MATERIALS AND METHODSCopy Number Variation assays (148) were pre-screened using the7900HT Fast Real-Time PCR System on a sample of normalhuman genomic DNA. Of these 148 assays, 48 were selected forour study (see Table 1). Referring to Figure 5, 6 replicate sub-arrays per assay (384 total PCR reactions) provides the level ofconfidence to observe low-fold differences between the cell lines.We chose to interrogate 8 human DNA samples: Normal Male andFemale from Lofstrand (Gaithersburg, MD); Raji, Hela, A431,K562, Jurkat, and MCF7 from Biochain (Hayward, CA). Eachsample was diluted down to the limiting range of approximately0.65 copies of DNA per reaction (based off of a NanoDrop® ODmeasurement).

Julie T. Kuykendall, Wensheng Nie, Ferrier Le, Chad Mooney, Junko F. Stevens, Jonathan T. Wang, Life T echnologies, 850 Lincoln Center Drive, Foster City, CA 94404, U.S.A.

ABSTRACTHere we describe an approach using Digital PCR to detectchromosomal aberration in tumor cell lines. We use 48 TaqMan®

Copy Number Assays (2 assays for 24 chromosomes – 22autosomes and 2 sex chromosomes) to count the copy number ofchromosomes. We characterized genomic DNA extracted from 6different human cancer cell lines. We also used genomic DNAnormal human cell lines (one female and one male) as controls.

INTRODUCTIONChromosomal rearrangement is a common feature in tumorderived cell lines. The inversion, translocation, deletion, orduplication of parts of the chromosome can lead to suchabnormalities. For example, a cancerous cell may have 3 copiesof Gene A while a normal cell will have 2. While these smalldifferences can be difficult to detect using traditional qPCR(quantitative real-time Polymerase Chain Reaction) methods,digital PCR technology offers the precision and accuracy requiredto discriminate these variations in copy number.

Digital PCR is an advanced utilization of PCR technology wherethe number of nucleic acid template is accurately counted. Thistechnology relies on diluting the nucleic acid template to thePoisson distribution limit where each PCR reaction contains atleast a single copy of template or none. The 1 and 0 scoring ofthe reactions gives the digital nature of this PCR application. The

Figure 4. 95% Confidence Limits

CONCLUSIONSWe demonstrated the use of the OpenArray® Real-Time PCR

Detection of Chromosome Aberration in Tumor Cell Lines Using the OpenArray® Real-Time PCR System

Figure 6. Heat Map – Normalized to Copies per Cell.

Table 1. Panel of 48 Assays

Expected 95% confidence limit for a target of 0.65 copies/reaction across thenumber of replicate sub-arrays.

Normalized data to compare each assay across the 8 cell lines. Normalization foreach sample was independent of the results from the other samples. Differencemap normalizes to normal female or male cell line.

RESULTS

the reactions gives the digital nature of this PCR application. Theaverage number of copies per reaction can then be calculatedusing Poisson statistics.

Additionally, the accuracy and precision of quantitation by digitalPCR is proportional to the number of reactions. Therefore, thehigh throughput OpenArray® Real-Time PCR System (3072 wellsper array and 3 arrays per instrument run) enables us to measurechromosomal rearrangements with confidence.

Digital PCR results for one run. On the left is a 2D representation of the positiveand negative reactions (white = positive). Each OpenArray® Digital PCR Platewas sub-divided into sections indicated by the red lines (6 sub-arrays per assay).To the right are the calculated results from the software, which include values forAverage Copies per Reaction and Confidence Interval Range.

Figure 5. Run Output from OpenArray ® Digital PCR Software

We demonstrated the use of the OpenArray® Real-Time PCRsystem to accurately count the number of copies for chromosomaltargets on 8 different cell lines. Abnormalities in the karyotypeswere extracted from the copies per reaction results and normalizedto one of the X-chromosome assays in order to compare copynumbers (Figure 6). This task was intuitive for the Normal Maleand Raji cell lines where the sex chromosomes were counted toequivalent values. As a result, we confirmed autosome numberaberrations in all six of the tested cancer cell lines. However, thesex chromosome numbers appear to be normal. The results areconsistent with the aneuploidy nature of these tumor derived celllines. The combination of chromosome numerical aberration andintra chromosome rearrangement pattern is unique to each cellline. This allows the identification of individual cell lines by a digitalPCR signature. The data also shows the linear nature of digitalcounting is capable of resolving high copy number variation (n>4)with precision. Combined with existing TaqMan® Copy Numberassays and cytogenetic techniques, the OpenArray® Real-TimePCR system can serve as a powerful tool to project cleaner andbrighter images of chromosomal aberrations.

REFERENCES1. Karapova MB , Schoumans J, Ernberg I, Henter JI, Nordenskjold M, & FadeelB. Raji revisited: cytogenetics of the original Burkitt's lymphoma cell line.Leukemia 2005; 19: 159-161.2. Macville M, Schrock E, Padilla-Nash H, Keck C, Ghadimi BM, Zimonjic D,Popescu N, & Ried T. Comprehensive and Definitive Molecular CytogeneticCharacterization of HeLa Cells by Spectral Karyotyping. Cancer Research 1999;59: 141-150.3. Fazekas de St. Groth, S. The Evaluation of Limiting Dilution Assays. Journal ofImmunological Methods 1982; 49: R11–22.

TRADEMARKS/LICENSINGFor research use only. Not intended for human or animal therapeutic or diagnosticuse.The trademarks mentioned herein are the property of Life TechnologiesCorporation or their respective owners.© 2010 Life Technologies Corporation. All rights reserved.TaqMan is a registered trademark of Roche Molecular System, Inc.

Figure 1. OpenArray ® Real-Time PCR Platform

Table 1. Panel of 48 Assays

Label Assay ID GeneCytogenetic

Band Label Assay ID GeneCytogenetic

Band1/3 Hs02024837_cn NFASC 1q32.1f 13/3 Hs00614394_cn DZIP1 13q32.1b1/5 Hs01518044_cn OSCP1 1p34.3d 13/4 Hs01061789_cn UPF3A 13q34d2/2 Hs05833409_cn LRP1B 2q22.1e 14/1 Hs02509147_cn GZMB 14q12a2/3 Hs02649413_cn FIGN 2p25.3g 14/2 Hs02603629_cn ZFYVE1 14q24.23/1 Hs04744091_cn TEX264 3p21.2b 15/1 Hs05380469_cn ATPBD4 15p13e3/3 Hs04751460_cn FOXP1 3p26.3c 15/7 Hs00689935_cn CHD2 15q26.1e4/1 Hs01271024_cn ZNF827 4q31.21c 16/1 Hs00257926_cn ITGAM 16p13.3f4/6 Hs01083969_cn C4orf17 4p16.3d 16/7 Hs03934353_cn 16q24.1c5/6 Hs00349464_cn PJA2 5p15.33e 17/1 Hs05512629_cn ACSF2 17q21.33b5/7 Hs06063082_cn SNX24 5p15.33e 17/3 Hs01959936_cn PIK3R6 17p13.3g6/1 Hs04902228_cn PHACTR1 6p25.3b 18/5 Hs06498356_cn ATP9B 18p11.32c6/3 Hs00528347_cn CD109 6p25.3b 18/7 Hs06449787_cn PTPN2 18p11.21d7/1 Hs04961331_cn PKD1L1 7p22.3d 19/1 Hs02134886_cn MEGF8 19q13.2c7/2 Hs04981288_cn C7orf10 7p22.3d 19/7 Hs02925411_cn SIN3B 19p13.11e8/1 Hs06220008_cn RAB2A 8q12.1d 20/2 Hs04053409_cn 20q13.12a8/4 Hs06163596_cn MSRA 8p23.1c 20/4 Hs04061467_cn TGM6 20p13f9/1 Hs05083990_cn PRUNE2 9q21.2a 21/2 Hs01315872_cn SYNJ1 21q22.11b9/5 Hs06836541_cn IFT74 9p21.2a 21/7 Hs02062007_cn KCNJ6 21q22.13b

10/2 Hs03073413_cn TIAL1 10q26.11d 22/1 Hs01285262_cn IL17RA 22q11.1d10/5 Hs01210989_cn BEND7 10p15.3d 22/5 Hs00954936_cn SBF1 22q13.33b11/3 Hs03789421_cn HNRNPUL2 11p15.5d X/1 Hs01967107_cn RGAG1 Xq23a11/5 Hs00979243_cn FADS3 11p15.5d X/3 Hs02185738_cn ZFX Xp22.11a12/4 Hs02118968_cn MGST1 12p13.33d Y/2 Hs04125045_cn RPS4Y2 Yq11.223a12/8 Hs06995341_cn YAF2 12p13.33d Y/4 Hs07226066_cn PCDH11Y Yp11.32c

Figure 2 is an example of qPCR as a function of both signal (Rn) andfrequency against Cycle (or CT). Single copy detection results in a widespread of CT values. Digital PCR makes use of the stoichastic distribution of 0,1, 2+ copies per reaction in determining absolute quantitation. Figure 3 is thedistribution of an experimental data set targeting a limiting dilution.

Figure 2. qPCR vs dPCR Figure 3. Poisson Distribution

~33nL reaction