Clin Genet 2007: 72: 441–447 # 2007 The Authors Printed in Singapore. All rights reserved Journal compilation # 2007 Blackwell Munksgaard CLINICAL GENETICS doi: 10.1111/j.1399-0004.2007.00882.x Short Report Formalin-fixed paraffin-embedded clinical tissues show spurious copy number changes in array-CGH profiles Mc Sherry EA, Mc Goldrick A, Kay EW, Hopkins AM, Gallagher WM, EA Mc Sherrya,b, A Mc Goldrickb, Dervan PA. Formalin-fixed paraffin-embedded clinical tissues show EW Kayc, AM Hopkinsb,d, spurious copy number changes in array-CGH profiles. WM Gallaghera and PA Dervanb,d Clin Genet 2007: 72: 441–447. # Blackwell Munksgaard, 2007 aUCD School of Biomolecular and b Formalin-fixed paraffin-embedded (FFPE) archival clinical specimens Biomedical Science, and UCD School of Medicine and Medical Science, are invaluable in discovery of prognostic and therapeutic targets for UCD Conway Institute, University College diseases such as cancer. However, the suitability of FFPE-derived genetic Dublin, Belfield, Dublin, Ireland, material for array-based comparative genomic hybridization (array- cDepartment of Histopathology, CGH) studies is underexplored. In this study, genetic profiles of matched Beaumont Hospital and The Royal FFPE and fresh-frozen specimens were examined to investigate DNA College of Surgeons in Ireland, integrity differences between these sample types and determine the Beaumont Hospital, Dublin, Ireland, and impact this may have on genetic profiles. Genomic DNA was extracted dMater Misericordiae Hospital, Dublin, from three patient-matched FFPE and fresh-frozen clinical tissue Ireland samples. T47D breast cancer control cells were also grown in culture and processed to yield a fresh T47D sample, a fresh-frozen T47D sample and a FFPE T47D sample. DNA was extracted from all the samples; array- Key words: comparative genomic hybridization – formalin fixation – CGH conducted and genetic profiles of matched samples were then genetic profiling compared. A loss of high molecular weight DNA was observed in the FFPE clinical tissues and FFPE T47D samples. A dramatic increase in Corresponding author: Professor William absolute number of genetic alterations was observed in all FFPE tissues Gallagher, UCD School of Biomolecular relative to matched fresh-frozen counterparts. In future, alternative and Biomedical Science, UCD Conway Institute, University College Dublin, fixation and tissue-processing procedures, and/or new DNA extraction Belfield, Dublin 4, and CGH profiling protocols, may be implemented, enabling Ireland. identification of changes involved in disease progression using stored Tel.: 1353 1 7166743; clinical specimens. fax: 1353 1 2837211; e-mail: [email protected] Received 14 May 2007, revised and accepted for publication 6 July 2007 Genomic instability is a hallmark of human cancer ization (CGH). Array-CGH is conducted by (1). Amplification or deletion of distinct sub- differentially labelling total genomic DNA from chromosomal areas can lead to over- or under- test and reference samples followed by cohybrid- expression of key genes, thus conferring a growth ization onto an array consisting of DNA probes advantage to malignant cells (2). Conversely, spanning the genome. The ratio of fluorescent deletions of tumour suppressor genes, such as intensities allows identification of copy number those involved in cell death, may also push the change between test and reference samples. genetic balance towards malignant growth. Ampli- Classically, this technique has been successfully fied genes are, therefore, important targets for used on DNA extracted from fresh-frozen clinical therapeutic invention, and identification of such tissue specimens, as these yield higher quality copy number alterations (CNAs) can help elucidate nucleic acids (3–5). However, as the availability potential mechanisms involved in tumour develop- of fresh-frozen tissue is often limited, studies of ment and progression and identify cancer subtypes. archival formalin-fixed paraffin-embedded (FFPE) The most widely used and successful technique clinical tissue specimens with accompanying ret- to study CNAs is comparative genomic hybrid- rospective data would be of enormous benefit in 441 Mc Sherry et al. elucidating key genes involved in disease pro- were immediately frozen using liquid nitrogen gression. Array-CGH studies using FFPE DNA upon reaching the histopathology laboratory have been reported (6–8), although little is known from surgery. Samples were stored at 280°C for of the true impact of tissue fixation and process- a maximum of 2 months prior to frozen sectioning ing on resulting genetic profiles. on a cryostat. DNA was immediately extracted The present study was primarily performed to after sectioning. Matched colon cancer FFPE assess the correlation of array-CGH profiles of tissues were fixed in 10% formalin for an average DNA extracted from matched FFPE and fresh- of 16 h before embedding in paraffin. Blocks were frozen tissues, investigate the number of CNAs stored at room temperature for a maximum of present in DNA extracted from FFPE tissues with 2 months prior to sectioning. DNA was then respect to their matched fresh-frozen counter- extracted within 1 week of sectioning. parts, and to assess any difference in DNA quality between these sample types. DNA extractions Materials and methods DNA from all above sources was extracted using Cell line and tissue preparation the NucleoSpin Tissue Kit (Clontech, Saint- T47D breast cells (European Collection of Cell German-en-laye, France). The following additional Cultures, UK) were cultured in Dulbecco’s Mod- protocol was applied prior to DNA extraction ified Eagle’s Medium 1 10% foetal bovine serum, from FFPE tissue. FFPE tissue sections were 200 mM L-glutamine. Three tissue culture flasks placed in 1.5 ml tubes. Xylene (1 ml) was added were grown to confluency. Three identical cell and mixed at room temperature for 30 min. Tubes T47D cell pellets were prepared, and genomic were microcentrifuged for 3 min at 11,000 3 g to DNA was immediately extracted from the first. A pellet tissue, and subsequent supernatant was FFPE cell sample was prepared from the second removed. Ethanol (96% v/v; 1 ml) was added to pellet, by first fixing T47D cells by suspension in each tissue, mixed, and microcentrifuged for 10% neutral buffered formalin (NBF) for 1 h at 3 min at 11,000 3 g then ethanol was removed. room temperature. The pellet was centrifuged at This step was repeated, and after removal of 18,000 3 g r.p.m. for 5 min, formalin removed, ethanol, the tissue was incubated at 37°C until all 70% ethanol added, mixed and left overnight. The ethanol had evaporated. The standard Nucleo- pellet was then dehydrated through increasing Spin Tissue Kit manual procedure was then concentrations of ethanol (70%, 95%, 100%, and applied to all tissues for extraction of DNA. 100%) at room temperature for 1 h 45 min each Extracted DNA was quantified using a NanoDrop and then finally in xylene at room temperature for ND-1000 Spectrophotometer (Nanodrop Tech- two steps of 1 h 45 min each. Liquid paraffin was nologies, Wilmington, DE). DNA integrity was added to the pellet, and cells were maintained at assessed by analysis of 100 ng DNA via 2%agarose 60°C (melting point of paraffin) and replaced gel electrophoresis. every 20 min for a period of 2 h to ensure removal of residual xylene. The pellet was then allowed to solidify at room temperature in fresh paraffin and Array-CGH was then implanted in a new paraffin block from which sections were then cut for DNA extraction The GenoSensor array-CGH platform (Vysis; from FFPE cells. Fresh-frozen T47D cells were Abbott Laboratories, Des Plaines, IL) was utilized prepared by snap-freezing the third pellet of T47D in this study. In brief, test and reference DNA cells in optimum cutting temperature (OCT) (100 ng) were labelled with Cy3 and Cy5 fluoro- matrix. The frozen block was stored at 280°C phores (1 mM). Labelled targets were DNase for 1 week, then trimmed to remove excess OCT treated for 1 h at 15°C then purified by sodium matrix from the frozen cell pellet, and fresh-frozen acetate and ethanol precipitation steps. The quality DNA was extracted. of purified, labelled DNA was assessed by agarose Three cases of patient-matched fresh-frozen gel electrophoresis. Equal concentrations of and FFPE colon cancer tissue samples were labelled test and reference DNA were then com- also acquired for array-CGH investigation from bined and loaded onto the GenoSensor Array 300 the Department of Histopathology, Beaumont array. Hybridization was conducted in the pre- Hospital, Dublin. DNA was extracted from a sence of humidified 23 standard saline citrate (SSC)/ 20 mm section of each fresh-frozen tissue and each 50% (v/v) formamide for 60 h at 37°C. Arrays matched FFPE tissue, which had been processed were washed in 23 SSC/50% (v/v) formamide, as follows. Colon cancer tissues for frozen storage followed by 13 SSC. Spots were counterstained 442 Array-CGH of FFPE tissues with DAPI, and scanning was performed using the differences in quality by agarose gel electropho- GenoSensor Reader System. resis (Fig. 1). Lanes 1–3 show a complete lack of high molecular weight DNA in FFPE clinical samples. Lanes 4–6 show the retention of high molecular weight DNA in fresh-frozen clinical Image acquisition and data analysis samples, with minor degradation as indicated by The GenoSensor reader was used to acquire the appearance of lower molecular weight DNA. a three-colour image of each microarray using DNA samples extracted from three alternatively a charge coupled device camera and a 175W xenon processed T47D breast cancer cell samples were light source. Automated analysis was then con- also compared. Lane 8 shows the retention of ducted by the GENOSENSOR READER software high molecular weight DNA in the DNA as follows. Spot segmentation and identification extracted from fresh T47D cells. Lane 9 shows was conducted to isolate and relate every DNA that snap-frozen OCT-embedded T47D cells spot on the microarray to a particular DNA target produced DNA of high molecular weight with clone.