Claire Seedhouse, Rowena Bainton, Michael Lewis, Alexander Harding, Nigel Russell, and Emma Das-Gupta

NEOPLASIA The genotype distribution of the XRCC1 gene indicates a role for base excision repair in the development of therapy-related acute myeloblastic leukemia Claire Seedhouse, Rowena Bainton, Michael Lewis, Alexander Harding, Nigel Russell, and Emma Das-Gupta Polymorphisms in several DNA repair genes have been described. These polymorphisms may affect DNA repair capacity and modulate cancer susceptibility by means of gene-environment interactions. We investigated DNA repair capacity and its association with acute myeloblastic leukemia (AML). We studied polymorphisms in 3 DNA repair genes: XRCC1, XRCC3, and XPD. We also assessed the incidence of a functional polymorphism in the NQO1 gene, which is involved in protection of cells from oxidative damage. We genotyped the polymorphisms by using polymerase chain reaction Introduction restriction fragment-length polymorphism analysis in 134 patients with de novo AML, 34 with therapy-related AML (t-aml), and 178 controls. The distributions of the XRCC3 Thr241Met and NQO1 Pro187Ser genotypes were not significantly different in patients and controls. However, the distribution of the XRCC1 Arg399Gln genotypes was significantly different when comparing the t-aml and control groups ( 2, P.03). The presence of at least one XRCC1 399Gln allele indicated a protective effect for the allele in controls compared with patients with t-aml (odds ratio 0.44; 95% confidence interval, 0.20-0.93). We found no interactions between the XRCC1 or XRCC3 and NQO1 genotypes. We also found no differences in the distribution of the XPD Lys751Gln or XRCC1 Arg194Trp genotypes. Our data provide evidence of a protective effect against AML in individuals with at least one copy of the variant XRCC1 399Gln allele compared with those homozygous for the common allele. (Blood. 2002;100: 3761-3766) 2002 by The American Society of Hematology Acute myeloblastic leukemia (AML) is a clonal hemopoietic disorder that is frequently associated with genetic instability characterized by a diversity of chromosomal and molecular changes. Most cases of AML arise de novo, with no known exposure to leukemogenic substances. However, approximately 10% to 20% of all cases of AML arise after therapy, most often chemotherapy, used to treat other malignant diseases (therapy-related AML [t-aml] 1 ). Many genes encode proteins that function to protect cells against genetic instability by means of many mechanisms, including DNA repair pathways and protection against oxidative stress. DNA repair pathways play a vital role in maintaining genetic integrity, and it is becoming clear that defects in repair pathways are connected to many different types of diseases, including leukemia and cancer. It is now thought that an individual s DNA repair capacity is genetically determined and is the result of combinations of multiple genes that may display subtle differences in their activity (see Mohrenweiser and Jones 2 for review). Inactivating mutations in DNA repair genes are rare, resulting in embryonic death or serious genetic diseases and reflecting the importance of the gene products; however, polymorphisms have been identified in several DNA repair genes. 3 Many of these polymorphisms result in amino acid substitutions and hence may alter wild-type (WT) protein function and affect cellular ability to repair endogenous and exogenous DNA damage, thereby contributing to disease susceptibility. XRCC1, XRCC3, and XPD are polymorphic genes belonging to 3 of the major DNA repair pathways. XRCC1 is involved in base excision repair (BER) and the repair of single-strand breaks. The XRCC1 gene product plays an important role in the pathway by acting as a scaffold for other DNA repair proteins, such as DNA polymerase 4 and DNA ligase III. 5 The protein also has a BRCA1 C-terminus (BRCT) domain, which is characteristic of proteins involved in the recognition of and response to DNA damage. XRCC1 interacts with poly (ADP-ribose) polymerase (PARP) by means of the BRCT domain to enable the recognition and subsequent repair of single-strand breaks. 6,7 DNA damage caused by a variety of internal and external factors, including ionizing radiation, alkylating agents, and oxidation, requires repair by the BER pathway. Several variants of XRCC1 have been described, including one affecting codon 399 in exon 10 that results in an arginine (Arg) to methionine (Met) substitution 3 and one affecting codon 194 in exon 6 that results in an Arg to tryptophan (Trp) substitution. Both codon 194 and codon 399 are conserved across species. Codon 194 resides in a linker region connecting the domains that interact with PARP and DNA polymerase, whereas codon 399 resides in the functionally important PARP-binding domain. 8 Chinese hamster ovary cell lines with nonconservative amino acid substitutions in the PARP-binding domain were found to have a decreased ability to repair DNA damage. 9 From the Division of Haematology, School of Clinical Laboratory Sciences, University of Nottingham, and Nottingham City Hospital, Nottingham, United Kingdom. Submitted April 16, 2002; accepted July 1, 2002. Prepublished online as Blood First Edition Paper, July 18, 2002; DOI 10.1182/blood-2002-04-1152. Sponsored partly by a grant from the Leukaemia Research Fund, United Kingdom. Reprints: Claire Seedhouse, Department of Academic Haematology, Clinical Sciences Building, Nottingham City Hospital, Nottingham, NG5 1PB, United Kingdom; e-mail: claire.seedhouse@nottingham.ac.uk. The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked advertisement in accordance with 18 U.S.C. section 1734. 2002 by The American Society of Hematology BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 3761

3762 SEEDHOUSE et al BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 The XRCC3 protein functions in the homologous DNA doublestrand break (DSB) repair pathway and directly interacts with and stabilizes Rad51, 10 one of the key components of the pathway. The homologous DNA DSB repair pathway uses the second, intact copy of a chromosome as a template to copy the information lost at the DSB site on the first chromosome, resulting in a high-fidelity process that has a vital role in preventing chromosomal aberrations. The threonine (Thr) to Met polymorphism at codon 241 of XRCC3 is a nonconservative change. Little is known about the functional consequence of this amino acid change, although positive associations between the variant allele and cancer have been observed in various investigations, including studies of bladder cancer 11 and melanoma skin cancer. 12 The XPD protein, a 5 to 3 DNA helicase involved in the nucleotide excision repair pathway, functions to remove bulky damage adducts from DNA. The XPD lysine (Lys) 751 glutamine (Gln) polymorphism does not reside in a known functional domain of XPD and was initially thought to be unlikely to result in an altered DNA repair capacity. 3,13 However, additional studies produced contrasting results, linking the XPD 751Gln variant allele to both reduced 14 and elevated 15 nucleotide excision repair capacity. Thus, more work is warranted to establish whether this polymorphism is clinically relevant. Nicotinamide adenine dinucleotide phosphate:quinone oxidoreductase 1 (NQO1) functions to protect cells from oxidative stress by detoxifying several compounds. The enzyme reduces reactive quinones to the less reactive hydroquinones and so prevents accumulation of reactive oxygen species that may then go on to damage DNA. A polymorphism has been identified in NQO1 at codon 187. 16 This polymorphism converts a proline (Pro) to a serine (Ser) residue and has been shown to result in inactivation of NQO1. 17 The incidence of this polymorphism was found to be significantly increased both in patients with de novo AML 18 and in those with t-aml, particularly those with chromosome 5/7 abnormalities. 19 In the current study, we investigated the genotype distributions of the XRCC1 Arg399Gln, XRCC3 Thr241Met, and NQO1 Pro187Ser polymorphisms in patients with AML, particularly those with t-aml, and in controls. We also studied the distribution of the XRCC1-194 and XPD-751 genotypes. In addition, because a malignant phenotype is likely to result from the accumulation of many minor genotypes, we assessed whether there was an association between DNA repair gene polymorphisms (XRCC1 and XRCC3) and a polymorphism in a protein involved with minimizing the effects of oxidative stress (NQO1). Materials and methods Study subjects Blood or bone marrow samples were obtained at the diagnosis of AML, after informed consent to participate was given according to the Declaration of Helsinki, from 168 white patients, including 34 with t-aml, presenting to the Haematology Department of Nottingham City Hospital. The diagnosis of AML was made by using the French-American-British criteria after conventional cytochemical and surface-marker analysis. The median age among patients with de novo AML was 63 years (range, 17-96 years); that among those with t-aml was 61.5 years (range, 37-88 years). Control peripheral blood samples were obtained from an equivalent number (n 178; median age, 51.5 years; range, 15-97 years) of whites with no known malignant diseases who resided in the same geographic community. Samples were not matched for sex; however, previous investigations consistently found no significant difference between male and female subjects in the distribution of the genotypes investigated in this study. 14,20 Genomic DNA was extracted from cells by using QIAamp blood DNA isolation kits (Qiagen, Crawley, United Kingdom) according to the manufacturer s protocol. Polymerase chain reaction (PCR) restriction fragment-length polymorphism (RFLP) genotyping analysis PCR followed by enzymatic digestion of the PCR products was used for genotyping the 3 polymorphisms. Approximately 50 ng genomic DNA was used as template in each of the PCR amplifications. The 50- L reaction also consisted of 150 M of each deoxynucleoside triphosphate (Amersham Biosciences, Little Chalfont, United Kingdom), 1 M of each primer, 1.5 to 2.0 mm/l magnesium chloride (MgCl 2 ), and 2 U Amplitaq Gold (PE Applied Biosystems, Warrington, United Kingdom) in the manufacturer s buffer. After an initial heat-activation step at 95 C for 10 minutes, amplification was performed in a PTC-100 programmable thermal controller (MJ Research, Watertown, MA). For XRCC1 Arg399Gln, XRCC1 Arg194Gln, XPD Gln751Lys, and XRCC3 Thr241Met, we used denaturation at 95 C for 1 minute, annealing at 60 C for 1 minute, and extension at 72 C for 1 minute for a total of 30 cycles ending with a final extension at 72 C for 10 minutes. For NQO1 Pro187Ser, we used denaturation at 94 C for 50 seconds, 52 C for 50 seconds, and 72 C for 30 seconds for a total of 35 cycles, followed by a final extension step at 72 C for 10 minutes. PCR products were purified by using a QIAquick PCR purification kit (Qiagen) before digestion and/or sequencing. The XRCC1 Arg399Gln polymorphism was amplified in a 616-bp fragment by using the following primers: xrcc1-399f 5 -TTGTGCTT- TCTCTGTGTCCA-3 and xrcc1-399r 5 -TCCTCCAGCCTTTACTGATA- 3 21 with 2 mm/l MgCl 2. The PCR product was digested with 10 U MspI (Helena Biosciences, Sunderland, United Kingdom) in the manufacturer s buffer at 37 C overnight. The recognition site for the MspI restriction endonuclease is present only in the Arg (WT) allele; hence, digestion of the Arg allele results in products of 376 bp and 240 bp, whereas the Gln allele remains undigested. A segment of the XRCC1 gene containing the Arg194Trp polymorphism was amplified by using the primer pairs xrcc1-194f 5 -GGTAAGCTGTAC- CTGTCACTC-3 and xrcc1-194r 5 -GACCCAGGAATCTGAGCC-3 with 1.5 mm/l MgCl 2. The PCR products were digested at 37 C overnight with 10 U MspI. The PCR product contains an internal MspI site, and products with the 194Arg genotype contain an additional MspI site, resulting in 20-, 117-, and 167-bp products. The 194Trp allele is digested only at the internal MspI site, resulting in 137- and 167-bp products. XRCC3 Thr241Met was amplified in a 415-bp product by using the primers xrcc3-241f 5 -GGTCGAGTGACAGTCCAAAC-3 and xrcc3-241r 5 -CTACCCGCAGGAGCCGGAGG-3 3 with 2 mm/l MgCl 2. The PCR products were digested at 37 C overnight with 10 U NlaIII (New England Biolabs, Hitchin, United Kingdom) in 1 buffer supplied with the enzyme and supplemented with 100 ng/ L bovine serum albumin. All XRCC3 PCR products contain an internal NlaIII site, and the presence of the Met polymorphism also generates an additional NlaIII site, resulting in 104-, 141-, and 170-bp products for the polymorphic allele and 141- and 274-bp products for the WT Thr allele. The XPD Lys751Gln polymorphism was amplified in a 344-bp fragment by using the following primers: xpd-751f 5 -TCAAACATCCT- GTCCCTACT-3 and xpd-751r 5 -CTGCCGATTAAAGGCTGTGGA-3 3 with 2 mm/l MgCl 2. The PCR product was digested with 10 U PstI (Helena Biosciences) at 37 C overnight. All PCR products contain an internal PstI site, resulting in 110- and 234-bp products in the 751Lys allele. In addition, an extra PstI site is present in the Gln allele, resulting in 63-, 110-, and 171-bp products. The NQO1 Pro187Ser polymorphism was amplified in a 304-bp fragment by using the primers nqo1f 5 -AAGCCCAGACCAACTTCT-3 and nqo1r 5 -TCTCCTCATCCTGTACCTCT-3 15 with 1.5 mm/l MgCl 2. The PCR products were digested overnight at 37 C with 10 U HinfI (New England Biolabs) in the manufacturer s buffer. All PCR products contain an internal HinfI site, and the Ser polymorphism also introduces an additional HinfI site, resulting in 33-, 120-, and 151-bp digested fragments, whereas the WT allele results in 33- and 271-bp products.

BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 POLYMORPHISMS IN AML 3763 The digested products were resolved on 3% agarose gels (Helena Biosciences), stained with ethidium bromide and analyzed under UV light. Two reviewers independently scored all genotypes, and samples that could not be scored were sequenced. Sequencing was also carried out on a 10% random sample population of control and AML PCR products. Sequencing reactions Sequencing reactions were set up with approximately 200 ng purified PCR product and 10 mm primer by using a Thermo Sequenase II dye terminator cycle sequencing premix kit according to the manufacturer s instructions (Amersham Biosciences). The primers used for sequencing were the same as those used for the PCR amplifications. The reactions were electrophoresed by using an ABI 377 automated DNA sequencer (PE Applied Biosystems, Foster City, CA) as recommended by the manufacturer. Statistical analysis The observed genotype frequencies of the 3 polymorphisms in the control cohorts were compared with those calculated by using the Hardy-Weinberg equilibrium (p2 q2 2pq 1; where p is the variant allele frequency). The distribution of genotypes in AML populations compared with the control population was assessed for significance by 2 testing, and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by logistic regression analysis and adjusted for the effect of age. P values of less than or equal to.05 were considered to represent significance. All analyses used the statistical package SPSS for Windows (version 9; SPSS, Chicago, IL). Results XRCC1 Arg399Gln, XRCC3 Thr241Met, and NQO1 Pro187Ser polymorphisms We examined the frequency of 3 polymorphisms in 134 patients with de novo AML (median age, 63 years) and 178 controls (median age, 51.5 years). We also analyzed a subgroup of 34 patients with t-aml (median age, 61.5 years). Because of an inadequate amount of DNA in some cases, several samples did not generate complete information for all 3 polymorphisms. Among the controls, the variant allele frequencies were 0.48 for XRCC1 399Gln, 0.30 for XRCC3 241Met, and 0.22 for NQO1 187Ser. All genotype frequencies in the control population were consistent with those expected from the Hardy-Weinberg equilibrium (XRCC1 Arg399Gln, 2 1.93 and P.38; XRCC3 Thr241Met, 2 1.22 and P.54; and NQO1 Pro187Ser, 2 0.94 and P.63). The distributions of the frequencies of the polymorphisms in AML cases and controls are shown in Table 1. The results for the XRCC1 Arg399Gln polymorphism showed that the proportion of AML and t-aml patients homozygous for the WT Arg allele was higher than in the control group, with the difference in the distribution of genotypes reaching statistical significance in the t-aml group (P.03 on 2 testing). Neither the XRCC3-241 nor the NQO1 genotype distributions showed any differences in either the de novo AML or t-aml group compared with the control group. The adjusted ORs for the individual genotypes are shown in Table 2. The XRCC1 399 homozygous variant genotype is a protective factor in both the de novo AML and t-aml patient groups. When the presence of at least one Gln allele is considered, it becomes apparent that the variant allele is significantly protective against the development of t-aml (OR, 0.46; 95% CI, 0.20-0.93). XRCC1 Arg194Trp and XPD Lys751Gln polymorphisms We also studied 2 additional polymorphisms, XRCC1 Arg194Trp and XPD Lys751Gln, in a smaller cohort of patients and controls. Table 1. Frequency of polymorphisms in AML and control populations Polymorphism Controls Patients with AML Patients with t-aml XRCC1-399 178 133 34 Arg/Arg 55 (30.9) 52 (39.1) 18 (52.9) Arg/Gln 76 (42.7) 57 (42.9) 12 (35.3) Gln/Gln 47 (26.4) 24 (18.0) 4 (11.8) P on 2 *.15.03 Arg/Arg 55 (30.9) 52 (39.1) 18 (52.9) Arg/Gln Gln/Gln 123 (69.1) 81 (60.9) 16 (47.1) P on 2 *.13.01 XRCC3-241 175 123 31 Thr/Thr 92 (52.6) 53 (43.1) 12 (38.7) Thr/Met 64 (36.6) 53 (43.1) 12 (38.7) Met/Met 19 (10.9) 17 (13.8) 7 (22.6) P on 2 *.27.14 Thr/Thr 92 (52.6) 53 (43.1) 12 (38.7) Thr/Met Met/Met 83 (47.4) 70 (56.9) 19 (61.3) P on 2 *.11.16 NQO1 175 134 33 Pro/Pro 110 (62.9) 95 (70.9) 19 (57.6) Pro/Ser 53 (30.3) 30 (22.4) 14 (42.4) Ser/Ser 12 (6.9) 9 (6.7) 0 (0) P on 2 *.29.16 Pro/Pro 110 (62.9) 95 (70.9) 19 (57.6) Pro/Ser Ser/Ser 65 (37.1) 39 (29.1) 14 (42.4) P on 2 *.14.57 XRCC1-194 87 112 14 Arg/Arg 78 (89.7) 100 (89.3) 12 (85.7) Arg/Trp 7 (8.0) 12 (10.7) 2 (14.3) Trp/Trp 2 (2.3) 0 (0) 0 (0) P on 2 *.23.65 Arg/Arg 78 (89.7) 100 (89.3) 12 (85.7) Arg/Trp Trp/Trp 9 (10.3) 12 (10.7) 2 (14.3) P on 2 *.93.66 XPD-751 73 107 15 Lys/Lys 30 (41.1) 40 (37.4) 4 (26.7) Lys/Gln 32 (43.8) 51 (47.7) 8 (53.3) Gln/Gln 11 (15.1) 16 (15.0) 3 (20) P on 2 *.86.58 Lys/Lys 30 (41.1) 40 (37.4) 4 (26.7) Lys/Gln Gln/Gln 43 (58.9) 67 (62.6) 11 (73.3) P on 2 *.62.30 Values are numbers (%) of controls or patients unless otherwise indicated. *For comparisons between patients and controls. Among the controls, the variant allele frequencies were 0.07 for XRCC1 194Trp and 0.38 for XPD 751Gln. The genotype distributions among the control population were consistent with those expected from the Hardy-Weinberg equilibrium (XRCC1 Arg194Trp, 2 2.49 and P.29; and XPD Lys751Gln, 2 0.25 and P.88). The distributions of the frequencies of the polymorphisms in AML patients and controls are shown in Table 1; adjusted ORs are shown in Table 2. No significant differences were observed between the AML patient cohort and the controls. Combined analysis of polymorphisms Table 3 shows the combined analysis for either the XRCC1 Arg399Gln or XRCC3 Thr241Met genotypes with the NQO1 Pro187Ser genotype. Because of the small numbers of t-aml samples available, these were grouped with the samples from the de novo AML cases for this analysis. The DNA repair gene polymorphisms did not show an interaction with the detoxification NQO1 Pro187Ser polymorphism.

3764 SEEDHOUSE et al BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 Table 2. Estimates of relative risk for AML associated with various genotypes Genotype Controls Patients with AML OR (95% CI)* P Patients with t-aml OR (95% CI)* P XRCC1-399 178 133 34 Arg/Arg 55 52 1.0 18 1.0 Arg/Gln 76 57 0.84 (0.50-1.42).52 12 0.54 (0.24-1.23).14 Gln/Gln 47 24 0.57 (0.30-1.08).09 4 0.28 (0.09-0.88).03 Arg/Gln Gln/Gln 123 81 0.74 (0.46-1.20).22 16 0.44 (0.20-0.93).03 XRCC3-241 175 123 31 Thr/Thr 92 53 1.0 12 1.0 Thr/Met 64 53 1.50 (0.90-2.51).12 12 1.55 (0.64-3.73).33 Met/Met 19 17 1.49 (0.70-3.20).31 7 2.66 (0.89-7.94).08 Thr/Met Met/Met 83 70 1.50 (0.93-2.42).10 19 1.82 (0.82-4.05).14 NQO1 175 134 33 Pro/Pro 110 95 1.0 19 1.0 Pro/Ser 53 30 0.66 (0.39-1.13).13 14 1.62 (0.74-3.54).22 Ser/Ser 12 9 0.71 (0.26-1.89).49 0 Pro/Ser Ser/Ser 65 39 0.67 (0.41-1.10).11 14 1.34 (0.62-2.91).45 XRCC1-194 87 114 14 Arg/Arg 78 100 1.0 12 1.0 Arg/Trp 7 12 2.79 (0.77-10.05).12 2 0.13 (0.00-43.31).49 Trp/Trp 2 0 0 Arg/Trp Trp/Trp 9 12 2.11 (0.64-6.98).22 2 0.12 (0.00-40.48).48 XPD-751 73 107 15 Lys/Lys 30 40 1.0 4 1.0 Lys/Gln 32 51 0.74 (0.31-1.77).50 8 9.66 (0.78-119.57).08 Gln/Gln 11 16 0.61 (0.18-2.04).42 3 1.13 (0.04-28.80).94 Lys/Gln Gln/Gln 43 67 0.70 (0.31-1.58).39 11 5.30 (0.54-52.26).15 *Adjusted for age. Used as the reference group. Discussion Although it has been established that mismatch repair is important in a subset of patients with AML, 22-24 little is known about the role of the other DNA repair pathways in this disease. In this study, we found that a polymorphism in a DNA repair gene belonging to the BER pathway, XRCC1 Arg399Gln, is associated with a protective effect against the development of AML, particularly t-aml. Hence, patients in whom AML develops as a result of therapy for a primary malignant disease are more likely to have the WT XRCC1 399 Arg allele. It is possible that differences between the cases and controls may have contributed to the observed association, but we believe that this is unlikely. Both populations were from the same small geographical area in the United Kingdom and had the same age range. Sex has been shown not to have any influence on the XRCC1 Arg399Gln genotype, 20 and although race was previously found to be an important factor in the distribution of the XRCC1 Arg399Gln genotype, 25,26 all our patients and controls were white. We did not find any differences between the distributions of the XRCC3 Thr241Met and NQO1 Pro187Ser genotypes. Others have observed a significant increase in the incidence of the variant NQO1 allele in patients with AML, particularly those with chromosome 5/7 abnormalities. 19 This association was not evident in our AML groups, although the number of patients with chromosome 5/7 abnormalities in our cohort was small and that may have accounted for the difference from previous studies. We were able to obtain cytogenetic information for approximately 100 of our patients, and only 10 of them had chromosome 5/7 abnormalities. We combined the de novo AML and t-aml groups to examine possible associations between either the XRCC1 Arg399Gln or XRCC3 Thr241Met genotypes and the NQO1 Pro187Ser genotypes. No significant association was observed, Table 3. Combined logistic regression analysis of the association between DNA repair and NQO1 genotype and the risk of AML Repair genotype Detoxification genotype Controls Patients with AML* OR (95% CI) P XRCC1-399 NQO1-187 WT WT 34 45 1 V WT 71 60 0.64 (0.36-1.14).13 WT V 18 18 0.74 (0.33-1.66).46 V V 43 33 0.62 (0.32-1.18).14 XRCC3-241 NQO1-187 WT WT 50 40 1 V WT 53 57 1.47 (0.83-2.62).19 WT V 36 19 0.76 (0.37-1.56).46 V V 24 24 1.27 (0.62-2.63).51 WT indicates wild-type genotype; V, presence of a variant allele (either heterozygous or homozygous). *The AML group in this analysis included both de novo AML and t-aml cases. Adjusted for age. Used as the reference group.

BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 POLYMORPHISMS IN AML 3765 suggesting that the phenotypes resulting from these proteins do not interact to increase the risk of AML. However, our number of t-aml samples was too small to allow us to conduct combined logistic regression analysis of this group alone. We also performed PCR-RFLP analysis of a smaller number of patients and controls to study the distribution of genotypes of an additional polymorphism in XRCC1 (XRCC1 Arg194Trp) and a polymorphism in XPD (XPD Lys751Gln), a gene involved in the nucleotide excision repair pathway. No significant differences were found between controls and patients with AML. Several previous studies assessed the functional relevance of polymorphisms in DNA repair genes. Lunn et al 21 measured genotoxic end points of DNA damage and showed that the XRCC1 399Gln allele was associated with increased placental aflatoxin DNA adducts and increased glycophorin A mutations in erythrocytes. They suggested that the Arg to Gln change at codon 399 may alter the phenotype of the XRCC1 protein, resulting in deficient DNA repair. In addition, Duell et al 27 measured DNA damage by using a sister chromatid exchange assay and also detected polyphenol DNA adducts. They found more damage in current smokers homozygous for the XRCC1 399Gln allele than in current smokers with the homozygous WT allele. A sister chromatid assay was also used by Abdel-Rahman and El-Zein, 28 who found that a higher level of DNA damage after treatment with a tobacco-specific nitrosamine occurred in XRCC1 399Gln homozygotes. Several polymorphism studies have observed a positive association of the XRCC1 399 Gln allele with various malignant diseases, including cancer of the head and neck, 20 breast, 25 lung, 26 and colon and rectum. 29 The researchers suggested that these results correlated with the in vitro functional data and illustrated the adverse effects of deficient DNA repair systems. However, other studies, including ours, had conflicting results. We found that the presence of the XRCC1 399Gln allele is protective against the development of t-aml. The same protective effect has been observed in studies of bladder cancer 30 and nonmelanoma skin cancer. 31 This may indicate the presence of a strong gene-environment interaction that is tumor specific. There is a fine balance between deficient and functional DNA repair systems. Although it is important for damaged DNA to be recognized and repaired, it is also vital for the cell to be able to recognize when the damage is too extensive to be repaired and to allow the apoptotic pathway to be stimulated; this prevents cells from potentially misrepairing damage and surviving with mutations. This process may be particularly important in the development of t-aml. Chemotherapy and radiotherapy both induce immense amounts of DNA damage and have the aim of achieving cell death. Even subtle changes in DNA repair capacity References are likely to be important when large external influences such as chemotherapy or radiotherapy are present. We hypothesize that when hematopoietic progenitor cells in the bone marrow are damaged by therapy, the cells with the XRCC1 399 Gln allele and resulting reduced DNA repair capacity are more likely to be driven toward apoptosis. This is contrary to what happens to cells with the WT genotype, which are more likely to repair their damage, possibly harbor mutations, and initiate clonal disease resulting in t-aml. Hence, we suggest the XRCC1 399 Gln allele is protective against the development of t-aml. A similar proposal was made by Nelson et al 31 in their large study of nonmelanoma skin cancer. Our hypothesis can be extended to the opposite situation observed in many primary malignant diseases in which the XRCC1 399Gln allele has been shown to be a risk factor. Although DNA damage caused by low-dose exogenous sources (for example, smoking and diet) is considered to be an important contributory factor in many malignant diseases, the resulting oxidative burden on the cells is significantly smaller than that following chemotherapy or radiotherapy, which can be considered to cause highdose damage. Hence, a BER pathway with a fully functional XRCC1 gene (XRCC1 399Arg) would be expected to repair the damage caused by low-dose damaging agents (smoking and diet), whereas a deficiency in the DNA repair pathway, such as an XRCC1 399Gln allele, would reduce the effectiveness of the pathway in repairing damage from low-dose sources and thus increase the risk of a primary malignant disease. Additional work is now required to determine whether polymorphisms do indeed lead to a reduced DNA repair capacity in vivo and to identify the consequences of these phenotypes in different environmental conditions. The identification of polymorphisms or mutations in many genes and the determination of their functional importance in AML will allow susceptibility-risk models to be designed for de novo and therapy-related disease. Intervention strategies and early detection approaches could then be targeted at those individuals genetically identified to be at higher risk of AML or t-aml. Acknowledgments We thank Tricia McKeever for assistance with the statistical analysis; Stephen Langabeer and the Kay Kendall Leukaemia Fund for t-aml DNA samples from the DNA/RNA banking facilities at University College Hospital, London; and the Medical Research Council (MRC) Acute Leukaemia Working Party for access to samples from patients with AML entered into MRC trials. 1. Pedersen-Bjergaard J, Andersen MK, Christiansen DH, Nerlov C. Genetic pathways in therapy-related myelodysplasia and acute myeloid leukemia. Blood. 2002;99:1909-1912. 2. Mohrenweiser HW, Jones IM. Variation in DNA repair is a factor in cancer susceptibility: a paradigm for the promises and perils of individual and population risk estimation? Mutat Res. 1998;400: 15-24. 3. Shen MR, Jones IM, Mohrenweiser H. Nonconservative amino acid substitution variants exist at polymorphic frequency in DNA repair genes in healthy humans. Cancer Res. 1998;58:604-608. 4. Kubota Y, Nash RA, Klungland A, Schar P, Barnes DE, Lindahl T. Reconstitution of DNA base excision-repair with purified human proteins: interaction between DNA polymerase and the XRCC1 protein. EMBO J. 1996;15:6662-6670. 5. Caldecott KW, McKeown CK, Tucker JD, Ljungquist S, Thompson LH. An interaction between the mammalian DNA repair protein XRCC1 and DNA ligase III. Mol Cell Biol. 1994;14:68-76. 6. Caldecott KW, Aoufouchi S, Johnson P, Shall S. XRCC1 polypeptide interacts with DNA polymerase and possibly poly (ADP-ribose) polymerase, and DNA ligase III is a novel molecular nicksensor in vitro. Nucleic Acids Res. 1996;24:4387-4394. 7. Masson M, Niedergang C, Schreiber V, Muller S, Menissier-de Murcia J, de Murcia G. XRCC1 is specifically associated with poly(adp-ribose) polymerase and negatively regulates its activity following DNA damage. Mol Cell Biol. 1998;18: 3563-3571. 8. Lamerdin JE, Montgomery MA, Stilwagen SA, et al. Genomic sequence comparison of the human and mouse XRCC1 DNA repair gene regions. Genomics. 1995;25:547-554. 9. Shen MR, Zdzienicka MZ, Mohrenweiser H, Thompson LH, Thelen MP. Mutations in hamster single-strand break repair gene XRCC1 causing defective DNA repair. Nucleic Acids Res. 1998; 26:1032-1037. 10. Bishop DK, Ear U, Bhattacharyya A, et al. Xrcc3 is required for assembly of Rad51 complexes in vivo. J Biol Chem. 1998;273:21482-21488. 11. Matullo G, Guarrera S, Carturan S, et al. DNA repair gene polymorphisms, bulky DNA adducts in white blood cells and bladder cancer in a casecontrol study. Int J Cancer. 2001;92:562-567. 12. Winsey SL, Haldar NA, Marsh HP, et al. A variant within the DNA repair gene XRCC3 is associated with the development of melanoma skin cancer. Cancer Res. 2000;60:5612-5616.

3766 SEEDHOUSE et al BLOOD, 15 NOVEMBER 2002 VOLUME 100, NUMBER 10 13. Broughton BC, Steingrimsdottir H, Lehmann AR. Five polymorphisms in the coding sequence of the xeroderma pigmentosum group D gene. Mutat Res. 1996;15:209-211. 14. Matullo G, Palli D, Peluso M, et al. XRCC1, XRCC3, XPD gene polymorphisms, smoking and (32)P-DNA adducts in a sample of healthy subjects. Carcinogenesis. 2001;22:1437-1445. 15. Lunn RM, Helzlsouer KJ, Parshad R, et al. XPD polymorphisms: effects on DNA repair proficiency. Carcinogenesis. 2000;21:551-555. 16. Rosvold EA, McGlynn KA, Lustbader ED, Buetow KH. Identification of an NAD(P)H:quinone oxidoreductase polymorphism and its association with lung cancer and smoking. Pharmacogenetics. 1995;5:199-206. 17. Siegel D, McGuinness SM, Winski SL, Ross D. Genotype-phenotype relationships in studies of a polymorphism in NAD(P)H:quinone oxidoreductase 1. Pharmacogenetics. 1999;9:113-121. 18. Smith MT, Wang Y, Kane E, et al. Low NAD(P)H: quinone oxidoreductase 1 activity is associated with increased risk of acute leukemia in adults. Blood. 2001;97:1422-1426. 19. Larson RA, Wang Y, Banerjee M, et al. Prevalence of the inactivating 609C3T polymorphism in the NAD(P)H:quinone oxidoreductase (NQO1) gene in patients with primary and therapy-related myeloid leukemia. Blood. 1999;94:803-807. 20. Sturgis EM, Castillo EJ, Li L, et al. Polymorphisms of DNA repair gene XRCC1 in squamous cell carcinoma of the head and neck. Carcinogenesis. 1999;20:2125-2129. 21. Lunn RM, Langlois RG, Hsieh LL, Thompson CL, Bell DA. XRCC1 polymorphisms: effects on aflatoxin B1-DNA adducts and glycophorin A variant frequency. Cancer Res. 1999;59:2557-2561. 22. Ben-Yehuda D, Krichevsky S, Caspi O, et al. Microsatellite instability and p53 mutations in therapy-related leukemia suggest mutator phenotype. Blood. 1996;88:4296-4303. 23. Zhu YM, Das-Gupta EP, Russell NH. Microsatellite instability and p53 mutations are associated with abnormal expression of the MSH2 gene in adult acute leukemia. Blood. 1999;94:733-740. 24. Das-Gupta EP, Seedhouse CH, Russell NH. Microsatellite instability occurs in defined subsets of patients with acute myeloblastic leukaemia. Br J Haematol. 2001;14:307-312. 25. Duell EJ, Millikan RC, Pittman GS, et al. Polymorphisms in the DNA repair gene XRCC1 and breast cancer. Cancer Epidemiol Biomarkers Prev. 2001;10:217-222. 26. Divine KK, Gilliland FD, Crowell RE, et al. The XRCC1 399 glutamine allele is a risk factor for adenocarcinoma of the lung. Mutat Res. 2001; 461:273-278. 27. Duell EJ, Wiencke JK, Cheng TJ, et al. Polymorphisms in the DNA repair genes XRCC1 and ERCC2 and biomarkers of DNA damage in human blood mononuclear cells. Carcinogenesis. 2000;21:965-971. 28. Abdel-Rahman SZ, El-Zein RA. The 399Gln polymorphism in the DNA repair gene XRCC1 modulates the genotoxic response induced in human lymphocytes by the tobacco-specific nitrosamine NNK. Cancer Lett. 2000;159:63-71. 29. Abdel-Rahman SZ, Soliman AS, Bondy ML, et al. Inheritance of the 194Trp and the 399Gln variant alleles of the DNA repair gene XRCC1 are associated with increased risk of early-onset colorectal carcinoma in Egypt. Cancer Lett. 2000;159:79-86. 30. Stern MC, Umbach DM, van Gils CH, Lunn RM, Taylor JA. DNA repair gene XRCC1 polymorphisms, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev. 2001;10:125-131. 31. Nelson HH, Kelsey KT, Mott LA, Karagas MR. The XRCC1 Arg399Gln polymorphism, sunburn, and non-melanoma skin cancer: evidence of gene-environment interaction. Cancer Res. 2002; 62:152-155.