What do we need to obtain high quality circulating tumor DNA (ctDNA) for routine diagnostic test in oncology? – Considerations on pre-analytical aspects by the IFCC workgroup cfDNA

The analysis of circulating cell free DNA is an important tool for the analysis of tumor resistance, tumor heterogeneity, detection of minimal residual disease and detection of allograft rejection in kidney or heart transplant patients. The proper use of this technique is important, and starts with considering pre-analytic aspects. The current paper addresses some important technical considerations to ensure the proper and harmonized use of cfDNA techniques


Introduction
Circulating cell-free DNA (cfDNA) was identified more than 50 years ago; it is derived from both normal and diseased cells and circulates in the blood.There are different mechanisms for DNA release from cells, such as apoptosis, necrosis and active secretion.Necrosis leads to the release of large DNA fragments (>10,000 bp) whereas apoptosis leads to smaller fragments (180-200 bp) [1][2][3], this size corresponding to the length of DNA wrapped around a single nucleosome.cfDNA -at lengths corresponding to two or three nucleosomes (i.e.340 and 510 bp) -may appear at lower quantities.Infections and exercise may also increase the release of cfDNA.The route of elimination of cfDNA is not clear.In addition to the kidney, also contributions of the liver and spleen have been discussed.The half-life of ctDNA is of relatively short, typically in the range of 90-120 min [4].cfDNA refers to cell-free DNA derived from blood, which includes DNA released from all cells, including normal cells and cells involved with pathologic processes (e.g.inflammation, neoplasia), while circulating tumor DNA (abbreviated ctDNA) refers to the fraction of tumor DNA in the cfDNA pool.ctDNA may contain a variety of genomic alterations (e.g. point mutations, chromosomal rearrangements, copy number variations) and the vast majority of ctDNA is believed not to be derived from circulating tumor cells (CTCs) but released apoptotic and/or necrotic cancer cells in the solid tumor [4].In addition to patients with cancer, who have elevated levels of cfDNA as compared to cancer-free subjects, also fetal-derived DNA is detectable in pregnant women and can be used for diagnostic purposes [5].
Due to the presence of tumor-specific variations in all cancers, ctDNA can serve as an extremely specific blood-based biomarker [6].In theory, this "liquid biopsy" can provide qualitative and quantitative information on a real-time basis, by the assessment of molecular tumor genotype and existing tumor burden [6,7].Historically, the major limitation for ctDNA as a biomarker has been related to a low detection rate; however, current and developing techniques, including digital polymerase chain reaction (dPCR) and targeted next-generation sequencing (NGS) have improved the detection sensitivity dramatically [6].Both the detectable fraction of mutated DNA over wild type is of importance, as well as the tumor fraction.The recent developments in massively parallel sequencing and digital genomic techniques support the evolving clinical validity of ctDNA as a liquid biopsy in human cancer [8].ctDNA detected in plasma can be used to non-invasively analyze tumor genomes, particularly if tissue biopsy is not feasible, and quantify tumor burden [7,8].In detail, the applications for ctDNA in plasma include identifying actionable genomic alterations.ctDNA may provide a noninvasive method to monitor treatment responses and unravel therapeutic resistance [9].Additionally, ctDNA may potentially detect disease progression before clinical and radiological confirmation and with enhanced sensitivity with respect to other biomarkers.Importantly, ctDNA measurements may be useful in monitoring treatment response with checkpoint inhibitors [10].
ctDNA may also be used to characterize tumor heterogeneity [11] and metastasis-specific mutations providing information to adapt the therapeutic management of patients [8,11].Pre-analytical variables as sample storage and DNA extraction methods can influence the quality and quantity of isolated DNA and affect molecular test performances; therefore, the adherence to standardized procedures is crucial to assure validity of test results [12].Here, we address some practical aspects to ensure high quality ctDNA for molecular analyses.

Technical considerations
For all its promises, ctDNA presents considerable challenges for laboratories.The amount of cfDNA is limited: usual estimates being in the range of 5-10 ng per ml plasma, which correspond to 1,500 to 3,000 copies of the haploid human genome [13].Furthermore, most of this DNA originates from normal cells and ctDNA represents only a minor fraction of it, possibly as little as 0.1% of the total in certain cases, if not less [14].Thus, in practice, a laboratory seeking a particular mutation may choose to attempt to reliably detect as few as 1 to 3 mutated copies per ml of plasma.For these reasons, rigorous collection and processing and the participation in external proficiency tests [15] is strongly recommended.

Blood drawing
The typical amount of blood drawn for many applications is 10 to 18 ml.Hemolysis is a major issue of concern; thus careful drawing must be applied in order to avoid it and tubes should be inspected for hemolysis after centrifugation.cfDNA in blood may be diluted by DNA released from nucleated blood cells following blood sampling, which means that plasma separation should be performed shortly after blood sampling.

Plasma or serum
Plasma remains the matrix of choice, since the higher cfDNA concentrations observed in serum depend on contamination by genomic DNA released by white blood cells during the clotting process [16,17].

Initial work up
If the blood is collected in standard K2EDTA tubes (i.e., BD Vacu-tainer® Blood Collection Tubes), sample is ideally processed within 2-3 h from collection; storage at room temperature is accepted (better at 4 • C) and transport through pneumatic tubes is also acceptable [18].EDTA appears to be the anticoagulant of choice since it blocks DNA degradation by chelating bivalent cations.Plasma should be separated by a first centrifugation not exceeding 1,900g for 10 min.Plasma is then transferred to 1.5 ml Eppendorf tubes and centrifuged again at 16,000g.Timely centrifugation is essential to prevent lysis of white blood cells, thereby increasing the amount on non-tumor derived DNA.Samples need to be immediately processed for cfDNA extraction or to be frozen at − 20 • C or − 80 • C for extended periods of time.Freezers should be equipped with temperature control system and acoustic alarm.Frozen samples can be thawed at RT and should then centrifuged at 16,000g for 5 min to remove cryoprecipitates.

Blood collection tubes.
In addition to K2EDTA tubes, other devices are available that contain preservative agents to minimize cell lysis and allow shipping at RT for periods up to 7 days before centrifugation and separation of plasma.Indeed, LBgard TM Blood Tube (Biomatrica), Streck cfDNA BCT® (Streck), PAXgene Blood ccfDNA Tubes (Qiagen), CellSave Preservative Tubes (CellSearch) and Cell-Free DNA Collection Tubes (Roche) appear to be superior to EDTA tubes in maintaining cfDNA yield and quality over 7 days at RT. Several studies are available in the literature and describe the performance of blood collection devices with respect to cfDNA recovery.Browne et al. [19] compared the performances of LBgard Blood Tube, Streck cfDNA BCT, PAXgene Blood ccfDNA Tube, Cell-Free DNA Collection Tube and BD Vacutainer Blood Collection Tubes and found that LBgard Blood Tube out-performed all tested blood tubes in inhibiting genomic DNA release for the longest duration (14 days) and across the widest temperature range (4 • C, 25 • C and 37 • C).LBgard Blood Tube shows equivalent inhibition of genomic DNA release to Streck cfDNA BCT for clinical samples, and both tubes out-perform EDTA blood tubes.LBgard Blood Tube also consistently shows better CTC stabilization and inhibition of hemolysis over Streck cfDNA BCT.A comparison between STRECK and PAXgene tubes in combination with isolation procedures using QIAgen, Maxwell or Chemagic kits was described by Lampignano et al. [20] .

Effect of time to processing samples on ctDNA concentrations in blood
A comparison of optimal conditions for blood collection for circulating tumor DNA (ctDNA) was carried out by Kang et al. [21].Blood was collected into Streck cfDNA BCT, standard BD Vacutainer Blood Collection Tubes, and CellSave Preservative Tubes and plasma was processed after 2, 6, and 48 h post-collection.Droplet dPCR (ddPCR) was used to quantify plasma ctDNA from patients who had tumor tissue mutations detected by commercially available ddPCR assays.ctDNA abundance was similar and stable for up to 6 h in all tube types, and there was no effect of storage temperature on the yield for Streck and EDTA tubes [21].After 48 h, however, one out of four patients with detectable ctDNA showed a 50% decline in ctDNA, and three out of six patients showed a 2-3-fold increase in wild-type DNA in the BD Vacutainer Blood Collection Tubes.Therefore, Streck cfDNA BCT, BD Vacutainer Blood Collection Tubes, and CellSave Preservative Tubes showed similar performance in preserving ctDNA for up to 6 h before plasma isolation but Streck cfDNA BCT, and CellSave Preservative Tubes more consistently stabilized ctDNA at 48 h than EDTA tubes [21].
Markus et al. [22] compared 3 blood collection protocols (BD Vacutainer Blood Collection Tubes processed within 1 h and Cell-free DNA Blood Collection Tubes processed within 24 and 72 h) using plasma samples from healthy volunteers and found no significant differences in cfDNA yield, fragment size and background noise between them.Nikolaev et al. [23] tested BD Vacutainer Blood Collection Tubes, Streck cfDNA BCT, PAXgene Blood ccfDNA Tubes, and Cell-Free DNA Collection Tubes to determine how effective the four types of tubes were in preventing contamination by genomic DNA, as a function of time and storage temperature (4 • C, room temperature, 30 • C and 3 • C) and investigated the effect of delaying plasma preparation after blood draw, as it might happen due to transportation.In BD Vacutainer Blood Collection Tubes kept at room temperature (22-24 • C), a noticeable contamination by high molecular weight DNA, most likely originating from leukocyte lysis, was apparent after 24 h.In contrast, blood sampled into specially formulated tubes (Streck cfDNA BCT, PAXgene Blood ccfDNA Tubes), ctDNA remained without noticeable contamination by high molecular weight DNA even after one week at 22 • C. In Cell-Free DNA Collection Tubes, no contamination was detectable after 1 week, however a small but statistically significant increase in the 305 bp/41 bp DNA ratio was detected by PCR, thus suggesting a small contamination by HMW DNA [23].
Parpart-Li et al. [24] evaluated BD Vacutainer Blood Collection Tubes and Streck cfDNA BCT tubes, on cfDNA and ctDNA recovery.A total of 45 samples from nine patients with cancer were collected in both tube types.Once collected, blood was processed into plasma immediately or kept at room temperature and processed into plasma at 1, 3, 5, or 7 days.As early as 24 h after collection, plasma isolated from blood collected in BD Vacutainer Blood Collection Tubes contained an elevated level of cfDNA that increased over time compared with Streck cfDNA BCT tubes where no significant increase in cfDNA levels was observed [24].When samples from an additional six patients with cancer, collected in the same manner, were stored at 4 • C in BD Vacutainer Blood Collection Tubes over the course of 3 days, total cfDNA and ctDNA levels were comparable between samples collected in Streck cfDNA BCT tubes.A decrease in ctDNA levels in both tubes that observed at day 3 that was more pronounced when measuring the mutant allele fraction for samples stored at 4 • C in BD Vacutainer Blood Collection Tubes [24].
In conclusion, EDTA tubes may be used only when cfDNA isolation is following within 6 h of venipuncture.For Streck cfDNA BCT, standard BD Vacutainer Blood Collection Tubes, CellSave Preservative Tubes, Streck cfDNA BCT and PAXgene Blood ccfDNA Tubes, stability is warranted for 24 h, and in some cases even up to one week at room temperature (Streck cfDNA BCT, PAXgene Blood ccfDNA Tubes).

Centrifugation
Sherwood et al. [12] collected blood from 20 NSCLC patients.Part of these samples were analyzed on DNA yields for the effects of using a single or a double centrifugation step.Centrifuging the plasma samples twice did not have a significant effect on DNA yield compared to single spin when processed within 2 h.After 72 h, a double centrifugation step resulted in decreased mean DNA yield compared to single centrifugation.DNA yields may thus vary, depending on the method by which plasma was processed.

Maternal and fetal blood derived cfDNA
Wong et al. [25] investigated the amount of cfDNA from maternal blood stored in Streck cfDNA BCT tubes for up to 7 days at ambient temperature and found stable concentrations of cfDNA.The amount of fetal DNA did not change over a broad range of storage temperatures (4 • C, 23 • C, 37 • C, 40 • C), but the amount of total (largely maternal) DNA increased in samples stored at 23 • C and above, indicating maternal cell degradation and genomic DNA release at elevated temperatures.Shipping maternal blood in Streck cfDNA BCT tubes did not affect sample quality.Therefore, maternal cfDNA stabilized for 0 to 7 days in Streck cfDNA BCT tubes can be used for non-invasive prenatal molecular applications, when temperatures are maintained within the broad parameters assessed their study [25].

cfDNA extraction
cfDNA is extracted by many commercially available kits and platforms.There are many methods for extracting the cfDNA, which include both the use of CE-IVD commercial kits and protocols developed by laboratories (non CE-IVD); however, the use of certified kits and platforms is strongly encouraged.Because of the small amount and the very fragmented nature of the cfDNA in plasma, extraction methods validated on tissue samples or other biological matrices must not be used for plasma specimens.The extraction method must be reliable and generate as much DNA as possible of the sample under examination, as not to compromise the analysis and generate false negative or positive results.It is highly recommended that an internal non-human DNA standard would be used to control for DNA extraction efficiency (Beck 2014 AACC Annual Meeting Abstracts: S194).
Methods for extraction and purification of cfDNA from plasma are either based on the use of columns with silica membranes, in association with a vacuum pump or centrifugation, or magnetic beads.These systems allow the simultaneous extraction of 1 to 24 samples and capture fragments of cfDNA starting from a minimum of 10 μl to a maximum of 10 ml of input plasma.In general, it is believed that 2-4 ml of plasma is the minimum volume needed to obtain accurate results using the different extraction methods.Furthermore, most of the kits are able to concentrate the eluate in a volume of elution of 20-150 μl.An excellent overview of a comparison of 6 different kits for ccfDNA extraction (QIAamp CAN, QIAsymphony, QIAampminielute, Maxwell AX1115, Maxwell AS1480 and Chemagic) was published recently [20].
Once extracted, the cfDNA must be quantified, for example using an Agilent 2100 Bioanalyzer, to discriminate the size range of DNA fragments.In general, the extraction methods mentioned above allow to obtain cfDNA samples of high quality and with a concentration above 5 ng/µl.The amount of DNA extracted is however influenced by the stage and type of the disease; for example, high amount of ctDNA is obtained from colorectal cancer patients, while in NSCLC subjects the amounts are generally low and correlate with tumor burden [26]

Conclusions
Given the challenges raised by the analysis of cfDNA, solid validation of pre-analytical procedures is a critical prerequisite in order to minimize the many sources of variation.The inclusion of tubes containing preservatives to reduce cell lysis was shown to improve the stabilization of peripheral blood compared to EDTA tubes when the delivery of the sample to the laboratory exceeds 2 hours after drawing.Additionally, the number of centrifugation steps, input plasma volume and extraction kit used may influence the cfDNA yield.A careful assessment of preanalytical variables in the laboratory and the adoption of written and validated procedures subjected to external quality assurance assessment schemes (EQAAS) will allow the laboratory to define the conditions of sample acceptance and to issue a set of recommendations to prescribing physicians to improve the likelihood of successful mutation detection from plasma, further reducing the need for invasive tumor biopsies.