1. Cell Free DNA (cfDNA) and Circulating Tumor DNA (ctDNA)
1.1. Sample Collection and Preparation Timeline
Although the cfDNA content is higher in serum than in plasma
[13][1], plasma is the preferred source for extraction, since there is less DNA contamination due to clotting, cell lysis, and the release of cfDNA from white blood cells
[14][2]. A standard volume for a sample collection has not yet been defined, but 20 mL blood samples are usually employed for a cfDNA analysis in most isolation protocols
[14][2].
The detection and identification of ctDNA is the preferred method of the liquid biopsy for lung cancer. As the levels of ctDNA in the plasma of NSCLC patients are very low (<0.5% of total cfDNA)
[15][3], and it has only a short half-life < 3 h
[16][4], time-sensitive isolation protocols should be implemented. The procedural steps in isolating ctDNA are: (1) drawing blood, (2) centrifugation, (3) DNA extraction, and (4) DNA analysis.
1.2. Collecting Tubes
Blood samples can be collected in either standard EDTA tubes or in specialized tubes containing a cfDNA preservative (see
Table 1). EDTA tubes are widely available and considerably less inexpensive than the specialized tubes with the preservative. Disadvantageously, the samples from EDTA tubes have to be processed within 1–2 h to avoid their contamination with cfDNA from leukocyte lysis
[17][5], but the samples from the tubes containing the cfDNA preservative can be stored at room temperature for up to 14 days
[18][6]. However, for cfDNA preservative tubes, the contamination with leucocyte DNA can commence earlier so that processing should begin within three days
[14][2]. The disadvantage with these tubes is that they require special processing and must often be sent to external laboratories, incurring delays and further costs.
Even in immediately processed samples, tumor-specific somatic mutations often represent < 1% of the total cfDNA for the region of interest, and any increase in the contamination with “diluting” cfDNA from in vitro lysis may result in false-negative results
[19][7]. Consequently, a two-step centrifugation of the EDTA samples is recommended to reduce the contamination from leukocyte DNA, prior to their freezing and storage
[19][7].
Table 1.
Comparison of blood collection tubes.
1.3. Methods for DNA Analysis
In total, two established methods are available for analyzing nucleic acids (e.g., DNA and RNA): targeted polymerase chain reaction (PCR)-based, and next generation sequencing (NGS)-based assays. Table 2 highlights the differences between the methods.
Table 2. Comparison of targeted PCR-based assays and NGS for DNA analysis (adopted from [9,12,22]). Comparison of targeted PCR-based assays and NGS for DNA analysis (adopted from [10][11][12]).
1.4. Real-Time (quantitative) PCR (qPCR) Assay
A qPCR is a widely used method for DNA analysis, which also allows for the semi-quantification of samples. The cost of the method has decreased markedly, and the results are easy to interpret. There are several commercially available FDA- and EMA-approved tests for detecting, activating, or resistance-causing EGFR mutations in NSCLC. Because only a single gene or an already known genetic alteration can be examined, the sensitivity is only 70–80% (detection limit 1–5%)
[23[13][14],
24], and using a qPCR alone carries the risk of false-negative results.
1.5. Digital Droplet PCR (ddPCR) Assay
A ddPCR is more sensitive than a qPCR (detection limit 0.1–1%)
[23][13]. In ddPCR, a sample oil emulsion is fractionated into many thousands of droplets, upon which a PCR is subsequently performed in the individual compartments of a microtiter plate. This method allows for the detection of even rare events and quantification at the level of a single molecule. BEAMing (beads, emulsions, amplifications, and magnetics) has further improved the ddPCR technique using DNA templates that are bound to magnetic beads
[25][15].
1.6. Next Generation Sequencing (NGS) Assay
NGS is a high-throughput method for sequencing DNA and RNA that allows for the detection of single nucleotide polymorphisms (SNPs), and small (insertions and deletions) as well as large (insertions, deletions, amplifications, inversions, and translocations) genetic alterations. The method is very sensitive (detection limit 0.001–2%) and can detect even unknown genetic alterations in small DNA fragments
[22][12]. Whole genome sequencing (WGS) and whole exome sequencing (WES) have not been established, as the amount of ctDNA obtained with liquid biopsies might be too low. So far, only hybrid capture-based and amplicon-based (PCR capture) NGS approaches are regularly used in routine testing for cancer diagnostics.
Hybrid Capture-Based NGS
The hybrid capture-based NGS approach is not based on primary amplification, and thus allows for a more reliable quantification of copy numbers than amplicon-based NGS. Targeted DNA sequences are hybridized (“captured”) to biotinylated probes, which are bound to magnetic beads. The beads are captured by magnets, and the non-hybridized DNA is washed off. Since the ctDNA plasma concentration is very low, and sequencing does not rely on prior amplification, this method is at a greater risk of sequencing errors, including false-positive results. This explains its low specificity (approx. 60%)
[12][11].
Amplicon-Based NGS (PCR Capture)
The amplicon-based NGS approach is based on the primary PCR amplification of specific genomic regions of interest, especially hotspot genes. Before sequencing, the amplified DNA sequences (“amplicons”) are multiplexed and marked with a distinct molecular barcode for identification. As this approach uses primary PCR amplification, it is valuable for liquid biopsies with their low amounts of ctDNA
[26][16]. The risk of false-positive results is effectively reduced using molecular barcodes. While amplification biases the quantification of allele frequencies and copy number variations (CNVs), this method is useful for the detection of SNPs, indels, or known gene fusions
[12][11].
1.7. Comparison of PCR-Based and NGS-Based Methods
PCR-based methods (qPCR, ddPCR) are well established, require short turnaround times of 2–3 days, and are fairly inexpensive. However, they allow for the detection of only a limited number of genetic alterations at a time (e.g., no multiplexing across different genes) and detect neither previously unknown alterations, nor gene fusions. NGS-based methods, on the other hand, have a longer turnaround time of about 1–2 weeks
[27][17], but are able to detect known and unknown genetic variants, including CNVs, SNPs, and gene fusions, which might promote the use of diagnostic algorithms for cancer diagnostics and treatment monitoring. Lately, the NGS of
ALK-positive NSCLC has offered a more detailed characterization of fusion partners than standard pathological techniques like IHC or FISH
[28][18]. For quantification analysis, NGS and ddPCR are more suitable than the semi-quantitative qPCR. Despite the high sensitivity of NGS-based methods, a single false-positive read (with an error rate of approximately 10
−3) can impact the result
[29][19]. Hence, error-proofing techniques and algorithms need to be implemented.
2. Circulating Tumor Cells (CTCs)
Circulating tumor cells are rare in peripheral blood (1 CTC per 10
6−7 leukocytes), where they occur as single tumor cells in cell clusters, as so-called circulating tumor microemboli
[30][20], or attached to stromal cells originating from the primary tumor
[31][21]. Their plasma half-life (1–2.4 h) is short, so that they must be promptly separated from other blood cells
[32][22]. CTCs can be isolated by using antibodies to identify the expression of intracellular (DNA) or transmembrane molecules (cytokeratins, epithelial adhesion molecules [EGDR and HER3], and CD45)
[33][23]. They can also be identified according to their density or size
[12][11]. Circulating epithelial cells are present in only 46% of blood samples from stage IV NSCLC patients, and circulating cells with an epithelial phenotype can also be found in 7% of healthy controls
[34][24]. Thus, the presence of CTCs in the bloodstream alone is an inadequate marker for the tumor burden. However, some studies have shown that the baseline CTC count is a prognostic marker for tumors other than lung cancer
[35,36,37][25][26][27].
2.1. Antigen-Based CTC Isolation
In the CellSearch
® CTC Test (Menarini Silicon Biosystems Inc., Castel Maggiore, Italy) the CTCs are separated magnetically from other blood cells using anti-EpCAM antibodies conjugated with magnetic nanoparticles. The separated cells are stained for nuclear DNA with 4′,6-diamidino-2-phenylindol (DAPI), and with fluorescently labeled antibodies for cytokeratins (CK) and CD45. They are then analyzed with automated fluorescence microscopy. CTCs are defined as DAPI
+, CK
+, and CD45
−, while leukocytes are DAPI
+ and CD45
+ [33][23]. While this widely used assay is considered the gold standard for CTC detection, it has, so far, only been approved by the FDA for routine use in metastatic breast, prostate, and colorectal cancer, but not yet in lung cancer.
A different commercially available test (AdnaTest
® (Qiagen GmbH, Hilden, Germany)) uses a combination of different antibodies (e.g., anti-EpCAM, anti-MUC1, anti-HER2, and anti-EGFR) conjugated with magnetic beads for cell separation, which increases the sensitivity of the CTC detection when only the antibodies against EpCAM are used
[38][28]. However, this test has not yet been approved by the FDA for routine use in cancer.
2.2. CTC Isolation Based on Biological and Physical Characteristics
Antigen-based isolation methods have a poor sensitivity and specificity
[31][21] due to the cells’ tendency for epithelial-to-mesenchymal transition (EMT)
[39][29], with a subsequent loss of epithelial surface markers
[40][30]. Focusing on the physical and biological characteristics of the CTCs, instead of the epithelial cell surface antigens, would bypass this problem.
CTCs can also be isolated by using the physical properties of the cells, i.e., size and density. One method employs a porous membrane (pore diameter 8 μm) (e.g., ISET
® system,(Rarecells Diagnostics SAS, Paris, France)) to isolate the CTCs. Here, leukocytes are not totally eliminated. Thus, a subsequent further cell characterization with a cytomorphological or immunocytochemical analysis is required
[41,42][31][32]. Another method is to centrifuge the plasma in a density gradient, e.g., a Ficoll-Paque
® solution, to separate the mononuclear cells (including the CTCs) from other blood cells (e.g., OncoQuick
® system, (Greiner BioOne International GmbH, Kremsmuenster, Austria)). The method is inexpensive but has a high rate of contamination with leukocytes
[43,44][33][34].
Microfluidic technologies have recently evolved as a very appealing approach to CTC isolation
[45][35]. Exemplarily, one such “CTC-chip” uses anti-EpCAM-coated microposts under precisely controlled laminar flow conditions to capture the CTCs
[46][36], while the vortex technology uses microscale vortices to isolate the CTCs based on their physical characteristics, such as size or compressibility
[12,47,48][11][37][38]. These approaches are antigen-independent, resulting in a high sensitivity (a low risk of false-negative results due to EMT) and high purity of the CTC isolates
[12][11].
2.3. Interpretation of CTCs
Apart from using the common morphological and biological features of malignant cells (e.g., nucleus size and chromatin structure) to discriminate the CTCs from other blood cells, immunocytochemistry can be used for CTC identification. Moreover, genomic analyses, such as DNA analysis, as well as RNA analysis (qPCR or RNA sequencing), can be performed on CTCs, and may reflect the tumor heterogeneity
[33][23]. So far, proteomics using mass spectrometry and immunoblotting have not been implemented in CTC analysis.
3. miRNA
MicroRNAs (miRNAs) are noncoding RNAs involved in the posttranscriptional regulation of gene expression that can act as tumor suppressors or oncogenes. Cell-free miRNA can be detected in various body fluids. It is passively released during cell lysis (e.g., necrosis and apoptosis) and actively secreted by cells for intercellular communication
[49][39]. miRNAs are usually packed into EVs (see below), or coupled with Argonaute2 (Ago2) protein
[50][40] or high-density lipoprotein (HDL)
[51][41], and dispersed into the extracellular environment. A genomic analysis can be performed using the same methods as those for cfDNA (see previous section).
4. Extracellular Vesicles (EVs)
EVs are small membrane particles that are shed by all living cells and mediate intercellular communication by carrying proteins, lipids, and nucleic acids (e.g., DNA, RNA) from the secreting to the surrounding cells. A total of two distinct EV populations have so far been described: small EVs (sEVs, also called exosomes) with a diameter between 50–150 nm, and larger microvesicles (intermediate size vesicles, IEVs) with a diameter between 100–1000 nm
[52][42]. sEVs are formed inside the cell through the inward budding of endosomal membranes, while IEVs bud directly from the plasma membrane
[53][43]. Numerous studies have demonstrated that tumor-derived EVs are crucial in establishing a favorable tumor microenvironment, and thus pave the way for metastatic spread
[54][44]. The large amount of EVs shed by tumor cells into the blood and other body fluids has opened up a new perspective on using EVs as cancer biomarkers in liquid biopsies, particularly in lung cancer. Due to their small size, the isolation and analysis of sEVs is time-consuming, and thus not suitable for clinical routine. In contrast, the extraction and analysis of IEVs with flow cytometry is less demanding and time-consuming
[52][42]. Although EVs can be isolated from body fluids
[55,56][45][46], an internationally standardized isolation method, which would be necessary for their clinical use, is lacking. Inherent to every method is the risk of co-isolating the other subtypes of EVs or lipoparticles, and the separation of different EV subtypes remains difficult.