Women are at higher risk for the development of lung adenocarcinoma than men; however, the mechanisms responsible for this are poorly understood. In lung adenocarcinoma cells, the estrogen receptor β (ERβ) is the predominating form. We found that 17β-estradiol enhanced proliferation of the putative cells of origin of lung adenocarcinoma, small airway epithelial cells (HPLD1), in response to the nicotine-derived nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK). Reverse-phase protein microarrays combined with Western blotting revealed that NNK induced phosphorylation of ERβ, an effect that involved stimulation of the adrenergic receptors β1 (β1AR). In transiently transfected cells, β1AR coprecipitated with ERβ, which increased with NNK treatment. ERβ enhanced NNK-induced cyclic AMP accumulation as well as Gαi-mediated mitogen-activated protein kinase/extracellular signal–regulated kinase (ERK) 1/2 activation. Coexpression of β1AR and ERβ activated NNK-mediated ERK1/2 cooperatively. ERβ gene knockdown, as well as coexpression of the dominant negative Ras and Raf, reduced stimulation of ERK1/2 by NNK. Whereas NNK phosphorylated Akt at Thr308 and Ser473, ERβ had no effect on this activity. Luciferase reporter assays showed that, in response to NNK, ERβ stimulated transcription of serum responsive element (SRE) but had a very small effect on the activity of estrogen responsive element (ERE). Together, the phosphorylation of ERβ, the dependence on Gαi proteins, the activation of ERK1/2, and the preferential targeting of SRE over the classic ERE pathway support a role for nongenomic ERβ in the development of smoking-associated lung cancer. This novel cooperation between β1AR and ERβ signaling may contribute to the prominence of lung adenocarcinoma in women. [Cancer Res 2007;67(14):6863–71]
Smoking is associated with ∼90% of all lung cancer cases. Adenocarcinoma of the lung is thought to arise from the epithelial lining cells of small airways and is the most common type of lung cancer today (1, 2). The risk for the development of lung adenocarcinoma is significantly greater in women than men; however, the reasons for this gender difference are poorly understood (3).
Estrogen signaling has been shown to play an important role in lung biology and pathology (4). Although the data are still conflicting, an association between high expression levels of estrogen receptors (ER) and occurrence of pulmonary adenocarcinoma has been reported (5–10). The lung is an estrogen-responsive organ, and ERβ is the predominant form in pulmonary adenocarcinoma (7). In female transgenic mice, inactivation of this receptor resulted in severe morphologic aberrations of the lung, thus revealing a crucial role for ERβ in lung biogenesis (11–13). The classic genomic estrogen pathway involves mostly the association of 17β-estradiol (E2) with nuclear ERα and ERβ receptors. ERs regulate gene transcription through direct interaction with specific estrogen responsive elements (ERE) and by interaction with other transcription factors (e.g., Sp1 and AP-1) bound to their response elements (14, 15). However, recent reports have also shown estrogen action at the membrane levels. Nongenomic ER activity induces rapid activation of multiple signal transduction pathways, including the mitogen-activated protein kinase (MAPK) extracellular signal-regulated kinase (ERK) 1/2 (16–19).
The tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is the most powerful cancer-causing agent in tobacco products. NNK induces development of lung adenocarcinoma in all laboratory rodents tested (20, 21). Its metabolites interact with DNA to form DNA methyl and pyridyloxobutyl adducts, which are thought to be crucial for its carcinogenic effects (22–24). In addition to its genotoxic effects, we have previously shown that NNK is an agonist for the adrenergic receptors β1 and β2 (β1AR and β2AR). Furthermore, we have shown that NNK stimulates proliferation of pulmonary adenocarcinoma and the putative cells of origin of this cancer, small airway epithelial cells, through a cross talk between β1AR and epidermal growth factor receptors (EGFR; refs. 25, 26). Other reports have shown that NNK-induced β-adrenergic signaling was associated with cell survival and growth regulation in several other non–small-cell and small-cell lung cancer cell lines and in colon cancer cells (27, 28).
To determine whether ERβ directly modulates NNK intracellular signaling, we overexpressed and knocked down these receptors in the human immortalized small airway epithelial cell line HPL1D and analyzed the responses to NNK exposure.
Materials and Methods
Cell line and tissue culture. The human immortalized small airway epithelial cell line HPL1D (29) was cultured at 37°C in phenol red–free Ham's F12 supplemented with 1% charcoal-stripped FCS, 5 μg/mL insulin, 5 μg/mL human transferrin, 50 nmol/L hydrocortisone hemisuccinate, 4.75 pmol/L 3,3′,5′-triiodo-l-thyronine, and 50 nmol/L Na-selenite.
Chemicals and antibodies. NNK was purchased from Chemsyn Laboratories. E2 and pertussis toxin were purchased from Calbiochem. Antiestrogen ICI 182780 and adrenergic receptor antagonists atenolol and propanolol were obtained from Tocris. ERK1/2, phospho-ERK1/2 (Thr202/Tyr204) antibodies, and horseradish peroxidase–conjugated antirabbit secondary antibodies were purchased from Cell Signaling Technology. ERβ, phospho-ERβ (Ser87), and ERα antibodies were purchased from Santa Cruz Biotechnology. Alexa Fluor 680 probes were obtained from Molecular Probes.
Treatment protocol. Cells were starved in basal medium without serum or additives for 16 to 24 h before treatment with 100 nmol/L E2 for 10 min or 1 μmol/L NNK for 1 h. Simultaneous exposure to E2 and NNK was done by pretreating with 100 nmol/L E2 for 10 min and then removing the medium and replacing it with basal medium containing 1 μmol/L NNK for 1 h.
Transient transfection and RNA interference induction. ERβ pBI-EGFP, ERα pBI-EGFP, pERE-tat-Luc, and pSRE-Luc were kindly provided by Dr. Jay Wimalasena. β1AR cDNA was provided by Dr. R.J. Lefkowitz (Department of Biochemistry, Duke University Medical Center, Durham, NC). Cells were plated at 4 × 105 per 10-cm-diameter plate and grown for 24 h. The cells were cotransfected with 2 μg of the indicated expression vector and 500 ng of pcMV-LacZ to monitor transfection efficiency using Lipofectamine reagents (Invitrogen). Cells transfected with empty vector (pBI-cDNA) served as controls. ERβ gene knockdown was accomplished by transfecting cells with 40 nmol/L ERβ Stealth RNA interference (RNAi; ref. 30). Small interfering RNA (siRNA) transfection was done in the presence of Lipofectamine 2000 according to the manufacturer's protocol. The siRNA and Lipofectamine were diluted with Opti-MEMI medium (Invitrogen), combined together, and incubated for 20 min before adding the mixture to the cells. To determine whether siRNA blocked ERβ expression, cells were harvested 72 h posttransfection for evaluation of ERβ proteins by Western blotting. As a negative control for the effects that may be associated with siRNA delivery, we transfected cells with RNAi that has high GC duplexes (Invitrogen). The primers used to target ERβ mRNA sequence for interference were sense, CCAGCAAUGUCACUAACAU, and antisense, AAGUUAGUGACAUUGCUGG.
Cell proliferation assay. The quantification of HPL1D cell proliferation was based on the measurement of bromodeoxyuridine (BrdUrd) incorporation according to manufacturer's instructions (Roche Applied Science). In 96-well plates, 10,000 cells were cultured in basal phenol red–free medium for 24 h and either left alone or exposed to 100 nmol/L or 1 μmol/L NNK for 1 h; pretreated with 100 nmol/L E2 for 10 min followed by NNK stimulation for 1 h; and pretreated with 100 nmol/L ICI 182780, or 1 μmol/L propanolol, or 1 μmol/L atenolol for 10 min before simultaneous stimulation by E2 and NNK. At the end of each treatment period, the medium was removed and replaced with basal medium without treatment agents. In parallel, cells were transfected with 40 nmol/L ERβ siRNA or ERα siRNA and treated under the same condition as their untransfected counterparts. Each experiment was run twice with three samples per treatment group and BrdUrd incorporation was determined (31). The statistical significance of differences was calculated by one-way ANOVA and two-tailed t test analysis; P < 0.05 was considered significant.
Cyclic AMP immunoassay. Cells were grown until 80% to 85% confluence before transfection with 2 μg of ERβ expression plasmid. Forty-eight hours posttransfection, cells were fed with phenol red–free basal medium for 24 h. Following two washes with 1× PBS, cells were treated with 1 μmol/L NNK for 1 h or pretreated with 100 nmol/L E2 for 10 min before NNK stimulation in medium containing 1 mmol/L I-methyl-3-isobutylxanthine. Then, cells were treated with 0.1 N HCl for 10 min, lysed by sonication, and samples were analyzed for cyclic AMP (cAMP) levels with a direct cAMP immunoassay kit according to the manufacturer's instructions (Assay Designs, Inc.). The statistical significance of differences was calculated by one-way ANOVA and two-tailed t test analysis; P < 0.05 was considered significant.
Reverse-phase protein microarray assay. Cells were lysed in a buffer containing 6 mol/L urea (Sigma), 65 mmol/L DTT, 2% Pharmalyte (pH 8–10.5), and 1% CHAPS (32). Samples were cleared by centrifugation at 4°C for 15 min. Approximately 60 nL of each sample were spotted onto nitrocellulose-coated glass slides (Schleicher & Schuell Bioscience) using a pin-in-ring format Affymetrix GMS 417 Arrayer (Affymetrix). Each sample was spotted in triplicate in a stepwise series of 2-fold dilutions 1, 2, 4, 8, 16, and 32. The arrayed slides were incubated with primary antibody for 1 h, washed, and reincubated in cognitive secondary antibody conjugated to Alexa Fluor 680 probe and scanned with GenePix 4000B microarray scanner (Axon Instruments). Spot images were converted to raw pixel values and analyzed by a modified version of GenePiX 5.1 software. The background was subtracted from the intensity values.
Western blot and immunoprecipitation. Equal amounts of protein were resolved on SDS-polyacrylamide gels (12% acrylamide), transferred to nitrocellulose membranes, and probed with the appropriate antibodies. For immunoprecipitation, cell lysates were incubated with 2 μg of the appropriate antibody and precleared with protein A/G-Sepharose for 2 h. Immunocomplexes were washed with 1× immunoprecipitation buffer [50 mmol/L Tris-base (pH 7.5), 5 mmol/L EDTA, 20 mmol/L β-glycerol-phosphate, 150 mmol/L NaCl, and 1% NP40 IP]. Proteins were eluted by boiling in 2× sample buffer, separated by SDS-PAGE (12%), and then transferred to nitrocellulose membrane. Enhanced chemiluminescence (ECL-plus) was used for detection.
ERK1/2 kinase assay. Equal amounts of proteins were incubated overnight with 15 μL of agarose hydrazide beads immobilized p44/42 (Cell Signaling Technology). Immunoprecipitates were washed thrice in 100 mmol/L Tris (pH 7.5), 1% NP40, 2 mmol/L sodium orthovanadate; once in 100 mmol/L Tris (pH 7.5), 0.5 mol/L lithium chloride; and once in kinase buffer [12.5 mmol/L MOPS (pH 7.5), 12.5 mmol/L β-glycerophosphate, 7.5 mmol/L MgCl2]. Proteins were incubated for 20 min at 30°C in a 30-μL kinase reaction containing 2-μg Elk-1 fusion protein (GST-Elk1 codons 307–428) and 10 μmol/L ATP. Proteins were separated on a 12% SDS-polyacrylamide gel, transferred to nitrocellulose membrane, and probed with anti–phospho-Elk (Ser383) antibody. After incubation with the secondary antibody, bands were visualized by enhanced chemiluminescence.
Luciferase reporter assays. Cells were cotransfected with 500 ng of pRSV-β-galactosidase and 1 μg of ERE-Luc, or serum responsive element (SRE)-Luc cDNAs, alone or along with 2 μg ERβ or 40 nmol/L ERβ siRNA constructs (33). Twenty-four hours posttransfection, cells were deprived of serum for 18 h before stimulation with the indicated agents. Untransfected and transfected cells with the empty vector, either stimulated with the same agents or left alone, served as controls. Cells were harvested 24 h later and luciferase and β-galactosidase activities were then measured using standard luciferase (Promega) and β-galactosidase detection kits (Applied Biosystems).
Statistical analysis. All data shown as bar graphs are expressed as the mean ± SE. The statistical significance of differences was calculated by one-way ANOVA and two-tailed t test analysis (34). Two-tailed P ≤ 0.05 was considered to be statistically significant.
E2 stimulates NNK-dependent BrdUrd incorporation. We have reported that NNK stimulated proliferation of the putative cells of origin of lung adenocarcinoma, small airway epithelial cells (HPLD1), via β-adrenergic receptor signaling (25, 26). To determine whether estrogen modulates this response, we analyzed proliferation of these cells after exposure to E2 for 10 min, exposure to NNK for 1 h, and pretreatment with E2 for 10 min before exposure to NNK for 1 h. In addition, cells expressing ERβ or ERα siRNA were treated in similar fashion. To determine whether siRNA blocked ERβ expression, cells were analyzed for ERβ proteins by Western blotting 72 h posttransfection. As a negative control for the effects that may be associated with siRNA delivery, we used oligos that consist of >50% GC content (Invitrogen). As expected, stimulation with 1 μmol/L NNK produced a 4-fold increase in BrdUrd incorporation (Fig. 1A). Pretreatment with E2 enhanced BrdUrd levels by almost 8-fold. Exposure to 100 nmol/L ICI 182780 for 10 min had a slight reducing effect on E2 + NNK stimulation. By contrast, treatment with 1 μmol/L β1 and β2AR antagonist, propranolol, or the site-selective antagonist for β1AR, atenolol, for 10 min reduced this response to E2 and NNK significantly (P < 0.05). BrdUrd incorporation in cells expressing ERβ siRNA or ERα siRNA treated simultaneously with E2 and NNK was reduced to 4-fold and 7-fold, respectively (Fig. 1B). These results suggest that in these cells, E2 and NNK stimulation promoted entry into S phase at a significantly higher rate than either compound alone. As indicated by the greater effect of ER-β silencing, this activity involves predominately ERβ signaling via β1AR.
NNK induces phosphorylation of ERβ. We focused on ERβ instead of ERα because in lung cancer cells, the former seems to be the predominating form (1, 10, 11). In addition, the tested cell line expresses more of ERβ proteins (data not shown). Furthermore, the proliferation results showed that interfering with ERα reduced the induction in response to E2 + NNK less than silencing of ERβ. First, we used reverse-phase protein microarrays to analyze phosphorylation of ERβ in response to 100 nmol/L and 1 μmol/L NNK stimulation. For validation purpose, increasing concentrations of HPL1D lysates (60 ng–1 μg) were spotted onto nitrocellulose-coated slides and probed with antimouse conjugated with Alexa Fluor 680. To emulate a cell lysate, Alexa labelings were carried out in a background of 1% bovine serum albumin. Control slides consisted of arrayed lysates without the primary antibody and unprinted slides with the primary antibody and the Alexa probe. The arrays were sensitive enough to detect 60 ng of spotted lysates. A linear increase in spot intensity was observed with increasing amounts of proteins. No immunoreactivity was observed in either control, which shows the specificity of the microarray (data not shown).
The results show that in response to 100 nmol/L and 1 μmol/L NNK exposure, phosphorylation of ERβ (Ser87) was stimulated by 1.5- and 5-fold, respectively (Fig. 2A). This response to NNK increased by 9-fold when cells were transfected with ERβ plasmid. By contrast, pretreatment with 1 μmol/L atenolol for 10 min reduced phosphorylation of ERβ (Ser87) by NNK significantly (P < 0.05). These findings are in accord with our previous observation that NNK stimulates cell proliferation by a mechanism that involves ligand binding of NNK to the β1AR (25, 26).
To further substantiate these observations, lysates from untransfected and ERβ transfected cells were immunoprecipitated with ERβ antibody, immunoblotted with p-ERβ (Ser87) antibody, and analyzed for phosphorylation with Western blots. The results were similar to those obtained by the reverse-phase approach, showing the validity of the assay (Fig. 2A). Together, these findings show that NNK induced phosphorylation of ERβ in a concentration-dependent manner, and that this effect required β1AR. The rapid ERβ phosphorylation and the dependency on membrane β1AR activation suggest that nongenomic ERβ modulates NNK intracellular signaling.
ERβ coprecipitates with β1AR and NNK further promotes this complex. Cells were transiently cotransfected with ERβ and β1AR plasmids and treated with 1 μmol/L NNK for 1 h. Immunoblot analysis of proteins immunoprecipitated with ERβ (Fig. 2B) or β1AR antibody (Fig. 2C) revealed that under basal conditions, both proteins coprecipitated. However, when cells were treated with NNK, complex formation between ERβ and β1AR was significantly evident (P < 0.05). These results suggest that NNK promotes ERβ-β1AR complex formation.
ERβ enhances NNK-dependent cAMP activity. Previous studies have reported increase of cAMP levels as a result of membrane E2 receptor activation (35, 36). To determine whether phosphorylation of ERβ by NNK used intracellular cAMP as a second messenger, we measured its levels in ERβ transfectants after exposure to 100 nmol/L E2 for 10 min; 1 μmol/L NNK for 1 h; or pretreatment with E2 before stimulation with NNK. Control samples consisted of untransfected cells and empty vector–transfected cells. NNK treatment of untransfected cells increased cAMP levels by 6-fold as compared with basal levels (Fig. 3). Pretreatment with E2 slightly increased cAMP response to NNK. On the other hand, ERβ transient expression enhanced NNK stimulation by almost 15-fold. E2 pretreatment had no significant effect on this activity. These findings show that overexpression of ERβ enhanced NNK-induced cAMP activation significantly (P < 0.05).
ERβ enhances NNK-dependent ERK1/2 enzymatic activity. Recently, we have reported that in these cells, NNK activates the MAPK ERK1/2 signaling pathway via β1AR-mediated transactivation of EGFR (26). To determine whether ERβ modulates ERK1/2 response to NNK, cells were transfected with 2-μg ERβ plasmid, serum starved, and then stimulated as indicated. ERK1/2 enzymatic activity was analyzed by an in vitro kinase assay. As shown by GST-Elk-1 fusion protein phosphorylation, E2 and NNK, alone, stimulated ERK1/2 by 2- and 3.5-fold, respectively (Fig. 4A). E2 and NNK cotreatment produced an 11-fold increase. Transient expression of ERβ enhanced ERK1/2 response to NNK by 9-fold.
The effects of pertussis toxin, an inhibitor for G protein Gαi, and ERβ gene knockdown were analyzed. We found that pertussis toxin exposure for 10 min blocked ERK1/2 response to E2 and NNK (Fig. 4B). We tested several Stealth RNAi oligos targeting different ERβ mRNA sequences and found one transcript that exerted the highest level of suppression of total ERβ expression. The results show that interfering with ERβ transcription reduced ERK1/2 response to E2 and NNK to almost half (P < 0.05). This suggests a functional role for nongenomic ERβ in NNK action on Gαi-mediated ERK1/2 signaling.
Coexpression of β1AR and ERβ enhances ERK1/2 activity synergistically. Cells were cotransfected with 2-μg ERβ and 2-μg β1AR, starved for 24 h, and stimulated with 100 nmol/L E2 for 10 min or 1 μmol/L NNK for 1 h. As shown in Fig. 4C, whereas E2 treatment of ERβ and β1AR cotransfectants enhanced ERK1/2 activity by 4-fold, NNK stimulation resulted in a 12-fold increase. This cooperative effect shows the involvement of β1AR in ERβ-mediated NNK signaling.
ERβ stimulation of NNK-dependent ERK1/2 requires Ras and Raf proteins. Recently, we have reported Raf overexpression in NNK-induced pulmonary adenocarcinoma in hamsters (37, 38). Therefore, we cotransfected cells with ERβ and the dominant negative Ras (Ras17N) or Raf (RafC4) and analyzed the effect of NNK on ERK1/2 activity. Figure 4D shows that transient expression of Ras17N as well as RafC4 blocked activation of ERK1/2 by E2 and NNK (P < 0.05). In addition, it interfered with the ability of ERβ to stimulate ERK1/2 in response to NNK. These findings indicate that ERβ enhances the actions of NNK on ERK1/2 via Ras and Raf proteins.
ERβ overexpression had no apparent effect on NNK-induced Akt phosphorylation. West et al. (39) have reported that in normal human airway epithelial cells, NNK activates Akt. As expected, we found that in these cells, NNK stimulated Akt phosphorylation at two sites, Thr308 and Ser473 (Fig. 5A and B). However, pretreatment with E2 or ERβ overexpression did not significantly increase this activity (P > 0.05), suggesting that in these cells, the actions of ERβ on NNK-induced ERK1/2 activity do not involve the Akt pathway.
ERβ transient expression constitutively activates Gαi-mediated serum responsive element transcription but slightly increases ERE activity. Phosphorylation of Elk-1 by ERK1/2 binds and stimulates transcription of the serum responsive elements (SRE; refs. 40, 41). Therefore, we analyzed the effects of ERβ overexpression and NNK stimulation on SRE transcription. Cells were either transfected with SRE-Luc reporter construct, alone or cotransfected with ERβ, or cotransfected with ERβ siRNA and treated with the indicated agents. The results show that in untransfected cells exposed to 10 nmol/L, 100 nmol/L, and 1 μmol/L NNK, SRE-Luc activity increased by 3-, 14-, and 17-fold, respectively (Fig. 6A). Whereas E2 alone (100 nmol/L) produced a little more than 2-fold induction, addition of 1 μmol/L NNK stimulated this activity by 27-fold. Pertussis toxin blocked this synergy. Exposure of ERβ transfectants to 10 nmol/L, 100 nmol/L, and 1 μmol/L NNK increased SRE-Luc activity by 4-, 19-, and 20-fold, respectively (Fig. 6B). ERβ gene knockdown reduced NNK-induced SRE transactivation by 2-, 5-, and 8-fold.
Lastly, we assessed the effect of ERβ overexpression on NNK-mediated ERE responses, the classic ligand-dependent E2-ER complexes (42). Contrary to the potent stimulation of SRE, we found a slight increase in ERE-Luc activity in response to E2 or NNK alone. This activity was not significantly enhanced by cotreatment with both agents (P > 0.05; Fig. 6C and D). Together, these findings show that the high increase of SRE transcriptional activity by E2 and NNK is regulated by Gαi. The fact that ERE transcription was not markedly enhanced suggests that NNK signaling involves mostly nongenomic ERβ.
Although the role and expression of ERs in lung tumors have been controversial for many years, recent studies have shown that ERβ is the predominating form (9–11). In this investigation, we used the human small airway epithelial cell line HPL1D to determine whether ERβ signaling modulates NNK-induced β-adrenergic signaling. We found that rapid exposure to E2 significantly enhanced cell proliferation in response to NNK. The molecular mechanisms underlying this cooperative effect remain to be defined. However, the magnitude of this response after a short-term application strongly suggests the involvement of nongenomic ERβ. The inhibitory effects of adrenergic receptor β-blockers, propanolol and atenolol, on E2/NNK synergy suggest that each agent acts at a distinct set of receptors at the membrane level, NNK at β1AR and E2 at ERβ. Thus, the signaling persistence at these sites may induce ERβ/β1AR complex formation with changes in the pattern of phosphorylation. This will result in the loss of normal regulatory constraints and confer a growth advantage. Alternatively, E2 and NNK may have affinity for the same receptors, including ERβ, and use the same signaling pathways to activate specific genes important for proliferation. Therefore, simultaneous treatment with E2 and NNK would produce higher proliferative responses than either compound alone. The fact that the classic antiestrogen ICI 182780 had only a slight negative effect on E2 + NNK–induced proliferation of these cells suggests the involvement of predominantly nongenomic ERβ.
Next, we analyzed the activation status of ERβ. We found that at nanomolar concentrations, NNK stimulated phosphorylation of ERβ (Ser87) in a dose-dependent manner. This activity was markedly enhanced by ERβ overexpression and substantially reduced after atenolol exposure. These findings suggest that the action of NNK on ERβ is regulated by the transmembrane receptor β1AR. This would support a role for nongenomic ERβ in NNK-induced signaling activities. Perhaps, NNK-activated β1AR recruits membrane or cytosolic ERβ, and this may explain the rapid action of NNK on the latter. Nevertheless, the results clearly indicate that β1AR contributed to the marked phosphorylation of ERβ by NNK, therefore suggesting a cross talk between the two pathways. Our coprecipitation experiments clearly show that in these cells, NNK enhanced the small relative amount of β1AR and ERβ that are complexed together. Additional studies will be needed to clarify the role of ERβ modulation of NNK-induced β1AR activity. Determination of the precise stereochemistry of β1AR/ERβ interaction would be of great interest, considering the biological consequences of this interaction.
When we exposed cells to both E2 and NNK, cAMP activity increased only slightly. A possible explanation for these results is that E2 and NNK stimulate different adenylyl cyclase isozymes contributing to basal cAMP production; therefore, costimulation with both agents may not produce a synergistic accumulation of cAMP. Alternatively, cAMP increase in response to NNK was already at maximum, or non–adenylyl cyclase–dependent signaling pathways may have contributed to the observed slight increase. ERβ transfectants stimulated with NNK exhibited the highest levels of cAMP, much more than their counterparts treated with E2. This suggests that the contribution of the nongenomic cAMP pathway of ERβ action is significant and further strengthens the argument for the implication of nonnuclear ERβ.
Recently, we have reported that NNK activates the MAPK/ERK1/2 pathway via β1AR transactivation of EGFR in HPL1D and human pulmonary adenocarcinoma cells (26). When we transfected cells with ERβ construct and applied NNK, ERK1/2 enzymatic activity increased pronouncedly. In contrast, ERβ gene knockdown reduced NNK-dependent ERK1/2 activation to about half. The fact that ERβ down-regulation did not completely abrogate ERK1/2 stimulation by NNK suggests that the latter targets nongenomic ERβ indirectly via cross talks with other membrane receptors. One scenario would be that NNK induces bidirectional signaling between nonnuclear ERβ and β1AR, and possibly others. This would explain the synergistic activation of ERK1/2 in β1AR and ERβ cotransfectants.
The Gαi inhibitor pertussis toxin interfered with the ability of E2 and NNK to activate ERK1/2 cooperatively. This indicates that E2-NNK signaling complexes operate via a Gαi-dependent mechanism. Perhaps, E2 exposure leads to Gαs activation, then NNK stimulation results in switching from Gαs to Gαi. In addition, we cannot exclude the possibility that both agents might induce sequential activation of Gαs and Gαi proteins. Similar mechanisms have been described during ERK1/2 activation by β-adrenergic receptors in Chinese hamster ovary cells (43, 44).
Activating mutations in K-Ras are one of the most common genetic alterations in human lung adenocarcinoma (45–47). It is evident from the observed blocking effects of Ras17N and RafC4 that the resulting signals from E2/ERβ and NNK stimulation converge onto the Ras-Raf–mediated ERK1/2 pathway.
Next, we found NNK-stimulated phosphorylation of Akt at Thr308 and Ser473; however; pretreatment with E2 or ERβ overexpression did not enhance this activity. Akt is commonly phosphorylated at Thr308 and Ser473 in tumor cells through activation of Ras. This protein has been shown to contribute to NNK-induced cell proliferation in a nicotinic acetylcholine receptor–dependent manner (39, 48). Ligand-binding of NNK to nicotinic acetylcholine receptors resulting in the activation of ERK1/2 and c-myc and subsequent stimulation of cell proliferation was first reported in human small-cell lung cancer cell lines and their putative cells of origin, pulmonary neuroendocrine cells (49, 50). The high affinity of NNK to members of both the β-adrenergic and nicotinic acetylcholine receptor families likely contributes to the exceptional carcinogenic potency of this tobacco nitrosamine. The fact that E2 did not enhance NNK-induced Akt phosphorylation suggests that ERβ cooperates with β-adrenergic but not nicotinic receptor signaling. This interpretation is in accord with a recent report showing that nicotinic receptor–mediated activation of Akt by NNK was not inhibited by β-adrenergic antagonists (48).
The final part of this investigation analyzed the transcriptional activity of SRE and ERE gene expression using luciferase reporter gene assays. We found that E2 and NNK cotreatment stimulated SRE synergistically and that ERβ overexpression increased this response significantly (P < 0.05, two-tailed t test). ERβ gene knockdown reduced NNK stimulation of SRE as to about half. In contrast to SRE, transcription of ERE was slightly affected by ERβ overexpression or by E2 and NNK costimulation. Because of the short duration of exposure, the involvement of G proteins, and the rapid activation of MAPK/ERK1/2, we expected a little to no effect on ERE transcription. One possible explanation for the slight increase in ERE is that once nongenomic ERβs are activated by NNK, they exert some residual actions on their nuclear counterparts. The data strongly suggest that NNK-induced ERβ activation targets predominantly SRE gene expression. ERβ modulates NNK activity, as evidenced by the reducing effect of ERβ gene knockdown on E2-and NNK-dependent activation of SRE gene expression.
We need to mention that we have tested the actions of ERα on NNK signaling in HPL1D cells. We found that ERα overexpression stimulated NNK actions very moderately. In addition, we found that ERα gene knockdown had very small reducing effects on NNK signaling, thus suggesting that this receptor is not the major estrogen player in NNK-mediated activity in HPL1D cells.
In summary, our data have identified a novel and hitherto unknown cooperation between NNK-induced β1AR and nongenomic ERβ mitogenic signaling, leading to up-regulation of the cAMP/Ras/Raf/ERK1/2/Elk1 signaling pathway in the cell type of origin of lung adenocarcinoma. Although further investigation is required to understand how precisely estrogen and smoking can affect tumor development, the results give new mechanistic insights into the prevalence of lung adenocarcinoma in women. Our current and published data (25, 26) suggest that in a subset of human patients, small airway epithelial cells and the adenocarcinomas derived from them are under β-adrenergic growth control and that NNK stimulates this pathway in cooperation with the ERβ. The levels of expression and the sensitivity of both β1AR and ERβ can be modulated by numerous factors associated with our modern environment and lifestyle. Among such factors are numerous decongestion, cold, and asthma medicines and dietary supplements that increase intracellular cAMP and thereby sensitize the β-adrenergic receptors. On the other hand, estrogen contraceptives or estrogen replacement therapy, as well as exposure to environmental toxicants with estrogenic activity, may further enhance signaling via stimulation of nonnuclear ERβ. Depending on the exposure history, there will thus be significant interindividual variations in β1AR and ERβ signaling in response to NNK. An in-depth understanding of the molecular basis of estrogenic influence on lung adenocarcinoma development in women is urgently needed to improve strategies for the prevention and therapy of this malignancy.
Grant support: National Cancer Institute grants RO1 CA088809 and RO1 CA096128.
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.
We thank Dr. Jay Wimalasena (UT Medical Hospital, Knoxville, TN) and Dr. H-C.R. Wang (College of Veterinary Medicine, Knoxville, TN) for providing us with expression vectors and for technical advice.