Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Improving the yield of circulating tumour cells facilitates molecular characterisation and recognition of discordant HER2 amplification in breast cancer

Improving the yield of circulating tumour cells facilitates molecular characterisation and... British Journal of Cancer (2010) 102, 1495 – 1502 & 2010 Cancer Research UK All rights reserved 0007 – 0920/10 $32.00 www.bjcancer.com Improving the yield of circulating tumour cells facilitates molecular characterisation and recognition of discordant HER2 amplification in breast cancer 1,2 3,5 3,4,5 1 4 3,4,5 3,5 LM Flores , DW Kindelberger , AH Ligon , M Capelletti , M Fiorentino , M Loda , ES Cibas , ,1,3,6 ,1,3,6 * * PA Janne and IE Krop 1 2 Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; Department of Biology, University of Massachusetts, Boston, MA, 3 4 USA; Harvard Medical School, Boston, MA, USA; Center for Molecular Oncologic Pathology, Dana-Farber Cancer Institute, Boston, MA, USA; Department of Pathology, Brigham and Women’s Hospital, Boston, MA, USA BACKGROUND: Circulating tumour cells (CTCs) offer a non-invasive approach to obtain and characterise metastatic tumour cells, but their usefulness has been limited by low CTC yields from conventional isolation methods. METHODS: To improve CTC yields and facilitate their molecular characterisation we compared the Food and Drug Administration- approved CellSearch Epithelial Kit (CEK) to a simplified CTC capture method, CellSearch Profile Kit (CPK), on paired blood samples from patients with metastatic breast (n ¼ 75) and lung (n ¼ 71) cancer. Molecular markers including Human Epidermal growth factor Receptor 2 (HER2) were evaluated on CTCs by fluorescence in situ hybridisation (FISH) and compared to patients’ primary and metastatic cancer. RESULTS: The median cell count from patients with breast cancer using the CPK was 117 vs 4 for CEK (Po0.0001). Lung cancer samples were similar; CPK: 145 cells vs CEK:4 cells (Po0.0001). Recovered CTCs were relatively pure (60–70%) and were evaluable by FISH and immunofluorescence. A total of 10 of 30 (33%) breast cancer patients with HER2-negative primary and metastatic tissue had HER2-amplified CTCs. CONCLUSION: The CPK method provides a high yield of relatively pure CTCs, facilitating their molecular characterisation. Circulating tumour cells obtained using CPK technology demonstrate that significant discordance exists between HER2 amplification of a patient’s CTCs and that of the primary and metastatic tumour. British Journal of Cancer (2010) 102, 1495 – 1502. doi:10.1038/sj.bjc.6605676 www.bjcancer.com & 2010 Cancer Research UK Keywords: circulating tumour cells; breast cancer; lung cancer; HER2; FISH The need for access to samples of metastatic cancer and the number of pure CTCs is necessary for these downstream inherent difficulty in obtaining these using conventional biopsies, applications. has led to interest in alternative sources of metastatic cells. One Circulating tumour cells have become increasingly accepted as such alternative is to analyse circulating tumour cells (CTCs). both an independent prognostic factor and as a marker of Circulating tumour cells are thought to be quite rare in the blood therapeutic response and attempts have been made to standardise of most patients with metastatic disease – it has been estimated their isolation and characterisation (Hayes and Smerage, 2008). that they are present at a frequency of 1 in 10 blood cells (Mostert The automated CellSearch System (Veridex; Warren, NJ, USA) is et al, 2009). Circulating tumour cells provide the potential ability the first method to be clinically approved by the United States to monitor metastatic disease, prognostic information and a Food and Drug Administration to capture and detect CTCs. This means to perform non-invasive molecular interrogation of cancers, system relies on immunomagnetic isolation followed by the use of as these cells can provide a molecular snapshot, or real-time fluorescence microscopy for analysis. Two different but related biopsy, of tumour cells. Circulating tumour cells can also methods of sample preparation, the CellSearch Epithelial Cell Kit potentially be used to determine the mechanisms of acquired (CEK) and the CellSearch Profile Kit (CPK) are currently available. resistance to targeted therapies, which only occur or evolve during Both methods use a positive immunomagnetic selection with anti- the course of drug treatment. The ability to isolate a relatively large epithelial cell adhesion molecule antibodies linked to iron particles to enrich for CTCs. In the CEK system, captured cells are permeabilised and labelled with cytokeratin and CD45-specific *Correspondence: Dr IE Krop; E-mail: ikrop@partners.org or antibodies, and the nuclear stain 4 -6-diamidino-2-phenylindole Dr PA Ja¨nne; E-mail: pjanne@partners.org (DAPI), and are analysed by semi-automated counting of These authors contributed equally to this manuscript appropriately labelled CTCs (Riethdorf et al, 2007). In contrast, Received 15 December 2009; revised 24 March 2010; accepted 29 the CPK procedure does not involve labelling or enumerating the March 2010 cells, but rather only uses the anti-epithelial cell adhesion molecule Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al antibody coupled to iron particles to yield an enriched population (Johnson & Johnson, 2008). Samples processed using the CPK of CTCs, which can be used for molecular analyses. However, in method (per manufacture’s instructions, (Johnson & Johnson, contrast to the CEK method, the CPK procedure is not Food and 2008)) were collected in a dilution buffer (phosphate buffered Drug Administration approved for clinical management. saline, 0.5% bovine serum albumin and 0.1% sodium azide) and The primary limitation of the CEK system is the low yield of immediately transferred to glass slides as cytospin preparations CTCs. In several large clinical trials of patients with metastatic using a ThermoFisher Cytospin 3. The centrifugation was breast cancer, the median number of isolated CTCs was five per performed at 500 g for 5 min with a cytology funnel, thin filter 7.5 ml of blood (Budd et al, 2006; Hayes et al, 2006; Hayes and (Thermo Fisher Scientific, Pittsburgh, PA, USA), and a SuperFrost Smerage, 2008). Although the low numbers of CTCs are sufficient Plus slide (Thermo Fischer Scientific). The cells on the resulting to provide clinical prognostic information, in which a cutoff of X5 cytospin slide were fixed with methanol for IF or immunohis- CTCs per 7.5 ml of blood is used to delineate a poor prognosis tochemistry or fixed with methanol:acetic acid (3 : 1) for FISH. population, the ability to perform molecular analyses of the CTCs Cells were stained with anti-cytokeratin phycoerythrin/DAPI/ is limited. In addition, it has been reported that the degree of CD45-allophycocyanin (Dako, Carpinteria, CA, USA) and then leukocyte contamination using the CEK system is quite high, manually counted using a standard fluorescence microscope which further complicates molecular analyses of CTCs. These (Olympus, Center Valley, PA, USA). Cytokeratin staining was limitations have led to the development of alternative CTC considered positive when cells displayed unequivocal cytoplasmic isolation methods, including technologies based on size filtration, cytokeratin staining in a ring-like pattern (confirming an intact density gradients, and microfluidic techniques, which have cytoplasm) surrounding an intact nucleus. demonstrated increased yields of CTCs (Vona et al, 2000; Nagrath et al, 2007; Adams et al, 2008; Williams et al, 2009). For example, a Cell lines and reagents microfluidic chip-based CTC technology has been developed, The breast cancer cell line SK-BR-3 was obtained from American which isolates relatively high numbers of CTCs and can be used to Type Culture Collection. The EGFR-mutant (del E746_A750) non-invasively detect drug-sensitive and -resistant epidermal NSCLC cell line, HCC827, has been described previously growth factor receptor (EGFR) mutations from non-small cell (Mukohara et al, 2005). The two cell lines were grown in lung cancer (NSCLC) patients (Nagrath et al, 2007; Maheswaran Dulbecco’s Modified Eagle Medium or Roswell Park Memorial et al, 2008). This method uses epithelial cell adhesion molecule- Institute 1640 (CellGrowth, Invitrogen, Life Technologies, specific antibodies that are conjugated to posts within the CTC Carlsbad, CA, USA) respectively, supplemented with 10% foetal chip. Currently this system is limited by the lack of widespread bovine serum. Gefitinib and lapatinib were obtained from commercial availability of the system, the inability to release the CTCs from the sources. Stock solutions of both drugs were prepared in dimethyl microfluidic posts, prolonged processing time (samples are run at –1 sulphoxide and stored at 201C. For in vitro recovery experiments 1–2 ml h ), and the requirement that samples be processed within SKBR3 cells were grown in DMEM with 10% foetal bovine serum, a few hours of collection (Nagrath et al, 2007). –1 10 mgml EGF, and 1.5 mML-glutamine. The HCC287 cells were To address some of the limitations of current CTC systems, we grown in ACL4 (Invitrogen, Life Technologies) with 10% foetal evaluated the CPK system as a method to isolate and study CTCs. bovine serum. For detailed methods see Supplementary methods. Unexpectedly, this method resulted in isolation of 420-fold more CTCs than the CEK method. We further characterised the purity of the isolated CTCs and evaluated them using immunofluorescence Fluorescence in situ hybridisation (FISH) (IF) and fluorescence in situ hybridisation (FISH). Our findings The EGFR/CEP 7 and Human Epidermal growth factor Receptor 2 suggest that a modification in the use of the current CellSearch (HER2)/Centromere Probe 17 (CEP17) bacterial artificial chromo- system can result in a significant increase in the ability to isolate and some (BAC) probes were obtained from Vysis Molecular (Abbot characterise CTCs from patients with breast cancer and NSCLC. Park, IL, USA). BAC mesenchymal-epithelial transition factor (MET) (RP11-95I20) was obtained from CHORI (Children Oakland Research Hospital, Oakland, CA, USA). Labelling was done with MATERIALS AND METHODS the Nick Translation Kit (Vysis, Abbott Molecular, Des Plaines, IL, USA). See Supplementary data for detailed methods of FISH Patients analysis of CTC, and for IF and immunohistochemistry staining Between December 2007 and March 2009 patients with metastatic methods. NSCLC (n¼ 71) or metastatic breast cancer (n¼ 75) were identified from the Thoracic Oncology and Women’s Cancer clinics at the RESULTS Dana Farber Cancer Institute. Characteristics of the patients are described in the Supplementary Table S1. The blood was collected The CPK method isolates a greater number of cells than from each donor into CellSave blood collection tubes (Veridex) or the CEK method ethylenediaminetetraacetic acid (EDTA) tubes, where specified. The blood samples were maintained at room temperature and processed We hypothesised that CTCs may be lost in the additional within a maximum of 72 h after collection. All samples from patients permeabilisation, labelling, and wash steps specific to the CEK with breast cancer were obtained at the time they were initiating a method, particularly if CTCs are more ‘fragile’ than other cell new therapy for the treatment of their disease. All of the samples types. To directly compare the yield of cells isolated using CEK and from patients with NSCLC were obtained while the patient was CPK procedures, we collected two peripheral blood samples from currently under therapy. Control blood samples were drawn from patients with metastatic breast cancer (n ¼ 75) and NSCLC healthy volunteers with no history of malignant disease. All patients (n ¼ 71), and processed them in parallel. Cells obtained using the provided written informed consent, and studies were approved by CPK method were counted manually and were compared with the the Dana Farber Cancer Institute Institutional Review Board (IRB). standard CEK counts obtained using the CellSearch system. The median cell count from patients with breast cancer using the CPK method was 116.5 per 7.5 ml of blood (range 4–2432) compared Sample processing and counting with 4 (range 0–57) using the CEK method (Po0.0001, Figure 1A). Samples were processed using the CEK method, including semi- The median cell count from patients with NSCLC processed using automated counting, according to the manufacture’s instructions the CPK method was 145 per 7.5 ml of blood (range 5–1801) British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK Improving yields of circulating tumour cells LM Flores et al CEK kit semi-automated count CPK kit-manual count 0 2 5 10 25 50 100 200 300 400 500 600 700 800 900 1000 2000 3000 No. of cells CEK semi-automated count CPK-manual count 0 2 5 10 25 50 100 200 300 400 500 600 700 800 900 1000 2000 No. of cells Figure 1 CPK method improves cell yields over CEK method. The blood samples from patients with (A) breast cancer (n ¼ 75) or (B) NSCLC (n ¼ 71) processed in parallel by the CEK method with semi-automated quantification (open columns) or the CPK method with manual quantification (closed columns). compared with 4 (range 0–53) using the CEK procedure (Po0.0001; Approximately 60% of cells isolated by CPK (range: 21–95%) were Figure 1B). We also compared these findings on a per patient basis identified as tumour cells by cytokeratin staining, whereas 6% (Supplementary Figures S1A and B) and demonstrated that for all (range: 1–22%) were identified as white blood cells (WBCs) by patients, the cell count by the CPK method exceeds the cell count by CD45 staining (Figure 2A and data not shown). These percentages the CEK method (breast cancer: R ¼ 0.004919; lung cancer: correspond to a median absolute number of cytokeratin positive R ¼ 0.006952). To determine whether the observed yield differences cells of 280 per sample (range 20–20 000) and a median of 34 between the CEK and CPK methods are due to differences in the (range 20–476) CD45-positive cells per sample. The remaining method of cell counting (manual vs semi-automatic), we compared 32% (range: 5–68%) of cells stained with DAPI only and could CTC counts obtained using the automated CEK method with that of not be classified as either tumour cells or leukocytes. As a a manual cell count of the same sample. The manual CEK count was complementary approach, we used fluorescence-activated cell performed by removing the enumerated samples from the ‘MagNest’ sorting on 11 patients (eight NSCLC and three breast cancer) to cartridges, processing them as cytospin preparations and counting compare the populations of CTCs with WBCs (Figure 2B). The them manually. In patients (n¼ 75 patients; 100 samples) with results were similar for cytokeratin (mean positive: 61%; range breast cancer (Supplementary Figure S1E), the median cell count 35–77%), except the frequency of WBCs was estimated to be even using the semi-automated CEK method was 4 per 7.5 ml of blood lower (mean positive: 1%; range 0–7%) than by IF (compare (range 0–57) compared with 5 per 7.5 ml of blood (range 0–61) Figure 2A and B). when the CEK processed specimens were counted manually As an additional method to evaluate the nature of the cells (P¼ NS). Similarly, in NSCLC patients (n¼ 71 patients; 100 isolated using the CPK method, we performed FISH using probes samples), the median cell count using the semi-automated CEK for HER2, EGFR, and MET. Given that a mean of 32% of cells had denuded membranes (DAPI-only cells; Figure 2A) and thus could method was four per 7.5 ml of blood (range 0–53) compared with five per 7.5 ml of blood (range 5–67) when the CEK processed not be definitely stained for cytokeratin, the identification of specimens were counted manually (P¼ not significant; Supplemen- aneuploidy and/or gene amplification in a high fraction of the cells tary Figure S1F). We further compared these findings on a per would strongly suggest that some or all of these cells were CTCs patient basis (Supplementary Figures S1C and S1D) and observed a and not WBCs (Figures 3A and B). Among the CTCs from patients strong correlation between the cell counts obtained using the semi- with breast cancers clinically defined as HER2 positive from automated CEK method with the manual CEK method (breast analyses of their primary tumours (FISH þ and/or IHC 3 þ by 2 2 cancer specimens: R ¼ 0.9930; lung cancer specimens: R ¼ 0.7200). local testing, n ¼ 24), 72% of all captured cells had a HER2 gene Together, our findings suggest that the CPK method isolates a copy number X4. This threshold has been used as an indicator of greater number of cells than the CEK method and that these HER2 amplification in previous studies (Meng et al, 2004). Only findings are not due to differences in the method of cell counting. 3% of the cells had two copies of HER2. In an additional six patients with clinically defined HER2-negative breast cancers, only 5% of captured cells had X4 copies of HER2 (Figure 3D). In The CPK method isolates a highly enriched population contrast, the majority (95%) of cells in those patients had a HER2 of CTCs gene copy number of 2 (Figure 3D). Similarly, among samples One explanation for the improved yields observed with the CPK from NSCLC patients (n ¼ 30), 53% had X4 copies of EGFR and method is that the isolated cells are not all CTCs but instead a 13% had X4 copies CEP 7 (Figure 3C). Only a minority of cells mixture of CTCs and contaminating leukocytes. This determina- (5%) had two copies of EGFR and CEP 7. tion is critical if the isolated cells are to be used for any subsequent As an additional control, we analysed 40 samples from healthy analyses. To determine the composition of the cells isolated using volunteers with no history of malignant disease. No tumour cells the CPK method, we performed IF using anti-CD45 and anti- were found with either the CEK semi-automated enumeration or cytokeratin (AE1/AE3) antibodies in 30 NSCLC and 30 breast with the CPK manual count (Supplementary Figure S2A). In cancer patients. Three distinct cell populations were observed: (1) samples processed using the CPK method, the median total CD45 positive, cytokeratin negative, (2) CD45 negative, cytokeratin number of cells was five per 7.5 ml of blood (range: 0–50). These positive, and (3) cells that were negative for both CD45 and cells were positive for CD45 and negative for cytokeratin using IF, cytokeratin but stained with the nuclear stain DAPI (Figure 2C). with only two gene copies of CEP7/EGFR or HER2 as detected by & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 No. of samples No. of samples Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al % Cytokeratin positive (Supplementary Figure S3C) with a low coefficient of variability % DAPI only (CV: 9.7%). We also evaluated the impact of time on the ability to % CD45 positive isolate CTCs using the CPK method. As many clinical trials involve multiple centres, it is important to determine the stability of unprocessed samples to determine whether the yield of CTCs declines over time. For these studies, we collected three CellSave (LLC a Johnson&Johnson company, Raritan, NJ, USA) or EDTA 0 tubes from seven NSCLC patients and incubated the samples at room 123456789 1011121314151617 18 19 20 21 22 23 24 25 26 27 28 29 30 temperature for 24–72 h before processing by the CPK method. We Patient specifically evaluated the collection of blood into both CellSave and % CD45 positive 100 EDTA tubes to determine the impact of the fixative (present in the % Cytokeratin positive CellSave tubes; absent in the EDTA tubes) on the impact of cell recovery over time. We isolated similar numbers of cells using both the CellSave (Supplementary Figure S3A) and EDTA tubes (Supplementary Figure S3B). There was no significant decline (coefficient of variation (CV) 9.1% and CV 6.5%) in the number of cells isolated with either method over the 72-h period and no decline in the number of cells isolated from CellSave tubes incubated at 123456789 10 11 room temperature for up to 144 h (Supplementary Figure S3A and Patient data not shown). However, there was significant decline in the number of cells isolated from EDTA tubes incubated at room temperature for X96 h before processing by the CPK method. Treatment affects yield of cell recovery Previous studies have found that when tumour cell lines are spiked into blood and isolated using the CEK method, the rate of cell recovery is high (85–95%; Riethdorf et al, 2007). However, the actual CTC yields from cancer patients using the CEK method are limited (Sleijfer et al, 2007). We hypothesised that one explanation for the low yields seen in clinical practice is that CTCs are more ‘fragile’ than cancer cell lines either due to apoptosis or damage acquired during transit through the bloodstream and, therefore, are not captured as efficiently with the anti-epithelial cell adhesion molecule antibodies and/or are lost in the post-capture washing and labelling steps. To investigate this possibility, we examined breast (SKBR3; HER2 amplified) and lung cancer (HCC827; EGFR Figure 2 The CPK method isolates a highly enriched population of del E746_750) cell lines either unprocessed or spiked into normal CTCs. Percentage of total cells captured by the CPK method from patients blood, then processed with the CPK or CEK methods. Cells were with NSCLC, staining for cytokeratin (AE1/AE3), CD45, or DAPI nuclear pretreated with the EGFR/HER2 inhibitor lapatinib (SKBR3) or the stain alone by (A) immunofluorescence or (B) FACS. Similar results were EGFR inhibitor gefitinib (HCC827) to induce apoptosis, or were seen with samples from patients with breast cancer. (C) Representative untreated. The cells were then stained with Ki67 or Apoptosis immunofluorescence image of CPK-captured cells from patient with Detection Using Terminal Transferase and Biotin-16-dUTP (TUNEL) NSCLC, labelled with cytokeratin (green), DAPI (blue), and CD45 (red). to evaluate proliferating and apoptotic cells, respectively. As expected, unprocessed cells demonstrated increased rates of apoptosis and decreased proliferation when treated with inhibitor FISH analysis, consistent with normal WBC. Collectively, these (Figure 4C). Interestingly, no apoptotic cells were noted in the data indicate that that majority of cells captured by the CPK samples processed with the CEK method in either untreated or system are in fact CTCs with only a minority of cells being WBCs. treated samples, although the treated cells that were captured by All of the samples described were prepared using the low-speed that method did have decreased levels of Ki67 staining, consistent cytospin method using a single cytology funnel with thin filter with the effect of the inhibitor (Figure 4A). In samples processed (see Materials and Methods section). We also evaluated other with the CPK method, a small percentage of TUNEL-positive cells techniques for preparing slides, including high-speed cytospin, were recovered in the untreated samples, which increased with different funnel types, a thin prep or cell block, and applying the inhibitor treatment, although not to the level of unprocessed samples to slides directly as a smear. In a comparative analysis, samples (Figure 4B). the low-speed cytospin method produced the highest yield, with We also evaluated the frequency of Ki67- and TUNEL-staining the least contamination, least cell damage, and the least slide-to- cells isolated from seven NSCLC patients using the CPK method. slide variation in comparison to other various methods (data not The frequency of TUNEL-staining cells was low (mean 1.9%; range shown). 0–6.4%) similar to what was observed in the SKBR3 and NSCLC cell lines (Figure 4D). The frequency of Ki67-staining cells was also lower (mean 0.7%; range 0.3–1.2%) in the CTCs from cancer The CPK method has low intrapatient variability and patients compared with the levels seen with cell lines (Figure 4D). samples are stable for at least 72 h When parallel samples were processed using the CEK method, no We next examined whether the number of cells isolated from TUNEL-positive cells were detected, but the fraction of Ki67- patients with cancer using the CPK method was reproducible and positive cells was numerically higher than in the CPK processed stable over time. We first obtained three samples from each of samples (Figure 4E). This observation is consistent with the seven patients with NSCLC at the same time point and processed hypothesis that potentially more ‘fragile’ apoptotic and non- them in parallel. The number of cells isolated was similar proliferating cells are disproportionably lost by the more British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK % of cells % of cells Improving yields of circulating tumour cells LM Flores et al Lymphocytes CTC % of cells with ≥4 of EGFR % of cells with ≥4 of CEP7 100 % of cells with ≥4 of MET 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Patient % of cells with 2 of HER2 % of cells with ≥4 of HER2 % of cells with ≥6 of HER2 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930 Patient HER2+ HER2– Figure 3 FISH analysis confirms that samples processed by the CPK method have a low percentage of contaminating normal cells. Representative FISH images of cells processed by CPK method, (A) lymphocyte with two copies of CEP7 (green), EGFR (red), and MET (blue), (B) CTC with amplified EGFR and MET. (C) Percentage of total cells captured by the CPK method from patients with NSCLC with X4 copies of EGFR, CEP7, and MET per nucleus. (D) Percentage of total cells captured by the CPK method from patients with clinically defined HER2-positive (patient number 1–24) or HER2-negative (patient number 25 –30) breast cancer, with the indicated copies of HER2 per nucleus. processing intensive CEK method. However, because of the small of their primary tumour (Meng et al, 2004; Wulfing et al, 2006; number of cells recovered, it was not possible to make a definitive Pestrin et al, 2009). No data exists examining the relationship comparison. Together with the data obtained from the cell lines, between the CTCs’ HER2 status and that of metastatic tissue. We these findings suggest that CTCs undergoing apoptosis in the hypothesised that the ability of the CPK technology to provide blood stream, either spontaneously or as result of treatment effect, relatively high numbers of CTCs and allow robust FISH analysis may be too fragile to survive the capture step and/or are being would facilitate the determination of the relationship between the eliminated during subsequent specimen processing. This effect HER2 status of a patients primary breast cancer, metastatic lesions, seems to be more pronounced with the CEK than CPK method and and CTCs. To test this hypothesis, we assessed the presence of may help explain the increased CTC yields seen with the latter HER2 amplification by FISH, normalised for CEP 17, on CTCs technique. These observations may also help to explain why from 75 women with breast cancer for whom the HER2 status of alternative CTC-capture techniques, such as the microfluidic chip the primary cancer and a metastatic biopsy sample was available or microfiltration systems that minimise the number of labelling (patient characteristics available in Supplementary Table S1). and wash steps are associated with an increase in CTC yield over For patients with HER2 gene amplification of their primary the CEK method. breast cancer, the degree of concordance with the HER2 status of their CTCs was high, with only 1 of 45 (2%) patients demonstrating loss of HER2 amplification in the corresponding CTCs (Table 1). In Characterisation of HER2 amplification of CTCs by that one patient with discordance, the metastatic biopsy was HER2 FISH analysis amplified, reflecting the primary cancer. In contrast, for those Several small studies have reported varying degrees of discordance patients in whom the primary breast cancer was HER2 negative, between the HER2 status of a patient’s breast cancer CTCs and that significant discordance between the primary, metastatic biopsy, & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 % of cells % of cells Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al 100 100 Tunel Tunel Table 1 Relationship between HER2 amplification of primary cancer and Ki67 Ki67 80 80 that of CTC 60 60 Total No. Patients Patients 40 40 Primary of with with 20 20 tumor patients HER2CTCs HER2+CTCs Discordance 0 0 HER2+ 45 1 44 2% HER2 30 20 10 33% Abbreviations: CTC ¼ circulating cancer cells; HER2 ¼ human epidermal growth CEK CPK factor receptor 2. HER2 positive defined as ratio of HER2/CEP17 X2.0 by FISH. 100 10 Tunel Tunel Ki67 80 Ki67 Cyclin D1 evaluate membrane staining of EGFR, pEGFR, HER2 and pHER in 20 CTCs from 20 patients with clinically HER2-positive breast cancer 0 (Supplementary Figure S5B). Similar to the CTCs isolated from NSCLC patients, we were able to detect EGFR, pEGFR, HER2, and 123456789 10 Patient pHER using IF in the CTCs and the frequency of staining was less than in the SKBR3 tumour cell line (Supplementary Figure S5B). In vitro propagation The standard CellSave tubes contain a No processing fixative that prevents the isolation of viable tumour cells. Having observed that CTCs can be efficiently isolated from the Method Number of KI67+ KI67+ KI67+ blood drawn in EDTA tubes that do not contain a fixative evaluable cells/total cells/total cells/sample samples cells (mean, %) cells (range, %) (Supplementary Figure S3B), we explored whether we could isolate SKBR3 and HCC827 cells using EDTA tubes and grow them CEK breast 28 21.1 10 – 44 1–6 in vitro. We spiked a range (5–1000) of SKBR3 or HCC827 cells CPK breast 75 7.1 1.1 – 13.2 2 – 480 CEK lung 21 32 10 – 64 2–8 into the normal blood drawn into EDTA tubes and processed them CPK lung 70 4.4 1.6 – 16 1–113 using the CPK method. We were able to isolate viable SKBR3 and HCC827 cells (Supplementary Figure S5) and propagate them Figure 4 Apoptotic CTCs are less effectively captured by CEK or CPK in vitro. However, this required a minimum input of 50 cells to be methods. SKBR3 (HER2 breast cancer) or HCC827 (EGFR mutant NSCLC) cells treated with vehicle or tyrosine kinase inhibitor (SKBR3: 1 mM processed (Supplementary Figure S5). Optimal growth occurred lapatinib and HCC827; 1 mM gefitinib) for 24 h and processed with (A) in the presence of enriched medium and cell adhesion matrix CEK method. (B) CPK method or (C) smeared directly on slide without (data not shown). processing. Plots depict percentage of cells ( 1 s.d.) staining for Ki67 (open bars, proliferation marker) or TUNEL (closed bars, apoptosis marker) by immunohistochemistry. (D) Percentage of CPK-processed DISCUSSION CTCs from patients with NSCLC staining positive for KI67 and TUNEL. (E) Comparison of KI67 expression in CTCs recovered by CEK or CPK One of the most important issues limiting the use of CTCs as an methods. alternative to invasive biopsies has been their relatively low yield. Using the CellSearch System (Veridex), the most widely available and CTC was observed. A total of 10 (33%) of these 30 patients technology, the median yield is reported to be in the range of 1 cell with HER2-negative primary cancers had HER2-positive CTCs per ml of blood screened (Cristofanilli et al, 2004). Our results with by FISH analysis. The median HER2/CEP17 ratio of the CTCs in patients with breast and lung cancer using the CellSearch CEK these patients was 7.1. Interestingly, in 9 of the 10 patients with platform are consistent with that finding. However, two seemingly discordance, the metastatic biopsy specimen was HER2 negative. paradoxical observations prompted us to explore this technology further. The first is that several studies using the CellSearch platform have demonstrated that the instrument was highly Evaluation of biologic properties of CTCs isolated using efficient at capturing immortalised breast cancer cell lines spiked the CPK method into the human blood, with yields of X85% of the input cells Expression of receptor tyrosine kinases The CPK method is able (Allard et al, 2004). We have replicated those studies with similar to isolate a highly enriched population of CTCs and greater results; a capture rate of 92% (data not shown). The second numbers of CTCs than the CEK method. These findings provide observation was the finding that the new CTC–chip microfluidic the opportunity to initiate the molecular characterisation of CTCs. technology reported mean CTC counts from patients with breast In addition to FISH (Figure 3), we evaluated the membrane cancer of 79 cells per ml of blood (Nagrath et al, 2007). If the expression of total and phosphorylated HER2 and EGFR using IF CellSearch technology is already highly efficient at capturing CTCs on the breast and lung cancer cell lines spiked into blood and in model systems using cancer cell lines, how can another recovered with the CPK (Supplementary Figure S4A). The IF technology report 10–100-fold higher yields of CTCs from patient analysis was superior to chromogenic immunohistochemical samples? We postulated that human CTCs, particularly those that staining, as the immunohistochemical substrate tended to bind have been subject to anti-cancer treatment, may be significantly non-specifically to ferroparticles, and resulted in high background more ‘fragile’ than the cancer cell lines used in the initial staining (data not shown). We evaluated the expression of EGFR evaluation of the CellSearch platform. Owing to this fragility, the and pEGFR from CTCs in seven NSCLC patients (Supplementary multiple processing steps involved in the CEK system, including Figure S4C). We were able to detect staining in all samples additional washes and labelling steps, could cause degradation and although the percent of cells staining for either EGFR or pEGFR loss of captured CTCs. was lower than in the HCC827 cells that had been spiked into We hypothesised that the CPK system, which involves blood and processed using the CPK method. We also used IF to significantly fewer processing steps, may provide improved yields British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK SKBR3 vehicle SKBR3 lapatinib HCC827 vehicle HCC827 iressa SKBR3 vehicle Untreated SKBR3 cells SKBR3 lapatinib Treated SKBR3 cells HCC827 vehicle Untreated HCC827 cells Treated HCC827 cells HCC827 iressa % of cells % of cells % of cells % of cells Improving yields of circulating tumour cells LM Flores et al of CTCs from patient samples. Indeed, we observed that the CPK status of CTCs by FISH and compared it with both the patient’s consistently provided a significantly greater yield of CTCs than primary breast cancer and tissue from matched distant metastatic that obtained with the CEK platform. The median difference was sites. In patients with clinically HER2-positive primary and 29-fold in breast cancer samples and 36-fold in lung cancer metastatic cancers, virtually all (98%) also had HER2-amplified samples (Figure 1). In addition, virtually all of the breast and lung CTCs. In contrast, in patients with clinically HER2-negative patients had detectable CTCs using the CPK method, whereas our primary cancers, we observed a significant number (33%) of results with the CEK system demonstrate only 72% of patients had discordant cases in which a patient’s CTCs had clear amplification detectable CTCs, consistent with the published results (Hayes et al, of the HER2 locus, despite an absence of HER2 amplification in 2006). The improvement in yield observed with the CPK was not the primary cancer. This result is consistent with the data from simply due to the differences between the CEK’s semi-automated several other small studies demonstrating discordant HER2 counting system and the CPK’s manual counts (Supplementary expression in CTCs and together provide the rationale for a Figures S1E and F). When the CTCs from both methods were clinical trial of a HER2-directed therapy in such patients to counted manually, the differences persisted. Instead, our observa- definitively test the clinical relevance of this observation (Meng tion that when cell lines were pretreated with apoptosis-inducing et al, 2004; Wulfing et al, 2006; Pestrin et al, 2009). To the best of agents (Figure 4), the yield of CTCs dropped dramatically, suggests our knowledge, our study is the first to also examine the that CTC fragility, coupled with the more intensive processing relationship between HER2 amplification in CTCs and that of inherent in the CEK method, likely accounts for the observed metastatic biopsy tissue. In six out of seven patients with HER2- differences between the CEK and CPK methods. negative primary cancers and HER2-amplified CTCs, the meta- Importantly, we also found that the purity of the CTCs obtained static biopsy tissue was HER2 negative, matching the primary from the CPK method was quite high, typically in the 60–70% range. cancer rather than the CTCs. This observation suggests that either This was somewhat unexpected as discussions of the CellSearch CTCs represent a separate population of cells distinct from that system in the literature cite purities in the 0.01–0.1% range, due to which makes up the bulk of the metastatic disease, or that the contaminating leukocytes. In our studies with the CPK system, the cancer developed HER2 amplification in the interval between number of contaminating leukocytes, using either patient samples or the patients’ metastatic biopsy and the CTC collection. If the latter blood from normal volunteers, and assessed using multiple explanation were correct, one might expect to see a correlation independent methods, (IF, fluorescence-activated cell sorting), between the time interval from biopsy to CTC collection and the ranged from 20 to 476, and was always less than 25% of total cells. likelihood that the CTC would gain HER2 amplification. Although The one study we identified describing the number of contaminating we did not observe such a trend (data not shown), our sample size leukocytes observed with the CPK method reported a range of 60– is too small to make any conclusions about the mechanism of this 929 CD45-positive cells per sample (Sieuwerts et al, 2009). This study discordance. did not report the percentage of CTCs vs leukocytes in their paper, so In summary, we have demonstrated that using the CellSearch their results cannot be compared with our data in that regard. CPK assay, an instrument that is currently available in hundreds As has previously been demonstrated with the CEK system, we of clinical and research laboratories, CTCs can be isolated in observed that the CPK method has high intrapatient reproduci- sufficient numbers and of sufficient purity to allow for their bility and that blood samples remain stable at room temperature molecular characterisation. This technology potentially has a large for up to 72 h before sample processing (Allard et al, 2004). As number of applications in investigating the biology of metastatic many clinical trials involve multiple centres, this sample stability cancer and in drug development in which it can be used to identify provides the opportunity to collect specimens from multiple sites predictive biomarkers, mechanisms of resistance, and facilitate and analyse them in one central location. This is an advantage over pharmacodynamic studies. the recently developed microfluidic-based CTC-chip method. That technology has a very low throughput and a requirement for samples to be run within 2 h of being drawn (Nagrath et al, 2007). ACKNOWLEDGEMENTS These limitations make it currently incompatible for use in multicentre clinical trials. This study is supported by grants from the National Cancer The CPK method clearly does not replace the CEK technique’s Institute Lung SPORE P50CA090578 (PAJ), the Department of well-validated usefulness as a prognostic tool for clinical samples. Defense W81XWH06-1-0303 (PAJ), the twoAM fund (IEK) and the Instead, we feel that the CPK technique’s relatively high yield, Hazel and Samuel Bellin research fund (PAJ). purity, and sample stability make it well suited for use in obtaining CTCs for molecular characterisation, facilitating clinical pheno- Supplementary Information accompanies the paper on British typing and investigational studies. As an example of the former, we Journal of Cancer website (http://www.nature.com/bjc) have used the technique to assess the HER2 gene amplification REFERENCES Adams AA, Okagbare PI, Feng J, Hupert ML, Patterson D, cells versus imaging – predicting overall survival in metastatic breast Gottert J, McCarley RL, Nikitopoulos D, Murphy MC, Soper SA (2008) cancer. Clin Cancer Res 12: 6403–6409 Highly efficient circulating tumor cell isolation from whole Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Matera J, Miller MC, Reuben blood and label-free enumeration using polymer-based micro- JM, Doyle GV, Allard WJ, Terstappen LW, Hayes DF (2004) Circulating fluidics with an integrated conductivity sensor. J Am Chem Soc 130: tumor cells, disease progression, and survival in metastatic breast cancer. 8633– 8641 N Engl J Med 351: 781–791 Allard WJ, Matera J, Miller MC, Repollet M, Connelly MC, Rao C, Hayes DF, Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Miller MC, Matera J, Tibbe AG, Uhr JW, Terstappen LW (2004) Tumor cells circulate Allard WJ, Doyle GV, Terstappen LW (2006) Circulating tumor cells at each in the peripheral blood of all major carcinomas but not in healthy follow-up time point during therapy of metastatic breast cancer patients subjects or patients with nonmalignant diseases. Clin Cancer Res 10: predict progression-free and overall survival. Clin Cancer Res 12: 6897– 6904 4218– 4224 Budd GT, Cristofanilli M, Ellis MJ, Stopeck A, Borden E, Miller MC, Matera J, Hayes DF, Smerage J (2008) Is there a role for circulating tumor cells in the Repollet M, Doyle GV, Terstappen LW, Hayes DF (2006) Circulating tumor management of breast cancer? Clin Cancer Res 14: 3646–3650 & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al Johnson & Johnson O-CD (2008) Instruction for use Circulating Tumor Cell between primary tumors and corresponding circulating tumor cells in Kit (Epithelial). Document MKG1417, Rev. 4 PI Component Code advanced breast cancer patients. Breast Cancer Res Treat 118: 523–530 7800047.0x Riethdorf S, Fritsche H, Muller V, Rau T, Schindlbeck C, Rack B, Janni W, Maheswaran S, Sequist LV, Nagrath S, Ulkus L, Brannigan B, Collura CV, Coith C, Beck K, Janicke F, Jackson S, Gornet T, Cristofanilli M, Pantel K Inserra E, Diederichs S, Iafrate AJ, Bell DW, Digumarthy S, Muzikansky A, (2007) Detection of circulating tumor cells in peripheral blood of patients Irimia D, Settleman J, Tompkins RG, Lynch TJ, Toner M, Haber DA (2008) with metastatic breast cancer: a validation study of the CellSearch Detection of mutations in EGFR in circulating lung-cancer cells. NEngl J system. Clin Cancer Res 13: 920–928 Med 359: 366 –377 Sieuwerts AM, Kraan J, Bolt-de Vries J, van der Spoel P, Mostert B, Martens Meng S, Tripathy D, Shete S, Ashfaq R, Haley B, Perkins S, Beitsch P, JW, Gratama JW, Sleijfer S, Foekens JA (2009) Molecular characteriza- Khan A, Euhus D, Osborne C, Frenkel E, Hoover S, Leitch M, Clifford E, tion of circulating tumor cells in large quantities of contaminating VitettaE,Morrison L,Herlyn D,TerstappenLW, Fleming T, leukocytes by a multiplex real-time PCR. Breast Cancer Res Treat 118: Fehm T, Tucker T, Lane N, Wang J, Uhr J (2004) HER-2 gene amplification 455–468 can be acquired as breast cancer progresses. Proc Natl Acad Sci USA 101: Sleijfer S, Gratama JW, Sieuwerts AM, Kraan J, Martens JW, Foekens JA 9393– 9398 (2007) Circulating tumour cell detection on its way to routine diagnostic Mostert B, Sleijfer S, Foekens JA, Gratama JW (2009) Circulating tumor implementation? Eur J Cancer 43: 2645–2650 cells (CTCs): detection methods and their clinical relevance in breast Vona G, Sabile A, Louha M, Sitruk V, Romana S, Schutze K, Capron F, cancer. Cancer Treat Rev 35: 463– 474 Franco D, Pazzagli M, Vekemans M, Lacour B, Brechot C, Paterlini- Mukohara T, Engelman JA, Hanna NH, Yeap BY, Kobayashi S, Lindeman N, Brechot P (2000) Isolation by size of epithelial tumor cells: a new method Halmos B, Pearlberg J, Tsuchihashi Z, Cantley LC, Tenen DG, Johnson BE, for the immunomorphological and molecular characterization of Janne PA (2005) Differential effects of gefitinib and cetuximab on non- circulating tumor cells. Am J Pathol 156: 57–63 small-cell lung cancers bearing epidermal growth factor receptor Williams A, Lin H, Zheng S, Szmulewitz R, Stadler W, Scher H, Fleisher M, mutations. J Natl Cancer Inst 97: 1185– 1194 Liu J, Nie S, Hawes D, Groshen S, Datar R, Tai Y-C, Cote R (2009) Nagrath S, Sequist LV, Maheswaran S, Bell DW, Irimia D, Ulkus L, Development of a membrane microfilter device for capture and Smith MR, Kwak EL, Digumarthy S, Muzikansky A, Ryan P, Balis UJ, characterization of circulating tumor cells in blood. Proceedings of the Tompkins RG, Haber DA, Toner M (2007) Isolation of rare circulating 100th Annual Meeting of the American Association for Cancer Research, tumour cells in cancer patients by microchip technology. Nature 450: Abstract 2608 1235– 1239 Wulfing P, Borchard J, Buerger H, Heidl S, Zanker KS, Kiesel L, Brandt B Pestrin M, Bessi S, Galardi F, Truglia M, Biggeri A, Biagioni C, Cappadona S, (2006) HER2-positive circulating tumor cells indicate poor clinical outcome Biganzoli L, Giannini A, Di Leo A (2009) Correlation of HER2 status in stage I to III breast cancer patients. Clin Cancer Res 12: 1715–1720 British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Improving the yield of circulating tumour cells facilitates molecular characterisation and recognition of discordant HER2 amplification in breast cancer

Loading next page...
 
/lp/springer-journals/improving-the-yield-of-circulating-tumour-cells-facilitates-molecular-Q40oCmH8fL

References (19)

Publisher
Springer Journals
Copyright
Copyright © 2010 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6605676
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2010) 102, 1495 – 1502 & 2010 Cancer Research UK All rights reserved 0007 – 0920/10 $32.00 www.bjcancer.com Improving the yield of circulating tumour cells facilitates molecular characterisation and recognition of discordant HER2 amplification in breast cancer 1,2 3,5 3,4,5 1 4 3,4,5 3,5 LM Flores , DW Kindelberger , AH Ligon , M Capelletti , M Fiorentino , M Loda , ES Cibas , ,1,3,6 ,1,3,6 * * PA Janne and IE Krop 1 2 Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; Department of Biology, University of Massachusetts, Boston, MA, 3 4 USA; Harvard Medical School, Boston, MA, USA; Center for Molecular Oncologic Pathology, Dana-Farber Cancer Institute, Boston, MA, USA; Department of Pathology, Brigham and Women’s Hospital, Boston, MA, USA BACKGROUND: Circulating tumour cells (CTCs) offer a non-invasive approach to obtain and characterise metastatic tumour cells, but their usefulness has been limited by low CTC yields from conventional isolation methods. METHODS: To improve CTC yields and facilitate their molecular characterisation we compared the Food and Drug Administration- approved CellSearch Epithelial Kit (CEK) to a simplified CTC capture method, CellSearch Profile Kit (CPK), on paired blood samples from patients with metastatic breast (n ¼ 75) and lung (n ¼ 71) cancer. Molecular markers including Human Epidermal growth factor Receptor 2 (HER2) were evaluated on CTCs by fluorescence in situ hybridisation (FISH) and compared to patients’ primary and metastatic cancer. RESULTS: The median cell count from patients with breast cancer using the CPK was 117 vs 4 for CEK (Po0.0001). Lung cancer samples were similar; CPK: 145 cells vs CEK:4 cells (Po0.0001). Recovered CTCs were relatively pure (60–70%) and were evaluable by FISH and immunofluorescence. A total of 10 of 30 (33%) breast cancer patients with HER2-negative primary and metastatic tissue had HER2-amplified CTCs. CONCLUSION: The CPK method provides a high yield of relatively pure CTCs, facilitating their molecular characterisation. Circulating tumour cells obtained using CPK technology demonstrate that significant discordance exists between HER2 amplification of a patient’s CTCs and that of the primary and metastatic tumour. British Journal of Cancer (2010) 102, 1495 – 1502. doi:10.1038/sj.bjc.6605676 www.bjcancer.com & 2010 Cancer Research UK Keywords: circulating tumour cells; breast cancer; lung cancer; HER2; FISH The need for access to samples of metastatic cancer and the number of pure CTCs is necessary for these downstream inherent difficulty in obtaining these using conventional biopsies, applications. has led to interest in alternative sources of metastatic cells. One Circulating tumour cells have become increasingly accepted as such alternative is to analyse circulating tumour cells (CTCs). both an independent prognostic factor and as a marker of Circulating tumour cells are thought to be quite rare in the blood therapeutic response and attempts have been made to standardise of most patients with metastatic disease – it has been estimated their isolation and characterisation (Hayes and Smerage, 2008). that they are present at a frequency of 1 in 10 blood cells (Mostert The automated CellSearch System (Veridex; Warren, NJ, USA) is et al, 2009). Circulating tumour cells provide the potential ability the first method to be clinically approved by the United States to monitor metastatic disease, prognostic information and a Food and Drug Administration to capture and detect CTCs. This means to perform non-invasive molecular interrogation of cancers, system relies on immunomagnetic isolation followed by the use of as these cells can provide a molecular snapshot, or real-time fluorescence microscopy for analysis. Two different but related biopsy, of tumour cells. Circulating tumour cells can also methods of sample preparation, the CellSearch Epithelial Cell Kit potentially be used to determine the mechanisms of acquired (CEK) and the CellSearch Profile Kit (CPK) are currently available. resistance to targeted therapies, which only occur or evolve during Both methods use a positive immunomagnetic selection with anti- the course of drug treatment. The ability to isolate a relatively large epithelial cell adhesion molecule antibodies linked to iron particles to enrich for CTCs. In the CEK system, captured cells are permeabilised and labelled with cytokeratin and CD45-specific *Correspondence: Dr IE Krop; E-mail: ikrop@partners.org or antibodies, and the nuclear stain 4 -6-diamidino-2-phenylindole Dr PA Ja¨nne; E-mail: pjanne@partners.org (DAPI), and are analysed by semi-automated counting of These authors contributed equally to this manuscript appropriately labelled CTCs (Riethdorf et al, 2007). In contrast, Received 15 December 2009; revised 24 March 2010; accepted 29 the CPK procedure does not involve labelling or enumerating the March 2010 cells, but rather only uses the anti-epithelial cell adhesion molecule Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al antibody coupled to iron particles to yield an enriched population (Johnson & Johnson, 2008). Samples processed using the CPK of CTCs, which can be used for molecular analyses. However, in method (per manufacture’s instructions, (Johnson & Johnson, contrast to the CEK method, the CPK procedure is not Food and 2008)) were collected in a dilution buffer (phosphate buffered Drug Administration approved for clinical management. saline, 0.5% bovine serum albumin and 0.1% sodium azide) and The primary limitation of the CEK system is the low yield of immediately transferred to glass slides as cytospin preparations CTCs. In several large clinical trials of patients with metastatic using a ThermoFisher Cytospin 3. The centrifugation was breast cancer, the median number of isolated CTCs was five per performed at 500 g for 5 min with a cytology funnel, thin filter 7.5 ml of blood (Budd et al, 2006; Hayes et al, 2006; Hayes and (Thermo Fisher Scientific, Pittsburgh, PA, USA), and a SuperFrost Smerage, 2008). Although the low numbers of CTCs are sufficient Plus slide (Thermo Fischer Scientific). The cells on the resulting to provide clinical prognostic information, in which a cutoff of X5 cytospin slide were fixed with methanol for IF or immunohis- CTCs per 7.5 ml of blood is used to delineate a poor prognosis tochemistry or fixed with methanol:acetic acid (3 : 1) for FISH. population, the ability to perform molecular analyses of the CTCs Cells were stained with anti-cytokeratin phycoerythrin/DAPI/ is limited. In addition, it has been reported that the degree of CD45-allophycocyanin (Dako, Carpinteria, CA, USA) and then leukocyte contamination using the CEK system is quite high, manually counted using a standard fluorescence microscope which further complicates molecular analyses of CTCs. These (Olympus, Center Valley, PA, USA). Cytokeratin staining was limitations have led to the development of alternative CTC considered positive when cells displayed unequivocal cytoplasmic isolation methods, including technologies based on size filtration, cytokeratin staining in a ring-like pattern (confirming an intact density gradients, and microfluidic techniques, which have cytoplasm) surrounding an intact nucleus. demonstrated increased yields of CTCs (Vona et al, 2000; Nagrath et al, 2007; Adams et al, 2008; Williams et al, 2009). For example, a Cell lines and reagents microfluidic chip-based CTC technology has been developed, The breast cancer cell line SK-BR-3 was obtained from American which isolates relatively high numbers of CTCs and can be used to Type Culture Collection. The EGFR-mutant (del E746_A750) non-invasively detect drug-sensitive and -resistant epidermal NSCLC cell line, HCC827, has been described previously growth factor receptor (EGFR) mutations from non-small cell (Mukohara et al, 2005). The two cell lines were grown in lung cancer (NSCLC) patients (Nagrath et al, 2007; Maheswaran Dulbecco’s Modified Eagle Medium or Roswell Park Memorial et al, 2008). This method uses epithelial cell adhesion molecule- Institute 1640 (CellGrowth, Invitrogen, Life Technologies, specific antibodies that are conjugated to posts within the CTC Carlsbad, CA, USA) respectively, supplemented with 10% foetal chip. Currently this system is limited by the lack of widespread bovine serum. Gefitinib and lapatinib were obtained from commercial availability of the system, the inability to release the CTCs from the sources. Stock solutions of both drugs were prepared in dimethyl microfluidic posts, prolonged processing time (samples are run at –1 sulphoxide and stored at 201C. For in vitro recovery experiments 1–2 ml h ), and the requirement that samples be processed within SKBR3 cells were grown in DMEM with 10% foetal bovine serum, a few hours of collection (Nagrath et al, 2007). –1 10 mgml EGF, and 1.5 mML-glutamine. The HCC287 cells were To address some of the limitations of current CTC systems, we grown in ACL4 (Invitrogen, Life Technologies) with 10% foetal evaluated the CPK system as a method to isolate and study CTCs. bovine serum. For detailed methods see Supplementary methods. Unexpectedly, this method resulted in isolation of 420-fold more CTCs than the CEK method. We further characterised the purity of the isolated CTCs and evaluated them using immunofluorescence Fluorescence in situ hybridisation (FISH) (IF) and fluorescence in situ hybridisation (FISH). Our findings The EGFR/CEP 7 and Human Epidermal growth factor Receptor 2 suggest that a modification in the use of the current CellSearch (HER2)/Centromere Probe 17 (CEP17) bacterial artificial chromo- system can result in a significant increase in the ability to isolate and some (BAC) probes were obtained from Vysis Molecular (Abbot characterise CTCs from patients with breast cancer and NSCLC. Park, IL, USA). BAC mesenchymal-epithelial transition factor (MET) (RP11-95I20) was obtained from CHORI (Children Oakland Research Hospital, Oakland, CA, USA). Labelling was done with MATERIALS AND METHODS the Nick Translation Kit (Vysis, Abbott Molecular, Des Plaines, IL, USA). See Supplementary data for detailed methods of FISH Patients analysis of CTC, and for IF and immunohistochemistry staining Between December 2007 and March 2009 patients with metastatic methods. NSCLC (n¼ 71) or metastatic breast cancer (n¼ 75) were identified from the Thoracic Oncology and Women’s Cancer clinics at the RESULTS Dana Farber Cancer Institute. Characteristics of the patients are described in the Supplementary Table S1. The blood was collected The CPK method isolates a greater number of cells than from each donor into CellSave blood collection tubes (Veridex) or the CEK method ethylenediaminetetraacetic acid (EDTA) tubes, where specified. The blood samples were maintained at room temperature and processed We hypothesised that CTCs may be lost in the additional within a maximum of 72 h after collection. All samples from patients permeabilisation, labelling, and wash steps specific to the CEK with breast cancer were obtained at the time they were initiating a method, particularly if CTCs are more ‘fragile’ than other cell new therapy for the treatment of their disease. All of the samples types. To directly compare the yield of cells isolated using CEK and from patients with NSCLC were obtained while the patient was CPK procedures, we collected two peripheral blood samples from currently under therapy. Control blood samples were drawn from patients with metastatic breast cancer (n ¼ 75) and NSCLC healthy volunteers with no history of malignant disease. All patients (n ¼ 71), and processed them in parallel. Cells obtained using the provided written informed consent, and studies were approved by CPK method were counted manually and were compared with the the Dana Farber Cancer Institute Institutional Review Board (IRB). standard CEK counts obtained using the CellSearch system. The median cell count from patients with breast cancer using the CPK method was 116.5 per 7.5 ml of blood (range 4–2432) compared Sample processing and counting with 4 (range 0–57) using the CEK method (Po0.0001, Figure 1A). Samples were processed using the CEK method, including semi- The median cell count from patients with NSCLC processed using automated counting, according to the manufacture’s instructions the CPK method was 145 per 7.5 ml of blood (range 5–1801) British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK Improving yields of circulating tumour cells LM Flores et al CEK kit semi-automated count CPK kit-manual count 0 2 5 10 25 50 100 200 300 400 500 600 700 800 900 1000 2000 3000 No. of cells CEK semi-automated count CPK-manual count 0 2 5 10 25 50 100 200 300 400 500 600 700 800 900 1000 2000 No. of cells Figure 1 CPK method improves cell yields over CEK method. The blood samples from patients with (A) breast cancer (n ¼ 75) or (B) NSCLC (n ¼ 71) processed in parallel by the CEK method with semi-automated quantification (open columns) or the CPK method with manual quantification (closed columns). compared with 4 (range 0–53) using the CEK procedure (Po0.0001; Approximately 60% of cells isolated by CPK (range: 21–95%) were Figure 1B). We also compared these findings on a per patient basis identified as tumour cells by cytokeratin staining, whereas 6% (Supplementary Figures S1A and B) and demonstrated that for all (range: 1–22%) were identified as white blood cells (WBCs) by patients, the cell count by the CPK method exceeds the cell count by CD45 staining (Figure 2A and data not shown). These percentages the CEK method (breast cancer: R ¼ 0.004919; lung cancer: correspond to a median absolute number of cytokeratin positive R ¼ 0.006952). To determine whether the observed yield differences cells of 280 per sample (range 20–20 000) and a median of 34 between the CEK and CPK methods are due to differences in the (range 20–476) CD45-positive cells per sample. The remaining method of cell counting (manual vs semi-automatic), we compared 32% (range: 5–68%) of cells stained with DAPI only and could CTC counts obtained using the automated CEK method with that of not be classified as either tumour cells or leukocytes. As a a manual cell count of the same sample. The manual CEK count was complementary approach, we used fluorescence-activated cell performed by removing the enumerated samples from the ‘MagNest’ sorting on 11 patients (eight NSCLC and three breast cancer) to cartridges, processing them as cytospin preparations and counting compare the populations of CTCs with WBCs (Figure 2B). The them manually. In patients (n¼ 75 patients; 100 samples) with results were similar for cytokeratin (mean positive: 61%; range breast cancer (Supplementary Figure S1E), the median cell count 35–77%), except the frequency of WBCs was estimated to be even using the semi-automated CEK method was 4 per 7.5 ml of blood lower (mean positive: 1%; range 0–7%) than by IF (compare (range 0–57) compared with 5 per 7.5 ml of blood (range 0–61) Figure 2A and B). when the CEK processed specimens were counted manually As an additional method to evaluate the nature of the cells (P¼ NS). Similarly, in NSCLC patients (n¼ 71 patients; 100 isolated using the CPK method, we performed FISH using probes samples), the median cell count using the semi-automated CEK for HER2, EGFR, and MET. Given that a mean of 32% of cells had denuded membranes (DAPI-only cells; Figure 2A) and thus could method was four per 7.5 ml of blood (range 0–53) compared with five per 7.5 ml of blood (range 5–67) when the CEK processed not be definitely stained for cytokeratin, the identification of specimens were counted manually (P¼ not significant; Supplemen- aneuploidy and/or gene amplification in a high fraction of the cells tary Figure S1F). We further compared these findings on a per would strongly suggest that some or all of these cells were CTCs patient basis (Supplementary Figures S1C and S1D) and observed a and not WBCs (Figures 3A and B). Among the CTCs from patients strong correlation between the cell counts obtained using the semi- with breast cancers clinically defined as HER2 positive from automated CEK method with the manual CEK method (breast analyses of their primary tumours (FISH þ and/or IHC 3 þ by 2 2 cancer specimens: R ¼ 0.9930; lung cancer specimens: R ¼ 0.7200). local testing, n ¼ 24), 72% of all captured cells had a HER2 gene Together, our findings suggest that the CPK method isolates a copy number X4. This threshold has been used as an indicator of greater number of cells than the CEK method and that these HER2 amplification in previous studies (Meng et al, 2004). Only findings are not due to differences in the method of cell counting. 3% of the cells had two copies of HER2. In an additional six patients with clinically defined HER2-negative breast cancers, only 5% of captured cells had X4 copies of HER2 (Figure 3D). In The CPK method isolates a highly enriched population contrast, the majority (95%) of cells in those patients had a HER2 of CTCs gene copy number of 2 (Figure 3D). Similarly, among samples One explanation for the improved yields observed with the CPK from NSCLC patients (n ¼ 30), 53% had X4 copies of EGFR and method is that the isolated cells are not all CTCs but instead a 13% had X4 copies CEP 7 (Figure 3C). Only a minority of cells mixture of CTCs and contaminating leukocytes. This determina- (5%) had two copies of EGFR and CEP 7. tion is critical if the isolated cells are to be used for any subsequent As an additional control, we analysed 40 samples from healthy analyses. To determine the composition of the cells isolated using volunteers with no history of malignant disease. No tumour cells the CPK method, we performed IF using anti-CD45 and anti- were found with either the CEK semi-automated enumeration or cytokeratin (AE1/AE3) antibodies in 30 NSCLC and 30 breast with the CPK manual count (Supplementary Figure S2A). In cancer patients. Three distinct cell populations were observed: (1) samples processed using the CPK method, the median total CD45 positive, cytokeratin negative, (2) CD45 negative, cytokeratin number of cells was five per 7.5 ml of blood (range: 0–50). These positive, and (3) cells that were negative for both CD45 and cells were positive for CD45 and negative for cytokeratin using IF, cytokeratin but stained with the nuclear stain DAPI (Figure 2C). with only two gene copies of CEP7/EGFR or HER2 as detected by & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 No. of samples No. of samples Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al % Cytokeratin positive (Supplementary Figure S3C) with a low coefficient of variability % DAPI only (CV: 9.7%). We also evaluated the impact of time on the ability to % CD45 positive isolate CTCs using the CPK method. As many clinical trials involve multiple centres, it is important to determine the stability of unprocessed samples to determine whether the yield of CTCs declines over time. For these studies, we collected three CellSave (LLC a Johnson&Johnson company, Raritan, NJ, USA) or EDTA 0 tubes from seven NSCLC patients and incubated the samples at room 123456789 1011121314151617 18 19 20 21 22 23 24 25 26 27 28 29 30 temperature for 24–72 h before processing by the CPK method. We Patient specifically evaluated the collection of blood into both CellSave and % CD45 positive 100 EDTA tubes to determine the impact of the fixative (present in the % Cytokeratin positive CellSave tubes; absent in the EDTA tubes) on the impact of cell recovery over time. We isolated similar numbers of cells using both the CellSave (Supplementary Figure S3A) and EDTA tubes (Supplementary Figure S3B). There was no significant decline (coefficient of variation (CV) 9.1% and CV 6.5%) in the number of cells isolated with either method over the 72-h period and no decline in the number of cells isolated from CellSave tubes incubated at 123456789 10 11 room temperature for up to 144 h (Supplementary Figure S3A and Patient data not shown). However, there was significant decline in the number of cells isolated from EDTA tubes incubated at room temperature for X96 h before processing by the CPK method. Treatment affects yield of cell recovery Previous studies have found that when tumour cell lines are spiked into blood and isolated using the CEK method, the rate of cell recovery is high (85–95%; Riethdorf et al, 2007). However, the actual CTC yields from cancer patients using the CEK method are limited (Sleijfer et al, 2007). We hypothesised that one explanation for the low yields seen in clinical practice is that CTCs are more ‘fragile’ than cancer cell lines either due to apoptosis or damage acquired during transit through the bloodstream and, therefore, are not captured as efficiently with the anti-epithelial cell adhesion molecule antibodies and/or are lost in the post-capture washing and labelling steps. To investigate this possibility, we examined breast (SKBR3; HER2 amplified) and lung cancer (HCC827; EGFR Figure 2 The CPK method isolates a highly enriched population of del E746_750) cell lines either unprocessed or spiked into normal CTCs. Percentage of total cells captured by the CPK method from patients blood, then processed with the CPK or CEK methods. Cells were with NSCLC, staining for cytokeratin (AE1/AE3), CD45, or DAPI nuclear pretreated with the EGFR/HER2 inhibitor lapatinib (SKBR3) or the stain alone by (A) immunofluorescence or (B) FACS. Similar results were EGFR inhibitor gefitinib (HCC827) to induce apoptosis, or were seen with samples from patients with breast cancer. (C) Representative untreated. The cells were then stained with Ki67 or Apoptosis immunofluorescence image of CPK-captured cells from patient with Detection Using Terminal Transferase and Biotin-16-dUTP (TUNEL) NSCLC, labelled with cytokeratin (green), DAPI (blue), and CD45 (red). to evaluate proliferating and apoptotic cells, respectively. As expected, unprocessed cells demonstrated increased rates of apoptosis and decreased proliferation when treated with inhibitor FISH analysis, consistent with normal WBC. Collectively, these (Figure 4C). Interestingly, no apoptotic cells were noted in the data indicate that that majority of cells captured by the CPK samples processed with the CEK method in either untreated or system are in fact CTCs with only a minority of cells being WBCs. treated samples, although the treated cells that were captured by All of the samples described were prepared using the low-speed that method did have decreased levels of Ki67 staining, consistent cytospin method using a single cytology funnel with thin filter with the effect of the inhibitor (Figure 4A). In samples processed (see Materials and Methods section). We also evaluated other with the CPK method, a small percentage of TUNEL-positive cells techniques for preparing slides, including high-speed cytospin, were recovered in the untreated samples, which increased with different funnel types, a thin prep or cell block, and applying the inhibitor treatment, although not to the level of unprocessed samples to slides directly as a smear. In a comparative analysis, samples (Figure 4B). the low-speed cytospin method produced the highest yield, with We also evaluated the frequency of Ki67- and TUNEL-staining the least contamination, least cell damage, and the least slide-to- cells isolated from seven NSCLC patients using the CPK method. slide variation in comparison to other various methods (data not The frequency of TUNEL-staining cells was low (mean 1.9%; range shown). 0–6.4%) similar to what was observed in the SKBR3 and NSCLC cell lines (Figure 4D). The frequency of Ki67-staining cells was also lower (mean 0.7%; range 0.3–1.2%) in the CTCs from cancer The CPK method has low intrapatient variability and patients compared with the levels seen with cell lines (Figure 4D). samples are stable for at least 72 h When parallel samples were processed using the CEK method, no We next examined whether the number of cells isolated from TUNEL-positive cells were detected, but the fraction of Ki67- patients with cancer using the CPK method was reproducible and positive cells was numerically higher than in the CPK processed stable over time. We first obtained three samples from each of samples (Figure 4E). This observation is consistent with the seven patients with NSCLC at the same time point and processed hypothesis that potentially more ‘fragile’ apoptotic and non- them in parallel. The number of cells isolated was similar proliferating cells are disproportionably lost by the more British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK % of cells % of cells Improving yields of circulating tumour cells LM Flores et al Lymphocytes CTC % of cells with ≥4 of EGFR % of cells with ≥4 of CEP7 100 % of cells with ≥4 of MET 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Patient % of cells with 2 of HER2 % of cells with ≥4 of HER2 % of cells with ≥6 of HER2 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930 Patient HER2+ HER2– Figure 3 FISH analysis confirms that samples processed by the CPK method have a low percentage of contaminating normal cells. Representative FISH images of cells processed by CPK method, (A) lymphocyte with two copies of CEP7 (green), EGFR (red), and MET (blue), (B) CTC with amplified EGFR and MET. (C) Percentage of total cells captured by the CPK method from patients with NSCLC with X4 copies of EGFR, CEP7, and MET per nucleus. (D) Percentage of total cells captured by the CPK method from patients with clinically defined HER2-positive (patient number 1–24) or HER2-negative (patient number 25 –30) breast cancer, with the indicated copies of HER2 per nucleus. processing intensive CEK method. However, because of the small of their primary tumour (Meng et al, 2004; Wulfing et al, 2006; number of cells recovered, it was not possible to make a definitive Pestrin et al, 2009). No data exists examining the relationship comparison. Together with the data obtained from the cell lines, between the CTCs’ HER2 status and that of metastatic tissue. We these findings suggest that CTCs undergoing apoptosis in the hypothesised that the ability of the CPK technology to provide blood stream, either spontaneously or as result of treatment effect, relatively high numbers of CTCs and allow robust FISH analysis may be too fragile to survive the capture step and/or are being would facilitate the determination of the relationship between the eliminated during subsequent specimen processing. This effect HER2 status of a patients primary breast cancer, metastatic lesions, seems to be more pronounced with the CEK than CPK method and and CTCs. To test this hypothesis, we assessed the presence of may help explain the increased CTC yields seen with the latter HER2 amplification by FISH, normalised for CEP 17, on CTCs technique. These observations may also help to explain why from 75 women with breast cancer for whom the HER2 status of alternative CTC-capture techniques, such as the microfluidic chip the primary cancer and a metastatic biopsy sample was available or microfiltration systems that minimise the number of labelling (patient characteristics available in Supplementary Table S1). and wash steps are associated with an increase in CTC yield over For patients with HER2 gene amplification of their primary the CEK method. breast cancer, the degree of concordance with the HER2 status of their CTCs was high, with only 1 of 45 (2%) patients demonstrating loss of HER2 amplification in the corresponding CTCs (Table 1). In Characterisation of HER2 amplification of CTCs by that one patient with discordance, the metastatic biopsy was HER2 FISH analysis amplified, reflecting the primary cancer. In contrast, for those Several small studies have reported varying degrees of discordance patients in whom the primary breast cancer was HER2 negative, between the HER2 status of a patient’s breast cancer CTCs and that significant discordance between the primary, metastatic biopsy, & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 % of cells % of cells Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al 100 100 Tunel Tunel Table 1 Relationship between HER2 amplification of primary cancer and Ki67 Ki67 80 80 that of CTC 60 60 Total No. Patients Patients 40 40 Primary of with with 20 20 tumor patients HER2CTCs HER2+CTCs Discordance 0 0 HER2+ 45 1 44 2% HER2 30 20 10 33% Abbreviations: CTC ¼ circulating cancer cells; HER2 ¼ human epidermal growth CEK CPK factor receptor 2. HER2 positive defined as ratio of HER2/CEP17 X2.0 by FISH. 100 10 Tunel Tunel Ki67 80 Ki67 Cyclin D1 evaluate membrane staining of EGFR, pEGFR, HER2 and pHER in 20 CTCs from 20 patients with clinically HER2-positive breast cancer 0 (Supplementary Figure S5B). Similar to the CTCs isolated from NSCLC patients, we were able to detect EGFR, pEGFR, HER2, and 123456789 10 Patient pHER using IF in the CTCs and the frequency of staining was less than in the SKBR3 tumour cell line (Supplementary Figure S5B). In vitro propagation The standard CellSave tubes contain a No processing fixative that prevents the isolation of viable tumour cells. Having observed that CTCs can be efficiently isolated from the Method Number of KI67+ KI67+ KI67+ blood drawn in EDTA tubes that do not contain a fixative evaluable cells/total cells/total cells/sample samples cells (mean, %) cells (range, %) (Supplementary Figure S3B), we explored whether we could isolate SKBR3 and HCC827 cells using EDTA tubes and grow them CEK breast 28 21.1 10 – 44 1–6 in vitro. We spiked a range (5–1000) of SKBR3 or HCC827 cells CPK breast 75 7.1 1.1 – 13.2 2 – 480 CEK lung 21 32 10 – 64 2–8 into the normal blood drawn into EDTA tubes and processed them CPK lung 70 4.4 1.6 – 16 1–113 using the CPK method. We were able to isolate viable SKBR3 and HCC827 cells (Supplementary Figure S5) and propagate them Figure 4 Apoptotic CTCs are less effectively captured by CEK or CPK in vitro. However, this required a minimum input of 50 cells to be methods. SKBR3 (HER2 breast cancer) or HCC827 (EGFR mutant NSCLC) cells treated with vehicle or tyrosine kinase inhibitor (SKBR3: 1 mM processed (Supplementary Figure S5). Optimal growth occurred lapatinib and HCC827; 1 mM gefitinib) for 24 h and processed with (A) in the presence of enriched medium and cell adhesion matrix CEK method. (B) CPK method or (C) smeared directly on slide without (data not shown). processing. Plots depict percentage of cells ( 1 s.d.) staining for Ki67 (open bars, proliferation marker) or TUNEL (closed bars, apoptosis marker) by immunohistochemistry. (D) Percentage of CPK-processed DISCUSSION CTCs from patients with NSCLC staining positive for KI67 and TUNEL. (E) Comparison of KI67 expression in CTCs recovered by CEK or CPK One of the most important issues limiting the use of CTCs as an methods. alternative to invasive biopsies has been their relatively low yield. Using the CellSearch System (Veridex), the most widely available and CTC was observed. A total of 10 (33%) of these 30 patients technology, the median yield is reported to be in the range of 1 cell with HER2-negative primary cancers had HER2-positive CTCs per ml of blood screened (Cristofanilli et al, 2004). Our results with by FISH analysis. The median HER2/CEP17 ratio of the CTCs in patients with breast and lung cancer using the CellSearch CEK these patients was 7.1. Interestingly, in 9 of the 10 patients with platform are consistent with that finding. However, two seemingly discordance, the metastatic biopsy specimen was HER2 negative. paradoxical observations prompted us to explore this technology further. The first is that several studies using the CellSearch platform have demonstrated that the instrument was highly Evaluation of biologic properties of CTCs isolated using efficient at capturing immortalised breast cancer cell lines spiked the CPK method into the human blood, with yields of X85% of the input cells Expression of receptor tyrosine kinases The CPK method is able (Allard et al, 2004). We have replicated those studies with similar to isolate a highly enriched population of CTCs and greater results; a capture rate of 92% (data not shown). The second numbers of CTCs than the CEK method. These findings provide observation was the finding that the new CTC–chip microfluidic the opportunity to initiate the molecular characterisation of CTCs. technology reported mean CTC counts from patients with breast In addition to FISH (Figure 3), we evaluated the membrane cancer of 79 cells per ml of blood (Nagrath et al, 2007). If the expression of total and phosphorylated HER2 and EGFR using IF CellSearch technology is already highly efficient at capturing CTCs on the breast and lung cancer cell lines spiked into blood and in model systems using cancer cell lines, how can another recovered with the CPK (Supplementary Figure S4A). The IF technology report 10–100-fold higher yields of CTCs from patient analysis was superior to chromogenic immunohistochemical samples? We postulated that human CTCs, particularly those that staining, as the immunohistochemical substrate tended to bind have been subject to anti-cancer treatment, may be significantly non-specifically to ferroparticles, and resulted in high background more ‘fragile’ than the cancer cell lines used in the initial staining (data not shown). We evaluated the expression of EGFR evaluation of the CellSearch platform. Owing to this fragility, the and pEGFR from CTCs in seven NSCLC patients (Supplementary multiple processing steps involved in the CEK system, including Figure S4C). We were able to detect staining in all samples additional washes and labelling steps, could cause degradation and although the percent of cells staining for either EGFR or pEGFR loss of captured CTCs. was lower than in the HCC827 cells that had been spiked into We hypothesised that the CPK system, which involves blood and processed using the CPK method. We also used IF to significantly fewer processing steps, may provide improved yields British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK SKBR3 vehicle SKBR3 lapatinib HCC827 vehicle HCC827 iressa SKBR3 vehicle Untreated SKBR3 cells SKBR3 lapatinib Treated SKBR3 cells HCC827 vehicle Untreated HCC827 cells Treated HCC827 cells HCC827 iressa % of cells % of cells % of cells % of cells Improving yields of circulating tumour cells LM Flores et al of CTCs from patient samples. Indeed, we observed that the CPK status of CTCs by FISH and compared it with both the patient’s consistently provided a significantly greater yield of CTCs than primary breast cancer and tissue from matched distant metastatic that obtained with the CEK platform. The median difference was sites. In patients with clinically HER2-positive primary and 29-fold in breast cancer samples and 36-fold in lung cancer metastatic cancers, virtually all (98%) also had HER2-amplified samples (Figure 1). In addition, virtually all of the breast and lung CTCs. In contrast, in patients with clinically HER2-negative patients had detectable CTCs using the CPK method, whereas our primary cancers, we observed a significant number (33%) of results with the CEK system demonstrate only 72% of patients had discordant cases in which a patient’s CTCs had clear amplification detectable CTCs, consistent with the published results (Hayes et al, of the HER2 locus, despite an absence of HER2 amplification in 2006). The improvement in yield observed with the CPK was not the primary cancer. This result is consistent with the data from simply due to the differences between the CEK’s semi-automated several other small studies demonstrating discordant HER2 counting system and the CPK’s manual counts (Supplementary expression in CTCs and together provide the rationale for a Figures S1E and F). When the CTCs from both methods were clinical trial of a HER2-directed therapy in such patients to counted manually, the differences persisted. Instead, our observa- definitively test the clinical relevance of this observation (Meng tion that when cell lines were pretreated with apoptosis-inducing et al, 2004; Wulfing et al, 2006; Pestrin et al, 2009). To the best of agents (Figure 4), the yield of CTCs dropped dramatically, suggests our knowledge, our study is the first to also examine the that CTC fragility, coupled with the more intensive processing relationship between HER2 amplification in CTCs and that of inherent in the CEK method, likely accounts for the observed metastatic biopsy tissue. In six out of seven patients with HER2- differences between the CEK and CPK methods. negative primary cancers and HER2-amplified CTCs, the meta- Importantly, we also found that the purity of the CTCs obtained static biopsy tissue was HER2 negative, matching the primary from the CPK method was quite high, typically in the 60–70% range. cancer rather than the CTCs. This observation suggests that either This was somewhat unexpected as discussions of the CellSearch CTCs represent a separate population of cells distinct from that system in the literature cite purities in the 0.01–0.1% range, due to which makes up the bulk of the metastatic disease, or that the contaminating leukocytes. In our studies with the CPK system, the cancer developed HER2 amplification in the interval between number of contaminating leukocytes, using either patient samples or the patients’ metastatic biopsy and the CTC collection. If the latter blood from normal volunteers, and assessed using multiple explanation were correct, one might expect to see a correlation independent methods, (IF, fluorescence-activated cell sorting), between the time interval from biopsy to CTC collection and the ranged from 20 to 476, and was always less than 25% of total cells. likelihood that the CTC would gain HER2 amplification. Although The one study we identified describing the number of contaminating we did not observe such a trend (data not shown), our sample size leukocytes observed with the CPK method reported a range of 60– is too small to make any conclusions about the mechanism of this 929 CD45-positive cells per sample (Sieuwerts et al, 2009). This study discordance. did not report the percentage of CTCs vs leukocytes in their paper, so In summary, we have demonstrated that using the CellSearch their results cannot be compared with our data in that regard. CPK assay, an instrument that is currently available in hundreds As has previously been demonstrated with the CEK system, we of clinical and research laboratories, CTCs can be isolated in observed that the CPK method has high intrapatient reproduci- sufficient numbers and of sufficient purity to allow for their bility and that blood samples remain stable at room temperature molecular characterisation. This technology potentially has a large for up to 72 h before sample processing (Allard et al, 2004). As number of applications in investigating the biology of metastatic many clinical trials involve multiple centres, this sample stability cancer and in drug development in which it can be used to identify provides the opportunity to collect specimens from multiple sites predictive biomarkers, mechanisms of resistance, and facilitate and analyse them in one central location. This is an advantage over pharmacodynamic studies. the recently developed microfluidic-based CTC-chip method. That technology has a very low throughput and a requirement for samples to be run within 2 h of being drawn (Nagrath et al, 2007). ACKNOWLEDGEMENTS These limitations make it currently incompatible for use in multicentre clinical trials. This study is supported by grants from the National Cancer The CPK method clearly does not replace the CEK technique’s Institute Lung SPORE P50CA090578 (PAJ), the Department of well-validated usefulness as a prognostic tool for clinical samples. Defense W81XWH06-1-0303 (PAJ), the twoAM fund (IEK) and the Instead, we feel that the CPK technique’s relatively high yield, Hazel and Samuel Bellin research fund (PAJ). purity, and sample stability make it well suited for use in obtaining CTCs for molecular characterisation, facilitating clinical pheno- Supplementary Information accompanies the paper on British typing and investigational studies. As an example of the former, we Journal of Cancer website (http://www.nature.com/bjc) have used the technique to assess the HER2 gene amplification REFERENCES Adams AA, Okagbare PI, Feng J, Hupert ML, Patterson D, cells versus imaging – predicting overall survival in metastatic breast Gottert J, McCarley RL, Nikitopoulos D, Murphy MC, Soper SA (2008) cancer. Clin Cancer Res 12: 6403–6409 Highly efficient circulating tumor cell isolation from whole Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Matera J, Miller MC, Reuben blood and label-free enumeration using polymer-based micro- JM, Doyle GV, Allard WJ, Terstappen LW, Hayes DF (2004) Circulating fluidics with an integrated conductivity sensor. J Am Chem Soc 130: tumor cells, disease progression, and survival in metastatic breast cancer. 8633– 8641 N Engl J Med 351: 781–791 Allard WJ, Matera J, Miller MC, Repollet M, Connelly MC, Rao C, Hayes DF, Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Miller MC, Matera J, Tibbe AG, Uhr JW, Terstappen LW (2004) Tumor cells circulate Allard WJ, Doyle GV, Terstappen LW (2006) Circulating tumor cells at each in the peripheral blood of all major carcinomas but not in healthy follow-up time point during therapy of metastatic breast cancer patients subjects or patients with nonmalignant diseases. Clin Cancer Res 10: predict progression-free and overall survival. Clin Cancer Res 12: 6897– 6904 4218– 4224 Budd GT, Cristofanilli M, Ellis MJ, Stopeck A, Borden E, Miller MC, Matera J, Hayes DF, Smerage J (2008) Is there a role for circulating tumor cells in the Repollet M, Doyle GV, Terstappen LW, Hayes DF (2006) Circulating tumor management of breast cancer? Clin Cancer Res 14: 3646–3650 & 2010 Cancer Research UK British Journal of Cancer (2010) 102(10), 1495 – 1502 Molecular Diagnostics Molecular Diagnostics Improving yields of circulating tumour cells LM Flores et al Johnson & Johnson O-CD (2008) Instruction for use Circulating Tumor Cell between primary tumors and corresponding circulating tumor cells in Kit (Epithelial). Document MKG1417, Rev. 4 PI Component Code advanced breast cancer patients. Breast Cancer Res Treat 118: 523–530 7800047.0x Riethdorf S, Fritsche H, Muller V, Rau T, Schindlbeck C, Rack B, Janni W, Maheswaran S, Sequist LV, Nagrath S, Ulkus L, Brannigan B, Collura CV, Coith C, Beck K, Janicke F, Jackson S, Gornet T, Cristofanilli M, Pantel K Inserra E, Diederichs S, Iafrate AJ, Bell DW, Digumarthy S, Muzikansky A, (2007) Detection of circulating tumor cells in peripheral blood of patients Irimia D, Settleman J, Tompkins RG, Lynch TJ, Toner M, Haber DA (2008) with metastatic breast cancer: a validation study of the CellSearch Detection of mutations in EGFR in circulating lung-cancer cells. NEngl J system. Clin Cancer Res 13: 920–928 Med 359: 366 –377 Sieuwerts AM, Kraan J, Bolt-de Vries J, van der Spoel P, Mostert B, Martens Meng S, Tripathy D, Shete S, Ashfaq R, Haley B, Perkins S, Beitsch P, JW, Gratama JW, Sleijfer S, Foekens JA (2009) Molecular characteriza- Khan A, Euhus D, Osborne C, Frenkel E, Hoover S, Leitch M, Clifford E, tion of circulating tumor cells in large quantities of contaminating VitettaE,Morrison L,Herlyn D,TerstappenLW, Fleming T, leukocytes by a multiplex real-time PCR. Breast Cancer Res Treat 118: Fehm T, Tucker T, Lane N, Wang J, Uhr J (2004) HER-2 gene amplification 455–468 can be acquired as breast cancer progresses. Proc Natl Acad Sci USA 101: Sleijfer S, Gratama JW, Sieuwerts AM, Kraan J, Martens JW, Foekens JA 9393– 9398 (2007) Circulating tumour cell detection on its way to routine diagnostic Mostert B, Sleijfer S, Foekens JA, Gratama JW (2009) Circulating tumor implementation? Eur J Cancer 43: 2645–2650 cells (CTCs): detection methods and their clinical relevance in breast Vona G, Sabile A, Louha M, Sitruk V, Romana S, Schutze K, Capron F, cancer. Cancer Treat Rev 35: 463– 474 Franco D, Pazzagli M, Vekemans M, Lacour B, Brechot C, Paterlini- Mukohara T, Engelman JA, Hanna NH, Yeap BY, Kobayashi S, Lindeman N, Brechot P (2000) Isolation by size of epithelial tumor cells: a new method Halmos B, Pearlberg J, Tsuchihashi Z, Cantley LC, Tenen DG, Johnson BE, for the immunomorphological and molecular characterization of Janne PA (2005) Differential effects of gefitinib and cetuximab on non- circulating tumor cells. Am J Pathol 156: 57–63 small-cell lung cancers bearing epidermal growth factor receptor Williams A, Lin H, Zheng S, Szmulewitz R, Stadler W, Scher H, Fleisher M, mutations. J Natl Cancer Inst 97: 1185– 1194 Liu J, Nie S, Hawes D, Groshen S, Datar R, Tai Y-C, Cote R (2009) Nagrath S, Sequist LV, Maheswaran S, Bell DW, Irimia D, Ulkus L, Development of a membrane microfilter device for capture and Smith MR, Kwak EL, Digumarthy S, Muzikansky A, Ryan P, Balis UJ, characterization of circulating tumor cells in blood. Proceedings of the Tompkins RG, Haber DA, Toner M (2007) Isolation of rare circulating 100th Annual Meeting of the American Association for Cancer Research, tumour cells in cancer patients by microchip technology. Nature 450: Abstract 2608 1235– 1239 Wulfing P, Borchard J, Buerger H, Heidl S, Zanker KS, Kiesel L, Brandt B Pestrin M, Bessi S, Galardi F, Truglia M, Biggeri A, Biagioni C, Cappadona S, (2006) HER2-positive circulating tumor cells indicate poor clinical outcome Biganzoli L, Giannini A, Di Leo A (2009) Correlation of HER2 status in stage I to III breast cancer patients. Clin Cancer Res 12: 1715–1720 British Journal of Cancer (2010) 102(10), 1495 – 1502 & 2010 Cancer Research UK

Journal

British Journal of CancerSpringer Journals

Published: May 11, 2010

There are no references for this article.