Retinoids, a class of natural and synthetic vitamin A analogs, exert profound effects on many biological processes, including cell proliferation and differentiation,1, 2 and have been recognized as promising agents for the prevention and treatment of a variety of human cancers.1, 2 Because epidemologic studies have demonstrated a close correlation between cancer development and dietary consumption of β-carotene and vitamin A,3, 4 retinoids have undergone extensive preventive and therapeutic trials against cancer.5–10 However, several trials that involved a large number of cigarette smokers demonstrated that neither vitamin A nor its metabolic precursor β-carotene had a beneficial effect on the incidence of lung cancer.9, 10
The effects of retinoids are mainly mediated by 2 classes of nuclear receptors: the retinoic acid receptors (RARs) and retinoid X receptors (RXRs).11–13 9-cis-retinoic acid (RA) is a high affinity ligand for both RARs and RXRs, whereas trans-RA is a ligand for only RARs. RARs and RXRs are encoded by 3 distinct genes (α, β and γ) and are members of the steroid/thyroid hormone receptor superfamily that function as ligand-activated transcription factors.11–13 RARs and RXRs primarily function as RXR/RAR heterodimers that bind to a variety of RA-response elements (RAREs) to regulate their transactivation activities.11–13 The effects of RAR/RXR heterodimers are mediated through binding of trans-RA to RAR, while RXR appears to function as a silent partner.11–13 However, activation of RXR by its ligand is required for the function of RXR homodimers14 and some of the RXR-containing heterodimers, such as RXR/TR3.15, 16
Recent studies have indicated that RARβ plays a critical role in mediating the growth inhibitory effect of retinoids in many different types of cancer cells,17–28 in part through its potent apoptosis-inducing effect20 and anti-AP-1 activity.29 Expression of RARβ is highly induced by trans-RA through the RARE (βRARE) present in its promoter,30–32 which is mainly activated by RAR/RXR heterodimers.33 In addition to RAR and RXR, several orphan receptors, such as COUP-TF (also called ARP-1 and ear-3)34, 35 and TR3 (also called NGFI-B and nur77),36–38 have been implicated in the regulation of RARβ expression and the retinoid responses. Expression of COUP-TF correlates with induction of RARβ by trans-RA in various cancer cell lines and is required for trans-RA to induce RARβ, growth inhibition and apoptosis.24, 39 COUP-TF, by binding to its response element COUP-TF-RE in the RARβ promoter, enhances the recruitment of a receptor coactivator to the RARβ promoter by the liganded βRARE-binding receptors.39 TR3, whose expression is induced by a variety of stimuli, including growth factors, TPA and Ca++,37, 38, 40 regulates retinoid responsiveness through its ability to heterodimerize with RXR15, 16, 23 and COUP-TF.24 Interaction of TR3 with RXR results in formation of a TR3/RXR heterodimer that binds strongly to the βRARE and activates the RARβ promoter in response to RXR ligands.23 In contrast, interaction of TR3 with COUP-TF prevents COUP-TF from binding to DNA.24
Tobacco consumption is recognized as a major risk factor for developing lung cancer.41, 42 Nicotine is the major alkaloid responsible for addiction to tobacco.43 It exerts its action mainly in brain through nicotine acetylcholine receptors (nAChR). Interestingly, nAChR is expressed in a variety of cell types, including human lung cancer cell lines with diverse histologies.44–47 Nicotine induced secretion of mitogenic hormones, stimulated cell proliferation,45, 48–54 and inhibited apoptosis of lung cancer cells,55, 56 probably through the mitogenic-activated protein kinase (MAPK) and protein kinase C pathways.57–59 These observations suggest that nicotine may act as a tumor promoter during lung carcinogenesis.
The reason for the failure of retinoids to demonstrate anti-cancer effect in cigarette smokers in clinical trials is unclear. Retinoids may be ineffective in suppressing the progression of neoplasia after exposure to carcinogens. Alternatively, cigarette smoking may be involved in regulating retinoid activities. In our study, we investigated the effect of nicotine on retinoid-induced growth inhibition. Our results demonstrated that nicotine could abrogate the growth inhibitory effect of trans-RA in lung cancer cells by inhibiting RARβ expression. The inhibition of RARβ expression is due to induction of TR3 expression by nicotine, which suppresses the transactivation function of COUP-TF on the RARβ promoter. Interestingly, nicotine failed to suppress the growth inhibition and the activation of RARβ promoter by an RXR-selective retinoid SR11237 that activated the RARβ promoter through the TR3/RXR heterodimer. Thus, RXR-selective retinoids may be effective agents for preventing and treating tobacco-associated cancers.
MATERIAL AND METHODS
CV-1 cells were grown in DME medium supplemented with 10% fetal calf serum (FCS). H460 and H157 lung cancer cells were grown in RPMI1640 supplemented with 10% FCS. Cells were treated with trans-RA (Sigma Chemical Co., St. Louis, MO), SR11237 or nicotine ([−]-1-methyl-2-[3-pyridyl]-pyrrolidine; Sigma) as indicated.
Growth inhibition assays
Cells were seeded at 1,000 cells per well in 96-well plates and treated with various concentrations of agents for 7 days. Media were changed every 48 hr. The number of viable cells was determined by MTT assay as described previously.60
For Northern blot analysis, total RNAs were prepared by RNeasy Mini Kit (Qiagen, Germany). Total RNAs (30 μg) from H460 cells treated with or without 10−6 M trans-RA and/or nicotine were analyzed by Northern blotting.20 An Eco RI fragment in the ligand-binding domain of RARβ or a Pst I fragment in the ligand-binding domain of TR3 cDNA was used as a probe to study expression of RARβ or TR3, respectively.20, 61 To determine that equal amounts of RNA were used, the expression of β-actin was studied.
Preparation of COUP-TF and TR3 proteins and gel shift assay
COUP-TF and TR3 protein were synthesized by in vitro transcription/translation system in rabbit reticulocyte lysates (Promega, Madison, WI) as described previously.39 The relative amount of translated protein was determined by 35S-methionine-labeled protein on SDS-PAGE, followed by quantitating the amount of incorporated radioactivity and normalizing it relative to the content of methionine in each protein. The gel shift assay using in vitro-synthesized proteins was performed as described.39
Plasmids and transient transfection assay
The RARβ promoter reporters, -126RARβCAT and -60RARβCAT and expression vectors for RARα, RXRα, COUP-TF and TR3 have been described previously.39 For transfection assay, CV-1 cells were plated at 1 × 105 cells per well in a 24-well plate for 16–24 hr before transfection by calcium phosphate precipitation procedure as described previously.39 The precipitates contained 100 ng reporter plasmid, 100 ng of β-galactosidase expression vector and various amount of receptor expression vectors. The total amount of transfected DNA was adjusted with pBluescript to 1 μg. For H460 cancer cells, 5 × 105 cells were seeded in 6-well plates. The precipitates contained 250 ng of reporter plasmid, 250 ng β-galactosidase expression vector and various amounts of receptor expression vectors. The total amount of transfected DNA was adjusted with pBluescript to 2.5 μg. Twenty hours after transfection, cells were washed and incubated in a median containing 10% charcoal-treated FCS with retinoids and/or nicotine for 24 hr. CAT activity was determined by using [3H] acelyl-CoA as substrate, normalized for transfection efficiency using β-gal activity.33
Effect of nicotine on trans-RA-induced growth inhibition in lung cancer cells
To investigate the possibility that nicotine modulates the growth inhibitory effect of trans-RA, we determined the effects of nicotine on growth inhibition of H460 and H157 human lung cancer cells by trans-RA. Cells were treated for 7 days with 10−6 M trans-RA, which exerted maximal growth inhibitory effect without apparent nonspecific toxic activity.60 As shown in Figure 1, trans-RA significantly inhibited the growth of these cells, with more than 50% inhibition. When it was used together with nicotine, its growth inhibitory effect was reduced in a nicotine concentration-dependent manner, while nicotine alone had no discernible effect on the growth of both cell lines. The antagonistic effect of nicotine on trans-RA activity in H157 cells was observed when 10−6 M nicotine was used, while 10−5 M nicotine was required to achieve the effect in H460 cells. At 10−3 M, nicotine almost completely prevented the growth inhibitory effect of trans-RA in both cell lines, with only 10% growth inhibition observed.
Inhibition of trans-RA-induced RARβ expression by nicotine
Since induction of RARβ is known to mediate the growth inhibitory effect of trans-RA in various cancer types,17–24, 62 we examined whether the suppressive effect of nicotine on trans-RA activity was due to its attenuation of RARβ expression by studying the effect of nicotine on RARβ expression in H460 cells. RARβ was only slightly expressed in H460 cells as revealed by Northern blotting (Fig. 2). However, its level of expression was dramatically enhanced when cells were treated with trans-RA. Treatment of cells with nicotine slightly inhibited RARβ expression. Interestingly, when cells were pretreated with nicotine, trans-RA-induced RARβ expression was strongly suppressed. Thus, inhibition of RARβ expression by nicotine may account for its suppressive effect on trans-RA-induced growth inhibition.
Induction of TR3 by nicotine in human lung cancer cells
Recently, nicotine was reported to induce expression of orphan receptor TR3 in PC12 pheochromocytoma cells and in animals.63 Since TR3 expression is associated with retinoid resistance in cancer cells 24 and nicotinic acetylcholine receptors are expressed in a variety of human lung cancer cell lines,45 we examined whether inhibition of trans-RA activity by nicotine could result from its induction of TR3. H460 cells were treated with various concentrations of nicotine and analyzed for TR3 expression by Northern blotting. TR3 was expressed in H460 cells in the absence of treatment. However, TR3 expression was significantly enhanced when cells were treated with nicotine (Fig. 3). Treatment of H460 cells with 10−5 M nicotine resulted in about 2-fold induction of TR3, while about 6-fold induction was found when cells were treated with 10−3 M nicotine. Addition of trans-RA did not modulate the effect of nicotine on TR3 induction (data not shown).
To determine how nicotine inhibits trans-RA-induced RARβ expression, we examined the effect of nicotine on RARβ promoter activity by transient transfection assay in H460 cells. A reporter construct containing the RARβ promoter sequence from −126 to +162 linked to the CAT gene (−126RARβCAT) 39 was transfected into H460 cells. The RARβ promoter sequence −126 to +162 encompasses both the RA response element (βRARE) and COUP-TF response element (COUP-TF-RE) required for optimal induction of the RARβ promoter activity by trans-RA.39 Indeed, treatment of H460 cells with trans-RA strongly induced the reporter transcriptional activity (Fig. 4a). However, when cells were pretreated with nicotine, trans-RA-induced reporter activity was reduced by nicotine in a concentration-dependent manner. Thus, nicotine may inhibit RARβ expression through its modulation of RARβ promoter activity.
We also examined the effect of nicotine on the transcriptional activity of −60RARβCAT reporter containing RARβ promoter sequence from −60 to +162 (Fig. 4b).39 The −60RARβ CAT reporter, which lacks the COUP-TF-RE, was activated by trans-RA in H460 cells, although the activation was lower than the activation of the −126RARβCAT, due to lack of the COUP-TF-RE.24 However, unlike its strong inhibitory effect on the −126RARβCAT reporter (Fig. 4a), nicotine did not suppress the activation by trans-RA on this promoter (Fig. 4b). Thus, sequence from −126 to −60 of the RARβ promoter is responsible for the inhibitory effect of nicotine.
TR3 inhibits trans-RA-induced RARβ promoter activity by antagonizing COUP-TF activity
The above observation that the sequence required for the nicotine effect contains the COUP-TF-RE suggested that nicotine might modulate RARβ promoter activity through COUP-TF that is known to enhance transactivation of the RARβ promoter by trans-RA.39 Because nicotine could induce expression of TR3 (Fig. 3), which is known to physically interact with COUP-TF,24 we next determined whether TR3 could inhibit the transactivation function of COUP-TF on the RARβ promoter. Cotransfection of COUP-TF expression vector with the RARα expression vector synergistically induced −126RARβCAT promoter activity in the presence of trans-RA (Fig. 5a), consistent with our previous observation.39 However, when TR3 expression vector was cotransfected, the activation of the RARβ promoter by COUP-TF and RARα was inhibited (Fig. 5a). The inhibitory effect of TR3 was due to its inhibition of COUP-TF but not RARα activity, since RARβ promoter activity induced by RARα alone was not affected (Fig. 5a). In addition, the inhibitory effect of TR3 was COUP-TF-RE dependent because cotransfection of TR3 did not inhibit the activity of the −60RARβCAT reporter that lacks the COUP-TF-RE (Fig. 5b). To determine whether TR3 protein could inhibit binding of COUP-TF to the COUP-TF-RE, COUP-TF protein was mixed with or without TR3 protein and analyzed for its binding to the COUP-TF-RE by the gel shift assay. COUP-TF bound strongly to the COUP-TF-RE (Fig. 6), as observed before.39 However, when COUP-TF protein was pre-incubated with TR3, the binding was significantly reduced in a TR3 protein concentration-dependent manner, while pre-incubation of COUP-TF with control protein did not show any effect. Together, these data demonstrate that nicotine may inhibit RARβ expression through its induction of TR3 expression, which in turn inhibits binding and activation of the COUP-TF-RE by COUP-TF.
Inhibition of TR3 expression abolishes nicotine effects
We previously obtained 2 stable clones expressing TR3 anti-sense RNA in H460 cells.60 To further determine the involvement of TR3 expression in nicotine-induced suppression of trans-RA effect, we examined whether nicotine could inhibit trans-RA-induced growth inhibition in H460 cells stably expressing TR3 anti-sense RNA. Treatment of 2 TR3 anti-sense stable clones (H460/A-TR3 #14 and H460/A-TR3 #15) and a control stable clone expressing the empty vector (H460/vector)60 with trans-RA inhibited their growth (Fig. 7). The growth inhibitory effect of trans-RA in the TR3 anti-sense stable clones (about 70% inhibition) was stronger than in the control stable clone (about 60% inhibition) (Fig. 7), suggesting that inhibition of TR3 expression enhances the sensitivity of cells to trans-RA effect. Interestingly, growth inhibition by trans-RA was only slightly affected by nicotine (10%) in the TR3 anti-sense stable clones, whereas nicotine significantly reversed growth inhibition by trans-RA in the control stable clone. These data indicate that expression of TR3 is required for nicotine to prevent trans-RA activity.
Nicotine fails to abrogate the growth inhibition by an RXR-selective retinoid
Recent studies have demonstrated that several RXR-selective retinoids are more effective than classical RAR-selective retinoids in inhibiting cancer cell growth in vitro23, 26 and in animal.64, 65 Therefore, we examined the effect of nicotine on an RXR-selective retinoid SR11237 that is known to inhibit H460 cell growth.26 As shown in Figure 8, SR11237 strongly inhibited H460 cell growth, as was observed for trans-RA (Fig. 1). However, nicotine did not prevent SR11237-induced growth inhibition, in contrast to its activity against trans-RA (Fig. 1). Thus, nicotine preferentially suppresses the growth inhibitory effect of RAR-selective retinoid trans-RA but not RXR-selective retinoid SR11237 in H460 cells.
RXR-selective retinoid SR11237 activates the RARβ promoter via TR3/RXR heterodimers
Induction of TR3 mediates the inhibitory effect of nicotine on trans-RA-induced growth inhibition and RARβ expression (Fig. 5). The observation that nicotine was unable to interfere with RXR-selective retinoid activity suggests that RXR-selective retinoid may induce growth inhibition and RARβ expression through a mechanism that is different from the 1 utilized by trans-RA and that the mechanism is not negatively regulated by nicotine-induced TR3 expression. We, therefore, studied the effect of TR3 induction on −126RARβ promoter activity in the absence or presence of SR11237 by transient transfection assay in CV-1 cells (Fig. 9). Cotransfection of RXR expression vector slightly induced RARβ promoter activity when cells were treated with SR11237. The RXR-induced RARβ promoter activity was not enhanced when COUP-TF was cotransfected, indicating that COUP-TF does not act to enhance RXR activity on the RARβ promoter. Because nicotine inhibited trans-RA activity by interfering with COUP-TF function in the RARβ promoter (Figs. 5, 6), the observation that COUP-TF was not required for RXR to activate the RARβ promoter explains the inability of nicotine to inhibit RXR ligand activity. Interestingly, when TR3 was cotransfected with RXR, the RARβ promoter activity was strongly induced by SR11237. This data is consistent with our previous observation that the TR3/RXR heterodimer can bind to and activate the βRARE in response to an RXR-selective ligand.23 Thus, nicotine, by inducing TR3 expression (Fig. 3), may enhance the ability of RXR-selective retinoids to induce RARβ expression through the TR3/RXR heterodimer. We also evaluated the effect of COUP-TF on TR3/RXR activity. Unlike its enhancing effect on trans-RA-induced RARα activity (Fig. 5a), cotransfection of COUP-TF inhibited SR11237-induced TR3/RXR activity when low concentrations (20 ng and 50 ng) of TR3 were used. However, the inhibitory effect of COUP-TF was reversed when high concentration of TR3 (100 ng) was used, indicating that overexpression of TR3 can overcome the inhibitory effect of COUP-TF on RXR ligand activity.
Recent studies have demonstrated lack of efficacy of vitamin A, its precursor β-carotene and its 13-cis isomer in preventing lung cancer in cigarette smokers.9, 10 These results suggest that components in tobacco may modulate retinoid activity. In an attempt to understand the mechanism by which cigarette smokers exhibit retinoid resistance, we investigated the effect of nicotine on retinoid activity and report here that nicotine strongly suppresses growth inhibition by trans-RA in lung cancer cells (Fig. 1). Our observation suggests that abrogation of retinoid activity by nicotine may contribute to the impaired effect of retinoids in cigarette smokers.
In studying how nicotine suppresses growth inhibition by trans-RA, we found that it strongly inhibited trans-RA-induced RARβ expression (Fig. 2). This observation is consistent with a recent study showing that lung tissue of ferrets exposed to cigarette smoke had significantly reduced levels of RARβ but not RARα and RARγ.28 The facts that RARβ plays a critical role in mediating growth inhibition by retinoids in many different types of cancer cells17–24, 62 and that up-regulation of RARβ is associated with a positive clinical response to retinoid in patients with premalignant oral lesions21 suggest that inhibition of RARβ expression may be an important mechanism by which nicotine antagonizes the anti-cancer activities of retinoids. RARβ is a potent AP-1 inhibitor29 and apoptosis inducer.20 Interestingly, lung tissue of ferrets exposed to tobacco smoke displayed enhanced AP-1 expression and activity.28 Thus, tobacco smoke may modulate lung cancer cell growth and apoptosis through its inhibition of RARβ expression.
We previously demonstrated that increased TR3 expression is associated with retinoid resistance in lung cancer cells.24 We showed here that the antagonistic effect of nicotine on trans-RA activity in H460 lung cancer cells was associated with TR3 induction (Fig. 3). Furthermore, we observed that nicotine failed to suppress the growth inhibitory effect of trans-RA in H460 cells stably expressing TR3 anti-sense RNA (Fig. 7). These observations suggest that TR3 plays a crucial role in mediating nicotine effect on inhibiting trans-RA activity in lung cancer cells. Blockade of the growth inhibitory effect of trans-RA by TR3 is expected to release a negative growth regulatory mechanism, resulting in the growth advantage of cancer cells. This may in part explain the mitogenic effect of TR3 in response to different growth stimuli.37, 38, 40 Paradoxically, TR3 expression is also required for apoptosis induced by a variety of apoptotic stimuli in various cancer cells, including lung cancer cells.60, 61, 66–68 Our recent finding that TR3 translocates from the nucleus to mitochondria to induce apoptosis61 suggests that TR3 may confer a growth advantage to cancer cells through its action in the nucleus. Thus, it is likely that TR3 induced by nicotine may act in the nucleus to suppress the growth inhibitory effect of trans-RA. This is supported by our observations that nicotine strongly inhibited trans-RA-induced RARβ promoter activity (Fig. 4) and RARβ expression (Fig. 2) in H460 cells.
How TR3 confers retinoid resistance remains unknown. Our study demonstrated that the inhibitory effect of nicotine on trans-RA-induced RARβ promoter activity is mediated by the COUP-TF-RE in the RARβ promoter (Fig. 4). The COUP-TF-RE is required for positive regulation of the RARβ promoter activity by orphan receptor COUP-TF, which binds to the element to enhance the recruitment of receptor co-activators to the RARβ promoter.39 In our study, we found that TR3 strongly inhibited the binding of COUP-TF to the COUP-TF-RE (Fig. 6), probably due to a direct TR3-COUP-TF interaction, as has been demonstrated previously.24 In addition, we observed that expression of TR3 significantly inhibited the transactivation function of COUP-TF in the RARβ promoter (Fig. 5). These data, therefore, demonstrate that TR3 confers retinoid resistance by inhibiting RARβ expression through its interference of COUP-TF activities.
How nicotine induces TR3 expression is unknown. Expression of TR3 is rapidly induced by serum growth factors, TPA, Ca++ and apoptotic stimuli.37, 38, 40, 60, 61, 66–68 The TR3 promoter contains multiple AP-1-like and GC-rich elements as well as a CArG-like element, which are known to be activated by growth factors, protein kinase C, calcium signals and apoptotic stimuli.69–71 Previous observations that nicotine stimulates MAP kinase activity in a Ca++ dependent manner and that nicotine can regulate lung cancer cell proliferation via protein kinase C pathway48, 53, 57–59 suggest that activation of MAP kinase and/or protein kinase C may be responsible for induction of TR3 expression by nicotine. The effects of nicotine are mainly mediated by its receptor nAChR. The nAChR was highly expressed in H157 cells, while it was not detectable in H460 cells.45 Our observation that the suppressive effect of nicotine on growth inhibition by trans-RA in H460 cells was slightly less than that in H157 cells (Fig. 1) suggests that nicotine may exert its effect in part through the nAChR. However, the fact that nicotine effectively induced TR3 expression in nAChR-negative H460 cells (Fig. 3) also demonstrates that nicotine can act in a nAChR-independent mechanism. This is supported by previous findings that nicotine could reverse the growth inhibition by morphine in both H460 and H157 cell lines45 and activate MAPK by both nAChR-dependent and –independent mechanisms.59
We previously reported that RXR-selective retinoids could exert growth inhibitory effect in both lung cancer26 and breast cancer23 cells. We showed here that the growth inhibitory effect of RXR-selective retinoid SR11237 was not blocked by nicotine (Fig. 8). The βRARE present in the RARβ promoter plays a crucial role in regulating RARβ expression.30–32 Activation of the βRARE by trans-RA is mainly mediated by RAR/RXR heterodimers, in which RXR functions as a silent partner.11–13 However, the βRARE can also bind TR3/RXR heterodimers that are activated by RXR-selective ligands,23 suggesting an alternative pathway of inducing RARβ. Indeed, we found that cotransfection of TR3 with RXR expression vectors enhanced RARβ promoter activity in response to RXR-selective retinoid SR11237 (Fig. 9). Thus, the requirement of TR3 for the activation of the RARβ promoter by SR11237 may explain the inability of nicotine, which induces TR3 expression (Fig. 3), to suppress growth inhibition by SR11237. This finding suggests an important new approach of using RXR-selective retinoids to inhibit the growth of tobacco-associated cancer cells.
In summary, we provide evidence that nicotine, by inhibiting RARβ expression, can antagonize the negative regulation of growth by trans-RA. In addition, we demonstrate that the inhibition of RARβ expression by nicotine is due to its induction of TR3 that, in turn, inhibits the ability of COUP-TF to transactivate the RARβ promoter. Moreover, we show that nicotine fails to suppress the growth inhibitory effect of an RXR-selective retinoid. Thus, RXR-selective retinoids may be more effective than classical RAR ligands in preventing and treating tobacco-associated cancers.