The detection of CTCs prior to and during therapy is an independent and strong prognostic marker, and it is predictive of poor treatment outcome. A major challenge is that different technologies are available for isolation and characterization of CTCs in peripheral blood (PB). We compare the CellSearch system and AdnaTest BreastCancer Select/Detect, to evaluate the extent that these assays differ in their ability to detect CTCs in the PB of MBC patients. CTCs in 7.5 ml of PB were isolated and enumerated using the CellSearch, before new treatment. Two cutoff values of ≥2 and ≥5 CTCs/7.5 ml were used. AdnaTest requires 5 ml of PB to detect gene transcripts of tumor markers (GA733-2, MUC-1, and HER2) by RT-PCR. AdnaTest was scored positive if ≥1 of the transcript PCR products for the 3 markers were detected at a concentration ≥0.15 ng/μl. A total of 55 MBC patients were enrolled. 26 (47%) patients were positive for CTCs by the CellSearch (≥2 cutoff), while 20 (36%) were positive (≥5 cutoff). AdnaTest was positive in 29 (53%) with the individual markers being positive in 18% (GA733-2), 44% (MUC-1), and 35% (HER2). Overall positive agreement was 73% for CTC≥2 and 69% for CTC≥5. These preliminary data suggest that the AdnaTest has equivalent sensitivity to that of the CellSearch system in detecting 2 or more CTCs. While there is concordance between these 2 methods, the AdnaTest complements the CellSearch system by improving the overall CTC detection rate and permitting the assessment of genomic markers in CTCs.
Metastatic breast cancer (MBC) is an incurable disease that is treated with palliative intent. The clinical significance of circulating tumor cells (CTCs) for the prognosis and prediction of survival outcomes in patients with MBC has been well documented.1–7 The detection of CTCs prior to and during therapy is an independent and strong prognostic marker, and it is predictive of poor treatment outcome. The prognostic implications of detecting CTCs prior to or during therapy for breast cancer raise important questions about the biological nature of these cells. A comprehensive analysis of CTCs may provide new insights into the biology of breast cancer that could have a potential impact on the clinical management of breast cancer patients.
Advances in technology have produced various methods of detecting CTCs in the blood of cancer patients. These technologies vary with regard to sensitivity, specificity, reproducibility, for isolation and characterization of CTCs in peripheral blood (PB). Current techniques used to detect CTCs in cancer patients rely on a cell-enrichment step followed by a detection step. Most enrichment steps utilize antibody-based magnetic capture with magnetically bound antibodies directed against the epithelial cell adhesion molecule-1 (EpCAM).8–14 CTCs can be detected after immunomagnetic capture using either direct, antibody-based methods such as immunocytochemistry (ICC), immunofluorescence (IF), and flow cytometry (FACS), or by indirect, nucleic acid-based methods which measure mRNA transcripts by reverse transcriptase-polymerase chain reaction (RT-PCR).8–14
The CellSearch system (Veridex, LLC, Warren, NJ, USA), which uses immunomagnetic separation to isolate tumor cells with subsequent visualization by microscopy and counting of cells that express cytokeratin by IF, has been cleared by the U.S. Food and Drug Administration for use in patient care.15, 16
The novel PCR-based assay developed by AdnaGen AG (Langenhagen, Germany) is performed on standard technology platforms and makes use of RT-PCR to identify putative transcripts of genes in EpCAM-positive cells that are isolated by a magnetic separation method.14, 17, 18 Because of the combination of different selection and tumor markers, both the heterogeneity of the tumor cells and possible individual or therapy-induced deviations in the expression patterns are taken into account.
The primary purpose of this study was to determine if the AdnaTest BreastCancer Select/Detect correlated with the CellSearch system at the defined cutoff values of 2 and 5 CTCs per 7.5 ml of blood. We also examined the individual gene transcript markers used in the AdnaTest to determine if HER2 gene amplification of the primary tumor correlated with HER2 protein expression in CTCs.
Patients and Methods
We conducted a prospective trial at The University of Texas MD Anderson Cancer Center to evaluate the CellSearch system and the AdnaTest for CTC detection. The CellSearch System, is an automated, standardized system for the immunocytochemical detection and quantification of CTC in blood and has been regulatory-approved. The AdnaTest Breast Cancer Select/Detect essentially involves the detection of tumor-associated transcripts by RT-PCR after an immunomagnetically enrichment of tumor cells. All patients with progressive, measurable MBC about to start a new treatment were included. All patients had Eastern Cooperative Oncology Group scores performance status of 0 to 2. Prior adjuvant treatment, treatment of metastatic disease, or both were permitted. The Institutional Review Board at MD Anderson approved the study protocol, and all patients provided written informed consent.
Isolation and Enumeration of CTCs by CellSearch
CTC isolation and enumeration were performed using the CellSearch system as previously described.1, 15, 16 Briefly, blood samples were drawn into 10-ml CellSave Vacutainer tubes (Becton Dickinson), which contained EDTA and a cell fixative. Samples were maintained at room temperature and processed within 72 h after collection. All evaluations were performed without knowledge of the clinical status of the patients.
The CellSearch system consists of a semiautomated system (CellPrep) for the preparation of the sample, and it is used with the CellSearch Epithelial Cell Kit. The CellPrep system enriches the sample for cells expressing EpCAM with antibody-coated ferrous particles, and it labels the cell nucleus with the fluorescent nucleic acid dye 4,2-diamidino-2-phenylindole dihydrochloride. Fluorescently labeled monoclonal antibodies specific for leukocytes (CD45-allophycocyan) and cytokeratins (CK 8, 18, 19-phycoerythrin) are used to distinguish epithelial cells from leukocytes. The identification and enumeration of CTCs were performed using the CellSpotter Analyzer, a semiautomated fluorescence-based microscopy system that permits computer-generated reconstruction of cellular images. CTCs were defined as nucleated cells that lacked CD45 and that expressed cytokeratin. Technical details of the CellSearch and CellSpotter systems, including accuracy, precision, linearity, and reproducibility, have been described elsewhere.19
Based on previous studies, we used 2 cutoff values to determine CTC positivity in samples: 2 or more CTCs per 7.5 ml of blood20 and 5 or more CTCs per 7.5 ml of blood.1 Healthy subjects and patients with nonmalignant diseases have less than 2 CTCs per 7.5 ml of blood. The median progression-free survival among patients reached a plateau at approximately 5 CTCs per 7.5 ml of blood and this value was chosen to distinguish patients with an unfavorable prognosis from patients with a favorable prognosis.
Detection of CTCs by AdnaTest BreastCancer Select/Detect
The AdnaTest can be performed on standard technology platforms. AdnaGen's two-step “Combination-of-Combinations Principle” is illustrated in Figure 1. It initially involves cell isolation, whereby tumor cells are enriched by an antibody-mix linked to magnetic particles and mRNA is isolated from the selected tumor cells, and subsequent molecular biological detection and analysis, whereby the isolated mRNA is transcribed into cDNA and a multiplex PCR is carried out for the analysis of tumor-associated gene expression.14, 17, 18
The AdnaTest has a reported analytical sensitivity of 2 CTCs per 5 ml of blood21
For the AdnaTest, blood samples (5 ml) were collected in EDTA tubes (AdnaCollect, AdnaGen), then immunomagnetic separation was performed using the AdnaTest BreastCancer Select. Briefly, epithelial cells are isolated by antibody-linked Dynalbeads directed against EpCAM (also known as GA733-2) and MUC-1. The immunomagnetic-captured cells were lysed. The mRNA was isolated from the cell lysates using the Dynabeads mRNA DIRECT Microkit (Dynal Biotech Gmbh, Hamburg, Germany) included in the AdnaTest BreastCancer Detect kit. The mRNA was then transcribed into cDNA which was then used as a template in a 2-step quantitative real-time PCR for selected markers including MUC-1, HER2 and GA733-2 (Fig. 2). The AdnaTest was considered to be positive if at least 1 or more of the 3 markers shown the expression signal intensity was equal to or greater than 0.15 ng/μl.
Estrogen receptor/progesterone receptor analysis
Immunostaining was performed on 5-μm-thick tissue sections of the primary tumor using primary antibodies directed against estrogen receptor (ER) (clone 6F11; Novocastra) and progesterone receptor (PR) (clone PgR1294; Dako Cytomation) using a Bond Max autostainer (Vision Biosystems) with epitope retrieval (citrate buffer at pH 6.0). Diaminobenzidine tetrahydrochloride was used for signal recognition of the antigen. Distinct nuclear staining in the tumor cells was regarded as positive for the marker. The number of tumor cells showing nuclear staining was recorded and expressed as a percentage. Expression of both markers in 1–9% of the tumor cells was regarded as low positive and in 10% or more as positive.
Fluorescence in situ hybridization analysis
The HER2/neu status of the primary tumor was determined by immunohistochemical staining of 4-μm-thick sections cut from a representative paraffin block of the invasive tumor. The slides were incubated with the anti-HER2/neu monoclonal antibody e2-4001 (1:100 dilution) on a Dako autostainer (Dako, Corp., Carpinteria, CA) with the LSAB2 peroxidase kit (Dako, Corp.) using 3,3′-diaminobenzidine. The percentage of cells displaying complete membranous staining and the intensity of staining were evaluated on a semiquantitative scale from 0 to 3+. In all cases exhibiting any positivity (1+, 2+, 3+), we confirmed the protein overexpression with fluorescence in situ hybridization of HER2/neu gene amplification. We used the PathVysion HER2/neu kit (Abbott Laboratories, Chicago, IL), which uses 2 directly labeled fluorescent DNA probes, one specific for the HER2/neu gene locus and one for the alpha satellite DNA sequence at the centromeric region of chromosome 17. Signals were counted for 50 nuclei of the tumor cells by using an epifluorescence microscope and the ratio of HER2/neu to chromosome 17 signals was calculated. A ratio greater than 2.2 was considered to represent HER2/neu gene amplification, which was regarded as a positive result.
We used the Pearson's χ2 test to examine the correlation between the results obtained using the AdnaTest and the CellSearch system. We also calculated the sensitivity and specificity of each method, as well as the overall agreement between the AdnaTest and CellSearch results, along with 95% confidence intervals (p < 0.01). All statistical analyses were performed using the SPSS software (version 17; SPSS, Chicago, IL).
A total of 55 MBC patients were enrolled between January 2006 and March 2007. Detection of CTCs was performed in all patients using both the CellSearch system and AdnaTest. Patient and tumor characteristics at the time of enrollment are shown in Table 1. The median age was 49 years (range, 21–85).
Table 1. Patient and tumor characteristics
Primary tumor hormone receptor status for ER/PR (detected by IHC) and HER2 overexpression (detected by IHC and FISH) of primary tumors were positive in 36 (65%) patients and 12 patients (22%), respectively. Triple-negative (ER-, PR- and HER2-negative) phenotype was present in 11/55 (20%) at the time of primary tumor diagnosis. Neither CTC detection method was found to be associated with ER, PR, or HER2 status of the primary tumor.
Baseline detection of CTCs by CellSearch
In the CellSearch system, patients were grouped according to the frequency of positive CTCs using the cutoff values of 2 or more CTCs per 7.5 ml of blood and 5 or more CTCs per 7.5 ml of blood. The CellSearch system reported 26/55 patients (47%) as CTC positive with the cutoff value of 2 or more CTCs per 7.5 ml of blood, and 20/55 patients (36%) as CTC positive with the cutoff value of 5 or more CTCs per 7.5 ml of blood (Table 2).
Table 2. Incidence of CTCs by CellSearch
Detection of CTCs by the AdnaTest
In the AdnaTest, blood samples were regarded as CTC positive if a PCR fragment of at least 1 of the 3 markers, GA733-2, MUC-1, or HER2, was detected at a concentration of 0.15 ng/μl or greater. The AdnaTest reported 29/55 patients (53%) as CTC positive, with the positive expression rates of 18% for GA733-2, 44% for MUC-1 and 35% for HER2 (Table 2). Of the 29 patients reported to be CTC positive by the AdnaTest, 19 were positive for either 2 or all 3 markers and the remaining 10 were positive for only 1 marker. The AdnaTest detected HER2 expression in CTCs in 19 patients; but HER2 status of the primary tumor was positive and concordant in only 5 patients, negative and discordant in 11, and unknown in 3 of these patients.
Comparison between the AdnaTest and CellSearch system
We found a positive agreement between the AdnaTest and CellSearch system (with a cutoff value ≥2 CTCs/7.5 ml of blood) of 73%, and a positive agreement between the AdnaTest and CellSearch system (with a cutoff value ≥5 CTCs/7.5 ml of blood) of 69% (Tables 3 and 4, respectively).
Table 3. Distribution of AdnaTest Results by CellSearch CTCs (<2 vs. ≥2)
Table 4. Distribution of AdnaTest Results by CellSearch CTCs (<5 vs. ≥5)
Our data indicate that the AdnaTest is at least equivalent to the CellSearch system for the detection of CTCs in patients with MBC. The data indicate detection of CTCs in higher proportion of cases compared to an appropriate agreement, considering the differences in methodologies between the 2 technologies. We found that 26/55 patients were reported as CTC positive by the CellSearch system (with a cutoff value of ≥2 CTCs/7.5 ml of blood), while 29/55 patients were reported as CTC positive by the AdnaTest, giving a positive agreement of 73% between the 2 methods. However, the AdnaTest was positive in 9 cases in which the CellSearch system was unable to detect circulating tumor cells, and the CellSearch system was positive in 6 cases in which the AdnaTest was considered negative considering possibility of complementary value. Based on recently reported data, the concordance rate of the Cell Search system and the AdnaTest is up to 88%.22–24
An interesting finding of this study is that among the 19 patients who were demonstrated to have CTC expression of HER2, 11 had no demonstrated HER2 gene amplification in the primary tumor. Assessing CTCs for HER2 overexpression could potentially help identify patients, who would not be considered otherwise candidates for HER-2 targeted therapy, because of negative disease status on the primary tumor. Discordance between HER2 amplification in the primary tumor and in CTCs has been previously reported. For instance, Pestrin et al. found that 29% of MBC patients with HER2-negative primary tumors had HER2-amplified CTCs, and 42% of patients with HER2-amplified primary tumors had HER2-negative CTCs.25 Similarly, Reuben et al. recently reported that although there was concordance in HER2 amplification between the primary and metastatic tumors of MBC patients, there was discordance in HER2 transcripts between the primary tumors and CTCs.26 These data suggest that there is a subset of patients with HER2-negative primary tumors who, as their disease progresses, develop CTCs with HER2 amplification and that such information can be important for treatment selection. Future prospective studies testing this hypothesis are currently being planned.
In consistency with our data, a degree of discordance in HER2 status (between 0 and 26%) in primary breast cancer and metastases have been reported.27 These findings suggest that breast cancer is a dynamic disease that evolves with time and as a function of therapy. The phenotypical changes or discrepancies between the primary tumor and the CTCs also support the notion that CTCs may represent a unique and heterogeneous tumor cell population with special biological properties that permit travel to distant sites and establishment of clinically disseminated disease. Reports indicate that CTCs are representative of the heterogeneity associated with cancer and are derived from clones in the primary tumors28, 29 Monitoring expression of HER2 on CTCs in patients with advanced cancer may be essential for rationally designed therapy since reliance on the immunophenotype of the primary tumor can be misleading.30
Using an immunomagnetic detection approach, Austrup et al. reported the prognostic significance of genomic alterations (e.g. HER2 overexpression) present in CTCs from the blood of patients with breast cancer.31 They found that the presence and the number of genomic imbalances measured in disseminated tumor cells were significantly associated with a worse prognosis.31Subsequently, Meng et al. demonstrated that CTCs recapitulate the HER2 status of the primary tumor.32 The authors demonstrated that a fraction of patients with HER2-negative primary tumors had detectable HER2 gene amplification in their CTCs, suggesting acquisition or selection of this phenotype during cancer progression. Intriguingly, the initiation of trastuzumab-based therapy in a few cases with altered HER2 status in CTCs associated with a clinical response was also observed.33 Furthermore, a number of other studies reported that up to one-third of patients whose primary tumors do not overexpress HER2 have CTCs with amplified HER227, 33, 34, 35 In view of that CTCs are likely to derive from clones in the primary tumors28, 29, our and others' data showing the discordance in HER2 status between the primary tumor and CTCs may indicate that tumor cells are undergoing genetic changes. Together, these important observations support the argument that detection of CTCs and determination of their gene amplification or expression can be used for better tailoring therapies.
Collecting representative tissue from solid tumor metastases usually requires more invasive procedures that increase the risk of complications and discomfort. Furthermore, these procedures may not provide an adequate specimen for detailed analysis and typically cannot be repeated for dynamic evaluation of the biological changes during treatments. Theoretically, CTC detection would allow specific genes (e.g., HER-2, epidermal growth factor receptor, and mammaglobin B) or more global gene expression to be analyzed while using specific targeted treatment for MBC based on the expression of the CTCs.33, 36, 37 This information could then be used to design specific treatments that more appropriately reflect the dynamics and heterogeneity of MBC.
The AdnaTest BreastCancer Select/Detect is a new approach for the detection of CTCs in patients with MBC. The AdnaTest has equivalent sensitivity to that of the CellSearch system in detecting 2 or more CTCs. The advantage of the AdnaTest over the CellSearch system is that it enhances the possibility of detecting very low numbers of CTCs or occult CTCs based on their expression of tumor-associated genes. The immunomagnetic cell capture technology, combined with multiplex RT-PCR in AdnaTest may potentially allow the detection of a broad range of molecular abnormalities in CTCs that can more accurately characterize the biological status of metastatic disease and support design of future studies testing CTCs directed therapies. AdnaTest complements the CellSearch system for CTC detection.