Final Report Summary - DIATOOLS (Tools for minimally invasive diagnostics)
Executive Summary:
The EU FP7 DiaTools project took as its aim to develop means for advanced, multimodular molecular analyses of patient samples obtainable by minimally invasive means. The successful development and implementation of such techniques can have profound impact on clinical medicine, but this will depend on progress along several dimensions that have all been organizing principles for DiaTools, as briefly described below:
Advanced molecular tools. Molecular recognition reactions assume central roles in diagnostics, and the partners have established a broad repertoire of techniques for analysis of nucleic acids and proteins. In situ PLA was used to characterize fusion proteins as diagnostic assays in chronic myelogenous leukemia patients through convenient flow cytometric analyses at UU. The proximity extension assay was used by Olink to successfully measure sets of 92 proteins from dried blood samples collected by finger pricks, enabling wellness studies. UU has further expanded the repertoire of molecular tools to include patented nFold, ExCirc, CutLig and superRCA probes of value in a growing range of analytic contexts. SU established highly multiplexed RNA analyses in cells and tissues through in situ sequencing of barcoded padlock probes, and they developed digital PLA protein assays for precise quantitation.
Advanced devices for analysis of cells and detection of signals. EPHESIA devices were developed at Institut Curie as highly efficient and convenient means to isolate circulating tumors cells from blood, for enumeration and subsequent molecular analysis, and Fluigent contributed complementary techniques required for trapping cells in fluidized beds. A major breakthrough was achieved with pioneering work at Stanford on multiparametric measurements of cells by the CyTOF, with applications to e.g. define shifts among different blood cells in disease. An amplified single molecule detector developed at Q-linea provided digital readout for a range of molecular detection reactions, and improved paramagnetic micro- and nanoparticles were constructed and made available to DiaTools participants by the partner in Prague.
Clinical validation. An important aspect of the project was contributed by clinical partners in Paris and Uppsala advising on clinical diagnostic needs and requirements. The partners in Paris also established an effective PCR-based approach to detect rare mutant sequences among cell free DNA in patients with uveal melanoma for disease follow-up, a growing diagnostic need. Other disease-associated features that were in focus were the circulating tumor cells whose isolation and enumeration provide access to tumor cells through a simple blood draw. Among many molecular features that were of diagnostic interest in cells isolated from peripheral blood was the HER2 status that correlates with prognosis in breast cancer, and fusion proteins heralding disease recurrence in CML and AML.
Impact. The DiaTools has made considerable progress along the several dimensions we set out to pursue. Besides important scientific publictions, several different analytic modalities were established, some of them ready for clinical implementation while others represent important research tools to validate biomarkers that only in a longer perspective will reach the clinic. Among our highlights are the early work on the CyTOF, now being rapidly implemented for research and in the clinic, and a range of novel tools that will permit e.g. multiplex analyses of rare mutant DNA, RNA or cells through liquid biopsies. Highly precise, parallel protein measurements in plasma developed by the partners are also rapidly making inroads in the clinic.
Through the above progress, the DiaTools project has provided a basis for industrial growth and expansion of markets, with application mainly in research and by big pharma in initial phases, but gradually also suited for clinical implementation. The project has also successfully promulgating the partners’ tradition of fostering researchers and entrepreneurs through research training, and via DiaTools courses and symposia.
Project Context and Objectives:
The overarching aim of the DiaTools project has been to enable multimodal molecular analyses of patient samples that are accessible by minimally invasive means. This is a central goal in capitalizing on in progress understanding the molecular underpinnings of normal and pathological physiology. Along with the vastly increased molecular insights, greatly improved techniques are becoming available for characterizing molecular states of tissues and body fluids with important consequence for future healthcare. The partners of this the DiaTools project have made important contributions to this technology development and to applications of the techniques. With the DiaTools project we have undertaken the important task of developing and adapting high-performance assays and equipment in order to enabling analyses of patient samples that can be readily accessed without surgery, and that can be used to investigate outpatients or to continuously monitor the effects of an ongoing therapy. Such analyses of samples obtainable through noninvasive measures will be a crucial element of a future healthcare with detailed, repeated molecular diagnostic investigations and personalized therapy.
The DiaTools project has developed tools to isolate rare cells from blood, for example circulating tumor cells, and to conveniently analyze their nucleic acids and proteins for precise diagnosis and therapy monitoring. We have also analyzed established techniques to analyze cells and individual macromolecules in high-throughput flow systems. Here we have used both standard flow cytometers, and one of the partners has also pioneered the use of a CyTOF instrument capable of 50-parameter molecular measurements of individual cells in very high throughput. Other partners have also developed a system to digitally count in flow individual detected macromolecules, represented by individual amplification products that can be enumerated in a microfluidic system with excellent specificity of detection.
Yet another aspect of the project has been geared towards comprehensive molecular analyses of body fluids, building in part on the highly resolving proximity ligation technique for protein detection developed by the partners. Our clinical partners supplied both relevant research challenges, medical and logistical insights, and suitable samples, and they have been pushing the sensitivity for detecting rare mutant DNA sequences to very low levels, an application that is assuming increasing importance for monitoring cancer therapy.
Project Results:
The DiaTools project has focused on development and application of methods, assays and devices for molecular analysis of patient samples obtained by minimally invasive means.
The small amount of sample available in this context puts high demands on the analysis technology in terms of specificity and sensitivity. The DiaTools consortium has addressed this challenge by working in parallel to optimize molecular analysis technologies as well as detection devices. The resulting assays and devices have been integrated and to some extent tested in a clinical setting, mainly in cancer diagnostics.
ADVANCED MOLECULAR TOOLS FOR PROTEIN DETECTION
The specificity and sensitivity of analysis of protein biomarkers has been pioneered by scientists at Uppsala university (UU). The central theme has been to utilize pairs of affinity reagents (antibodies) to increase specificity and sensitivity.
Proximity ligation and extension reactions for protein detection.
Antibodies with attached DNA strands provide new, valuable opportunities to engineer high-performance protein assays. In these assays, protein recognition via sets of antibodies with attached oligonucleotides is combined with molecular genetic reactions to copy, ligate, amplify and distinguishing reporter DNA sequences for efficient readout in solution or in situ, individually or in high multiplex. Proximity ligation and extension reactions, developed at UU and at Olink are excellently suited for minimally invasive protein diagnostics, and they have been a central aspect of the DiaTools project.
Proximity extension assays for biomarker discovery and validation.
Proximity extension assays (PEA), developed at and commercialized by Olink employ pairs of oligonucleotide-modified antibodies directed against the same protein for detection in solution phase. An initial incubation is followed by a reagent addition step that serves to initiate a DNA extension reaction between oligonucleotides on proximate antibody pairs. The addition also has the effect of reducing background signals. The resulting DNA strands are then amplified in realtime PCR reactions to identify and quantify detected proteins. The DNA extension and amplification reactions are designed only to detect products from cognate pairs of antibodies. This has the important consequence of allowing extensive multiplexing without any increased risks of formation of crossreactive products between all the reagent pairs combined in a single reaction. In this manner, each 1 µl sample of plasma or cell lysate can be interrogated using antibody pairs directed against 92 proteins, with another four assays included as internal controls.
Diagnostic applications of PEA.
The lab at UU has collaborated with Olink to establish a set of assays for more than 300 proteins, with the aim to monitor plasma from venous blood samples for levels of these proteins as possible indicator of a wide variety of disease processes or tissue trauma. In one study, a set of 250 proteins was measured in microdialysis samples from the brains of five neurotrauma patients. The samples were collected at 3 hr intervals during up to 6 days of care at the neurotrauma unit at Uppsala University Hospital. The assays revealed striking variation of levels of several or the investigated proteins over time as well as between different patients; in decreasing trends over time for some markers and with gradual increases among the up to 40 different samples for other proteins. These assays may provide valuable opportunities to assess the state of health of the investigated patients, and to estimate the individual prognoses and select optimal care.
Fine needle biopsies and single cells.
The PEA assays are also useful for measuring protein levels in lysates of tissues, and fine needle biopsies represent a useful means of accessing tumors via a minimally invasive procedure. Lysates from tissue left in the cannula after samples have been ejected for cytology and provide more than sufficient material for analyses via multiplex PEA. In fact, the sensitivity of PEA is often sufficient to measure protein levels in single cells, and this provides a valuable means to investigate intratumor heterogeneity and characterize the different clones of cells present. UU is currently conducting a study where individual cells from short term cultures of gliomas have been isolated and investigated with respect to 88 proteins and 88 transcript each. The results reveal differences between cells in susceptibility to treatment with bone morphogenetic factor 4, providing a much better resolved view of the therapeutic potential of the agent in different patients.
Ultrasensitive protein detection.
There are now extensive studies underway in Uppsala, also outside the DiaTools project, where the PEA technique is used to screen panels of protein biomarkers for differences in levels between patients and controls, with so far 20 000 samples analyzed and 1.5 million proteins measured. These studies regularly identify promising markers, but in order to fully take advantage of the potential for early detection of disease using such markers it is important to further increase sensitivity of detection of markers present at very low levels. To this end, UU has established a variant of proximity ligation referred to PLARCA, where target molecules are captured from 50 µl sample volumes, before being interrogated using pairs of oligonucleotide-modified antibodies that jointly give rise to a circular DNA strand, for rolling circle amplification. This mechanism is identical to that used in in situ PLA, and the assay provides added specificity and sensitivity due to the requirement for three interactions, and an ability to measure larger sample volumes than in PEA, amongst other features. While PEA typically offers sensitivity similar to standard sandwich ELISA reactions, albeit in multiplex and very small sample volumes, PLARCA routinely results in 1 000-fold greater sensitivity than ELISA. This improved sensitivity may translate to earlier detection of disease-specific markers during the course of disease.
New molecular tools for enhanced diagnostics.
The group at UU has developed a series of molecular tools that have become widely used in research, licensed by several biotech companies, and formed the basis of four spin-out companies, including the DiaTools partner Olink. A central aim of the DiaTools project has been to now also enable diagnostic applications of the previously developed padlock and proximity probes. Meanwhile, the DiaTools project has also provided a valuable opportunity to develop new classes of molecular tools for enhanced molecular analysis in both research and diagnostics. A total of seven patent applications have been filed, although most of the procedures described therein have not yet been published in the academic literature. Very briefly, we developed procedures for multiplexed in situ PLA for parallel visualization of proteins and their interactions in cells and tissues in pathology examinations. Unfolding probes are designed to detect either nucleic acids or proteins with any desired levels of specificity, singly or in high multiplex, producing circular reporter DNA strands for convenient detection. ExCirc and CutLig probes serve to detect RNA and DNA sequences, respectively, capturing part of the target sequences in circular DNA strands. Super rolling circle amplification (sRCA) and padlock sRCA are variant techniques that permits even single detected molecules to be visualized as large clusters of DNA strands, easily detected and quantified with excellent precision using e.g. flow cytometry. The technology also provides exceptional specificity in analyses of rare mutant DNA sequences in a vast excess of normal sequences. RCA reporters, finally, are probes that promptly initiate an RCA upon encountering their target sequences for rapid, convenient detection.
While all these probes and techniques are expected to find applications on their own merits, there is a particular value in combining several of these for particular diagnostic applications. Examples of applications include investigations of circulating cell free DNA in plasma for mutations that may reflect the presence of a tumor in a patient, highly specific detection and quantitation of circulating tumor cells in blood, or rapid high-performance assays at the point of care. These are all important objects of the DiaTools project, and the new molecular tools may provide new and further improved means to develop such diagnostic assays, beyond the conclusion of the DiaTools project.
Proximity Ligation Assay (PLA) for detection of BCR/ABL fusion protein in CML patients.
A PLA method for detection of fusion protein was one of the central aims in the plans for the DiaTools project. Fusion proteins are a hallmark of some types of blood cancer including chronic myeloid leukemia (CML). Detection of fusion proteins are central in diagnostics of CML and have so far mainly been performed through detection of the fusion RNA using polymerase chain reaction. We have now successfully developed an assay to directly detect BCR-ABL fusion protein, present in malignant cells in most individuals with chronic myeloid leukemia (CML) The PLA method efficiently detects the fusion protein both in cell lines and CML patients.
The method is now being evaluated in clinical material from CML patients and the results are compared with routine PCR analysis. So far 10 patients with low tumor burden have been tested, as well as 3 patients with new diagnosis, and high tumor burden. The results comparing the fusion protein detection in flow cytometry and PCR detection of fusion transcripts are similar, and in some samples the fusion protein is detected when the transcript is undetectable in the patients. Direct detection of fusion proteins in the CML patient cells will hopefully speed up and improve the diagnosis of this disease. In addition, PLA for multiplex testing of fusion proteins in leukemia has been developed and evaluated. Multiplex detection of different fusion proteins at the same time will further improve diagnosis and in other forms of blood cancer including Acute Myeloid Leukemia and Acute lymphoblastic leukemia. In the DiaTools project, antibodies for fusion proteins involved in AML and ALL have been tested in 4 different cell lines, each carrying a different fusion protein. The assays have been validated using in situ PLA with microscope readout or flow cytometry readout.
In addition to development of diagnostic assays for blood cancers, researches at UU have investigated the pathways and mechanisms underlying the disease. Using methods developed in the DiaTools project, mutations affecting the signaling pathways in cancer cells have been identified and the effects have been examined in patient samples. Since the NF-kB signaling pathway is constitutively activated in chronic lymphocytic leukemia (CLL) cells, we first utilized the selector-based targeted enrichment technology, developed in-house, and performed next-generation sequencing of 18 NF-kB core complex genes in a discovery and validation CLL cohort (in total 315 patients). Our prime finding was a recurrent mutation within IkBe (encoded by NFKBIE), a negative regulator of NF-kB in B-cells, that was found recurrently mutated in 21/315 cases; 13 carried an identical 4bp frame-shift deletion. By screening an additional 416 CLL cases, the NFKBIE deletion was predominantly seen in poor-prognostic patients, and minor subclones and/or clonal evolution were also observed, thus linking this recurrent event to disease progression. This 4 frame-shift deletion leads to introduction of a stop codon and subsequent loss of the ankyrin repeat region, resulting in a truncated form lacking this domain, which is a prerequisite for protein-protein interaction, i.e. binding to p65 or other transcription factors. To further explore if the physical interaction between IκBε and p65 is influenced, we applied proximity ligation assays, as a novel means to study their interaction, revealing that this truncating mutation resulted in reduced IkBe/p65 interaction, at the single cell level, compared to wild-type patients. Finally, we analyzed the presence of the NFKBIE-deletion in different lymphoma entities (in total 372 lymphoma samples) and this genetic event was also identified as recurrent, pointing to a common mechanism of NF-kB deregulation during lymphomagenesis.
While some research groups have concentrated their efforts on development of novel molecular detection technologies, other DiaTools partners have developed new methods to read-out and record the detected molecules.
Digital protein analysis for non-invasive diagnostics.
Researchers at Stockholm University (SU) have developed a truly digital readout for PLA reactions that can be readout by Qlinea’s Amplified Single Molecule Detector (ASMD) reader. We converted two PCR-based solid phase PLA assays to Rolling Circle Amplification (RCA) based reactions. The PLA assays targeted IL-6 and VEGF and the PLA ligation product was digested and circularized according to the circle-to-circle amplification (C2CA) scheme. The formed circles were amplified by RCA and digitally recorded by ASMD. The digital quantification of PLA products resulted in an extrapolated limit of detection of approximately 5 fM (0.1 pg/ml) and average CVs of 7% across a dynamic range of 5 logs. PCR readout performed in parallel resulted in 1.5 fold less sensitive limit of detection of approximately 9 fM, and the average CVs for the PCR readout was 18% over a comparable dynamic range. ASMD thus provided at least two fold improved precision compared to the PCR-based readout. A similar quantitative precision was observed in detection of the cytokine vascular epithelial growth factor (VEGF) in buffer.
ASSAYS FOR MULTIPLEX PROTEIN DETECTION
PEA development and optimization
Better capabilities to diagnose disease at the early and most curable stages will greatly improve human health and reduce health care costs. Patient stratification is also in need of better diagnostics to facilitate the selection of appropriate patient care. Preliminary data in the literature claim improved diagnostics with the use of multiple complementary biomarkers, as no marker can today single-handedly diagnose all cases with desired accuracy for a certain cancer type. Therefore, a specific and sensitive immunoassay at high multiplex level would allow for more rapid identification of new biomarkers and combinations of biomarkers.
One unique feature of the proximity extension assay (PEA) described above is the extremely low sample consumption, which makes it especially suitable for biomarker studies of scarce samples, such as small needle biopsies and lysates of single-cells to mention a few. PEA possesses high sensitivity, high specificity, great multiplex capability, and potential for rapid assay development time and high throughput screening.
In the DiaTools project, the biotech company Olink has developed and optimized the PEA technology. One of Olink’s current products on the market, Proseek multiplex, allows for detection and quantification of 92 proteins in only 1 µL sample. However, the initial version of the product did not have sufficient inter-assay precision to generate quantitative data across a large set of samples at sufficient precision, but the DiaTools project has enabled us to develop a new inter-plate control and a new normalization procedure that decreased the inter-assay and batch-to-batch variation.
Furthermore, in order to generate PEA quantification at the single-cell level, high intra-assay precision is critical. During the course of the DiaTools project we identified a position-dependent variation in signal that seemed to occur during the qPCR detection step. This led to worsened precision and could, in worst case, result in misinterpretation of the data. During the project we managed to optimize the detection protocol, which minimized the position effect, and thereby improve the intra-assay precision.
At the conclusion of this collaborative project, the research and innovation conducted within the consortium has allowed Olink to further strengthen our competitiveness in the international arena of high-throughput biomarker research.
PEA single cell development
As mentioned above, Uppsala University has in collaboration with Olink generated a protocol for protein quantification in single-cells using PEA. In brief, single cells are sorted by flow cytometry into a 96-well plate containing lysate buffer, after which 1 µL is transferred to a new plate for standard PEA analysis. The protocol is very easy to apply to the standard Proseek Multiplex assay and demonstrating sufficient sensitivity and precision to enable multiplex protein quantification at the single-cell level.
Optimization of PLA and Padlock probes (PLP) for flow cytometry
A protocol was developed combining the unique aspects of the PLA and PLP assays into one, without significantly decreasing the sensitivity or specificity of the individual assays. The assay was developed to allow higher multiplexing degrees. The combined assay performs well when done on slide or in solution, allowing for quantification (counting RCA products) using a microscope, or doing cell population profiles using flow cytometry.
The PLA/PLP multiplex protocol was developed to be performed within regular working hours and to not deviate too far from the standard PLA and PLP protocols, in context of number of steps and time intervals. The resulting protocol is actually shorter (7-8 hours) and easier to perform than doing the PLA and PLP assays separately (either simultaneously or sequentially).
ADVANCED MOLECULAR TOOLS FOR DETECTION OF DNA AND RNA
Digital DNA analysis for non-invasive diagnostics
SU researchers have also developed an approach for non-invasive prenatal diagnosis (NIPD). Our approach is based on quantification of differentially methylated CpG sites on chromosomes 21 and 18 between maternal and fetal DNA, using padlock probes and RCA using a digital ASMD readout. We show that fetal DNA can be enriched from maternal DNA by a factor of 5, and that fetal DNA in fractions down to 30-40% can be detected using this pilot assay configuration targeting 7 CpG sites. The sensitivity of the approach can be improved by targeting many differentially methylated CpG sites.
Circulating tumor DNA
In addition to detecting circulating tumour cells, described below, we have developed methods to detect tumour cell DNA present in the blood. We have developed an automated C2CA device for diagnostic mutation detection in circulating tumor DNA. It is a multi-step, bead-based protocol that gains in automation to enable clinical use at a point-of-care. Efficient re-suspension and mixing of the beads is critical for assay performance. We have therefore integrated the molecular protocol into a microfluidic device which keeps the beads in dynamic suspension based on a fluidized bed approach that enabled us to automate and simplify the full assay. The fluidized bed system was developed at Institut Curie and was operated by a pressure control system from Fluigent. Within the DiaTools project it was optimized for DNA analysis and fabricated in a thermo-resistant polymer to function at elevated temperatures in our protocol.
A multiplex padlock and C2CA assay was successfully developed for genotyping of the seven most common mutations in the frequently mutated oncogene Kirsten rat sarcoma viral oncogene homolog (KRAS). The protocol was adapted to fit the microfluidic chip format in two stages. A model assay was first developed and integrated in the fluidized bed chip for a 1RCA protocol with target capture and RCA amplification on beads followed by release of the products and fluorescent ASM readout. The target capture efficiency was 93% with the magnetic bead plug. Next we developed a new solid-phase C2CA protocol which simplified integration of the whole protocol on chip. The target capture, upconcentration and the first amplification is performed on magnetic beads in the fluidized bed and the monomers from the digested RCA products are re-captured to probes on the surface of a second chamber where the second amplification and labeling takes place. Amplification products are visualized in a microscope or in the future with a dedicated array reader (developed by Qlinea). A fully integrated polymer chip was successfully designed and fabricated for the solid-phase protocol. The surface in the second chamber was functionalized through direct UV binding of the probes and the complete assay tested on chip.
Detection of somatic mutations in intact cells
In addition to detecting mutations in freely circulating DNA, we have also applied our in situ genotyping method for detection of somatic mutations in intact cells. We developed assays for oncogenic mutations in the KRAS and EGFR genes and verified the performance on a range of clinically relevant patient samples including sections of fresh-frozen tissues, sections of formalin fixed paraffin-embedded tissue and cytological tumor imprints. The results were 100% concordant with results from the routinely used diagnostic method applied on DNA extracted from the same tumors. We demonstrated the sensitivity of the method by detecting rare mutated cells spiked into a background of wild-type cells. By targeting the tumor suppressor gene TP53, we also successfully detected loss-of heterozygosity of this gene. The method should be directly applicable for analysis of intact captured circulating tumour cells (CTCs).
Sequencing of RNA in intact cells
We have developed a technology that, for the first time, allows sequencing of RNA directly in fixed cells and tissues. The technology uses padlock probes and RCA to generate targeted sequencing libraries in situ. The RCA products are sequenced using sequencing-by-ligation chemistry allowing reads of 6 bases to be recorded in situ. This read-length is sufficient for scoring mutations in oncogenes, and in one experiment we detected 5 KRAS mutated cells in the background of 5 000 KRAS WT cells. The read length is also suitable for reading barcodes in the probe library. With 6 bases read-length it is possible to discriminate 4 096 different sequence barcoded probes. We demonstrated this strategy by performing targeted expression profiling of 39 transcripts in intact sections of breast cancer tumor tissue with cellular resolution. The technology should be directly applicable for analysis of intact captured CTCs.
MICROFLUIDIC DEVICE FOR CAPTURE OF CIRCULATING TUMOUR CELLS
In the framework of the DiaTools project, researchers at the Institut Curie in Paris have developed and validated a microfluidic platform dedicated to circulating tumor cells (CTC) capture from patient blood samples. It is a huge challenge as we are looking a few CTCs in billions of “interfering cells” (white blood cells, red blood cells…). The microfluidic platform we developed is called EPHESIA and it relies on immunosorting (related to the presence of specific antigen on CTC membrane).
The system we developed combines an unprecedented level of automation, integration and cost-effectiveness. It essentially involves the self-assembly, in a microfluidic system, of a regular array of columns made of magnetic microparticles coated with antibodies or other ligands directed against the CTCs. It captures cells by surface antigens. It does not require any complicated microfabrication, since the physics of self-assembly does the job of making very regular, high aspect ratio capture columns. Very importantly, all the delicate biofunctionalization steps can be performed in large batches out of the system, reducing cost and increasing reproducibility.
Beyond this, however, the EPHESIA technology completely reverses the paradigm of magnetic sorting, bringing in a dramatic increase in sorting efficiency and sensitivity. In our method, the magnetic particles are first immobilized, and then the cells are flown through. This presents two major advantages: first, the number of particles necessary is proportional essentially to the number of cells to be captured, and not to the sample volume (as in batch methods). This is interesting not only for cost, but also because the use of a large excess of particles leads to cell damage, and renders accurate morphological observation impossible. Second, in our system the interaction between the cells and the immobilized particles is not induced by Brownian motion, but by the well controlled hydrodynamic forces inside the array. Additionally, our system allows the application of all the staining and observation protocols currently used in optical cytology.
During the project, we have first demonstrated the ability of this platform to capture cells presenting different specific antigen expression in order to evaluate our platform capture efficiency in different cellular conditions. We have thus validated the EPHESIA platform with real patient samples in blind study. Experiments have been performed to compare our device with the current standard method for CTC analysis (Cellsearch, Veridex). Very high capture efficiency and purity were achieved with conventional CTC model cells lines (90%, 0,02% of purity). We have obtained similar or better CTC capture with our microfluidic platform both with breast and prostate cancer patient. The platform automation has been greatly improved, especially the fluidic control of the experiment as well as the cell staining and imaging performed after cell capture.
Even if being able to capture a few CTC from blood is already an important information for clinicians concerning patients prognosis, it turns out that getting more information on the proteomic or genetic status of CTCs could help clinicians in treatment orientation. The first CTC characterization we performed on chip is related to immuno-staining to confirm their CTC status. Besides, we have also integrated a molecular analysis method (in situ PLA) developed by DiaTools partners to investigate protein-protein interaction on the cancer cell membrane. Indeed, it is now well establish that protein dimerization occurs on cell membranes and they could be the starting point of signaling pathway that will modify the cell fate. Here we have been interested in HER2/HER3 dimerization. After an important optimization process, we have successfully demonstrated the possibility to integrate Proximity Ligation Assay (PLA) on captured CTCs to evaluate their level of HER2/HER3 dimerization.
Finally, working with columns made of magnetic particles offers the possibility to release the magnetic beads by switching of the magnetic field. We have taken benefit of this to retrieve cells after their capture for subsequent CTC DNA analysis. Thanks to the high purity achieved with our EPHESIA platform, we have been able to detect PI3CKA mutation in less than three cells captured in our platform.
During the project, the microfluidic platform has been greatly improved. Especially, chip material has been investigated to get materials compatible both with mass production, high temperatures and good optical quality. We have thus used thermoplastic materials and we have put efforts both on microfabrication process and on surface treatment strategies. Indeed, we have been working with COC as material since COC is hydrophobic and inert. We have also developed new coating strategies based on physisorption to avoid non specific adsorption of beads and antibodies on the microfluidic chip surface.
DEVELOPMENT OF SINGLE CELL ANALYSIS BY MASS CYTOMETRY
The Nolan Lab at Stanford University is focused on technologies that analyze biological events at the single cell versus an average of such events within a pooled population of many cells. Such technologies can therefore study rare cells, such as stem cells or drug-resistant cells, whose biological attributes would otherwise be overshadowed by more abundant cell types within a mixed cell population. Single cells can be considered as the basic units of organs and tissues akin to individual bricks that comprise the structure of a building.
The Nolan laboratory studies phospho-immune cell and cancer signaling and other metabolic parameters by analysis of biochemical functions at the single cell in primary cell populations. From around 2009, the Nolan Lab introduced and developed mass cytometry, a technology platform that allows one to survey and quantify 40-plus cellular markers on a cell-by-cell basis. Mass cytometry is a novel adaptation of atomic mass spectrometry that enables high-dimensional flow cytometry applications. It employs antibodies conjugated to stable isotopes of the transition elements as reporters, replacing fluorophores previously used in traditional flow cytometry. Overlap of fluorescence emission signals precludes the use of fluorophores to measure more than 15 or so parameters per single cell. In mass cytometry, given that each stable isotope has a distinct mass (no overlap of “mass”) there is a potential to exceed upwards of 100 parameters per cell. The mass cytometer has been developed for high-accuracy elemental analysis. The abundances of the metal reporters bound to cells via antibodies are determined and consequently the levels of a protein to which each antibody binds can be determined. Antibody panels are designed to recognize surface proteins, to delineate a cell type (e.g immune cell subset, epithelial cell, neuron) and their intracellular signaling pathways (e.g pathways involved in proliferation, apoptosis, metabolism, protein translation, transcription).
As with any quantitative technology, there is a stringent requirement for internal and external reagent standards for data normalization. During the DiaTools project, several have been developed in the Nolan Lab and have been made available to the scientific community:
i) Variation in instrument performance can be caused by factors such as instrument calibration, fluctuations in plasma and build-up of cellular debris in the sample introduction components. In order to normalize for these factors, polystyrene beads infused with precise amounts of several lanthanide isotopes are acquired simultaneously with every sample.
ii) In any biological sample dead cells often give false positive results, as they tend to bind non-specifically to many reagents. Therefore, removing dead cells from mass cytometry data is a critical step to ensure accurate results and analysis. A viability dye is used to identify dead cells and operates on the principle that a compromised cell with a damaged plasma membrane permits reagent entry into the cytoplasm, whereas a healthy cell with an intact plasma membrane does not. Recently, a protocol was described for using cisplatin to determine cell viability. It enters dead cells and the platinum can be detected by the mass cytometer and the signal used to quantify them.
iii) Mass-tag cell barcoding labels individual cell samples with a unique combination of six palladium barcodes. Samples can then be pooled for processing and measurement as a single multiplexed sample. The barcoding procedure eliminates variability between samples in antibody staining and instrument sensitivity. We developed a debarcoding algorithm to identify each sample.
New Assay formats
Proximity ligation assay, developed by Professor Ulf Landegren’s group at the Uppsala University is a technology that detects, quantifies and determines cellular localization with high specificity and sensitivity. Protein targets can be readily detected and localized with single molecule resolution and objectively quantified in fixed cells and tissue samples. Over the course of the DiaTools project, we have modified the assay for both fluorescence and mass cytometry to detect: a) messenger RNA in single cells – up to twenty different RNAs simultaneously. b) messenger RNA and protein simultaneously in single cells- up to twenty RNA and proteins simultaneously c) nuclear localization of proteins – the importance of this adaptation lies in the ability to see nuclear localization by a non-imaging technology. Professor Landegren’s group provided the foundational work upon which our adaptations are based.
Mass Cytometry data – lineage analysis
B-cell development as described by trajectory detection at the single cell level. Tissue regeneration is an orchestrated progression of cells from an immature stage to a mature one. A continuum of transitional states exists between these discrete stages. In a recent publication, multi-parametric single cell data was analyzed using an algorithm designed to leverage this continuity. Applied to B-cell lymphopoiesis, the algorithm constructed trajectories spanning from hematopoietic stem cells to naïve B cells. Within these trajectories, there were nascent fractions of B cell progenitors along with regulatory signaling and functional events such as immunoglobulin rearrangement. The study provided a comprehensive analysis of “healthy” B cell lymphopoiesis providing a reference against which to evaluate aberrant B-cell maturation and function in diseases such as B-cell leukemias. Furthermore, the approach maybe applied to other cell lineages.
Computational Analyses
The Nolan Lab continues to develop computational tools with which to analyze high dimensional single cell data. At their core these algorithms are designed to evaluate i) clusters of cells based on their similarity to one another ii) or each individual cell and how it looks in relation to others in a sample and across samples in multidimensional space. The computational analysis tools developed during the project will now be made available to the scientific community. We have several publications and manuscripts in preparation around each type of computational approach.
MAGNETIC PARTICLES FOR USE IN DIAGNOSTIC ASSAYS
Many of the diagnostic assays developed in the DiaTools project require magnetic micro- and nanoparticles with defined surface characteristics and specific binding properties.
Within the DiaTools project, new innovative surface-modified magnetic microparticles and nanoparticles have been developed at the Institute of Macromolecular Chemistry (IMC) in Prague, which suitably complement materials available nowadays on the market. While size of the nanoparticles was in tens of nanometers, typical diameter of the microspheres was in a range of micrometers. Generally, size of the particles was controllable and uniform which ensures the same physical, chemical and biological properties in the applications. The particles contain a reactive functional group available for attachment of a target biomolecule, typically an antibody or nucleic acid. Industrial production of these new particles is a subject of further development.
Magnetic microparticles
(i) Magnetic bead cellulose was developed by modification of sol-gel transition of viscose in the presence of maghemite (γ-Fe2O3) nanoparticles. Though the particle size was rather large (15-40 µm in diameter) and the particle size distribution was broad, the beads made of natural material have very high content of water (i.e. porosity) which makes them ideal material for fast and low-cost high-capacity separations of biomolecules together with the superior flow properties in the chromatography columns. The number of hydroxyl groups in polysaccharide chains makes the beads highly hydrophilic and therefore biocompatible and well tolerated by biosystems. The magnetic bead cellulose was activated with tosyl groups and exploited at Uppsala University as a solid support for sensitive protein detection in femtomolar concentration range using proximity ligation assay.
(ii) By elaboration of multiple swelling polymerization, magnetic monodispersed poly(2-hydroxyethyl methacrylate) microspheres were developed which were then modified with streptavidin. Size of the microspheres was controlled by size of starting monosized seeds and by degree of swelling with activation agents, monomers and porogens. High content of the magnetic compoud in the particles was achieved by repeated imbibing of Fe salts in the pores and precipitation with a base. The particles are easily separated by a magnet and exhibit low non-specific protein sorption if modified by poly(ethylene glycol). Thier performance and the binding efficiency were tested at Stockholm University by the circle-to-circle amplification.
Magnetic nanoparticles
Significant advances were achieved in fabrication of the advanced functional magnetic nanoparticles. Their size, particle size distribution, colloidal stability and magnetic properties were regulated depending on the reaction parameters during the synthesis and the selected production method, i.e. co-precipitation of Fe salts vs. thermal decomposition of Fe organic precursors. Size of the particles according to the transmission electron microscopy was controlled in the range of 10-30 nm.
Functionalization and applications
Surface coating of the particles is of key importance for their applications since it prevents particles from the aggregation, reduces undesirable nonspecific protein adsorption and enables immobilization of a target biomarker via reactive groups. Particles coated with several different compounds are now available at the end of the project including dopamine-hyaluronate conjugate, poly(N,N-dimethylacrylamide), poly(L-lysin), poly(ethylene glycol) derivatives containing small amounts of reactive functional groups (typically carboxyl, amino or tosyl), streptavidin, numerous silica derivatives of controlled thickness, etc. Newly, polyzwiterions were attached from the microsphere surface by reversible addition-fragmentation chain transfer polymerization. While the magnetic nanoparticles were able to successfully label various cells enabling their monitoring by magnetic resonance imaging, the magnetic microspheres were very useful for selective molecular diagnosis of various diseases including via rolling circle amplification. Antibody-bound magnetic particles are prone to easy manipulation and fast isolation of target biomolecules by a magnet and are amenable to automatic processes.
CLINICAL VALIDATION OF DIAGNOSTIC ASSAYS
A number of diagnostic assays have been clinically validated on patient samples at Curie Hospital in Paris. Assays for detection of nucleic acids in solution phase were validated for diagnosis of the eye cancer uveal melanoma. We firstly determined the prevalence of the Q209 mutations of GNAQ and GNA11 genes in 197 uveal melanoma tumors. We found that about 85 % of the tumors harbored only 3 nt mutations. This opened up the possibility to use the detection of these mutations to monitor circulating tumour DNA (ctDNA). For that purpose we set up a technique to detect point mutations in a complex DNA mixture. We chose to use the pyrophosphorolysis-activated polymerization method (bi-PAP) and we demonstrated the capacity of this method to detect tumor DNA (ctDNA) with a one-copy sensitivity. We applied this method to uveal melanoma xenografts by showing a very good correlation between LINE-1 PCR assay and bi-PAP PCR assays. We applied this method to 22 patients with metastatic uveal melanoma (MUM) and found that 21 patients harbored detectable ctDNA. Levels of ctDNA were correlated to the tumor burden.
Second, we compared the prognosis significance of CTC as measured by the Veridex technique (current gold standard for CTC detection) versus ctDNA in metastatic uveal melanoma. GNAQ/GNA11 mutations were characterized in archived tumor tissue. Using the above-mentioned bi-PAP technique, GNAQ626A>T, GNAQ626A>C and GNA11626A>T copy numbers were quantified in plasma from 12 ml of blood. CTC were detected at the same time in 7.5 ml of blood by the CellSearch® technique. Patient characteristics and outcome were prospectively collected. ≥1 CTC were detected in 12 of the 40 included patients (30%; range 1-20). Among the 26 patients with known detectable mutations, ctDNA was detected and quantified in 22 (84%; range 4-11421 copies/mL). CTC count and ctDNA levels were associated with the presence of military hepatic metastasis (p=0. 004 & 0.03 respectively), with metastasis volume (p=0.005 & 0.004) and with each other (p<0.0001). CTC count and ctDNA levels were both strongly associated with progression-free survival (p=0.003 & 0.001) and overall survival (p=0.0009 & <0.0001). In multivariate analyses, ctDNA appeared to be a better prognostic marker than CTC. ctDNA and CTC are correlated and both have poor prognostic significance. In conclusion, CTC detection can be performed in every patient but, in patients with detectable mutations, ctDNA was more sensitive than CTC and possibly has more prognostic value. This first clinical study has been published.
Correlation with clinical disease evolution and metastatic relapses
We are currently building up a very large cohort (target accrual: 800 patients diagnosed with localized non-metastatic uveal melanoma) whereby plasma samples from which circulating DNA is extracted are being collected, stored and analyzed for GNAQ/GNA11 mutations in batch. 500 patients have been included to date (December 2014) in the study, which has been ethically approved. Blood sampling occurs at diagnosis, before and after treatment, and at regular intervals or whenever the patient is monitored at Institut Curie. From this patient set, we expect approximately 25% will develop metastasis. We have already analyzed the ctDNA load in pre-metastatic patients for the first 66 patients without finding true positive samples.
Analysis of circulating tumor cells in breast cancer: HER2 status and inhibition
The assessment of HER2 status is a key parameter in breast cancer management at both early and metastatic stages. While our team reported that HER2 immunofluorescence on CTC is not reliable for HER2 status assessment of the tumor, our first aim was to implement HER2 Fluorescence in situ hybridization (FISH) on CTCs isolated by our in-house developed EPHESIA technique. Curie hospital provided clinical samples of HER2+ breast cancer that allowed the setting up of a new microfluidic chip in which high resolution imaging was possible. This low cost microfluidic chip integrating 3D micro-chambers allowed the capture and analysis of tumor cells, as well as the implementation of standard biological protocols in a chip format with low volume consumption. To validate the performances of our device, immuno-fluorescence labeling and FISH analysis on cancer cell lines and on patient pleural effusion samples were performed successfully.
More specific work on FISH implementation in microfluidic chips was then performed. Namely, we developed a protocol performed entirely in the liquid phase, allowing the immobilization and fixation of cells and their quantitative characterization by FISH. We demonstrated first in cell lines and in clinical cases the potential of this method to perform quantitative copy number measurement and clinical scoring of the amplification of the HER2 gene. This validation was performed in a blinded protocol in two clinical case studies (samples being provided by the Institut Curie hospital), in reference to the gold standard and clinically used method based on glass slides. We obtained a comparable reproducibility and a minor difference in apparent amplification, which can be corrected by internal calibration. The method thus reaches the standard of robustness needed for clinical use. The protocol can be fully automated, and its consumption of samples and DNA probes is reduced as compared to glass slide protocols by a factor of at least 10.
Building on these results, two main projects are now ongoing at the Institut Curie:
1/ The CirCe T-DM1 trial is a phase 2 trial (funded by a pharmaceutical company), which aims at treating patients with HER2-negative breast cancer and HER2-amplified circulating tumor cells with a pure anti-HER2 drug, T-DM1. This trial started in early 2014 in several cancer centers in France. Briefly, patients with HER2 negative metastatic breast cancers are screened for HER-amplified CTC by HER2 FISH on the captured CTC and, if positive, are eligible for anti-HER2 therapy. This trial has been ethically approved.
2/ The CirCe PLA observational study will start in 2015 and is funded by a pharmaceutical company. In this study, HER2-HER3 heterodimers on tumor cells (from both the tumor and the CTCs) will be quantified by PLA (proximity-ligation assay) before and during treatment. We aim at demonstrating that HER2-HER3 heterodimerization is inhibited by pertuzumab, a new anti-HER2 treatment, and to establish HER2-HER3 PLA as a predictive biomarker of efficacy of pertuzumab in HER2-positive metastatic breast cancer.
DEVELOPMENT OF DEVICES FOR ISOLATION AND DETECTION
In addition to molecular methods and assays for molecular diagnostics, advanced devices for isolation and detection have been developed within the DiaTools project. Within the framework of other activities, the company Q-linea has developed technology for detecting so-called Amplified Single Molecules (ASM). These consist of multiple repeated DNA strands with either the exact sequence, or the mirror (or homologous) sequence, of an identified target molecule, collapsed into spheroids large enough to be detected by standard microscopy technologies. The target molecule may be a unique DNA sequence identifying e.g. a specific tumor type, or a specific pathogen. Q-linea has developed a device for detecting and enumerating these entities. The sum procedure is termed ASMD (Amplified Single Molecule Detection), and is characterized by high sensitivity, high quantitativity, and high intrinsic multiplexability. In the DiaTools project the potential of this device to analyze larger objects such as cells for the presence of individual ASM associated with each cell was explored. Matching a specific ASM to a certain cell type may serve to identify specific cancer cells, isolated from blood circulation.
The standard Q-linea ASMD detector was evaluated with respect to performance in analyzing isolated cells. The performance was characterized, and a number of potential improvements were identified, the implementation of which was expected to improve performance.
In the latter part of this effort, a breadboard setup corresponding to a modified ASMD detector was built. This setup realized for most of the proposed improvements, and was at project end ready for evaluation. It should be noted, however, that as the DiaTools project has evolved, it has proven more fruitful to pursue an approach to detect rare circulating tumor cells that is based on microarray technology.
ASMD multiplex reader
Within DiaTools, different approaches to detect both protein and DNA signatures of e.g. tumor cell material have been developed and evaluated. Detection of a rare signature involves recognition, amplification, and readout of the target. With regard to protein and DNA target detection, the recognition step differs, but the amplification and readout steps can be essentially the same. Common to these approaches is the need of a dedicated, efficient amplification and readout device. As part of the project, Q-linea has developed two distinct systems to perform these tasks.
The initially developed system amplifies targets into distinct, countable ASM, which are read by a microscopy-based fluidics readout, passing the ASM in a liquid flow in front of a detector. Characteristics of this system are short time to answer, high sensitivity, quantifiable results, high throughput, but low to medium multiplexability.
As the project matured, it became clear that higher multiplex levels than achievable with the initially developed system was needed. This led to the development of a second detection system, based on microarray detection. This final system has delivered multiplex levels of 99+, with a potential of several thousand. The system also comprises sample preparation from blood samples, making a complete sample-to-result system. As this direction of work was set in the latter part of the project, an integrated demonstrator was not achieved, however the concept is demonstrated by a prototype of automatable modules. This concept in itself has been shown promising, and will be developed further within Q-linea.
Within the DiaTools project, the company Fluigent has developed a prototype device for isolation and concentration of proteins and nucleic acids from blood and from fine needle aspirates. The device is based on membrane filtration through a tubular membrane. An integrated module consisting of the required pumps, valves and other components, with the possibility to vary experimental conditions such as time or flow rate has been constructed and tested.
With this concentration module, a 10-fold concentration of proteins with 71 % recovery has been achieved within a duration of the global process of 5 minutes. Under the conditions used, the membrane has a life-time of > 150 hours.
This module can be adapted to molecules to be concentrated (modification of the pore of the membrane) and specifications of the process (duration, volume to concentrate, efficiency…).
Regarding the capture of rare cells from biological samples, Fluigent together with Institut Curie has developed a fully automated platform. Different tools and platforms have been successfully developed and integrated (electronic, mechanical and software integration): flow-rate platform (to monitor and control the flow-rate of liquid flowing in a microfluidic system) and valving platform (to inject sequentially different fluids in a microfluidic system).
Thanks to the different platforms and the development of a script editing tool, the whole sample processing can be automated.
Potential Impact:
The DiaTools project has already had considerable impact and we are confident that its effects will continue to be felt over coming years. Several of our projects are now ready to be tested in clinical trials to assess their relative merits over older methods currently in use. The following are some of the highlights of research applications that are close to clinical implementation.
Proximity ligation and proximity extension techniques have been positioned as world-leading methods for high-throughput, low sample consumption protein marker analysis in plasma, and very large-scale investigations are now in progress. So far some 23 000 plasma samples have been interrogated for a total of more than 2 million protein determinations at Uppsala University. On the basis of findings a panel of around 20 markers is currently being assembled for clinical implementation in cardiovascular disease.
The CyTOF technique, pioneered by one of the participants in the course of the DiaTools project is rapidly assuming a central role in a current drive towards single cell biology, and the technique has been successfully combined with the in situ PLA technique in collaboration between two partner labs, greatly extending the range of markers that are accessible to analysis. A molecular method for detection of fusion proteins in flow has been established and is ready for clinical studies to assess its practical utility, estimated to be significant.
The isolation of single tumor cells has been successfully accomplished using novel affinity matrices, and the cells have been interrogated using molecular tools provided by other partners.
Important technology has also been established and perfected for imaging mutation specific transcripts in individual cells, which can be retrieved by methods such as fine needle biopsy to assess the contribution of malignant cells in biopsy material, and analogous methods are pursued by other partners to allow assessment of cellular signaling via protein interactions and modifications.
The methods, assays and devices developed in DiaTools will ultimately result in improved diagnostics and treatment, primarily of different forms of cancer. In addition, these tools assays and devices will broaden the possibilities of the companies participating in the project and in the long run also strengthen the pharmaceutical industry through the availability of novel tool for evaluation of treatment response of novel drug candidates.
The DiaTools project has strived to disseminate its results and findings through open symposia, practical workshops, research exchanges and scientific publishing. Through these dissemination actions, there DiaTools project has helped to spread the knowledge about novel tools for minimally invasive diagnostics to the research community, healthcare professionals and the biotech and pharmaceutical industry. Moreover, through its cross-disciplinary nature the project has been instrumental in fostering a next generation of scientist and biotech entrepreneurs. In total, the DiaTools project has resulted in more than 60 scientific publications and nine patents or registered designs.
A broad repertoire of molecular tools, established both before and during the DiaTools project, are now nearing clinical routine applications. The great increase in capacity for protein measurement in e.g. plasma and in other body fluids, including microdialysis filtrates from lesions, is described elsewhere herein. In the course of the project in situ Proximity Ligation Assay (PLA) has been shown to provide unique value for monitoring protein activation markers in patient material as superior prognostic markers for therapy response, and techniques have been established to permit parallel analyses of many such markers. The set of molecular tools have been implemented on each of three flow systems pursued by other partners. Finally, DiaTools has provided resources for initiating work on a new generation of molecular tools that will allow minimally invasive tumor diagnostics by assessing the presence of tumor specific mutations at very low concentrations among already scarce cell free DNA in blood. The technique combines our earlier selector probes for parallel and specific capture of target DNA sequences with a novel technique called superRCA that very accurately interrogates targeted molecules for sequence variation, and in the process, creates large, easily detectable fluorescent products. This work initiated during DiaTools is now being taken forward towards a radically new and valuable minimally invasive diagnostic.
The DiaTools project has also resulted in deeper knowledge about how signaling pathways are disrupted in leukemia and lymphoma cells. This knowledge will be implemented in diagnostics and treatment of these diseases. By selector-based re-sequencing of the NF-kB pathway with proximity-ligation assays, we could for the first time provide a molecular basis for NF-kB activation in CLL; the prime finding being the recurrent truncating mutations of the cytoplasmic inhibitor IkBe (encoded by NFKBIE) in CLL. We could show that NFKBIE mutations were highly enriched in poor-prognostic, stereotyped subset patients, probably contributing to their very dismal prognosis, and resulted in reduced IκBε protein levels, loss of physical interaction between IκBε/p65, as revealed by proximity ligation assays, and increased p65 phosphorylation. Considering that IkBe truncating mutations were also identified in other B-cell-derived malignancies, hence alluding to a common pathobiology for NF-kB activation, components of the NF-kB signaling pathway emerge as possible targets for future targeted therapies in CLL and, possibly, also other mature B-cell lymphomas.
The DiaTools project has lead to introduction of a new class of targeted digital analysis methods for non-invasive diagnostics. These technologies will be competitive compared to NGS-based technology because of their ease of use, and scalability in number of processed samples.
Automated molecular analysis
We have developed the first integrated device for C2CA analysis of tumor DNA in solution. The microfluidic set-up simplifies the multi-step protocol making it suitable for clinical implementation. The established concepts and modules are generic and open for further applications.
Diagnostics of cells and tissues
The development of the EPHESIA technology for isolation and analysis of circulating tumour cells will have clear impact on diagnosis and treatment of several forms of cancer. The benefit of the EPHESIA system lies in the fact that the captured cells can be directly analyzed on the columns. This opens possibilities for faster and more detailed characterization of CTC compared to the systems currently in use.
We have further worked on further developments of the in situ genotyping technology that was originally developed within the FP7 READNA project. We have demonstrated somatic mutation detection in morphologically intact cells, and we have developed an in situ sequencing method that allows highly multiplexed molecular profiling in intact cells and tissue sections. We are currently exploring these sets of tools and technologies to achieve cellularly resolved molecular profiling of tumor tissue sections to help dissect different molecular profiles obtained in bulk extracts from tumors to distinct populations of cell-types and –states, and subclones of cancer cells.
The project has lead to extensive collaboration between research groups at Stockholm University and Institut Curie which will continue after completion of the DiaTools projects.
Our work on in situ mutation analysis has lead to further developments within the EU FP7 CareMore project. In this project the technology developed in the DiaTools project will be further optimized for mutation analysis in morphologically intact CTCs. For example, to improve sensitivity preamplification and direct RNA detection are explored. We are also exploring in situ mutation analysis as a pre-amplification step in a CTC monitoring biosensor system developed in the FP7 CANDO project.
With regards to magnetic micro- and nanoparticles developed in DiaTools, rather complex techniques of multiple swelling and polymerization was elaborated and substantially improved. Great advantage of the newly developed microspheres consists in their cheap production compared with the commercial products. Their use in the molecular diagnostics may provide tailored and cost-efficient treatment for infectious disease and cancer. The particles enable target-specific isolation of proteins from complex clinical samples and removal of reagents, inhibitors and contaminants between the reaction steps. This may be beneficial for work in any clinical or biochemical laboratory. Moreover, synthesis of surface modified magnetic nanoparticles was mastered.
Thanks to the DiaTools, cooperation with the Uppsala University, the Stockholm University and the Curie Institute was established and since October 2014, the Institute of Macromolecular Chemistry participates in the new COST project “Multifunctional nanoparticles for magnetic hyperthermia and indirect radiation therapy (RADIOMAG)”.
Potential impact of the CyToF technology
The impact of the accomplishments from the Nolan lab at Stanford University resulting from the DiaTools project allowed us to expand our capabilities for measuring some new attributes of single cells beyond the scope of protein expression and protein activation states.
For example, measurements of gene expression are a fundamental tool to understand how genetic networks coordinately function in normal cells and tissues and how they malfunction in disease. Therefore including measurements of messenger RNA molecules and proteins simultaneously will greatly affect our conceptual understanding of diverse biological processes with broad implications for both basic and clinical research.
Additionally, including subcellular localization of proteins, nucleotides in our assays enables single-cell, subcellular visual and spatial information to be obtained from a non-visualizable platform with throughput and detection limit significantly improved relative to microscopy. This new attribute shows the relationships between signaling pathway activation and nuclear entry and transcriptional activation again with significant implications to increase our understanding of biological processes in health and disease. In its current form can detect proteins in the nucleus and cytoplasm but are currently developing additional reagents with which to look at localization to other intracellular organelles.
Given the complexity of the large datasets, measurements of upwards of 40 parameters per single cell, new computational tools are required to unravel the biology addressed by mass cytometry assays. Our approach is to develop multiple complementary approaches that will recapitulate know features while simultaneously uncovering new biology. The DiaTools project hasve enabled us to make a significant contribution to biomedical research with opportunities for continuation on multiple levels.
Commercial impact Olink:
The DiaTools project has given Olink a great opportunity to improve our main product line, Proseek multiplex. First, we designed and evaluated a new inter-plate control sample (IPC) with new normalization procedure that reduced the inter-plate and batch-to-batch variation. Furthermore, we have improved the qPCR detection protocol in multiplex PEA, which improved the intra-plate precision. The IPC sample and new qPCR detection protocol were further developed by our production department outside of the DiaTools project and finally introduced to the Proseek multiplex products and launched in October 2013 and July 2014, respectively.
The DiaTools project has helped Olink Bioscience to develop a prototype protocol for combined multiplexed in situ proximity ligation (PLA) and padlock probing (PLP) with either readout by flow cytometry or microscopy. This protocol has been transferred to another FP7 project that Olink is coordinating (CareMore), in which it is being further developed for characterization of ER and pHer2 protein expression levels by PLA, and mutation status of a relevant transcription factor by PLP in circulating tumor cells from metastatic breast cancer patients. The combined PLA/PLP protocol also holds great potential for future development into an in vitro diagnostic reagent for molecular pathology.
The single-cell protocol developed within the scope of the DiaTools project makes the multiplex PEA protein quantification of a suitable method for characterization of rare cells, such as studied on circulating tumor cells and other.
Commercial impact Qlinea:
While the cell imaging capabilities of Q-lineas modified reader has no straightforward commercialization potential, both reader concepts have been taken further. The ASMD amplification/readout system first designed has been taken to a TRL7 level, and will be further developed within the IPODS project, sponsored by the EDA.
The microarray-based amplification and detection system has clear clinical diagnostics potential, and is being developed to its full potential within Q-linea. Furthermore Q-linea has developed a significantly better performing single use microarray consumable compared to current state of the art. This will likely add to the commercial impact of future Q-linea ivd system development. The microarray consumable also enables it to be a product in its own right to be sold to other third parties.
Initially targeted for the defence market, Q-linea´s evolved ASMD amplification/readout system has gathered interest in the homeland security setting, and has been successfully evaluated in a relevant setting, biothreat agent dissemination in a subway station.
The high-multiplex system, on the other hand, has raised interest from an infectious agent ID point of view, with focus on microbial ID and antibiotics susceptibility assessment. The characteristics of this platform potentially sets it apart from competing technologies within these fields, providing a new tool with substantially shorter lead times to pathogen ID and resistance profiling. Further development of this platform is ongoing.
The sample preparation technology developed, capable of processing large volumes ~5ml of whole blood, is judged to be better than current state of the art. This can be produced as a standalone product but also enables Q-linea to enter blood testing ivd markets
Commercial impact Fluigent:
Two products which have been partially developed through the DiaTools project have been commercialized by Fluigent: the Easy Switch Solutions and the Flow-Rate Platform, both hardware and software. The ESS platform enables the user to inject sequentially different fluids in a microfluidic system. Thanks to the Flow-Rate Platform, the user can monitor the flow-rate in its microfluidic environment. Within the project, needs for individual flow control units have emerged as well as pre-integrated flow control platform. Specific studies will be part of the new developments.
List of Websites:
http://www.diatools.org/
Project Coordinator:
Professor Ulf Landegren
Department of Immunology, Genetics and Pathology and SciLifeLab, Uppsala University
Box 815, SE-751 08 Uppsala, SWEDEN
E-mail: ulf.landegren@igp.uu.se
Tel: +46 18 4714910
Project manager:
Associate Professor Erik Ullerås
Department of Immunology, Genetics and Pathology
Uppsala University
Box 815
SE751 08 UPPSALA, Sweden
+46 18 471 44 93
+46 704 30 74 85
erik.ulleras@igp.uu.se
The EU FP7 DiaTools project took as its aim to develop means for advanced, multimodular molecular analyses of patient samples obtainable by minimally invasive means. The successful development and implementation of such techniques can have profound impact on clinical medicine, but this will depend on progress along several dimensions that have all been organizing principles for DiaTools, as briefly described below:
Advanced molecular tools. Molecular recognition reactions assume central roles in diagnostics, and the partners have established a broad repertoire of techniques for analysis of nucleic acids and proteins. In situ PLA was used to characterize fusion proteins as diagnostic assays in chronic myelogenous leukemia patients through convenient flow cytometric analyses at UU. The proximity extension assay was used by Olink to successfully measure sets of 92 proteins from dried blood samples collected by finger pricks, enabling wellness studies. UU has further expanded the repertoire of molecular tools to include patented nFold, ExCirc, CutLig and superRCA probes of value in a growing range of analytic contexts. SU established highly multiplexed RNA analyses in cells and tissues through in situ sequencing of barcoded padlock probes, and they developed digital PLA protein assays for precise quantitation.
Advanced devices for analysis of cells and detection of signals. EPHESIA devices were developed at Institut Curie as highly efficient and convenient means to isolate circulating tumors cells from blood, for enumeration and subsequent molecular analysis, and Fluigent contributed complementary techniques required for trapping cells in fluidized beds. A major breakthrough was achieved with pioneering work at Stanford on multiparametric measurements of cells by the CyTOF, with applications to e.g. define shifts among different blood cells in disease. An amplified single molecule detector developed at Q-linea provided digital readout for a range of molecular detection reactions, and improved paramagnetic micro- and nanoparticles were constructed and made available to DiaTools participants by the partner in Prague.
Clinical validation. An important aspect of the project was contributed by clinical partners in Paris and Uppsala advising on clinical diagnostic needs and requirements. The partners in Paris also established an effective PCR-based approach to detect rare mutant sequences among cell free DNA in patients with uveal melanoma for disease follow-up, a growing diagnostic need. Other disease-associated features that were in focus were the circulating tumor cells whose isolation and enumeration provide access to tumor cells through a simple blood draw. Among many molecular features that were of diagnostic interest in cells isolated from peripheral blood was the HER2 status that correlates with prognosis in breast cancer, and fusion proteins heralding disease recurrence in CML and AML.
Impact. The DiaTools has made considerable progress along the several dimensions we set out to pursue. Besides important scientific publictions, several different analytic modalities were established, some of them ready for clinical implementation while others represent important research tools to validate biomarkers that only in a longer perspective will reach the clinic. Among our highlights are the early work on the CyTOF, now being rapidly implemented for research and in the clinic, and a range of novel tools that will permit e.g. multiplex analyses of rare mutant DNA, RNA or cells through liquid biopsies. Highly precise, parallel protein measurements in plasma developed by the partners are also rapidly making inroads in the clinic.
Through the above progress, the DiaTools project has provided a basis for industrial growth and expansion of markets, with application mainly in research and by big pharma in initial phases, but gradually also suited for clinical implementation. The project has also successfully promulgating the partners’ tradition of fostering researchers and entrepreneurs through research training, and via DiaTools courses and symposia.
Project Context and Objectives:
The overarching aim of the DiaTools project has been to enable multimodal molecular analyses of patient samples that are accessible by minimally invasive means. This is a central goal in capitalizing on in progress understanding the molecular underpinnings of normal and pathological physiology. Along with the vastly increased molecular insights, greatly improved techniques are becoming available for characterizing molecular states of tissues and body fluids with important consequence for future healthcare. The partners of this the DiaTools project have made important contributions to this technology development and to applications of the techniques. With the DiaTools project we have undertaken the important task of developing and adapting high-performance assays and equipment in order to enabling analyses of patient samples that can be readily accessed without surgery, and that can be used to investigate outpatients or to continuously monitor the effects of an ongoing therapy. Such analyses of samples obtainable through noninvasive measures will be a crucial element of a future healthcare with detailed, repeated molecular diagnostic investigations and personalized therapy.
The DiaTools project has developed tools to isolate rare cells from blood, for example circulating tumor cells, and to conveniently analyze their nucleic acids and proteins for precise diagnosis and therapy monitoring. We have also analyzed established techniques to analyze cells and individual macromolecules in high-throughput flow systems. Here we have used both standard flow cytometers, and one of the partners has also pioneered the use of a CyTOF instrument capable of 50-parameter molecular measurements of individual cells in very high throughput. Other partners have also developed a system to digitally count in flow individual detected macromolecules, represented by individual amplification products that can be enumerated in a microfluidic system with excellent specificity of detection.
Yet another aspect of the project has been geared towards comprehensive molecular analyses of body fluids, building in part on the highly resolving proximity ligation technique for protein detection developed by the partners. Our clinical partners supplied both relevant research challenges, medical and logistical insights, and suitable samples, and they have been pushing the sensitivity for detecting rare mutant DNA sequences to very low levels, an application that is assuming increasing importance for monitoring cancer therapy.
Project Results:
The DiaTools project has focused on development and application of methods, assays and devices for molecular analysis of patient samples obtained by minimally invasive means.
The small amount of sample available in this context puts high demands on the analysis technology in terms of specificity and sensitivity. The DiaTools consortium has addressed this challenge by working in parallel to optimize molecular analysis technologies as well as detection devices. The resulting assays and devices have been integrated and to some extent tested in a clinical setting, mainly in cancer diagnostics.
ADVANCED MOLECULAR TOOLS FOR PROTEIN DETECTION
The specificity and sensitivity of analysis of protein biomarkers has been pioneered by scientists at Uppsala university (UU). The central theme has been to utilize pairs of affinity reagents (antibodies) to increase specificity and sensitivity.
Proximity ligation and extension reactions for protein detection.
Antibodies with attached DNA strands provide new, valuable opportunities to engineer high-performance protein assays. In these assays, protein recognition via sets of antibodies with attached oligonucleotides is combined with molecular genetic reactions to copy, ligate, amplify and distinguishing reporter DNA sequences for efficient readout in solution or in situ, individually or in high multiplex. Proximity ligation and extension reactions, developed at UU and at Olink are excellently suited for minimally invasive protein diagnostics, and they have been a central aspect of the DiaTools project.
Proximity extension assays for biomarker discovery and validation.
Proximity extension assays (PEA), developed at and commercialized by Olink employ pairs of oligonucleotide-modified antibodies directed against the same protein for detection in solution phase. An initial incubation is followed by a reagent addition step that serves to initiate a DNA extension reaction between oligonucleotides on proximate antibody pairs. The addition also has the effect of reducing background signals. The resulting DNA strands are then amplified in realtime PCR reactions to identify and quantify detected proteins. The DNA extension and amplification reactions are designed only to detect products from cognate pairs of antibodies. This has the important consequence of allowing extensive multiplexing without any increased risks of formation of crossreactive products between all the reagent pairs combined in a single reaction. In this manner, each 1 µl sample of plasma or cell lysate can be interrogated using antibody pairs directed against 92 proteins, with another four assays included as internal controls.
Diagnostic applications of PEA.
The lab at UU has collaborated with Olink to establish a set of assays for more than 300 proteins, with the aim to monitor plasma from venous blood samples for levels of these proteins as possible indicator of a wide variety of disease processes or tissue trauma. In one study, a set of 250 proteins was measured in microdialysis samples from the brains of five neurotrauma patients. The samples were collected at 3 hr intervals during up to 6 days of care at the neurotrauma unit at Uppsala University Hospital. The assays revealed striking variation of levels of several or the investigated proteins over time as well as between different patients; in decreasing trends over time for some markers and with gradual increases among the up to 40 different samples for other proteins. These assays may provide valuable opportunities to assess the state of health of the investigated patients, and to estimate the individual prognoses and select optimal care.
Fine needle biopsies and single cells.
The PEA assays are also useful for measuring protein levels in lysates of tissues, and fine needle biopsies represent a useful means of accessing tumors via a minimally invasive procedure. Lysates from tissue left in the cannula after samples have been ejected for cytology and provide more than sufficient material for analyses via multiplex PEA. In fact, the sensitivity of PEA is often sufficient to measure protein levels in single cells, and this provides a valuable means to investigate intratumor heterogeneity and characterize the different clones of cells present. UU is currently conducting a study where individual cells from short term cultures of gliomas have been isolated and investigated with respect to 88 proteins and 88 transcript each. The results reveal differences between cells in susceptibility to treatment with bone morphogenetic factor 4, providing a much better resolved view of the therapeutic potential of the agent in different patients.
Ultrasensitive protein detection.
There are now extensive studies underway in Uppsala, also outside the DiaTools project, where the PEA technique is used to screen panels of protein biomarkers for differences in levels between patients and controls, with so far 20 000 samples analyzed and 1.5 million proteins measured. These studies regularly identify promising markers, but in order to fully take advantage of the potential for early detection of disease using such markers it is important to further increase sensitivity of detection of markers present at very low levels. To this end, UU has established a variant of proximity ligation referred to PLARCA, where target molecules are captured from 50 µl sample volumes, before being interrogated using pairs of oligonucleotide-modified antibodies that jointly give rise to a circular DNA strand, for rolling circle amplification. This mechanism is identical to that used in in situ PLA, and the assay provides added specificity and sensitivity due to the requirement for three interactions, and an ability to measure larger sample volumes than in PEA, amongst other features. While PEA typically offers sensitivity similar to standard sandwich ELISA reactions, albeit in multiplex and very small sample volumes, PLARCA routinely results in 1 000-fold greater sensitivity than ELISA. This improved sensitivity may translate to earlier detection of disease-specific markers during the course of disease.
New molecular tools for enhanced diagnostics.
The group at UU has developed a series of molecular tools that have become widely used in research, licensed by several biotech companies, and formed the basis of four spin-out companies, including the DiaTools partner Olink. A central aim of the DiaTools project has been to now also enable diagnostic applications of the previously developed padlock and proximity probes. Meanwhile, the DiaTools project has also provided a valuable opportunity to develop new classes of molecular tools for enhanced molecular analysis in both research and diagnostics. A total of seven patent applications have been filed, although most of the procedures described therein have not yet been published in the academic literature. Very briefly, we developed procedures for multiplexed in situ PLA for parallel visualization of proteins and their interactions in cells and tissues in pathology examinations. Unfolding probes are designed to detect either nucleic acids or proteins with any desired levels of specificity, singly or in high multiplex, producing circular reporter DNA strands for convenient detection. ExCirc and CutLig probes serve to detect RNA and DNA sequences, respectively, capturing part of the target sequences in circular DNA strands. Super rolling circle amplification (sRCA) and padlock sRCA are variant techniques that permits even single detected molecules to be visualized as large clusters of DNA strands, easily detected and quantified with excellent precision using e.g. flow cytometry. The technology also provides exceptional specificity in analyses of rare mutant DNA sequences in a vast excess of normal sequences. RCA reporters, finally, are probes that promptly initiate an RCA upon encountering their target sequences for rapid, convenient detection.
While all these probes and techniques are expected to find applications on their own merits, there is a particular value in combining several of these for particular diagnostic applications. Examples of applications include investigations of circulating cell free DNA in plasma for mutations that may reflect the presence of a tumor in a patient, highly specific detection and quantitation of circulating tumor cells in blood, or rapid high-performance assays at the point of care. These are all important objects of the DiaTools project, and the new molecular tools may provide new and further improved means to develop such diagnostic assays, beyond the conclusion of the DiaTools project.
Proximity Ligation Assay (PLA) for detection of BCR/ABL fusion protein in CML patients.
A PLA method for detection of fusion protein was one of the central aims in the plans for the DiaTools project. Fusion proteins are a hallmark of some types of blood cancer including chronic myeloid leukemia (CML). Detection of fusion proteins are central in diagnostics of CML and have so far mainly been performed through detection of the fusion RNA using polymerase chain reaction. We have now successfully developed an assay to directly detect BCR-ABL fusion protein, present in malignant cells in most individuals with chronic myeloid leukemia (CML) The PLA method efficiently detects the fusion protein both in cell lines and CML patients.
The method is now being evaluated in clinical material from CML patients and the results are compared with routine PCR analysis. So far 10 patients with low tumor burden have been tested, as well as 3 patients with new diagnosis, and high tumor burden. The results comparing the fusion protein detection in flow cytometry and PCR detection of fusion transcripts are similar, and in some samples the fusion protein is detected when the transcript is undetectable in the patients. Direct detection of fusion proteins in the CML patient cells will hopefully speed up and improve the diagnosis of this disease. In addition, PLA for multiplex testing of fusion proteins in leukemia has been developed and evaluated. Multiplex detection of different fusion proteins at the same time will further improve diagnosis and in other forms of blood cancer including Acute Myeloid Leukemia and Acute lymphoblastic leukemia. In the DiaTools project, antibodies for fusion proteins involved in AML and ALL have been tested in 4 different cell lines, each carrying a different fusion protein. The assays have been validated using in situ PLA with microscope readout or flow cytometry readout.
In addition to development of diagnostic assays for blood cancers, researches at UU have investigated the pathways and mechanisms underlying the disease. Using methods developed in the DiaTools project, mutations affecting the signaling pathways in cancer cells have been identified and the effects have been examined in patient samples. Since the NF-kB signaling pathway is constitutively activated in chronic lymphocytic leukemia (CLL) cells, we first utilized the selector-based targeted enrichment technology, developed in-house, and performed next-generation sequencing of 18 NF-kB core complex genes in a discovery and validation CLL cohort (in total 315 patients). Our prime finding was a recurrent mutation within IkBe (encoded by NFKBIE), a negative regulator of NF-kB in B-cells, that was found recurrently mutated in 21/315 cases; 13 carried an identical 4bp frame-shift deletion. By screening an additional 416 CLL cases, the NFKBIE deletion was predominantly seen in poor-prognostic patients, and minor subclones and/or clonal evolution were also observed, thus linking this recurrent event to disease progression. This 4 frame-shift deletion leads to introduction of a stop codon and subsequent loss of the ankyrin repeat region, resulting in a truncated form lacking this domain, which is a prerequisite for protein-protein interaction, i.e. binding to p65 or other transcription factors. To further explore if the physical interaction between IκBε and p65 is influenced, we applied proximity ligation assays, as a novel means to study their interaction, revealing that this truncating mutation resulted in reduced IkBe/p65 interaction, at the single cell level, compared to wild-type patients. Finally, we analyzed the presence of the NFKBIE-deletion in different lymphoma entities (in total 372 lymphoma samples) and this genetic event was also identified as recurrent, pointing to a common mechanism of NF-kB deregulation during lymphomagenesis.
While some research groups have concentrated their efforts on development of novel molecular detection technologies, other DiaTools partners have developed new methods to read-out and record the detected molecules.
Digital protein analysis for non-invasive diagnostics.
Researchers at Stockholm University (SU) have developed a truly digital readout for PLA reactions that can be readout by Qlinea’s Amplified Single Molecule Detector (ASMD) reader. We converted two PCR-based solid phase PLA assays to Rolling Circle Amplification (RCA) based reactions. The PLA assays targeted IL-6 and VEGF and the PLA ligation product was digested and circularized according to the circle-to-circle amplification (C2CA) scheme. The formed circles were amplified by RCA and digitally recorded by ASMD. The digital quantification of PLA products resulted in an extrapolated limit of detection of approximately 5 fM (0.1 pg/ml) and average CVs of 7% across a dynamic range of 5 logs. PCR readout performed in parallel resulted in 1.5 fold less sensitive limit of detection of approximately 9 fM, and the average CVs for the PCR readout was 18% over a comparable dynamic range. ASMD thus provided at least two fold improved precision compared to the PCR-based readout. A similar quantitative precision was observed in detection of the cytokine vascular epithelial growth factor (VEGF) in buffer.
ASSAYS FOR MULTIPLEX PROTEIN DETECTION
PEA development and optimization
Better capabilities to diagnose disease at the early and most curable stages will greatly improve human health and reduce health care costs. Patient stratification is also in need of better diagnostics to facilitate the selection of appropriate patient care. Preliminary data in the literature claim improved diagnostics with the use of multiple complementary biomarkers, as no marker can today single-handedly diagnose all cases with desired accuracy for a certain cancer type. Therefore, a specific and sensitive immunoassay at high multiplex level would allow for more rapid identification of new biomarkers and combinations of biomarkers.
One unique feature of the proximity extension assay (PEA) described above is the extremely low sample consumption, which makes it especially suitable for biomarker studies of scarce samples, such as small needle biopsies and lysates of single-cells to mention a few. PEA possesses high sensitivity, high specificity, great multiplex capability, and potential for rapid assay development time and high throughput screening.
In the DiaTools project, the biotech company Olink has developed and optimized the PEA technology. One of Olink’s current products on the market, Proseek multiplex, allows for detection and quantification of 92 proteins in only 1 µL sample. However, the initial version of the product did not have sufficient inter-assay precision to generate quantitative data across a large set of samples at sufficient precision, but the DiaTools project has enabled us to develop a new inter-plate control and a new normalization procedure that decreased the inter-assay and batch-to-batch variation.
Furthermore, in order to generate PEA quantification at the single-cell level, high intra-assay precision is critical. During the course of the DiaTools project we identified a position-dependent variation in signal that seemed to occur during the qPCR detection step. This led to worsened precision and could, in worst case, result in misinterpretation of the data. During the project we managed to optimize the detection protocol, which minimized the position effect, and thereby improve the intra-assay precision.
At the conclusion of this collaborative project, the research and innovation conducted within the consortium has allowed Olink to further strengthen our competitiveness in the international arena of high-throughput biomarker research.
PEA single cell development
As mentioned above, Uppsala University has in collaboration with Olink generated a protocol for protein quantification in single-cells using PEA. In brief, single cells are sorted by flow cytometry into a 96-well plate containing lysate buffer, after which 1 µL is transferred to a new plate for standard PEA analysis. The protocol is very easy to apply to the standard Proseek Multiplex assay and demonstrating sufficient sensitivity and precision to enable multiplex protein quantification at the single-cell level.
Optimization of PLA and Padlock probes (PLP) for flow cytometry
A protocol was developed combining the unique aspects of the PLA and PLP assays into one, without significantly decreasing the sensitivity or specificity of the individual assays. The assay was developed to allow higher multiplexing degrees. The combined assay performs well when done on slide or in solution, allowing for quantification (counting RCA products) using a microscope, or doing cell population profiles using flow cytometry.
The PLA/PLP multiplex protocol was developed to be performed within regular working hours and to not deviate too far from the standard PLA and PLP protocols, in context of number of steps and time intervals. The resulting protocol is actually shorter (7-8 hours) and easier to perform than doing the PLA and PLP assays separately (either simultaneously or sequentially).
ADVANCED MOLECULAR TOOLS FOR DETECTION OF DNA AND RNA
Digital DNA analysis for non-invasive diagnostics
SU researchers have also developed an approach for non-invasive prenatal diagnosis (NIPD). Our approach is based on quantification of differentially methylated CpG sites on chromosomes 21 and 18 between maternal and fetal DNA, using padlock probes and RCA using a digital ASMD readout. We show that fetal DNA can be enriched from maternal DNA by a factor of 5, and that fetal DNA in fractions down to 30-40% can be detected using this pilot assay configuration targeting 7 CpG sites. The sensitivity of the approach can be improved by targeting many differentially methylated CpG sites.
Circulating tumor DNA
In addition to detecting circulating tumour cells, described below, we have developed methods to detect tumour cell DNA present in the blood. We have developed an automated C2CA device for diagnostic mutation detection in circulating tumor DNA. It is a multi-step, bead-based protocol that gains in automation to enable clinical use at a point-of-care. Efficient re-suspension and mixing of the beads is critical for assay performance. We have therefore integrated the molecular protocol into a microfluidic device which keeps the beads in dynamic suspension based on a fluidized bed approach that enabled us to automate and simplify the full assay. The fluidized bed system was developed at Institut Curie and was operated by a pressure control system from Fluigent. Within the DiaTools project it was optimized for DNA analysis and fabricated in a thermo-resistant polymer to function at elevated temperatures in our protocol.
A multiplex padlock and C2CA assay was successfully developed for genotyping of the seven most common mutations in the frequently mutated oncogene Kirsten rat sarcoma viral oncogene homolog (KRAS). The protocol was adapted to fit the microfluidic chip format in two stages. A model assay was first developed and integrated in the fluidized bed chip for a 1RCA protocol with target capture and RCA amplification on beads followed by release of the products and fluorescent ASM readout. The target capture efficiency was 93% with the magnetic bead plug. Next we developed a new solid-phase C2CA protocol which simplified integration of the whole protocol on chip. The target capture, upconcentration and the first amplification is performed on magnetic beads in the fluidized bed and the monomers from the digested RCA products are re-captured to probes on the surface of a second chamber where the second amplification and labeling takes place. Amplification products are visualized in a microscope or in the future with a dedicated array reader (developed by Qlinea). A fully integrated polymer chip was successfully designed and fabricated for the solid-phase protocol. The surface in the second chamber was functionalized through direct UV binding of the probes and the complete assay tested on chip.
Detection of somatic mutations in intact cells
In addition to detecting mutations in freely circulating DNA, we have also applied our in situ genotyping method for detection of somatic mutations in intact cells. We developed assays for oncogenic mutations in the KRAS and EGFR genes and verified the performance on a range of clinically relevant patient samples including sections of fresh-frozen tissues, sections of formalin fixed paraffin-embedded tissue and cytological tumor imprints. The results were 100% concordant with results from the routinely used diagnostic method applied on DNA extracted from the same tumors. We demonstrated the sensitivity of the method by detecting rare mutated cells spiked into a background of wild-type cells. By targeting the tumor suppressor gene TP53, we also successfully detected loss-of heterozygosity of this gene. The method should be directly applicable for analysis of intact captured circulating tumour cells (CTCs).
Sequencing of RNA in intact cells
We have developed a technology that, for the first time, allows sequencing of RNA directly in fixed cells and tissues. The technology uses padlock probes and RCA to generate targeted sequencing libraries in situ. The RCA products are sequenced using sequencing-by-ligation chemistry allowing reads of 6 bases to be recorded in situ. This read-length is sufficient for scoring mutations in oncogenes, and in one experiment we detected 5 KRAS mutated cells in the background of 5 000 KRAS WT cells. The read length is also suitable for reading barcodes in the probe library. With 6 bases read-length it is possible to discriminate 4 096 different sequence barcoded probes. We demonstrated this strategy by performing targeted expression profiling of 39 transcripts in intact sections of breast cancer tumor tissue with cellular resolution. The technology should be directly applicable for analysis of intact captured CTCs.
MICROFLUIDIC DEVICE FOR CAPTURE OF CIRCULATING TUMOUR CELLS
In the framework of the DiaTools project, researchers at the Institut Curie in Paris have developed and validated a microfluidic platform dedicated to circulating tumor cells (CTC) capture from patient blood samples. It is a huge challenge as we are looking a few CTCs in billions of “interfering cells” (white blood cells, red blood cells…). The microfluidic platform we developed is called EPHESIA and it relies on immunosorting (related to the presence of specific antigen on CTC membrane).
The system we developed combines an unprecedented level of automation, integration and cost-effectiveness. It essentially involves the self-assembly, in a microfluidic system, of a regular array of columns made of magnetic microparticles coated with antibodies or other ligands directed against the CTCs. It captures cells by surface antigens. It does not require any complicated microfabrication, since the physics of self-assembly does the job of making very regular, high aspect ratio capture columns. Very importantly, all the delicate biofunctionalization steps can be performed in large batches out of the system, reducing cost and increasing reproducibility.
Beyond this, however, the EPHESIA technology completely reverses the paradigm of magnetic sorting, bringing in a dramatic increase in sorting efficiency and sensitivity. In our method, the magnetic particles are first immobilized, and then the cells are flown through. This presents two major advantages: first, the number of particles necessary is proportional essentially to the number of cells to be captured, and not to the sample volume (as in batch methods). This is interesting not only for cost, but also because the use of a large excess of particles leads to cell damage, and renders accurate morphological observation impossible. Second, in our system the interaction between the cells and the immobilized particles is not induced by Brownian motion, but by the well controlled hydrodynamic forces inside the array. Additionally, our system allows the application of all the staining and observation protocols currently used in optical cytology.
During the project, we have first demonstrated the ability of this platform to capture cells presenting different specific antigen expression in order to evaluate our platform capture efficiency in different cellular conditions. We have thus validated the EPHESIA platform with real patient samples in blind study. Experiments have been performed to compare our device with the current standard method for CTC analysis (Cellsearch, Veridex). Very high capture efficiency and purity were achieved with conventional CTC model cells lines (90%, 0,02% of purity). We have obtained similar or better CTC capture with our microfluidic platform both with breast and prostate cancer patient. The platform automation has been greatly improved, especially the fluidic control of the experiment as well as the cell staining and imaging performed after cell capture.
Even if being able to capture a few CTC from blood is already an important information for clinicians concerning patients prognosis, it turns out that getting more information on the proteomic or genetic status of CTCs could help clinicians in treatment orientation. The first CTC characterization we performed on chip is related to immuno-staining to confirm their CTC status. Besides, we have also integrated a molecular analysis method (in situ PLA) developed by DiaTools partners to investigate protein-protein interaction on the cancer cell membrane. Indeed, it is now well establish that protein dimerization occurs on cell membranes and they could be the starting point of signaling pathway that will modify the cell fate. Here we have been interested in HER2/HER3 dimerization. After an important optimization process, we have successfully demonstrated the possibility to integrate Proximity Ligation Assay (PLA) on captured CTCs to evaluate their level of HER2/HER3 dimerization.
Finally, working with columns made of magnetic particles offers the possibility to release the magnetic beads by switching of the magnetic field. We have taken benefit of this to retrieve cells after their capture for subsequent CTC DNA analysis. Thanks to the high purity achieved with our EPHESIA platform, we have been able to detect PI3CKA mutation in less than three cells captured in our platform.
During the project, the microfluidic platform has been greatly improved. Especially, chip material has been investigated to get materials compatible both with mass production, high temperatures and good optical quality. We have thus used thermoplastic materials and we have put efforts both on microfabrication process and on surface treatment strategies. Indeed, we have been working with COC as material since COC is hydrophobic and inert. We have also developed new coating strategies based on physisorption to avoid non specific adsorption of beads and antibodies on the microfluidic chip surface.
DEVELOPMENT OF SINGLE CELL ANALYSIS BY MASS CYTOMETRY
The Nolan Lab at Stanford University is focused on technologies that analyze biological events at the single cell versus an average of such events within a pooled population of many cells. Such technologies can therefore study rare cells, such as stem cells or drug-resistant cells, whose biological attributes would otherwise be overshadowed by more abundant cell types within a mixed cell population. Single cells can be considered as the basic units of organs and tissues akin to individual bricks that comprise the structure of a building.
The Nolan laboratory studies phospho-immune cell and cancer signaling and other metabolic parameters by analysis of biochemical functions at the single cell in primary cell populations. From around 2009, the Nolan Lab introduced and developed mass cytometry, a technology platform that allows one to survey and quantify 40-plus cellular markers on a cell-by-cell basis. Mass cytometry is a novel adaptation of atomic mass spectrometry that enables high-dimensional flow cytometry applications. It employs antibodies conjugated to stable isotopes of the transition elements as reporters, replacing fluorophores previously used in traditional flow cytometry. Overlap of fluorescence emission signals precludes the use of fluorophores to measure more than 15 or so parameters per single cell. In mass cytometry, given that each stable isotope has a distinct mass (no overlap of “mass”) there is a potential to exceed upwards of 100 parameters per cell. The mass cytometer has been developed for high-accuracy elemental analysis. The abundances of the metal reporters bound to cells via antibodies are determined and consequently the levels of a protein to which each antibody binds can be determined. Antibody panels are designed to recognize surface proteins, to delineate a cell type (e.g immune cell subset, epithelial cell, neuron) and their intracellular signaling pathways (e.g pathways involved in proliferation, apoptosis, metabolism, protein translation, transcription).
As with any quantitative technology, there is a stringent requirement for internal and external reagent standards for data normalization. During the DiaTools project, several have been developed in the Nolan Lab and have been made available to the scientific community:
i) Variation in instrument performance can be caused by factors such as instrument calibration, fluctuations in plasma and build-up of cellular debris in the sample introduction components. In order to normalize for these factors, polystyrene beads infused with precise amounts of several lanthanide isotopes are acquired simultaneously with every sample.
ii) In any biological sample dead cells often give false positive results, as they tend to bind non-specifically to many reagents. Therefore, removing dead cells from mass cytometry data is a critical step to ensure accurate results and analysis. A viability dye is used to identify dead cells and operates on the principle that a compromised cell with a damaged plasma membrane permits reagent entry into the cytoplasm, whereas a healthy cell with an intact plasma membrane does not. Recently, a protocol was described for using cisplatin to determine cell viability. It enters dead cells and the platinum can be detected by the mass cytometer and the signal used to quantify them.
iii) Mass-tag cell barcoding labels individual cell samples with a unique combination of six palladium barcodes. Samples can then be pooled for processing and measurement as a single multiplexed sample. The barcoding procedure eliminates variability between samples in antibody staining and instrument sensitivity. We developed a debarcoding algorithm to identify each sample.
New Assay formats
Proximity ligation assay, developed by Professor Ulf Landegren’s group at the Uppsala University is a technology that detects, quantifies and determines cellular localization with high specificity and sensitivity. Protein targets can be readily detected and localized with single molecule resolution and objectively quantified in fixed cells and tissue samples. Over the course of the DiaTools project, we have modified the assay for both fluorescence and mass cytometry to detect: a) messenger RNA in single cells – up to twenty different RNAs simultaneously. b) messenger RNA and protein simultaneously in single cells- up to twenty RNA and proteins simultaneously c) nuclear localization of proteins – the importance of this adaptation lies in the ability to see nuclear localization by a non-imaging technology. Professor Landegren’s group provided the foundational work upon which our adaptations are based.
Mass Cytometry data – lineage analysis
B-cell development as described by trajectory detection at the single cell level. Tissue regeneration is an orchestrated progression of cells from an immature stage to a mature one. A continuum of transitional states exists between these discrete stages. In a recent publication, multi-parametric single cell data was analyzed using an algorithm designed to leverage this continuity. Applied to B-cell lymphopoiesis, the algorithm constructed trajectories spanning from hematopoietic stem cells to naïve B cells. Within these trajectories, there were nascent fractions of B cell progenitors along with regulatory signaling and functional events such as immunoglobulin rearrangement. The study provided a comprehensive analysis of “healthy” B cell lymphopoiesis providing a reference against which to evaluate aberrant B-cell maturation and function in diseases such as B-cell leukemias. Furthermore, the approach maybe applied to other cell lineages.
Computational Analyses
The Nolan Lab continues to develop computational tools with which to analyze high dimensional single cell data. At their core these algorithms are designed to evaluate i) clusters of cells based on their similarity to one another ii) or each individual cell and how it looks in relation to others in a sample and across samples in multidimensional space. The computational analysis tools developed during the project will now be made available to the scientific community. We have several publications and manuscripts in preparation around each type of computational approach.
MAGNETIC PARTICLES FOR USE IN DIAGNOSTIC ASSAYS
Many of the diagnostic assays developed in the DiaTools project require magnetic micro- and nanoparticles with defined surface characteristics and specific binding properties.
Within the DiaTools project, new innovative surface-modified magnetic microparticles and nanoparticles have been developed at the Institute of Macromolecular Chemistry (IMC) in Prague, which suitably complement materials available nowadays on the market. While size of the nanoparticles was in tens of nanometers, typical diameter of the microspheres was in a range of micrometers. Generally, size of the particles was controllable and uniform which ensures the same physical, chemical and biological properties in the applications. The particles contain a reactive functional group available for attachment of a target biomolecule, typically an antibody or nucleic acid. Industrial production of these new particles is a subject of further development.
Magnetic microparticles
(i) Magnetic bead cellulose was developed by modification of sol-gel transition of viscose in the presence of maghemite (γ-Fe2O3) nanoparticles. Though the particle size was rather large (15-40 µm in diameter) and the particle size distribution was broad, the beads made of natural material have very high content of water (i.e. porosity) which makes them ideal material for fast and low-cost high-capacity separations of biomolecules together with the superior flow properties in the chromatography columns. The number of hydroxyl groups in polysaccharide chains makes the beads highly hydrophilic and therefore biocompatible and well tolerated by biosystems. The magnetic bead cellulose was activated with tosyl groups and exploited at Uppsala University as a solid support for sensitive protein detection in femtomolar concentration range using proximity ligation assay.
(ii) By elaboration of multiple swelling polymerization, magnetic monodispersed poly(2-hydroxyethyl methacrylate) microspheres were developed which were then modified with streptavidin. Size of the microspheres was controlled by size of starting monosized seeds and by degree of swelling with activation agents, monomers and porogens. High content of the magnetic compoud in the particles was achieved by repeated imbibing of Fe salts in the pores and precipitation with a base. The particles are easily separated by a magnet and exhibit low non-specific protein sorption if modified by poly(ethylene glycol). Thier performance and the binding efficiency were tested at Stockholm University by the circle-to-circle amplification.
Magnetic nanoparticles
Significant advances were achieved in fabrication of the advanced functional magnetic nanoparticles. Their size, particle size distribution, colloidal stability and magnetic properties were regulated depending on the reaction parameters during the synthesis and the selected production method, i.e. co-precipitation of Fe salts vs. thermal decomposition of Fe organic precursors. Size of the particles according to the transmission electron microscopy was controlled in the range of 10-30 nm.
Functionalization and applications
Surface coating of the particles is of key importance for their applications since it prevents particles from the aggregation, reduces undesirable nonspecific protein adsorption and enables immobilization of a target biomarker via reactive groups. Particles coated with several different compounds are now available at the end of the project including dopamine-hyaluronate conjugate, poly(N,N-dimethylacrylamide), poly(L-lysin), poly(ethylene glycol) derivatives containing small amounts of reactive functional groups (typically carboxyl, amino or tosyl), streptavidin, numerous silica derivatives of controlled thickness, etc. Newly, polyzwiterions were attached from the microsphere surface by reversible addition-fragmentation chain transfer polymerization. While the magnetic nanoparticles were able to successfully label various cells enabling their monitoring by magnetic resonance imaging, the magnetic microspheres were very useful for selective molecular diagnosis of various diseases including via rolling circle amplification. Antibody-bound magnetic particles are prone to easy manipulation and fast isolation of target biomolecules by a magnet and are amenable to automatic processes.
CLINICAL VALIDATION OF DIAGNOSTIC ASSAYS
A number of diagnostic assays have been clinically validated on patient samples at Curie Hospital in Paris. Assays for detection of nucleic acids in solution phase were validated for diagnosis of the eye cancer uveal melanoma. We firstly determined the prevalence of the Q209 mutations of GNAQ and GNA11 genes in 197 uveal melanoma tumors. We found that about 85 % of the tumors harbored only 3 nt mutations. This opened up the possibility to use the detection of these mutations to monitor circulating tumour DNA (ctDNA). For that purpose we set up a technique to detect point mutations in a complex DNA mixture. We chose to use the pyrophosphorolysis-activated polymerization method (bi-PAP) and we demonstrated the capacity of this method to detect tumor DNA (ctDNA) with a one-copy sensitivity. We applied this method to uveal melanoma xenografts by showing a very good correlation between LINE-1 PCR assay and bi-PAP PCR assays. We applied this method to 22 patients with metastatic uveal melanoma (MUM) and found that 21 patients harbored detectable ctDNA. Levels of ctDNA were correlated to the tumor burden.
Second, we compared the prognosis significance of CTC as measured by the Veridex technique (current gold standard for CTC detection) versus ctDNA in metastatic uveal melanoma. GNAQ/GNA11 mutations were characterized in archived tumor tissue. Using the above-mentioned bi-PAP technique, GNAQ626A>T, GNAQ626A>C and GNA11626A>T copy numbers were quantified in plasma from 12 ml of blood. CTC were detected at the same time in 7.5 ml of blood by the CellSearch® technique. Patient characteristics and outcome were prospectively collected. ≥1 CTC were detected in 12 of the 40 included patients (30%; range 1-20). Among the 26 patients with known detectable mutations, ctDNA was detected and quantified in 22 (84%; range 4-11421 copies/mL). CTC count and ctDNA levels were associated with the presence of military hepatic metastasis (p=0. 004 & 0.03 respectively), with metastasis volume (p=0.005 & 0.004) and with each other (p<0.0001). CTC count and ctDNA levels were both strongly associated with progression-free survival (p=0.003 & 0.001) and overall survival (p=0.0009 & <0.0001). In multivariate analyses, ctDNA appeared to be a better prognostic marker than CTC. ctDNA and CTC are correlated and both have poor prognostic significance. In conclusion, CTC detection can be performed in every patient but, in patients with detectable mutations, ctDNA was more sensitive than CTC and possibly has more prognostic value. This first clinical study has been published.
Correlation with clinical disease evolution and metastatic relapses
We are currently building up a very large cohort (target accrual: 800 patients diagnosed with localized non-metastatic uveal melanoma) whereby plasma samples from which circulating DNA is extracted are being collected, stored and analyzed for GNAQ/GNA11 mutations in batch. 500 patients have been included to date (December 2014) in the study, which has been ethically approved. Blood sampling occurs at diagnosis, before and after treatment, and at regular intervals or whenever the patient is monitored at Institut Curie. From this patient set, we expect approximately 25% will develop metastasis. We have already analyzed the ctDNA load in pre-metastatic patients for the first 66 patients without finding true positive samples.
Analysis of circulating tumor cells in breast cancer: HER2 status and inhibition
The assessment of HER2 status is a key parameter in breast cancer management at both early and metastatic stages. While our team reported that HER2 immunofluorescence on CTC is not reliable for HER2 status assessment of the tumor, our first aim was to implement HER2 Fluorescence in situ hybridization (FISH) on CTCs isolated by our in-house developed EPHESIA technique. Curie hospital provided clinical samples of HER2+ breast cancer that allowed the setting up of a new microfluidic chip in which high resolution imaging was possible. This low cost microfluidic chip integrating 3D micro-chambers allowed the capture and analysis of tumor cells, as well as the implementation of standard biological protocols in a chip format with low volume consumption. To validate the performances of our device, immuno-fluorescence labeling and FISH analysis on cancer cell lines and on patient pleural effusion samples were performed successfully.
More specific work on FISH implementation in microfluidic chips was then performed. Namely, we developed a protocol performed entirely in the liquid phase, allowing the immobilization and fixation of cells and their quantitative characterization by FISH. We demonstrated first in cell lines and in clinical cases the potential of this method to perform quantitative copy number measurement and clinical scoring of the amplification of the HER2 gene. This validation was performed in a blinded protocol in two clinical case studies (samples being provided by the Institut Curie hospital), in reference to the gold standard and clinically used method based on glass slides. We obtained a comparable reproducibility and a minor difference in apparent amplification, which can be corrected by internal calibration. The method thus reaches the standard of robustness needed for clinical use. The protocol can be fully automated, and its consumption of samples and DNA probes is reduced as compared to glass slide protocols by a factor of at least 10.
Building on these results, two main projects are now ongoing at the Institut Curie:
1/ The CirCe T-DM1 trial is a phase 2 trial (funded by a pharmaceutical company), which aims at treating patients with HER2-negative breast cancer and HER2-amplified circulating tumor cells with a pure anti-HER2 drug, T-DM1. This trial started in early 2014 in several cancer centers in France. Briefly, patients with HER2 negative metastatic breast cancers are screened for HER-amplified CTC by HER2 FISH on the captured CTC and, if positive, are eligible for anti-HER2 therapy. This trial has been ethically approved.
2/ The CirCe PLA observational study will start in 2015 and is funded by a pharmaceutical company. In this study, HER2-HER3 heterodimers on tumor cells (from both the tumor and the CTCs) will be quantified by PLA (proximity-ligation assay) before and during treatment. We aim at demonstrating that HER2-HER3 heterodimerization is inhibited by pertuzumab, a new anti-HER2 treatment, and to establish HER2-HER3 PLA as a predictive biomarker of efficacy of pertuzumab in HER2-positive metastatic breast cancer.
DEVELOPMENT OF DEVICES FOR ISOLATION AND DETECTION
In addition to molecular methods and assays for molecular diagnostics, advanced devices for isolation and detection have been developed within the DiaTools project. Within the framework of other activities, the company Q-linea has developed technology for detecting so-called Amplified Single Molecules (ASM). These consist of multiple repeated DNA strands with either the exact sequence, or the mirror (or homologous) sequence, of an identified target molecule, collapsed into spheroids large enough to be detected by standard microscopy technologies. The target molecule may be a unique DNA sequence identifying e.g. a specific tumor type, or a specific pathogen. Q-linea has developed a device for detecting and enumerating these entities. The sum procedure is termed ASMD (Amplified Single Molecule Detection), and is characterized by high sensitivity, high quantitativity, and high intrinsic multiplexability. In the DiaTools project the potential of this device to analyze larger objects such as cells for the presence of individual ASM associated with each cell was explored. Matching a specific ASM to a certain cell type may serve to identify specific cancer cells, isolated from blood circulation.
The standard Q-linea ASMD detector was evaluated with respect to performance in analyzing isolated cells. The performance was characterized, and a number of potential improvements were identified, the implementation of which was expected to improve performance.
In the latter part of this effort, a breadboard setup corresponding to a modified ASMD detector was built. This setup realized for most of the proposed improvements, and was at project end ready for evaluation. It should be noted, however, that as the DiaTools project has evolved, it has proven more fruitful to pursue an approach to detect rare circulating tumor cells that is based on microarray technology.
ASMD multiplex reader
Within DiaTools, different approaches to detect both protein and DNA signatures of e.g. tumor cell material have been developed and evaluated. Detection of a rare signature involves recognition, amplification, and readout of the target. With regard to protein and DNA target detection, the recognition step differs, but the amplification and readout steps can be essentially the same. Common to these approaches is the need of a dedicated, efficient amplification and readout device. As part of the project, Q-linea has developed two distinct systems to perform these tasks.
The initially developed system amplifies targets into distinct, countable ASM, which are read by a microscopy-based fluidics readout, passing the ASM in a liquid flow in front of a detector. Characteristics of this system are short time to answer, high sensitivity, quantifiable results, high throughput, but low to medium multiplexability.
As the project matured, it became clear that higher multiplex levels than achievable with the initially developed system was needed. This led to the development of a second detection system, based on microarray detection. This final system has delivered multiplex levels of 99+, with a potential of several thousand. The system also comprises sample preparation from blood samples, making a complete sample-to-result system. As this direction of work was set in the latter part of the project, an integrated demonstrator was not achieved, however the concept is demonstrated by a prototype of automatable modules. This concept in itself has been shown promising, and will be developed further within Q-linea.
Within the DiaTools project, the company Fluigent has developed a prototype device for isolation and concentration of proteins and nucleic acids from blood and from fine needle aspirates. The device is based on membrane filtration through a tubular membrane. An integrated module consisting of the required pumps, valves and other components, with the possibility to vary experimental conditions such as time or flow rate has been constructed and tested.
With this concentration module, a 10-fold concentration of proteins with 71 % recovery has been achieved within a duration of the global process of 5 minutes. Under the conditions used, the membrane has a life-time of > 150 hours.
This module can be adapted to molecules to be concentrated (modification of the pore of the membrane) and specifications of the process (duration, volume to concentrate, efficiency…).
Regarding the capture of rare cells from biological samples, Fluigent together with Institut Curie has developed a fully automated platform. Different tools and platforms have been successfully developed and integrated (electronic, mechanical and software integration): flow-rate platform (to monitor and control the flow-rate of liquid flowing in a microfluidic system) and valving platform (to inject sequentially different fluids in a microfluidic system).
Thanks to the different platforms and the development of a script editing tool, the whole sample processing can be automated.
Potential Impact:
The DiaTools project has already had considerable impact and we are confident that its effects will continue to be felt over coming years. Several of our projects are now ready to be tested in clinical trials to assess their relative merits over older methods currently in use. The following are some of the highlights of research applications that are close to clinical implementation.
Proximity ligation and proximity extension techniques have been positioned as world-leading methods for high-throughput, low sample consumption protein marker analysis in plasma, and very large-scale investigations are now in progress. So far some 23 000 plasma samples have been interrogated for a total of more than 2 million protein determinations at Uppsala University. On the basis of findings a panel of around 20 markers is currently being assembled for clinical implementation in cardiovascular disease.
The CyTOF technique, pioneered by one of the participants in the course of the DiaTools project is rapidly assuming a central role in a current drive towards single cell biology, and the technique has been successfully combined with the in situ PLA technique in collaboration between two partner labs, greatly extending the range of markers that are accessible to analysis. A molecular method for detection of fusion proteins in flow has been established and is ready for clinical studies to assess its practical utility, estimated to be significant.
The isolation of single tumor cells has been successfully accomplished using novel affinity matrices, and the cells have been interrogated using molecular tools provided by other partners.
Important technology has also been established and perfected for imaging mutation specific transcripts in individual cells, which can be retrieved by methods such as fine needle biopsy to assess the contribution of malignant cells in biopsy material, and analogous methods are pursued by other partners to allow assessment of cellular signaling via protein interactions and modifications.
The methods, assays and devices developed in DiaTools will ultimately result in improved diagnostics and treatment, primarily of different forms of cancer. In addition, these tools assays and devices will broaden the possibilities of the companies participating in the project and in the long run also strengthen the pharmaceutical industry through the availability of novel tool for evaluation of treatment response of novel drug candidates.
The DiaTools project has strived to disseminate its results and findings through open symposia, practical workshops, research exchanges and scientific publishing. Through these dissemination actions, there DiaTools project has helped to spread the knowledge about novel tools for minimally invasive diagnostics to the research community, healthcare professionals and the biotech and pharmaceutical industry. Moreover, through its cross-disciplinary nature the project has been instrumental in fostering a next generation of scientist and biotech entrepreneurs. In total, the DiaTools project has resulted in more than 60 scientific publications and nine patents or registered designs.
A broad repertoire of molecular tools, established both before and during the DiaTools project, are now nearing clinical routine applications. The great increase in capacity for protein measurement in e.g. plasma and in other body fluids, including microdialysis filtrates from lesions, is described elsewhere herein. In the course of the project in situ Proximity Ligation Assay (PLA) has been shown to provide unique value for monitoring protein activation markers in patient material as superior prognostic markers for therapy response, and techniques have been established to permit parallel analyses of many such markers. The set of molecular tools have been implemented on each of three flow systems pursued by other partners. Finally, DiaTools has provided resources for initiating work on a new generation of molecular tools that will allow minimally invasive tumor diagnostics by assessing the presence of tumor specific mutations at very low concentrations among already scarce cell free DNA in blood. The technique combines our earlier selector probes for parallel and specific capture of target DNA sequences with a novel technique called superRCA that very accurately interrogates targeted molecules for sequence variation, and in the process, creates large, easily detectable fluorescent products. This work initiated during DiaTools is now being taken forward towards a radically new and valuable minimally invasive diagnostic.
The DiaTools project has also resulted in deeper knowledge about how signaling pathways are disrupted in leukemia and lymphoma cells. This knowledge will be implemented in diagnostics and treatment of these diseases. By selector-based re-sequencing of the NF-kB pathway with proximity-ligation assays, we could for the first time provide a molecular basis for NF-kB activation in CLL; the prime finding being the recurrent truncating mutations of the cytoplasmic inhibitor IkBe (encoded by NFKBIE) in CLL. We could show that NFKBIE mutations were highly enriched in poor-prognostic, stereotyped subset patients, probably contributing to their very dismal prognosis, and resulted in reduced IκBε protein levels, loss of physical interaction between IκBε/p65, as revealed by proximity ligation assays, and increased p65 phosphorylation. Considering that IkBe truncating mutations were also identified in other B-cell-derived malignancies, hence alluding to a common pathobiology for NF-kB activation, components of the NF-kB signaling pathway emerge as possible targets for future targeted therapies in CLL and, possibly, also other mature B-cell lymphomas.
The DiaTools project has lead to introduction of a new class of targeted digital analysis methods for non-invasive diagnostics. These technologies will be competitive compared to NGS-based technology because of their ease of use, and scalability in number of processed samples.
Automated molecular analysis
We have developed the first integrated device for C2CA analysis of tumor DNA in solution. The microfluidic set-up simplifies the multi-step protocol making it suitable for clinical implementation. The established concepts and modules are generic and open for further applications.
Diagnostics of cells and tissues
The development of the EPHESIA technology for isolation and analysis of circulating tumour cells will have clear impact on diagnosis and treatment of several forms of cancer. The benefit of the EPHESIA system lies in the fact that the captured cells can be directly analyzed on the columns. This opens possibilities for faster and more detailed characterization of CTC compared to the systems currently in use.
We have further worked on further developments of the in situ genotyping technology that was originally developed within the FP7 READNA project. We have demonstrated somatic mutation detection in morphologically intact cells, and we have developed an in situ sequencing method that allows highly multiplexed molecular profiling in intact cells and tissue sections. We are currently exploring these sets of tools and technologies to achieve cellularly resolved molecular profiling of tumor tissue sections to help dissect different molecular profiles obtained in bulk extracts from tumors to distinct populations of cell-types and –states, and subclones of cancer cells.
The project has lead to extensive collaboration between research groups at Stockholm University and Institut Curie which will continue after completion of the DiaTools projects.
Our work on in situ mutation analysis has lead to further developments within the EU FP7 CareMore project. In this project the technology developed in the DiaTools project will be further optimized for mutation analysis in morphologically intact CTCs. For example, to improve sensitivity preamplification and direct RNA detection are explored. We are also exploring in situ mutation analysis as a pre-amplification step in a CTC monitoring biosensor system developed in the FP7 CANDO project.
With regards to magnetic micro- and nanoparticles developed in DiaTools, rather complex techniques of multiple swelling and polymerization was elaborated and substantially improved. Great advantage of the newly developed microspheres consists in their cheap production compared with the commercial products. Their use in the molecular diagnostics may provide tailored and cost-efficient treatment for infectious disease and cancer. The particles enable target-specific isolation of proteins from complex clinical samples and removal of reagents, inhibitors and contaminants between the reaction steps. This may be beneficial for work in any clinical or biochemical laboratory. Moreover, synthesis of surface modified magnetic nanoparticles was mastered.
Thanks to the DiaTools, cooperation with the Uppsala University, the Stockholm University and the Curie Institute was established and since October 2014, the Institute of Macromolecular Chemistry participates in the new COST project “Multifunctional nanoparticles for magnetic hyperthermia and indirect radiation therapy (RADIOMAG)”.
Potential impact of the CyToF technology
The impact of the accomplishments from the Nolan lab at Stanford University resulting from the DiaTools project allowed us to expand our capabilities for measuring some new attributes of single cells beyond the scope of protein expression and protein activation states.
For example, measurements of gene expression are a fundamental tool to understand how genetic networks coordinately function in normal cells and tissues and how they malfunction in disease. Therefore including measurements of messenger RNA molecules and proteins simultaneously will greatly affect our conceptual understanding of diverse biological processes with broad implications for both basic and clinical research.
Additionally, including subcellular localization of proteins, nucleotides in our assays enables single-cell, subcellular visual and spatial information to be obtained from a non-visualizable platform with throughput and detection limit significantly improved relative to microscopy. This new attribute shows the relationships between signaling pathway activation and nuclear entry and transcriptional activation again with significant implications to increase our understanding of biological processes in health and disease. In its current form can detect proteins in the nucleus and cytoplasm but are currently developing additional reagents with which to look at localization to other intracellular organelles.
Given the complexity of the large datasets, measurements of upwards of 40 parameters per single cell, new computational tools are required to unravel the biology addressed by mass cytometry assays. Our approach is to develop multiple complementary approaches that will recapitulate know features while simultaneously uncovering new biology. The DiaTools project hasve enabled us to make a significant contribution to biomedical research with opportunities for continuation on multiple levels.
Commercial impact Olink:
The DiaTools project has given Olink a great opportunity to improve our main product line, Proseek multiplex. First, we designed and evaluated a new inter-plate control sample (IPC) with new normalization procedure that reduced the inter-plate and batch-to-batch variation. Furthermore, we have improved the qPCR detection protocol in multiplex PEA, which improved the intra-plate precision. The IPC sample and new qPCR detection protocol were further developed by our production department outside of the DiaTools project and finally introduced to the Proseek multiplex products and launched in October 2013 and July 2014, respectively.
The DiaTools project has helped Olink Bioscience to develop a prototype protocol for combined multiplexed in situ proximity ligation (PLA) and padlock probing (PLP) with either readout by flow cytometry or microscopy. This protocol has been transferred to another FP7 project that Olink is coordinating (CareMore), in which it is being further developed for characterization of ER and pHer2 protein expression levels by PLA, and mutation status of a relevant transcription factor by PLP in circulating tumor cells from metastatic breast cancer patients. The combined PLA/PLP protocol also holds great potential for future development into an in vitro diagnostic reagent for molecular pathology.
The single-cell protocol developed within the scope of the DiaTools project makes the multiplex PEA protein quantification of a suitable method for characterization of rare cells, such as studied on circulating tumor cells and other.
Commercial impact Qlinea:
While the cell imaging capabilities of Q-lineas modified reader has no straightforward commercialization potential, both reader concepts have been taken further. The ASMD amplification/readout system first designed has been taken to a TRL7 level, and will be further developed within the IPODS project, sponsored by the EDA.
The microarray-based amplification and detection system has clear clinical diagnostics potential, and is being developed to its full potential within Q-linea. Furthermore Q-linea has developed a significantly better performing single use microarray consumable compared to current state of the art. This will likely add to the commercial impact of future Q-linea ivd system development. The microarray consumable also enables it to be a product in its own right to be sold to other third parties.
Initially targeted for the defence market, Q-linea´s evolved ASMD amplification/readout system has gathered interest in the homeland security setting, and has been successfully evaluated in a relevant setting, biothreat agent dissemination in a subway station.
The high-multiplex system, on the other hand, has raised interest from an infectious agent ID point of view, with focus on microbial ID and antibiotics susceptibility assessment. The characteristics of this platform potentially sets it apart from competing technologies within these fields, providing a new tool with substantially shorter lead times to pathogen ID and resistance profiling. Further development of this platform is ongoing.
The sample preparation technology developed, capable of processing large volumes ~5ml of whole blood, is judged to be better than current state of the art. This can be produced as a standalone product but also enables Q-linea to enter blood testing ivd markets
Commercial impact Fluigent:
Two products which have been partially developed through the DiaTools project have been commercialized by Fluigent: the Easy Switch Solutions and the Flow-Rate Platform, both hardware and software. The ESS platform enables the user to inject sequentially different fluids in a microfluidic system. Thanks to the Flow-Rate Platform, the user can monitor the flow-rate in its microfluidic environment. Within the project, needs for individual flow control units have emerged as well as pre-integrated flow control platform. Specific studies will be part of the new developments.
List of Websites:
http://www.diatools.org/
Project Coordinator:
Professor Ulf Landegren
Department of Immunology, Genetics and Pathology and SciLifeLab, Uppsala University
Box 815, SE-751 08 Uppsala, SWEDEN
E-mail: ulf.landegren@igp.uu.se
Tel: +46 18 4714910
Project manager:
Associate Professor Erik Ullerås
Department of Immunology, Genetics and Pathology
Uppsala University
Box 815
SE751 08 UPPSALA, Sweden
+46 18 471 44 93
+46 704 30 74 85
erik.ulleras@igp.uu.se