Genomic DNA extraction methods using formalin-fixed paraffin-embedded tissue

Accepted Manuscript Genomic DNA Extraction Methods Using FFPE Tissue Keerti Potluri, Ahmed Mahas, Michael N. Kent, Sameep Naik, Michael Markey PII: DOI: Reference:

S0003-2697(15)00322-X http://dx.doi.org/10.1016/j.ab.2015.06.029 YABIO 12128

To appear in:

Analytical Biochemistry

Received Date: Revised Date: Accepted Date:

8 April 2015 5 June 2015 23 June 2015

Please cite this article as: K. Potluri, A. Mahas, M.N. Kent, S. Naik, M. Markey, Genomic DNA Extraction Methods Using FFPE Tissue, Analytical Biochemistry (2015), doi: http://dx.doi.org/10.1016/j.ab.2015.06.029

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Title: Genomic DNA Extraction Methods Using FFPE Tissue Subject category: DNA Recombinant Techniques and Nucleic Acids Authors: Keerti Potluri1, Ahmed Mahas1, Michael N Kent2,3, Sameep Naik2, Michael Markey1 (corresponding author) Affiliations: 1Wright State University, Department of Biochemistry and Molecular Biology, 2Wright State University Boonshoft School of Medicine, Department of Dermatoology3Dermatopathology Laboratory of Central States Address: 3640 Colonel Glenn Hwy, Dayton, OH 45435; 7835 Paragon Road Dayton, OH 45459 Address to which proofs should be mailed: 3640 Colonel Glenn Hwy, Dayton, OH 45435 Corresponding author contact: Phone (937) 775-4536, Fax (937) 775-4494, [email protected]

1

Abstract: As new technologies come within reach for the average cytogenetic laboratory, the study of chromosome structure has become increasingly more sophisticated. Resolution has improved from karyotyping (in which whole chromosomes are discernible), to fluorescence in situ hybridization and comparative genomic hybridization (CGH, with which specific megabase regions are visualized), arraybased CGH (aCGH, examining hundreds of base pairs), and next-generation sequencing (providing single base pair resolution). Whole genome next generation sequencing remains a cost-prohibitive method for many investigators. Meanwhile, the cost of aCGH has been reduced in recent years, even as resolution has increased and protocols have simplified. However, aCGH presents its own set of unique challenges. DNA is required of sufficient quantity and quality to hybridize to arrays and provide meaningful results. This is especially difficult for DNA from formalin-fixed paraffin-embedded (FFPE) tissues. Here, we compare three different methods for acquiring DNA of sufficient length, purity, and “amplifiability” for aCGH and other downstream applications. Phenol-chloroform extraction and column-based commercial kits were compared to Adaptive Focused Acoustics (AFA). Of the three extraction methods, AFA samples showed increased amplicon length and decreased PCR failure rate. These findings support AFA as an improvement over previous DNA extraction methods for FFPE tissues.

2

Introduction For studies of human disease, it is critical to examine actual human tissues representative of the disease state. These materials are often difficult for the researcher to acquire for a number of reasons. First, recruiting patients to donate tissue is a complex process that requires the cooperation of surgeons, pathologists, patients, and oversight by institutional review boards (IRBs). Second, some conditions are rarer and require a long time to accumulate prospectively. Additionally, any longitudinal study will require clinical data that may not be available for some time, such as how the subject responded to treatment. Fortunately, a great wealth of human biospecimens exists in the archives of many hospitals, tissue banks, and pathology laboratories in the form of formalin-fixed paraffin-embedded (FFPE) tissue blocks. In many cases, banked FFPE specimens can solve some of these difficulties. The College of American Pathologists requires specimens to be retained for 10 years following diagnosis [1], thus retrospective longitudinal data can be available. Rarer specimens can be accumulated over time. Importantly, these specimens lend themselves to de-identification, which streamlines the IRB process. Therefore, FFPE tissue banks can be a ready resource for researchers, if only we can unlock useful materials from the preserved specimens. While FFPE tissues are stable for decades [2], this preservation presents significant challenges. Formalin treatment cross-links biological molecules such as DNA and proteins [3]. Additionally, longer nucleic acids like DNA and RNA, in particular, are fragmented in the preservation process, leading to poor performance in downstream analyses [4, 5]. To optimize utility of nucleic acids from FFPE specimens for aCGH, we must reverse these cross links and avoid further degradation during the DNA or RNA extraction process. Several previous studies have compared protocols for DNA extraction from FFPE tissues, with varying results. Senguven et al. compared several variations on the CTAB method and a hot alkali method with 3

the Qiagen FFPE DNA kit [6]. They found that the Qiagen columns performed best, and that DNA yield became especially low in samples more than 50 years old. There is some controversy between laboratories, however, as other studies have compared the Qiagen kit with phenol-chloroform and found either the columns [7] or the phenol method [8]to be superior in yield and amplifiability. Other methods less frequently usednot measured in our study include the Maxwell 16 method by Promega (Mannheim, Germany), which was shown to perform well against other automated Qiagen kits [9], and the Chelex bead method (Bio-Rad, Berkeley, California) which has been used with some success in manual or automated extractions from FFPE tissues [10]. A small number of studies have shown that the DNA obtained by a recent sonication-based method called adaptive focused acoustics (AFA) is more amenable to next generation sequencing than DNA from other methods, resulting in greater coverage of the genome and facilitating easier assembly and alignment [11-13]. To our knowledge, ours is the first study to compare AFA for use in aCGH. Here, we compare three methods for extracting quality DNA from FFPE tissues: phenol-chloroform extraction, commercial column-based purification, and AFA extraction. Metrics of quantity and quality are considered for each method. Materials and Methods Specimen collection and preparation Following IRB review, specimens were selected from a large archive of FFPE skin biopsies collected at a national dermatopathology laboratory (Dermatopathology Laboratory of Central States, DLCS, Dayton, OH). Patients ranged in age from 16 to 75 years. The biopsy specimens were collected in 2007 or 2004, making them 8 or 11 years old. Specimens were stored in a temperature-controlled environment. De-identified retrospective clinical data were obtained from clinical databases and patient health records at DLCS. For all experiments, 10 µm thick sections were taken for each sample from paraffin blocks by using a microtome with disposable blades. Care was taken to avoid contamination between 4

the specimens by wearing gloves when handling the blocks, cleaning the microtome after cutting a block, and using fresh blades for each specimen. The sections were placed in a warm water bath and then mounted on slides. Slides were air dried. The first and last slides from each block were stained with H&E to verify that the region of interest (consisting of cellular matieral) had not been exhausted. Tissue was then scraped from slides using sterile scalpel blades into 1.5 mL microcentrifuge tubes for DNA extraction. The number of sections taken per sample varies between methods, as described below. DNA Isolation from FFPE tissues DNA isolation was carried out by three extraction methods: Qiagen QIAamp DNA FFPE tissue kit, phenolchloroform extraction, and adaptive focused acoustics (AFA)- based extraction using the Covaris truXTRAC FFPE DNA kit.

For the Qiagen method, 24 sections of 10 µm thickness FFPE tissues were used as suggested by the manufacturer with some modifications as follows. Deparaffinization of FFPE tissues was performed by incubating twice in 1 mL xylene, then in a descending concentration of ethanol (100, 75%, then 50%). The tissue was then incubated in 300 µL Qiagen buffer ATL plus 40 µL proteinase K (20mg/mL, 5 PRIME Inc., Gaithersburg, MD, USA) for 72 hours. An additional 30 µL proteinase K was added at 24 hours, and another 30 µL at 48 hours. After 72 hours digestion, samples were washed in Qiagen DNeasy Mini Spin Columns with buffers AW1 and AW2. An extra wash with AW2 buffer was used to further reduce the carry-over of solvents. Finally, samples were eluted in 100 µl ATE buffer. Phenol-chloroform-isoamyl alcohol 25:24:1 (Fisher Scientific, Fair Lawn, New Jersey, USA) was used as described in Isola et al. [14]. 24 sections of 10 µm thickness were used from each FFPE sample. FFPE sections were scraped off of the air-dried slides using sterile scalpel blades and collected into 1.5 mL microcentrifuge tubes. Deparaffinization and digestion with proteinase K was performed as in the Qiagen protocol, above. An extra washing of the precipitated pellet with 70% ethanol was included to 5

further remove residual phenol-chloroform. Finally, DNA was resuspended in 40 µl TE buffer (Tris 10Mm, EDTA 1Mm pH8.0) and incubated overnight at room temperature to resuspend the DNA. The third extraction method used adaptive focused acoustics (AFA) technology. This was performed using FFPE tissues with 10 µm thickness for 8-10 sections, depending on the size of the tissue, to obtain approximately 5 mg of tissue. The extraction was performed according to the protocol suggested by Covaris (Woburn, MA, USA) in the truXTRAC FFPE DNA kit. Slides were warmed on a heat block to 37 °C for 30 seconds. FFPE tissue was then scraped from the slides, avoiding paraffin, using Covaris SectionPicks. Sections were collected into Screw-Cap microTUBES by using FFPE SectionPicks provided by Covaris. AFA was performed per manufacturer’s instructions (“protocol C”) on a Covaris M220 Focused-Ultrasonicator. Homogenized tissue was then digested for 2 hours in Covaris proteinase K at 56 °C. DNA was finally isolated from lysates using the columns of the Covaris truXTRAC FFPE DNA kit and eluted in 100 µL Covaris BE buffer. DNA was concentrated by speedvac without heat. Quality Control A Nanodrop spectrophotometer was used to quantify DNA concentration as well as determine the A260/A230 and A260/A280 ratios. For a more accurate quantitation, the Qubit® dsDNA BR Assay Kit (Thermo Fisher Scientific, Waltham, MA) was used to check the concentrations of dsDNA for all samples. Genomic DNA fragment sizes were first estimated by agarose gel electrophoresis of 250 ng DNA using 1% agarose gels (90 mM Tris –borate, 2 mM EDTA, 1% agarose). Samples with visible DNA fragments as large as 23,000 base pairs (bp) were processed further by randomly amplified polymorphic DNA PCR (RAPD-PCR) to directly determine the ability of each sample to produce high molecular weight amplicons (“amplifiability”). Non-specific primers and PCR conditions (below) were used to produce multiple amplicons from each sample. Agarose gel electrophoresis was used to determine the range of product sizes and only samples with amplicons larger than 500 bp were used for microarray analysis. 6

RAPD-PCR reactions were carried out in a 20 µL volume containing 25 ng DNA and using 10µl of GoTaq 2X Green Master Mix( Promega, Madison, WI USA). PCR was performed in 0.2mL tubes in a GeneAmp PCR System 9700 thermocycler (Life Technologies, Carlsbad, CA, USA). Primers used for RAPD-PCR were generated by Eurofins MWG Operon Inc (Huntsville, AL, USA). Sequences for the primer pairs and cycling parameters were as follows: 5’-AATCGGGCTG-3’ and 5’GAAACGGGTG-3’, 94°C for 2.5 minutes, then 45 cycles of 1 minute 94°C, 1 minute 55°C and 2 minutes 72°C, then 7 minutes 72°C and holding at 4°C; or 5’- TGTGCCCAGTGAAGACTCAG-3’ and 5’GAGTGAGCGGAGAGGGAACT-3’, 45 cycles of 94° C for 1 minute, 35° C for 1 minute, and 72° C for 2 minute. PCR products were resolved on 3% TBE agarose plus SYBR Safe dye (Life Technologies). Gels were visualized with a GE ImageQuant LAS-3000 camera (GE Healthcare Life Sciences, Piscataway, NJ, USA). Microarrays A larger set of 63 skin biopsy specimens were processed and hybridized to Affymetrix SNP6.0 microarrays. A subset of nine of these were extracted by AFA, while the others were processed using columns, as above. All samples passed RAPD-PCR quality control. 0.5 µg of genomic DNA was processed using the SNP 6.0 protocol and microarrays by Affymetrix (Affymetrix, Santa Clara, CA, USA) with some modifications to the standard protocol. The input amount of DNA was increased from 250ng per restriction enzyme (Nsp1 and Sty1) to 500ng each. The number of PCR reactions was doubled from the suggested three for Sty1 and four for Nsp1 to six for Sty1 and eight for Nsp1. It is important to note that the number of reactions was increased; the number of cycles in each reaction remained the same. The additional PCR reactions were combined as in the standard protocol. PCR cleanup was performed using isopropanol extraction (refer to Affymetrix User Bulletin 2: Improvements to step 7 of the SNP Assay 6.0,

7

PCR cleanup, using an isopropanol precipitation method, P/N 702968 Rev. 1). Hybridization and scanning of the arrays followed the manufacturer’s protocol. Data Analysis Statistical significance of comparisons between methods was determined using a two-tailed paired student’s t-test of the mean and standard error for 27 samples isolated by each of three methods. Microarray data were processed in Affymetrix Power Tools. Contrast Quality Control (CQC) was calculated and compared to restriction fragment sizes (for Nsp1 and Sty1) to measure the performance of DNA on the arrays. CQC is a measure of the ability of the data to resolve different alleles of the same target sequence. A CQC value > 0.4 is considered sufficient [15].

Results Twenty-seven benign melanocytic biopsy specimens were identified in the archives of the Dermatopathology Lab of Central States (Dayton, OH). These specimens (Table 1) were preserved in FFPE and stored at room temperature for an average of 8.67 years. The average patient age was 38. Specimens represent biopsies from a variety of locations on the skin. All were found to be large enough to take a sufficient number of slides for all three DNA extraction methods. The first slide and the last slide taken from each specimen were examined by staining with hemotoxylin and eosin (H&E) to verify the presence of cellular material in all sections. After extracting DNA from the same 27 samples using phenol- chloroform-isoamyl alcohol, Qiagen columns, or adaptive focused acoustics (AFA), the DNA was separated by electrophoresis to determine the genomic DNA size range. Four representative samples are shown in Figure 1a by each method. Every sample was further quantified by spectrophotometry and fluorometry (Table 2), thenused in 8

RAPD-PCR to determine the range of amplicon sizes that could be produced from DNA extracted by each method. Figure 1b shows RAPD-PCR results for these same four samples across each method. The quantity and quality of the extracted DNA was compared across all samples. Interestingly, there was a significant difference in the DNA yield per section between methods. The commercially available column-based kit produced significantly more DNA per section than phenol -chloroform or AFA (Figure 2, p < 10-4 ). The purity of the extracted DNA was compared by spectrophotometry to measure absorbance at 260, 280 and 230 nanometers. All three methods produced samples that would typically be called pure with A260/A280 ratios greater than 1.8, and an A260/A230 ratio indicative of minor contamination at approximately 1.7 (Figure 3). In order to measure the performance of the DNA in downstream PCR, RAPD-PCR was performed for each sample across all three extraction methods. Here, DNA produced by AFA showed amplicons of significantly greater length than DNA extracted by other means (Figure 4, p ≤ 0.04). These findings are summarized in Table 2. This comparison was expanded into a separate panel of 63 skin biopsy specimens examined by array comparative genomic hybridization (aCGH). Of these, nine were prepared by AFA and the remainder by columns. Contrast QC was calculated across the two groups of arrays and compared against restriction fragment size (Figure 5). A CQC above 0.4 shows sufficient hybridization on the arrays to resolve differences between the A and B alleles [15].

Discussion In this study, we compared two standard methods for isolation of DNA from FFPE tissues with a recent method based on sonication. FFPE tissues are some of the most readily available biomaterials, due in part to their long shelf life and accreditation requirements for their storage. The College of American

9

Pathologists (CAP) Commission on Laboratory Accreditation currently requires surgical pathology records and paraffin blocks to be retained for 10 years [1]. However, these blocks may be released for research purposes as long as 1) HIPAA requirements for patient privacy are observed, 2) the diagnostic laboratory retains sufficient tissue for diagnostic purposes, 3) the diagnostic laboratory has provisions to retrieve any material leftover after use in research, and 4) the requirements of applicable institutional review boards (IRB) and state and local laws are observed. Additionally, CAP currently encourages the retention of these archived materials beyond 10 years if possible, in appreciation of the increasing utility of these specimens as new biomarkers emerge. Therefore, many diagnostic laboratories represent de facto biorepositories that can be of use to researchers. Additionally, relevant clinical data are often already tied to these specimens, such as patient age, diagnosis, staging, and sometimes drug response and follow-up. Despite the availability of archived specimens, the process of fixation complicates the use of nucleic acids from these tissue blocks [16]. In order to overcome these barriers, it is important to use efficient methods for the extraction of quality nucleic acids, especially when the available tissue sections are small and irreplaceable. When comparing methods, we should consider the total yield and purity of the nucleic acids. The ability of the DNA to be amplified by PCR is a measure of its purity, as common contaminants (such as alcohols, xylene, and salts) often inhibit the PCR reaction. It is also important to consider the amount of tissue required to obtain sufficient DNA, especially given the regulatory requirement to retain enough tissue for future diagnostic work. In this study, we compared DNA extracted from the same specimens by three different methods and found that the extraction method can significantly affect the quality and quantity of DNA obtained from a given specimen.

10

Several measures were used as quality controls for this study. First, we examined the size distribution of genomic DNA obtained by each method (Figure 1a). Extraction by AFA showed slightly higher size distributions, but these results were mostly equivocal, suggesting that any method worked about as well as the other. However, when we use this DNA for RAPD-PCR we get a different picture: more products of a larger size were able to be amplified from DNA obtained by AFA than by other methods (Figure 1b, Figure 4). This is important because most applications using gDNA from clinical specimens will begin with PCR amplification, and RAPD-PCR is useful for predicting the suitability of a DNA sample for downstream PCR [17]. While RAPD-PCR does not reliably produce the same specific bands on a gel, we have found RAPD-PCR to be a rapid method to screen many amplicons using a single PCR reaction, and that these results are generally a good measure of performance in downstream applications (specifically aCGH). Introducing RAPD-PCR into our aCGH workflow reduced the rate of failed amplifications. [7-10, 18]We also examined the DNA yield per section (Figure 2). This controlled for the number of slides used in each extraction as well as the elution volumes used for each method. A variety of incubation times have been suggested in the literature for deparaffinization and proteinase K digestion [14, 18-20]. Increasing the proteinase K digestion time from one hour to overnight or even 72 hours has been shown to increase the amount of retrieved DNA in some studies [14, 18]; therefore, in this study a prolonged proteinase K digestion time was afforded to both the phenol-chloroform and column methods to compare best performance against the newer AFA method. Here there was a clear difference between the column-based method and the other two methods. The commonly used column-based kit retrieved approximately twice as much DNA per section as phenol-chloroform or AFA. It is possible that the yield by phenol-chloroform extraction could be improved by using phase-lock gel, which can increase nucleic acid recovery by 20-30% (5Prime, Gaithersburg, MD). For the recovered DNA to be of any use, it has to exist as a polymer; individual nucleotides or very small fragments will tell us little about our biological specimen. Therefore, it is important to compare the fragment length of the 11

recovered DNA in addition to quantity. Table 2 shows the quantification of recovered DNA by both a fluorometric method (Qubit) and a spectrophotometric method (Nanodrop). Nanodrop will include these smallest fragments as an increased DNA concentration, and we do observe an over-estimation by Nanodrop for all three extraction methods. Additionally, Figure 1 suggests that the DNA recovered by Qiagen columns had a greater constituency of small fragments; this could be partially responsible for the observed increase in concentration. Importantly, purity of the DNA is critical for most applications. Inhibitors of PCR can be carried over from the purification process and become apparent first by spectroscopy as lower A260/A230 ratios. In Figure 3, we see that all three methods achieved ratios of approximately 1.7, which is sufficient for most applications. It is unlikely, therefore, that what we describe as “PCR failure” (Table 2) is due to the presence of inhibitors in the DNA preparations We defined PCR failure as a RAPD-PCR resulting in amplicons less than 300 bp in length; we do not proceed to use these samples as this predicts poor performance in downstream PCR [17]. The rate of these failures was over 25% by phenol-chloroform and over 22% by columns. However, only one of 27 samples (3.7%) failed by AFA. An alternative explanation for failure to produce larger amplicons is that each method varies in its ability to extract intact DNA molecules of sufficient length. This is suggested by a greater proportion of gDNA below 2,000 bp in Figure 1 for phenol-chloroform and column methods. One possibility is that the reduced processing times involved in AFA reduces the likelihood that any longer DNA molecules present in the FFPE sample will be fragmented during the extraction. . Finally, a separate and larger set of FFPE skin biopsy specimens were processed and hybridization to Affymetrix SNP6.0 aCGH microarrays. Array CGH has been shown to be of use in the diagnosis ambiguous benign and malignant melanocytic lesions [21, 22], as the diagnosis of melanoma by qualified pathologists can be challenging. Veenhuizen et al. described 13-25% of ambiguous cases

12

submitted to a melanoma review board that were diagnosed as benign had been underdiagnosed and were in fact melanomas, while 14-36% of suspected or invasive melanomas had actually been over diagnosed and were benign [23]. Advances in molecular techniques suggest adjunctive molecular testing may provide means to reduce this diagnostic variability. Copy number variation has been shown to be a distinguishing characteristic of melanoma, and a potential way of discriminating between the cancer and ambiguous benign lesions [21, 24-27]. The reliability and accuracy of these molecular tests depends on our ability to obtain quality DNA from the same FFPE biopsy material that the pathologist is provided. Here we analyzed 54 specimens extracted using columns and another 9 extracted by AFA. Interestingly, none of the samples prepared by phenol-chloroform performed well in the PCR amplification step of the SNP6 protocol. Even when sufficient amplicon length was observed, the total DNA yield was insufficient to move forward with hybridization to arrays. During aCGH, genomic DNA is first digested with restriction enzymes and then these fragments are amplified from adapters ligated to these restriction sites. Therefore, if a region between restriction sites is broken in our input, that region will not be represented in the material hybridized to the array. The expected restriction fragments are binned by size and graphed in Figure 5 versus their abundance on hybridized arrays. Comparing the contrast QC between the Qiagen columns and AFA demonstrates larger restriction fragment lengths were better represented on the arrays using AFA (Figure 5). This was especially pronounced for the Nsp1 digestions, where the column DNA showed CQC > 0.4 for fragments 400-500 bp in length, but AFA DNA showed CQC > 0.4 for DNA fragments 600-700 bp long. Table 2 summarizes these findings.

The Qiagen column-based extraction method resulted in greater DNA yield than other methods. However, when determining what method to use these gains must be weighed against the greater

13

amenability to PCR, lower PCR failure rates, and recovery of larger chromosomal fragments when DNA is obtained by AFA.

Acknowledgments The authors would like to thanks Dr. John Moad and the Dermatopathology Lab of Central States (Dayton, OH) for providing specimens and expert dermatopathological assistance for these studies. This work was funded by the Research Challenge Ohio Third Frontier Grant Program and the IRS Qualified Therapeutic Discovery Program.

References 1. 2.

3. 4. 5.

6. 7. 8.

9.

Pathologists, C.o.A., CAP Acceditation Program Anatomic Pathology Checklist. 2014, Northfield, IL 60093-2750: College of American Pathologists. Shibata, D., W.J. Martin, and N. Arnheim, Analysis of DNA sequences in forty-year-old paraffinembedded thin-tissue sections: a bridge between molecular biology and classical histology. Cancer Res, 1988. 48(16): p. 4564-6. Srinivasan, M., D. Sedmak, and S. Jewell, Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol, 2002. 161(6): p. 1961-71. Bonin, S., et al., PCR analysis in archival postmortem tissues. Mol Pathol, 2003. 56(3): p. 184-6. Tuefferd, M., et al., Genome-wide copy number alterations detection in fresh frozen and matched FFPE samples using SNP 6.0 arrays. Genes Chromosomes Cancer, 2008. 47(11): p. 95764. Paireder, S., et al., Comparison of protocols for DNA extraction from long-term preserved formalin fixed tissues. Anal Biochem, 2013. 439(2): p. 152-60. Wang, J.H., et al., DNA extraction from fresh-frozen and formalin-fixed, paraffin-embedded human brain tissue. Neurosci Bull, 2013. 29(5): p. 649-54. Rabelo-Goncalves, E., et al., Evaluation of five DNA extraction methods for detection of H. pylori in formalin-fixed paraffin-embedded (FFPE) liver tissue from patients with hepatocellular carcinoma. Pathol Res Pract, 2014. 210(3): p. 142-6. Heydt, C., et al., Comparison of pre-analytical FFPE sample preparation methods and their impact on massively parallel sequencing in routine diagnostics. PLoS One, 2014. 9(8): p. e104566.

14

10.

11.

12. 13.

14.

15. 16. 17. 18. 19. 20. 21.

22. 23.

24. 25. 26. 27.

van Eijk, R., et al., Assessment of a fully automated high-throughput DNA extraction method from formalin-fixed, paraffin-embedded tissue for KRAS, and BRAF somatic mutation analysis. Exp Mol Pathol, 2013. 94(1): p. 121-5. Ling Lin, et al. Improved FFPE DNA extraction for next generation sequencing using adaptive focused acoustics technology. in 105th Annual Meeting of the American Association for Cancer Research. 2014. San Diego, CA: Cancer Research. Nelson, D.S., et al., Somatic activating ARAF mutations in Langerhans cell histiocytosis. Blood, 2014. 123(20): p. 3152-5. Yap, K.L., et al., Whole-Exome Sequencing of Muscle-Invasive Bladder Cancer Identifies Recurrent Mutations of UNC5C and Prognostic Importance of DNA Repair Gene Mutations on Survival. Clin Cancer Res, 2014. 20(24): p. 6605-17. Isola, J., et al., Analysis of changes in DNA sequence copy number by comparative genomic hybridization in archival paraffin-embedded tumor samples. Am J Pathol, 1994. 145(6): p. 13018. Affymetrix, I., Genotyping Console 4.0 User Manual. 2009. Ben-Ezra, J., et al., Effect of fixation on the amplification of nucleic acids from paraffin-embedded material by the polymerase chain reaction. J Histochem Cytochem, 1991. 39(3): p. 351-4. Siwoski, A., et al., An efficient method for the assessment of DNA quality of archival microdissected specimens. Mod Pathol, 2002. 15(8): p. 889-92. Senguven, B., et al., Comparison of methods for the extraction of DNA from formalin-fixed, paraffin-embedded archival tissues. Int J Med Sci, 2014. 11(5): p. 494-9. Ludyga, N., et al., Nucleic acids from long-term preserved FFPE tissues are suitable for downstream analyses. Virchows Arch, 2012. 460(2): p. 131-40. Specht, K., et al., Quantitative gene expression analysis in microdissected archival formalin-fixed and paraffin-embedded tumor tissue. Am J Pathol, 2001. 158(2): p. 419-29. Chandler, W.M., et al., Differentiation of malignant melanoma from benign nevus using a novel genomic microarray with low specimen requirements. Arch Pathol Lab Med, 2012. 136(8): p. 947-55. Beroukhim, R., et al., Assessing the significance of chromosomal aberrations in cancer: methodology and application to glioma. Proc Natl Acad Sci U S A, 2007. 104(50): p. 20007-12. Veenhuizen, K.C., et al., Quality assessment by expert opinion in melanoma pathology: experience of the pathology panel of the Dutch Melanoma Working Party. J Pathol, 1997. 182(3): p. 266-72. Bastian, B.C., et al., Chromosomal gains and losses in primary cutaneous melanomas detected by comparative genomic hybridization. Cancer Res, 1998. 58(10): p. 2170-5. Bastian, B.C., et al., Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol, 2003. 163(5): p. 1765-70. Dalton, S.R., et al., Use of fluorescence in situ hybridization (FISH) to distinguish intranodal nevus from metastatic melanoma. Am J Surg Pathol, 2010. 34(2): p. 231-7. Gerami, P., et al., Fluorescence in situ hybridization for distinguishing nevoid melanomas from mitotically active nevi. Am J Surg Pathol, 2009. 33(12): p. 1783-8.

15

Figure Legends Figure 1. DNA extracted from FFPE biopsy specimens. A) Genomic DNA was extracted from four specimens (numbered 1-4) by three different methods as described in Methods. Phenol chloroform isoamyl alcohol (“Phenol”), commercial column-based extraction kit (“Column”) and adaptive focused acoustics (“AFA”). Total extracted DNA was separated by electrophoresis on 1% agarose. B) gDNA was used for randomly amplified polymorphic DNA (RAPD) PCR to produce a variety of amplicons. The same four samples (1-4) isolated by three different methods were compared side-by-side. PC, positive control Jurkat genomic DNA. Figure 2. DNA yield per section. The DNA of 27 specimens was extracted by each of three different methods (x-axis). The total yield in nanograms (y-axis) is adjusted to control for the number of slides used in each extraction. Error bars show standard error for each method, and p-values were determined by two-tailed paired student’s t-test. Figure 3. Comparison of extracted DNA by spectrophotometry. The DNA extracted by each of three methods was analyzed by spectrophotometry to determine absorbance at 260 and 230 nm. Error bars show standard error and statistical significance was determined using a two-tailed paired student’s ttest. Figure 4. Maximum RAPD-PCR amplicon length. DNA extracted by each of three methods was used for RAPD-PCR to generate a variety of amplicons. The largest amplicon (in base pairs, bp) produced for each sample is predictive of performance in array comparative genomic hybridization (REF). Error bars show standard error and statistical significance was determined using a two-tailed paired student’s t-test. Figure 5. Array CGH contrast QC comparison. DNA was extracted by either the column-based method (A-B) or by AFA (C-D). DNA was hybridized to Affymetrix SNP6.0 microarrays. Contrast QC, the ability to

16

resolve between alleles on the arrays, was calculated for each probe and plotted against the length of the restriction fragment containing that target. Restriction for these arrays is performed with the enzyme Sty1 (A, C) or Nsp1 (B,D). Fragment length was divided into 100 bp bins (x-axis). Median CQC is shown in the bar graph (left y-axis). Distribution of the data is shown by the superimposed box plot (right y-axis).

17

Table 1. Details of specimens used in this study.

Sample number 1 2 3 4 5 6 7 8 9 10

Patient age 30 25 16 35 38 21 26 32 31 69

40 44 22 56 47 35 29 34 44 75 36 60 16 41 16 35 74 Average: 38.04 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Sex F M M F M M M F M F F F F M F F F F F F M F F F F M F

Specimen age (years) 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 11 11 11 11 11 11 8.67

Clinical notes Biopsy, nose, left side dermal nevus Biopsy, scalp, right post dermal nevus Biopsy, abdomen left dermal nevus Biopsy, back, left center dermal nevus Biopsy, scalp dermal nevus Biopsy, back right middle dermal nevus Biopsy, abdomen dermal nevus Biopsy, cheek, right dermal nevus Biopsy, axilla, right dermal nevus Biopsy, back, right upper dermal nevus Biopsy, axillary area. Anterior nevus lipomatosus superficialis Biopsy, calf, left post dermal nevus Biopsy, chest, left lateral neurotized dermal nevus Biopsy, cheek, left intradermal melanocytic nevus Biopsy, axilla, left dermal nevus Biopsy, chest dermal nevus Biopsy, back,midline lower neurotized dermal nevus Biopsy, deltoid left anterior dermal nevus Biopsy, lip,left upper dermal nevus Biopsy, knee, left medial dermal nevus Biopsy, axilla right dermal nevus Biopsy, back dermal nevus Biopsy, back, left upper dermal nevus Biopsy, forehead,right neurotized dermal nevus Biopsy, back, inferior lower dermal nevus Excision, malar, left dermal nevus Biopsy, back, left upper dermal nevus

18

Table 2. Summary of method comparisons.

DNA (ng)/section (Qubit) DNA (ng)/section (Nanodrop)

Phenol 106.2 ± 33.1 702.1± 200.2

Column 264.3 ± 35.4 716.0 ± 116.1

AFA 134.6 ± 18.1 453.8 ± 53.9

A

1.94 ± 0.02

2.04 ± 0.03

1.90 ± 0.02

1.71 ± 0.18 346.7 ± 24.1

1.71 ± 0.12 347.4 ± 21.4

1.75 ± 0.43 401.9 ± 10.2

260

/A

260

280 230

A /A Max. amplicon (bp)

19

20

22

23

24

25