Molecular and Functional Characterization of Circulating Tumor Cells: From Discovery to Clinical Application

Abstract

BACKGROUND

One of the objectives for the liquid biopsy is to become a surrogate to tissue biopsies in diagnosis of cancer as a minimally invasive method, with clinical utility in real-time follow-ups of patients. To achieve this goal, it is still necessary to achieve a better understanding of the mechanisms of cancer and the biological principles that govern its behavior, particularly with regard to circulating tumor cells (CTCs).

CONTENT

The isolation, enumeration, detection, and characterization of CTCs have already proven to provide relevant clinical information about patient prognosis and treatment prediction. Moreover, CTCs can be analyzed at the genome, proteome, transcriptome, and secretome levels and can also be used for functional studies in in vitro and in vivo models. These features, taken together, have made CTCs a very valuable biosource.

SUMMARY

To further advance the field and discover new clinical applications for CTCs, several studies have been performed to learn more about these cells and better understand the biology of metastasis. In this review, we describe the recent literature on the topic of liquid biopsy with particular focus on the biology of CTCs.

Liquid biopsy has gained a prominent standing among the different diagnostics tools in cancer diagnosis. This relatively new approach originally referred to the analysis of circulating tumor cells (CTCs)³ and then was extended to the analysis of circulating cell-free DNA, cell-free microRNA, extracellular vesicles, and tumor-educated platelets, which can be found in blood, urine, cerebrospinal fluid, and other fluids in cancer patients (1). The present review will focus on the detection and molecular characterization of CTCs in blood.

The aim of liquid biopsy analysis is to provide accurate clinical and pathological information that contributes to the personalized management of patients, mainly in the field of oncology. This personalized management includes early detection of cancer, determination of prognosis, identification of predictive markers for targeted therapies, real-time monitoring of therapy efficacy, and identification of the emergence of resistance mechanisms. Within this context, CTCs can be used as biomarkers related to the metastatic cascade, and, owing to their biological functions, the analysis of these cells has revealed important key mechanisms of cancer progression and metastasis (1).

To analyze CTCs, these cells must be enriched and detected; the methodology for this has been extensively reviewed elsewhere (2). In short, the enrichment of CTCs is based on biological and physical properties that can distinguish them from other blood cells. The mainstay of this is epithelial cell adhesion molecule–based enrichment followed by detection based on cytokeratin production in patients with epithelial tumors. However, not all CTCs produce these markers owing to changes in phenotype that allow migration (3). Different markers and methods have been established as alternatives in CTCs with reduced or null production of cytokeratin and/or epithelial cell adhesion molecules (2). For this reason, an understanding of the biological processes involved in the origin, survival, and colonization of CTCs is fundamental to the further development of clinical applications. Moreover, new innovative personalized therapies against CTCs in the bloodstream may surge as a result.

The purpose of this review is to highlight recent advances in the understanding of the biology of CTCs at the molecular and proteomic level and to emphasize the use of functional studies of CTCs to unravel the mechanisms behind metastasis that will give rise to better enrichment, detection, and characterization methods, as well as new therapeutic approaches (Fig. 1).

Molecular and functional characterization of CTCs. Because repeated tissue biopsies are invasive and not always feasible, the assessment of tumor characteristics by CTC analysis could contribute to improvements in personalized medicine.

DNA studies in CTCs

Technological advancements enabling the preservation and capture of intact CTC populations (single cells or clusters) have enabled subsequent molecular studies. Characterization of CTCs provides a better insight into the genomic profiles of these cells, using a variety of different technologies such as immunofluorescence (for CTC detection) combined with array comparative genome hybridization, next-generation sequencing, and fluorescence in situ hybridization.

In-depth genomic analysis of single CTCs requires whole-genome amplifications to obtain sufficient amounts of DNA for subsequent comparative genome hybridization or next-generation sequencing analysis (1). Although the amplification methods for single cells may induce changes in the sequencing results, one of the advantages of analyzing single CTCs from a patient is the possibility to assess the inter- and intratumoral heterogeneity (ITH); it has been suggested that the ITH is the result of clonal evolution and presence of cancer stem cells (4, 5). Many studies have demonstrated ITH in CTCs by using next-generation sequencing technologies: Luca et al. observed high discordance of single-nucleotide variants between CTCs and primary tumors in patients with breast cancer (6). Furthermore, in prostate cancer, Lohr et al. reported that CTCs shared 90% of the mutations present in the primary site and shared 70% of mutations with metastatic sites (7). This study suggests that CTCs might originate from both the primary site and metastatic tumors. However, to address the problem of artifacts generated by whole-genome amplification, Heitzer et al. used comparative genome hybridization in colorectal cancer between the primary tumor, metastatic tumor, and a homogeneous control sample, by reanalysis with deep sequencing and then comparison with the CTCs' genomic profile, in this way demonstrating that CTCs reflect the whole genomic profile of the different subclones that are present in the primary tumor (8).

It has been suggested that the mechanisms for metastasis are shaped by epigenetic modifications, and that CTCs therefore must be subject to these alterations. Recently, the methylation status in CTCs has also been evaluated. Gkoutela et al. evaluated the different methylation profiles in breast cancer of single CTCs and clustered CTCs, observing that genes related to stemness and proliferation were hypomethylated in cluster CTCs (9) and, furthermore, that the formation of CTC clusters (10), as well as stemness and proliferation features, was associated with a higher metastatic potential (11). Therefore, taking together the evidence of heterogeneity, it can be suggested that genomic changes give rise to different subclones within the tumor and CTCs and that the subsequent metastasis process is, at least in part, guided by epigenetic changes. Moreover, computational models have suggested the emergence of subclones with driver mutations and the capability to form metastases appearing at early stages of the disease (5), which is in agreement with observations of “CTC-like” cells in patients with a higher risk of cancer but no evidence of tumors (12). It is important to note, however, that there are previous reports of circulating epithelial cells in benign conditions (13). More research is needed to elucidate at what point in time CTCs are first released from the primary tumor and the exact mechanisms for this release.

Most of clinical studies evaluating the CTC genome focus on alterations that can affect the efficacy of target therapies, such as mutations, rearrangements, or amplifications in EGFR⁴, KRAS, AR, HER2, ERBB2, PIK3CA, ER, BRAF, ALK, and ROS1, among others.

EGFR

Targeted therapies using tyrosine kinase inhibitors against mutated EGFR have been shown to increase the survival rate in non–small cell lung cancer (NSCLC). These mutations can be evaluated in CTCs as well; Maheswaran et al. were able to identify drug-sensitive mutations in CTCs, such as in-frame deletions in exon 19 and missense mutation in L858R and T790M; this last one is associated with therapy resistance (14). Likewise, Marchetti et al. developed a method allowing performance of NGS on CTCs after they are enumerated with the CellSearch^® system, with a sensitivity and specificity of 84% and 100%, respectively, for the detection of mutations in EGFR, as compared to the primary tumors (15). In contrast, Sundaresan et al. showed that CTCs could be isolated in only 28 of 37 patients, and among the 21 patients with both CTC and tissue genotyping, only 12 patients (57%) had concordant results (16). Furthermore, to fully include liquid biopsies in the clinical practice for detection of EGFR mutations in NSCLC, the assays must have an acceptably low rate of false-negative results. So far, circulating tumor DNA (ctDNA) is the only liquid biopsy biosource used clinically for EGFR mutation detection in blood plasma, but this approach (ctDNA) has a limited negative predictive value (17). Thus, one can envisage that a combined approach (of CTCs and ctDNA) may increase the precision of liquid biopsy assays for EGFR mutation detection.

KRAS

KRAS is part of the downstream mechanism of EGFR; mutations in KRAS block the efficacy of anti-EGFR therapies in colorectal cancer (CRC). The ITH of KRAS mutations has been demonstrated in tumor tissues and CTCs; Liu et al. reported in a meta-analysis that at early stages of CRC there is a higher rate of discrepancies in the mutation profile of KRAS between CTCs and primary tumors from the same patients (18). Mutations in KRAS codon 12 are related to a higher aggressiveness of CRC owing to KRAS-mutated CTCs' ability to escape from the anti-EGFR therapy. Therefore, tracing this mutation in CTCs might work as an early indicator for resistance to EGFR therapy. Dynamic changes in KRAS mutations in tumor clones during different disease stages can be revealed by real-time CTC analysis, which could thereby provide helpful information to guide anti-EGFR-directed therapy (18). Other driver gene mutations of CRC (APC or PIK3CA) can be found in CTCs and in the primary and metastatic tumors of the same patient. Interestingly, amplification of CDK8, a therapeutic target for CDK inhibitors, has been also reported in CTCs from CRC (8).

AR

Castration-resistant prostate cancer mutations in the AR gene are associated with resistance against hormone therapy; therefore, blockade therapy against AR can improve the survival rate of patients. These mutations have been assessed in CTCs as well (19). Moreover, in the absence of mutation, AR amplifications can facilitate the uptake of androgen hormones in patients after castration, and these alterations can be also detected in CTCs (20).

HER2

In breast cancer, resistance against HER2 target therapy can occur as a result of the mutations in PIK3CA. In CTCs of metastatic breast cancer patients, PIK3CA mutations have been found in 15.9% of patients (21). In patients with more than 5 CTCs per 7.5 mL and with an original diagnosis of HER2 negative but CTC-HER2 positive, the rate of PIK3CA mutations has been reported to be higher still, at 36.4% (22), similar to previous reports in isolated single CTCs (23). Furthermore, a concordance in the HER2 status between ctDNA and CTCs of 91% has been reported (24).

ESR

Other mutations in breast cancer, such as mutations in ESR1, emerge as acquired endocrine resistance mechanisms against endocrine agents used in ER-positive metastatic breast cancer. This mutation has been evaluated in CTCs (25); however, the detection rate in CTCs might be lower than in ctDNA during progression in metastatic breast cancer (26).

BRAF

BRAF mutations are predictors for target therapies in colorectal cancer and melanoma; however, resistance occurs in most of the patients treated (27). In CTCs, it is possible to evaluate the active changes in the mutation status of BRAF in melanoma (28) and colorectal cancer (18, 24). This ability might allow the use of CTCs to guide target therapy in these malignancies.

ALK and ROS1 REARRANGEMENTS

Genomic rearrangements are ideally detected in tissue by fluorescence in situ hybridization; therefore, in the absence of enough tissue material for a biopsy, CTCs can be used as a surrogate for this aim. In NSCLC patients, Pailler et al. were able to detect ALK and ROS1 rearrangements in CTCs (29); these rearrangements are related to sensitivity against crizotinib, and therefore, CTCs could help to identify patients at risk of early resistance to target therapy. Likewise, Tan et al. found a high concordance in the detection of ALK-rearrangements between tumor biopsies and CTCs (30). Moreover, both studies reported higher production of markers related to epithelial-mesenchymal transition in CTCs with ALK rearrangement (29, 30).

RNA studies in CTCs

The recent advances in single-cell RNA sequencing have allowed a better understanding of the transcriptomics in CTCs. ITH has been shown in CTCs transcriptomics as well; in hepatocellular carcinoma, D'Avola et al. isolated CTCs by imaging flow cytometry and detected a higher production in these cells for hepatocellular carcinoma upregulated long noncoding RNA; furthermore, an overproduction of IGF-2 was reported, which might make a suitable target for future therapies (31). MicroRNAs (miRs) are also present in CTCs; some examples are microRNAs −10b, −16, −21, −31, −200, and −210; they have functions in promoting epithelial-mesenchymal transition, a metastatic mechanism that allows migrations of cells through the loss of epithelial polarity. However, a comprehensive analysis of these subtypes of RNA in a single CTC is still challenging (32, 33).

The production of mRNA in CTCs for specific biomarkers has been used to predict the efficiency of targeted therapies. For instance, in prostate cancer, the expression of AR-V7, which is a truncated form of androgen receptors that works in a constitutive manner without the need of a ligand, is predictor for the failure of antiandrogen therapy (e.g., enzalutamide or abiraterone) in metastatic castration-resistant prostate cancer (34). Additional mechanisms of resistances related with noncanonical Wnt pathways have been demonstrated in CTCs as well (35). It is not only specific mRNA that can be assessed in CTCs: the RNA profile has also shown its potential as a clinical marker. For example, in breast cancer CTCs, the RNA profile allows to define molecular subtypes based on the production of estrogen receptors, progesterone receptors, and HER2 (36). Moreover, PIK3CA pathway mRNA is overproduced in CTCs, and others in mTOR signals have been linked to poor response to therapy (37). In melanoma, Hong et al. determined the RNA profile of CTCs and were able to correlate the number of CTCs during melanoma progression after treatment to immunotherapy, thus identifying early markers of therapy success that are difficult to identify by other means (38).

Protein Studies in CTCs

Surface biomarkers of CTCs could be key candidates for targeted therapies, as CTCs can be sampled and characterized repeatedly during therapy and represent the tumor heterogeneity.

The CellSearch® system characterizes CTCs as the subset of epithelial cell adhesion molecule-captured cells that are confirmed positive for cytokeratin (CK8, CK18, and CK19) and negative for CD45 (specific for leukocytes). Additionally, capture and characterization of CTCs are possible, based on other biomarkers like ER, HER2, and PD-L1 (39, 40).

ER

Some of the most prominent therapeutic agents for breast cancer are directed against ER, which is produced in 70%–80% of primary tumors, (41) and HER2, which is overproduced in approximately 20% of primary breast carcinoma (42). Detection of either of these markers in the primary tumor or metastatic lesions is currently used to help decide whether to submit a patient for endocrine or targeted therapy. Breast cancer patients with ER-positive primary tumors can harbor ER-negative CTCs, showing a significant heterogeneity between ER production in CTCs and primary tumor/metastatic biopsies (43). The heterogeneity at the ER protein level in CTCs and therapeutic resistance mechanisms needs further evaluation in larger prospective clinical trials (43, 44).

HER2

The presence of CTCs and overproduction of HER2 have been shown to be associated with a poor prognosis in patients with breast cancer (45). There is growing evidence that the HER2 status of the primary tumor may vary during cancer recurrence or progression and that changes can occur during treatment. Therefore, CTC phenotyping can be used to reevaluate HER2 status and potentially guide treatment decisions. The data from different clinical trials confirmed the frequent discordance in HER2 status in CTCs compared to the primary tumor (46, 47). Contradictory results in terms of progression-free survival and overall survival rate were observed regarding HER2 production in CTCs (45, 48, 49), which suggest the need for careful interpretation of the data and a higher number of patients in the clinical trials. In addition, a strong association between tumor aggressiveness and HER2-positive CTCs has been reported (50).

In 2 multi-institutional studies, HER2-positive CTCs were evaluated in patients with HER2-negative primary tumors, showing a higher survival under trastuzumab treatment. Consequently, the presence of HER2 production in CTCs allows further insights into the effects of trastuzumab therapy in the context of neoadjuvant therapy (51, 52). Jaeger et al. evaluated the HER2 status of CTCs before adjuvant chemotherapy. They reported that 57.8% of patients had HER2-positive CTCs, a discrepancy between the HER2 production of CTCs and the primary tumor, and no significant association between tumor size, histopathological grade, or hormone receptor status (40). These findings strongly suggested that the HER2 status of patients needs to be reassessed to determine the appropriate use of HER2-targeted therapy.

Paoletti et al. have proposed a multiparameter CTC endocrine therapy index that combines CTC enumeration and the production of ER, HER2, Ki67, and BCL2 in metastatic breast cancer. The clinical implications of the CTC endocrine therapy index are being evaluated in an ongoing prospective study (43).

PD-1/PD-L1

PD-1/PD-L1 checkpoint blockade immunotherapy is a promising therapeutic strategy (53). Interestingly, CTCs are more often PD-L1 positive than tissue (54). High numbers of PD-L1-positive CTCs (before treatment) are associated with poor prognosis in patients treated with PD-1 inhibitors (54), which makes it an ideal target for cancer therapy.

Mazel et al. showed PD-L1 is frequently produced on CTCs in metastatic breast cancer patients (39). Moreover, the dynamic change of PD-L1 production has been observed by Yu et al., who reported it in 35 patients with different advanced gastrointestinal tumors under PD-1 blockade therapy. It has been demonstrated that PD-L1 quantification in CTCs can be correlated with PD-1-inhibited therapies (55).

Also, immunotherapy resistance of PD-L1-positive CTCs in metastatic NSCLC patients indicated that CTCs may mirror a mechanism of immune escape (56). In another clinical study of nonmetastatic NSCLC patients, PD-L1 production in CTCs was monitored before, during, and after treatment. When patients received radiotherapy, the proportion of PD-L1-positive CTCs increased significantly. Among patients undergoing chemoradiation with concurrent carboplatin and paclitaxel, 7 out of 8 patients showed increased PD-L1 production during treatment. It has been suggested that with or without concurrent chemotherapy, PD-L1 production in CTC increases during radiation; therefore, it might be worthwhile to monitor PD-L1 production in CTCs throughout treatment (57).

AR/PSMA

In prostate cancer, CTCs showed dynamic phenotypical composition of AR-on to the AR-off in patients over the course of androgen depletion therapy; moreover, prostate-specific antigen was upregulated after AR activation, and prostate-specific membrane antigen was upregulated as well when AR was suppressed (58). Thus, prostate-specific antigen/prostate-specific membrane antigen–based measurements on CTCs appear to be surrogates for AR signaling in CTCs, and this information might help to predict the outcome of AR-based therapy (58). AR has also been identified in breast cancer CTCs as well (59).

Functional Studies in CTCs

So far it has been possible to explore some of the possible clinical uses for the detection and analysis of CTCs. This investigation has given insights into the use of these technologies in clinic decisions. However, there is still a lack of understanding of the basic biology behind CTCs that has hampered the realization of clinical applications. To better comprehend the mechanism behind the metastatic cascade and the origin of CTCs, in vivo and in vitro expansion models have been successfully applied (60). Furthermore, new approaches to characterize single CTCs by their secretion or metabolism are currently in development (1).

Short-term and long-term expansion of CTCs has been obtained from breast cancer, colorectal cancer, nonsmall cell lung cancer, small cell lung cancer, and prostate cancer, among others (60).

To expand use of CTCs, it is necessary that the methods for their capture preserve the viability of the cells, making this procedure challenging. Moreover, high numbers of CTCs are required, which are only available in a few patients with advanced cancer. Despite these issues, there have been insights into the biology of CTCs. For instance, from colorectal cancer, a CTC in vitro cell line was obtained that displayed features of an intermediate epithelial/mesenchymal phenotype, as well as stem cell properties and osteogenesis potential, suggesting that the original single CTC from which the expansion began was a cancer stem cell. Moreover, these cells formed tumors after being implanted in immunodeficient mice (61). These cells display the production of genes related to the regulation of energy metabolism, DNA repair, and stemness (11). Additionally, from the same patient, subsequent CTC cell lines were obtained in which it was demonstrated that, after the first cycles of chemotherapy, CTCs presented higher aggressiveness and stem cell features (62).

In breast cancer, the expansion of CTCs enabled the identification of a subpopulation capable of brain metastasis with a stem cell phenotype (63). Also, epigenetic changes related to cluster formation and stemness were reported in CTCs in vitro and then correlated with CTCs in cancer patients (9). Clustering was also reported to be a factor that increased the chances of metastasis and in vitro expansion of CTCs (64). CTC clusters are associated with neutrophils in blood; this association increases the metastatic potential, as has been proven in xenograft models (65).

Another alternative to performing functional studies with CTCs is the EPISPOT (Epithelial Immuno-SPOT assay), which detects the active secretion, release, or shedding of molecules by the viable cells, including for example, CK19 and HER2 in breast cancer (66, 67), CK19 only in colorectal cancer (68), and prostate-specific antigen and fibroblast growth factor 2 in prostate cancer (69). This assay has demonstrated a higher number of viable CTCs from colorectal cancer (releasing CK19) in the mesenteric veins than in peripheral veins; therefore, the liver must work as a filter for these cells in blood circulation (68). Currently, a new approach called EPIDROP (EPISPOT in a Drop) is in development; with this method, it will be possible to evaluate single CTCs in a cost-efficient manner and explore the molecular pathways of viable CTCs. The distinctive metabolomic activity of cancer cells has also been used to identify viable CTCs in lung cancer. In this method, cells are passed to microwells in a chip and then exposed to a fluorescent glucose analog; cancer cells or CTCs increase the uptake of this analog by means of the Warburg effect, allowing cells then to be visualized by fluorescent microscopy (70).

Conclusions and Perspectives

The applications of real-time liquid biopsy in the clinic have become a reality. The continuous increase in understanding of the cellular and molecular mechanisms behind each one of the biosources in this field (e.g., CTCs, ctDNA, tumor-educated platelets, and extracellular vesicles) has been fundamental to the discovery of new clinical applications. Because metastases cause the majority of cancer mortalities, CTCs are particularly useful in providing information regarding prognosis and treatment response in cancer patients.

The rise of clinical and basic studies on CTCs has shown how these cells present phenotypes and genotypes different to those of the primary tumor and, in some cases, different to the metastatic lesions, which is consistent with the concept of tumor evolution. In particular, CTC analyses provide compelling evidence of tumoral heterogeneity, which would be impossible to demonstrate in patients by current clinical diagnostic procedures. This heterogeneity is due, mostly, to phenotypic changes required for tumor cell dissemination, such as induction of epithelial-mesenchymal transition and cancer stemness characteristics, leading to selection of the fittest clone able to home to and colonize distant organs. All these different features have been demonstrated in CTCs. Furthermore, the upsurge for in vivo and in vitro models, as well as functional studies in CTCs, is beginning to yield evidence to unravel the distinct biology of cancer cells during the metastatic cascade, so far showing specific genomic and epigenetic profiles and how these profiles are related to cell-to-cell interactions that can increase the metastatic potential of CTCs (9, 65).

Despite these impressive advances and the demonstrated clinical validity for the enumeration of CTCs (71–74), assessment of CTCs for staging is only included in the 4th edition of the WHO's Classification of Tumors of the Breast; however, most of the clinical guidelines for different types of cancer still do not support their routine evaluation (75). Recent ongoing clinical trials aim to demonstrate the clinical utility of CTCs to determine treatment in breast cancer (STIC CTC study; NCT01710605) and prostate cancer (TACTIK study; NCT03101046). These early clinical trials have already shown promising results and are examples of how the enumeration of CTCs can provide relevant clinical information about the status of an individual patient.

In the future, the molecular characterization of CTCs will add another level of clinically relevant information, in particular, regarding therapeutic targets or resistance mechanisms (e.g., for HER2 or ARv7 production), which can be used to tailor therapy to the specific tumor characteristics of individual cancer patients. Moreover, a better understanding of the intrinsic mechanisms allowing CTCs to survive in the bloodstream and home to distant organs will open new avenues for innovative strategies in cancer treatment with the goal of blocking metastatic progression at a subclinical stage before incurable overt metastases appear (1).

Footnotes

Footnotes

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: K. Pantel, guest editor, Clinical Chemistry, AACC.

Consultant or Advisory Role: None declared.

Stock Ownership: None declared.

Honoraria: K. Pantel, Agena, Novartis, Roche, Sanofi, Novintum.

Research Funding: L.E. Cortés-Hernández, ELBA - Innovative Training Networks (ITN) H2020 - European Liquid Biopsies Academy project - Towards widespread clinical application of blood- based diagnostic tools. H2020-MSCA-ITN-2017 (http://elba.uni-plovdiv.bg); Z. Eslami-S, the ELBA - Innovative Training Networks (ITN) H2020 - European Liquid Biopsies Academy project - Towards widespread clinical application of blood- based diagnostic tools. H2020-MSCA-ITN- 2017; K. Pantel, CANCER-ID, an Innovative Medicines Initiative Joint Undertaking under grant agreement no. 115749, resources of which are from the European Union's Seventh Framework Program (FP7/2007–2013) (www.cancer-id.eu) and EFPIA companies' in-kind contribution; C. Alix-Panabières, CANCER-ID, an Innovative Medicines Initiative Joint Undertaking under grant agreement no. 115749, resources of which are from the European Union's Seventh Framework Program (FP7/2007–2013) (www.cancer-id.eu) and EFPIA companies' in-kind contribution. The National Institute of Cancer (INCa, http://www.e-cancer.fr).

Expert Testimony: None declared.

Patents: K. Pantel, EPO patent application No. 17157020.3 1405 “Method of detecting cancer or cancer cells,” EPO patent application No. 2016128125 A1 “Immobilization of cells or virus particles on protein structures using a microfluidic chamber”; C. Alix-Panabières, PCT/EP2017/059209.

References

Author notes