DNA extracted from Formalin-Fixed Paraffin-Embedded (FFPE) tissue blocks is increasingly being used for applications such as Next Generation Sequencing (NGS) both in oncology research and clinical diagnostics. However, the accuracy and reproducibility of this analysis, as well as its resilience to variations in the quality of the samples used, remain to be ascertained. IBBL recently released its 100th publication, reviewing specifically the preanalytical factors affecting the quality of DNA extracted from fixed tissue. Dr. Olga Kofanova, Biospecimen Quality Team Leader at IBBL, explains the main findings.
Interview by Federica Amato (FA)
FA: In this publication you focus on DNA extraction from fixed tissues. Can you explain what fixation entails and why it is an important factor to consider when evaluating the quality of the extracted DNA?
OK: Before tissue samples can be used for diagnostics or research purposes, they need to be preserved and be made amenable to further analyses. This is done either chemically (by fixation or stabilisation) or cryogenically (by freezing). Typically, most tissue biospecimens are fixed using formalin and are then embedded in paraffin, which results in what is known as a formalin-fixed paraffin-embedded (FFPE) tissue block. Despite the many advantages of this method, the process of formalin fixation often induces the fragmentation of the extracted DNA, as well as chemical modifications and crosslinking, which can negatively affect the efficiency of the extraction and of the amplification of the extracted DNA. Non-formalin alternatives such as the PAXgene tissue fixative were developed to address these shortcomings, resulting in so-called PAXgene-fixed paraffin-embedded (PFPE) tissue blocks.
FA: In addition to the type of fixative, what are the other critical parameters that can affect the quality and quantity of the DNA extracted from FFPE tissues?
OK: Considerable work has been done by the international community, including the National Cancer Institute (NCI) and the European Committee for Standardization (CEN), on this subject. The most critical preanalytical variables to be considered are tissue fixation time, storage conditions of the FFPE blocks and the choice of DNA extraction method. Indeed, the goal is to have enough DNA of sufficient integrity after extraction, if we want to obtain reliable results when performing downstream NGS analyses. This is why, prior to NGS, it is important to carry out Quality Control of the extracted DNA to assess its yield and integrity. The former is typically done by techniques such as spectrofluorometry or quantitative PCR, while DNA integrity is evaluated by the DNA Integrity Number (DIN). In this paper, we specifically assessed the impact of different types of fixatives – namely FFPE and PFPE – and DNA extraction methods on the quality of the extracted DNA.
FA: Can you briefly describe the experimental setup?
OK: The idea behind the study design is simple. We took 15 FFPE tissue blocks and 15 PFPE tissue blocks, extracted DNA using four different DNA extraction methods and then performed a series of analyses to assess whether the type of fixative and extraction technique affected various quality parameters in the extracted DNA. Specifically, we measured DNA concentration and purity by spectrophotometry and spectrofluorometry, we assessed DNA integrity and the extent of crosslinking by PCR and Whole Genome Amplification (WGA) and carried out NGS followed by bioinformatics analysis to estimate any damage in the DNA samples which would result in incorrect sequencing data. We also compared the results with 15 paired “gold standard” tissue samples, collected in a nucleic acid stabiliser.
FA: What did you find out?
OK: In terms of DNA yield and purity, we found that FFPE and PFPE processing resulted in lower quantities of extracted DNA compared to the reference samples. When looking specifically at DNA extracted from FFPE samples, higher yields were observed when using the QIAamp DNA FFPE extraction method, compared to the other extraction techniques. Conversely, we found the yield of DNA extracted from PFPE tissue to be higher when using the PAXgene Tissue DNA and Chemagic DNA extraction methods. With regard to the purity of the extracted DNA, the best quality DNA from fixed tissue was obtained with the PAXgene Tissue DNA kit from PFPE tissue and with the QIAamp DNA FFPE kit from FFPE tissue. As for DNA integrity, the best quality DNA from fixed material was DNA extracted with the Chemagic DNA Tissue kit from either PFPE or FFPE tissue. Essentially, we showed that the combination of fixative and DNA extraction method affect the yield, purity and integrity of the resulting DNA. In particular, the use of non-formalin fixatives such as PAXgene coupled with the Chemagic extraction method has shown better results in terms of DNA yield and integrity, but also accuracy of NGS analyses.
FA: And what are the implications for NGS and Quality Control?
OK: Once we have measured parameters such as the ones discussed above, we can use these values to detect the DNA samples of higher quality for the purpose of NGS analysis. Namely, the best ‘diagnostic’ performance in identifying the DNA samples that would be best suited for NGS was obtained with the combination of double stranded (ds) DNA proportion and DNA integrity. Specifically, we found that the DNA samples most amenable to NGS should contain at least 25% of double-stranded DNA (dsDNA), with DNA fragments having a length of at least 200 base pairs – as measured by multiplex PCR – and at least an ’intermediate’ integrity score – as measured by WGA. Researchers can use this finding as a useful combinatorial Quality Control (QC) assay to assess whether their extracted DNA samples will give reliable results when being subjected to NGS.
FA: This paper is a particularly special one for IBBL!
OK: Indeed it is! It is our 100th publication, a considerable milestone! From a scientific standpoint, the paper reviews critical aspects pertaining to the preanalytics of DNA extraction and proposes a combinatorial QC metric that guides researchers in the qualification of DNA samples and in the assessment of their fitness for purpose in terms of NGS. In addition, the publication is the result of fruitful collaborations with partners from the Greater Region and from the Luxembourg biomedical research environment, specifically the Luxembourg Institute of Health (LIH), which assisted us with bioinformatics analyses. Moreover, the publication is part of a collection of articles published in the context of the European project SPIDIA4P.
FA: Thank you very much Olga for taking the time to answer my questions!
The detailed results of the study can be found in the full paper here.