Cold Formalin Fixation Guarantees DNA Integrity in Formalin Fixed Paraffin Embedded Tissues: Premises for a Better Quality of Diagnostic and Experimental Pathology With a Specific Impact on Breast Cancer

Abstract

Formalin fixation and paraffin embedding (FFPE) represent the standard method to preserve tissue specimens for diagnostic pathology, however formalin fixation induces severe fragmentation of nucleic acids. We investigated whether formalin fixation at 4°C could preserve DNA integrity in FFPE specimens. Paired samples from 38 specimens were formalin fixed at room temperature (stdFFPE) and at 4°C (coldFFPE), respectively. Two independent cohorts were prospectively collected, cohort A (collected 6 years prior to the study, n = 21), cohort B (collected at time of the study, n = 17). DNA was extracted and its integrity evaluated with a qPCR-based assay that produces a normalized integrity index, the QC score (ratio between the quantity of a long and a short amplicon of the same gene). We observed higher QC scores in coldFFPE compared to stdFFPE samples (mean values: 0.69 vs. 0.36, p < 0.0001) and stdFFPE breast cancer specimens showed the most detrimental effect overall. Comparable QC scores were obtained between coldFFPE tissues of both cohorts; conversely, DNA integrity of stdFFPE was significantly lower in cohort A compared to cohort B (p < 0.0001). Of note, QC scores of stdFFPE (but not of coldFFPE) samples were significantly reduced following 6 months of storage (p = 0.0001). Monitored formalin fixation at 4°C outperforms standard fixation in ensuring high-quality DNA, which is key to feasibility of downstream high-throughput molecular analyses. An important effect was observed over storage time, thus suggesting a likely better preservation of archival samples when this cold fixation protocol is used.

Keywords: DNA fragmentation; biomarkers; breast cancer; cold formalin; diagnostic accuracy; fixation; molecular diagnostics; oncology.

Figure 1

Design of the study. The full cohort comprised 38 cancer specimens that were collected and sampled in parallel to allow standard fixation (i.e., at room temperature) and cold formalin fixation (i.e., at 4°C). Following processing and tissue sectioning the H&E slides were reviewed to identify the tumor area that was mesodissected for DNA extraction. Two independent cohorts were prospectively collected: Cohort A, whose sampled were collected 6 years prior to the study and DNA extraction performed at present time; Cohort B, whose samples were collected at time of the present study with contextual DNA extraction. For 14 samples from Cohort B, i.e., corresponding to those samples for which at least 6 months elapsed from collection, two DNA extractions were performed: at baseline (at time of collection/fixation) and after 6 months of storage/archival. On the total 90 DNA samples we performed fluorometric and spectrophotometric quantifications and we ran a qPCR with the DEPArray™ FFPE QC Kit. RT, room temperature; NBF, neutral buffered formalin; EtOH, ethanol.

Figure 2

Quantity and quality of DNA extracted from parallel samples. (A) DNA quantification of samples. Violin plots representing the DNA μg purified from the specimens, grouped according to the fixation methods (stdFFPE and coldFFPE) and the time of cohort collection [Cohort A (collected 6 years prior to the study with DNA extraction at present time); Cohort B (collected at time of the present study with contextual DNA extraction)]. Cohort A was characterized by higher DNA yields for coldFFPE compared to stdFFPE samples (4.1 μg vs. 2.1 μg, p = 0.04). As for Cohort B, violin plots showed a median-around density distribution for both stdFFPE and coldFFPE samples that was significantly higher compared the same fixation protocol of Cohort A (p = 0.03 for cold fixation and p =0.0001 for standard fixation). (B) Violin plots representing the QC score distribution in the groups clustered according to fixation method. The stdFFPE cohort was characterized by heterogeneous QC scores with several samples in the low range (wider diameters of the black violin) and the DNA fragmentation was statistically significantly higher than in coldFFPE (gray violin). (C) Dot plot illustrating the Pearson correlation between the QC scores of corresponding samples fixed with the two protocols. The absence of a robust correlation suggests the reduced influence of the intra-individual fragmentation on the QC score. (D) QC distribution among the sample sets. Violin plots representing the QC score of the samples, which are grouped according to the fixation methods (stdFFPE and coldFFPE) and the time of collection [Cohort A (collected 6 years prior to the study with DNA extraction at present time); Cohort B (collected at time of the present study with contextual DNA extraction)]. The QC scores of coldFFPE samples are comparable between the two cohorts, regardless of time of collection/fixation (p = 0.79), whereas the QC score of stdFPPE samples are significantly lower in Cohort A compared to Cohort B (p < 0.0001). Notably, the DNA purified from Cohort B samples displayed a better preservation in both stdFFPE and coldFFPE, nevertheless coldFFPE samples still showed a higher QC distribution. n.s. not significant, *p < 0.05, ***p < 0.001.

Figure 3

QC score and QC score fold change in breast, colon, and lung carcinoma samples. The heatmap represents the QC score for purified DNA from stdFFPE and coldFFPE samples. The histogram shows the fold changes of the QC score in coldFFPE compared to the corresponding stdFFPE samples. The order of the samples in the heatmap is defined by the color in the histogram. All the breast stdFFPE samples shows the lowest QC score level, which is less heterogenous among the coldFFPE specimens. Nine of the first ten samples with and increased level of DNA integrity have a mammary site of origin, confirming the specific improvement for breast cancer tissue.

Figure 4

QC score distribution in the subgroup analysis of the 14 tissue pairs with a second DNA extraction after 6 months from collection/fixation. (A) Violin plots representing the QC scores in a head to head comparison between DNA extraction at baseline and after 6 months, subdivided by standard fixation (stdFFPE) and cold fixation (coldFFPE). Significantly lower QC scores are observed after 6 months in stdFFPE samples, whereas comparable QC scores are displayed for coldFFPE samples. (B) Bubble plot of the QC score for each patient. The plot showed the basal QC score on the Y axes and the 6-month QC score on the X axes, paired for each sample. The bubble size illustrates the drop in terms of QC score for each couple of DNA samples purified after 6 months of archival. n.s. not significant, ***p < 0.001.