Faculty Bibliography

Long-term exposure to estrogens influences the development of breast cancer in women, but the precise mechanisms involved are not clearly defined. Our working hypothesis is that estrogen modulates this process by two separate processes. One involves the binding of estradiol to estrogen receptor (ER) alpha with stimulation of cell proliferation. Errors in DNA occurring during replication result in fixed mutations when not repaired. The other process results from the formation of genotoxic metabolites of estradiol, which can bind to DNA, cause depurination, and result in mutations. Our collaborative group, funded by a Department of Defense Center of Excellence grant, has examined this hypothesis using a variety of experimental methods. Estradiol and its catechol-estrogen metabolite 4-OH-estradiol causes mutations in cell culture systems and can transform benign MCF-10F cells, allowing them to cause tumors in SCID mice. We have demonstrated loss of heterozygosity and gains and losses of DNA segments by comparative genomic hybridization methodology. The depurinated estradiol-guanine and -adenine adducts are measurable in MCF-7 breast cancer cells in culture and in mouse mammary tissue. The double transgenic, alpha estrogen receptor knockout/Wnt-1 knockin mouse model allows us to dissect out the separate effects of ER-mediated and ER-independent actions of estradiol. Knock out of the ER alpha delays the onset of breast tumors in this model, demonstrating a role of receptor-mediated actions. Oophorectomy delays the onset of tumors and reduces overall incidence, providing evidence for an ER-independent effect. Taken together, these data suggest that both ER-dependent and genotoxic ER-independent effects of estradiol mediate breast cancer development

The aryl hydrocarbon receptor (AhR), a client protein of heat shock protein 90 (HSP90), plays a significant role in polycyclic aromatic hydrocarbon (PAH)-induced carcinogenesis. Tobacco smoke, a source of PAHs, activates the AhR, leading to enhanced transcription of CYP1A1 and CYP1B1, which encode proteins that convert PAHs to genotoxic metabolites. The main objectives of this study were to determine whether HSP90 inhibitors suppress PAH-mediated induction of CYP1A1 and CYP1B1 or block benzo(a)pyrene [B(a)P]-induced formation of DNA adducts. Treatment of cell lines derived from oral leukoplakia (MSK-Leuk1) or esophageal squamous cell carcinoma (KYSE450) with a saline extract of tobacco smoke, B(a)P, or dioxin induced CYP1A1 and CYP1B1 transcription, resulting in enhanced levels of message and protein. Inhibitors of HSP90 [17-allylamino-17-demethoxygeldanamycin (17-AAG); celastrol] suppressed these inductive effects of PAHs. Treatment with 17-AAG and celastrol also caused a rapid and marked decrease in amounts of AhR protein without modulating levels of HSP90. The formation of B(a)P-induced DNA adducts in MSK-Leuk1 cells was inhibited by 17-AAG, celastrol, and alpha-naphthoflavone, a known AhR antagonist. The reduction in B(a)P-induced DNA adducts was due, at least in part, to reduced metabolic activation of B(a)P. Collectively, these results suggest that 17-AAG and celastrol, inhibitors of HSP90, suppress the activation of AhR-dependent gene expression, leading, in turn, to reduced formation of B(a)P-induced DNA adducts. Inhibitors of HSP90 may have a role in chemoprevention in addition to cancer therapy.

Previously we showed that the organoselenium compound, 1,4-phenylenebis(methylene)selenocyanate (p-XSC)(1) inhibits 4-nitroquinoline-N-oxide (4-NQO)-induced tongue tumorigenesis in Fisher rats. Here we investigate possible mechanisms of this inhibition by monitoring mutagenesis and p53 protein levels in lacI and conventional Fisher rats treated with: (1) a carcinogenic dose of 4-NQO for 10 weeks in drinking water, (2) 4-NQO+p-XSC (15 ppm as selenium), and (3) 4-NQO followed by p-XSC. For mutagenesis studies, rats were euthanized at 7, 12 or 23 weeks after the start of 4-NQO. For studies on p53 levels, rats were euthanized at 11, 15 and 23 weeks. Appropriate controls were also monitored. In the 4-NQO-alone groups, the mutant fraction (MF) in the cII gene in tongue increased at least 50x background level. The MF (in units of mutants/10(5) plaque forming units) for the 7, 12, and 23 weeks 4-NQO groups were respectively, 184 +/- 88, 237 +/- 105, and 329 +/- 110. Thus, mutagenesis increased with length of exposure and post-treatment time. p-XSC modestly (ca. 15-30%) inhibited mutagenesis under all conditions. The inhibition reached significance at the last time point. When p-XSC was administered after 4-NQO, the MF was also modestly reduced. In 4-NQO-alone animals, levels of p53 in tongue (determined by Western blotting) were 1, 1.5 and 2.4 control levels at 10, 15 and 23 weeks, respectively. In the p-XSC+4-NQO group, the enhancement in p53 levels by 4-NQO treatment was decreased about 90% at 15 weeks and 45% (P<0.05) at 23 weeks, and by slightly smaller percentages in corresponding post-treatment groups. p-XSC alone did not alter p53 levels. As p53 levels generally increase in response to DNA damage, these results suggest that p-XSC reduces 4-NQO-induced DNA damage, resulting in reduced 4-NQO-induced mutagenesis and carcinogenesis. However, the fact that p-XSC is also effective when administered after 4-NQO, suggests additional mechanisms of inhibition exist

The dietary and environmental agent, 6-nitrochrysene (6-NC) is a powerful mammary carcinogen and mutagen in rats. It is known to be metabolized by ring-oxidation, nitro-reduction and a combination of the two pathways. In order to determine the ultimate mutagenic metabolites, we have compared the previously determined mutational profile of 6-NC in rat mammary gland [T. Boyiri, et al. (2004) Carcinogenesis, 25, 637-643] with that of five of its known metabolites in the cII gene of lacI mammary epithelial cells in vitro. In vivo, 6-NC gives rise to three major mutations, AT > GC, AT > TA and GC > TA (in decreasing order) which comprise >70% of the mutations. The metabolite whose mutational profile was most similar to that of 6-NC in vivo was trans-1,2-dihydroxy-1,2-dihydro-N-hydroxy-6-aminochrysene (1,2-DHD-6-NHOH-C) which arises from both ring-oxidation and nitro-reduction. However, metabolites arising from either ring-oxidation or nitro-reduction alone exhibited some similarities to mutational profile of 6-NC. These results, taken in conjunction with previous data showing that the major DNA adducts in mammary tissue of rats treated with 6-NC are products of the reaction of 1,2-DHD-6-NHOH-C with guanine and adenine, make a strong case that 1,2-DHD-6-NHOH-C is the ultimate genotoxic metabolite from 6-NC.

Exposure to estrogens is associated with increased risk of breast and other types of human cancer. Estrogens are converted to metabolites, particularly the catechol estrogen-3,4-quinones (CE-3,4-Q), that can react with DNA to form depurinating adducts. These adducts are released from DNA to generate apurinic sites. Error-prone base excision repair of this damage may lead to the mutations that can initiate breast, prostate and other types of cancer. The reaction of CE-3,4-Q with DNA forms the depurinating adducts 4-hydroxyestrone(estradiol) [4-OHE1(E2)-1-N3Ade and 4-OHE1(E2)-1-N7Gua. These two adducts constitute more than 99% of the total DNA adducts formed. Increased levels of these quinones and their reaction with DNA occur when estrogen metabolism is unbalanced. Such an imbalance is the result of overexpression of estrogen activating enzymes and/or deficient expression of the deactivating (protective) enzymes. This unbalanced metabolism has been observed in breast biopsy tissue from women with breast cancer, compared to control women. Recently, the depurinating adduct 4-OHE1(E2)-1-N3Ade has been detected in the urine of prostate cancer patients, but not in urine from healthy men. Mutagenesis by CE-3,4-Q has been approached from two different perspectives: one is mutagenic activity in the lacI reporter gene in Fisher 344 rats and the other is study of the reporter Harvey-ras gene in mouse skin and rat mammary gland. A-->G and G-->A mutations have been observed in the mammary tissue of rats implanted with the CE-3,4-Q precursor, 4-OHE2. Mutations have also been observed in the Harvey-ras gene in mouse skin and rat mammary gland within 6-12 h after treatment with E2-3,4-Q, suggesting that these mutations arise by error-prone base excision repair of the apurinic sites generated by the depurinating adducts. Treatment of MCF-10F cells, which are estrogen receptor-alpha-negative immortalized human breast epithelial cells, with E2, 4-OHE2 or 2-OHE2 induces their neoplastic transformation in vitro, even in the presence of the antiestrogen ICI-182,780. This suggests that transformation is independent of the estrogen receptor. The transformed cells exhibit specific mutations in several genes. Poorly differentiated adenocarcinomas develop when aggressively transformed MCF-10F cells are selected and injected into severe combined immune depressed (SCID) mice. These results represent the first in vitro/in vivo model of estrogen-induced carcinogenesis in human breast epithelial cells. In other studies, the development of mammary tumors in estrogen receptor-alpha knockout mice expressing the Wnt-1 oncogene (ERKO/Wnt-1) provides direct evidence that estrogens may cause breast cancer through a genotoxic, non-estrogen receptor-alpha-mediated mechanism. In summary, this evidence strongly indicates that estrogens can become endogenous tumor initiators when CE-3,4-Q react with DNA to form specific depurinating adducts. Initiated cells may be promoted by a number of processes, including hormone receptor stimulated proliferation. These results lay the groundwork for assessing risk and preventing disease.

Estrogens are hypothesized to contribute to breast cancer via estrogen receptor-mediated increases in cell proliferation and via genotoxic processes leading to mutations. In this latter process, estradiol (E(2)) is thought to be oxidized to 4-hydroxyestradiol and then to E(2)-3,4-quinone, which reacts with DNA leading to apurinic sites. These sites represent premutagenic lesions. Additionally, E(2)-3,4-quinone can undergo redox cycling with E(2)-3,4-hydroquinone, leading to the release of reactive oxygen species. Although there is evidence that estradiol and E(2)-3,4-quinone are carcinogenic or mutagenic in several systems, 4-hydroxyestradiol, a key intermediate in the proposed genotoxic pathway, has thus far been negative in mutagenesis assays. Another major metabolite of estradiol, 2-hydroxyestradiol, is essentially inactive in carcinogenicity or mutagenicity assays. Here, we report that when using multiple low-dose exposures 4-hydroxyestradiol is mutagenic in the cII assay in BB rat2 cells. Under similar conditions, 2-hydroxyestradiol is inactive. Furthermore, the mutational spectrum of 4-hydroxyestradiol contains a considerable proportion of mutations at A:T base pairs, consistent with the known ability of E(2)-3,4-quinone to form a significant fraction of DNA adducts at adenines. Thus, the results of this study support the proposal that estradiol can contribute to carcinogenesis via a genotoxic pathway.

The mutational spectrum of bleomycin was compared with the spontaneous mutational spectrum in lacZ mouse kidney. Mice were treated with four 20 mg/kg of doses of bleomycin over a two-week period, leading to a mutant fraction several times greater than that of controls. The major class of bleomycin-induced mutations consisted of small deletions, in particular -1 deletions at AT base pairs and hot spots for deletions at 5'-GTC-3' sequences. Smaller, but significant fractions of GC > AT followed by GC > TA substitutions were also observed. In untreated mice, the major class of mutations consisted of GC > AT substitutions followed by GC > TA mutations, and a much smaller fraction of deletions. Other than the specificity of bleomycin for AT base pairs and the 5'-GTC-3' hotspots, the mutational spectrum of bleomycin in mice is similar to that reported for ionizing radiation. However, bleomycin initially mediates the formation of oxidized DNA via reduction of molecular oxygen, as opposed to the radiolysis of water. In this respect mutagenesis induced by bleomycin may be more similar to that induced by endogenous reactive oxygen species (ROS) than mutagenesis induced by ionizing radiation. If bleomycin-induced mutagenesis is an appropriate model for mutagenesis induced by ROS, then, based on the difference between the mutational spectrum of bleomycin and spontaneous mutagenesis, the latter appears not to result predominantly from ROS, at least in mouse kidney.

We have studied the effects of three chemopreventive agents alone or in binary combinations on benzo[a]pyrene (BaP)-induced mutagenesis in the oral cavity and esophagus of lacZ mice using galE(-) selection. The mice were fed diets supplemented with 1,4-phenylenebis(methylene)selenocyanate (p-XSC) at 2.5 and 10 ppm Se, selenium-enriched yeast (SeY) at 2.5 and 10 ppm Se, and 3H-1,2-dithiole-3-thione (D3T) at 65 and 250 ppm, for 6 weeks. Two weeks after the start of the dietary regimen, mice were gavaged with five doses of 125 mg/kg BaP over 2 weeks, and the experiment was terminated 2 weeks later. Mutagenesis was measured in tongue, other pooled oral tissues (OTs), and esophagus. In mice treated with BaP alone, mutagenesis in the above tissues was in the range of 21-32 mutants/10(5)pfu (ca. 6-10 background levels for the corresponding tissues). p-XSC modestly inhibited mutagenesis (10-33% inhibition) in all tissues, but statistical significance was only observed at the low dose in esophagus, and pooled OT. SeY was not inhibitory alone. Greater inhibitory effects were observed with D3T, and inhibition was statistically significant at the high dose in tongue and esophagus (ca. 33%). Two combinations of low doses of the inhibitors were tested, and the D3T + SeY mix was most effective, leading to statistically significant inhibition in all three tissues (ca. 30-40% inhibition). The mixture D3T + p-XSC was of similar effectiveness as the low dose of D3T alone. This study combined with those previously done in our laboratory demonstrates effectiveness of D3T and to a lesser extent, p-XSC in the inhibition of mutagenesis, and provides support for the use of certain combinations of inhibitors as a means to increase effectiveness and reduce the dose of chemopreventive agents.

We determined the mutant fractions (MF) and mutational specificities in the cII gene in histologically confirmed normal, non-involved and tumor mammary tissues of female transgenic (Big Blue F344 x Sprague-Dawley)F1 rats treated with the environmental pollutant 6-nitrochrysene (6-NC). At 30 days of age, three groups were set up for oral treatment with 6-NC dissolved in trioctanoin, or trioctanoin alone once a week for 8 weeks. Two dose levels of 6-NC (100 and 200 micromol/rat) were selected on the basis of our previous carcinogenicity bioassays with CD rats. The rats were decapitated 32 weeks after the last carcinogen dose. Both incidence and multiplicity of mammary adenocarcinomas were significantly elevated in the high dose (36%, 0.57, P < 0.01) group but at the low dose these outcomes (16%, 0.23, P < 0.1) were not significantly different from those of control rats (3%, 0.03). The MF in normal, non-involved and tumor tissues from the mammary glands of 6-NC-treated rats were comparable. At the high and low doses, respectively (4.8 +/- 2.0, 3.2 +/- 2.1) the MF of 6-NC-treated rats, were significantly higher (P < 0.05) than that observed in control rats (1.2 +/- 0.6). Control mutants consisted primarily of GC --> AT transitions, whereas 6-NC-induced mutants were comprised of several major classes of mutations with GC --> TA, GC --> CG, AT --> GC and AT --> TA as the most prevalent. Further studies indicated that the structures of 6-NC-DNA adducts in the mammary tissue are consistent with the mutational specificities. This is the first report that defines the relationship between carcinogenesis and mutagenesis, as well as between structures of 6-NC-DNA adducts and mutation characteristics in the target organ in vivo.