessible online.
- Received December 28, 2010.
- Revision received January 30, 2011.
- Accepted January 30, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved