Priority Report Mutational Analysis of Thirty-two Double-Strand DNA Break Repair Genes in Breast and Pancreatic Cancers Xianshu Wang,1 Csilla Szabo,1 Chiping Qian,3 Peter G. Amadio,1 Stephen N. Thibodeau,1 James R. Cerhan,2 Gloria M. Petersen,2 Wanguo Liu,3 and Fergus J. Couch1 Departments of 1Laboratory Medicine and Pathology and 2Health Sciences Research, Mayo Clinic College of Medicine, Rochester, Minnesota and 3Department of Genetics, Louisiana State University Health Sciences Center, New Orleans, Louisiana Abstract that mutations in other DNA damage repair genes may predispose and/or contribute to breast cancer. Similarly, the recent discovery Inactivating mutations in several genes that encode compo- BRCA2, FANCC FANCG nents of the DNA repair machinery have been associated with that mutations in , and (2, 11, 12) are an increased risk of breast cancer. To assess whether associated with pancreatic cancer suggests that mutations in other alterations in other DNA repair genes contribute to breast repair genes may contribute to pancreatic cancer risk. To identify cancer and to further determine the relevance of these genes other DNA repair genes associated with breast and pancreatic to pancreatic cancer, we performed mutational analysis of 32 cancer, we performed a mutation screen of the coding regions of 32 genes involved in DSB signaling and repair in 38 breast tumors, DNA double-strand break repair genes in genomic DNA from BRCA1/ BRCA1/ 48 pancreatic tumors, and germline DNA from 10 non- 38 breast tumors, 48 pancreatic tumors, and 10 non- BRCA2 BRCA2 hereditary breast cancer patients. A total of 494 coding hereditary breast cancer patients. We report the identifi- exons were screened by denatured high-performance liquid cation of several truncating and missense mutations. chromatography and direct DNA sequencing. Two inactivating mutations were identified in breast tumor samples, a germline Materials and Methods single-nucleotide deletion in POLQ (c.3605delT) and a somatic Samples for mutation screen. Genomic DNA was obtained from an nonsense change in PRKDC (c.2408C>A, p.Ser803X). Two unselected series of 48 pancreatic tumors and 38 breast tumors and from germline-inactivating mutations in RAD50 (c.1875C>G, blood samples from probands of 10 high-risk breast cancer families without p.Tyr625X and IVS14+1G>A) were also detected in separate BRCA1 and BRCA2 mutations collected at Mayo Clinic. In addition, 48 pancreatic tumor samples. In addition, 35 novel nonsynon- control DNAs were obtained from blood samples of noncancer patients ymous amino acid substitutions, resulting from two in-frame attending Internal Medicine clinics. DNA was extracted from cryocut sections of frozen tumors or buffy coats with Easy-DNA kits (Invitrogen). deletions and 33 single nucleotide alterations, were identified. Mutation screening by denatured high-performance liquid chro- Seven of these were predicted to influence protein function. A matography. PCR primers were designed for all coding exons of 32 genes CLSPN separate analysis of the c.3839C>T (rs35490896) variant (Supplementary Table S1). PCR products were generated using AmpliTaq that was observed more frequently in breast tumors than in Gold (Roche) or HotStar Taq (Invitrogen), 10 ng genomic DNA, and 35 to 40 pancreatic tumors or normal controls failed to detect a cycles of amplification. Denatured high-performance liquid chromatography significant association with breast cancer risk in a Mayo Clinic (DHPLC) analysis was performed on automated DHPLC instruments (Trans- breast cancer case-control study. In conclusion, this screen of genomic or Varian). DNA samples displaying alterations on DHPLC were DNA repair genes implicates PRKDC and POLQ as candidate reamplified and sequenced on an ABI Prism 377 Sequencer (Perkin-Elmer). tumor suppressor genes involved in breast cancer and Coding single-nucleotide polymorphisms selection and PMut suggests that inactivating mutations in RAD50 predispose to analysis. All known single-nucleotide polymorphisms (SNP) in the coding regions of the 32 genes were obtained from Hapmap release 21 (13). Only pancreatic cancer as well as breast cancer. [Cancer Res SNPs with validation and defined minor allele frequency in Caucasian 2008;68(4):971–5] populations were used as positive controls to validate the sensitivity of the DHPLC analysis. Amino acid substitutions were predicted to be neutral or Introduction pathologic using PMut4 prediction scores and an associated reliability Double-strand DNA break (DSB) repair pathways are essential index, which compares favorably to PolyPhen and SIFT (14). Loss of heterozygosity and mRNA transcription analysis. Loss of for the prevention of genomic instability. Previous studies have BRCA1 BRCA2 heterozygosity (LOH) analysis was performed on laser capture micro- shown that genetic defects in and , key components dissected pancreatic tumor and matched normal stromal tissues. DNA and of homologous recombination repair, are associated with breast, RNA were isolated from 4,000 tumor cells captured on a Veritas Laser ovarian, and pancreatic cancer risk (1–3). Rare inactivating Capture Microdissection instrument (Molecular Devices Corporation) using mutations in several DNA damage repair and signaling genes, QIAamp DNA Microkit (Qiagen) and PicoPure RNA Isolation kit (Molecular ATM (4), BRIP1 (5), CHEK2 (6), PALB2 (7), NBS1 (8), RAD50 (9), and Devices Corporation). PCR and reverse transcription–PCR (RT-PCR) were TP53 (10), have also been implicated in breast cancer. This suggests conducted using primers shown in Supplementary Table S2. Case and control population. Genotyping of the CLSPN rs35490896 SNP was conducted using 798 invasive breast cancer cases and 843 matched controls from an Institutional Review Board–approved Mayo Clinic breast Note: Supplementary data for this article are available at Cancer Research Online cancer case-control series (15). Collection and characteristics of cases and (http://cancerres.aacrjournals.org/). controls are described in the Supplementary Materials and Methods. Requests for reprints: Fergus J. Couch, Department of Laboratory Medicine and Pathology, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN 55905. Phone: 507-284-3623; Fax: 507-538-1937; E-mail: [email protected]. I2008 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-07-6272 4 http://mmb2.pcb.ub.es:8080/PMut/ www.aacrjournals.org 971 Cancer Res 2008; 68: (4). February 15, 2008 Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2008 American Association for Cancer Research. Cancer Research TaqMan genotyping and data analysis. Genomic DNAs from breast sufficiently sensitive to detect the majority of putative disease- cancer cases and controls were genotyped by the 5¶ nuclease assay associated mutations. (TaqMan), using the 7900HT Real-Time PCR System (Applied Biosystems), Novel deleterious mutations. The mutation screen of 96 in 384-well format. Duplicate samples (5%) were included and all displayed samples resulted in the identification of 73 unique alterations in 100% concordance for the CLSPN c.3839C>T genotype. Unconditional 17 genes. These variants included three truncating mutations, an logistic regression was used to estimate odds ratios and 95% confidence intervals (CI) under general (df, 2) and ordinal models. Analyses were in-frame splicing variant, two in-frame three-nucleotide deletions, adjusted for common breast cancer risk factors and performed using and 33 novel nonsynonymous changes. In addition, 154 intronic SAS software, version 8.2 (SAS Institute, Inc.). changes were detected. A summary of the results of this extensive mutation screen is shown in Tables 1 and 2. One truncating mutation, resulting in a premature stop codon Results p.Ser803X, was detected in PRKDC in a breast tumor sample Sensitivity of DHPLC analysis. We screened a total of 1.7 Â 107 (Table 2). PCR and DNA sequencing of blood-based DNA from the bases of DNA in 494 coding exons from 32 genes to identify same patient confirmed that the mutation was of somatic origin. mutations associated with breast and pancreatic cancer. To assess Another truncating mutation, c.3605delT, was identified in POLQ in the sensitivity of the DHPLC screen, we compared the variants blood DNA from a proband of a high-risk breast cancer family detected in the mutation screen with all known SNPs from these (Table 2). The single-nucleotide deletion caused a frameshift and exons. Among 60 SNPs listed in dbSNP as validated in at least two truncation of the protein and exclusion of the POLQ C-terminal studies of Caucasians, 54 were detected in this mutation screen. DNA polymerase and 3¶-5¶ exonuclease domains. These mutations Four of the six SNPs not identified exhibited minor allele likely inactivate the DNA-PK and POLQ DNA repair proteins. frequencies of <2% and were excluded. The detection rate of Two deleterious mutations were also identified in the RAD50 96.4% for known SNPs indicated that the DHPLC technique was gene in DNA from pancreatic tumors (Table 2). A c.1875C>G Table 1. Mutation screening of DNA damage repair genes in breast and pancreatic cancers DNA repair genes Coding exons screened Known SNPs in databases Known SNPs detected Novel alterations detected ABL1 11/11 4 4 6 ATF2 12/12 2 0 2 ATR 47/47 7 7 7 ATRX 35/35 1 0 4 CLSPN 25/25 1 1 4 EEF1E1 4/4 0 0 0 FAAP24 4/4 4 4 0 H2AFX 1/1 0 0 0 HUS1 8/8 1 1 0 LIG4 1/1 3 3 0 MRE11 19/19 0 0 0 NBN 16/16 4 3 1 POLQ 31/31 10 10 13 RAD1 5/5 1 1 1 RAD50 25/25 0 0 2 RAD51 9/9 0 0 0 RAD51C 9/9 0 0 0 RAD51L3 10/10 2 2 0 RPA1 17/17 2 2 1 SFN 1/1 0 0 0 SMC1A 25/25 0 0 0 TERF2 10/10 0 0 1 TP53BP1 27/27 5 5 7 TREX1 1/1 1 1 2 XPA 6/6 0 0 0 XPC 16/16 4 4 6 XRCC2 3/3 1 1 0 XRCC3 7/7 1 1 0 XRCC4 7/7 3 3 0 XRCC5 21/21 2 1 4 XRCC6 12/12 1 0 2 XRCC7* 69/86 NA NA 10 Total 494 60 54 73 *For XRCC7 (PRKDC), only 69 of 86 coding exons were completed.