|Home | About | Journals | Submit | Contact Us | Français|
Circulating Tumor Cells (CTCs) represent a “liquid biopsy of the tumor” which might allow real‐time monitoring of cancer biology and therapies in individual patients. CTCs are extremely rare in the blood stream and their analysis is technically challenging.
The CellSearch® system provides the enumeration of CTCs with prognostic significance in patients with metastatic breast cancer (mBC), but it does not allow their molecular characterization, which might be useful to identify therapeutically relevant targets for individualized treatment. Combining the CellSearch® and DEPArray™ technologies allows the recovery of single CTCs as a pure sample for molecular analysis.
The purpose of the study was to investigate the heterogeneity of PIK3CA mutational status within single CTCs isolated from individual mBC patients.
CTCs were enriched and enumerated by CellSearch® in blood samples collected from 39 mBC patients. In 20 out of 39 patients enriched samples with ≥5 CTCs were sorted using DEParray™ to isolate single CTCs or pools of CTCs to be submitted to Whole Genome Amplification (WGA) before sequencing analysis. In 18 out of 20 patients, it was possible to perform PIK3CA sequencing on exons 9 and 20.
Twelve subjects were wild type (wt) for the PIK3CA gene. PIK3CA status could also be assessed in pools of CTCs in seven of these patients, with consistent wt status found. Six patients (33%) had a PIK3CA mutation identified. In 2 of the six patients, molecular heterogeneity was detected when mutational analysis was performed on more than one single CTC, including one patient with loss of heterozygosity on both single and pooled CTCs, and one patient with three different PIK3CA variants on single CTCs but PIK3CA wt status on pooled CTC samples. In six out of the 18 cases PIK3CA status was also evaluable on a primary tumor sample. In one of the six cases a discordance in PIK3CA status between the primary (wild‐type) and the matched CTC (exon 20 mutation) was observed.
This study demonstrates the feasibility of a non‐invasive approach based on the liquid biopsy in mBC patients.
Moreover, our data suggest the importance of characterizing CTCs at the single cell level in order to investigate the molecular heterogeneity within cells from the same patient.
Considerable progress has been made towards elucidating the basic biology of primary cancers, however, the molecular characterization of metastatic disease, which generally occurs months or years after primary tumor excision, remains limited. Therapeutic decisions in patients with metastatic cancer continue to be based largely on biomarker expression in the primary tumor, despite frequent discordance in biological characteristics between primary and metastatic lesions (Amir et al., 2008). When patients with disseminated cancer undergo re‐biopsy of metastatic disease, this generally happens for a single lesion even if multiple foci are present, with the majority of metastatic lesions not biopsied due to anatomic inaccessibility and/or associated morbidity of the procedure. Thus, the biological characterization of multiple metastatic sites is rarely performed, although recent data suggest that the selection of therapies according to the alterations of the metastatic lesions, rather than primary tumors, could be beneficial for the patient (Flaherty et al., 2012; Gerlinger et al., 2012; Kwak et al., 2010).
Alternatively, CTCs may offer a readily accessible means of evaluating the biology of metastatic cells, providing a non‐invasive approach that could potentially identify drug sensitivity and resistance‐associated markers, guiding therapeutic decisions. For this reason, the evaluation of CTCs is often referred to as a “liquid biopsy” (Alix‐Panabières and Pantel, 2013). While being a promising approach, CTC analysis, as a means of assessing the biological characteristics of metastatic disease, involves challenging technical aspects and requires further clinical validation.
The enumeration of CTCs, as performed by the CellSearch® System (Veridex LLC, Raritan, NJ, USA), has FDA‐approved clinical utility as a prognostic marker in patients with metastatic breast, colon and prostate cancers (Cristofanilli et al., 2005; Cohen et al., 2008; de Bono et al., 2008). Using this platform, the phenotypic characterization of therapeutically or biologically relevant biomarkers expressed by CTCs has been shown to be feasible (Pestrin et al., 2012; Smerage et al., 2013). However, the presence of leucocytes in CTC enriched samples and the low number of isolated CTCs substantially limits subsequent molecular analyses (Sieuwerts et al., 2009, 2011).
The recently developed DEPArray™ system (Di‐Electro‐Phoretic Array system; Silicon Biosystems, Bologna Italy) is based on an electrokinetic principle, called dielectrophoresis (DEP), which can be used to trap cells in DEP “cages” and then move the trapped cells by manipulating the applied electric field pattern (Fuchs et al., 2006). DEPArray™ combines imaging technologies with the ability to manipulate and recover individual cells from a heterogenous sample, providing pure material for further analyses, including the molecular characterization of single CTCs (Fabbri et al., 2013; Peeters et al., 2013).
To date, the genetic variability among different single CTCs within the same sample has not been extensively evaluated, predominantly due to the technical challenges involved in this approach. An advantage of single cell analysis is that it allows the evaluation of heterogeneity of biological characteristics, including genetic variability of the CTC population for an individual patient (Wills et al., 2013; De Souza, 2013; Shapiro et al., 2013).
The clonal expansion theory describes cancers arising as a result of the sequential acquisition of driver mutations. In breast cancer PIK3CA has been identified as a gene harboring driver mutations. (Stephens et al., 2012).
After the TP53 suppressor gene, the PIK3CA oncogene is the most frequently mutated gene in human breast cancer; mutations are observed in 20–40% of cases (Saal et al., 2005; Stemke‐Hale et al., 2008), with the majority being activating mutations. As tumors harboring PIK3CA activating mutations have constitutive activation of the PI3K signaling cascade, they are predicted to be sensitive to agents targeting this pathway. Moreover, agents targeting components of the PI3K pathway have shown activity in preclinical models (Serra et al., 2008) and clinical trials (Baselga et al., 2012; Hudes et al., 2007; Yao et al., 2011). The allosteric inhibitors of mTOR, everolimus and temsirolimus, are now FDA approved for the treatment of advanced renal, breast, and pancreatic neuroendocrine cancers. However, despite some success with this therapeutic strategy, the benefit in unselected patients remains suboptimal. Furthermore, the prognostic significance of PIK3CA mutations is not well understood, with recent reports that PIK3CA activating mutations may in fact confer improved outcomes in luminal breast cancer subtypes (Gonzalez‐Angulo and Blumenschein, 2013; Loi et al., 2013).
In this study, we aimed at investigating the heterogeneity of PIK3CA mutational status within single CTCs isolated from individual mBC patients. In order to reach this objective the feasibility of a protocol that combines the CellSearch® and DEPArray™ to molecularly characterize single tumor cells was evaluated on mBC patient's single CTCs. We studied also pools of CTCs in order to investigate the importance of performing analyses at the single cell level.
MDA‐MB‐231 cells were obtained from Di.V.A.L. Toscana (Laboratory for Drug Validation and Antibody production, University of Florence) and maintained in DMEM (Lonza, Basel, Switzerland) + 10% FBS (PAA Laboratories, Austria) at 37 °C in 5% CO2.
Cells were detached from plastic surface by Accutase (PAA Laboratories) to better preserve membrane antigens. A known number of MDA‐MB‐231 cells (n = 130) were spiked in blood from a healthy volunteer collected in a CellSave® tube (Veridex LLC, Raritan, NJ) and processed as described below.
Eligible patients had metastatic breast cancer and were free from systemic anti‐cancer therapies for at least three weeks. The study protocol was approved by the local ethical committee and all included patients gave a written informed consent.
For each enrolled patient two 10 mL blood samples were drawn before the start of the planned systemic treatment, for CTC enrichment, enumeration and characterization using the CellSave™ tube (Veridex LLC).
Samples from patients were analyzed using the experimental procedure described in details below.
When available, a representative primary tumor tissue block was collected.
CTC enrichment and enumeration was performed by the FDA‐approved CellSearch® System. 7.5 mL of whole blood were processed using the CellSearch® Epithelial Cell kit (Veridex LLC), which selects EpCAM positive cells using ferrofluids particles coated with EpCAM antibody. In order to increase the probability of processing at least one sample per patient an additional 7.5 mL sample of whole blood was processed using a “combined kit” which enriches for EpCAM and CD146 positive CTCs (Mostert et al., 2011). In both kits cells are stained with the nuclear dye 4′,6′‐diamino‐2‐phenylindole (DAPI), anti‐cytokeratin 8, 18 and 19‐phycoerythrin (PE) labeled antibodies, and CD45 antibody labeled with allophycocyanin (APC). In addition, in the CellSearch® Epithelial Cell kit, CellSearch® Tumor Phenotyping Reagent HER‐2/neu FITC conjugated (Veridex LLC) was added to evaluate the HER2 status on CTCs, while in the “combined kit”, Ab anti‐CD34 FITC‐conjugated (BD Biosciences) was added to exclude circulating endothelial cells (CECs), given that CD146 enriches for CEC. After enrichment, isolated and stained cells were resuspended in the MagNest Device (Veridex LLC), labeled cells were analyzed in the CellTracks® Analyzer II (Veridex LLC) and CTCs were identified and enumerated according to the criteria specified by the manufacturer's instructions.
CellSearch® cartridges were stored protected from light at 4 °C before analysis and sorting with DEPArray™. Since the DEPArray™ System analyzes 66% (Peeters et al., 2013) of the sample volume loaded into the A300K cartridge, only patients with ≥5 CTCs, after CellSearch® enrichment, were analyzed with DEPArray™.
Each CTC‐enriched sample was recovered from the Veridex cartridge and loaded into the DEPArray™ A300K chip (Silicon Biosystems) according to the manufacturer's instructions. The chip was set into the DEPArray™ system. Chip scanning was performed by an automated fluorescence microscope to generate an image gallery.
An event was considered a CTC according to its morphology (round shape, round nucleus within the cytoplasm) and staining pattern deriving from that of the CellSearch® system: DAPI positive, CK8, CK18, CK19 positive, CD45 negative, and CD34 negative. After CTC identification, single cells or pools of cells were recovered into 200 μl tubes.
As the A300K cartridge allows up to 15 recoveries, and considering that three of them are employed to recover single leucocytes, required as a wild‐type control for molecular analyses, the remaining 12 were used to isolate single or pooled CTCs.
The recovery ratio (Rr), defined as the ratio between the number of single CTCs identified by DEPArray™ and the total number of CTCs isolated by CellSearch® was calculated.
Single CTCs were submitted to whole genome amplification (WGA) using the Ampli1™ WGA kit (Silicon Biosystems) according to the manufacturer's instructions, in order to obtain a sample suitable for sequencing analysis. As a control, WGA was performed on a matched leukocyte isolated from the same enriched sample.
Amplified DNA from CTCs was used for PIK3CA gene analysis which was performed by Sanger sequencing. Briefly, target hot‐spot regions in exons 9 and 20 of the PIK3CA gene were amplified by the Ampli1™ PIK3CA Seq Kit (Silicon Biosystems) according to the manufacturer's instructions. PCR products were purified using the HiYield Gel/PCR DNA Fragments Extraction Kit (RBC Bioscience) and sequenced using the BigDye Terminator 1.1 CycleSequencing kit (Applied Biosystems). The sequence reaction was purified using the ZR DNA Sequencing Clean‐Up Kit (Zymo Research) and analyzed using an ABI PRISM 310 Genetic Analyzer (Applied Biosystems).
DNA from primary tumor tissues was extracted using the FFPE Tissue kit (QIAgen, Hilden, Germany); 50 ng DNA were amplified by the Ampli1™ PIK3CA Seq Kit and sequenced according to same protocol adopted for CTCs.
After enrichment and immunolabeling by CellSearch®, 86 of 130 spiked MDA‐MB 231 cells were identified by DEPArray™ (Rr of 66%). 10 single cells were recovered by DEPArray™ and submitted to WGA using the Ampli1™ WGA kit (Silicon Biosystems) according to the manufacturer's instructions. Exons 9 and 20 of the PIK3CA gene were sequenced in all the sorted cells demonstrating the feasibility of single cell molecular analysis and concordance with the expected wild‐type genotype.
From 07/2012 to 04/2013, 39 HER2 negative mBC patients, were consecutively enrolled.
Patient characteristics by number of CTCs isolated through the CellSearch® are reported in Table 1.
Patients' baseline characteristics by CTC number assessed through the CellSearch® system.
CTC number correlated with the number of tumor sites (p = 0.0448 Fisher's Exact Test) and the presence of bone metastases (p = 0.0170 Fisher's Exact Test).
Twenty patients (51%) had at least one CellSearch® cartridge with ≥5 CTCs per 7.5 mL whole blood and were selected for DEPArray™ sorting. Six (30%) out of these 20 patients had at least one HER2 positive CTC with 8% median percentage of HER2 positive CTCs (range: 4–51%).
CTC count by CellSearch® ranged from 5 to 660 (median 40 per 7.5 mL) and CTCs recovered by DEPArray™ ranged from 1 to 98 (median 13 per 7.5 mL).
Median CTC Rr was 31%, with considerable variability among samples (range 2–76%).
Eighteen patients (49%) had at least one single CTC recovered by DEPArray™ that was sequenced for PI3KCA gene. In 9 patients CTCs were recovered also in pools made of 2–29 cells.
Figure 1 reports the study diagram and corresponding patients' flow.
Study diagram and patients' flow.
For this study overall 221 single CTCs were sorted from eighteen mBC patients and 115 were successfully sequenced for PIK3CA (median percentage of sequenced CTCs per patient is 50%, range: 17–100%, Table 2). Sequencing failure could be due to cell loss during sample manipulation after DEPArray™ sorting, WGA reaction or PCR amplification failure.
Mutational status of the PIK3CA gene (exons 9 and 20) in single or pooled CTCs and primary tumors.
In addition, 15 pools of CTCs from a subset of 9 patients were recovered and successfully sequenced.
CTCs from 12 of 18 breast cancer patients were classified as wild type (wt) for PIK3CA exons 9 and 20. CTCs from the remaining 6 patients demonstrated different mutations in exon 9 or exon 20 of the PIK3CA gene; CTCs from two of the six patients with PIK3CA mutation exhibited heterogeneity (Table 2).
In details, for patient I a c.1633G>A (p.Glu545Lys) mutation in exon 9 of the PIK3CA gene was detected in all the four analyzed CTCs.
Patient III had a c.3140A>G (p.His1047Arg) mutation in exon 20 in all the four analyzed CTCs.
For patient VI, a c.3140A>G (p.His1047Arg) mutation in exon 20 of the PIK3CA gene was found in the single evaluable CTC.
In patient IX, 5 out of 8 single sequenced CTCs were HER2 positive. A c.1624G>A (p.Glu542Lys) mutation in exon 9 of the PIK3CA gene was detected in all the CTCs. In all samples, except a single HER2 positive CTC and one HER2 negative pool, the mutation was present in homozygosis, suggesting a hypothetical loss of heterozygosity (LOH) of the wt allele and consequently a molecular heterogeneity of CTCs.
In patient X a single CTC showed a c.3145G>A (p.Gly1049Ser) mutation in exon 20 of the PIK3CA gene.
Patient XVII showed heterogeneity in the PIK3CA gene mutational status on CTCs. Among the single CTCs (n = 19), 5 were HER2 positive. Three different sequence variants of the PIK3CA gene were identified in three distinct single HER2 negative CTCs: i) c.3037C>A (p.Leu1013Ile) in exon 20; ii) c.1641G>T (p.Glu547Asp) in exon 9 and iii) an intronic variant IVS 9 + 46 G>T. The remaining 16 CTCs were wt for PIK3CA, as were the two pools of analyzed CTCs.
Figure 2 depicts an example of cell imaging and molecular analysis from a patient sample.
Imaging and molecular analysis of a single CTC. Single CTC staining pattern obtained by the CellSearch® (A) and the DEPArray™ (B): the cell shows positive fluorescent signals for DAPI and PE and no signal for APC. (C) Examples of PIK3CA ...
Primary tumor samples were available for six patients and consequently were analyzed for PIK3CA mutational status. In four of six patients both CTCs and tissues were wild‐type for exons 9 and 20 of the PIK3CA gene. In one patient the same sequence variant was found in CTCs and primary tumor, while in the remaining case a mutation of exon 20 was detected in the single evaluable CTC despite a wild type status for both exons 9 and 20 at the primary tumor site (Table 2).
There is increasing evidence that distant metastases can bear unique genomic alterations different from the original primary tumor and that the characterization of metastases, rather than the primary tumor, could provide important information to guide targeted therapy in metastatic disease (Flaherty et al., 2012). Moreover, recent work has shown that different metastatic sites harbor different genomic aberrations (Gerlinger et al., 2012) and biopsy of one or two accessible metastases may not be representative of the whole tumor burden. In this context, the analysis of CTCs might allow real‐time monitoring of cancer biology with a minimally invasive and easily repeatable evaluation in individual patients with metastatic cancer.
Sequencing of single CTCs is likely to improve three major fields of oncology: cancer detection and diagnosis, evaluation of disease progression, and prediction of therapeutic efficacy. Technical advances have enabled genomic analyses at the single‐cell level allowing the profiling of rare cancer cells in clinical samples (Navin and Hicks, 2011). CTC monitoring and the detection of rare clones resistant to chemotherapy might enable clinicians to more effectively tailor therapeutic regimens for individual patients.
At present, the CellSearch® is the only FDA‐approved diagnostic tool for CTC detection and enumeration and can be used to monitor disease progression and efficacy of therapy in mBC patients (Cristofanilli et al., 2005). Besides enumeration, the molecular characterization of CTCs might be useful to identify therapeutically relevant targets on CTCs.
The primary aim of our work was to demonstrate the feasibility of a protocol for non‐invasive genotyping of single CTCs isolated from mBC patients in order to study the heterogeneity of CTCs.
Following CTC enrichment by the CellSearch® system and sorting of pure CTCs (single and pooled) with DEPArray™, we were able to perform genetic analyses upon WGA. In this study we focused our attention on PIK3CA, as somatic mutations in this gene have been found in a substantial fraction of breast cancer tissues (8–35% depending on the intrinsic breast cancer subtype). PIK3CA mutations have prognostic significance and potentially are predictive for response to agents targeting the PI3K pathway. However, while PIK3CA is known to play a vital role in cancer cell proliferation, metabolism and survival (Fu et al., 2013), the clinical relevance of PIK3CA analysis is not yet completely understood and requires further evaluation.
Currently, there are few techniques available for the isolation of single cells, including micromanipulation or laser micro‐dissection, however these methods are performed manually, are time‐consuming and have a low throughput (Shapiro et al., 2013; Navin and Hicks, 2011). Alternatively, high throughput cell isolation may be achieved by FACS (fluorescence‐activated cell sorting), although the disadvantage of this method is the requirement for a large number of input cells (Shapiro et al., 2013), a relatively rare situation when dealing with CTCs. To date, the above cited approaches have been used successfully mainly on cell lines‐spiked samples (Peeters et al., 2013; Kroneis et al., 2011). There are limited studies evaluating these methods for the molecular characterisation of single CTCs using samples from patients, although some data are reported in studies on subjects with colorectal cancer (Fabbri et al., 2013; Heitzer et al., 2013; Gasch et al., 2013).
DEPArray™ is a novel semi‐automated CTC isolation technique that provides cell selection through a multiparametric immunofluorescent staining with high practicability (Peeters et al., 2013). A combination of CellSearch® and DEPArray™ allows the recovery of pure single CTCs that are suitable for molecular characterization, as previously demonstrated on cell lines spiked samples (Peeters et al., 2013; Kroneis et al., 2011) and in metastatic colorectal cancer patients (Fabbri et al., 2013).
We have been able to demonstrate the feasibility of isolation and subsequent molecular analysis of single CTCs and/or pools of pure CTCs from patients with mBC. Considering the fact that the DEPArray System™ analyzes only 66% of the sample volume provided by the CellSearch® cartridge and that additional cell loss might occur during sample preparation, it is expected that the recovery ratio, expressed as the number of CTCs identified with DEPArray™ by the number of CTCs isolated with CellSearch®, will never be 100% and that in the best scenario it will approach 66%. As a consequence we decided to analyze samples with a minimum of five CTC/7.5 mL identified by CellSearch® in order to enhance the probability to recover at least one CTC by DEPArray™.
It can be argued that the evaluation of biomarkers on a limited number of CTCs could not be enough representative of the metastatic tumor biology and complexity. Only future studies correlating bio‐marker status assessed on CTCs with clinical outcome will be able to define the clinical relevance of the presented approach. Despite this potential limitation, to the best of our knowledge, this is the first time that the approach of combining CellSearch® and DEPArray™ for single CTC isolation from blood samples and PIK3CA status evaluation has been applied in mBC. Schneck et al. (2013) recently reported a study of PIK3CA mutation analysis on CTCs, finding that PI3KCA was mutated in nearly 16% of mBC patients. In this case, however, they used SNaPshot methodology, whereby PIK3CA mutational analysis was performed directly on the whole enriched material obtained from the CellSearch® cartridge, rather than assessing mutation status of individual cells. Furthermore, the SNaPshot assay combines multiplex PCR amplifications of exons 9 and 20 of the PI3KCA gene combined with a multiplex primer extension assay, with targeted detection of several mutations in one reaction, while discovery of unknown mutations is not possible. Importantly, when using the SNaPshot method, enriched samples obtained from the CellSearch® cartridge still contain leucocytes, potentially confounding subsequent molecular analyses.
In our study cohort, 33% of patients had a PI3KCA mutation identified. The higher rate of PIK3CA mutation detection, compared with that reported by Schneck and colleagues, may possibly be a result of working on pure CTC samples and at the single cell level. We found six different mutations in the PIK3CA gene. Among them c.3140A>G (p.His1047Arg), c.1633G>A (p.Glu545Lys) and c.1624G>A (p.Glu542Lys) are the most frequently reported mutations in breast cancer, with a frequency of 55%, 20% and 11%, respectively, (www.mycancergenome.org). In addition, the mutation c.3037C>A (p.Leu1013Ile) found in patient XVII is not yet described in the COSMIC (Catalogue of Somatic Mutations in Cancer) database.
The ability to perform a mutational analysis on single CTCs allowed us to assess the presence of CTC heterogeneity within an individual patient and to compare the mutational status between CTCs and matched primary tumors.
In our series, two of 18 patients (11%) presented heterogeneity in PIK3CA status. Since using this technical approach we analyse a subgroup of CTCs with homogenous features (due to antibody‐mediated cell capture), the observed heterogeneity can be considered as a marker of a higher variability in the whole CTC population.
Further investigation of the incidence as well as the clinical relevance of PIK3CA mutational heterogeneity would be of interest, particularly in relation to the response to PI3K pathway‐targeting agents. The use of enriched samples or of pools of CTCs could be an easy and potentially valuable way to analyse DNA variants in the circulation of patients with cancer. Importantly, our results indicate that it is advisable to study CTCs at the single cell level for two main reasons: 1) since the conventional sequencing methods have a maximum sensitivity of 5–10% of mutated allele in a wild type background, the analysis of a heterogeneous sample may impair the detection of rare mutations; 2) it is not possible to correctly determine the presence of multiple different mutations in a CTC pool, with the potential loss of information in this setting (e.g. patient XVII).
In the six cases where we evaluated PIK3CA status in the primary tumor and in the matched CTCs, we found one case with discordance in PIK3CA status. In this case the mutation in the primary tumor might have been missed due to the presence of only a minor sub‐clone bearing the mutation. Alternatively CTCs might gain additional genomic characteristics over time that could be distinct from those of the primary tumor (Alix‐Panabières and Pantel, 2013). However, this has to be considered a preliminary finding requiring confirmation in a large series.
In this study we have shown that the molecular characterization of CTCs in mBC patients can be achieved with a non‐invasive approach, and that the isolation and molecular analysis of single CTCs is feasible in mBC patients with ≥5 CTCs. Further, to the best of our knowledge, we describe for the first time the mutational status of PIK3CA in single CTCs isolated from mBC patients. This study focused on PIK3CA mutations only, however the methodology employed could also be applied to other clinically relevant genetic mutations or potentially even to the evaluation of gene signatures. As the next step, we are now planning a prospective clinical trial to investigate the clinical relevance of PI3KCA gene mutations detected on CTCs, in ER+ HER2 negative advanced breast cancer patients treated with endocrine therapy.
This work has received financial support from Ente Cassa di Risparmio di Firenze (Firenze, Italy) (2011.1035), “CYTOPEM” POR CRO FSE 2007‐2013 Regione Toscana, Breast Cancer Research Foundation (New York, USA) and Associazione Italiana Ricerca Cancro (Milan, Italy (10016).
Pestrin Marta, Salvianti Francesca, Galardi Francesca, De Luca Francesca, Turner Natalie, Malorni Luca, Pazzagli Mario, Di Leo Angelo, Pinzani Pamela, (2015), Heterogeneity of PIK3CA mutational status at the single cell level in circulating tumor cells from metastatic breast cancer patients, Molecular Oncology 9, doi: 10.1016/j.molonc.2014.12.001.