Final Report Summary - CAREMORE (Cancer Responsiveness Monitoring based on Resistance mutations in CTCs)
Executive Summary:
CareMore was a three-year project supported by the European Community's 7th Framework Program (FP7/2007-2013), under grant agreement n°601760 (CareMore). The main goal was to improve the diagnosis and subsequent treatment of women with metastatic breast cancer through advanced molecular analysis of circulating tumour cells (CTCs) in blood. Our many objectives were successfully met in most areas, as summarized below:
Multiplex in situ detection of 4 proteins in CTCs processed according to CellSearch®-based protocol developed in CareMore.
A four-plex in situ PLA™ assay was developed to report four target proteins in fluorescence in parallel with two identification proteins and a nuclear stain, in total thus seven colour fluorescence simultaneously on single cell CTCs. The target proteins were selected for comparison to the primary tumour (ER), for presence and activity status of HER2 (pHER2), and two proteins (p4E-bp and pS6) which were included during the project, to determine PIK3CA mutation, replacing the in-situ mRNA detection originally planned. By measuring the activated and phosphorylated forms of these two proteins found downstream in the signal transduction pathway of PIK3CA, we might have detected an indicator of potential therapy resistance. Clinical data collection is still ongoing as well as the six-month follow-up, due to delays by low patient inclusion and the extra proteins postponing the validation phase. All assay protocols and reagents are available, awaiting execution on the full patient numbers to be carried out outside the CareMore project. Each of these PLA-assays consists of two antibodies targeting the same protein, thus achieving a double recognition and consequently a higher specificity. In total four oligo-nucleotide reporter systems and eight antibodies were used in the PLA-assay in situ successfully.
Improved collection and isolation of CTCs from blood on two platforms.
A new method, implemented on the CytoTrack™ platform was benchmarked against the only clinically approved CTC-profiling method available today, i.e. CellSearch. This new method includes placing cells from a buffy coat on a glass-CD after pre-processing the blood sample and cytokeratin immunostaining, allowing for a fast two-minute scan on the glass CD to identify candidate CTCs. In addition, we were able to add immunostaining for EpCAM receptor. This platform was extended with a so-called CytoPicker™, a technology that allows for a single cell isolation step, whereby characterized CTCs can be picked up and placed in PCR tubes. These cells are then analysed by digital PCR in a single cell analysis to detect the different mutation variations of the PIK3Ca gene.
The existing CellSearch platform at EMC was also extended to allow for a downstream analysis of four target proteins. Starting from the CellSearch Profile kit a more efficient way of collecting and isolating CTCs was developed, whereby a larger quantity of CTCs could be collected from a larger blood volume. The CareMore protocol also includes a modified fixation before arranging the cells on a standard microscope slide through magnetically-assisted centrifugation step. This allows the slides to be scanned in a whole slide fluorescence digital scanner.
Multiplex detection through fluorescence microscopy scanning and subsequent image analysis.
The project planned for and successfully detected seven markers at the same time on the slides from CareMore processed CellSearch samples. To that end all fluorescent labels were carefully selected for their spectral properties in combination with selected light sources and emission filters to be able to quantify the respective signals without interference. Also the cellular location of the biomarker was considered when allocating the different signals to each biomarker unambiguously.
We developed algorithms and established an image analysis pipeline that can pre-process 7-channel fluorescence images from the digital slide scanner, resulting in a score for each biomarker in each cell and cellular compartment. By using reference cell lines that were selected as positive and negative controls, we were able to calibrate the PLA assays. The algorithms can be applied to any target in a multiplex fluorescence assay as long as reference cell lines with a ground truth biomarker expression are available for calibration.
Clinical outcome.
All patient samples from the two clinical studies that were designed and initiated are stored until the completion of the studies and the CareMore CellSearch protocol and CareMore PLA assay can be performed without further modification. It was decided to not run the assay on a smaller group of patients to ensure that the exactly same assay conditions will be applied to all samples and reliable conclusions can be drawn. The studies without the CareMore assay will still reveal if patients with a HER2-positive CTC-sample and a HER2-negative primary tumour benefitted from a trastuzumab treatment, which would not have been given without HER-characterization performed in Study2. We are confident that the tools and data produced by the CareMore project will trigger others to set up molecular diagnostic tests on CTCs for therapy selection in patients with metastatic breast cancer.
Project Context and Objectives:
BACKGROUND ON DIAGNOSIS FOR METASTATIC BREAST CANCER
Today's treatment of metastatic breast cancer is guided by characterisation of the primary tumour, while 90% of deaths due to breast cancer occur as a consequence of metastases. Given important differences in molecular characteristics between metastatic and primary tumour tissue, characterization of metastatic tumour tissue instead of the primary tumour may provide better treatment guidance for patients with metastatic disease. Metastatic tumours are, however, difficult to access using currently available techniques. In the CareMore project, we proposed to guide treatment decisions in women with metastatic breast cancer based on molecular characterization of circulating tumour cells (CTC), used as proxies for the metastatic tumour. From a clinical point of view, mutations in PIK3CA are considered to be important biomarkers that can help the physician to guide the treatment of the patient, i.e. to choose the right treatment at the right point in time. For the sensitive detection of genetic mutations in circulating tumour cells two unique methods were combined, i.e. cell-based detection of genetic mutations and improved circulating tumour cell sampling and identification. Both techniques were provided by the participating SMEs Olink and CytoTrack. Together with leading academic and clinical scientists at the Erasmus University Medical Center and an advanced industrial partner, Philips, the consortium was ideally suited to develop and validate the methods in order to reach proof of concept stage before the methods can be tested for clinical utility, the final aim of the CareMore project.
CONCEPTS AND OBJECTIVES
To advance Circulating Tumour Cell (CTC)-based cancer diagnostics as the original objective, the CareMore project aimed high by developing assays to be applied in two parallel prospective clinical studies on women with advanced breast cancer.
More specifically, we aimed to advance CTC-based cancer diagnostics by:
• Introducing multiplex in situ testing for activating mutations in PIK3CA gene transcripts, for HER2 receptor activation and for estrogen receptor (ER) protein expression using Padlock Probe (PLP) technology (Stockholm University) and Proximity Ligation Assay (PLA®) technology (Olink) in combination with digital slide scanning and separating 8 colours in situ by image analysis (Philips);
• Improving the collection of CTCs (Cytotrack) by increasing the number of isolated CTCs from each patient beneficial for subsequent molecular characterisation and by increasing speed of isolation and identification;
• Overcoming the current limitation of EpCAM dependent enrichment and benchmark this approach against the CellSearch® technology of Janssen Diagnostics, LLC. (EMC and CytoTrack);
• Verifying the clinical utility of these biomarkers in CTCs for therapy guidance in patients with metastatic breast cancer in two prospective clinical studies, study 1 and 2, with clinical follow-up data collected 6 months after the start of therapy (EMC and Philips).
During the first phase of the project it turned out that the PLP assay was not sensitive enough for the targets to be detected (i.e. mutant PIK3CA transcripts). Therefore, we abandoned this strategy and entirely changed the method to determine activation of the PIK3CA pathway in two ways:
• Detect two phosphorylated proteins, pS6 and p4E-BP1, that signal downstream of PIK3CA. The extra protein targets resulted in a 4-plex instead of 2-plex protein detection assay. This assay was developed using PLA technology.
• Detect the PIK3CA mutations using a single-cell digital PCR assay on isolated CTCs. This was enabled by a cell picking methodology developed by CytoTrack to isolate single CTCs after staining and identification, resulting in the Cytopicker™, an integrated part of the CT11™ scanner.
In addition, it was decided to include EpCAM staining in the clinical fluorescent staining assay performed at CytoTrack, to gain knowledge on the percentage of CTCs (determined by CK staining) lacking EpCAM and to allow benchmarking the CytoTrack assay against CTC enumeration by the FDA approved standard Cellsearch assay.
To determine clinical validity of the developed CareMore assays 2 prospective clinical trials were executed within the project. Study 1 aimed at assessing the impact of HER2 status in CTCs on outcome to aromatase-inhibitors in metastatic breast cancer patients with ER-positive/HER2 negative primary tumours; and study 2 aimed at determining the success rate of trastuzumab-based chemotherapy in metastatic breast patients with HER2-negative primary tumours but HER2-positive CTCs. The inclusion criteria and the objectives are listed below:
Study 1
Metastatic breast cancer postmenopausal patients with ER-positive/HER2-negative primary tumours, to receive treatment with aromatase-inhibitors.
• Primary objective: to determine the impact of HER2 expression in CTCs taken at base-line on the clinical outcome with respect to treatment aromatase-inhibitors.
• Secondary objectives: to determine the impact of ER, pHER2, and PIK3CA mutational status in CTCs taken at baseline on the clinical outcome to aromatase-inhibitors treatment; separate analyses for the 3 factors and a combined analysis.
Study 2
Metastatic breast cancer pre- and postmenopausal patients with HER2-negative primary tumours but with HER2-positive CTCs, to receive treatment with trastuzumab-based chemotherapy.
• Primary objective: to determine if these patients benefit from trastuzumab-containing chemotherapy
• Secondary objective: to determine impact of pHER2 and PIK3CA mutations in CTCs on the clinical outcome to trastuzumab based chemotherapy.
The final aim of the project is to show the value of the developed assays by applying them directly to clinical specimens in the two clinical studies executed within the CareMore project. The studies aim to determine the assays’ power to predict outcome of metastatic breast cancer patients treated with endocrine or Herceptin based targeted therapy.
Project Results:
The main goal of the CareMore project was to develop methods to better guide treatment of women with metastatic breast cancer. Nowadays treatment options are based on the characteristics of the primary tumour. However, metastases often show discordant features when compared to their primary tumour. Within CareMore, we aim to use CTCs as proxies for metastases and investigate if indeed their characteristics are deviating in a similar fashion. In view of the substantial discrepancies that exist between primary tumour tissue and metastatic cells, it is likely that treatment decision-making in metastatic breast cancer based on molecular characterization of CTCs is superior to decision-making based on primary tumour tissue, as is currently most frequently done. CareMore will provide means, far beyond current state-of-the-art, to characterize CTCs from patients and score markers known to affect treatment outcome and provide it in a format suitable for clinical use.
The CareMore project aimed to improve the state of the art in the following seven areas:
1. Clinical validation of current research-use-only PLA and PLP technology on CTCs to determine (1) activation status of the HER2 receptor, and (2) presence of PIK3CA mutations.
2. Clinical validation of existing proprietary instrumentation and algorithms for non-EpCAM dependent isolation and analysis of CTCs.
3. Development of a prototype instrument, based on the existing CytoTrack instrument, suitable for next generation molecular characterization methods of CTCs.
4. Development and validation of a complete workflow for clinical application relating to techniques and quality control criteria for isolation and preparation of CTCs with preserved mRNA quality (used for genotyping), multiplex in situ staining of CTCs, as well as software for (automated) fluorescence image analysis and subsequent clinical interpretation.
5. Validation of therapy response markers in patients with metastatic breast cancer.
6. Generation of a prototype diagnostic systems for therapy guidance in post-menopausal breast cancer patients with ER-positive but HER2 negative primary tumours.
7. Generation of a prototype diagnostic system for therapy guidance in patients with metastatic breast cancer with HER2-negative primary tumours but HER2-positive CTCs.
Below the scientific development and technological achievements within the project are outlined, subdivided in three different important areas:
• Improve the collection of CTCs using two different collection methods
• Developing the multiplex in situ assay (the CareMore assay)
• Clinical utility for better management of metastatic breast cancer patients
During the CareMore project, 2 methods (called project “arms”) were developed for CTC isolation and characterization and validated on blood spiked cell lines and patient samples:
• The CellSearch Arm (EpCAM dependent) consists of the following sample processing steps: CellSearch enumeration and parallel isolation of CTCs from 30 mL of blood followed by transferring of enriched cells to a microscope slide with minimal cell loss using a magnetic device developed by Philips and EMC.
• The CytoTrack Arm (EpCAM independent) consists of the following sample processing steps: Cytotrack CTC isolation and enumeration (including EpCAM characterization) followed by single cell picking and determination of the PIK3CA mutational status using digital PCR.
For the two investigated CTC-collection platforms, CytoTrack and CellSearch, collection protocols had to be improved and adjusted, each to be tailor made for both mRNA or protein quantification. The project aimed to detect in one and the same sample preparation in situ all markers in the same cell on both platforms.
1. CTC enumeration and characterization using a multiplex in situ assay (CareMore assay) in the CellSearch arm
The objective of the CellSearch arm was to enumerate CTCs using the FDA approved CellSearch platform followed by isolation and preparation of CTCs for downstream protein and molecular characterization.
1.1.1 Protocol development and optimization
To increase the CTC yield, a method was developed and validated to isolate CTCs for downstream analysis from 30 mL of blood instead of the traditional 7,5 mL. The same was done in the CytoTrack arm.
After Immuno-magnetic enrichment of CTCs from (spiked) blood samples using the CellSearch profile kit according to manufactures instructions, the enriched cells were transferred to a microscope slide with minimal cell loss using a magnetic device developed by Philips/EMC. This assay was validated for optimal recovery with spiked reference samples in healthy donor blood.
For optimal preservation of CTCs on slide for both protein and molecular (DNA and RNA) characterization many different fixation methods, incubation times and storage conditions for blood samples and slides were tested using reference samples followed by multiples PLA and PLP assays.
1.1.2. Improved transfer of CellSearch-enriched CTC slides
Different ways to transfer the CellSearch enriched samples to slides were investigated to increase CTC recovery. Also different fixation methods of cells were compared as well as their effect on the stability of stored analytes. The developed protocol was finalised before patient inclusion was started in 2015. The method finally developed allowed the cells to be cytospun on a slide and thus easier to be investigated with subsequent detection techniques, i.e. the CareMore protein staining assay and automated slide scanning. The end product of CTCs cytospun on a slide was used in the proof-of-concept validation using selected reference cell lines spiked in healthy blood.
Before continuing on clinical material a proof-of-concept test was performed with the newly developed protocol using CellSearch-enriched fractions which were cytospun on slides to ensure CTC recovery was feasible with limited loss using the new procedure.
We were able to establish a fixation method that preserved both proteins and mRNA in cells spiked in blood.
In conclusion, the storage conditions for the CareMore protocol as well as a protocol to successfully transfer samples to standard microscope slides were defined in the beginning of 2015 after which inclusion of the first patient samples was started.
1.2. Developing the multiplex in situ assay (the CareMore assay)
The original objectives were to improve the collection and identification of CTCs, and subsequently detect two selected proteins (estrogen receptor (ER) and phosphorylated Human Epidermal growth factor Receptor 2 (pHER2)) at the same time as detecting 4 mutations (in mRNA) from a potential cancer marker gene (PIK3CA). Mutations in PIK3CA are considered to be an important biomarker that might help physicians guide the treatment of a patient, i.e. to choose the right treatment at the right point of time. The in situ mutation detection assay to be developed in CareMore is a PLP assay from Olink developed with knowledge from Stockholm University (SU), and will allow for the identification of Phosphoinositide 3-kinase (PI3K; the protein product of the PIK3CA gene) activating PIK3CA mutations. At the same time, using the PLA® technology ER- and pHER2-proteins will be visualized in CTCs isolated from blood. Detection was planned to be in situ with subsequent visualization using a whole slide microscope scanner which could handle 8-9 colours simultaneously. The protein assay was finally developed to use 7 colours of which 4 instead of 2 proteins are quantifiable in situ. The PLP mRNA mutation analysis was swapped for a single cell PCR.
1.2.1 multiplex PIK3CA mutation detection at the mRNA level using PLP
The original objective was to detect mRNA from the PIK3CA gene, and its mutated variants, in situ at the same time as detecting two proteins (ER and pHER2). Padlock probes (PLP) have been used to detect mRNA in tissue before, and in this project its performance on CTCs was investigated. Thorough investigation however did not lead to the successful in situ detection of PIK3CA mRNA in any form. It is still unclear why the PIK3CA mutant transcript detection did not work, since PLP is a proven technology and has been used before to detect point mutations in transcripts of other genes expressed at same level as PIK3CA in situ on microscope slides.
The efforts of detecting the different PIK3CA mutations using mRNA PLP was a main focus for almost two years. In parallel to the PIK3Ca work, we used PLP on well-known genes to find a common fixation method for proteins and mRNA together. Moreover, many other different approaches were investigated with the goal to detect PIK3CA mRNA but without success. Among other things, we developed a more stringent PLA-protocol regarding specificity because of the requirement that the PLP assay should work simultaneous with the PLA-assays. The original objective was thus changed from in situ to in solution detection after a project decision, explained in more detail in the single cell PCR section.
Parallel to the project, the group at SU has also investigated several more genes at the same expression level of PIK3CA in combination with several ways to isolate CTCs from blood, i.e. from different platforms available on the market. This research is to be presented in other EU-funded projects and will add to the knowledge of how different genes behave differently in biological samples. The knowledge gained will thus be used to continue development of the method, for other selected target genes.
The PLAs developed within CareMore are described below, as well as the respective downstream analysis for the two platforms.
1.3.1 multiplex protein detection using PLA
When the PIK3CA mRNA turned out to be impossible to detect (see above), the project decided jointly to add two extra protein targets to the PLA-assay as a compensation, and at a later state we also tried to detect PIK3CA but using single cell PCR (see below). The two added proteins were selected to reflect the activation status of PIK3CA by detecting proteins that are activated downstream of PI3K (the PIK3CA protein product) signalling.
The specificity of the protein assays was investigated on cell lines, both before and after the fixation conditions were established. Since originally there was also a need to adjust the PLA staining protocol to function in tandem with the PLP assay, necessary optimisations were made to the protocol to, despite of the already specific dual recognition of the antibodies, change to a more specific ligase enzyme and to make the annealing temperature higher to prevent any non-specific hybridisation.
The reference cell lines that were used as controls for the protein assays were selected based on their known status regarding the proteins of interest. When ER was concerned, there were already known guide lines for staining and subsequent interpretation of the results, even though the IVD was developed for use on tissue. In like manner, a lot is known about the expression of HER2, but less is known about the HER2 pool which is actually phosphorylated. As for the two lesser known proteins, phospho-S6 (pS6) and phospho-4E-binding protein 1 (p4E-BP1), a selection of reference cell lines was made after an extensive screening of about 55 cell lines per protein The challenge with these abundantly expressed proteins was to find cell lines that could be used as negative controls. Thus, we chose low expressing reference cell lines as our negative controls for these latter 2 proteins.
Multiplex staining and detection using fluorescent labels necessary for quantification of the 4 proteins, i.e. ER, pHER2, pS6 and p4E-BP1, was performed. The four respective proteins were detected through PLA® technology using 4 oligo detection systems whereby each protein is bound by 2 antibodies (4 times 2 oligonucleotides for conjugation). Each antibody is conjugated with the respective oligo and each set of 2 matching oligos is amplified by rolling circle amplification followed by simultaneous detection using specific probes each having an excitation at a distinctive fluorescent wavelength. One major aim of the CareMore project was to analyse PIK3CA mutations in situ in the CTCs and simultaneously detect ER and phosphorylated HER2. Since we decided to not proceed with the in situ PLP, we instead expanded the PLA assay to also include two additional proteins to represent activation of the PI3K pathway by the downstream proteins pS6 and p4E-bp by PLA. unfortunately, it turned out that the 4-plex PLA could not recognize its epitopes in the frozen samples prepared for the CytoTrack platform. Therefore, the CareMore protein assay was used on the CellSearch platform purified samples only. Instead, the PIK3CA single cell PCR analyses were performed on the CytoTrack platform in the final proof-of-concept validation of mRNA and proteins, respectively.
Within the CareMore project a 4-plex PLA was developed, enabling quantification of 4 proteins simultaneously at single CTC level. In addition to the 4 proteins to be detected by PLA, immunofluorescence was used to simultaneously detect cytokeratin (CK) and CD45 while DAPI was used to detect cellular nuclei. CK and CD45 are used in downstream analysis to separated CTCs, which are CK-positive and CD45-negative from leukocytes, which are CK-negative and CD45-positive. For 3 of the 4 proteins included in the evaluation tests the aim was to detect the phosphorylation status, which is difficult to detect in blood stored for a long time. However, when using PLA, the signal is amplified and quantifiable. To ensure the detected cells are CTCs, the analysed cells must show a cell-shaped nuclei in the DAPI channel, a distinct Cytokeratin staining and no CD45 staining.
This project introduces a completely novel method to quantify 4 proteins simultaneously, in situ in individual CTCs. This is enabled by using 4 different PLAs, usable in parallel, developed by Olink. The PLA assays are immunoassays where each protein is detected using two antibodies. In order to work simultaneously, the 8 antibodies binding the 4 protein targets in a PLA are directly conjugated with DNA oligonucleotides unique to the 4 labelling systems. These oligo systems are not locked to the specific antibodies and their protein targets tested here and can be transferred to other markers of interest. As long as there are specific antibodies in a purified state, they can be used in any of the 4 oligo systems, in any combination.
1.4. Conclusion for protein Multiplex assays
The PLAs work as a proof-of-concept evaluation and can report quantitative values for 4 of the targets in cell lines spiked in blood. Two of the targets consist of epitopes that seem to be severely affected by the long storage time before sample preparation, thus resulting in values that are too low to be reliable. In addition, the 4 oligo systems are proven to be able to report only after dual antibody recognition, which should enable higher specificity. In the future, the reporting systems can be used on epitopes that are investigated to be more stable than for example ER.
2. The CytoTrack arm for CTC enumeration, EpCAM quantification and molecular characterization by single cell PCR
The objective for the CytoTrack arm was to enumerate CTCs, to stain them for EpCAM as well as various molecular markers of the PI3K pathway. Improvements which were made are discussed in detail below and relate to:
1. Blood sampling and preservation
2. CTC enumeration, including EpCAM characterization
3. Retracing, CytoPicking, and PIK3CA mutation analysis
2.1.1. Blood sampling for the CytoTrack arm
The average amount of CTCs in 10 ml blood is low, often just a few CTCs are detected. There is therefore a widespread but until now unmet need for processing larger blood sample volumes. Recently, we showed that modifying the sample preparation techniques by adding a pre-processing step to enrich for mononuclear cells, CTC recovery could be increased to 300% because 30 mL of whole blood could be analysed instead of the traditional 10 ml.
This application is commercially available as a supplementary CytoTrack application note.
2.1.2. Blood preservation for the CytoTrack arm
A major challenge in the detection, enumeration and characterization of CTCs in blood samples is the limited storage potential of commercial blood sample tubes. For biobanking, scientific or clinical purposes, preservation can provide opportunities and allow batch processing of blood samples over time in the future. At CytoTrack, a method for cryopreservation of CTCs maintaining compatibility with the CytoTrack instrument and methods of detection, enumeration and characterization was successfully developed and validation showing CTC recoveries of 89-91%.
This cryopreservation method was used in the clinical study and published in Biopreservation & Biobanking 2016.
2.2.1. Development of a CytoDisc staining lid for on disc in-situ assays
At Philips Research a novel staining method for CytoDiscs was developed. It comprised a lid that could be applied to the disc after the buffy coat was applied and dried. The lid contains a 1 cm hole in the center and is placed on 6-8 round spacers of 50-180 µm thick with diameters of about 5 mm. This creates a cavity above the dried buffy coat with the height of the spacers. When the disc with the lid was placed on a spin coating device the whole system could be spun around at various speeds. When liquids were applied to the centre hole in the lid, the spinning of the disc would force the liquid into and out of the cavity where the buffy coat is. The flow stopped automatically when the hole was completely empty due to the capillary force acting inside the cavity. In this way liquids could sequentially be applied to the CytoDisc surface without the risk of drying the cells. By sequentially applying different liquids like washing and antibody solutions combined with incubation steps, staining assays could be efficiently performed with minimal use of reagents. After an assay the disc and lid could be rotated at high speeds to dry the sample completely before applying a coverslip. The staining lid system was developed for a short amount of time but abandoned because the effort to get it at the same functional level as normal, manual on-disc staining was judged to be too great for this project alone.
2.2.2. CytoTrack concept for retracing individual CTCs on CytoDisc
Both On-Disc and Off-Disc characterization require retracing of individual CTCs, i.e. the ability to find the original location of each CTC after removing the CytoDisc and re-inserting it into the CytoTrack instrument. This was accomplished using a ‘notch’ on the design of the CytoDisc, such that it locks into the same position every time. The location of each CTC is displayed on a digital ‘CytoMap’, which may be printed to serve as template for the CytoDisc. The print works as an underlay to the CytoDisc, and shows where the target cells are on the disc. In this way only the areas where the cells are positioned can be stained, reducing the use of (expensive) staining reagents.
Validation of this concept showed >96% recovery and the results were published in Journal of Circulating Biomarkers.
2.2.3. CytoTrack On-Disc characterization (for using HER2 as PoP)
The On-Disc procedure involves the disc to be inserted into the CytoTrack instrument twice. The first procedure is to detect the target cells on the disc, i.e. to identify the position of the target cells (digital mapping), and to enumerate the number of CTCs. Hereafter, the disc is taken out of the instrument, and the CoverDisc is removed. Now the cells on the CytoDisc are stained, either with IF or FISH. A new CoverDisc is applied, and the CytoDisc is reinserted into the instrument. The instrument can now go directly back to the position of CTCs on the disc, since the positions on the disc are already captured and stored. Now it is possible to detect the additional staining and characterisation of the target cells. CTC retracing and the associated digital CTC map form the basis for downstream characterization of CTCs while they are attached to the CytoDisc. The characterization reagents are applied to the surface of the CytoDisc and may be either antibodies (IF) or fluorescence in situ hybridization (FISH) probes. Results are read after reinserting the CytoDisc into the CytoTrack instrument. The concept was developed using HER2 as biomarker as HER2 is both an important breast cancer marker with therapeutic application and a biomarker that may be measured by both immunofluorescence IF and FISH.
2.2.4. On-disc EpCAM characterization of CTCs using multiplex immunofluorescence
EpCAM characterization of CTCs as a multiplex assay component combined with CTC detection was developed as a means to simultaneously measure EpCAM expression levels in individual CTCs and was included into the CareMore clinical study. The accomplishment has been commercialized as a CTC StainEpCAM reagent and the associated protocol and clinical performance have, among others, been presented at the European Cancer Congress in Vienna in 2015.
As part of the protocol all patient samples are screened for HER2-positivity by the CellSearch HER2-kit. In the project all patients have also been tested in parallel on disc for the same markers including EpCAM. The on-disc assay allows us to enumerate all cells stained by CK irrespective of whether cells also stain positive for the CellSearch capturing marker EpCAM. The amount of CTCs identified by each of the two methods was also compared.
2.3.1. Off-Disc characterization: CytoPicker for isolation of individual CTCs
A major accomplishment was the discovery that individual CTCs may be ‘lifted’ up from the CytoDisc using a combination of micromanipulation tools and proprietary ‘cell picking’ methodology.
A cell picking device was developed during the project that could transfer stained single CTCs from a disc in the CT11™ scanner and into a PCR tube. Using the CytoPicker innovation, the picking of individual CTCs was optimized to a point where more than 80% of CTCs could be picked within 2-8 min per CTC and free of contaminating white blood cells deposited into (for example) PCR tubes for subsequent analysis. The achievement of this milestone forms the basis for single cell analysis of PIK3CA mutations; the last and equally challenging step towards completing the CytoTrack arm with single CTC analysis.
This innovation was built into the CytoTrack instrument and presented to the public for the first time at the 7th International Conference of Contemporary Oncology, Poznan, Poland, in March 2015 after protecting the intellectual property rights with a patent application. Validation studies were used to generate specifications for the CytoPicker that among others are summarized in a CytoTrack brochure/flyer.
The CytoPicker can be used to isolate single target cells. The cells are lifted from the disc surface and are then placed in PCR tubes (one cell per tube). This is done without any contamination of white blood cells.
2.3.2 Single cell PCR for PIK3CA mutations for individual CTCs
As an alternative to the PLP assay, we developed a procedure to detect PIK3CA mutations in clinical samples on single cells using a digital PCR method analysing 4 common activating PIK3CA mutations (leading to the following amino acid changes in PIK3CA: E542K, E545K, H1047R and H1047L). We were able to validate the procedure on control cell lines with known mutations spiked into healthy donor blood as well as to successfully determine the PIK3CA mutation status on single CTCs picked by the CytoTrack system, from 5 patients. To conclude, a digital PCR method covering the 4 most common PIK3CA mutations was successfully developed and validated on cell lines with known mutational status spiked in healthy blood samples which consider an important accomplishment, moreover, it completed the development of the CytoTrack arm.
2.4. General conclusion for the Cytotrack arm
A new model of the CytoTrack instrument and corresponding CTC characterization techniques have successfully been developed and validated for use in the clinical validation study and in parallel was made commercially available for research use only.
This complete solution integrated into a compact bench-top instrument which allows CTCs to be detected, enumerated and characterized either by multiplexing or by downstream characterization performed either on the CytoDisc (On-Disc) or after isolating the individual CTCs.
For the CytoTrack instrument the following novelties were developed: (1) Concept for retracing individual CTCs on CytoDisc, (2) CytoPicker for isolation of individual CTCs and (3) CTC rating and improved imaging for faster results. With regard to methodology the following improvements were realised: (1) Sample volume to be used increased to 30 ml, (2) Cryopreservation of blood samples, (3) multiplex EpCAM characterization of CTCs, (4) On-Disc characterization, (5) Off-Disc characterization followed by CytoPicking and (6) Single cell characterisation by PCR for PIK3CA mutations for individual CTCs.
In summary, the CytoTrack arm comprises the key accomplishment described above. It has successfully been developed and validated for use in the clinical validation study.
3.1 Multiplexed detection of the developed CareMore situ assay
In order to achieve detection of 7 biomarkers, i.e. Cell nucleus, CK, CD45, ER, pHER2, p4E-BP1 and pS6 on a single microscope slide, we used 7 fluorophores with well distinguishable spectral properties: DAPI, Atto488, Atto647n, Atto550, Atto490LS, Atto430LS and Atto590, respectively. DAPI is used to detect cell nuclei and is the basis of the cell identification algorithm. Based on DAPI a nuclear and cytosolic/membrane compartment is defined. CK is used to identify cells of epithelial origin, while CD45 indicates that a cell originates from the blood lineage. CTCs are defined as CK positive and CD45 negative cells. Both CK and CD45 are stained using standard immunofluorescence (IF) protocols. The remaining protein markers (ER, pHER2, p4E-BP1 and pS6) provide further characterization of the CTCs and are stained using PLAs as described above.
The 7 fluorophores were chosen based on the spectral constraints of the light source of the fluorescent scanner which required them to be excitable in the 400 - 650 nm range. The fluorophores mentioned above meet this requirement. We made use of the available spectrum to the fullest by adding 2 long Stokes shift (LS) fluorophores to the set (Atto430LS and Atto490LS). These molecules show similar excitation spectra to DAPI and Atto488, respectively, but their emission is much more redshifted. Consequently, they can be spectrally separated from the other fluorophores quite well using proper emission filters.
Both the excitation and emission spectra of the fluorophores typically have a spectral width of approximately 50 nm (full width at half maximum). This means that there is considerable spectral overlap between them. Combined with the fact that we are bound by what filter set suppliers can offer, typically leading to slightly sub-optimal wavelength cut-offs, we were left with significant crosstalk between the 7 detection channels. To minimize the risk of misinterpreting signals, we assigned the channels with the strongest cross-talk to markers that were least likely to co-localize, e.g. pHER2-Atto490LS leaks into ER-Atto550 but pHER2 is membrane localized while ER is predominantly present in the nucleus. The filters were designed, purchased and mounted in the scanner. Cross talk analysis revealed small but unfavourable effects that could be solved by biochemically tuning relative signal brightness and by characterizing said cross talk and applying a correcting algorithm that was developed in house as part of the image analysis pipeline.
Thus through carefully selected filters combined with a multi-band LED light source, 7 fluorophores could be separately analysed after being scanned in a whole slide fluorescent scanner. Judging the fluorophores according to the known cellular localization of the targets helped the developed software to correct for the leakage of signals between a few of the filters. Thresholds for potential cut-offs for a positive and negative signal for every marker were determined by using the selected reference cell lines as positive and negative controls. Combined with an automated detection of each cellular nucleus the scores could be determined for each cell compartment and allowed signal intensity histograms to be produced for each sample. In this way an assay was judged by comparing the signal patterns with the expected values for the ground truth samples.
In conclusion, all 7 biomarkers could be detected accurately.
3.2. Automated scoring of the developed CareMore in situ assay
The objective of the scoring part was to provide algorithms that: (1) Extract required parameters from the multicolour fluorescence images with dot-like signals obtained from the stained samples, and (2) Stratify the patients according to the test results.
There were several different applications for which algorithms have been developed. Due to the need for a high multiplex (7 signals) and spectral limitations in the excitation light source, the fluorescence filter sets and the CCD detector system, spectral overlap between some fluorophores was unavoidable. This resulted in crosstalk of signals. This crosstalk was quantified and an algorithm was developed that minimizes it, resulting in properly separated signals that could be further processed. To detect the dot-like signals an algorithm was developed that minimizes the influence of interfering signals and that finds the peak of each dot. To identify CTCs an algorithm that detects nuclei was developed together with an algorithm that estimates the location of cell membranes. The nucleus detection algorithm was aimed at being independent of any possible background signals. The estimation of the cell membranes was done by taking the nuclei outlines as a starting point, and then increasing the outlines in size. These detected nuclei and estimated membranes could then be used to quantify the dot-like or immunofluorescence signals within the created compartment boundaries. All these algorithms were combined to quantify the validation samples for the CareMore assay. Those quantification results were extensively visualized and statistically analysed to assess the performance of the assay.
No patient stratification was done due to low patient inclusion in the clinical trial and the decision to not stain patient samples yet with the CareMore assay. In conclusion, the algorithms were successfully applied to assess the performance of the assay. Unfortunately, due to low patient inclusion in the clinical trial, no stratification of patients could be performed.
4. Clinical utility testing
4.1. Introduction
To collect patient samples for clinical validation of the designed assays and to obtain clinical results, two clinical studies were designed. Both studies were approved by the ethical board of the Erasmus MC in Rotterdam. Next, the rationale of the studies and a description of both studies is given.
Metastatic breast cancer patients are still treated on primary tumour characteristics, while it is now increasingly appreciated that important characteristics like ER- and HER2-expression can differ between the primary tumour and the metastasis. Circulating tumour cells (CTCs) are present in the peripheral blood of patients with metastatic breast cancer and are thought to represent the characteristics of the metastases. From previous studies, we know that 18% of the patients with an HER2-negative primary tumour do have HER2-positive CTCs.
CareMore study 1
From literature it is known that patients with breast cancer with a ER-positive and HER2-postive primary tumour have a worse response on endocrine therapy compared to patients with an ER-positive but HER2-negative primary tumour. Therefore, we hypothesize that patients with an ER-positive and HER2-negative primary tumour, but with HER2-positive CTCs will also have a worse response to endocrine therapy. So the primary objective of this study is to determine the impact of HER2-expression in CTCs taken at baseline on outcome to aromatase inhibitor (AI) therapy in metastatic breast cancer patients with an ER-positive primary tumour. To answer this primary objective, 30 patients with HER2-positive CTC’s will be enrolled in this study. From all patients with breast cancer, 65% has detectable CTCs. From the patients with detectable CTCs, 28% has HER2-positive CTCs. Therefore, to include 30 patients with HER2-positive CTCs in this study, 165 patients have to be screened. Ethical approval of this study was obtained at the 10th of February 2015. The clinical study opened in March 2015 and the first patients were included in the same month.
CareMore study 2
Trastuzumab is a targeted therapy, directed against HER2. This therapy is given to breast cancer patients with a HER2-positive primary tumour. There is accumulating evidence that there are patients with a HER2-negative primary tumour who respond to trastuzumab-based chemotherapy. A group of patients who might benefit from trastuzumab-based therapy are patients with HER2-positive CTCs. Therefore, we hypothesize that patients with a HER2-negative primary tumour but with at least one HER2-postive CTC will benefit from HER2-targeted therapy. The primary objective of this study is to determine if metastatic breast cancer patients with HER2-negative primary tumors but at least one HER2-positive CTC benefit from chemotherapy with trastuzumab. To answer this primary objective, 18 patients should be treated with trastuzumab. Therefore, 99 patients should be screened for HER2-positive CTCs. Ethical approval of this study was obtained at the 18th of February 2015. The clinical study opened in March 2015 and the first patients were included in the same month.
4.2. Patient inclusion
Patient inclusion CareMore study 1
The following hospitals from the Netherlands and Belgium are participating in the CareMore study 1:
• Erasmus MC (Rotterdam)
• Ikazia Hospital (Rotterdam)
• Saint Franciscus Gasthuis (Rotterdam)
• Jeroen Bosch Hospital (Den Bosch)
• Albert Schweitzer Hospital (Dordrecht)
• Gelre Hospital (Apeldoorn)
• Vlietland Hospital (Schiedam)
• GZA Hospitals (Antwerp)
As mentioned before, 165 patients should be screened in this study. The result of the HER2-CTC enumeration is not reported back to the hospitals. We will look at this data at the end of the enrolment of the 165 patients. Currently, 54 patients are included in the study. We did not manage to include all 165 patients in the three-year period of the CareMore-project. Mostly due to the fact that the clinical studies could not be opened before March 2015, as the first 14 months of the project were needed to optimize and validate patient sample collection and preservation methods for downstream analysis, so the study was only open for 18 months and patient accrual was slower than expected. We will continue to collect samples of the patients after the CareMore-project has finished. All participating hospitals will join in the completion of the patients’ sample collection, and even new centres are currently being attracted to participate in this study, to further boost accrual.
Patient inclusion CareMore study 2
The following hospitals from the Netherlands are participating in the CareMore study 2:
• Erasmus MC (Rotterdam)
• Ikazia Hospital (Rotterdam)
• Saint Franciscus Gasthuis (Rotterdam)
• Vlietland Hospital (Schiedam)
In total, 29 patients have been screened for HER2-positive CTCs, of those only 4 patients have been treated or are currently being treated with chemotherapy and trastuzumab. This means 14% of the patients had HER2-positive CTCs, this is a bit lower than the expected 18%. But this number will probably increase with a higher number of inclusions. The aim was to screen 99 patients in the CareMore-project. We were unable to fulfil this aim, because also this study was only open for 18 months of the CareMore-project, because of the mentioned reasons above. Also for the CareMore-trastuzumab, accrual will continue after the CareMore-Project has finished. All participating hospitals, and also some new hospitals will contribute to the patient inclusion.
4.3. Patient sample processing
Patient sample processing CareMore study 1
In the CareMore study 1, 80 mL of blood was drawn from all patients at baseline. The blood is collected in 2 x 10 mL CellSave tubes and 6 x 10 mL EDTA tubes. After a sample is drawn, the blood is collected by a courier at the participating hospitals, allowing the blood to arrive in time in the laboratory to be processed within the required 24 hours. For all patient samples, we did manage to process the EDTA blood within 24 hours.
The 2 CellSave tubes are collected for CTC enumeration. One tube is processed in the laboratory for Translational Cancer Genomics and Proteomics in the Erasmus MC in Rotterdam, the Netherlands. Here the CellSave tube is processed in the only FDA-approved CTC enumeration system: the CellSearch system (Janssen Diagnostics, Raritan, NJ, USA). During the enumeration, a HER2-staining is performed, so HER2-positive and HER2-negative CTCs can be determined. The second CellSave tube is send to the CytoTrack Laboratory in Denmark. There CTCs are enumerated using the CytoTrack system and after enumeration, single CTCs are picked and processed for further molecular characterization. The EDTA tubes were processed in the laboratory for Translational Cancer Genomics and Proteomics in the Erasmus MC in Rotterdam, the Netherlands. From 30 mL of EDTA blood CTCs were isolated using the CellSearch system and after isolation the CTCs were processed on a glass slide for further PLA processing, the other 30 mL of blood was stored in DMSO so this could be send to CytoTrack for single cell picking in batches.
We were able to collect all the samples from the patients that participated in the study. We were also able to send the CellSave tubes to the CytoTrack Laboratory in Denmark and process the samples there within 96 hours. There were no major issues in processing the samples and all samples have been stored for further processing.
Patient sample processing CareMore study 2
In the CareMore study 2, 20 mL of CellSave blood is collected from all patients at baseline. After the blood draw, the blood is picked up by a courier so the blood could send to the CytoTrack Laboratory in Denmark on time. We managed to send the CellSave tubes to Denmark and process them within 96 hours for all but one patient. When a patient had HER2-positive CTCs, another blood draw was performed to collect the 60 mL of EDTA blood. This blood was processed the same way as the EDTA blood in the CareMore study 1. Also for this study, there were no major issues in processing the samples and all samples were stored at the moment for further processing in the future.
4.4 Clinical data collection
For both studies, clinical data of all patients is up to date. At the moment, we are not able to provide information about the progression-free survival at 6 months (PFS6months) from all patients, because multiple patients are still being treated. Because of the technical difficulties, as described in the previous paragraphs, we are not able to correlate the PFS6months with the characteristics of the CTCs that have been isolated.
As a consequence of a delay of the development of assay and preservation methods leading to the required changes as outlined above, and due to slower patient inclusion than predicted, it was decided to amend the project´s end point regarding validation on clinical material requiring the six-month patient follow-up which could not be reached for a majority of patients within the time line of the project. Therefore, and also for ethical reasons, the consortium decided not to interfere with potential outcome of the two ongoing patient studies by performing the CareMore assay (CTC identification and subsequent quantifying 4 proteins) on the limited number of patients collected within the scope of the project. Instead we changed the end points regarding the CareMore protein assay from staining all clinical samples to a proof-of-concept validation on well-defined reference cell lines spiked into healthy blood to show feasibility of the developed assay to detect their target molecules towards they were developed. Recruitment of patients and subsequent sample collection started in early 2015 is still ongoing and will continue beyond the end date of the project. All samples are stored according to the developed storage conditions which maintain analyte stability.
In summary, the clinical samples will be preserved for future use and will not be investigated using the developed CareMore assay, consequently neither was the six-month follow-up objective investigated within the scope of this project. All reagents are stored and have been passed on to EMC so that the assays can be performed in batch on the samples when the target number of patient samples are collected, which at today’s rate of recruitment would be approximately in October 2017.
The final statistical analysis of patient outcome in relation to the investigated markers will be available only 6-months after the last patient has been included in the study, this is also true for the 2 clinical studies. Due to unforeseen circumstances including delays in assay development, slower than expected patient accrual combined with the fact that fewer patients than expected showed CTCs positive for HER2, final analysis will be outside the project´s 36 months’ time frame. We don’t know why fewer patients than expected have HER2 positive CTCs, however recent experience indicates HER2-postivity of CTCs in a similar patient population is indeed lower than we anticipated at the start of the project. Of course the slower recruitment means that the sample collection is taking longer to complete. All patient data acquired within the time frame of the CareMore project are presented in a conscious way, well-aware that it represents only a part of the entire cohort, awaiting a significant number of cases to be included later than planned. A fraction of the clinical samples was also investigated for PIK3CA mutations. This was done with cells already identified as CTCs after going through the Cytotrack platform including staining of HER2, EpCAM and cytokeratin. This showed the feasibility of mutation calling in clinical specimens included in the CareMore studies.
Potential Impact:
CytoTrack
With regard to Sales & Marketing, CytoTrack will continue to use the results - CytoTrack instrument and CTC characterization methods – to support sales and marketing in the form of Press Releases, Product Specifications, Web Site, Scientific Publications and Poster Presentations. Examples of such dissemination activities initiated during the project period are available online. Furthermore, the CytoTrack instrument and accompanying methods are made commercially available and thus allow researchers and clinicians around the world to perform their own CTC studies.
In addition a CTC Centre of Excellence was initiated which means the CytoTrack instrument and CTC characterization capabilities are already available as a service from the Department of Clinical Biochemistry, North Zealand Hospital, University of Copenhagen Hillerød, Denmark, where the CytoTrack instruments and methods are implemented, servicing the clinical and scientific community as a reference laboratory where CTC analysis may be performed on blood samples shipped from clinical studies around the EU. For researcher and clinicians and other centres are being explored in Poland and Germany. This way researchers and clinicians may immediately benefit from capabilities offered by the CytoTrack instrument and methods. The CTC Centre of Excellence is also used to process the blood samples from the clinical validation study.
With regard to grant application, the results have and will continue to be included in company presentations and grant applications.
With regard to Intellectual Property, the patent protecting the CytoPicker will be applied for in the EU and optionally the US for future protection of the invention made during the project.
So far, CytoTrack instruments and method have been developed to form basis for the generation of the socio-economic impact and the wider societal implications and the CareMore clinical study on breast cancer has been initiated to provide the first clinical evidence. In parallel, other clinical studies are being outlined for other cancer types and/or with other cancer centres in EU.
In conclusion, the CareMore project has successfully enabled CytoTrack to develop and commercialize instruments and methods for characterization of CTCs to a technology readiness level where they can be validated for clinical use. As such, CytoTrack has transformed from a technology company to a provider of clinical research products and is ready for exploring means for entering the lucrative clinical diagnostic market. Impacted by the encouraging results from the CareMore project, a Horizon phase 1 grant was awarded to CytoTrack in 2015 for preparation of a Business Plan for CytoTrack to bridge the gap - extensive clinical studies and regulatory approvals – between the research and clinical market. Now, the CareMore project extended with a Business Plan (Horizon 2020) together provide foundation for preparation of an application for a Horizon phase 2 grant with focus on a multicentre clinical study in Europe providing the clinical evidence for a better and more cost-efficient personalized approach to cancer therapy and for CytoTrack to obtain regulatory approvals; the ‘license’ to enter the clinical diagnostic cancer market.
Philips
At Philips Research we have equipped our 3DHistech Pannoramic Midi scanner with 7 additional fluorescent filter sets and software was developed to reduce the inevitable crosstalk between the 7 resulting fluorescent channels due to spectral overlap. An algorithm was developed to identify cell nuclei in the DAPI channel and thereby define a nuclear compartment. Subsequently a virtual membrane was algorithmically generated around the nuclei to allow for the definition of a cytosolic compartment. This allowed us to quantify all the fluorescent channels other than DAPI in the nucleus and cytosol compartments for both immunofluorescent (homogeneous, smooth) and PLA (discrete, dot-like) signals. We developed code to be able to visualize the individual cells sorted based on any of the quantified fluorescent signals or morphological parameters such as nucleus size. Moreover, we were able to rank and present cells based on multiple parameters facilitating visual inspection of the most interesting cells first. Based on the cytokeratin (CK) status spiked tumour cell lines (and in the future possibly real CTCs) could readily be identified amongst a background of CK-negative white blood cells. Statistical methods were applied to investigate the power of the individual CareMore assays to separate positive from negative cells for 7 biomarkers in total.
We will continue to use this high multiplexing capability in cell line experiments and perhaps in the future for the characterization of CTCs as well. The algorithm created to identify cells and quantify fluorescent signals of different nature (homogenous and dotted) in high multiplex was successfully adapted for use with formalin-fixed, paraffin-embedded tissue slides and enables us to investigate any type of cancer tissue. We performed multiplexed mRNA FISH experiments and were able to use the new filters in combination with the cell and dot detection algorithms to quantify the mRNA generated by the activity of oncogenic pathways at the cellular level. Some of this work is going to be published. The software for exporting and analyzing (annotated) scanned images from the 3D Histech scanner are used frequently and will be used for the foreseeable future in several research projects. The tool set enables us to carry out translational research collaborations with clinical sites to test biomarker candidates for prognostic and therapy prediction assays.
We also collaborated with EMC to develop a device that allowed for the centrifugation of a Hettich Cyto chamber (a device often used in cytology) while keeping it inside a magnetic field. The field was generated by a permanent magnet built into a PMMA holder. The holder fits a Cyto chamber on top of a standard microscope slide and was designed specifically for this purpose but it could also be used without it. For the latter case we designed small cups that could be attached to a microscope slide using a non-permanent adhesive tape. These cups can be filled with immunomagnetically labelled cells, centrifuged with or without the magnet and they facilitate subsequent fixation, staining and/or lysing of cell samples. The method combines a magnetic field gradient with centrifugal force in order to optimize the recovery of immunomagnetically labelled cells by pulling them towards a microscope slide with a larger force than any of the two methods separately. The magnetic device can be taken out of the centrifuge together with the microscope slide containing the cells. This means that the cells never leave the magnetic field gradient and can be further processed while being kept firmly in place by the magnet.
This magnetic device was also used at Philips Research Eindhoven in another CTC project focussed on prostate cancer and aided in obtaining high CTC recoveries. In the future we may well use the device again for cytology related applications.
EMC
At Erasmus MC, a digital PCR assay was developed for sensitive detection of the 4 most frequent PIK3CA mutations on single CTCs purified using the Cytopicker from Cytotrack. The assay is being exploited for other scientific question as well and we have extended our repertoire of digital PCR assays for other relevant markers (e.g. ESR1), beyond the current scope of the project.
The clinical study will continue beyond the current project potentially providing impact using the assays developed within the CareMore project. The technological expertise (e.g. single cell PCR) gained within the project will be used in other grant applications.
Olink
The four oligo systems used in the four PLA assays in CareMore protein assay was developed initially in the DIATOOLS project (FP7-259796), except Duolink that was already a commercial product from Olink. During CareMore, the three additional oligo systems were optimised to function together in a higher specificity mode, and examined and cleared for cross-reporting. In summary, the stringency is higher now due to protocol optimisation performed by changing to a more specific ligation enzyme, increasing the amplification enzyme and increasing the temperature during parts of the protocol.
In addition, we now have eight antibodies and four protein assays that have gone through thorough testing. These protein assay, and especially the early developed ones pHER2 and ER, have been performing well particularly in the test systems with shorter transportation time than 24h and are now available for further use together or by themselves. Moreover, since the staining has not yet been performed on the clinical samples, there is a possibility to find even better assays but to use the reporting system optimised within the project.
Future projects focusses on developing an enzyme free detection system and therefore well suited for automatic staining instruments. The protein assays that the CareMore project developed, and thus all 8 antibodies involved, can now be used and detected with the future method, to be benchmarked to the Duolink system as a golden standard.
Stockholm University
The work on establishing in situ mutation detection in intact CTC:s did not result in a useful assay within this project, but was picked up as a new approach in the FP7 CANDO project. In this project in situ KRAS mutation detection has been achieved in CTC:s captured from pancreatic cancer patients on in vivo cell collectors from Gilupi, In parallel, SU has collaborated with researchers in Graz to detect a splice variant of the AR transcript in CTC:s captured from prostate cancer patients on in vivo collectors. A patent has been filed on in situ mutation detection in CTC:s captured on in vivo collectors.
List of Websites:
CareMore project home page: http://www.caremorectc.eu/
CytoTrack contact: service@cytotrack.com for further information about CTC testing service.
Philips contact: reinhold.wimberger-friedl@philips.com
Olink contact: http://www.olink.com/olink-bioscience/ Ann-Catrin Andersson, ann-catrin.andersson@olink.com (ann-catrin.andersson@comhem.se) Peter Åsberg peter.asberg@olinkbioscience.com Erik Ullerås erik.ullerås@olinkbioscience.com
EMC Contact: John Martens j.martens@erasmusmc.nl Jaco Kraan j.kraan@erasmusmc.nl Inge de Kruijff i.dekruijff@erasmusmc.nl
Stockholm University Contact: Mats Nilsson mats.nilsson@scilifelab.se
CareMore was a three-year project supported by the European Community's 7th Framework Program (FP7/2007-2013), under grant agreement n°601760 (CareMore). The main goal was to improve the diagnosis and subsequent treatment of women with metastatic breast cancer through advanced molecular analysis of circulating tumour cells (CTCs) in blood. Our many objectives were successfully met in most areas, as summarized below:
Multiplex in situ detection of 4 proteins in CTCs processed according to CellSearch®-based protocol developed in CareMore.
A four-plex in situ PLA™ assay was developed to report four target proteins in fluorescence in parallel with two identification proteins and a nuclear stain, in total thus seven colour fluorescence simultaneously on single cell CTCs. The target proteins were selected for comparison to the primary tumour (ER), for presence and activity status of HER2 (pHER2), and two proteins (p4E-bp and pS6) which were included during the project, to determine PIK3CA mutation, replacing the in-situ mRNA detection originally planned. By measuring the activated and phosphorylated forms of these two proteins found downstream in the signal transduction pathway of PIK3CA, we might have detected an indicator of potential therapy resistance. Clinical data collection is still ongoing as well as the six-month follow-up, due to delays by low patient inclusion and the extra proteins postponing the validation phase. All assay protocols and reagents are available, awaiting execution on the full patient numbers to be carried out outside the CareMore project. Each of these PLA-assays consists of two antibodies targeting the same protein, thus achieving a double recognition and consequently a higher specificity. In total four oligo-nucleotide reporter systems and eight antibodies were used in the PLA-assay in situ successfully.
Improved collection and isolation of CTCs from blood on two platforms.
A new method, implemented on the CytoTrack™ platform was benchmarked against the only clinically approved CTC-profiling method available today, i.e. CellSearch. This new method includes placing cells from a buffy coat on a glass-CD after pre-processing the blood sample and cytokeratin immunostaining, allowing for a fast two-minute scan on the glass CD to identify candidate CTCs. In addition, we were able to add immunostaining for EpCAM receptor. This platform was extended with a so-called CytoPicker™, a technology that allows for a single cell isolation step, whereby characterized CTCs can be picked up and placed in PCR tubes. These cells are then analysed by digital PCR in a single cell analysis to detect the different mutation variations of the PIK3Ca gene.
The existing CellSearch platform at EMC was also extended to allow for a downstream analysis of four target proteins. Starting from the CellSearch Profile kit a more efficient way of collecting and isolating CTCs was developed, whereby a larger quantity of CTCs could be collected from a larger blood volume. The CareMore protocol also includes a modified fixation before arranging the cells on a standard microscope slide through magnetically-assisted centrifugation step. This allows the slides to be scanned in a whole slide fluorescence digital scanner.
Multiplex detection through fluorescence microscopy scanning and subsequent image analysis.
The project planned for and successfully detected seven markers at the same time on the slides from CareMore processed CellSearch samples. To that end all fluorescent labels were carefully selected for their spectral properties in combination with selected light sources and emission filters to be able to quantify the respective signals without interference. Also the cellular location of the biomarker was considered when allocating the different signals to each biomarker unambiguously.
We developed algorithms and established an image analysis pipeline that can pre-process 7-channel fluorescence images from the digital slide scanner, resulting in a score for each biomarker in each cell and cellular compartment. By using reference cell lines that were selected as positive and negative controls, we were able to calibrate the PLA assays. The algorithms can be applied to any target in a multiplex fluorescence assay as long as reference cell lines with a ground truth biomarker expression are available for calibration.
Clinical outcome.
All patient samples from the two clinical studies that were designed and initiated are stored until the completion of the studies and the CareMore CellSearch protocol and CareMore PLA assay can be performed without further modification. It was decided to not run the assay on a smaller group of patients to ensure that the exactly same assay conditions will be applied to all samples and reliable conclusions can be drawn. The studies without the CareMore assay will still reveal if patients with a HER2-positive CTC-sample and a HER2-negative primary tumour benefitted from a trastuzumab treatment, which would not have been given without HER-characterization performed in Study2. We are confident that the tools and data produced by the CareMore project will trigger others to set up molecular diagnostic tests on CTCs for therapy selection in patients with metastatic breast cancer.
Project Context and Objectives:
BACKGROUND ON DIAGNOSIS FOR METASTATIC BREAST CANCER
Today's treatment of metastatic breast cancer is guided by characterisation of the primary tumour, while 90% of deaths due to breast cancer occur as a consequence of metastases. Given important differences in molecular characteristics between metastatic and primary tumour tissue, characterization of metastatic tumour tissue instead of the primary tumour may provide better treatment guidance for patients with metastatic disease. Metastatic tumours are, however, difficult to access using currently available techniques. In the CareMore project, we proposed to guide treatment decisions in women with metastatic breast cancer based on molecular characterization of circulating tumour cells (CTC), used as proxies for the metastatic tumour. From a clinical point of view, mutations in PIK3CA are considered to be important biomarkers that can help the physician to guide the treatment of the patient, i.e. to choose the right treatment at the right point in time. For the sensitive detection of genetic mutations in circulating tumour cells two unique methods were combined, i.e. cell-based detection of genetic mutations and improved circulating tumour cell sampling and identification. Both techniques were provided by the participating SMEs Olink and CytoTrack. Together with leading academic and clinical scientists at the Erasmus University Medical Center and an advanced industrial partner, Philips, the consortium was ideally suited to develop and validate the methods in order to reach proof of concept stage before the methods can be tested for clinical utility, the final aim of the CareMore project.
CONCEPTS AND OBJECTIVES
To advance Circulating Tumour Cell (CTC)-based cancer diagnostics as the original objective, the CareMore project aimed high by developing assays to be applied in two parallel prospective clinical studies on women with advanced breast cancer.
More specifically, we aimed to advance CTC-based cancer diagnostics by:
• Introducing multiplex in situ testing for activating mutations in PIK3CA gene transcripts, for HER2 receptor activation and for estrogen receptor (ER) protein expression using Padlock Probe (PLP) technology (Stockholm University) and Proximity Ligation Assay (PLA®) technology (Olink) in combination with digital slide scanning and separating 8 colours in situ by image analysis (Philips);
• Improving the collection of CTCs (Cytotrack) by increasing the number of isolated CTCs from each patient beneficial for subsequent molecular characterisation and by increasing speed of isolation and identification;
• Overcoming the current limitation of EpCAM dependent enrichment and benchmark this approach against the CellSearch® technology of Janssen Diagnostics, LLC. (EMC and CytoTrack);
• Verifying the clinical utility of these biomarkers in CTCs for therapy guidance in patients with metastatic breast cancer in two prospective clinical studies, study 1 and 2, with clinical follow-up data collected 6 months after the start of therapy (EMC and Philips).
During the first phase of the project it turned out that the PLP assay was not sensitive enough for the targets to be detected (i.e. mutant PIK3CA transcripts). Therefore, we abandoned this strategy and entirely changed the method to determine activation of the PIK3CA pathway in two ways:
• Detect two phosphorylated proteins, pS6 and p4E-BP1, that signal downstream of PIK3CA. The extra protein targets resulted in a 4-plex instead of 2-plex protein detection assay. This assay was developed using PLA technology.
• Detect the PIK3CA mutations using a single-cell digital PCR assay on isolated CTCs. This was enabled by a cell picking methodology developed by CytoTrack to isolate single CTCs after staining and identification, resulting in the Cytopicker™, an integrated part of the CT11™ scanner.
In addition, it was decided to include EpCAM staining in the clinical fluorescent staining assay performed at CytoTrack, to gain knowledge on the percentage of CTCs (determined by CK staining) lacking EpCAM and to allow benchmarking the CytoTrack assay against CTC enumeration by the FDA approved standard Cellsearch assay.
To determine clinical validity of the developed CareMore assays 2 prospective clinical trials were executed within the project. Study 1 aimed at assessing the impact of HER2 status in CTCs on outcome to aromatase-inhibitors in metastatic breast cancer patients with ER-positive/HER2 negative primary tumours; and study 2 aimed at determining the success rate of trastuzumab-based chemotherapy in metastatic breast patients with HER2-negative primary tumours but HER2-positive CTCs. The inclusion criteria and the objectives are listed below:
Study 1
Metastatic breast cancer postmenopausal patients with ER-positive/HER2-negative primary tumours, to receive treatment with aromatase-inhibitors.
• Primary objective: to determine the impact of HER2 expression in CTCs taken at base-line on the clinical outcome with respect to treatment aromatase-inhibitors.
• Secondary objectives: to determine the impact of ER, pHER2, and PIK3CA mutational status in CTCs taken at baseline on the clinical outcome to aromatase-inhibitors treatment; separate analyses for the 3 factors and a combined analysis.
Study 2
Metastatic breast cancer pre- and postmenopausal patients with HER2-negative primary tumours but with HER2-positive CTCs, to receive treatment with trastuzumab-based chemotherapy.
• Primary objective: to determine if these patients benefit from trastuzumab-containing chemotherapy
• Secondary objective: to determine impact of pHER2 and PIK3CA mutations in CTCs on the clinical outcome to trastuzumab based chemotherapy.
The final aim of the project is to show the value of the developed assays by applying them directly to clinical specimens in the two clinical studies executed within the CareMore project. The studies aim to determine the assays’ power to predict outcome of metastatic breast cancer patients treated with endocrine or Herceptin based targeted therapy.
Project Results:
The main goal of the CareMore project was to develop methods to better guide treatment of women with metastatic breast cancer. Nowadays treatment options are based on the characteristics of the primary tumour. However, metastases often show discordant features when compared to their primary tumour. Within CareMore, we aim to use CTCs as proxies for metastases and investigate if indeed their characteristics are deviating in a similar fashion. In view of the substantial discrepancies that exist between primary tumour tissue and metastatic cells, it is likely that treatment decision-making in metastatic breast cancer based on molecular characterization of CTCs is superior to decision-making based on primary tumour tissue, as is currently most frequently done. CareMore will provide means, far beyond current state-of-the-art, to characterize CTCs from patients and score markers known to affect treatment outcome and provide it in a format suitable for clinical use.
The CareMore project aimed to improve the state of the art in the following seven areas:
1. Clinical validation of current research-use-only PLA and PLP technology on CTCs to determine (1) activation status of the HER2 receptor, and (2) presence of PIK3CA mutations.
2. Clinical validation of existing proprietary instrumentation and algorithms for non-EpCAM dependent isolation and analysis of CTCs.
3. Development of a prototype instrument, based on the existing CytoTrack instrument, suitable for next generation molecular characterization methods of CTCs.
4. Development and validation of a complete workflow for clinical application relating to techniques and quality control criteria for isolation and preparation of CTCs with preserved mRNA quality (used for genotyping), multiplex in situ staining of CTCs, as well as software for (automated) fluorescence image analysis and subsequent clinical interpretation.
5. Validation of therapy response markers in patients with metastatic breast cancer.
6. Generation of a prototype diagnostic systems for therapy guidance in post-menopausal breast cancer patients with ER-positive but HER2 negative primary tumours.
7. Generation of a prototype diagnostic system for therapy guidance in patients with metastatic breast cancer with HER2-negative primary tumours but HER2-positive CTCs.
Below the scientific development and technological achievements within the project are outlined, subdivided in three different important areas:
• Improve the collection of CTCs using two different collection methods
• Developing the multiplex in situ assay (the CareMore assay)
• Clinical utility for better management of metastatic breast cancer patients
During the CareMore project, 2 methods (called project “arms”) were developed for CTC isolation and characterization and validated on blood spiked cell lines and patient samples:
• The CellSearch Arm (EpCAM dependent) consists of the following sample processing steps: CellSearch enumeration and parallel isolation of CTCs from 30 mL of blood followed by transferring of enriched cells to a microscope slide with minimal cell loss using a magnetic device developed by Philips and EMC.
• The CytoTrack Arm (EpCAM independent) consists of the following sample processing steps: Cytotrack CTC isolation and enumeration (including EpCAM characterization) followed by single cell picking and determination of the PIK3CA mutational status using digital PCR.
For the two investigated CTC-collection platforms, CytoTrack and CellSearch, collection protocols had to be improved and adjusted, each to be tailor made for both mRNA or protein quantification. The project aimed to detect in one and the same sample preparation in situ all markers in the same cell on both platforms.
1. CTC enumeration and characterization using a multiplex in situ assay (CareMore assay) in the CellSearch arm
The objective of the CellSearch arm was to enumerate CTCs using the FDA approved CellSearch platform followed by isolation and preparation of CTCs for downstream protein and molecular characterization.
1.1.1 Protocol development and optimization
To increase the CTC yield, a method was developed and validated to isolate CTCs for downstream analysis from 30 mL of blood instead of the traditional 7,5 mL. The same was done in the CytoTrack arm.
After Immuno-magnetic enrichment of CTCs from (spiked) blood samples using the CellSearch profile kit according to manufactures instructions, the enriched cells were transferred to a microscope slide with minimal cell loss using a magnetic device developed by Philips/EMC. This assay was validated for optimal recovery with spiked reference samples in healthy donor blood.
For optimal preservation of CTCs on slide for both protein and molecular (DNA and RNA) characterization many different fixation methods, incubation times and storage conditions for blood samples and slides were tested using reference samples followed by multiples PLA and PLP assays.
1.1.2. Improved transfer of CellSearch-enriched CTC slides
Different ways to transfer the CellSearch enriched samples to slides were investigated to increase CTC recovery. Also different fixation methods of cells were compared as well as their effect on the stability of stored analytes. The developed protocol was finalised before patient inclusion was started in 2015. The method finally developed allowed the cells to be cytospun on a slide and thus easier to be investigated with subsequent detection techniques, i.e. the CareMore protein staining assay and automated slide scanning. The end product of CTCs cytospun on a slide was used in the proof-of-concept validation using selected reference cell lines spiked in healthy blood.
Before continuing on clinical material a proof-of-concept test was performed with the newly developed protocol using CellSearch-enriched fractions which were cytospun on slides to ensure CTC recovery was feasible with limited loss using the new procedure.
We were able to establish a fixation method that preserved both proteins and mRNA in cells spiked in blood.
In conclusion, the storage conditions for the CareMore protocol as well as a protocol to successfully transfer samples to standard microscope slides were defined in the beginning of 2015 after which inclusion of the first patient samples was started.
1.2. Developing the multiplex in situ assay (the CareMore assay)
The original objectives were to improve the collection and identification of CTCs, and subsequently detect two selected proteins (estrogen receptor (ER) and phosphorylated Human Epidermal growth factor Receptor 2 (pHER2)) at the same time as detecting 4 mutations (in mRNA) from a potential cancer marker gene (PIK3CA). Mutations in PIK3CA are considered to be an important biomarker that might help physicians guide the treatment of a patient, i.e. to choose the right treatment at the right point of time. The in situ mutation detection assay to be developed in CareMore is a PLP assay from Olink developed with knowledge from Stockholm University (SU), and will allow for the identification of Phosphoinositide 3-kinase (PI3K; the protein product of the PIK3CA gene) activating PIK3CA mutations. At the same time, using the PLA® technology ER- and pHER2-proteins will be visualized in CTCs isolated from blood. Detection was planned to be in situ with subsequent visualization using a whole slide microscope scanner which could handle 8-9 colours simultaneously. The protein assay was finally developed to use 7 colours of which 4 instead of 2 proteins are quantifiable in situ. The PLP mRNA mutation analysis was swapped for a single cell PCR.
1.2.1 multiplex PIK3CA mutation detection at the mRNA level using PLP
The original objective was to detect mRNA from the PIK3CA gene, and its mutated variants, in situ at the same time as detecting two proteins (ER and pHER2). Padlock probes (PLP) have been used to detect mRNA in tissue before, and in this project its performance on CTCs was investigated. Thorough investigation however did not lead to the successful in situ detection of PIK3CA mRNA in any form. It is still unclear why the PIK3CA mutant transcript detection did not work, since PLP is a proven technology and has been used before to detect point mutations in transcripts of other genes expressed at same level as PIK3CA in situ on microscope slides.
The efforts of detecting the different PIK3CA mutations using mRNA PLP was a main focus for almost two years. In parallel to the PIK3Ca work, we used PLP on well-known genes to find a common fixation method for proteins and mRNA together. Moreover, many other different approaches were investigated with the goal to detect PIK3CA mRNA but without success. Among other things, we developed a more stringent PLA-protocol regarding specificity because of the requirement that the PLP assay should work simultaneous with the PLA-assays. The original objective was thus changed from in situ to in solution detection after a project decision, explained in more detail in the single cell PCR section.
Parallel to the project, the group at SU has also investigated several more genes at the same expression level of PIK3CA in combination with several ways to isolate CTCs from blood, i.e. from different platforms available on the market. This research is to be presented in other EU-funded projects and will add to the knowledge of how different genes behave differently in biological samples. The knowledge gained will thus be used to continue development of the method, for other selected target genes.
The PLAs developed within CareMore are described below, as well as the respective downstream analysis for the two platforms.
1.3.1 multiplex protein detection using PLA
When the PIK3CA mRNA turned out to be impossible to detect (see above), the project decided jointly to add two extra protein targets to the PLA-assay as a compensation, and at a later state we also tried to detect PIK3CA but using single cell PCR (see below). The two added proteins were selected to reflect the activation status of PIK3CA by detecting proteins that are activated downstream of PI3K (the PIK3CA protein product) signalling.
The specificity of the protein assays was investigated on cell lines, both before and after the fixation conditions were established. Since originally there was also a need to adjust the PLA staining protocol to function in tandem with the PLP assay, necessary optimisations were made to the protocol to, despite of the already specific dual recognition of the antibodies, change to a more specific ligase enzyme and to make the annealing temperature higher to prevent any non-specific hybridisation.
The reference cell lines that were used as controls for the protein assays were selected based on their known status regarding the proteins of interest. When ER was concerned, there were already known guide lines for staining and subsequent interpretation of the results, even though the IVD was developed for use on tissue. In like manner, a lot is known about the expression of HER2, but less is known about the HER2 pool which is actually phosphorylated. As for the two lesser known proteins, phospho-S6 (pS6) and phospho-4E-binding protein 1 (p4E-BP1), a selection of reference cell lines was made after an extensive screening of about 55 cell lines per protein The challenge with these abundantly expressed proteins was to find cell lines that could be used as negative controls. Thus, we chose low expressing reference cell lines as our negative controls for these latter 2 proteins.
Multiplex staining and detection using fluorescent labels necessary for quantification of the 4 proteins, i.e. ER, pHER2, pS6 and p4E-BP1, was performed. The four respective proteins were detected through PLA® technology using 4 oligo detection systems whereby each protein is bound by 2 antibodies (4 times 2 oligonucleotides for conjugation). Each antibody is conjugated with the respective oligo and each set of 2 matching oligos is amplified by rolling circle amplification followed by simultaneous detection using specific probes each having an excitation at a distinctive fluorescent wavelength. One major aim of the CareMore project was to analyse PIK3CA mutations in situ in the CTCs and simultaneously detect ER and phosphorylated HER2. Since we decided to not proceed with the in situ PLP, we instead expanded the PLA assay to also include two additional proteins to represent activation of the PI3K pathway by the downstream proteins pS6 and p4E-bp by PLA. unfortunately, it turned out that the 4-plex PLA could not recognize its epitopes in the frozen samples prepared for the CytoTrack platform. Therefore, the CareMore protein assay was used on the CellSearch platform purified samples only. Instead, the PIK3CA single cell PCR analyses were performed on the CytoTrack platform in the final proof-of-concept validation of mRNA and proteins, respectively.
Within the CareMore project a 4-plex PLA was developed, enabling quantification of 4 proteins simultaneously at single CTC level. In addition to the 4 proteins to be detected by PLA, immunofluorescence was used to simultaneously detect cytokeratin (CK) and CD45 while DAPI was used to detect cellular nuclei. CK and CD45 are used in downstream analysis to separated CTCs, which are CK-positive and CD45-negative from leukocytes, which are CK-negative and CD45-positive. For 3 of the 4 proteins included in the evaluation tests the aim was to detect the phosphorylation status, which is difficult to detect in blood stored for a long time. However, when using PLA, the signal is amplified and quantifiable. To ensure the detected cells are CTCs, the analysed cells must show a cell-shaped nuclei in the DAPI channel, a distinct Cytokeratin staining and no CD45 staining.
This project introduces a completely novel method to quantify 4 proteins simultaneously, in situ in individual CTCs. This is enabled by using 4 different PLAs, usable in parallel, developed by Olink. The PLA assays are immunoassays where each protein is detected using two antibodies. In order to work simultaneously, the 8 antibodies binding the 4 protein targets in a PLA are directly conjugated with DNA oligonucleotides unique to the 4 labelling systems. These oligo systems are not locked to the specific antibodies and their protein targets tested here and can be transferred to other markers of interest. As long as there are specific antibodies in a purified state, they can be used in any of the 4 oligo systems, in any combination.
1.4. Conclusion for protein Multiplex assays
The PLAs work as a proof-of-concept evaluation and can report quantitative values for 4 of the targets in cell lines spiked in blood. Two of the targets consist of epitopes that seem to be severely affected by the long storage time before sample preparation, thus resulting in values that are too low to be reliable. In addition, the 4 oligo systems are proven to be able to report only after dual antibody recognition, which should enable higher specificity. In the future, the reporting systems can be used on epitopes that are investigated to be more stable than for example ER.
2. The CytoTrack arm for CTC enumeration, EpCAM quantification and molecular characterization by single cell PCR
The objective for the CytoTrack arm was to enumerate CTCs, to stain them for EpCAM as well as various molecular markers of the PI3K pathway. Improvements which were made are discussed in detail below and relate to:
1. Blood sampling and preservation
2. CTC enumeration, including EpCAM characterization
3. Retracing, CytoPicking, and PIK3CA mutation analysis
2.1.1. Blood sampling for the CytoTrack arm
The average amount of CTCs in 10 ml blood is low, often just a few CTCs are detected. There is therefore a widespread but until now unmet need for processing larger blood sample volumes. Recently, we showed that modifying the sample preparation techniques by adding a pre-processing step to enrich for mononuclear cells, CTC recovery could be increased to 300% because 30 mL of whole blood could be analysed instead of the traditional 10 ml.
This application is commercially available as a supplementary CytoTrack application note.
2.1.2. Blood preservation for the CytoTrack arm
A major challenge in the detection, enumeration and characterization of CTCs in blood samples is the limited storage potential of commercial blood sample tubes. For biobanking, scientific or clinical purposes, preservation can provide opportunities and allow batch processing of blood samples over time in the future. At CytoTrack, a method for cryopreservation of CTCs maintaining compatibility with the CytoTrack instrument and methods of detection, enumeration and characterization was successfully developed and validation showing CTC recoveries of 89-91%.
This cryopreservation method was used in the clinical study and published in Biopreservation & Biobanking 2016.
2.2.1. Development of a CytoDisc staining lid for on disc in-situ assays
At Philips Research a novel staining method for CytoDiscs was developed. It comprised a lid that could be applied to the disc after the buffy coat was applied and dried. The lid contains a 1 cm hole in the center and is placed on 6-8 round spacers of 50-180 µm thick with diameters of about 5 mm. This creates a cavity above the dried buffy coat with the height of the spacers. When the disc with the lid was placed on a spin coating device the whole system could be spun around at various speeds. When liquids were applied to the centre hole in the lid, the spinning of the disc would force the liquid into and out of the cavity where the buffy coat is. The flow stopped automatically when the hole was completely empty due to the capillary force acting inside the cavity. In this way liquids could sequentially be applied to the CytoDisc surface without the risk of drying the cells. By sequentially applying different liquids like washing and antibody solutions combined with incubation steps, staining assays could be efficiently performed with minimal use of reagents. After an assay the disc and lid could be rotated at high speeds to dry the sample completely before applying a coverslip. The staining lid system was developed for a short amount of time but abandoned because the effort to get it at the same functional level as normal, manual on-disc staining was judged to be too great for this project alone.
2.2.2. CytoTrack concept for retracing individual CTCs on CytoDisc
Both On-Disc and Off-Disc characterization require retracing of individual CTCs, i.e. the ability to find the original location of each CTC after removing the CytoDisc and re-inserting it into the CytoTrack instrument. This was accomplished using a ‘notch’ on the design of the CytoDisc, such that it locks into the same position every time. The location of each CTC is displayed on a digital ‘CytoMap’, which may be printed to serve as template for the CytoDisc. The print works as an underlay to the CytoDisc, and shows where the target cells are on the disc. In this way only the areas where the cells are positioned can be stained, reducing the use of (expensive) staining reagents.
Validation of this concept showed >96% recovery and the results were published in Journal of Circulating Biomarkers.
2.2.3. CytoTrack On-Disc characterization (for using HER2 as PoP)
The On-Disc procedure involves the disc to be inserted into the CytoTrack instrument twice. The first procedure is to detect the target cells on the disc, i.e. to identify the position of the target cells (digital mapping), and to enumerate the number of CTCs. Hereafter, the disc is taken out of the instrument, and the CoverDisc is removed. Now the cells on the CytoDisc are stained, either with IF or FISH. A new CoverDisc is applied, and the CytoDisc is reinserted into the instrument. The instrument can now go directly back to the position of CTCs on the disc, since the positions on the disc are already captured and stored. Now it is possible to detect the additional staining and characterisation of the target cells. CTC retracing and the associated digital CTC map form the basis for downstream characterization of CTCs while they are attached to the CytoDisc. The characterization reagents are applied to the surface of the CytoDisc and may be either antibodies (IF) or fluorescence in situ hybridization (FISH) probes. Results are read after reinserting the CytoDisc into the CytoTrack instrument. The concept was developed using HER2 as biomarker as HER2 is both an important breast cancer marker with therapeutic application and a biomarker that may be measured by both immunofluorescence IF and FISH.
2.2.4. On-disc EpCAM characterization of CTCs using multiplex immunofluorescence
EpCAM characterization of CTCs as a multiplex assay component combined with CTC detection was developed as a means to simultaneously measure EpCAM expression levels in individual CTCs and was included into the CareMore clinical study. The accomplishment has been commercialized as a CTC StainEpCAM reagent and the associated protocol and clinical performance have, among others, been presented at the European Cancer Congress in Vienna in 2015.
As part of the protocol all patient samples are screened for HER2-positivity by the CellSearch HER2-kit. In the project all patients have also been tested in parallel on disc for the same markers including EpCAM. The on-disc assay allows us to enumerate all cells stained by CK irrespective of whether cells also stain positive for the CellSearch capturing marker EpCAM. The amount of CTCs identified by each of the two methods was also compared.
2.3.1. Off-Disc characterization: CytoPicker for isolation of individual CTCs
A major accomplishment was the discovery that individual CTCs may be ‘lifted’ up from the CytoDisc using a combination of micromanipulation tools and proprietary ‘cell picking’ methodology.
A cell picking device was developed during the project that could transfer stained single CTCs from a disc in the CT11™ scanner and into a PCR tube. Using the CytoPicker innovation, the picking of individual CTCs was optimized to a point where more than 80% of CTCs could be picked within 2-8 min per CTC and free of contaminating white blood cells deposited into (for example) PCR tubes for subsequent analysis. The achievement of this milestone forms the basis for single cell analysis of PIK3CA mutations; the last and equally challenging step towards completing the CytoTrack arm with single CTC analysis.
This innovation was built into the CytoTrack instrument and presented to the public for the first time at the 7th International Conference of Contemporary Oncology, Poznan, Poland, in March 2015 after protecting the intellectual property rights with a patent application. Validation studies were used to generate specifications for the CytoPicker that among others are summarized in a CytoTrack brochure/flyer.
The CytoPicker can be used to isolate single target cells. The cells are lifted from the disc surface and are then placed in PCR tubes (one cell per tube). This is done without any contamination of white blood cells.
2.3.2 Single cell PCR for PIK3CA mutations for individual CTCs
As an alternative to the PLP assay, we developed a procedure to detect PIK3CA mutations in clinical samples on single cells using a digital PCR method analysing 4 common activating PIK3CA mutations (leading to the following amino acid changes in PIK3CA: E542K, E545K, H1047R and H1047L). We were able to validate the procedure on control cell lines with known mutations spiked into healthy donor blood as well as to successfully determine the PIK3CA mutation status on single CTCs picked by the CytoTrack system, from 5 patients. To conclude, a digital PCR method covering the 4 most common PIK3CA mutations was successfully developed and validated on cell lines with known mutational status spiked in healthy blood samples which consider an important accomplishment, moreover, it completed the development of the CytoTrack arm.
2.4. General conclusion for the Cytotrack arm
A new model of the CytoTrack instrument and corresponding CTC characterization techniques have successfully been developed and validated for use in the clinical validation study and in parallel was made commercially available for research use only.
This complete solution integrated into a compact bench-top instrument which allows CTCs to be detected, enumerated and characterized either by multiplexing or by downstream characterization performed either on the CytoDisc (On-Disc) or after isolating the individual CTCs.
For the CytoTrack instrument the following novelties were developed: (1) Concept for retracing individual CTCs on CytoDisc, (2) CytoPicker for isolation of individual CTCs and (3) CTC rating and improved imaging for faster results. With regard to methodology the following improvements were realised: (1) Sample volume to be used increased to 30 ml, (2) Cryopreservation of blood samples, (3) multiplex EpCAM characterization of CTCs, (4) On-Disc characterization, (5) Off-Disc characterization followed by CytoPicking and (6) Single cell characterisation by PCR for PIK3CA mutations for individual CTCs.
In summary, the CytoTrack arm comprises the key accomplishment described above. It has successfully been developed and validated for use in the clinical validation study.
3.1 Multiplexed detection of the developed CareMore situ assay
In order to achieve detection of 7 biomarkers, i.e. Cell nucleus, CK, CD45, ER, pHER2, p4E-BP1 and pS6 on a single microscope slide, we used 7 fluorophores with well distinguishable spectral properties: DAPI, Atto488, Atto647n, Atto550, Atto490LS, Atto430LS and Atto590, respectively. DAPI is used to detect cell nuclei and is the basis of the cell identification algorithm. Based on DAPI a nuclear and cytosolic/membrane compartment is defined. CK is used to identify cells of epithelial origin, while CD45 indicates that a cell originates from the blood lineage. CTCs are defined as CK positive and CD45 negative cells. Both CK and CD45 are stained using standard immunofluorescence (IF) protocols. The remaining protein markers (ER, pHER2, p4E-BP1 and pS6) provide further characterization of the CTCs and are stained using PLAs as described above.
The 7 fluorophores were chosen based on the spectral constraints of the light source of the fluorescent scanner which required them to be excitable in the 400 - 650 nm range. The fluorophores mentioned above meet this requirement. We made use of the available spectrum to the fullest by adding 2 long Stokes shift (LS) fluorophores to the set (Atto430LS and Atto490LS). These molecules show similar excitation spectra to DAPI and Atto488, respectively, but their emission is much more redshifted. Consequently, they can be spectrally separated from the other fluorophores quite well using proper emission filters.
Both the excitation and emission spectra of the fluorophores typically have a spectral width of approximately 50 nm (full width at half maximum). This means that there is considerable spectral overlap between them. Combined with the fact that we are bound by what filter set suppliers can offer, typically leading to slightly sub-optimal wavelength cut-offs, we were left with significant crosstalk between the 7 detection channels. To minimize the risk of misinterpreting signals, we assigned the channels with the strongest cross-talk to markers that were least likely to co-localize, e.g. pHER2-Atto490LS leaks into ER-Atto550 but pHER2 is membrane localized while ER is predominantly present in the nucleus. The filters were designed, purchased and mounted in the scanner. Cross talk analysis revealed small but unfavourable effects that could be solved by biochemically tuning relative signal brightness and by characterizing said cross talk and applying a correcting algorithm that was developed in house as part of the image analysis pipeline.
Thus through carefully selected filters combined with a multi-band LED light source, 7 fluorophores could be separately analysed after being scanned in a whole slide fluorescent scanner. Judging the fluorophores according to the known cellular localization of the targets helped the developed software to correct for the leakage of signals between a few of the filters. Thresholds for potential cut-offs for a positive and negative signal for every marker were determined by using the selected reference cell lines as positive and negative controls. Combined with an automated detection of each cellular nucleus the scores could be determined for each cell compartment and allowed signal intensity histograms to be produced for each sample. In this way an assay was judged by comparing the signal patterns with the expected values for the ground truth samples.
In conclusion, all 7 biomarkers could be detected accurately.
3.2. Automated scoring of the developed CareMore in situ assay
The objective of the scoring part was to provide algorithms that: (1) Extract required parameters from the multicolour fluorescence images with dot-like signals obtained from the stained samples, and (2) Stratify the patients according to the test results.
There were several different applications for which algorithms have been developed. Due to the need for a high multiplex (7 signals) and spectral limitations in the excitation light source, the fluorescence filter sets and the CCD detector system, spectral overlap between some fluorophores was unavoidable. This resulted in crosstalk of signals. This crosstalk was quantified and an algorithm was developed that minimizes it, resulting in properly separated signals that could be further processed. To detect the dot-like signals an algorithm was developed that minimizes the influence of interfering signals and that finds the peak of each dot. To identify CTCs an algorithm that detects nuclei was developed together with an algorithm that estimates the location of cell membranes. The nucleus detection algorithm was aimed at being independent of any possible background signals. The estimation of the cell membranes was done by taking the nuclei outlines as a starting point, and then increasing the outlines in size. These detected nuclei and estimated membranes could then be used to quantify the dot-like or immunofluorescence signals within the created compartment boundaries. All these algorithms were combined to quantify the validation samples for the CareMore assay. Those quantification results were extensively visualized and statistically analysed to assess the performance of the assay.
No patient stratification was done due to low patient inclusion in the clinical trial and the decision to not stain patient samples yet with the CareMore assay. In conclusion, the algorithms were successfully applied to assess the performance of the assay. Unfortunately, due to low patient inclusion in the clinical trial, no stratification of patients could be performed.
4. Clinical utility testing
4.1. Introduction
To collect patient samples for clinical validation of the designed assays and to obtain clinical results, two clinical studies were designed. Both studies were approved by the ethical board of the Erasmus MC in Rotterdam. Next, the rationale of the studies and a description of both studies is given.
Metastatic breast cancer patients are still treated on primary tumour characteristics, while it is now increasingly appreciated that important characteristics like ER- and HER2-expression can differ between the primary tumour and the metastasis. Circulating tumour cells (CTCs) are present in the peripheral blood of patients with metastatic breast cancer and are thought to represent the characteristics of the metastases. From previous studies, we know that 18% of the patients with an HER2-negative primary tumour do have HER2-positive CTCs.
CareMore study 1
From literature it is known that patients with breast cancer with a ER-positive and HER2-postive primary tumour have a worse response on endocrine therapy compared to patients with an ER-positive but HER2-negative primary tumour. Therefore, we hypothesize that patients with an ER-positive and HER2-negative primary tumour, but with HER2-positive CTCs will also have a worse response to endocrine therapy. So the primary objective of this study is to determine the impact of HER2-expression in CTCs taken at baseline on outcome to aromatase inhibitor (AI) therapy in metastatic breast cancer patients with an ER-positive primary tumour. To answer this primary objective, 30 patients with HER2-positive CTC’s will be enrolled in this study. From all patients with breast cancer, 65% has detectable CTCs. From the patients with detectable CTCs, 28% has HER2-positive CTCs. Therefore, to include 30 patients with HER2-positive CTCs in this study, 165 patients have to be screened. Ethical approval of this study was obtained at the 10th of February 2015. The clinical study opened in March 2015 and the first patients were included in the same month.
CareMore study 2
Trastuzumab is a targeted therapy, directed against HER2. This therapy is given to breast cancer patients with a HER2-positive primary tumour. There is accumulating evidence that there are patients with a HER2-negative primary tumour who respond to trastuzumab-based chemotherapy. A group of patients who might benefit from trastuzumab-based therapy are patients with HER2-positive CTCs. Therefore, we hypothesize that patients with a HER2-negative primary tumour but with at least one HER2-postive CTC will benefit from HER2-targeted therapy. The primary objective of this study is to determine if metastatic breast cancer patients with HER2-negative primary tumors but at least one HER2-positive CTC benefit from chemotherapy with trastuzumab. To answer this primary objective, 18 patients should be treated with trastuzumab. Therefore, 99 patients should be screened for HER2-positive CTCs. Ethical approval of this study was obtained at the 18th of February 2015. The clinical study opened in March 2015 and the first patients were included in the same month.
4.2. Patient inclusion
Patient inclusion CareMore study 1
The following hospitals from the Netherlands and Belgium are participating in the CareMore study 1:
• Erasmus MC (Rotterdam)
• Ikazia Hospital (Rotterdam)
• Saint Franciscus Gasthuis (Rotterdam)
• Jeroen Bosch Hospital (Den Bosch)
• Albert Schweitzer Hospital (Dordrecht)
• Gelre Hospital (Apeldoorn)
• Vlietland Hospital (Schiedam)
• GZA Hospitals (Antwerp)
As mentioned before, 165 patients should be screened in this study. The result of the HER2-CTC enumeration is not reported back to the hospitals. We will look at this data at the end of the enrolment of the 165 patients. Currently, 54 patients are included in the study. We did not manage to include all 165 patients in the three-year period of the CareMore-project. Mostly due to the fact that the clinical studies could not be opened before March 2015, as the first 14 months of the project were needed to optimize and validate patient sample collection and preservation methods for downstream analysis, so the study was only open for 18 months and patient accrual was slower than expected. We will continue to collect samples of the patients after the CareMore-project has finished. All participating hospitals will join in the completion of the patients’ sample collection, and even new centres are currently being attracted to participate in this study, to further boost accrual.
Patient inclusion CareMore study 2
The following hospitals from the Netherlands are participating in the CareMore study 2:
• Erasmus MC (Rotterdam)
• Ikazia Hospital (Rotterdam)
• Saint Franciscus Gasthuis (Rotterdam)
• Vlietland Hospital (Schiedam)
In total, 29 patients have been screened for HER2-positive CTCs, of those only 4 patients have been treated or are currently being treated with chemotherapy and trastuzumab. This means 14% of the patients had HER2-positive CTCs, this is a bit lower than the expected 18%. But this number will probably increase with a higher number of inclusions. The aim was to screen 99 patients in the CareMore-project. We were unable to fulfil this aim, because also this study was only open for 18 months of the CareMore-project, because of the mentioned reasons above. Also for the CareMore-trastuzumab, accrual will continue after the CareMore-Project has finished. All participating hospitals, and also some new hospitals will contribute to the patient inclusion.
4.3. Patient sample processing
Patient sample processing CareMore study 1
In the CareMore study 1, 80 mL of blood was drawn from all patients at baseline. The blood is collected in 2 x 10 mL CellSave tubes and 6 x 10 mL EDTA tubes. After a sample is drawn, the blood is collected by a courier at the participating hospitals, allowing the blood to arrive in time in the laboratory to be processed within the required 24 hours. For all patient samples, we did manage to process the EDTA blood within 24 hours.
The 2 CellSave tubes are collected for CTC enumeration. One tube is processed in the laboratory for Translational Cancer Genomics and Proteomics in the Erasmus MC in Rotterdam, the Netherlands. Here the CellSave tube is processed in the only FDA-approved CTC enumeration system: the CellSearch system (Janssen Diagnostics, Raritan, NJ, USA). During the enumeration, a HER2-staining is performed, so HER2-positive and HER2-negative CTCs can be determined. The second CellSave tube is send to the CytoTrack Laboratory in Denmark. There CTCs are enumerated using the CytoTrack system and after enumeration, single CTCs are picked and processed for further molecular characterization. The EDTA tubes were processed in the laboratory for Translational Cancer Genomics and Proteomics in the Erasmus MC in Rotterdam, the Netherlands. From 30 mL of EDTA blood CTCs were isolated using the CellSearch system and after isolation the CTCs were processed on a glass slide for further PLA processing, the other 30 mL of blood was stored in DMSO so this could be send to CytoTrack for single cell picking in batches.
We were able to collect all the samples from the patients that participated in the study. We were also able to send the CellSave tubes to the CytoTrack Laboratory in Denmark and process the samples there within 96 hours. There were no major issues in processing the samples and all samples have been stored for further processing.
Patient sample processing CareMore study 2
In the CareMore study 2, 20 mL of CellSave blood is collected from all patients at baseline. After the blood draw, the blood is picked up by a courier so the blood could send to the CytoTrack Laboratory in Denmark on time. We managed to send the CellSave tubes to Denmark and process them within 96 hours for all but one patient. When a patient had HER2-positive CTCs, another blood draw was performed to collect the 60 mL of EDTA blood. This blood was processed the same way as the EDTA blood in the CareMore study 1. Also for this study, there were no major issues in processing the samples and all samples were stored at the moment for further processing in the future.
4.4 Clinical data collection
For both studies, clinical data of all patients is up to date. At the moment, we are not able to provide information about the progression-free survival at 6 months (PFS6months) from all patients, because multiple patients are still being treated. Because of the technical difficulties, as described in the previous paragraphs, we are not able to correlate the PFS6months with the characteristics of the CTCs that have been isolated.
As a consequence of a delay of the development of assay and preservation methods leading to the required changes as outlined above, and due to slower patient inclusion than predicted, it was decided to amend the project´s end point regarding validation on clinical material requiring the six-month patient follow-up which could not be reached for a majority of patients within the time line of the project. Therefore, and also for ethical reasons, the consortium decided not to interfere with potential outcome of the two ongoing patient studies by performing the CareMore assay (CTC identification and subsequent quantifying 4 proteins) on the limited number of patients collected within the scope of the project. Instead we changed the end points regarding the CareMore protein assay from staining all clinical samples to a proof-of-concept validation on well-defined reference cell lines spiked into healthy blood to show feasibility of the developed assay to detect their target molecules towards they were developed. Recruitment of patients and subsequent sample collection started in early 2015 is still ongoing and will continue beyond the end date of the project. All samples are stored according to the developed storage conditions which maintain analyte stability.
In summary, the clinical samples will be preserved for future use and will not be investigated using the developed CareMore assay, consequently neither was the six-month follow-up objective investigated within the scope of this project. All reagents are stored and have been passed on to EMC so that the assays can be performed in batch on the samples when the target number of patient samples are collected, which at today’s rate of recruitment would be approximately in October 2017.
The final statistical analysis of patient outcome in relation to the investigated markers will be available only 6-months after the last patient has been included in the study, this is also true for the 2 clinical studies. Due to unforeseen circumstances including delays in assay development, slower than expected patient accrual combined with the fact that fewer patients than expected showed CTCs positive for HER2, final analysis will be outside the project´s 36 months’ time frame. We don’t know why fewer patients than expected have HER2 positive CTCs, however recent experience indicates HER2-postivity of CTCs in a similar patient population is indeed lower than we anticipated at the start of the project. Of course the slower recruitment means that the sample collection is taking longer to complete. All patient data acquired within the time frame of the CareMore project are presented in a conscious way, well-aware that it represents only a part of the entire cohort, awaiting a significant number of cases to be included later than planned. A fraction of the clinical samples was also investigated for PIK3CA mutations. This was done with cells already identified as CTCs after going through the Cytotrack platform including staining of HER2, EpCAM and cytokeratin. This showed the feasibility of mutation calling in clinical specimens included in the CareMore studies.
Potential Impact:
CytoTrack
With regard to Sales & Marketing, CytoTrack will continue to use the results - CytoTrack instrument and CTC characterization methods – to support sales and marketing in the form of Press Releases, Product Specifications, Web Site, Scientific Publications and Poster Presentations. Examples of such dissemination activities initiated during the project period are available online. Furthermore, the CytoTrack instrument and accompanying methods are made commercially available and thus allow researchers and clinicians around the world to perform their own CTC studies.
In addition a CTC Centre of Excellence was initiated which means the CytoTrack instrument and CTC characterization capabilities are already available as a service from the Department of Clinical Biochemistry, North Zealand Hospital, University of Copenhagen Hillerød, Denmark, where the CytoTrack instruments and methods are implemented, servicing the clinical and scientific community as a reference laboratory where CTC analysis may be performed on blood samples shipped from clinical studies around the EU. For researcher and clinicians and other centres are being explored in Poland and Germany. This way researchers and clinicians may immediately benefit from capabilities offered by the CytoTrack instrument and methods. The CTC Centre of Excellence is also used to process the blood samples from the clinical validation study.
With regard to grant application, the results have and will continue to be included in company presentations and grant applications.
With regard to Intellectual Property, the patent protecting the CytoPicker will be applied for in the EU and optionally the US for future protection of the invention made during the project.
So far, CytoTrack instruments and method have been developed to form basis for the generation of the socio-economic impact and the wider societal implications and the CareMore clinical study on breast cancer has been initiated to provide the first clinical evidence. In parallel, other clinical studies are being outlined for other cancer types and/or with other cancer centres in EU.
In conclusion, the CareMore project has successfully enabled CytoTrack to develop and commercialize instruments and methods for characterization of CTCs to a technology readiness level where they can be validated for clinical use. As such, CytoTrack has transformed from a technology company to a provider of clinical research products and is ready for exploring means for entering the lucrative clinical diagnostic market. Impacted by the encouraging results from the CareMore project, a Horizon phase 1 grant was awarded to CytoTrack in 2015 for preparation of a Business Plan for CytoTrack to bridge the gap - extensive clinical studies and regulatory approvals – between the research and clinical market. Now, the CareMore project extended with a Business Plan (Horizon 2020) together provide foundation for preparation of an application for a Horizon phase 2 grant with focus on a multicentre clinical study in Europe providing the clinical evidence for a better and more cost-efficient personalized approach to cancer therapy and for CytoTrack to obtain regulatory approvals; the ‘license’ to enter the clinical diagnostic cancer market.
Philips
At Philips Research we have equipped our 3DHistech Pannoramic Midi scanner with 7 additional fluorescent filter sets and software was developed to reduce the inevitable crosstalk between the 7 resulting fluorescent channels due to spectral overlap. An algorithm was developed to identify cell nuclei in the DAPI channel and thereby define a nuclear compartment. Subsequently a virtual membrane was algorithmically generated around the nuclei to allow for the definition of a cytosolic compartment. This allowed us to quantify all the fluorescent channels other than DAPI in the nucleus and cytosol compartments for both immunofluorescent (homogeneous, smooth) and PLA (discrete, dot-like) signals. We developed code to be able to visualize the individual cells sorted based on any of the quantified fluorescent signals or morphological parameters such as nucleus size. Moreover, we were able to rank and present cells based on multiple parameters facilitating visual inspection of the most interesting cells first. Based on the cytokeratin (CK) status spiked tumour cell lines (and in the future possibly real CTCs) could readily be identified amongst a background of CK-negative white blood cells. Statistical methods were applied to investigate the power of the individual CareMore assays to separate positive from negative cells for 7 biomarkers in total.
We will continue to use this high multiplexing capability in cell line experiments and perhaps in the future for the characterization of CTCs as well. The algorithm created to identify cells and quantify fluorescent signals of different nature (homogenous and dotted) in high multiplex was successfully adapted for use with formalin-fixed, paraffin-embedded tissue slides and enables us to investigate any type of cancer tissue. We performed multiplexed mRNA FISH experiments and were able to use the new filters in combination with the cell and dot detection algorithms to quantify the mRNA generated by the activity of oncogenic pathways at the cellular level. Some of this work is going to be published. The software for exporting and analyzing (annotated) scanned images from the 3D Histech scanner are used frequently and will be used for the foreseeable future in several research projects. The tool set enables us to carry out translational research collaborations with clinical sites to test biomarker candidates for prognostic and therapy prediction assays.
We also collaborated with EMC to develop a device that allowed for the centrifugation of a Hettich Cyto chamber (a device often used in cytology) while keeping it inside a magnetic field. The field was generated by a permanent magnet built into a PMMA holder. The holder fits a Cyto chamber on top of a standard microscope slide and was designed specifically for this purpose but it could also be used without it. For the latter case we designed small cups that could be attached to a microscope slide using a non-permanent adhesive tape. These cups can be filled with immunomagnetically labelled cells, centrifuged with or without the magnet and they facilitate subsequent fixation, staining and/or lysing of cell samples. The method combines a magnetic field gradient with centrifugal force in order to optimize the recovery of immunomagnetically labelled cells by pulling them towards a microscope slide with a larger force than any of the two methods separately. The magnetic device can be taken out of the centrifuge together with the microscope slide containing the cells. This means that the cells never leave the magnetic field gradient and can be further processed while being kept firmly in place by the magnet.
This magnetic device was also used at Philips Research Eindhoven in another CTC project focussed on prostate cancer and aided in obtaining high CTC recoveries. In the future we may well use the device again for cytology related applications.
EMC
At Erasmus MC, a digital PCR assay was developed for sensitive detection of the 4 most frequent PIK3CA mutations on single CTCs purified using the Cytopicker from Cytotrack. The assay is being exploited for other scientific question as well and we have extended our repertoire of digital PCR assays for other relevant markers (e.g. ESR1), beyond the current scope of the project.
The clinical study will continue beyond the current project potentially providing impact using the assays developed within the CareMore project. The technological expertise (e.g. single cell PCR) gained within the project will be used in other grant applications.
Olink
The four oligo systems used in the four PLA assays in CareMore protein assay was developed initially in the DIATOOLS project (FP7-259796), except Duolink that was already a commercial product from Olink. During CareMore, the three additional oligo systems were optimised to function together in a higher specificity mode, and examined and cleared for cross-reporting. In summary, the stringency is higher now due to protocol optimisation performed by changing to a more specific ligation enzyme, increasing the amplification enzyme and increasing the temperature during parts of the protocol.
In addition, we now have eight antibodies and four protein assays that have gone through thorough testing. These protein assay, and especially the early developed ones pHER2 and ER, have been performing well particularly in the test systems with shorter transportation time than 24h and are now available for further use together or by themselves. Moreover, since the staining has not yet been performed on the clinical samples, there is a possibility to find even better assays but to use the reporting system optimised within the project.
Future projects focusses on developing an enzyme free detection system and therefore well suited for automatic staining instruments. The protein assays that the CareMore project developed, and thus all 8 antibodies involved, can now be used and detected with the future method, to be benchmarked to the Duolink system as a golden standard.
Stockholm University
The work on establishing in situ mutation detection in intact CTC:s did not result in a useful assay within this project, but was picked up as a new approach in the FP7 CANDO project. In this project in situ KRAS mutation detection has been achieved in CTC:s captured from pancreatic cancer patients on in vivo cell collectors from Gilupi, In parallel, SU has collaborated with researchers in Graz to detect a splice variant of the AR transcript in CTC:s captured from prostate cancer patients on in vivo collectors. A patent has been filed on in situ mutation detection in CTC:s captured on in vivo collectors.
List of Websites:
CareMore project home page: http://www.caremorectc.eu/
CytoTrack contact: service@cytotrack.com for further information about CTC testing service.
Philips contact: reinhold.wimberger-friedl@philips.com
Olink contact: http://www.olink.com/olink-bioscience/ Ann-Catrin Andersson, ann-catrin.andersson@olink.com (ann-catrin.andersson@comhem.se) Peter Åsberg peter.asberg@olinkbioscience.com Erik Ullerås erik.ullerås@olinkbioscience.com
EMC Contact: John Martens j.martens@erasmusmc.nl Jaco Kraan j.kraan@erasmusmc.nl Inge de Kruijff i.dekruijff@erasmusmc.nl
Stockholm University Contact: Mats Nilsson mats.nilsson@scilifelab.se