Obtaining high-quality genomic DNA from tumor tissues is an essential first step for analyzing genetic alterations using next generation sequencing. In this article, we present a simple and rapid method to enrich tumor cells and obtain intact DNA from touch imprint cytology specimens.
It is critical to determine the mutational status in cancer before administration and treatment of specific molecular targeted drugs for cancer patients. In the clinical setting, formalin-fixed paraffin-embedded (FFPE) tissues are widely used for genetic testing. However, FFPE DNA is generally damaged and fragmented during the fixation process with formalin. Therefore, FFPE DNA is sometimes not adequate for genetic testing because of low quality and quantity of DNA. Here we present a method of touch imprint cytology (TIC) to obtain genomic DNA from cancer cells, which can be observed under a microscope. Cell morphology and cancer cell numbers can be evaluated using TIC specimens. Furthermore, the extraction of genomic DNA from TIC samples can be completed within two days. The total amount and quality of TIC DNA obtained using this method was higher than that of FFPE DNA. This rapid and simple method allows researchers to obtain high-quality DNA for genetic testing (e.g., next generation sequencing analysis, digital PCR, and quantitative real time PCR) and to shorten the turnaround time for reporting results.
Next generation sequencing technology has provided researchers significant advancements in analyzing genome information in genetic variations, Mendelian disease, hereditary predisposition, and cancer 1,2,3. The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) have pursued the identification of genetic alterations in several types of common cancers4. Hundreds of essential cancer driver genes have been successfully identified, and some of these molecules are being targeted for drug development1,5,6.
In the clinical setting, FFPE specimens are commonly used for pathological diagnosis and molecular testing for various diseases, including cancer. However, during the fixation process with formalin, DNA-protein or DNA-DNA cross-linking occurs and DNA fragmentation is induced. Thus, FFPE DNA samples are not always suitable for genetic analysis because of low quality and quantity of DNA7,8,9. Additionally, it takes several days to prepare FFPE specimens, and technical skill is necessary to accurately prepare the sections. Therefore, it is desirable to develop a simple and rapid method for obtaining high-quality intact DNA.
Cytology is an alternative method for pathological diagnosis. Cytological sample preparation is a simpler, less expensive, and more rapid approach compared with FFPE preparation10. The TIC technique has been performed on sentinel lymph nodes and marginal tissues from breast cancer patients for intraoperative rapid diagnosis for some years11,12. However, there are few reports that have examined whether high-quality genomic DNA can be extracted from TIC specimens and used for subsequent genetic analysis. Cytological specimens are commonly stained with Papanicolaou (Pap) or Giemsa staining, and we previously reported that the amount and quality of DNA extracted from TIC specimens (especially Giemsa-stained samples) are superior to samples obtained from FFPE tissues13. Compared with Pap staining, Giemsa staining has an advantage in requiring less staining procedures. In Pap staining, after the samples have been fixed and stained, they must be mounted with mounting medium (e.g., Malinol) for distinguishing sample contents, such as tumor cells, normal cells, and inflammatory cells under a microscope. If the Pap specimen is prepared without the mounting step, it is almost impossible to observe the cells under a microscope because the specimen is dried. In comparison, Giemsa staining can be observed in the dried state, therefore, the mounting step is not necessary for quick cellular evaluation. For microdissection, Giemsa staining is more suitable because it requires dry specimens.
In this report, we introduce a simple and rapid method for preparing TIC specimens with Giemsa staining and demonstrate that TIC is a better source for DNA compared with FFPE specimens.
1. TIC Preparation for Quick Microscopic Assessment Using Normal Glass Slides
2. Preparation of the PEN Membrane Slide Film for Genetic Testing
3. DNA Extraction
4. Estimation of DNA Quality by Quantitative Real Time PCR
5. Preparation of the Next Generation Sequencing Library
6. Quantify the Library Concentration by Quantitative Real Time PCR
7. Next Generation Sequencing
Figure 1 shows the entire process from preparing TIC specimens to DNA extraction. Notably, the procedure takes only two days to obtain genomic DNA from TIC samples. We evaluated any effects of tumor storage before the slide processing. We found that tumor cells were attached onto the glass slide when tissue specimens were immediately touched onto the slide, and when tissues were kept in saline moistened sterile-gauze for 1 h (Figure 2). However, when tissues were kept at room temperature, the tumor cells were not well attached to the slide. Prepared slides could be stored at 4 °C for three months. Therefore, it is important not to allow the specimens to dry.
We examined the utility of our method and compare it with specimens acquired using FFPE. TIC and FFPE samples were prepared from 14 tumor specimens, and we confirmed that tumor contents and purity could be assessed from the TIC specimens. After we performed Giemsa staining, the number of tumor cells and the tumor morphology could be assessed by microscopy. We were thus able to routinely evaluate the tumor cells and subsequently perform a DNA quality check.
The DNA quantity and quality were estimated by quantitative real time PCR13,14,15. We determined the absolute DNA quantities and the RQ value, which is an indicator of the degradation level of genomic DNA. The results showed that a higher DNA yield was achieved using the TIC specimens compared with the FFPE specimens (Table 1). In addition, the RQ values of the TIC DNA were significantly higher compared with that of the FFPE DNA (Figure 3, p = 2.3 x 10-8, two-tailed Student's t tests). We also assessed the RQ values of the TIC and FFPE DNA extracted from different tumor types and found that the TIC DNA was higher in quality compared to the FFPE DNA (Figure 3). These results indicated that the TIC DNA was less fragmented than the FFPE DNA.
We next assessed whether the TIC DNA could be used for next generation sequencing analysis. FFPE and TIC DNA were prepared from primary colorectal cancer and metastatic liver cancer obtained from a patient (Figure 4A). PCR amplification and library preparation were conducted using the Cancer Hotspot Panel and subsequently targeted sequencing was performed. As a result, APC Q1367* was identified in both FFPE and TIC DNA extracted from Site 1 (Table 2). Furthermore, APC S1356*, KRAS G12D, and TP53 M237I were detected in both FFPE and TIC DNA extracted from Site 2, 3, and 4 (Table 2). These results suggested that the identical somatic mutations were identified in paired FFPE and TIC DNA samples prepared from the same tumor site (Figure 4B and Table 2). Notably, the same somatic mutations were detected between primary colorectal cancer (Site 2, not Site 1) and two metastatic liver cancer samples, suggesting that the tumor clones from Site 2 metastasize to liver (Figure 4B and Table 2). Together these findings suggest that the TIC DNA is high-quality and is suitable for a wide range of genetic testing including next generation sequencing.
Figure 1.Schematic for obtaining DNA from a TIC sample preparation. The tumor tissue is touched on the slide glass to prepare the TIC sample. After assessment of the tumor morphology and contents under a microscope, the tumor DNA can be extracted and used for genetic testing. By using TIC, the tumor DNA can be obtained from a clinical pathological specimen within two days. Please click here to view a larger version of this figure.
Figure 2. Preparation of TIC samples. Resected tumor specimens were immediately touched onto a normal glass slide (left panel), kept in saline moistened sterile-gauze for 1 h (middle), and kept at room temperature for 1 h (right). Sample #1 was hepatocellular carcinoma and sample #2 was breast cancer. Microscopic pictures were captured using a digital camera mounted to a microscope. Scale bar: 100 µm. Please click here to view a larger version of this figure.
Figure 3. DNA quality check estimated by quantitative real-time PCR analysis. TIC and FFPE DNA were extracted from 14 tumor tissues from colorectal (n = 8), stomach (n = 4), and metastatic liver cancer (n = 2). Comparison of the relative quantification scores between TIC-Giemsa and FFPE-HE samples. RQ values of TIC DNA were significantly higher than that of FFPE DNA. Statistical analysis between the two groups was performed and p-values were calculated by unpaired two-tailed Student's t test using Excel. Please click here to view a larger version of this figure.
Figure 4. Next generation sequencing analysis data using TIC and FFPE DNA. (A) Macroscopic and microscopic images. Representative images of TIC-Giemsa and FFPE-HE staining from the colorectal cancer (site 1 and 2) and metastatic liver carcinoma samples (site 3 and 4). Scale bar in macroscopic images: 1 cm, Scale bar in microscopic images: 100 µm. (B) Heat map showing the distribution of somatic mutations for each tumor site (n = 8). Identical mutations were detected among paired TIC and FFPE DNA samples. Values of allelic fractions are indicated in graduation color scale from 1% (light pink) to 100% (pink). Gray columns showed no identified mutation. Please click here to view a larger version of this figure.
TIC-Giemsa (n=14) | FFPE-HE (n=14) | |||||||||
Total DNA (ng) | Total DNA (ng) | |||||||||
Sample | Tumor site | Short | Long | RQ | Short | Long | RQ | |||
Site1 | Colon | 1495 | 1247 | 0.83 | 939 | 349 | 0.37 | |||
Site2 | Colon | 991 | 1057 | 1.07 | 556 | 204 | 0.37 | |||
Site3 | Liver | 467 | 511 | 1.09 | 130 | 39 | 0.3 | |||
Site4 | Liver | 2172 | 2115 | 0.97 | 488 | 127 | 0.26 | |||
Site5 | Colon | 749 | 598 | 0.8 | 529 | 205 | 0.39 | |||
Site6 | Stomach | 330 | 286 | 0.86 | 211 | 98 | 0.46 | |||
Site7 | Stomach | 636 | 499 | 0.78 | 154 | 84 | 0.55 | |||
Site8 | Stomach | 27 | 27 | 1.01 | 135 | 81 | 0.6 | |||
Site9 | Colon | 1986 | 1611 | 0.81 | 476 | 163 | 0.34 | |||
Site10 | Colon | 280 | 218 | 0.78 | 209 | 83 | 0.39 | |||
Site11 | Colon | 1546 | 575 | 0.37 | 366 | 159 | 0.43 | |||
Site12 | Stomach | 1501 | 1200 | 0.8 | 274 | 132 | 0.48 | |||
Site13 | Colon | 1556 | 1404 | 0.9 | 326 | 179 | 0.55 | |||
Site14 | Colon | 1565 | 1210 | 0.77 | 680 | 295 | 0.43 | |||
Mean±SD | 1093±682 | 897±600 | 0.85±0.18 | 391±253 | 157±86 | 0.42±0.10 |
Table 1. DNA quality data. Paired TIC (n = 14) and FFPE specimens (n = 14) were prepared from colon, liver, and stomach cancers. TIC samples were stained with Giemsa, and FFPE samples were stained with hematoxylin and eosin (HE). DNA samples were extracted and quantified by quantitative real-time PCR with two primer pairs amplifying RNaseP locus (long amplicon (268 bp) and short amplicon (87 bp)). RQ values were calculated as follow: the mean value of long amplicon divided bythe mean value of short amplicon. SD, standard deviation; RQ, relative quantitation
Sample Name | Location | Preparation | Gene symbol | Mutation | Position | Reference | Variant | Coding | Coverage | Allelic fraction | |
Primary colorectal cancer | Site 1 | FFPE | APC | Q1367* | chr5:112175390 | C | T | c.4099C>T | 1999 | 45% | |
Primary colorectal cancer | Site 1 | TIC | APC | Q1367* | chr5:112175390 | C | T | c.4099C>T | 1011 | 72% | |
Primary colorectal cancer | Site 2 | FFPE | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1968 | 77% | |
Primary colorectal cancer | Site 2 | FFPE | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1993 | 52% | |
Primary colorectal cancer | Site 2 | FFPE | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1991 | 73% | |
Primary colorectal cancer | Site 2 | TIC | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1967 | 89% | |
Primary colorectal cancer | Site 2 | TIC | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1994 | 56% | |
Primary colorectal cancer | Site 2 | TIC | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1995 | 86% | |
Metastatic liver cancer | Site 3 | FFPE | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1974 | 92% | |
Metastatic liver cancer | Site 3 | FFPE | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1995 | 62% | |
Metastatic liver cancer | Site 3 | FFPE | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1296 | 91% | |
Metastatic liver cancer | Site 3 | TIC | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1964 | 97% | |
Metastatic liver cancer | Site 3 | TIC | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1993 | 64% | |
Metastatic liver cancer | Site 3 | TIC | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1998 | 95% | |
Metastatic liver cancer | Site 4 | FFPE | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1965 | 93% | |
Metastatic liver cancer | Site 4 | FFPE | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1992 | 60% | |
Metastatic liver cancer | Site 4 | FFPE | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1993 | 92% | |
Metastatic liver cancer | Site 4 | TIC | APC | S1356* | chr5:112175358 | C | A | c.4067C>A | 1960 | 94% | |
Metastatic liver cancer | Site 4 | TIC | KRAS | G12D | chr12:25398284 | C | T | c.35G>A | 1992 | 62% | |
Metastatic liver cancer | Site 4 | TIC | TP53 | M237I | chr17:7577570 | C | A | c.711G>T | 1995 | 95% |
Table 2. Comparison of mutations in paired FFPE and TIC DNA detected by targeted sequencing analysis. Paired TIC and FFPE samples were prepared from 2 primary colorectal cancers (Site 1 and 2) and 2 metastatic liver cancers (Site 3 and 4). Targeted sequencing was performed with these DNA samples and somatic mutations were identified. Mutation profiles were identical between paired TIC and FFPE DNA samples.
In this study, we presented an alternative method for obtaining tumor DNA from clinical pathological specimens using TIC. TIC preparation is very simple and needs less time compared with FFPE methods, without the requirement for special instruments10. All procedures from the TIC preparation to DNA extraction can be completed within two days (Figure 1). This method thus shortens the turnaround time for performing genetic testing. Notably, this provides a significant advantage in shortening the number of days required for molecular analysis. This short turnaround time allows us to immediately offer an appropriate therapy for progressive cancer patients who require molecularly targeted drugs. This method thus provides benefits for analyses of genetic alterations in tumors and administration of molecularly targeted drugs for cancer therapy.
There are some key points for the success of using TIC DNA for genetic analysis. Tumor tissues may be necrotic because of treatment such as chemotherapy or other treatments. Thus, caution is needed in preparing TIC specimens to avoid sampling from necrotic sections in the tumor as much as possible. Further, initial microscopic assessment is very important for subsequent procedures. In addition, preventing the drying of tumor tissues is required for successful TIC sample preparation. If the tumor tissues are dried, this results in fewer attached cells on the glass slide. It will also be difficult to observe the tumor cellularity and morphology because drying leads to degeneration of tumor tissues.
There are some potential benefits to using TIC DNA. First, we can check the tumor cellularity during the first assessment with microscopy. If normal cells, such as lymphocytes and stromal cells, were abundant in the tissue samples, the contamination of these normal cells would prevent the ability to detect somatic mutations in the tumor cells. Quick microscopic assessment of the samples may help the evaluation of the tumor cellularity and estimation of whether adequate tumor DNA samples were obtained from the TIC samples before DNA extraction. Second, our results showed that the TIC DNA is both high in quality and quantity. FFPE DNA has been used for next generation sequencing analysis including targeted sequencing, exome sequencing, and whole genome sequencing16,17,18,19,20. Archival FFPE DNA is also routinely used for genetic analysis21,22, however FFPE DNA can be fragmented during formalin fixation. This can result in problems in PCR amplification or DNA extraction, causing a lack of sequence coverage, uniformity at target regions, and increase risk of sequence error23,24. In contrast, TIC DNA is not fragmented, which may be due to the alcohol fixation that has less of an influence on the nucleic acid. As TIC DNA is not fragmented like the FFPE DNA, these samples will be more suitable for next generation sequencing analysis, real-time PCR, and digital PCR. Indeed, we previously showed that TIC DNA from a tumor specimen can be used in next-generation sequencing analysis, a method called "TIC-seq", and the results could precisely capture the tumor somatic mutations13. Third, this technique is applicable for a wide range of materials. In the current report, we used surgical specimens. In addition to surgical tissues, metastatic lymph node, bronchial or endoscopic biopsy can be used for preparing TIC DNA.
In conclusion, we present a simple and rapid method for preparing high-quality tumor DNA using TIC samples. This method will expand new possibilities in the field of genetic analysis and help promote precision medicine in the clinical setting.
The authors have nothing to disclose.
We thank all the medical and ancillary staff of the hospital and the patients for consenting to participate. We thank Gabrielle White Wolf, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this report. This study was supported by a Grant-in-Aid for Genome Research Project from the Yamanashi Prefecture (Y.H. and M.O.) and a grant from The YASUDA Medical Foundation (Y.H.).
FINE FROST white20 micro slide glass | Matsunami Glass ind, Ltd | SFF-011 | |
Arcturus PEN Membrane Glass Slides | Thermo Fisher Scientific | LCM0522 | |
Cyto Quick A solution | Muto Pure Chemicals | 20571 | |
Cyto Quick B solution | Muto Pure Chemicals | 20581 | |
May-Grunwald Solution | Muto Pure Chemicals | 15053 | |
Giemsa solution | Muto Pure Chemicals | 15002 | |
QIAamp DNA FFPE tissue kit | Qiagen | 56404 | |
TaqMan Fast Advanced Master Mix | Thermo Fisher Scientific | 4444557 | |
TaqMan RNase P Detection Reagents Kit | Thermo Fisher Scientific | 4316831 | |
TaqMan Assay from FFPE DNA QC Assay v2 | Thermo Fisher Scientific | 4324034 | |
MicroAmp Fast Optical 96-Well Reaction Plate | Thermo Fisher Scientific | 4346907 | |
MicroAmp optical Adhesive Film | Thermo Fisher Scientific | 4311971 | |
MicroMixer E36 | TITEC | 0027765-000 | |
ViiA 7 Real-Time PCR System | Thermo Fisher Scientific | VIIA7-03 | |
Himac CF16RXII | Hitachi-koki | CF16RII | |
Ion Library TaqMan Quantitation Kit | Thermo Fisher Scientific | 4468802 | |
Ion AmpliSeq Cancer Hotspot Panel v2 | Thermo Fisher Scientific | 4475346 | |
Ion AmpliSeq Library Kit 2.0 | Thermo Fisher Scientific | 4480442 | |
Ion Xpress Barcode Adapters 1-16 Kit | Thermo Fisher Scientific | 4471250 | |
Ion PGM Hi-Q View Sequencing Kit (200 base) | Thermo Fisher Scientific | A30044 | |
Ion Chef System | Thermo Fisher Scientific | 4484177 | |
Veriti 96-well Thermal Cycler | Thermo Fisher Scientific | Veriti200 | |
Ion 318 Chip Kit v2 BC | Thermo Fisher Scientific | 4488150 | |
Ion PGM System | Thermo Fisher Scientific | PGM11-001 | |
Ion PGM Wash 2 Bottle kit | Thermo Fisher Scientific | A25591 | |
Agencourt™ AMPure™ XP Kit | Beckman Coulter | A63881 | |
16-position Magnetic Stand | Thermo Fisher Scientific | 4457858 | |
Nonstick, RNase-free Microfuge Tubes, 1.5 mL (Low binding tube) | Thermo Fisher Scientific | AM12450 | |
Nuclease-free water | Thermo Fisher Scientific | AM9938 | |
MicroAmp™ Optical 96-well Reaction Plates | Thermo Fisher Scientific | 4306737 | |
MicroAmp™ Clear Adhesive Film | Thermo Fisher Scientific | 4306311 | |
Agencourt™ AMPure™ XP Kit | Beckman Coulter | A63881 | |
Ethanol(99.5) | Nacalai Tesque | 08948-25 | |
Sodium hydroxide (10M) | Sigma | 72068 | |
DTU-Neo | TAITEC | 0063286-000 | |
E-36 | TAITEC | 0027765-000 | |
ECLIPSE Ci-L | Nikon | 704354 | |
Pipet-Lite LTS Pipette L-2XLS+ | METTLER TOLEDO | 17014393 | |
Pipet-Lite LTS Pipette L-10XLS+ | METTLER TOLEDO | 17014388 | |
Pipet-Lite LTS Pipette L-20XLS+ | METTLER TOLEDO | 17014392 | |
Pipet-Lite LTS Pipette L-100XLS+ | METTLER TOLEDO | 17014384 | |
Pipet-Lite LTS Pipette L-200XLS+ | METTLER TOLEDO | 17014391 | |
Pipet-Lite LTS Pipette L-1000XLS+ | METTLER TOLEDO | 17014382 | |
petit-change | WAKEN | MODEL8864 | Mini centrifuge |
petit-incubator | WAKEN | WKN-2290 | Air incubator |
SensiCare Powder-free Nitrile Exam Gloves | MEDLINE | SEM486802 | |
Sterile gauze | Osaki | 11138 | |
Refrigerator MediCool | SANYO | MPR-312DCN-PJ | |
FEATHER TRIMMING BLAD | FEATHER | No.130 | |
FEATHER TRIMMING BLAD | FEATHER | No.260 | |
FEATHER S | FEATHER | FA-10 | |
Vortex Genius 3 | IKA | 41-0458 | Vortex mixer |
Pincette | NATSUME | A-5 | |
1.5 mL microtube | BIOBIK | RC-0150 |