Heat Shock Protein 90 Inhibitors Suppress Aryl Hydrocarbon Receptor–Mediated Activation of CYP1A1 and CYP1B1 Transcription and DNA Adduct Formation Free

The aryl hydrocarbon receptor (AhR), a ligand-activated member of the basic helix-loop-helix family of transcription factors (1), binds with high affinity to the environmental toxin 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD or dioxin; ref. 2). Polycyclic aromatic hydrocarbons (PAH) in tobacco smoke, coal tar, automobile exhaust, wood smoke, and charbroiled food also bind to and activate the AhR. Most of the diverse biochemical, biological, and toxicologic responses caused by exposure to PAHs and polychlorinated dioxins are mediated, at least in part, by the AhR (3). The AhR is believed to be involved in the genotoxic actions of environmental carcinogens. The potential significance of the AhR in carcinogenesis is supported by recent studies of engineered mice. A constitutively active AhR expressed in transgenic mice induced tumors in the glandular stomach (4). Constitutive activation of AhR also renders mice more susceptible to carcinogens (5). Mice engineered to be AhR deficient were protected against benzo(a)pyrene [B(a)P]–induced skin tumors (6).

In the absence of ligand, the AhR is present in the cytosol as a component of a complex with a dimer of the chaperone heat shock protein 90 (HSP90; refs. 7, 8), XAP2 (9), and p23 (10, 11). Upon ligand binding, the AhR translocates into the nucleus (12) where it dissociates from its chaperone complex and forms a heterodimer with the AhR nuclear transporter (13, 14). The AhR–AhR nuclear transporter dimer then binds to the upstream regulatory region of genes containing xenobiotic responsive elements (XRE; refs. 13, 14), resulting in the transcriptional activation of a network of genes encoding enzymes including CYP1A1 and CYPB1. The activation of AhR-mediated signaling leading to induction of xenobiotic metabolism can provide a first line of defense against certain toxic environmental carcinogens (15). However, the induction of xenobiotic-metabolizing enzymes by ligand-activated AhR can also increase the production of highly carcinogenic metabolites, creating a link between the AhR and chemical carcinogenesis. For example, B(a)P, a potent ligand of the AhR, induces its own metabolism to noncarcinogenic B(a)P phenols (16) and a toxic metabolite, anti-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo(a)pyrene (BPDE; ref. 17), which covalently binds to DNA, forming bulky DNA adducts that induce mutations (18). It is possible, therefore, that agents that suppress the activation of AhR signaling will possess chemopreventive properties.

Inhibitors of HSP90 possess anticancer activity and are being extensively evaluated for the treatment of cancer (19). Synthetic HSP90 inhibitors, including 17-allylamino-17-demethoxygeldanamycin (17-AAG), bind within the ATP-binding pocket of the NH₂-terminal domain of HSP90 (20) and thereby inhibit HSP90 activity. Recently, celastrol and gedunin (Fig. 1), two natural products derived from plants of the Celastracae and Meliacae families, were found to inhibit HSP90 (21), albeit via a different mechanism than other known HSP90 inhibitors. Inhibition of HSP90 induces the degradation of a large number of client proteins with oncogenic properties. Although the AhR is a client protein of HSP90, little is known about whether 17-AAG, celastrol, or gedunin can suppress the activation of AhR signaling or alter carcinogen metabolism.

In the present study, we first determined that both synthetic and natural inhibitors of HSP90 suppressed tobacco smoke, B(a)P, and dioxin-mediated induction of CYP1A1 and CYP1B1 transcription. This appeared to reflect the ability of HSP90 inhibitors to induce the rapid degradation of the AhR. Importantly, inhibitors of HSP90 also suppressed the formation of B(a)P-induced DNA adducts in a cellular model of oral leukoplakia. Taken together, these findings both strengthen the rationale of targeting AhR as a chemopreventive approach and suggest the potential use of HSP90 inhibitors for this purpose.

Materials. Keratinocyte growth medium was obtained from Clonetics. DMEM, fetal bovine serum, and Lipofectamine 2000 were from Invitrogen. Antibody to β-actin, Lowry protein assay kits, and B(a)P were obtained from Sigma Chemical. Antiserum to CYP1B1 was provided by Dr. Craig B. Marcus (University of New Mexico, Albuquerque, NM). Antibodies to CYP1A1, AhR, and HSP90 were obtained from Santa Cruz Biotechnology, and antibody to p23 was obtained from Affinity Bioreagents. Western blot analysis detection reagents (enhanced chemiluminescence) were from Amersham Biosciences. Nitrocellulose membranes were from Schleicher and Schuell. DNA and RNA were prepared using kits from Qiagen. PCR primers were synthesized by Sigma Genosys. Murine leukemia virus reverse transcriptase, Taq polymerase, and deoxynucleotide triphosphates were purchased from Applied Biosystems. Reagents for the luciferase assay were from Analytical Luminescence. 17-AAG was from Biomol International LP. Celastrol and gedunin were purchased from Calbiochem.

Cell culture. MSK-Leuk1 cells were established from a premalignant dysplastic leukoplakia lesion adjacent to a squamous cell carcinoma of the tongue (22). Cells were routinely maintained in keratinocyte growth medium, grown to 60% confluency, and trypsinized with 0.125% trypsin-2 mmol/L EDTA solution. KYSE 450 esophageal squamous cell carcinoma cells (23) were maintained in DMEM with low glucose, supplemented with 10% fetal bovine serum, 100 IU/mL penicillin, and 100 μg/mL streptomycin. In all experiments, cells were grown in basal medium for 24 h before treatment. Treatments were carried out in growth factor–free basal medium.

Preparation of tobacco smoke. Cigarettes (2R4F, Kentucky Tobacco Research Institute) were smoked in a Borgwaldt piston-controlled apparatus (model RG-1) using the Federal Trade Commission standard protocol. Cigarettes were smoked one at a time in the apparatus and the smoke was drawn under sterile conditions into premeasured amounts of sterile PBS (pH 7.4). This smoke in PBS represents whole trapped mainstream smoke (TS). Quantitation of smoke content is expressed in puffs/mL of PBS with one cigarette yielding about 8 puffs drawn into a 5-mL volume. The final concentration of TS in the cell culture medium is expressed as puffs/mL medium. All treatments were carried out with 0.03 puffs/mL TS because this concentration of TS was previously found to induce CYP1A1 and CYP1B1 (24, 25). As in our previous studies (24, 25), TS was stored at −80°C until use.

Western blot analysis. Cell lysates were prepared by treating cells with lysis buffer [150 mmol/L NaCl, 100 mmol/L Tris (pH 8.0), 1% Tween 20, 50 mmol/L diethyldithiocarbamate, 1 mmol/L phenylmethylsulfonyl fluoride, 10 μg/mL aprotinin, 10 μg/mL trypsin inhibitor, and 10 μg/mL leupeptin]. Lysates were sonicated for 3 × 10 s on ice and centrifuged at 14,000 × g for 10 min at 4°C to sediment the particulate material. The protein concentration of the supernatant was measured by the method of Lowry (26). SDS-PAGE was done under reducing conditions on 10% polyacrylamide gels. The resolved proteins were transferred onto nitrocellulose sheets and then incubated with antisera to CYP1A1, CYP1B1, AhR, HSP90, XAP-2, p23, and β-actin. Secondary antibody to IgG conjugated to horseradish peroxidase was used. The blots were then reacted with the ECL Western blot detection system, according to the manufacturer's instructions.

Analysis of CYP1A1 and CYP1B1 mRNA. Total cellular RNA was isolated using the RNeasy Mini-kit with on-column DNA digestion using RNase-free DNase according to the manufacturer's instructions. Reverse transcription was done using 1 μg RNA per 50 μL reaction. The reaction mixture contained 1× PCR Buffer II, 2.5 mmol/L MgCl₂, 500 μmol/L deoxynucleotide triphosphates, 50 units RNase inhibitor, 125 units MuLV reverse transcriptase, and 2.5 μmol/L random hexamers. Samples were incubated for 10 min at room temperature, then cycled at 42°C for 15 min and 95°C for 10 min. CYP1A1 and CYP1B1 expression was then determined by reverse transcription-PCR. An aliquot of 2 μL of cDNA was subjected to 30 cycles of PCR in a 25-μL reaction mixture (1× PCR Buffer II, 2 mmol/L MgCl₂, 400 μmol/L deoxynucleotide triphosphates, 2.5 units Taq polymerase, and upstream and downstream primers). Primers used were as follows: CYP1A1, forward 5′-TCATGCTTTTCCCAATCTCC-3′, reverse 5′-CTCCTGCAACGTGCTTATCA-3′; CYP1B1, forward 5′-AACGTCATGAGTGCCGTGTGT-3′, reverse 5′-GGCCGGTACGTTCTCCAAATC-3′; β-actin, forward 5′-GGTCACCCACACTGTGCCCAT-3′, reverse 5′-GGATGCCACAGGACTCCATGC-3′. Thermal cycling conditions were as follows: 95°C for 2 min, followed by 30 s at 95°C, 30 s at 62°C and 45 s at 72°C for 30 cycles and then 72°C for 10 min. Importantly, the signal for each product was linear with the amount of RNA analyzed. PCR products were electrophoresed on a 1% agarose gel with ethidium bromide and photographed under UV light. The identity of each PCR product was confirmed by direct sequencing.

Transient transfection. MSK-Leuk1 cells were seeded at a density of 8 × 10⁵ per well in six-well dishes and grown to 60% confluence. In each well, cells were cotransfected with 1.8 μg of the XRE luciferase plasmid pGudLuc6.1 (a gift of Dr Michael S. Denison, University of California at Davis, Davis, CA) and 0.2 μg of pSVβgal using Lipofectamine 2000 as per the manufacturer's instructions. After 6 h of incubation, the medium was replaced with basal medium and the cells were allowed to grow for 24 h. The activities of luciferase and β-galactosidase were measured in cellular extract.

Chromatin immunoprecipitation assay. Chromatin immunoprecipitation assay was done with a kit (Upstate) according to the manufacturer's instructions. Cells (2 × 10⁶) were cross-linked in a 1% formaldehyde solution for 10 min at 37°C. Cells were then lysed in 200 μL of SDS buffer and sonicated to generate 200- to 1,000-bp DNA fragments. After centrifugation, the cleared supernatant was diluted 10-fold with chromatin immunoprecipitation buffer and incubated with 1.5 μg of the indicated antibody at 4°C. Immune complexes were precipitated, washed, and eluted as recommended. DNA-protein cross-links were reversed by heating at 65°C for 4 h, and the DNA fragments were purified and dissolved in 50 μL of water. Ten microliters of each sample were used as a template for PCR amplification. CYP1A1 oligonucleotide sequences for PCR primers were forward 5′-ACCCGCCACCCTTCGACAGTTC-3′ and reverse 5′-TGCCCAGGCGTTGCGTGAGAAG-3′ (27). Forward and reverse primers used to amplify CYP1B1 were 5′-GTTCCCTTATAAAGGGAG-3′ and 5′-CTGCGATGGAAGCCGTTG-3′ (28). PCR was done at 94°C for 30 s, 60°C for 30 s, and 72°C for 45 s for 30 cycles. The PCR products generated from the chromatin immunoprecipitation template were sequenced, and the identity of the CYP1A1 and CYP1B1 promoters was confirmed.

DNA adducts and B(a)P-tetrols. MSK-Leuk1 cells were grown to 50% confluence and placed in 10 mL KBM. Cells were pretreated with 17-AAG, celastrol, α-napthoflavone, or vehicle (0.1% DMSO) for 2 h. Subsequently, [³H]B(a)P (G.E. Healthcare) was diluted to a specific activity of 12.8 Ci/mmol B(a)P with unlabeled B(a)P and added to a final concentration of 1 μmol/L, resulting in a total amount of 128 μCi [³H]-B(a)P per dish. Twelve h later, cells were harvested, rinsed with PBS, and DNA was prepared using a Recoverase kit (Stratagene) as per the manufacturer's instructions. The DNA was then extracted several times with water-saturated ethyl acetate until radioactivity was undetectable. The volume of the DNA solution was brought to 25 μL, the pH was adjusted to 1.0 with HCl and the solution was heated at 90°C for 3 h to release B(a)P-tetrols. To identify peaks containing radioactivity, the hydrolysate was added to markers of the B(a)P-tetrols produced by the spontaneous hydrolysis of ±anti-7,8,dihydroxy-9,10-epoxy-7,8,9,10-tetrahydo-B(a)P (Chemical Carcinogen Reference Standard Repository, Midwest Research Institute). The tetrols, designated B(a)P-tetrol I-1 and I-2, are derived from the major adducts found in human DNA samples (29).

After hydrolysis, the DNA was applied to a 2 × 150 mm Nova-Pak C18 column and eluted isocratically using a Waters 501 pump at a flow rate of 0.15 mL/min with a mobile phase containing 24% acetonitrile and 10 mmol/L ammonium formate (pH 5.1). The eluate was monitored on a Hitachi F-1080 fluorescence detector using 325-nm-excitation and 400-nm-emission wavelengths. Fractions were collected every minute and counted in a Packard 1600 TR liquid scintillation analyzer.

To determine the concentration of DNA in each sample, an aliquot of the hydrolysate was applied to an Alltima HP C18 3μ 2.1 × 150 mm column (Alltech Associates) and the eluate was monitored using a Waters 440 absorbance detector at 254 nm. The mobile phase was 1.5% acetonitrile in 50 mmol/L ammonium formate (pH 5.1) and the flow was 0.12 mL/min. The DNA concentration was calculated from the concentration of adenine in the hydrolysate.

B(a)P tetrols in the medium result from the spontaneous decomposition of BPDE and were monitored by separating a fraction of the medium containing nearly all of the radioactivity in B(a)P-tetrol I-1 and I-2. For this separation, 500 μL of medium were removed when the DNA isolation began and the medium was centrifuged at 14,000 rpm for 5 min to remove and cells or debris. The supernatant was extracted with 0.5 volume of methylene chloride, which removed B(a)P and other very hydrophobic solutes, but not the tetrols, and contained 10% to 20% of the total radioactivity in the medium. The aqueous layer was then extracted twice with 2 volumes of water-saturated ethyl acetate, which extracted about one third of the remaining radioactivity. Fluorescence assays for the internal standard showed that about 95% of tetrols were extracted in ethyl acetate under these conditions. For several samples, an aliquot of the ethyl acetate fraction was injected onto the high-performance liquid chromatography system and analyzed as above. About 90% of the radioactivity was contained in the B(a)P-tetrol peaks.

Statistics. Comparisons between groups were made by Student's t test. A difference between groups of P < 0.05 was considered significant.

HSP90 inhibitors suppress TS-mediated induction of CYP1A1 and CYP1B1 mRNA and protein in MSK-Leuk1 and KYSE450 cells. Initially, we evaluated whether either 17-AAG or the natural inhibitors (celastrol, gedunin) of HSP90 suppressed TS-mediated induction of CYP1A1 and CYP1B1. Western blot analysis was carried out on lysates prepared from MSK-Leuk1 cells, a cell line that was established from an oral leukoplakia lesion. As shown in Fig. 2A and B, each of these HSP90 inhibitors caused concentration-dependent inhibition of TS-mediated induction of CYP1A1 and CYP1B1. To confirm that these effects were not unique to MSK-Leuk1 cells, similar experiments were carried out in KYSE450 cells, a cell line derived from human esophageal squamous cell carcinoma. Once again, 17-AAG, celastrol, and gedunin caused concentration-dependent inhibition of TS-mediated induction of CYP1A1 and CYP1B1 (Fig. 2A and B). In both cell lines, 17-AAG was more potent than either celastrol or gedunin, with inhibitory effects found in the nanomolar range. Because the suppressive effects of celastrol and gedunin were comparable, we used celastrol in subsequent experiments. To determine if the suppressive effects of 17-AAG and celastrol were pretranslational, levels of mRNA were measured by RT-PCR. Consistent with the Western blot results, both 17-AAG and celastrol inhibited TS-mediated induction of the CYP1A1 and CYP1B1 message (Fig. 3A and B). The HSP90 inhibitors exhibited a similar suppressive effect in both MSK-Leuk1 and KYSE450 cell lines. To complement the studies of TS, we also evaluated B(a)P and TCDD, prototypic ligands of the AhR. B(a)P is found in tobacco smoke and charbroiled food, whereas TCDD is a potent environmental toxin. As shown in Fig. 4A and B, treatment of MSK-Leuk1 cells with either B(a)P or TCDD induced CYP1A1 and CYP1B1. Consistent with the TS findings, both 17-AAG and celastrol blocked B(a)P- and TCDD-mediated induction of CYP1A1 and CYP1B1 (Fig. 4A and B).

HSP90 inhibitors suppress the activation of AhR-dependent gene expression. Experiments were next carried out to determine whether the HSP90 inhibitors blocked the induction of CYP1A1 and CYP1B1 by suppressing transcription. Transient transfections were done using an XRE-luciferase promoter construct. This promoter construct was selected because ligand-activated AhR binds to XREs in the promoters of CYP1A1 and CYP1B1, leading to increased transcription. Treatment with TS, B(a)P, or TCDD led to a marked increase in XRE-luciferase activity, an effect that was suppressed by either 17-AAG or celastrol (Fig. 5A and B). To further evaluate effects on transcription, chromatin immunoprecipitation assays were done. Protein-DNA complexes were immunoprecipitated with antibodies to AhR, and bound DNA fragments were recovered and subjected to semiquantitative RT-PCR with oligonucleotides specific for the CYP1A1 and CYP1B1 promoters. TS induced the binding of AhR to the CYP1A1 and CYP1B1 promoters, an effect that was blocked by both 17-AAG and celastrol (Fig. 5C-F).

To investigate the mechanism by which HSP90 inhibitors blocked AhR ligand–mediated activation of gene expression, levels of proteins known to be involved in AhR signaling were determined. Treatment with concentrations of 17-AAG or celastrol that blocked the induction of CYP1A1 and CYP1B1 led to a rapid and pronounced decrease in amounts of AhR protein (Fig. 6A). In contrast, HSP90 inhibitors did not affect the amounts of other proteins that are important for AhR signaling, including HSP90, p23, and XAP-2 (Fig. 6A).

B(a)P-induced DNA adduct formation is suppressed by HSP90 inhibitors. Both CYP1A1 and CYP1B1 can metabolize PAHs, including B(a)P, into DNA-reactive species that form adducts (30, 31). The formation of adducts can, in turn, lead to mutations. Because HSP90 inhibitors blocked PAH-mediated induction of CYP1A1 and CYP1B1, we postulated that 17-AAG and celastrol might suppress the formation of DNA adducts. For cells treated with 1 μmol/L B(a)P alone, the levels of B(a)P-adducts were in the range of 1 adduct/10⁷ nucleotides on a mole basis. Both 17-AAG and celastrol inhibited B(a)P-induced DNA adduct formation. 17-AAG, at 25 and 50 nmol/L, inhibited DNA adduct formation by ∼65% and 90%, respectively (P < 0.05; Fig. 6B). Celastrol, at 0.5 and 1 μmol/L, inhibited DNA adduct formation by ∼50% and 75%, respectively (P < 0.05; Fig. 6B). α-Napthoflavone (1 μmol/L), a known AhR antagonist, served as a positive control, and inhibited DNA adduct formation by >90% (P < 0.01; data not shown).

The major pathway by which B(a)P is metabolized to a genotoxic metabolite involves its oxidation to BPDE. The epoxidation reactions are catalyzed by the cytochrome P450 mixed function oxidase system, with CYP1A1 and CYP1B1 playing a major role. BPDE is extremely reactive and undergoes spontaneous hydrolysis in aqueous medium to B(a)P-tetrols. To determine if the observed suppression of DNA adducts mediated by HSP90 inhibitors reflected changes in B(a)P metabolism, the levels of B(a)P-tetrols were measured in the cell medium. As shown in Fig. 6C, 25 and 50 nmol/L 17-AAG inhibited B(a)P-tetrol formation by ∼50% and 80%, respectively (P < 0.05). Celastrol, at 0.5 and 1 μmol/L, inhibited B(a)P-tetrol formation by ∼50% and 62%, respectively (P < 0.05; Fig. 6C). Lastly, 1 μmol/L α-napthoflavone inhibited B(a)P-tetrol formation by ∼75% (P < 0.05; data not shown). Taken together, HSP90 inhibitors appear to suppress the metabolic activation of B(a)P, resulting in the suppression of B(a)P-induced DNA adducts.

Although smoking cessation is optimal, it is not possible for everyone. Hence, developing strategies to reduce the procarcinogenic effects of PAHs remains a significant unmet need. As detailed above, multiple lines of evidence suggest that targeting the AhR is a potential pharmacologic approach to reducing PAH-induced cancer (32). In this study, we found that both synthetic and natural inhibitors of HSP90 suppressed amounts of the AhR, blocked PAH-mediated activation of CYP1A1 and CYP1B1 transcription, and reduced the formation of DNA adducts in B(a)P-treated cells.

HSP90 is an evolutionarily conserved and ubiquitously expressed molecular chaperone that modulates client protein folding and prevents the nonspecific aggregation of misfolded or unfolded proteins. There are numerous client proteins of HSP90, including many whose deregulation has been linked to carcinogenesis, such as Bcr-Abl, mutated p53, ErbB2, Akt, Flt3, HIF-1α, and B-Raf (33, 34). Inhibitors of HSP90 suppress levels of numerous client proteins and are being actively investigated for the treatment of cancer (33, 35, 36). Here, we focused on using synthetic and natural inhibitors of HSP90 to disrupt the actions of AhR, another HSP90 client protein. Treatment with 17-AAG, a synthetic inhibitor of HSP90, led to a rapid decrease in amounts of the AhR. Consistent with this effect, 17-AAG also suppressed PAH-mediated activation of CYP1A1 and CYP1B1 transcription, leading to a marked reduction in levels of message and protein. These findings are consistent with prior evidence that geldanamycin, a benzoquinone ansamycin that directly associates with the ATP/ADP binding site of HSP90, suppressed levels of AhR and reduced dioxin-mediated induction of CYP1A1 (37, 38).

Little is known about the cellular targets of celastrol and gedunin but both possess antioxidant and anti-inflammatory properties. Recently, a chemical genomic approach was used to identify these compounds as inhibitors of HSP90, albeit by a different mechanism than 17-AAG or geldanamycin (21). It was of interest, therefore, to evaluate whether these natural HSP90 inhibitors blocked the activation of AhR-dependent gene expression. Similar to 17-AAG, both celastrol and gedunin blocked TS-mediated induction of CYP1A1 and CYP1B1 proteins. Because the magnitude of the suppressive effects of celastrol and gedunin were similar, additional mechanistic studies were only carried out with celastrol. Although less potent than 17-AAG, celastrol also caused a rapid decrease in amounts of AhR and suppressed PAH-mediated induction of CYP1A1 and CYP1B1 transcription. Whether celastrol and gedunin, natural inhibitors of HSP90, offer any advantages over 17-AAG or related compounds is uncertain and deserves further study.

The major pathway by which B(a)P is metabolized to a genotoxic metabolite involves its oxidation to BPDE. The epoxidation reactions are catalyzed by the cytochrome P450 mixed function oxidase system, with CYP1A1 and CYP1B1 playing a major role (30, 31). BPDE is extremely reactive and spontaneously hydrolyzes in aqueous medium to B(a)P-tetrols. A small fraction reacts with macromolecules, including DNA. The major site of adduction in mammalian systems is the exocyclic N² position of guanine (39). Mispairing as a result of replication across the adducted guanine is thought to lead to mutations and initiation of carcinogenesis (40). Agents that reduce the formation of BPDE inhibit the formation of DNA adducts and production of B(a)P-tetrols. Here, we show that HSP90 inhibitors suppress the formation of both products.

As B(a)P is stable in the cell medium in the absence of cells, B(a)P-tetrols in the medium result from intracellular metabolism of B(a)P and simple diffusion into the medium. The amount of total radioactivity in the medium represented only about 20% of the input radioactivity, presumably because most of the B(a)P and its metabolites were sequestered within the cells. There was good concordance between the effects of inhibitors on levels of B(a)P-tetrols in the medium and adducts in the DNA. This suggests that levels of B(a)P-tetrols in the medium serve as markers for cellular metabolism of B(a)P to BPDE. Collectively, our results suggest that 17-AAG and celastrol block PAH-mediated induction of CYP1A1 and CYP1B1, resulting in reduced B(a)P metabolism and DNA adduct formation. An alternative explanation, such as the induction of DNA repair activity by 17-AAG or celastrol, would not explain the strong reduction in levels of B(a)P-tetrols found in the medium following treatment of cells with HSP90 inhibitors.

At first glance, it appears predictable that suppressing the induction of CYP1A1 and CYP1B1, carcinogen-activating enzymes, should reduce the formation of DNA adducts and inhibit carcinogenesis. However, several lines of evidence suggest that this may not be the case in vivo. Pretreatment of rats with PAH results in reduced tissue levels of orally administered B(a)P because of enhanced metabolism (41). Moreover, studies with CYP1A1 and CYP1B1 knockout mice have suggested that these proteins protect against B(a)P toxicity because in their absence, levels of DNA adducts are higher (42–44). Finally, some but not all models of PAH-induced carcinogenesis suggest that activation of AhR signaling may protect against the carcinogenic effects of PAH (15). The balance between tissue-specific expression of the CYP1A1 and CYP1B1 enzymes may regulate sensitivity to B(a)P toxicity and perhaps carcinogenicity (44). In addition to suppressing PAH-mediated induction of CYP1A1 and CYP1B1, inhibitors of HSP90 have multiple other effects. It is predictable, for example, that HSP90 inhibitors will down-regulate levels of multiple client proteins and suppress the induction of other AhR-dependent genes. Given the potential limitations of our in vitro findings, additional preclinical studies involving animal models will be needed to determine whether HSP90 inhibitors suppress PAH-induced DNA adduct formation, mutagenesis, and carcinogenesis. If inhibiting HSP90 reduces the development of PAH-induced malignancies in preclinical models, it will be important to consider clinical scenarios in which these agents might be used. Elevated levels of CYP1A1 and CYP1B1 are detected in the upper aerodigestive tracts of smokers (24, 25, 45). Patients with a history of head and neck cancer who continue to smoke are at increased risk for second primary tumors (46). Whether administering inhibitors of HSP90 will reduce the risk of second primaries in this patient population could be evaluated. In this context, the possibility of administering HSP90 inhibitors topically as a rinse or lozenge to reduce toxicity should be considered. Certainly, this approach would minimize the theoretical concern that HSP90 inhibitors might increase PAH-induced DNA adduct formation in internal organs (41–44). Our findings have other potential clinical implications. In addition to being able to activate carcinogens, CYP1A1 and CYP1B1 play a role in the metabolism and clearance of a variety of drugs, including anticancer agents (47, 48), and many xenobiotics and endogenous substrates (48). It is possible, therefore, that inhibitors of HSP90 will affect the metabolism and clearance of other therapeutic drugs, which could affect both their efficacy and toxicity. Similarly, the metabolism and clearance of xenobiotics and endogenous substrates could be affected (48). In future studies of HSP90 inhibitors, our findings suggest that it will be important to monitor blood levels of coadministered drugs that are metabolized by enzymes encoded by AhR-dependent genes.

No potential conflicts of interest were disclosed.