Abstract
Objective The present study aimed to investigate the influence of several important preanalytical factors (storage period of the tumor block, maximal diameter of the tumor circled area, tumor volume and tumor fraction) on the isolated DNA from formalin-fixed paraffin-embedded (FFPE) tissues in a series of thyroid carcinomas.
Design Our study included 212 FFPE blocks, archived in the Department of Pathology, Târgu-Mureș Emergency County Hospital for up to 10 years. DNA isolation was performed using a commercially available kit (MasterPure DNA purification kit, Epicentre). The DNA parameters (concentration and purity) were determined using a spectrophotometer and the Qubit 2.0 Fluorometer (Thermo Fisher Scientific) for an accurate and sensitive DNA quantification.
Results The mean DNA concentration and purity for the study cases were 489.3±372.6 ng/µl and 1.667±0.1912, respectively. The DNA concentration was correlated with the maximal diameter of the tumor circled area (p<0.0001), the tumor volume (p<0.0001) and tumor fraction (p=0.0462). No statistically significant differences both in terms of DNA concentration (p=0.374) and purity (p=0.125) in relation with the storage period of the tumor blocks were observed. When using a fluorometric quantification method, the DNA concentration was lower (mean DNA concentration: 47.15±32.85 ng/µl), but similar correlations with the morphological factors were observed. Apart for three cases, the real-time PCR amplification of the BRAF gene was successfully assessed in all cases.
Conclusion The maximal diameter of the tumor circled area, tumor volume and tumor fraction are important morphological factors that correlate with the DNA concentration and should be carefully assessed in routine practice prior to performing DNA isolation from FFPE tissues.
Significance of this study
What is already known about this subject?
In recent years, together with the development of targeted, personalized therapies, isolation and preservation of tumor DNA has become an important matter in molecular oncology.
Formalin-fixed paraffin-embedded (FFPE) tissues are now used by many laboratories as a powerful source of biological material both for clinical diagnosis and research.
Isolation of an adequate genomic DNA from FFPE tissues is not always easily achieved and might represent a challenging task.
Preanalytical factors play an important role in successful DNA extraction from FFPE specimens and downstream molecular analysis.
What are the new findings?
DNA isolation from achieved FFPE tissues using commercially accessible DNA isolation kits in laboratories with basic equipment is possible and feasible.
The DNA concentration obtained from FFPE tissue blocks is highly dependent on a constellation of morphological preanalytical factors, which include the maximal diameter of the tumor circled area, the tumor volume and the tumor fraction.
All these parameters can easily be evaluated by the pathologist on the mirror H&E staining prior to performing DNA isolation from FFPE specimens.
Significance of this study
How might these results change the focus of research or clinical practice?
All these morphological parameters should be carefully assessed in routine practice prior to performing DNA isolation from FFPE specimens.
To our knowledge, there are very little published data that make such a correlation, whereas the idea that DNA yield might be predicable from a digital scan of an H&E slide is something that molecular pathologists are asking
Introduction
Tissue samples in routine pathology setting are almost always fixed in formalin and stored as formalin-fixed paraffin-embedded (FFPE) blocks.1 Compared with fresh-frozen samples, FFPE blocks are easy to handle, broadly available and imply low expenses for long-term storage. Due to all these, they represent the preferred method for preserving tissues in pathology departments all over the world.2–4
In recent years, together with the development of targeted, personalized therapies, isolation and preservation of tumor DNA has become an important matter in molecular oncology. FFPE tissues are now used by many laboratories as a powerful source of biological material both for clinical diagnosis and research.1 2 5–9 However, DNA recovery from FFPE tissues is a challenging task, highly dependent on a number of preanalytical factors. These include: tissue processing (formalin fixation) and storage (period and conditions), type of specimen (surgical resection vs superficial biopsy), tumor type and tumor site, tumor size and cellularity, tumor fraction and viability, decalcification procedures (bone) and other factors (eg, blood, mucin).10
DNA extraction from FFPE tissues is challenging mainly because formalin fixation induces chemical cross-linking of protein and nucleic acids which might compromise the DNA integrity.10–13 On the other hand, FFPE tissues also have many advantages. When performing DNA extraction from FFPE tissues, the tumor fraction, an important preanalytical parameter can also be estimated by using a representative mirror H&E-stained section and sequential cuts from the FFPE tissue blocks. Despite the fact that frozen tissue provides unfragmented and high-quality DNA, it does not always allow an accurate, reliable evaluation of the tumor (presence of the tumor and the tumor fraction).10
The relationship between the storage period of the tumor block and its effect on the extracted macromolecules has received considerable attention in the last period.1 3 5 14–16 Archived FFPE tissues are required in clinical care for the development of target treatments that often imply molecular analysis of tumor tissue samples several years after the diagnosis. Older FFPE tissues are also required for research purposes of new molecular targets in large retrospective cohort studies, with long-term follow-up data.1 However, further studies and investigations are needed in order to determine the optimal storage duration of the FFPE tissues in this fit-for-purpose approach.17 18
All in all, preanalytical tissue selection plays an important role for successful downstream molecular analysis. The present study aimed to evaluate the influence of several important preanalytical factors on the concentration and purity of the DNA extracted from FFPE specimens in a series of thyroid carcinoma cases archived in our department for up to 10 years. The following factors were evaluated: the storage period of the tumor block, the maximal diameter of the tumor circled area, the tumor volume and the tumor fraction.
Material and methods
Ethics
The study complies with the Declaration of Helsinki. Informed consent was obtained from all the subjects included in the study.
Samples
The FFPE specimens included in the study were all retrieved from the archives of the Pathology Department, Târgu-Mureş Emergency County Hospital, Romania.
Based on the availability of remnant resections of the surgical specimens, 212 FFPE blocks corresponding to malignant thyroid tumors (stored for 7–10 years, 4–6 years and 1–3 years, respectively) were analyzed in this study (table 1).
The FFPE blocks were stored at a temperature ranging between 17°C and 22°C. All tissue samples were fixed immediately following surgery in fresh neutral buffered formalin. The fixation period was 48–72 hours, depending on the size of the tissue sample. The formalin volume was at least five times greater than the volume of the specimen. Regarding the postfixation processing technique, all samples were processed in the same pathology laboratory at Târgu-Mureș Emergency County Hospital, using long-established, consistent protocols which encompassed dehydration, clearing and paraffin embedding.
Pathological diagnosis
The histological type of tumor was established in compliance with the 2004 WHO Classification of Tumors of Endocrine Organs.19 Both papillary thyroid carcinoma (PTC) and follicular thyroid carcinoma (FTC) cases were included in the study. Papillary thyroid microcarcinomas (PTMCs) were defined as PTCs incidentally discovered, measuring 1 cm or less in diameter.
DNA extraction from FFPE tissues
Two pathologists (ANB and AB), specialized in endocrine pathology, analyzed and reviewed all the cases included in the study. They selected one tissue block per case and circumscribed the tumor area on the H&E stained slides. In order to obtain better results regarding the DNA concentration, more cellular tumor areas were selected in favor of fibrotic areas. Hemorrhagic or calcified tumor areas were attentively avoided from being selected for the molecular assay, since decalcification procedures or hem in erythrocytes may inhibit PCR reaction.10 For each case, sections (4×4 µm thick) were cut from the tissue blocks. Using a standard microscope and the H&E stain as a guide the tumor area was manually macrodissected. The tissue fragments were scraped from the preselected area and transferred to an Eppendorf tube for DNA isolation using the MasterPure DNA purification kit (Epicentre, Madison), as previously described.20 The DNA elution volume was 60 µL. Given that photometry is also influenced by RNA that co-elutes with the DNA, a RNase digest was incorporated into the DNA digestion protocol.
In all cases, the maximal diameter of the circled tumor area and the tumor volume were assessed. The tumor volume was obtained by multiplying the tumor area surface by 0.004 mm (the thickness of a FFPE section) and by 4 (the number of FFPE sections used per case). The tumor fraction was also approximated from the mirror HE section and defined as the percentage of the tumor cells in relation with the tumor stroma.
DNA concentration and purity
The DNA parameters (concentration and purity) were first assessed using a spectrophotometer (Eppendorf BioSpectrometer, Hamburg, Germany). The DNA concentration was determined by measuring the absorbance at 260 nm (A260) and DNA purity by calculating the A260/A280 ratio.
In order to assess the specific amount of intact double standard DNA (dsDNA), we have also further quantified the isolated DNA samples by means of a fluorometric assessment method using the Qubit 2.0 Fluorometer (Life Technologies, Thermo Fisher Scientific, Massachusetts, USA) and a high-sensitivity (HS) assay kit. The Qubit dsDNA HS assay kit is selective for dsDNA because the kit contains fluorescent dyes that are specific to the target molecule of interest and the fluorescence is emitted only when the dyes bound to the specific (dsDNA) target molecule.
Real-time PCR amplification
All cases were further subjected to real-time PCR amplification targeting the BRAF V600E somatic mutation. Amplification reactions were performed using a 7500 Fast Dx real-time PCR system (Applied Biosystems, USA) and the Thyroid Cancer Mutation Analysis Kit (EntroGen, USA, BRAF cosmic ID 476) for most of the cases. Some cases (n=25 PTMC cases) were amplified using a LightCycler 480 system (Roche Diagnostics, Vienna, Austria) and a pair of primers (forward and reverse) targeting the BRAF V600E point mutation (amplificon length 109).
Statistical analysis
Statistical analysis was performed with the Statistical Package for Social Sciences software, V.20 (Chicago, Illinois, USA). Data were labeled as nominal or quantitative variables, characterizing the nominal variables by their mean frequencies. The Kolmogorov-Smirnov test was used for assessing the distribution of quantitative variables. Whenever appropriate, the quantitative variables were described by mean±SD or median and centiles. The χ2 test was used to compare frequencies in nominal variables while t test and the analysis of variance test were applied to analyze differences in the mean or median between groups.
The linear relationship between a dependent variable (Y-mean tumor volume) and independent variables (X –DNA concentration and purity) were calculated using linear regression. A linear regression model was applied to describe the dependent variable using a straight line, defined by the equation Y=a + b× X (Y is the intersection of the line, b is the line’s slope). In order to assess and better define the independent relationship between different variables (eg, the tumor fraction and the tumor volume) Spearman’s correlation was also applied. A p value lower than 0.05 was considered statistically significant.
Results
We investigated the purity and quantity of the DNA isolated from 212 FFPE tissue blocks that had been archived in the pathology department for 1–3 years, 4–6 years and 7–10 years, respectively (table 1). The cases corresponded to 163 PTCs, 43 PTMCs and 6 FTCs.
Table 2 shows the DNA concentrations (both spectrophotometric and fluorometric assessments) and DNA purity values for the study cases in relation with the factors that were evaluated.
The spectrophotometric assessment revealed good quantity and purity for obtained DNA from all of the cases included in our study (mean DNA concentration: 489.3±372.6 ng/µl; mean A260/A280 ratio: 1.667±0.191). The DNA concentration was significantly higher among cases with a larger mean tumor volume (ranging from 273.5±249.2 ng/µl for cases with a mean tumor volume between 0.06 mm3 and 1.99 mm3 to 956.4±346.9 ng/µl for cases with a mean tumor volume between 6.00 mm3 and 8.99 mm3; p<0.0001). The same results were observed for the maximal diameter of the circled tumor area (p<0.0001). Figure 1 demonstrates the statistically significant linear increasing trend of the mean DNA concentration depending on the tumor volume (R2=0.3425, p<0.0001).
Most of the cases included in the present study were follicular variants of PTC cases that are generally characterized by rich cellularity and reduced stroma. As a consequence, more than half of the study cases were consistent with a high tumor fraction of more than 80% (n=133/212, 62.7%). As expected, these cases also had the highest mean DNA concentration (see table 2) and the differences were statistically significant when compared with cases with lower tumor fractions (p=0.0462). The Spearman correlation between the tumor fraction as independent variable (X) and tumor volume as dependent variable (Y) demonstrated only a weak a correlation (R=0.3192, p<0.001) (figure 2).
No statistically significant differences regarding the DNA quantity were noticed between FFPE tumor blocks that were stored for 1–3 years, 4–6 years and 7–10 years (p=0.3744), respectively.
The highest purity (1.716±0.1442) was obtained for cases with a mean tumor volume ranging between 2.00 mm3 and 3.99 mm3; when compared with smaller (0.06–1.99 mm3) or larger (4.00–8.99 mm3) tumor volume cases the obtained p value was 0.0089. No statistically significant linear differences of the DNA purity trend dependent on the tumor volume were observed for the study cases (R2=0.092, p<0.1626) (figure 3). No differences in terms of DNA purity in relation with the pathological diagnosis for the study cases (p=0.4402), tumor fraction (p=0.6671) and storage period of the tumor blocks (p=0.124) were observed.
When using a fluorometric quantification method, as expected, the DNA concentrations were lower (mean DNA concentration: 47.15±32.85 ng/µl) (see table 2). Nevertheless, similar correlations with the morphological parameters were observed. The DNA concentration was significantly higher among cases with a higher maximal diameter of the tumor circled area (p=0.0129), a larger mean tumor volume (p=0.0200) and a higher tumor fraction (p=0.0023). Regarding the storage period of the FFPE tumor block, the DNA concentration was higher among tumor blocks that were stored for 1–3 years compared with 4–6 years and 7–10 years, respectively, but the differences did not achieve statistical significance (p=0.5650).
For the majority of the cases, the real-time PCR amplification of the BRAF gene (targeting the BRAF V600E point mutation) was successfully assessed, apart for three cases. These three cases corresponded to FFPE blocks that were stored for 8 years, 3 years and 3 years, respectively, in our department. The purity of the isolated DNA for these cases was 1.50, 1.84 and 1.77, respectively.
Out of the 209 remaining cases, 65 cases (31.1%) were BRAF V600E positive (figure 4), whereas 144 cases (68.9%) were BRAF negative (non-mutated).
Discussion
An important component of the developing personalized cancer treatment in solid tumors is represented by molecular investigation of gene mutations.16 In pathology departments all over the world the majority of tissue samples are processed and stored as FFPE specimens. Thus, FFPE tissues are the standard substrate for most clinical laboratory testing in routine practice.10 FFPE tissue blocks can be stored at room temperatures for long periods of time, at low expenses, and can be used both for histological examinations as well as for ancillary studies (special staining, immunohistochemistry and recently, molecular testing).10 The collection and storage of frozen tissues, on the other hand, is relatively costly, resource-intensive and not always available.3 5 21 Nevertheless, isolation of an adequate genomic DNA from FFPE tissues (both in terms of quality and quantity) can sometimes represent an extremely challenging work. The success of the next-generation sequencing (NGS) mutation analysis, for example, is affected by the DNA input quantity and also by other factors such as the DNA quality, the specimen type and the cellularity of the investigated tumor.10 Small tissue specimens (here included extremely small tumor sections with less than 10 mm2 of selected tumor rich tissue) can also be problematic. They often fail to provide an adequate DNA, even if multiple (10 or 20) unstained FFPE sections are used for DNA extraction.16
In the present research, the DNA obtained from a series of FFPE tissues corresponding to thyroid carcinomas that had been archived in our department for up to 10 years was examined, aiming to establish if the preananytical factors had a significant impact on the extracted DNA. The main strengths of our study are the large number of samples, the macrodissection of the tumor-only fraction and the use of fixation times that are longer than 24 hours (48–72 hours). We demonstrated that DNA isolation from FFPE tissue blocks by using commercially accessible DNA isolation kits in laboratories with basic equipment is possible and feasible. Further on, important parameters regarding the DNA quantity and quality can also be assessed, allowing consecutive efficient PCR amplification.
We obtained good DNA quantity and purity for isolated DNA from all of the cases included in our study. When using a fluorometric method for an accurate and sensitive DNA quantification and assessment of the dsDNA, lower DNA concentrations were observed. This is most probably related to the DNA degradation, a well-known problem when dealing with samples represented by FFPE tissues since DNA extracted from FFPE tissues is a mixture of single stranded DNA and dsDNA. However, although the DNA concentration was lower compared with the spectrophotometric method, similar correlations with the morphologic parameters were observed.
Our results have shown that the DNA concentration is highly dependent on a constellation of morphologic preanalytical factors, which include the maximal diameter of the tumor circled area, the tumor volume and the tumor fraction. All these parameters can easily be evaluated by the pathologist on the mirror H&E stain prior to performing DNA isolation from FFPE specimens. To our knowledge, there are very little published data that make such a correlation, whereas the idea that DNA yield might be predicable from a digital scan of the H&E slide is something that molecular pathologists are asking.
The size of the circled tumor area is often related to the type of surgical procedure that was performed and should be larger than 10 mm2 (corresponding approximately to a 3 mm maximal tumor diameter).10 Surgical resections and excisions are usually consistent with a large tumor size and elevated DNA yield, whereas superficial and image-guided biopsies are generally performed for small or extremely small tumors and provide a reduced sample of tumor tissue, being thus associated with a low DNA yield.16 Tumor fraction determinants, on the other hand, include: tumor type (solid vs cystic vs sclerotic vs mucinous), primary versus metastasis, decalcification procedures and associated stromal and inflammatory cells.10 Previous studies have demonstrated that the tumor fraction has an important effect on the successful rate of the NGS tests for mutational analysis16 22 and should be minimum 10%–20%.10 In our study group most of the cases were follicular variants of PTCs, tumors that are generally characterized by rich cellularity and reduced stroma. Moreover, in order to obtain better results regarding the DNA concentration, when performing the DNA extraction steps, more cellular tumor areas were selected in favor of fibrotic areas. As a consequence, more than half of the cases in our study were consistent with a high tumor fraction of more than 80% (n=133/212, 62.7%). These cases also had the highest mean DNA concentration and the differences were statistically significant when compared with cases with lower tumor fractions (p=0.0462 spectrophotometric assessment, p=0.0023 fluorometric quantification, respectively).
Regarding the tumor type, as expected, PTMCs were consistent with the lowest DNA concentrations, when compared with PTC and FTC cases (p<0.0001). The lower DNA yields in these cases came as a consequence of the smaller number of cells, related to the smaller tumor volumes in PTMC cases.
No differences in terms of DNA purity in relation with the pathological diagnosis for the study cases, the tumor fraction and the storage period of the tumor blocks were observed. Also, the DNA purity did not correlate with the mean tumor volume. These results highlight a reliable and reproducible DNA isolation technique, which further warrant an efficient PCR amplification. The DNA extraction was performed in a consistent manner for all the studied cases regardless of the pathological diagnosis, the tumor fraction, the tumor volume and the storage period of the FFPE blocks.
Long-term storage of paraffin blocks has received considerable attention in the last years.17 18 Nevertheless, literature data are controversial. Some studies have reported that FFPE blocks can be stored for several years with minor or no effects on the subsequent DNA analysis.5 23 23–25 Kokkat et al,5 for example, have reported no differences in DNA quality and quantity in relation with the storage period of the FFPE tissues they had examined. DNA isolated from FFPE blocks archived for 11–12 years, 5–7 years and 1–2 years, respectively, demonstrated no differences compared with DNA obtained from very recent, less than a year-old FFPE blocks. By contrast, other studies have reported a diminished ability to amplify the DNA when obtained from FFPE tissues that are 5 years old or older,3 or a significant degradation of the DNA isolated from the same FFPE block after few years of storage (4–6 years).1 Given all these controversies, our study, with the large sample size, is a potential useful addition to this set of data.
We noticed no statistically significant differences related to the storage period of the tumor blocks among the study groups. Our results have shown that DNA obtained from FFPE blocks archived in the pathology department for up to 10 years is comparable to the DNA isolated from much more recent FFPE tissues, stored for 1 year, 2 years or 3 years, both in terms of quantity (p=0.374 spectrophotometric assessment, p=0.565 fluorometric quantification, respectively) and purity (p=0.124). All tissue samples in this study were fixed immediately following surgery in fresh neutral buffered formalin and the fixation period was 48–72 hours, depending on the size of the tissue sample. Although we have used blocks with a long fixation time (>24 hours), this aspect did not appear to have influenced the DNA stability over time. Moreover, real-time PCR amplification of the BRAF gene (targeting the BRAF V600E point mutation) was successful for all the cases, except for three cases, despite the fact that there were no statistically significant differences regarding the preanalytical factors and DNA extraction results compared with the other cases. DNA degradation, an important issue related to FFPE tissues11 26 27 could explain the poor-quality PCR amplification in these particular cases. Approximately a third (31.1%) of the study cases were BRAF V600E positive, a result that is consistent with our previous publication20 including 25 PTMCs of which 8 cases (33.3%) harbored the BRAF V600E mutation.
Conclusions
We demonstrated that the maximal diameter of the tumor circled area, the tumor volume and the tumor fraction are important factors that correlate with the DNA concentration and should be carefully assessed in routine practice. No statistically significant differences related to the storage period of the tumor blocks were observed among the study groups regarding both DNA concentration and purity. Nevertheless, the DNA purity did not reveal significant differences neither in relation to the pathological diagnosis for the study cases, nor in relation with the tumor volume or the tumor fraction. These results highlight a reliable and reproducible DNA isolation method, which further warrant an efficient PCR amplification
Footnotes
Contributors ANB: histological examination of the cases, data collection and analysis, molecular analysis, study design and drafting the manuscript; CB: molecular analysis and major contribution to study design and in writing the article; VM: molecular analysis; AEZ: data collection and analysis; ES: histological examination of the cases; AB: histological examination of the cases and coordination of the study, with a major contribution in study design and in writing the article. All the authors read and approved the final manuscript.
Funding This work was supported by the George Emil Palade University of Medicine, Pharmacy, Sciences and Technology of Târgu-Mureş, Romania, Research Grant No. 275/1/11.01.2017.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval Ethics Committee of George Emil Palade University of Medicine, Pharmacy, Sciences and Technology of Târgu-Mureş, Romania (letter of approval no.36/07.03.2017).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.