Purpose: Numerous chemotherapeutic agents are cytotoxic through generation of reactive species, and variability in genes related to oxidative stress may influence disease-free survival (DFS). We examined relationships between DFS and variants in NOS3, as well as NQO1, NQO2, and CBR3, among treated and untreated breast cancer patients in a Southwest Oncology Group clinical trial (S8897).

Experimental Design: In the parent trial, women were assigned according to prognostic features; the high-risk group was randomized to cyclophosphamide, i.v. methotrexate, and 5-fluorouracil or to cyclophosphamide, i.v. doxorubicin, and 5-fluorouracil ± tamoxifen, and the low-risk group did not receive adjuvant therapy. We extracted DNA from normal lymph node tissue and examined functional polymorphisms in NOS3, NQO1, NQO2, and CBR3, in relation to DFS, using Cox proportional hazard model.

Results: There were significant interactions between DFS, adjuvant therapy, and NOS3 Glu298Asp and −786 polymorphisms, alone and in combination (P for interaction = 0.008). When NOS3 genotypes were combined, women with genotypes encoding for lower nitric oxide who received chemotherapy had a >2-fold increase in hazard of progression (hazard ratio, 2.32; 95% confidence interval, 1.26-4.25), whereas there was reduced risk for those who did not receive adjuvant therapy (hazard ratio, 0.42; 95% confidence interval, 0.19-0.95). There were no associations between the other genotypes and DFS in either group.

Conclusion: Variants encoding lower activity of NOS3 may affect outcomes in breast cancer patients, with the direction of risk differing depending on chemotherapy status. These results may mirror the known dual functions of nitric oxide and nitric oxide synthase, depending on oxidative environment. (Clin Cancer Res 2009;15(16):5258–66) (Clin Cancer Res 2009;15(16):5258–)

Translational Relevance

In this ancillary study to a clinical trial, germ-line variants encoding lower activity of NOS3 are related to better prognosis in untreated patients, but conversely, with worse outcomes in patients who are treated with chemotherapy, perhaps by altering sensitivity normally induced by NOS3-generated production of nitric oxide. The heterogeneity of effects by whether or not patients received chemotherapy cautions against studies of predictors of breast cancer survival among groups of patients with varied treatments, or for whom treatment regimens are unknown, because the effects can vary dramatically based on treatment status. This growing body of evidence of the important role of endogenous oxidants in chemotherapy treatment outcomes warrants increased investigation of potential interactions with supplement use by patients undergoing chemotherapy treatment, and may add weight to the recommendations that patients receiving adjuvant chemotherapy should be cautioned about the use of antioxidant supplements.

Breast cancer is the most commonly diagnosed cancer and the second leading cause of cancer-related death in American women (1, 2). Adjuvant systemic therapy, in particular chemotherapy and endocrine therapy for hormone receptor–positive patients, reduces the risk of recurrence and mortality for women with early-stage disease (3). However, only a fraction of patients benefit from adjuvant systemic treatment, either because their prognosis is so favorable that they are not likely to relapse without treatment or because the therapy is ineffective. Many somatic factors, such as nodal status, tumor size and grade, and expression of hormone receptors and HER2, are used to determine prognosis and prediction for specific therapies (4). The role of inherited genetic factors has not been thoroughly investigated in relation to breast cancer treatment outcomes.

Cyclophosphamide and anthracyclines exert their cytotoxic effects through apoptosis via intrinsic mitochondrial pathways, mediated through the production of reactive oxygen species (ROS) and concomitant oxidative stress (5). As an activated analogue of cyclophosphamide, 4-hydroperoxy-cyclophosphamide mediates caspase-independent T-cell apoptosis involving oxidative stress–induced nuclear relocation of mitochondrial apoptosis (5). Murata and colleagues showed generation of hydrogen peroxide during degradation of 4-hydroperoxy-cyclophosphamide and showed that blocking this by addition of catalase abrogated DNA damage in a leukemic cell line (6). It has also been shown that cyclophosphamide-mediated apoptosis in human granules cells in culture involves oxidative stress and glutathione depletion (7). Among various combinations of intracellular mechanisms for cytotoxic effects of anthracyclines, a number of oxidoreductases catalyze the one-electron reduction of the quinone moiety of anthracyclines (ring C) to generate a semiquinone free radical that covalently interacts with and damages DNA. In the presence of oxygen, this free radical can be oxidized back to the quinone, concomitantly generating ROS such as superoxide. Doxorubicin semiquinone can also be subjected to reductive deglycosidation to form 7-deoxyaglycone, which in turn forms ROS by intercalating into biological membranes (reviewed in ref. 8).

Thus, common variants in genes that generate or protect from ROS may play a role in breast cancer treatment outcomes. Among them, endothelial nitric oxide synthase (NOS3), NAD(P)H dehydrogenase, quinone 1 (NQO1), NAD(P)H dehydrogenase, quinone 2 (NQO2), and carbonyl reductase 3 (CBR3) are plausible candidates because of their potential role during the detoxification of quinones and oxidative stress (reviewed in ref. 9). NOS3 generates low amounts of short-lived nitric oxide (NO) by converting l-arginine to citrulline in endothelial tissue (10). The NOS3 786 T > C promoter polymorphism results in reduced promoter activity, and a 894 G > T polymorphism in NOS3 exon 7 results in an amino acid substitution that alters susceptibility to cleavage, leading to decreased endothelial NO production (1113). Similar to NOS3, NQO1 and NQO2, as oxidoreductases, catalyze the one-electron reduction of the quinone moiety of anthracyclines (doxorubicin), which is associated with metabolic detoxification or activation of quinones and quinone-based antitumor drugs (8, 14). A 609 C > T polymorphism in NQO1 leads to a proline-to-serine substitution (15) that results in rapid degradation of the variant enzyme (16, 17). CBR activity metabolizes a broad spectrum of xenobiotic carbonyl compounds, including various pharmacologic agents (18) such as haloperidol and doxorubicin (19). CBRs catalyze the two-electron reduction of the side-chain C-13 carbonyl group to form anthracycline-alcohol metabolites, transforming doxorubicin to doxorubicinol, which has diminished antineoplastic activity (8, 18). Variable cytosolic CBR activities have been documented in breast and lung tumors, as well as in human liver (20, 21). Blanco and colleagues found that a common genetic polymorphism in CBR3 (CBR3 V244M) results in protein isoforms with distinctive catalytic properties toward the quinone menadione, doxorubicin, and the NADP(H) cofactor (22).14

In this study, in the context of a completed clinical trial [Southwest Oncology Group (SWOG) 8897], we investigated whether women with functional genetic polymorphisms that result in less protection from chemotherapy-induced oxidative damage would presumably have better tumor cell kill and, subsequently, less disease recurrence than those with common alleles. We hypothesized that genotypes would affect outcomes of patients treated with chemotherapy through indirect effects on treatment-induced cytotoxicity, but would have no effects among untreated patients.

Clinical information

These analyses were conducted within the context of a completed clinical trial for breast cancer (S8897), which was led by SWOG within the North American Breast Cancer Intergroup (INT0102). Methods and results are reported according to the REMARK criteria (23). Complete details of S8897 have been reported elsewhere (24). As shown in Fig. 1, women with node-negative breast cancer who were eligible for the trial were assigned to groups based on presumed risk of recurrence (low, indeterminate, high) according to standard prognostic features (tumor size and hormone receptor status). Those in the indeterminate group were then assigned to either the low-risk or the high-risk group based on flow cytometry results. Those in the low-risk group (n = 1208) were assigned to follow-up with no adjuvant therapy, and those in the high-risk group (n = 1153) were randomly assigned to adjuvant chemotherapy consisting of cyclophosphamide, i.v. methotrexate, and 5-fluorouracil (CMF) or cyclophosphamide, i.v. doxorubicin, and 5-fluorouracil (CAF). CMF chemotherapy was administered per guidelines established by Bonadonna et al. (25), with cyclophosphamide 100 mg/m2 orally on days 1 to 14, methotrexate 40 mg/m2 i.v. on days 1 and 8, fluorouracil 600 mg/m2 i.v. on days 1 and 8, repeated at 28-d intervals for six cycles; and CAF chemotherapy per guidelines established by Bull et al. (26), with cyclophosphamide 100 mg/m2 orally on days 1 to 14, doxorubicin 30 mg/m2 i.v. on days 1 and 8, fluorouracil 500 mg/m2 i.v. on days 1 and 8, repeated at 28-d intervals for six cycles. Doses were delayed for up to 2 wk for granulocytopenia or thrombocytopenia, and then given at reduced levels if granulocytopenia or thrombocytopenia persisted. Doses on day 8 were decreased or withheld accordingly.

Fig. 1.

Schema in SWOG S8897.

Fig. 1.

Schema in SWOG S8897.

Close modal

Hematopoietic growth factors were not used. In tamoxifen arms, tamoxifen 20 mg/d was started on day 29 of cycle 6 and was continued for 5 y. Distributions of age, race, and type of primary surgery were similar among all treatment groups. The end points for the trial were disease-free survival (DFS) as defined prospectively in the S8897 protocol. DFS was calculated from the date of randomization to the date of first recurrence or death due to any cause, and included local, regional, or distant recurrence or a new breast primary, but not a new nonbreast primary cancer. As previously reported (24), the median follow-up time was 10.8 y, and 10-y estimates indicated that CAF was not significantly better than CMF for DFS.

Specimens and genotyping

As shown in Fig. 1 and as previously reported (27), only patients with archived, uninvolved lymph node tissue available in the SWOG tissue bank from which DNA extraction could be done were included in the current genotyping study. This group included women initially assigned to the low-risk group (n = 1208), and those who were in the indeterminate group, subsequently assigned to low-risk (n = 547) or high-risk (n = 528) category. Thus, participants in the high-risk group included in our analyses all had hormone receptor–positive tumors, generally between 1 to 2 cm in size, whereas those in the low-risk group had either hormone receptor–positive tumors or tumors too small for evaluation. Patients consented to the use of their tissue for ancillary research studies, and this pharmacogenetic study was approved by the Institutional Review Board at Roswell Park Cancer Institute. Two 5-μm slides of normal lymph node tissue were procured from the SWOG tissue bank in San Antonio, were deparaffinized, and removed from slides; DNA was extracted as previously described (28). DNAs were plated in a blinded fashion for genotyping, with duplicates from 5% of samples included for quality control. Genotyping for NOS3, NQO1, NQO2, and CBR3 was done using Sequonom high-throughput matrix assisted laser desorption/ionization time-of-flight mass spectrometry. A total of 458 high-risk and 874 low-risk participants had sufficient amplifiable DNA, and genotyping rates were ≥97.5% for all polymorphisms assessed, with differences in genotyping rates between those in the treated and those in the untreated arms not exceeding 3.2%.

Statistical analyses

The effects of genetic polymorphisms on DFS were evaluated by contrasting those with genotypes associated with higher ROS, assumed to have better tumor cell kill and thus better survival, against women with lower oxidative stress–related genotypes. Kaplan-Meier estimates of DFS for the subgroups of patients with common and variant alleles were generated, with differences tested using log-rank tests. Cox regression models were used to estimate hazard ratios (HR) and 95% two-sided confidence intervals, with adjustments for age (continuous), race (Caucasian versus non-Caucasian), menopausal status, time from initial surgery to treatment in the treated group, and randomization assignment (CAF versus CMF; tamoxifen versus no tamoxifen) in the treated group. Separate analyses were fit for the treated (high-risk) and untreated (low-risk) groups. We also used a composite model that tested an explicit interaction between treatment and genotypes on DFS. We further evaluated if relationships varied by whether women received CAF or CMF or were assigned to tamoxifen or not. To evaluate whether multiple polymorphisms in the same pathway may have synergistic or additive effects, “high-risk” alleles of each gene were tallied. The number of high-risk alleles for each individual ranged from 0 to 12 and was subsequently categorized as 0 to 3, 4, 5, and ≥6.

The characteristics of the SWOG S8897 participants for whom genotyping and vital status data were available are shown in Table 1. Women in the treated group tended to be younger, premenopausal, and to have larger tumor size than women in the untreated group. When examining associations between genotypes and DFS, we noted a statistically significant interaction (P for interaction = 0.003) between treatment status and the NOS3 Glu298Asp (G to T) polymorphism, with diametrically opposite associations between the treated and untreated groups. As shown in Fig. 2, women who were homozygous for T alleles, associated with lower activity, and who received no adjuvant systemic treatment had better survival than those with GG/GT genotypes [adjusted HR, 0.67; 95% confidence interval (95% CI), 0.42-1.07; Table 2]. In contrast, women treated with CMF or CAF who were homozygous for the variant TT genotype had poorer survival compared with those with GG/GT genotypes (HR, 1.76; 95% CI, 1.14-2.92; Table 2). A second promoter polymorphism in NOS3 (−786 T to C) was also associated with better survival among women who did not receive treatment (HR, 0.59; 95% CI, 0.38-0.91), but had no effect in the treated arm (HR, 1.15; 95% CI, 0.69-1.91; P for interaction = 0.049; Table 2). When NOS3 genotypes were combined, associations were further strengthened. As shown in Table 3, women who did not receive adjuvant therapy and had both low-activity genotypes (NOS3 −786 CC and 894 TT) were at significantly decreased risk of recurrence (HR, 0.42; 95% CI, 0.19-0.95) compared with women with other genotypes (−786 TT + TC and 894 GG + GT). Women assigned to chemotherapy who had NOS3 −786 CC and 894 TT genotypes had a >2-fold increase in risk of recurrence compared with those with common alleles (HR, 2.32; 95% CI, 1.26-4.25; P for interaction = 0.008).

Table 1.

Demographic and clinical characteristics of patients participating in SWOG 8897 from whom DNA was available

Characteristic% Treated group (n = 458)% Untreated group (n = 874)
Age 
    <40 15 
    40-49 37 31 
    50-59 23 25 
    60-69 19 24 
    ≥70 12 
    Median (range) 49 (27-85) 54 (24-89) 
Race 
    White, non-Hispanic 88 92 
    Black, non-Hispanic 
    Hispanic 
    Other 
Menopausal status 
    Premenopausal 51 38 
    Postmenopausal 49 62 
Postmenopausal estrogen 
    Yes 14 18 
    No 86 82 
Primary treatment 
    BSP, delayed RT 18 23 
    BSP, RT before registration 12 13 
    Mastectomy 70 64 
Receptor status 
    Receptor positive 100 − 
    ER and PgR negative − 
Node involvement 
    No 100 100 
    Yes 
Tumor size, cm 
    ≤1 25 59 
    1.1-1.9 75 40 
Characteristic% Treated group (n = 458)% Untreated group (n = 874)
Age 
    <40 15 
    40-49 37 31 
    50-59 23 25 
    60-69 19 24 
    ≥70 12 
    Median (range) 49 (27-85) 54 (24-89) 
Race 
    White, non-Hispanic 88 92 
    Black, non-Hispanic 
    Hispanic 
    Other 
Menopausal status 
    Premenopausal 51 38 
    Postmenopausal 49 62 
Postmenopausal estrogen 
    Yes 14 18 
    No 86 82 
Primary treatment 
    BSP, delayed RT 18 23 
    BSP, RT before registration 12 13 
    Mastectomy 70 64 
Receptor status 
    Receptor positive 100 − 
    ER and PgR negative − 
Node involvement 
    No 100 100 
    Yes 
Tumor size, cm 
    ≤1 25 59 
    1.1-1.9 75 40 

Abbreviations: ER, estrogen receptor; PgR, progesterone receptor; BSP, breast sparing surgery; RT, radiation therapy.

Fig. 2.

Disease-free survival by NOS genotypes.

Fig. 2.

Disease-free survival by NOS genotypes.

Close modal
Table 2.

Disease-free survival hazard ratios for NOS3 genotypes among women randomized to CMF or CAF ± tamoxifen and among women not receiving adjuvant therapy

GenotypesTreated (n = 458)Untreated (n = 874)
CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
NOS3 −786 T > C   n = 445   n = 862 
    TT 123 (38.2) 56 (45.5) 1.0 230 (37.7) 112 (44.4) 1.0 
    TC 154 (47.8) 49 (39.8) 0.74 (0.50-1.12) 278 (45.6) 117 (46.4) 0.89 (0.68-1.15) 
    CC 45 (14.0) 18 (14.6) 0.98 (0.57-1.70) 102 (16.7) 23 (9.1) 0.55 (0.35-0.87) 
    Ptrend   0.575   0.015 
    Pinteraction      0.089 
    TT + TC 277 (86.0) 105 (85.4) 1.0 508 (83.3) 229 (90.9) 1.0 
    CC 45 (14.0) 18 (14.6) 1.15 (0.69-1.91) 102 (16.7) 23 (9.1) 0.59 (0.38-0.91) 
    Pinteraction      0.049 
NOS3 894 G > T   n = 438   n = 862 
    GG 154 (48.7) 65 (53.3) 1.0 290 (47.5) 132 (52.4) 1.0 
    GT 131 (41.5) 36 (29.5) 0.71 (0.47-1.08) 241 (39.5) 101 (40.1) 0.97 (0.75-1.26) 
    TT 31 (9.8) 21 (17.2) 1.54 (0.94-2.53) 79 (13.0) 19 (7.5) 0.66 (0.41-1.07) 
    Ptrend   0.526   0.171 
    Pinteraction      0.005 
    GG + GT 285 (90.2) 101 (82.8) 1.0 531 (87.1) 233 (92.5) 1.0 
    TT 31 (9.8) 21 (17.2) 1.76 (1.10-2.83) 79 (13.0) 19 (7.5) 0.67 (0.42-1.07) 
    Pinteraction      0.003 
GenotypesTreated (n = 458)Untreated (n = 874)
CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
NOS3 −786 T > C   n = 445   n = 862 
    TT 123 (38.2) 56 (45.5) 1.0 230 (37.7) 112 (44.4) 1.0 
    TC 154 (47.8) 49 (39.8) 0.74 (0.50-1.12) 278 (45.6) 117 (46.4) 0.89 (0.68-1.15) 
    CC 45 (14.0) 18 (14.6) 0.98 (0.57-1.70) 102 (16.7) 23 (9.1) 0.55 (0.35-0.87) 
    Ptrend   0.575   0.015 
    Pinteraction      0.089 
    TT + TC 277 (86.0) 105 (85.4) 1.0 508 (83.3) 229 (90.9) 1.0 
    CC 45 (14.0) 18 (14.6) 1.15 (0.69-1.91) 102 (16.7) 23 (9.1) 0.59 (0.38-0.91) 
    Pinteraction      0.049 
NOS3 894 G > T   n = 438   n = 862 
    GG 154 (48.7) 65 (53.3) 1.0 290 (47.5) 132 (52.4) 1.0 
    GT 131 (41.5) 36 (29.5) 0.71 (0.47-1.08) 241 (39.5) 101 (40.1) 0.97 (0.75-1.26) 
    TT 31 (9.8) 21 (17.2) 1.54 (0.94-2.53) 79 (13.0) 19 (7.5) 0.66 (0.41-1.07) 
    Ptrend   0.526   0.171 
    Pinteraction      0.005 
    GG + GT 285 (90.2) 101 (82.8) 1.0 531 (87.1) 233 (92.5) 1.0 
    TT 31 (9.8) 21 (17.2) 1.76 (1.10-2.83) 79 (13.0) 19 (7.5) 0.67 (0.42-1.07) 
    Pinteraction      0.003 

NOTE: HR controlling for age, race, menopausal status, time between surgery and registration, and types of treatment among the treated arm.

Abbreviation: aHR, adjusted hazard ratio.

Table 3.

Associations between NOS3 combined genotypes and DFS, treated and untreated arms in a recessive model

GenotypeTreatedUntreated
NOS3 −786NOS3 894CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
TT + TC GG + GT 253 (80.6) 93 (78.2) 1.0 467 (76.8) 215 (85.7) 1.0 
TT + TC TT 18 (5.7) 8 (6.7) 1.19 (0.57-2.47) 39 (6.4) 13 (5.2) 0.84 (0.48-1.48) 
CC GG + GT 30 (9.6) 6 (5.0) 0.63 (0.27-1.44) 62 (10.2) 17 (6.8) 0.67 (0.41-1.10) 
CC TT 13 (4.1) 12 (10.1) 2.32 (1.26-4.25) 40 (6.6) 6 (2.4) 0.42 (0.19-0.95) 
P for interaction       0.008 
GenotypeTreatedUntreated
NOS3 −786NOS3 894CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
TT + TC GG + GT 253 (80.6) 93 (78.2) 1.0 467 (76.8) 215 (85.7) 1.0 
TT + TC TT 18 (5.7) 8 (6.7) 1.19 (0.57-2.47) 39 (6.4) 13 (5.2) 0.84 (0.48-1.48) 
CC GG + GT 30 (9.6) 6 (5.0) 0.63 (0.27-1.44) 62 (10.2) 17 (6.8) 0.67 (0.41-1.10) 
CC TT 13 (4.1) 12 (10.1) 2.32 (1.26-4.25) 40 (6.6) 6 (2.4) 0.42 (0.19-0.95) 
P for interaction       0.008 

NOTE: HR controlling for age, race, menopausal status, time between surgery and registration, and types of treatment among the treated arm.

No significant associations were observed between the genotypes for NQO1 Pro187Ser, NQO2 Lys47Phe, and CBR3 Val244Met and Val93Val and breast cancer survival in either group (Table 4). When calculating the total number of “at-risk” alleles (those associated with reduced activity of NOS3, NQO1, NQO2, and CBR3), and categorizing into four groups (0-3, 4, 5, and 6+), women in the untreated arm with more than six risk alleles had a 2-fold risk reduction in mortality compared with women having zero to three risk alleles (P for trend = 0.001), although the magnitude of reduction was not greater than that associated with the NOS3 CC and TT genotypes combined (data not shown). Among women who were randomized to chemotherapy, there were no significant effects of combined variant alleles (P for trend = 0.244).

Table 4.

Disease-free survival hazard ratios for NQO1, NQO2, and CBR3 genotypes among women randomized to CMF or CAF ± tamoxifen and among women not receiving adjuvant therapy

GenotypesTreated (n = 458)Untreated (n = 874)
CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
NQO1 609 C > T   n = 447   n = 863 
    CC 210 (65.2) 90 (72.0) 1.0 388 (63.7) 172 (67.7) 1.0 
    CT 97 (30.1) 28 (22.4) 0.71 (0.47-1.10) 191 (31.4) 71 (28.0) 0.85 (0.65-1.13) 
    TT 15 (4.7) 7 (5.6) 1.10 (0.51-2.39) 30 (4.9) 11 (4.3) 0.74 (0.40-1.36) 
    Ptrend   0.368   0.162 
    Pinteraction      0.522 
    CC + CT 307 (95.3) 118 (94.4) 1.0 579 (95.1) 243 (95.7) 1.0 
    TT 15 (4.7) 7 (5.6) 1.21 (0.56-2.60) 30 (4.9) 11 (4.3) 0.77 (0.42-1.42) 
    Pinteraction      0.345 
NQO2 14055 T > C  n = 450   n = 867  
    TT 209 (64.5) 86 (68.3) 1.0 366 (60.0) 166 (66.1) 1.0 
    TC 105 (32.4) 37 (29.4) 0.87 (0.59-1.28) 212 (34.8) 73 (29.1) 0.77 (0.58-1.01) 
    CC 10 (3.1) 3 (2.4) 0.76 (0.24-2.39) 32 (5.3) 12 (4.8) 0.81 (0.45-1.45) 
    Ptrend   0.407   0.082 
    Pinteraction      0.905 
    TT + TC 314 (96.9) 123 (97.6) 1.0 578 (94.8) 239 (95.2) 1.0 
    CC 10 (3.1) 3 (2.4) 0.79 (0.25-2.49) 32 (5.3) 12 (4.8) 0.88 (0.49-1.58) 
    Pinteraction      0.894 
CBR3 ex1-11 C > T  n = 447   n = 860  
    CC 116 (36.0) 48 (38.4) 1.0 194 (31.8) 87 (34.8) 1.0 
    CT 143 (44.4) 54 (43.2) 0.90 (0.61-1.33) 311 (51.0) 131 (52.4) 0.93 (0.71-1.22) 
    TT 63 (19.6) 23 (18.4) 0.93 (0.56-1.53) 105 (17.2) 32 (12.8) 0.74 (0.49-1.11) 
    Ptrend   0.692   0.167 
    Pinteraction      0.706 
    CC + CT 259 (80.4) 102 (81.6) 1.0 505 (82.8) 218 (87.2) 1.0 
    TT 63 (19.6) 23 (18.4) 0.98 (0.63-1.55) 105 (17.2) 32 (12.8) 0.73 (0.52-1.08) 
    Pinteraction      0.566 
CBR3 ex3-155G > A  n = 445   n = 863  
    GG 125 (39.1) 46 (36.8) 1.0 268 (43.9) 100 (39.7) 1.0 
    GA 144 (45.0) 55 (44.0) 1.01 (0.68-1.50) 270 (44.2) 125 (49.6) 1.14 (0.88-1.49) 
    AA 51 (15.9) 24 (19.2) 1.19 (0.76-1.86) 73 (11.9) 27 (10.7) 1.01 (0.66-1.55) 
    Ptrend   0.543   0.613 
    Pinteraction      0.596 
    GG + GA 269 (84.1) 101 (80.8) 1.0 538 (88.1) 225 (89.3) 1.0 
    AA 51 (15.9) 24 (19.2) 1.19 (0.76-1.86) 73 (11.9) 27 (10.7) 0.95 (0.63-1.41) 
    Pinteraction      0.404 
GenotypesTreated (n = 458)Untreated (n = 874)
CensoredFailuresaHR (95% CI)CensoredFailuresaHR (95% CI)
NQO1 609 C > T   n = 447   n = 863 
    CC 210 (65.2) 90 (72.0) 1.0 388 (63.7) 172 (67.7) 1.0 
    CT 97 (30.1) 28 (22.4) 0.71 (0.47-1.10) 191 (31.4) 71 (28.0) 0.85 (0.65-1.13) 
    TT 15 (4.7) 7 (5.6) 1.10 (0.51-2.39) 30 (4.9) 11 (4.3) 0.74 (0.40-1.36) 
    Ptrend   0.368   0.162 
    Pinteraction      0.522 
    CC + CT 307 (95.3) 118 (94.4) 1.0 579 (95.1) 243 (95.7) 1.0 
    TT 15 (4.7) 7 (5.6) 1.21 (0.56-2.60) 30 (4.9) 11 (4.3) 0.77 (0.42-1.42) 
    Pinteraction      0.345 
NQO2 14055 T > C  n = 450   n = 867  
    TT 209 (64.5) 86 (68.3) 1.0 366 (60.0) 166 (66.1) 1.0 
    TC 105 (32.4) 37 (29.4) 0.87 (0.59-1.28) 212 (34.8) 73 (29.1) 0.77 (0.58-1.01) 
    CC 10 (3.1) 3 (2.4) 0.76 (0.24-2.39) 32 (5.3) 12 (4.8) 0.81 (0.45-1.45) 
    Ptrend   0.407   0.082 
    Pinteraction      0.905 
    TT + TC 314 (96.9) 123 (97.6) 1.0 578 (94.8) 239 (95.2) 1.0 
    CC 10 (3.1) 3 (2.4) 0.79 (0.25-2.49) 32 (5.3) 12 (4.8) 0.88 (0.49-1.58) 
    Pinteraction      0.894 
CBR3 ex1-11 C > T  n = 447   n = 860  
    CC 116 (36.0) 48 (38.4) 1.0 194 (31.8) 87 (34.8) 1.0 
    CT 143 (44.4) 54 (43.2) 0.90 (0.61-1.33) 311 (51.0) 131 (52.4) 0.93 (0.71-1.22) 
    TT 63 (19.6) 23 (18.4) 0.93 (0.56-1.53) 105 (17.2) 32 (12.8) 0.74 (0.49-1.11) 
    Ptrend   0.692   0.167 
    Pinteraction      0.706 
    CC + CT 259 (80.4) 102 (81.6) 1.0 505 (82.8) 218 (87.2) 1.0 
    TT 63 (19.6) 23 (18.4) 0.98 (0.63-1.55) 105 (17.2) 32 (12.8) 0.73 (0.52-1.08) 
    Pinteraction      0.566 
CBR3 ex3-155G > A  n = 445   n = 863  
    GG 125 (39.1) 46 (36.8) 1.0 268 (43.9) 100 (39.7) 1.0 
    GA 144 (45.0) 55 (44.0) 1.01 (0.68-1.50) 270 (44.2) 125 (49.6) 1.14 (0.88-1.49) 
    AA 51 (15.9) 24 (19.2) 1.19 (0.76-1.86) 73 (11.9) 27 (10.7) 1.01 (0.66-1.55) 
    Ptrend   0.543   0.613 
    Pinteraction      0.596 
    GG + GA 269 (84.1) 101 (80.8) 1.0 538 (88.1) 225 (89.3) 1.0 
    AA 51 (15.9) 24 (19.2) 1.19 (0.76-1.86) 73 (11.9) 27 (10.7) 0.95 (0.63-1.41) 
    Pinteraction      0.404 

When evaluating potential modification of results by type of chemotherapy received and further randomization to tamoxifen, we noted that there were no differences by whether or not women received CAF or CMF, although there was a suggestion of stronger effects for CBR3 among women who received the anthracycline-containing regimen (data not shown). Women receiving CAF and homozygous for the CBR3 A allele had a nonsignificantly increased risk of recurrence and mortality [odds ratio (OR), 1.54; 95% CI, 0.86-2.77] compared with G allele carriers, but risks were not elevated for those receiving CMF adjuvant chemotherapy (OR, 0.86; 95% CI, 0.42-1.74). For NOS3 894T polymorphism, relationships were strongest among women who were further randomized to tamoxifen treatment (OR, 2.44, 95% CI 1.17-5.05) than those who were not (OR, 1.48, 0.79-2.75). Type of chemotherapy and tamoxifen treatment did not modify any of the null relationships observed between DFS and NQO1 and NQO2 polymorphisms.

In this ancillary study to SWOG clinical trial 8897, we noted associations between DFS and inherited germ-line polymorphisms in both the promoter and coding regions of NOS3 that result in decreased activity. Combined risk variants were associated with a >2-fold reduction in hazard of disease-related death among women who did not receive adjuvant systemic therapy, but among women who received alkylating agent–based adjuvant chemotherapy, low-activity NOS genotypes were associated with a 2-fold increase in risk of recurrence. We did not find any statistically significant associations between other candidate genes (NQO1, NQO2, and CBR3) and breast cancer outcomes among both untreated and treated groups, although there was a suggestion of an association between DFS and CBR3 polymorphisms among women who received CAF, in contrast to those receiving CMF.

The differential effects observed with NOS3 are likely explained by the multiple roles of NO in tumorigenesis and progression, which seem to be substantially modified by tumor environment, cell type, NO concentrations produced, and type of adjuvant treatments received (29, 30). Among women in the untreated group, relationships between NOS3 variants resulting in lower endothelial NO levels and better DFS may be due to the role of NO in disease progression, which seem to be modified by concentrations produced. Low levels have been associated with increased vascular endothelial growth factor synthesis and neovascularization, tumor angiogenesis, blood flow, immune surveillance, enhanced tumor growth, invasion, and metastasis (29, 3133). Low levels of NO can also enhance the nitrosylation and activation of Ras oncogenes, which are required for tumor growth and maintenance (34, 35). Furthermore, increased NO levels are associated with better tumor vasculature; thus, in the untreated group, higher levels could result in increased angiogenesis. These mechanism could result in slower tumor growth and reduced risk of metastasis among those with genotypes encoding lower levels of NO, resulting in better prognosis.

The contrasting data in the chemotherapy-treated patients, showing that lower levels of NOS3 were associated with poorer survival, are consistent with the theory that NOS3 enzymatic activity may mediate sensitivity to alkylating agent–based chemotherapy by virtue of production of NO. Several studies have shown that NO production interacts with other chemotherapy-induced radicals to form peroxynitrite, which damages lipids, DNA, and proteins via direct oxidative reactions or indirect, radical-mediated mechanisms (36, 37). In addition, enhanced tumor vasculature with higher levels of NO would result in better drug delivery; in this case, independent of prognosis in untreated women, those who receive chemotherapy would be expected to have a worse outcome if they have variants encoding lower levels of NOS3, compared with those who produce ROS, which is precisely what we observed.

Our data are consistent with previously published human studies and in vitro data. Co-author Choi previously reported that NOS3 haplotypes containing low-activity −786C or 894T alleles were associated with greater risk of recurrence among 873 Asian breast cancer patients who received adjuvant chemotherapy (38). Furthermore, NOS3 protein expression either in the tumor itself or in the surrounding vasculature has been associated with favorable prognosis in patients who received chemotherapy (39, 40). Taken together with the present results, these data suggest that patients with major alleles, and therefore normal NOS3 enzymatic expression and activity, seem to benefit from chemotherapy more than those with genotypes encoding lower enzyme expression.

Findings from several cell line studies also provide support for an interaction between NOS3 function and chemotherapy on therapeutic outcomes, potentially mediated by hypoxia-induced chemoresistance through the low-concentration NO-cyclic guanosine 3′,5′-monophosphate signaling pathway, which leads to the phosphorylation of various target molecules that regulate cell function and gene expression (41). Matthews et al. (42) showed that hypoxia-mediated acquisition of resistance to doxorubicin and 5-fluorouracil in human breast carcinoma could be attenuated by low concentrations of the NO mimetics glyceryl trinitrate or diethylenetriamine-NO adduct, suggesting that suppression of endogenous NO production is important in the development of hypoxia-induced drug resistance. Recently, NO mimetics were also found to attenuate drug resistance associated with spheroid culture of human MDA-MB-231 breast carcinoma cells, which are associated with a high level of resistance to antitumor drugs (43). Interestingly, incubation with glyceryl trinitrate or diethylenetriamine-NO adduct decreased survival of cancer cells by up to 47%, but the survival of MDA-MB-231 cells was not affected by incubation with glyceryl trinitrate or diethylenetriamine-NO adduct in the absence of doxorubicin (43). The inverse association between low concentrations of NO and increasing rates of chemoresistance is unlikely specific to a particular chemoregimen because NO mimetics have been shown to be effective in reducing hypoxia-induced acquired resistance to doxorubicin, 5-fluorouracil, and paclitaxel (4143), supporting our findings that the low-activity NOS3 894 T allele was associated with worse DFS for both CAF- and CMF-treated patients.

In addition to NOS3, there are two other isoforms of NOS; neuronal NOS (NOS1 or nNOS) and inducible NOS (NOS2 or iNOS). NOS1 and NOS3 are constitutive and expressed in neuronal and vascular endothelial cells, respectively. NOS2, however, is transcriptionally regulated and induced by factors related to hypoxia and oxidative stress. In addition to the research showing dual roles for NOS3, there has also been much interest and research in NOS2 in cancer etiology and treatment outcomes due to its role in immunologic responses. Similar to the proneoplastic and antineoplastic properties of NOS3 and NO, iNOS has been shown to exhibit the same characteristics. There have been numerous in vitro and in vivo studies that have shown the tumorigenic properties of iNOS, particularly in rodent models (reviewed in ref. 44). Other studies have shown that iNOS inhibits tumor formation, growth, and metastasis (44). It is likely that the tumor microenvironment also influences the effects of high levels of NO generated by iNOS, and in fact, there is preclinical research to exploit iNOS in chemotherapy regimens.

The effect of NOS3 genotypes on DFS seemed to be more pronounced among women receiving tamoxifen, possibly because tamoxifen treatment interrupts the activation of NOS3 by estrogens, which would be beneficial for preventing hypoxia-related chemoresistance. Estradiol, and potentially other steroid hormones such as dehydroepiandrosterone sulfate, can stimulate NOS3 expression in breast cancer cell lines through both estrogen receptor–dependent and estrogen receptor–independent phosphorylation by the phosphoinositide 3-kinase-Akt pathway (4550). Therefore, disruption of the estrogen-estrogen receptor complex by tamoxifen might lead to reduced levels of NO generated from NOS3, which would further reduce NO levels in low-activity NOS3 894 T allele carriers, resulting in worse disease prognosis. Because aromatase inhibitors directly deplete estrogen, the potential modification of aromatase inhibitor efficacy by NOS3 polymorphisms merits investigation.

Consistent with the role of CBR3 in the pharmacodynamics of doxorubicin, increased hazards associated with the CBR3 −155 A allele, although not statistically significant, were only observed among those receiving CAF chemotherapy, with no associations observed among the CMF or the untreated group. These findings would be consistent with the expectation that the high activity CBR3 A allele would lead to faster conversion of doxorubicin to doxorubicinol, which is less antineoplastic but plays a key role in anthracycline-related chronic cardiotoxicity (51, 52). A recent study examining CBR1 and CBR3 genotypes with the pharmacokinetics and pharmacodynamics of doxorubicin in a group of 101 breast cancer patients found the CBR3 A variant to be associated with higher doxorubicinol area under the curve and CBR3 expression in breast tumor tissues (53). CBR3 variants may play a more important role in treatment-related toxicities than in tumor cell kill. There is relatively low expression of CBR3 in liver, and the activity of this enzyme in vitro is very modest. However, due to the relatively small number of women who were homozygous for the CBR3 variant alleles, these findings need to be confirmed in a larger population. Furthermore, because epirubicin has been shown to provide some survival advantage over methotrexate (54) and the anthracycline used in this trial, doxorubicin, did not, it is possible that results may not be generalizable to all anthracyclines used in chemotherapy regimens.

In this study, we used genetic polymorphisms as surrogates for systemic enzymatic levels. Because we did not have tumor tissue available, we were not able to measure mRNA or protein in the breast tumors or to correlate levels with genotypes. However, there are ample data supporting the functional effects of both polymorphisms evaluated (refs. 1113), and expression levels of NOS3 in blood mononuclear cells were lower among those with T−786C alleles, in a dose-dependent manner, in both patients with Alzheimer's disease and healthy controls (55). NOS3 gene expression levels and protein concentrations and enzyme activity were also genotype dependent in human cultured endothelial cells (56).

It is possible that the differential results for NOS3 genotypes among treated and untreated groups could result from bias if the frequencies of the genotypes varied with disease characteristics, which were determinants of whether or not patients were randomized to adjuvant chemotherapy. This could be a concern given that all treated patients had larger (1-2 cm) hormone receptor–positive tumors, whereas the low-risk untreated group consisted of patients with smaller tumors (≤1 cm). However, frequencies of the variant minor alleles for the NOS3 894 G > T and −786 T > C polymorphisms were virtually identical (−786 C allele: 37.0% in treated, 37.4% in untreated; 894 T allele: 30.9% in treated, 31.2% in untreated), eliminating this potential source of bias. Because one of the aims of this study was to assess modification of outcomes by adjuvant treatment regimens received, DFS, rather than mortality, was used as the outcome of interest because patients may choose to cross over to another therapeutic agent if recurrence occurred, which would complicate the assessment of outcome associations from the first agent (57). A key strength of this pharmacogenetic study was the population source, which provided comparisons between individuals with similar risk for recurrence based on their tumor characteristics and the standardization of treatments received, thereby eliminating biases and confounding due to treatment heterogeneity (58). As well, the presence of an untreated group in the study population was unique, which allowed for the examination of genotype as a modifier of treatment outcomes.

The results of this correlative study, which are consistent with our a priori hypothesis, suggest that inherited, germ-line variants encoding lower activity of the oxidative stress gene NOS3 may affect outcomes in patients with early-stage breast cancer. These variants may be related to better prognosis in untreated patients by virtue of reducing malignant tumor biology mediated by the activity of this gene, but conversely with worse outcomes in patients who are treated with chemotherapy, perhaps by altering sensitivity normally induced by NOS3 generated production of NO. The heterogeneity of effects by whether or not patients received chemotherapy cautions against studies of predictors of breast cancer survival among groups of patients with varied treatments, or for whom treatment regimens are unknown, because the effects can vary dramatically based on treatment status.

It is likely that oral cyclophosphamide was an important contributor to antitumor effect in these patients because the differences observed apply equally in the arms with and without doxorubicin, and 5-fluorouracil effects are not thought to be mediated through generation of ROS. Whereas we do not recommend any change in standard clinical practice based on these results, this growing body of evidence of the important role of endogenous oxidants in chemotherapy treatment outcomes warrants increased investigation of potential interactions with supplement use by patients undergoing chemotherapy treatment. These findings, and other investigations showing that genotypes related to higher levels of oxidative stress are associated with better survival (59), may add weight to the recommendations that patients receiving adjuvant chemotherapy should be cautioned about the use of antioxidant supplements (60).

Our results suggest that although inhibition of ROS might be an attractive strategy to prevent cancer, it may have adverse effects on the efficacy of adjuvant chemotherapy. However, to date, there are few rigorous data in the literature to support these recommendations. Indeed, based on these findings and others regarding endogenous antioxidants, SWOG is now conducting a prospective clinical trial addressing nonprescribed antioxidant therapy in women participating in a randomized chemotherapy trial (S0221).

No potential conflicts of interest were disclosed.

1
Greenlee
RT
,
Hill-Harmon
MB
,
Murray
T
,
Thun
M
. 
Cancer statistics, 2001
.
CA Cancer J Clin
2001
;
51
:
15
36
.
2
Jemal
A
,
Siegel
R
,
Ward
E
,
Murray
T
,
Xu
J
,
Thun
MJ
. 
Cancer statistics, 2007
.
CA Cancer J Clin
2007
;
57
:
43
66
.
3
Abe
O
,
Abe
R
,
Enomoto
K
, et al
. 
Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials
.
Lancet
2005
;
366
:
2087
106
.
4
Harris
L
,
Fritsche
H
,
Mennel
R
, et al
. 
American Society of Clinical Oncology 2007 update of recommendations for the use of tumor markers in breast cancer
.
J Clin Oncol
2007
;
25
:
5287
312
.
5
Strauss
G
,
Westhoff
M
,
Fischer-Posovszky
P
, et al
. 
4-Hydroperoxy-cyclophosphamide mediates caspase-independent T-cell apoptosis involving oxidative stress-induced nuclear relocation of mitochondrial apoptogenic factors AIF and EndoG
.
Cell Death Differ
2008
;
15
:
332
43
.
6
Murata
M
,
Suzuki
T
,
Midorikawa
K
,
Oikawa
S
,
Kawanishi
S
. 
Oxidative DNA damage induced by a hydroperoxide derivative of cyclophosphamide
.
Free Radic Biol Med
2004
;
37
:
793
802
.
7
Tsai-Turton
M
,
Luong
BT
,
Tan
Y
,
Luderer
U
. 
Cyclophosphamide-induced apoptosis in COV434 human granulosa cells involves oxidative stress and glutathione depletion
.
Toxicol Sci
2007
;
98
:
216
30
.
8
Minotti
G
,
Menna
P
,
Salvatorelli
E
,
Cairo
G
,
Gianni
L
. 
Anthracyclines: molecular advances and pharmacologie developments in antitumor activity and cardiotoxicity
.
Pharmacol Rev
2004
;
56
:
185
229
.
9
Choi
J
,
Nowell
SA
,
Blanco
JG
,
Ambrosone
CB
. 
The role of genetic variability in drug metabolism pathways in breast cancer prognosis
.
Pharmacogenomics
2006
;
7
:
613
24
.
10
Moncada
S
,
Higgs
A
. 
The l-arginine-nitric oxide pathway
.
N Engl J Med
1993
;
329
:
2002
12
.
11
Ahsan
A
,
Norboo
T
,
Baig
MA
,
Pasha
MAQ
. 
Simultaneous selection of the wild-type genotypes of the G894T and 4B/4A polymorphisms of NOS3 associate with high-altitude adaptation
.
Ann Hum Genet
2005
;
69
:
260
7
.
12
Tesauro
M
,
Thompson
WC
,
Rogliani
P
,
Qi
L
,
Chaudhary
PP
,
Moss
J
. 
Intracellular processing of endothelial nitric oxide synthase isoforms associated with differences in severity of cardiopulmonary diseases: cleavage of proteins with aspartate vs. glutamate at position 298
.
Proc Natl Acad Sci U S A
2000
;
97
:
2832
5
.
13
Nakayama
M
,
Yasue
H
,
Yoshimura
M
, et al
. 
T−786→C mutation in the 5′-flanking region of the endothelial nitric oxide synthase gene is associated with coronary spasm
.
Circulation
1999
;
99
:
2864
70
.
14
Celli
CM
,
Tran
N
,
Knox
R
,
Jaiswal
AK
. 
NRH:quinone oxidoreductase 2 (NQO2) catalyzes metabolic activation of quinones and anti-tumor drugs
.
Biochem Pharmacol
2006
;
72
:
366
76
.
15
Traver
RD
,
Siegel
D
,
Beall
HD
, et al
. 
Characterization of a polymorphism in NAD(P)H:quinone oxidoreductase (DT-diaphorase)
.
Br J Cancer
1997
;
75
:
69
75
.
16
Ross
D
,
Traver
RD
,
Siegel
D
,
Kuehl
BL
,
Misra
V
,
Rauth
AM
. 
A polymorphism in NAD(P)H:Quinone oxidoreductase (NQO1): relationship of a homozygous mutation at position 609 of the NQO1 cDNA to NQO1 activity
.
Br J Cancer
1996
;
74
:
995
6
.
17
Siegel
D
,
Anwar
A
,
Winski
SL
,
Kepa
JK
,
Zolman
KL
,
Ross
D
. 
Rapid polyubiquitination and proteasomal degradation of a mutant form of NAD(P)H:quinone oxidoreductase 1
.
Mol Pharmacol
2001
;
59
:
263
8
.
18
Minotti
G
,
Recalcati
S
,
Menna
P
,
Salvatorelli
E
,
Corna
G
,
Cairo
G
. 
Doxorubicin cardiotoxicity and the control of iron metabolism: quinone-dependent and independent mechanisms
.
Methods Enzymol
2004
;
340
61
.
19
Hoffmann
F
,
Maser
E
. 
Carbonyl reductases and pluripotent hydroxysteroid dehydrogenases of the short-chain dehydrogenase/reductase superfamily
.
Drug Metab Rev
2007
;
39
:
87
144
.
20
Lopez de Cerain
A
,
Marin
A
,
Idoate
MA
,
Tunon
MT
,
Bello
J
. 
Carbonyl reductase and NADPH cytochrome P450 reductase activities in human tumoral versus normal tissues
.
Eur J Cancer
1999
;
35
:
320
4
.
21
Covarrubias
VG
,
Lakhman
SS
,
Forrest
A
,
Relling
MV
,
Blanco
JG
. 
Higher activity of polymorphic NAD(P)H:Quinone oxidoreductase in liver cytosols from blacks compared to whites
.
Toxicol Lett
2006
;
164
:
249
58
.
22
Lakhman
SS
,
Ghosh
D
,
Blanco
JG
. 
Functional significance of a natural allelic variant of human carbonyl reductase 3 (CBR3)
.
Drug Metab Dispos
2005
;
33
:
254
7
.
23
McShane
LM
,
Altman
DG
,
Sauerbrei
W
, et al
. 
Reporting recommendations for tumor marker prognostic studies (REMARK)
.
J Natl Cancer Inst
2005
;
97
:
1180
4
.
24
Hutchins
LF
,
Green
SJ
,
Ravdin
PM
, et al
. 
Randomized, controlled trial of cyclophosphamide, methotrexate, and fluorouracil versus cyclophosphamide, doxorubicin, and fluorouracil with and without tamoxifen for high-risk, node-negative breast cancer: treatment results of intergroup protocol INT-0102
.
J Clin Oncol
2005
;
23
:
8313
21
.
25
Bonadonna
G
,
Brusamolino
E
,
Valagussa
P
. 
Combination chemotherapy as an adjuvant treatment in operable breast cancer
.
N Engl J Med
1976
;
294
:
405
10
.
26
Bull
JM
,
Tormey
DC
,
Li
SH
. 
A randomized comparative trial of Adriamycin versus methotrexate in combination drug therapy
.
Cancer
1978
;
41
:
1649
57
.
27
Ambrosone
CB
,
Barlow
WE
,
Reynolds
W
, et al
. 
Myeloperoxidase genotypes and enhanced efficacy of chemotherapy for early stage breast cancer in SWOG 8897
.
J Clin Oncol
.
In press 2009
.
28
Rae
JM
,
Cordero
KE
,
Scheys
JO
,
Lippman
ME
,
Flockhart
DA
,
Johnson
MD
. 
Genotyping for polymorphic drug metabolizing enzymes from paraffin-embedded and immunohistochemically stained tumor samples
.
Pharmacogenetics
2003
;
13
:
501
7
.
29
Jenkins
DC
,
Charles
IG
,
Thomsen
LL
, et al
. 
Roles of nitric oxide in tumor growth
.
Proc Natl Acad Sci U S A
1995
;
92
:
4392
6
.
30
Lancaster
JR
 Jr.
,
Xie
K
. 
Tumors face NO problems?
Cancer Res
2006
;
66
:
6459
62
.
31
Rigas
B
. 
Novel agents for cancer prevention based on nitric oxide
.
Biochem Soc Trans
2007
;
35
:
1364
8
.
32
Lala
PK
,
Orucevic
A
. 
Role of nitric oxide in tumor progression: lessons from experimental tumors
.
Cancer Metastasis Rev
1998
;
17
:
91
106
.
33
Jadeski
LC
,
Chakraborty
C
,
Lala
PK
. 
Nitric oxide-mediated promotion of mammary tumour cell migration requires sequential activation of nitric oxide synthase, guanylate cyclase and mitogen-activated protein kinase
.
Int J Cancer
2003
;
106
:
496
504
.
34
Lim
KH
,
Ancrile
BB
,
Kashatus
DF
,
Counter
CM
. 
Tumour maintenance is mediated by eNOS
.
Nature
2008
;
452
:
646
9
.
35
Shang
ZJ
,
Li
JR
. 
Expression of endothelial nitric oxide synthase and vascular endothelial growth factor in oral squamous cell carcinoma: its correlation with angiogenesis and disease progression
.
J Oral Pathol Med
2005
;
34
:
134
9
.
36
Felley-Bosco
E
. 
Role of nitric oxide in genotoxicity: implication for carcinogenesis
.
Cancer Metastasis Rev
1998
;
17
:
25
37
.
37
Ambs
S
,
Hussain
SP
,
Harris
CC
. 
Interactive effects of nitric oxide and the p53 tumor suppressor gene in carcinogenesis and tumor progression
.
FASEB J
1997
;
11
:
443
8
.
38
Choi
JY
,
Lee
KM
,
Noh
DY
, et al
. 
Genetic polymorphisms of eNOS, hormone receptor status, and survival of breast cancer
.
Breast Cancer Res Treat
2006
;
100
:
213
8
.
39
Mortensen
K
,
Holck
S
,
Christensen
IJ
, et al
. 
Endothelial cell nitric oxide synthase in peritumoral microvessels is a favorable prognostic indicator in premenopausal breast cancer patients
.
Clin Cancer Res
1999
;
5
:
1093
7
.
40
Martin
JH
,
Begum
S
,
Alalami
O
,
Harrison
A
,
Scott
KW
. 
Endothelial nitric oxide synthase: correlation with histologic grade, lymph node status and estrogen receptor expression in human breast cancer
.
Tumour Biol
2000
;
21
:
90
7
.
41
Frederiksen
LJ
,
Sullivan
R
,
Maxwell
LR
, et al
. 
Chemosensitization of cancer in vitro and in vivo by nitric oxide signaling
.
Clin Cancer Res
2007
;
13
:
2199
206
.
42
Matthews
NE
,
Adams
MA
,
Maxwell
LR
,
Gofton
TE
,
Graham
CH
. 
Nitric oxide-mediated regulation of chemosensitivity in cancer cells
.
J Natl Cancer Inst
2001
;
93
:
1879
85
.
43
Muir
CP
,
Adams
MA
,
Graham
CH
. 
Nitric oxide attenuates resistance to doxorubicin in three-dimensional aggregates of human breast carcinoma cells
.
Breast Cancer Res Treat
2006
;
96
:
169
76
.
44
Huerta
S
,
Chilka
S
,
Bonavida
B
. 
Nitric oxide donors: novel cancer therapeutics [review]
.
Int J Oncol
2008
;
33
:
909
27
.
45
Haynes
MP
,
Sinha
D
,
Russell
KS
, et al
. 
Membrane estrogen receptor engagement activates endothelial nitric oxide synthase via the PI3-kinase-akt pathway in human endothelial cells
.
Circ Res
2000
;
87
:
677
82
.
46
Haynes
MP
,
Li
L
,
Sinha
D
, et al
. 
Src kinase mediates phosphatidylinositol 3-kinase/Akt-dependent rapid endothelial nitric-oxide synthase activation by estrogen
.
J Biol Chem
2003
;
278
:
2118
23
.
47
Loibl
S
,
Bratengeier
J
,
Farines
V
, et al
. 
Investigations on the inducible and endothelial nitric oxide synthases in human breast cancer cell line MCF-7—estrogen has an influence on e-NOS, but not on i-NOS
.
Pathol Res Pract
2006
;
202
:
1
7
.
48
Nakatani
K
,
Horinouchi
J
,
Yabu
Y
,
Wada
H
,
Nobori
T
. 
Expression of endothelial nitric oxide synthase is induced by estrogen with glycogen synthase 3β phosphorylation in MCF-7 cells
.
Oncol Rep
2004
;
12
:
833
6
.
49
Simoncini
T
,
Mannella
P
,
Fornari
L
,
Varone
G
,
Caruso
A
,
Genazzani
AR
. 
Dehydroepiandrosterone modulates endothelial nitric oxide synthesis via direct genomic and nongenomic mechanisms
.
Endocrinology
2003
;
144
:
3449
55
.
50
Yallampalli
C
,
Dong
Y
. 
Estradiol-17β inhibits nitric oxide synthase (NOS)-II and stimulates NOS-III gene expression in the rat uterus
.
Biol Reprod
2000
;
63
:
34
41
.
51
Olson
LE
,
Bedja
D
,
Alvey
SJ
,
Cardounel
AJ
,
Gabrielson
KL
,
Reeves
RH
. 
Protection from doxorubicin-induced cardiac toxicity in mice with a null allele of carbonyl reductase
.
Cancer Res
2003
;
63
:
6602
6
.
52
Mordente
A
,
Meucci
E
,
Martorana
GE
,
Giardina
B
,
Minotti
G
. 
Human heart cytosolic reductases and anthracycline cardiotoxicity
.
IUBMB Life
2002
;
52
:
83
8
.
53
Fan
L
,
Goh
BC
,
Wong
CI
, et al
. 
Genotype of human carbonyl reductase CBR3 correlates with doxorubicin disposition and toxicity
.
Pharmacogenet Genomics
2008
;
18
:
621
31
.
54
Early Breast Cancer Trialists' Collaborative Group (EBCTCG)
. 
Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials
.
Lancet
2005
;
365
:
1687
717
.
55
Venturelli
E
,
Galimberti
D
,
Lovati
C
, et al
. 
The T−786C NOS3 polymorphism in Alzheimer's disease: association and influence on gene expression
.
Neurosci Lett
2005
;
382
:
300
3
.
56
Song
J
,
Yoon
Y
,
Park
KU
, et al
. 
Genotype-specific influence on nitric oxide synthase gene expression, protein concentrations, and enzyme activity in cultured human endothelial cells
.
Clin Chem
2003
;
49
:
847
52
.
57
Tang
S
. 
Reducing the risk of distant metastases: a better end point in adjuvant aromatase inhibitor breast cancer trials?
Cancer Invest
2008
;
26
:
481
90
.
58
James
RM
,
Reid
EM
,
Rebbeck
T
,
Ambrosone
CB
,
Shields
PG
. 
Trials and interventions in molecular epidemiology
.
Molecular epidemiology: applications in cancer and other human diseases
.
Informa Healthcare
; 
2008
, p.
29
40
.
59
Ambrosone
CB
,
Ahn
J
,
Singh
KK
, et al
. 
Polymorphisms in genes related to oxidative stress (MPO, MnSOD, CAT) and survival after treatment for breast cancer
.
Cancer Res
2005
;
65
:
1105
11
.
60
Cassileth
BR
,
Vickers
AJ
. 
High prevalence of complementary and alternative medicine use among cancer patients: implications for research and clinical care
.
J Clin Oncol
2005
;
23
:
2590
2
.

Competing Interests

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.