Tumor metastases account for approximately 90% of cancer-related deaths [1, 2]. Each malignancy has preferential organs and/or tissues for the metastatic dissemination and consequent colonization (Table 1) [3, 4]. The mechanism of the various steps that are involved in the metastatic process is still mostly unclear [5-7]. Overall, the multistep metastatic process involves the following stages: detachment of malignant cells from the primary tumor; extravasation of some cancer cells in the bloodstream; various factors promote the survival of a number of circulating tumor cells (CTCs) during the migration in the vascular system; another extravasation of malignant cells to distant organs and/or tissues; colonization and proliferation of disseminated tumor cells (DTCs) within the affected distant organs (Fig 1) [8]. In this scenario, CTCs may be considered an important factor that takes part in the dissemination of metastases in patients with cancer. The detachment of malignant cells from the primary tumor requires a gradual epithelial-mesenchymal transition (EMT) (Fig 2, panels A and B) [9]. Epithelial cells are polarized and interact with a basement membrane. EMT consists of various biochemical modifications to confer a mesenchymal-like cell phenotype to transformed epithelial cells (Fig 2, panels A and B). As the transition to the mesenchymal-like phenotype proceeds, modified and transformed epithelial cells tend to acquire enhanced invasiveness and migratory ability. At this stage, malignant cells detach from the basement membrane and can move away from the epithelial layer and become CTCs (Fig 2, panel A).

Current technologies allow for the isolation of CTCs from the peripheral blood of patients, which can be subsequently analyzed at the genetic and molecular level for the detection of tumor-associated genes and/or antigens [10, 11]. Protocols for the analysis of tumor associated markers in clinical samples have been described elsewhere. The isolation of CTCs from peripheral blood is essentially based on epithelial markers and other factors, which will be described in the next section (see “Enrichment, isolation and characterization of CTCs”).

[enlarge]

Figure 1. A vascularized primary tumor. Circulating tumor cells (CTCs) detach from the primary tumor, enter the blood vessel, migrate through the vascular system. Another extravasation CTCs to distant organs may take place, followed by colonization and proliferation of disseminated tumor cells (DTCs) within the affected distant organs. This leads to the constitution of a metastatic site. The formation of metastatic sites within certain organs and/or tissues are tumor specific (Table 1).

The detection of CTCs in the peripheral blood of patients with cancer is a relatively easy system that does not involve an invasive intervention, such as biopsies from the patients [12]. Protocols have been clinically validated for the prognosis/staging of a variety of solid tumors, such as malignancies of the breast, prostate, colon, pancreas, lung, melanoma, glioblastoma or multiple myeloma [13]. Clinical studies are currently in progress to develop protocols for the monitoring of CTCs in patients with different types of sarcomas. For these types of malignancies, a major challenge arises from the lack of specific markers on the membrane of sarcoma cells. Circulating sarcoma cells are currently isolated either by using so-called non-specific parameters, or by detecting specific chromosomal translocations, which take place in transformed cells and are not present in normal tissues and/or cells. Such chromosomal translocations may result in the expression of fusion genes that can be detected by reverse transcriptase polymerase chain reaction (RT-PCR).

[enlarge]

Figure 2. Panel A. Epithelial cells are polarized and interact with the basement membrane. The malignant cell transformation is a multistep process, which involves a number of mutations. Malignant cells retain epithelial cell markers and remain attached to the basement membrane and/or to other cells of the primary tumor. The epithelial mesenchymal transition (EMT) is a necessary step for the detachment and migration of cancer cells, which begin to lose epithelial markers and acquire a mesenchymal phenotype. The basement membrane is represented by a solid blue line. Abbreviations: CTCs = circulating tumor cells; DTCs = disseminated tumor cells. Panel B. This panel shows the gradual loss of the epithelial-like cell phenotype in malignant cells, which gradually gain a mesenchymal-like cell phenotype. In an advanced EMT, epithelial markers such as EpCAM and/or CKs are poorly expressed, or completely missing.

CTCs can be observed in the peripheral blood either as single cells, or in small clusters. CTCs clusters are more sporadic than single CTCs. Typically, the percentage of CTCs clusters is in the range of 2 to 5% of the total number of CTCs. CTCs clusters comprise more than two or three malignant cells and exhibit strong cell-to-cell adhesion. Some clusters may contain up to 100 cells. In addition to malignant cells, clusters include other cell types, such as cells of the immune compartment, platelets and tumor-associated fibroblasts. CTCs clusters are also termed microemboli. The formation of CTCs clusters, or microemboli, provides a microenvironment that protects to a certain extent circulating malignant cells from the action of the immune system, shear stress and anoikis, which is a type of apoptosis that may derive from the absence of cell adhesion [14]. Furthermore, circulating tumor microemboli may facilitate the colonization of the organ by DTCs. However, in spite of the mentioned protective functions of cell-to-cell adhesion, CTCs clusters exhibit shorter life span in the bloodstream than single CTCs, as shown by in vivo flow cytometry conducted on breast cancer patient-derived samples, which were subsequently injected into the tail vein of immunodeficient mice. In this animal model, the CTCs clusters clearance rates were three times faster than single CTCs. Probably, this finding can be attributed to the entrapment of CTCs clusters inside blood capillaries.

As revealed by clinical studies in patients with breast cancer and in animal models, CTCs clusters are highly metastatic [15-19]. These findings were also confirmed in patients with other tumors, such as advanced colorectal cancer [20, 21], pancreatic adenocarcinoma [22], melanoma [23], non-small cell lung cancer (NSCLC) [24, 25], small cell lung cancer [26] and prostate cancer [27]. Clinical studies are in progress to assess the diagnostic relevance of CTCs clusters, or microemboli, in other types of malignancies [28].

This review provides lists of reagents, along with the clinical protocols that can be used to isolate and monitor CTCs in various malignancies, in order to follow up the clinical course of the disease and to assess the response to the therapy.

The isolation of CTCs simply requires a sample of peripheral blood from oncological patients. CTCs are very rare and have a short half-life in the bloodstream. Usually, between 1 to 10 CTCs can be detected per milliliter of peripheral blood of patients with metastatic stage cancer [29, 30]. The ratio of CTCs and peripheral blood mononuclear cells may range between 1 to 105 and 1 to 106 [31]. For this reason, peripheral blood-derived CTCs fractions require an enrichment before they can be analyzed. Methods for enrichment and analysis are based on physical and/or biological properties of CTCs, such as size, deformability, density, polarity and electrical charge, epithelial cell adhesion molecule (EpCAM), cytokeratins (CKs) and tumor-associated markers expression [32]. CTC-derived cell lines can also be used [15, 33]. For example, Gkountela S et al obtained CTC-derived cell lines BR16, Brx50, Brx07, and Brx68 from other investigators to study CTC clusters [15].

EpCAM is a transmembrane glycoprotein that is expressed in epithelial and adenomatous cells and is overexpressed in a variety of carcinomas like prostate cancer [34, 35], whereas CKs are intermediate filaments that are present in the cytoplasm of epithelial cells [36, 37]. Epithelial markers render CTCs distinguishable from blood cells. However, epithelial markers expression decreases in CTCs with the progression of the epithelial mesenchymal transition (EMT) (Fig 2, panels A and B). On these grounds, some CTCs may not be detectable in various assays.

The main characteristics and drawbacks of detection and isolation systems for the analysis of CTCs are summarized in Table 2 and 3, respectively.

This method for CTCs enrichment from peripheral blood utilizes an inert polysucrose termed Ficoll (GE Healthcare Bio-Sciences, Pittsburg, PA; BD Biosciences, San Jose, CA), which was originally produced for the isolation of mononuclear blood cells from whole blood [38, 39]. The Ficoll protocol was then optimized for the separation of CTCs from blood cells and relies on differential migration properties that derive from diverse cell-type dependent buoyant densities [40]. A density gradient centrifugation system named OncoQuick is commercially available for the Ficoll-based separation of CTCs from peripheral blood [41] (Greiner Bio One, Frickenhausen, Germany). A flaw of the density-based cell separation method is the possible loss of some CTCs.

This type of CTCs enrichment is based on size and deformability of epithelial cancer cells [44-46]. A 10-cm silicon wafer device was produced with standard micromachining techniques. This device consists of four segments of microfluidic channels placed on a silicon wafer, which are attached by adjacent little tanks that contain a fluid inlet and outlet [44].

Another microdevice contains multiple arrays of crescent-shaped isolation wells for the isolation of epithelial cancer cells from blood cells. Each trap has a gap of 5μm, which allows for the exit of blood components, including larger white blood cells that are more deformable than malignant cells. Prefilters with 20μm gaps prevent the clogging of the microdevice, by removing larger clumps and/or debris prior to the filtration. Microdevices are engineered to enhance hydrodynamic efficiency and to facilitate the optimal retrieval of CTCs.

A microfilter made of a parylene membrane was produced for the isolation, electrolysis and genomic analysis of CTCs. Parylene is provided by Specialty Coating Systems (Indianapolis, IN), or by Para-Coat Technologies (Johnstown, PA).

Size-based enrichment systems are limited by the significant variations in cell size among cancer cell populations, because of tumor heterogeneity [47]. This may result in the loss of some CTCs.

A negative enrichment technique consists of the depletion of most of the leukocytes and erythrocytes by means of RosetteSep (StemCell Technologies, Vancouver, BC, Canada), which utilizes antibodies that target hematopoietic cells of human whole blood and crosslink them to multiple erythrocytes. The crosslink, in turn, leads to the formation of immunorosettes. Centrifugation over the Ficoll-Paque buoyant density medium allows for the precipitation of immunorosettes and unbound red blood cells, while CTCs fractions can be recovered from the medium. This protocol is limited by the possible loss of some CTCs.

Ferromagnetic beads coated with anti-CD45 antibodies can be utilized as a further purification step of CTCs fractions derived from density gradient centrifugation procedures [48]. CD45 is a protein tyrosine phosphatase, receptor type, C (PTPRC) that is expressed on the membrane of hematopoietic cells, with the exception of erythrocytes and plasma cells [49]. CD45 is also missing in epithelial cells. Thus, anti-CD45 immunomagnetic beads bind only blood cells, which can be subsequently removed with a magnetic field. The capture efficiency of this protocol ranges from 52% to 88%. Anti-CD45 immunomagnetic beads are provided by Myltenyi Biotec (San Diego, CA) and Thermo Fisher Scientific (Monza MB, Italy). This technique is limited by the possible loss of some CTCs and by the presence of impurities, as not all CD45-positive cells are removed during the procedure.

In this negative enrichment cell sorting system, cells are labeled with magnetic nanoparticles. Magnetophoretic mobility is referred to the differential migration of magnetically labelled cells within a magnetic field [50]. A Quadrupole Magnetic Cell Sorter (QMS) was produced for the rapid enrichment and separation of CTCs from hematopoietic cells (Ikotech, Greenville, IN). QMS can sort 10 million cells per second, with a 99% depletion efficiency of CD34(+) and CD45(+) peripheral blood cells. This system allows only for the depletion of unwanted cells. After the enrichment step, other detection techniques must be applied for the isolation of pure CTCs populations.

This method is also known as magnetic activated cell sorting (MACS). Ferromagnetic beads coated with anti-EpCAM antibodies can be used to isolate CTCs, following a density gradient centrifugation step. EpCAM-positive CTCs can be automatically isolated with the MagSweeper, which utilizes a magnetic arm to harvest the ferromagnetic beads [51, 52]. The isolation of CTCs based on MACS depends on the single target molecule that is used in the protocol. In addition, some CTCs with low EpCAM expression and/or CTCs that have undergone advanced EMT might be lost.

In microfluidic devices, the blood flows on so-called CTC-chips that are made of equilateral triangle structures, which contain an assortment of anti-EpCAM antibody-coated microposts [53] or chips with other configurations such as graphene oxide chips [42]. Interestingly, microfluidific devices exhibit an efficacy of capture rate that is comparable among cancer cell lines with different levels of EpCAM expression [53]. For instance, in the first studies a group of investigators used a panel of human cancer cell lines with variable amounts of EpCAM expression, such as NSCLC NCI-H1650 and breast cancer SKBr-3 cell lines (with more than 500,000 EpCAM molecules per cell), prostate cancer PC3-9 cell line (with 50,000 EpCAM molecules per cell) and bladder cancer T-24 cell line (with roughly 2,000 EpCAM molecules per cell) [54]. Most likely, this finding is due to the increased interactions between cells and the substrate of the CTC-chip. However, the enhanced sensitivity for the capture of cell lines with low EpCAM expression might not be sufficient for the selection of biopsy-derived CTCs that have undergone advanced EMT.

A more advanced microfluidic system, CTC-iChip, combines size-based separation of nucleated cells (leukocytes and CTCs) from red blood cells, platelets and plasma, and depletion of magnetically labeled white blood cels to enrich CTCs [55]. Ebright RY et al, for exmaple, further stained the enriched products with EpCAM, cadherin 11, and HER2 antibodies for CTCs and with CD45, CD14 and CD16 antibodies for white blood cells to obtain individual CTC under a fluorescent microscope with an Eppendorf Transfer-Man NK 2 micromanipulator [56]. W Huang et al devised a similar size- and magnetics-based micro-aperture platform to capture both cells and plasma proteins in separate chambers [57]. M Radovich et al applied such a micro-aperture process to measure circulating tumor cells in order to evaluate the prognostic value of circulating tumor cells in breast cancer patients in a clinical trial [58].

This protocol was developed by Cytelligen, San Diego, CA. For the description of the protocol, see Table 3 and the paragraph “Integrating antigen-independent subtraction enrichment and immunostaining-FISH (SE-iFISH)”, under “Protocols for CTCs detection”.

Quantitative RT-PCR is the preferred nucleic acid-based method for the analysis of CTCs associated markers [59, 60]. The expression of messenger RNA implies that most likely cells were still alive at the time of the isolation and analysis. In fact, RNA is rapidly destroyed when cells die. In addition, studies showed that RT-PCR is more sensitive than immunohistochemistry for the probing of cell specific markers [61, 62]. Following CTCs isolation, quantitative RT-PCR was utilized for the detection of MUC-1 in breast cancer patients [63], alpha-fetoprotein (AFP) in primary liver cancer patients [64], human telomerase reverse transcriptase (hTERT) in patients with gastric tumor [65], cytokeratin (CK) 19 and CK 20 in colorectal cancer patients [66] and prostate-specific antigen (PSA) in patients with prostate cancer (Table 2) [67-69]. RT-PCR may be susceptible to provide false positives, because of sample contamination (Table 2) [70]. There is also the possibility that the target gene is expressed in some normal cells [70]. Another disadvantage of RT-PCR is that analyzed cells cannot be used for any other type of study.

This technique utilizes fluorescent labeled DNA probes to detect specific DNA sequences within certain chromosomes [71]. FISH analysis is very precise and can detect a variety of genetic abnormalities that are related to human illnesses [71, 72]. However, the experimental procedure requires highly trained personnel, is labor intensive and sometimes may not provide clear results. In this respect, an alternative optical imaging system named Ikoniscope® (Ikonisis, New Haven, CT) was developed to detect rare cell populations [73-75]. Interestingly, the Ikoniscope® imaging system can detect one CTC per milliliter of peripheral blood [75]. However, analyzed cells cannot be utilized for subsequent studies, as they are no longer viable after FISH analysis.

This technique relies on an enrichment step for CTCs, in which cells retain viability and are suitable for primary tumor cell culture conditions (Table 3). The enrichment step is followed by immunostaining for epithelial markers combined with FISH analysis for chromosomal aneuploidy (Table 2) [76]. SE-iFISH technology can be supported by Metafer-iFISH automated CTC imaging system, in order to provide a platform for the characterization of each CTC subtype in the progression of the clinical course of the disease.

The subtraction enrichment procedure utilizes a non-toxic matrix with a specific density to remove red blood cells from white blood cells and CTCs after a centrifugation (Cytelligen). Hypotonic solutions are not required in this protocol, which could potentially damage some epithelial-derived cells. White blood cells are then removed with magnetic beads conjugated with anti-CD45 antibodies.

The iFISH method allows for the in situ phenotyping and karyotyping of CTCs, along with the classification of various CTCs subtypes, following the detection of tumor biomarkers expression levels.

FACS relies on the use of antibodies and is commonly utilized for the separation of a specific cell type, which can be isolated with high purity from the general cell population [77]. Forward scatter and side scatter allow for the characterization of physical properties such as cell size and internal cellular complexity, respectively [77, 78]. An electrostatic detection system leads to the accurate isolation of a determined cell fraction, based on the electric charge of that particular cell population. Cellular electric charges depend on the specific genes that are expressed in each cellular context. Up to 50,000 cells per second can be sorted with advanced flow cytometers (Abcam, Cambridge, MA; BIO-RAD, Hercules, CA; BD Sciences, San Jose, CA). FACS is a versatile technique that has a wide range of applications, as several parameters can be simultaneously analyzed [79]. FACS has a limited throughput, as cells are sorted individually. FACS technology may be affected by another shortcoming. In some cases, flow sorting conditions may be detrimental to particular types of cells and, therefore, impede subsequent studies [80].

The Parsortix Cell Separation System employs microfluidic particle separation technology to capture tumor cells based on their large size and resistance to compressiona from bodily fluids (e.g., blood, bone marrow, and ascites) [81]. The capture does not require antibodies or other cell surface affinity agents; however, the captured cells can be directly stained inside the cassettes. Gkountela S et al, in their studies of CTC clustering, captured CTC from human and mouse blood with a Parsortix GEN3D6.5 Cell Separation Cassette, and directly stained the captured CTC with EpCAM-AF488 (CellSignaling, Cat# CST5198), HER2-AF488 (#324410, BioLegend), EGFR-FITC (GeneTex, Cat# GTX11400) and CD45-BV605 (Biolegend, Cat# 304042 (anti-human); Cat# 103140 (anti-mouse)) antibodies [15].

FAST cytometry was developed to speed up the analyses of immunofluorescent-labeled cells in large volumes of peripheral blood (SRI International, Menlo Park, CA) [82, 83]. This system allows for the detection and isolation of CTCs from large volumes of peripheral blood. A laser printing technique is utilized to localize rare subpopulations of immunofluorescent-labeled cells on glass substrates. The laser printing optic has the ability to excite 300,000 cells per second and the emission spectrum is monitored on a very wide field of view. The scan rates of FAST analyses can be 500-fold faster than automated digital microscopy. The enhanced screening rate combined with the high sensitivity provides the desired performance for the detection in peripheral blood of early-stage tumors. The possibility to detect and isolate CTCs in a single step minimizes the chance of losing cells. FAST cytometry was used to monitor CTCs in the peripheral blood of patients with stage IV breast cancer.

A device termed CellSearch Circulating Tumor Cell Test is commercially available (Menarini-Silicon Biosystems, Bologna, Italy and San Diego, CA) and has been approved by the U.S. Food and Drug Administration (FDA) for clinical testing of blood samples of patients with breast cancer [16, 84, 85], colorectal cancer [86] and prostate cancer [87]. A ferrofluid linked to anti-EpCAM antibodies is utilized for the initial enrichment of EpCAM-positive cells. Subsequently, a magnetic field is applied for the isolation of the ferrofluid-EpCAM-positive cells complex. The CTCs detection is carried out with a pool of phycoerythrin (PE)-labelled anti-CK antibodies and allophycocyanin (APC)-labeled anti-CD45 antibodies, mixed with a nuclear dye (4′,6-diamidino-2-phenylindole (DAPI)) and a permeabilization buffer for the entry of anti-cytokeratin antibodies into the cells [88, 89]. Finally, immunofluorescence is carried out for the enumeration of EpCAM-positive CK-positive CD45-negative CTCs. All the above-mentioned components and reagents necessary for the CTCs analysis are included in the CellSearch Epithelial Cell kit (Veridex, Warren, NJ). The CellSearch system can only detect EpCAM-positive and CK-positive CTCs. This system is not suitable for CTCs that either do not express epithelial phenotypes, or have lost EpCAM and CK expression, due to advanced EMT.

CellSearch can also be augmented with a second approach such as the DEPArray technology to increase the chance of detecting CTCs [43].

ISET enriches epithelial cancer cells with a blood filtration, which takes place through a membrane that has pores with a diameter of 8 μm. Larger epithelial cancer cells remain on the filter and, therefore, can be stained either for immunocytochemistry, or cytomorphological analysis [90]. ISET has the ability to detect a single CTC from 1 ml of peripheral blood. ISET is commercially available (RareCells, Paris, France) and has been used for the detection of CTCs in blood samples of patients with tumors of the breast, lung and prostate. A comparative analysis revealed some discrepancies in the enumeration of CTCs between ISET and CellSearch. These discrepancies may be attributable to the heterogeneity of tumor cell size and/or to differential expression of EpCAM among cancer cells. The latter is a limitation of CellSearch, whereas the former is a drawback of the isolation by cell size.

This test is commercially available (AdnaGen AG, Langenhagen, Germany) and has been used for the analysis of CTCs in blood samples derived from patients with cancers of the breast, colon and prostate [91-93]. CTCs enrichment is conducted with magnetic separation of EpCAM-positive cells, whereas the detection is carried out by RT-PCR for the identification of tumor-associated transcripts. A clinical study conducted on patients with metastatic breast cancer used RT-PCR for the analysis of the following transcripts: GA733-2, MUC-1, and HER2. Both AdnaTest and CellSearch exhibited concordant data in terms of CTCs enumeration and sensitivity. Both systems were able to detect 2 CTCs in 7.5 ml of patient-derived peripheral blood. However, cells are no longer viable after the AdnaTest and, therefore, they cannot be used for other studies. In addition, RT-PCR techniques may provide false positives, because of nucleic acid-related contaminations.

EPISPOT detects specific secreted proteins and stands for epithelial immunospot, which derives from enzyme-linked immuno-assays [94]. An initial enrichment step is carried out by immunomagnetic depletion of CD45-positive cells, which is then followed by immunomagnetic positive enrichment by selecting CXCR4-positive cells. At this stage, only viable cells can secrete proteins for EPISPOT assay. MUC1 secretion was detected in 100% of patients with metastatic breast cancer, whereas PSA was observed in 83.3% of prostate cancer patients [95, 96]. This system is restricted to the analysis of CXCR4-positive CTCs and, therefore, CXCR4-negative CTCs are lost.

The Ariol system is commercially available (Genetix USA Inc, San Jose, CA, USA) and consists of an automated cell image capture for the analysis of CTCs placed on glass slides [97]. The protocol was set up for the analysis of blood samples of patients with breast cancer. For the enrichment step, peripheral blood is initially treated with a lysis buffer to remove erythrocytes (CTC Enrichment and Detection Kit, Genetix, New Milton, UK). Cells were then centrifuged at 700 × g for 10 minutes at room temperature and the cell pellet was gently resuspended with 0.5 ml dilution buffer (Carcinoma Cell Enrichment and Detection Kit, Miltenyi Biotec, Bergisch Gladbach, Germany). Cell blocking, permeation and fixation were carried out following the manufacturer’s instructions (Carcinoma Cell Enrichment and Detection Kit, Miltenyi Biotec, Bergisch Gladbach, Germany). CTCs enrichment was carried out with anti-CK antibodies either alone, or in combination with anti-EpCAM antibodies, following the instructions of the CTC Enrichment and Detection Kit (Genetix, New Milton, UK). As further controls, labeled anti-CD45 antibodies and DAPI were included in the assay [97]. A Leica DM6000 with software for clinical breast cancer panel analysis was utilized for image capture (Leica Microsystems, Wetzlar, Germany). The Ariol system allows for the detection of CK-positive EpCAM-positive and CK-positive EpCAM-negative CTCs. However, cells are not viable after the Ariol system assay.