Introduction: DNA extracted from cytologic samples is occasionally used for various molecular tests. The aim of this study was to evaluate DNA extracted from differently prepared cytologic slides that can be used for PCR-based molecular tests. Methods: For each 23 cases of papillary thyroid carcinoma or colorectal adenocarcinoma tissues, six touch-imprinted cytological slides were prepared (group 1∼3), and remnant tissues were blocked for FFPE tissue (group 4). Cytologic slides were grouped by preparation methods: air-dried slides (group 1), fixed slides (group 2), and stained slides (group 3). Fixed slides were classified as 95% ethanol fixed (group 2A) and CytoRich Red Preservative solution fixed (group 2B). Stained slides were divided in 3 ways: Giemsa, Pap, and H&E stained (group 3A, 3B, and 3C, respectively). DNA extracted from each group was evaluated for concentration, 260/280 ratio, DNA Integrity Number (DIN) value, and mutation. Results: DNA concentration was highest in group 1 and lowest in group 2B. DIN value was highest in group 2A and lowest in group 2B. A mutation of BRAF or KRAS genes was detected in 18 FFPE tissue samples. Matched DNA extracts from groups 1, 2A, and 3 produced results consistent with FFPE tissue results, while mutation testing was successful for only four samples of DNA from group 2B. Conclusion: The mutation tests worked well for most samples except CytoRich Red Preservative-fixed slides. This study indicates that stained and unstained cytologic slides are a suitable source of PCR-based molecular tests as long as they are fixed in ethanol or stored for a short time in an air-dried condition.
Many genetic modifications continue to be discovered in cancers, and these modifications are changing therapeutic paradigms [1, 2]. Drugs have been developed to target specific protein products of these mutated genes (so-called “druggable targets”) [1, 2], and thus, it has become common practice for pathology laboratories to test for mutations of several genes simultaneously in cancer. For example, lung adenocarcinoma can harbor abnormalities in EGFR, BRAF, ROS1, ALK, or other genes, and these can be tested concurrently using immunohistochemistry-based and molecular tests . As more molecular tests are being incorporated into clinical practice, laboratories require larger amounts of DNA, but occasionally, samples are insufficient for multiple genetic tests because only small biopsy specimens are acquired, especially from patients with advanced-stage cancers [4, 5] and when minimally invasive sampling methods are used [6, 7]. On the other hand, cytologic slides prepared from bile, peritoneal, pleural fluid, sputum, and urine are more easily obtained than tissue biopsy samples and are valuable resources for molecular tests in the absence of formalin-fixed paraffin-embedded (FFPE) tissues [8‒12]. Furthermore, the advantage of using cytologic specimens for DNA testing is that they have not undergone formalin-induced protein and nucleic acid crosslinking, and the nuclei of cells on cytologic slides are intact, while those in FFPE tissues are transected during microtome sectioning.
Cytologic slide preparation requires sample acquisition, smearing slides, fixation, staining, mounting, and coverslipping , and each step has several options and affects the quality of DNA extracted from slides. Fixation can be performed by spraying with a cytofixative or submerging in ethanol or a commercial agent such as CytoRich Red Preservation solution, and slides can then be stained with Giemsa, hematoxylin and eosin (H&E), or Papanicolaou (Pap) stain. Lastly, slides can be sealed with mounting medium and coverslips.
Although several authors have noted that DNA extracted from cytology samples is suitable for various molecular tests [6‒10], we focused on cytology slides, which offer a more practical source of DNA. Cell blocks were not included because they were prepared from fine needle aspiration. We also tested the most commonly used fixatives, that is, ethanol and CytoRich Red Preservation solution, and compared results with unfixed samples. H&E, Pap, and Giemsa stains were also compared to unstained samples. Touch-imprinted preparations from papillary thyroid carcinoma and adenocarcinoma obtained by thyroidectomy and colectomy were used to standardize starting materials. BRAF or KRAS mutations were investigated and compared with the results of matched FFPE tissues. These genes were chosen because they are detected at relatively high rates in papillary thyroid carcinoma and colorectal adenocarcinoma, respectively, and the testing involved is straightforward.
Materials and Methods
Tissues (2 mm3) were obtained as soon as specimens had been received in our pathology department and inspected by a pathologist. Tissue fragments were gently dabbed onto glass slides to cover at least 60% of slide surfaces. Six touch-imprinted cytologic slides were prepared for each tumor sample, and remnant tissues were blocked for paraffin embedding (Fig. 1). One of the 6 slides was air-dried and stored, one was immersed in 95% ethanol, and another in CytoRich Red Preservative solution and stored until required. The other three slides were stained with Giemsa, Pap, or H&E after 12 h of 95% ethanol fixation. Slides were grouped by preparation method and stored under ambient conditions for 1 week as follows (Fig. 1): group 1, air-dried; fixation group 2A, 95% ethanol; group 2B, CytoRich Red preservation solution; stained group 3A, Giemsa; group 3B, Pap; group 3C, H&E; and group 4, FFPE tissue. The slides in group 3 were digitalized using Pannoramic 250 Flash III scanner (3DHistech, Budapest, Hungary). This study was approved by the Institutional Review Board of Pusan National University Yang-san Hospital.
Slides of group 2A were air-dried, and slides of group 2B were washed in 99% ethanol and air-dried. Slides of group 3 (stained) were immersed in xylene for 1 h, coverslips were removed, and slides were immersed in xylene for 5 min to remove remaining mounting solution and then briefly immersed in 99% alcohol and air-dried. A small amount of lysis buffer was then dropped on all slides of groups 1–3. Cells were then scraped off using a sterile pipette tip into 1.5-mL tubes and incubated in lysis buffer containing proteinase K for an hour at 56°C. Finally, DNA was extracted using a Maxwell 16 Cell DNA Purification Kit (Promega, USA). For FFPE tissues of group 4, 5–10 sections (10 µm) were placed in 1.5-mL tubes, treated with 20 µL proteinase K and 180 µL incubation buffer overnight at 72°C, and 400 µL lysis buffer was then added. Finally, DNA was extracted using Maxwell 16 FFPE Tissue LEV DNA Purification Kits (Promega, USA). The concentrations and purities of DNA solutions were determined using a NanoDrop-2000 (Thermo Fisher Scientific, Wilmington, DE, USA). DNA Integrity Number (DIN) values were determined using an Agilent TapeStation and Genomic DNA ScreenTape (Agilent Technologies, Santa Clara, CA, USA). Mutational analyses for KRAS (codon 12 and 146) and BRAF (V600) genes were performed using real-time quantitative PCR with peptide nucleic acid (PNA)-mediated clamping using PNAClamp KRAS and BRAF Mutation Detection kits (PANAGENE, Daejeon, Korea), as previously described .
Of the 23 samples, 14 were from males and 9 were from females. Donor ages ranged from 39 to 78 years (mean, 59 years). There were 11 cases of papillary thyroid carcinoma and 12 cases of colorectal adenocarcinoma. Six imprinted slides and one FFPE tissue block were prepared from each sample (Fig. 1). Colorectal adenocarcinoma tissues were soft and easily imprinted, while papillary thyroid carcinoma tissues were firm and adhered less well to slides. Smears from colon cancer were thicker and clumped more than those from papillary thyroid carcinoma, which produced thinly spread cells on slides (Fig. 2).
For air-dried slides of group 1, the concentration of DNA was 137.5 ± 132.4 ng/µL. For unstained slides in the 95% ethanol (group 2A) or CytoRich Red Preservative solution (group 2B), mean DNA concentrations were 118.2 ± 88.9 ng/µL and 26.4 ± 18.2 ng/µL, respectively. The mean concentrations of DNA extracted from stained slides of group 3 were 80.6 ± 55.3 ng/µL (group 3A), 73.0 ± 34.3 ng/µL (group 3B), and 80.2 ± 56.9 ng/µL (group 3C), respectively. FFPE tissues from group 4 had a mean DNA concentration of 75.0 ± 70.2 ng/µL. Cells on group 3 slides were more strongly adherent than unstained cells. The A260/A280 ratios of DNA extracted from groups 1, 2A, and 3A were ≥1.8. Mean DINs of groups 1, 2A, and 3A were 8.1, 8.4, and 7.2, respectively, and gel images of DIN value exhibited clear bands. In groups 3B and 3C, DNA was degraded (DIN values were 4.2 and 1.8, respectively), and bands were frequently smeared. In group 2B, DIN was not measurable due to extensive degradation. Mean DIN value in group 4 was 3.5, and bands were indistinct. Results are summarized in Table 1 and Figure 3.
Twelve of the 23 cases were found to harbor mutation in the BRAF or KRAS genes when analyzed using FFPE tissues (Table 2). Matched DNAs extracted from cytologic preparations were used for mutation testing, except for case 3 in groups 3A and 3B, for which tumor cells were insufficient. KRAS or BRAF mutation was detected in all groups 1, 2A, and 3 samples, and results matched FFPE results. In group 2B, mutation was detected in 4 of 12 cases; amplification failed for the others.
In addition, Ct values determined by non-PNA probe real-time PCR were compared to evaluate DNA quality. All groups except group 2B had appropriate values. Ct values generally corresponded with DIN values and indirectly reflected DNA quality (Fig. 4).
In this era of precision medicine for cancer treatment, various molecular tests are being implemented in daily practice [1‒3], and not uncommonly, multiple molecular tests are performed on single small biopsy specimens, which means appropriate amounts of nucleic acid must be obtained for successful testing . For example, recurrent cancers after curative surgery may arise in challenging locations, and securing adequate samples may be difficult, and cancers that recur after chemotherapy or radiotherapy may contain few cancer cells. Liquid biopsy provides an alternative means of detecting mutations in blood, but success rates vary and depend on tumor volume . On the other hand, cytologic specimens can be obtained less invasively and repeatedly and can be prepared from biopsy samples using the touch-imprint technique. We considered that if DNA of appropriate quality could be extracted from cytology slides, it might be suitable for molecular tests.
In the present study, touch-imprint cytology slides were prepared from papillary thyroid carcinoma and colorectal adenocarcinoma tissues obtained from thyroidectomy or colectomy specimens. Slides were grouped 1 to 3 according to the fixation and staining methods used; samples of groups 1 and 2 were unstained, and those of group 3 were stained. Unstained groups were tested to evaluate the effect of fixation. Group 1 (unfixed) contained DNA of unexpectedly good quality and at higher concentrations than the other groups. Its DIN value was 8.13, which was second only to ethanol-fixed unstained slides. This result indicates that cells on slides maintained their DNA integrity at least for several days under ambient conditions but that long-term storage would be inappropriate. Slides of group 2 were fixed in ethanol (group 2A) or CytoRich Red Preservation solution (group 2B) for a week to test the effect of fixatives. While the qualities of DNA extracted from group 2A were similar to those of group 1, group 2B DNA was markedly deteriorated, which was probably due to prolonged fixation, as it has been reported that the quality of DNA in CytoRich Red Preservation solution decreases gradually with fixation time [15‒17] and that the formalin in CytoRich Red Preservation solution may reduce DNA integrity . The stained slides of group 3 showed a ca. 40% reduction in DNA concentration versus groups 1 and 2A. We suspect that the lower DNA concentration obtained for group 3 was due to loss of cells during coverslip removal. DIN values of group 3 were significantly different, and the DIN value was greatest (at 7.2) in group 3A (the Giemsa-stained group) and lowest (at 1.8) in group 3C (the H&E group). The reason why DNA extracted from H&E-stained slides had a low DIN value is probably associated with xylene, as slides were soaked in xylene for 5 s, 30 s, and 4 min for Giemsa, Pap, and H&E staining, respectively.
To investigate how successfully PCR reactions were performed, Ct values of non-PNA probes for BRAF or KRAS that amplify target regions regardless of gene mutation status were calculated. We found that a DIN value of >1.0 predicted successful PCR reaction with Ct numbers <30 (Fig. 4). However, 260/280 purity ratio was less reliable for predicting Ct values. For example, a sample in group 2B had a 260/280 ratio of 1.77 and a Ct value >30, but in other groups, all samples with a 260/280 ratio of ≤1.60 had Ct values <30. In daily practice, DIN values are not determined for all samples because of the associated cost, and thus, 260/280 ratio is a more accessible quality metric. Pathologists and laboratory technicians should bear in mind that 260/280 ratio only provides an estimate of nucleic acid quantity with respect to other substances in solution and is not an indicator of DNA integrity, which is crucial for successful molecular testing. If the DIN value cannot be measured, Ct provides a possible means of checking the qualities of DNA extracted from cytologic preparations.
Molecular testing is usually performed during the diagnostic process. Cytologic specimens are frequently obtained at the same time as tissue biopsy specimens which are sometimes inadequate for molecular testing. Thus, cytologic specimens may be a valuable source of DNA. We evaluated the quality of DNA stored for a week in a variety of cytologic preparations. Ethanol fixation was found to be much better for molecular testing than CytoRich Red Preservation solution, in which DNA quality deteriorated within 24 h [15‒17]. Nonfixated slides were also acceptable, but in practice, further fixation and staining are necessary to evaluate the cellularities of cancer cells. As regards to staining, Giemsa staining was superior to Pap or H&E staining in terms of DNA quality and concentration, which may have been due to different adherences to slides or xylene soak times. Although DNA qualities varied, mutation tests worked well on most samples, which indicated the staining methods examined are suitable for routine molecular testing.
Recently, the use of next-generation sequencing (NGS) has increased rapidly in clinical practice, and this technique requires greater amounts of higher DNA quality. Several authors have recommended the use of cytological slides for NGS [18, 19]. The limitation of our study is that we tested only two genes by PCR. For further NGS-based assay, vigorous validation will be necessary . Moreover, due to recent advances in digital pathology, cytologic slides can be freely utilized without being archived in themselves. Therefore, they can be digitalized using a slide scanner before DNA extraction for future review.
This study confirms that cytologic slides that have undergone appropriate processing are good sources of DNA for real-time PCR-based tests. Ethanol fixation was excellent for DNA preservation, while CytoRich Red preservation solution was inappropriate for extended storage. Giemsa staining was associated with better DNA quality than Pap or H&E staining. We suggest that cytologic samples should be considered whenever tissue samples are inadequate for molecular testing.
Statement of Ethics
This study was approved by the Institutional Review Board of Pusan National University Yang-san Hospital (approval number 05-2020-046) and was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.
Conflict of Interest Statement
The authors declare no conflicts of interest.
This work was supported by a 2-Year Research Grant of Pusan National University.
Hyo Sum Kwon conducted the experimental work and manuscript preparation. Dong Hoon Shin designed and supervised the experiment. So Young Kim, Chung Su Hwang, and Arhong Kim analyzed data. Hyun Jung Lee, Joo Young Na, Jung Hee Lee, and Jee Yeon Kim critically reviewed the manuscript, and all authors approved the final version of the manuscript.
Data Availability Statement
All data analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.