Regulation of IKB Kinase E Expression by the Androgen Receptor and the Nuclear Factor-KB Transcription Factor in Prostate Cancer

Although several genes have been associated with prostate cancer progression, it is clear that we are far from understanding all the molecular events implicated in the initiation and progression of the disease to a hormone-refractory state. The androgen receptor is a central player in the initiation and proliferation of prostate cancer and its response to hormone therapy. Nuclear factor-KB has important proliferative and antiapoptotic activities that could contribute to the development and progression of cancer cells as well as resistance to therapy. In this study, we report that IKB kinase E (IKKE), which is controlled by nuclear factor-KB in human chondrocytes, is expressed in human prostate cancer cells. We show that IKKE gene expression is stimulated by tumor necrosis factor-A treatment in LNCaP cells and is inhibited by transfection of a dominant-negative form of IKBA, which prevents the nuclear translocation of p65. Furthermore, we found that tumor necrosis factor-A–induced IKKE expression is inhibited by an androgen analogue (R1881) in androgen-sensitive prostate cancer cells and that this inhibition correlates with the modulation of IKBA expression by R1881. We also noted constitutive IKKE expression in androgen-independent PC-3 and DU145 cells. To our knowledge, this is the first report of an IKB kinase family member whose expression is modulated by androgen and deregulated in androgen receptor–negative cells. (Mol Cancer Res 2007;5(1):87–94) Introduction Prostate cancer is the most common malignant disease among men in the Western world. The mainstay for prostate cancer control is radical surgery or radiotherapy for tumors confined to the prostate, whereas hormone therapy is commonly used alone or in combination with other treatments in advanced or high-risk prostate cancer. Eventually, prostate cancer stops responding to hormone therapy, yielding aggressive malignancies described as androgen independent or hormone refractory (1-3). Most men with hormone-refractory prostate cancer will die from their disease within 1 to 2 years (4). Understanding the biological mechanisms involved in prostate inflammation, androgen-independent growth, tumor progression, and metastasis has emerged as fundamental issues in prostate cancer research. It is known that members of the Rel/nuclear factor-nB (NF-nB) family play an important role in the development and progression of several human malignancies. NF-nB gene products have also been shown to have important proliferative and antiapoptotic activities that could contribute to the development, progression, and resistance to therapy of tumor cells (5, 6). Previous studies have observed high activity and nuclear translocation of NF-nB in prostate cancer cells (7-13) and found that NF-nB nuclear localization was strongly predictive of recurrence in patients following radical prostatectomy (14, 15). Prominent constitutive activation of NF-nB was also observed in the PC-3 and DU145 prostate cancer cell lines lacking androgen receptor expression, whereas only low NF-nB activity was seen in the LNCaP androgen-sensitive cell line (16). Androgen receptor, which is a member of the steroid hormone receptor family of ligand-activated nuclear transcription factors, is central to the initiation and growth of prostate cancer and to its response to hormone therapy (17). The DNAbinding activity of NF-nB in CL2 cells, hormone refractor (HR) derivative of LNCaP cells, was found to be higher than in the parental cell line (9). These data suggest an antagonistic effect between androgen receptor and NF-nB activity and an inverse correlation between androgen receptor expression and constitutive NF-nB activity in prostate cancer cells. In fact, some suggest that constitutive activation of NF-nB may play a role in the progression of prostate cancer and contribute to prostate cancer cell survival following androgen withdrawal (7-13). In this regard, we and others have recently found that NF-nB nuclear localization/activity in primary prostate cancer tissues correlates with poor patient outcome and bone Received 5/19/06; revised 9/27/06; accepted 11/21/06. Grant support: Canadian Uro-Oncology/AstraZeneca award. J-S. Diallo is a recipient of Canderel and Marc Bourgie studentships. L. Lessard is a recipient of Canadian Prostate Cancer Research Initiative studentship from the Canadian Health Institutes of Research. F. Saad is the recipient of the Université de Montréal chair in Prostate Cancer Research. 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. Requests for reprints: Anne-Marie Mes-Masson, Centre de Recherche du Centre Hospitalier de l’Université de Montréal/Institut du cancer de Montréal, 1560 rue Sherbrooke Est (Y-4609), Montréal, Québec, Canada H2L 4M1. Phone: 514890-8000, ext. 25496; Fax: 514-412-7703. E-mail: anne-marie.mes-masson@ umontreal.ca Copyright D 2007 American Association for Cancer Research. doi:10.1158/1541-7786.MCR-06-0144 Mol Cancer Res 2007;5(1). January 2007 87 Research. on October 18, 2017. © 2007 American Association for Cancer mcr.aacrjournals.org Downloaded from metastasis (10-18). Furthermore, we have found an increased nuclear NF-nB localization in lymphocytes and malignant cells in prostate cancer metastases containing pelvic lymph nodes (18). We also observed nuclear localization of both canonical and noncanonical NF-nB subunits in prostate cancer tissues, which suggests that different NF-nB pathways may be activated in prostate cancer progression (19). The classic NF-nB transcription factor is a heterodimer composed of p50 and p65 (20). In unstimulated cells, NF-nB is sequestered in the cytoplasm through an interaction with InBa. On stimulation of cells by specific stimuli, such as tumor necrosis factor-a (TNF-a), InBa is phosphorylated on Ser and Ser by the cytoplasmic InB kinase (IKK) complex, which consists of the IKKa and IKKh kinases and the NF-nB essential modulator/IKKg regulatory protein (reviewed in ref. 21). Degradation of InBa via the ubiquitin-proteasome pathway (22, 23) allows a rapid but transient translocation of NF-nB to the nucleus, where it binds to nB consensus sites and interacts with coactivators to promote transcription (23, 24). Recently, two noncanonical homologues of IKKs [i.e., IKKq and TANK-binding kinase-1 (TBK-1)] have been identified as NF-nB activators (25). Both kinases can phosphorylate InBa but only on Ser. Neither IKKq nor TBK-1 can phosphorylate InBa on Ser, a phosphorylation site necessary for the degradation of InBa by the ubiquitination pathway (26, 27). NF-nB-dependent gene expression is impaired in embryonic fibroblasts from TBK-1-deficient mice, which die as a result of apoptotic liver degeneration (28). TBK-1 is ubiquitously expressed whereas IKKq is constitutively expressed in lymphoid cells and human fibroblast-like synoviocytes from rheumatoid arthritis patients, although inducible in other cell types (29, 30). It has been shown that the NF-nB p65 subunit is involved in the transcriptional regulation of the IKKq promoter in human chondrocytes (31). IKKq mRNA synthesis can be induced in many cell types in response to inflammatory cytokines (TNF-a and interleukin1h) and lipopolysaccharides, indicating that proinflammatory agent-mediated stimuli can modulate its expression (32, 33). Despite the importance of IKKq in the NF-nB pathway, little is known about IKKq expression in prostate cancer. In this study, we looked at IKKq expression in prostate cancer cell lines and how its expression varies in response to TNF-a and androgen. Results TNF-a Induction of IKKe Expression in the LNCaP Cell Line As NF-nB p65 subunit is involved in the transcriptional regulation of IKKq promoter in human chondrocytes and IKKq could be responsible for the activation of NF-nB, a component of prostate cancer progression, we investigated IKKq expression in relation to NF-nB nuclear translocation in the androgensensitive LNCaP cells. Under standard culture conditions, no endogenous IKKq expression was detectable by Western blot analysis in the cytosolic fraction of LNCaP cells (Fig. 1A). Similarly, no NF-nB p65 subunit expression was detectable in the nuclear fraction. The addition of TNF-a rapidly induced p65 nuclear translocation (Fig. 1B). Although increased endogenous IKKq expression was visible for at least 24 h, it was first detectable in the cytosolic fraction only after 4 h of TNF-a treatment (Fig. 1A). Hence, p65 translocation occurs before IKKq expression. In addition, IKKq expression did not increase p65 translocation (Fig. 1B). Interestingly, TBK-1 expression was constitutive and was not modulated by TNF-a stimulation. Correlation between IjBa and IKKe mRNA Expression Following stimulation of nearly confluent LNCaP cells with TNF-a, IKKe mRNA expression was quantified by real-time PCR. Similarly to what was observed by Western blot, we found that IKKe expression was dramatically increased by TNF-a treatment, 6-fold after 2 h and 8-fold after 4 h of TNF-a treatment, compared with mock-treated cells (Fig. 2A). High IKKe mRNA levels were still detected at 8 h after TNF-a treatment. We also measured IjBa mRNA expression in parallel as a means to follow the NF-nB transcriptional activity. A TNF-a treatment induced a rapid translocation of NF-nB (Fig. 1A) concomitant with an increase in IjBa mRNA transcription (Fig. 2B). InBa mRNA seems to increase at the same time as IKKe mRNA after TNF-a treatment, although InBa gene activation was less important compared with mocktreated cells (5-fold after 2 h). The increase in the expression of this gene was also maintained 8 h after TNF-a treatment. Implication of NF-jB p65 Subunit in the Activation of IKKe Expression in LNCaP Cell Line To test the implication of NF-nB in IKKq gene activation in the LNCaP cell line, we transiently transfected the pCMVInBadn construct in these cells. The InBadn is a dominantnegative construct of the NF-nB inhibitor, which cannot be phosphorylated and thereby inhibits the activation of NF-nB. As a control for this experiment, we used the pCMV-Neo plasmid. Transfection of InBadn dramatically blocked p65 nuclear translocation after TNF-a treatment (Fig. 3A). Eight hours (Fig. 3B) or 24 h (data not shown) after stimulation, IKKq expression was observed in the control cells, whereas no IKKq protein could be detected in the LNCaP pCMV-InBadn cells after TNF-a treatment (Fig. 3B). Degradation of endogenous InBa was shown as a control of TNF-a stimulation (Fig. 3B). The inhibition of NF-nB nuclear translocation had no effect on TBK-1 expression. Implication of Androgen Receptor in the Regulation of IKKe Expression in Prostate Cancer Cell Lines Given these observations suggesting that the NF-nB p65 subunit is involved in TNF-a–induced IKKq expression in androgen-sensitive LNCaP cell line and because of several studies have suggested an antagonistic relationship between androgen receptor and NF-nB, we chose to examine the effect of androgen receptor stimulation by androgen on IKKq expression. Other studies have suggested that androgen receptor stimulation by androgen can maintain InBa levels and consequently block NF-nB activity (34, 35). To determine if androgen receptor stimulation by androgen could control InBa expression levels, we used cycloheximide to block translation in LNCaP cells. Figure 4 shows that inhibition of Péant et al. Mol Cancer Res 2007;5(1). January 2007 88 Research. on October 18, 2017. © 2007 American Association for Cancer mcr.aacrjournals.org Downloaded from translation prevents an androgen analogue-induced (R1881) increase in InBa protein (Fig. 4, lanes 4 and 5). To determine whether androgen modulation of InBa could affect the induction of IKKq by TNF-a, we pretreated our cells for 2 h with R1881 before TNF-a stimulation. As LNCaP cells do not constitutively express IKKq, TNF-a stimulation was required to observe the effects of androgen on IKKq protein levels. We analyzed IKKq and InBa protein levels 8 h following TNF-a stimulation (Fig. 4). R1881 treatment alone did not induce IKKq expression (Fig. 4, lane 4) as opposed to a clear induction by TNF-a (Fig. 4, lane 2). A decrease in IKKq up-regulation by TNF-a, which correlated with an increase in InBa protein, was observed when cells were pretreated with R1881 for 2 h before TNF-a stimulation (Fig. 4, lane 2 versus Fig. 4, lane 6). Interestingly, 8 h following TNF-a R1881 costimulation, IKKq levels were equivalent to that observed after TNF-a treatment alone (Fig. 4, lane 8 versus Fig. 4, lane 2) despite a modulation of InBa levels. IKKe Expression in Hormone-Resistant Versus HormoneSensitive Prostate Cancer Cells Next, we determined the level of IKKq expression in PC-3 and DU145 androgen-independent cells compared with 22Rv1 and LNCaP androgen-sensitive cells. We observed high constitutive IKKq expression in PC-3 and DU145 cells that was not affected by TNF-a treatment (Fig. 5A). Moreover, IKKq expression in unstimulated DU145 and PC-3 cells was higher than in TNF-a–stimulated LNCaP and 22Rv1 cells. TNF-a treatment also induced a weak IKKq expression in the 22Rv1 cell line, which was only detectable after long exposures (Fig. 5B). Differences observed between IKKq expression in androgen-sensitive and androgen-independent cell lines were not correlated with variations in NF-nB activity (Fig. 5C). In fact, PC-3 and 22Rv1 NF-nB activity was only weakly affected by TNF-a treatment (1.4and 1.3-fold, respectively) compared with DU145 and LNCaP cells (3and 6.8-fold, respectively). After TNF-a stimulation, NF-nB activity in LNCaP cells dramatically increased and was higher than in all other cell lines studied. Comparable NF-nB activities were measured for DU145 and LNCaP cells in unstimulated conditions (Fig. 5C). Although NF-nB activity was stimulated by TNF-a in DU145 cells, no increase in IKKq expression levels was observed. Finally, TBK-1 expression was not affected by TNF-a treatment in any of the cell lines studied. However, TBK-1 expression level differed among the cell lines and did not correlate with IKKq expression or NF-nB activity levels. Discussion In the present study, we report for the first time NF-nB regulation of IKKq expression in androgen-dependent prostate cancer cell lines. It is known that IKKq is predominantly expressed in cells and tissues of the immune system, such as peripheral blood leukocytes, thymus and spleen (26-32), and murine embryonic fibroblasts (33). Recently, constitutive IKKq expression was also detected in human chondrocytes from cartilage and in the C28/I2 human chondrocyte cell line (31). Our data show differential expression of IKKq in the prostate cancer cell lines tested, and expression levels seem to be linked to the androgen receptor status of the cells. In androgensensitive cell lines, such as LNCaP and 22Rv1, IKKq expression can be induced in response to TNF-a stimulation. In androgen-independent PC-3 and DU145 cell lines, IKKq is constitutively expressed. These observations contrast with results presented in a recent report that showed an equal and constitutive IKKq expression in PC-3 and LNCaP cells (36). TBK-1 has been reported to be ubiquitously expressed (26-32). Although we observed some variations in TBK-1 levels between prostate cancer cell lines tested, neither TNF-a nor R1881 treatment had an effect on its expression. Using LNCaP cells, we found that IKKq mRNA and protein synthesis correlate with p65 nuclear translocation and InBa mRNA synthesis in response to TNF-a treatment. These results suggest a role for NF-nB in the regulation of IKKq expression, in contrast to previous findings in murine embryonic fibroblasts Figure 1. Effect of TNF-a on IKKq protein expression in LNCaP cells. Cells were grown to 80% confluency and treated with TNF-a (10 ng/mL) for the indicated times. Equal amounts of nuclear and cytosolic proteins were resolved by SDS-PAGE, transferred onto nitrocellulose membranes, and probed with appropriate antibodies. A. Western blot analysis of IKKq, TBK-1, p100/p52, and p65 levels in cytosolic extracts (30 Ag of cytosolic proteins). Equal loading was tested with anti-actin and anti-a-tubulin antibodies. B. Western blot analysis of nuclear fractions for p65 and p100/p52 nuclear translocation (10 Ag of nuclear proteins). Equal loading was tested with an anti-actin antibody, and cytosolic contamination was controlled with an anti-a-tubulin antibody. NF-nB Regulation of IKKq Expression in Prostate Cancer Mol Cancer Res 2007;5(1). January 2007 89 Research. on October 18, 2017. © 2007 American Association for Cancer mcr.aacrjournals.org Downloaded from (33). Moreover, IKKq is thought to have a significant kinase activity when expressed and isolated from unstimulated cells (27-38). In addition, another study showed that expression of IKKq is sufficient to induce phosphorylation, nuclear translocation, and DNA binding of IRF-3 and IRF-7 (39). Moreover, a recent study showed a correlation between IKKq expression and the phosphorylation of p65 on Ser in several non– prostate cancer cell lines as opposed to prostate cancer cell lines where this correlation was not observed (36). Consequently, we were expecting to observe NF-nB activation and nuclear translocation during IKKq expression. In this study, we failed to find any p65 nuclear accumulation following an increase in IKKq synthesis after TNF-a treatment of LNCaP cells. In fact, we observed a decrease in nuclear p65 levels 6 h after TNF-a stimulation, whereas IKKq protein levels increased. These results suggest that the TBK-1/IKKq complex does not activate p65 translocation and canonical NF-nB activity in prostate cancer cells. The role of NF-nB protein in the induction of IKKq expression was verified by transfecting a dominant-negative form of InBa, the inhibitor of p65/p50 NF-nB dimer, in LNCaP cells. pCMV-InBadn inhibited the induction of p65 nuclear translocation by TNF-a treatment and blocked IKKq expression. These observations correlate well with a previous study, which showed that the interaction between NF-nB p65 protein and 833/ 847 nB sites on the IKKq promoter occurs and this interaction leads to the activation of the IKKe gene (31). Because NF-nB is thought to be constitutively activated in androgen-independent prostate cancer cells (7-9) and IKKq is a target gene of NF-nB, it is tempting to speculate that constitutive expression of IKKq in PC-3 and DU145 cells is the consequence of elevated NF-nB activity. However, this hypothesis is too simplistic as we observed equivalent NF-nB activity in LNCaP and DU145 cells in unstimulated conditions (Fig. 5C), although DU145 cells constitutively express IKKq whereas this protein is not detectable in LNCaP cells under these conditions. Likewise, the small difference in NF-nB activity observed between PC-3 and LNCaP cells could not completely explain the difference in IKKq expression in these cell lines. These observations lead us to investigate a role for the androgen receptor in the modulation of IKKq expression. Cross-modulation, transcriptional interference, and physical interaction between androgen receptor and NF-nB have been described (40, 41). This interaction could explain the downregulation in IKKq expression after androgen receptor stimulation by the androgen analogue R1881 in LNCaP cells. One study showed that direct protein-protein inhibition of NF-nB by androgen receptor does not occur in the nucleus and incubation of LNCaP cells with dihydrotestosterone before NF-nB activation inhibited NF-nB-DNA complex formation (35). These authors suggested that NF-nB may be sequestered Figure 2. Effect of TNF-a on InBa and IKKq mRNA expression in LNCaP cells. Cells were grown to 80% confluency and treated with TNF-a (10 ng/mL) or mock treated for 8 h. Each mRNA sample (2 Ag) was used to synthesize cDNAs, which were used for quantitative real-time PCR analysis. Each sample was tested in duplicate. C t values were determined and used to calculate relative mRNA expression (sample/actinB ratio). Fold induction was calculated relative to untreated cells for each gene. Columns, mean of two independent experiments; bars, SE. Figure 3. Effect of p65 nuclear translocation inhibition on IKKq expression. Cells were grown to 80% confluency and transfected with the pCMV-InBadn vector or the pCMV-Neo vector as a control. Thirty-six hours later, cells were treated with TNF-a (10 ng/mL) for either 30 min or 8 h. Equal amounts of nuclear and cytosolic proteins were migrated by SDS-PAGE, transferred onto nitrocellulose membranes, and probed with appropriate antibodies. A. Membrane containing nuclear extracts was immunoblotted with anti-p65 antibody. Equal loading was tested with an anti-actin antibody, and cytosolic contamination was controlled with an anti-a-tubulin antibody. B. Membrane containing cytosolic extracts was successively immunoblotted (after stripping) with anti-IKKq/TBK-1 and anti-p65. Anti-InBa antibody recognized endogenous InBa but failed to detect InBadn. Equal loading was tested with an anti-a-actin antibody. Péant et al. Mol Cancer Res 2007;5(1). January 2007 90 Research. on October 18, 2017. © 2007 American Association for Cancer mcr.aacrjournals.org Downloaded from in the cytoplasm after androgen receptor stimulation. Other studies have suggested that androgen receptor stimulation by androgen can maintain InBa levels through induction of protein expression or the inhibition of InBa phosphorylation (34, 35). Our finding that R1881 treatment induced an increase in InBa levels could explain androgen receptor control on NF-nB activity and, indirectly, on IKKq expression by NF-nB sequestration in the cytosol. Moreover, we observed an accumulation of InBa in cells treated with R1881 compared with control cells or cells pretreated with cycloheximide before R1881 stimulation. These observations imply that InBa is likely produced de novo after R1881 stimulation. In the same experiment, we also failed to note any protective effect of R1881 pretreatment on TNF-a–induced degradation of InBa. These observations favor a role for the androgen receptor on InBa protein expression rather than on InBa degradation. Further studies need to be conducted to clarify the androgendependent control of androgen receptor on IKKq expression. In summary, we have identified and characterized NF-nB as a regulator of the IKKe gene in human prostate cancer cells. We have also shown that androgens modulate IKKq regulation in androgen receptor–positive hormone-sensitive prostate cancer cells and that this is potentially mediated by InBa synthesis in response to androgen stimulation (Fig. 6). Moreover, we present, for the first time, the deregulated expression of an InB kinase member in androgen-independent prostate cancer cells. The study of this phenomenon and the interactions between NF-nB, androgen receptor, and IKKq will certainly be critical in improving our understanding of the development of hormone-independent cells and metastatic prostate disease. Materials and Methods Cell Lines and Cell Culture Androgen-independent PC-3 and DU145 cells and androgen-sensitive 22Rv1 and LNCaP cells were purchased from the American Type Culture Collection (ATCC CRL-1435, ATCC HTB-81, ATCC CRL-2505, and ATCC CRL-1740, respectively; Manassas, VA). Cells were routinely grown in RPMI 1640 (Wisent, Inc., St-Bruno, Quebec, Canada) supplemented with 100 Ag/mL gentamicin, 0.25 Ag/mL amphotericin B (Invitrogen, Paisley, United Kingdom), and 10% FCS. Human recombinant TNF-a was purchased from Roche Applied Science (Indianapolis, IN). Androgen analogue R1881 (methyltrienolone) was obtained from Perkin-Elmer (Wellesley, MA), and cycloheximide was from Supelco (Bellefonte, PA). Transfection and Luciferase Assays Transfection of LNCaP cells with the plasmid pCMVInBadn (40 Ag/plate), obtained from Clontech (Mountain View, CA), was done in 150-mm tissue culture plates when cells reached 80% to 90% confluence using LipofectAMINE reagent (Invitrogen) according to the manufacturer’s instructions. For luciferase assays, prostate cancer cells were plated on 48-well plates (3 10 per well) and incubated with RPMI 1640 containing 10% FCS for 24 h. Transfections were done Figure 4. Inhibition of IKKq expression by androgen receptor (AR ) stimulation. Western blot analysis of IKKq/TBK-1, androgen receptor, and InBa protein levels after treatments by R1881 and TNF-a. Cells were grown to 80% confluency and pretreated 4 h with cycloheximide (50 Ag/mL). R1881 (10 nmol/L) was added to the medium 2 h before TNF-a stimulation (T-2 ) or at the same time (T0 ). Cells were treated with TNF-a (10 ng/mL) for 8 h. Equal amounts of proteins from whole-cell extracts were resolved by SDS-PAGE, transferred onto nitrocellulose membranes, and probed with appropriate antibodies. Equal loading was tested with an anti-actin antibody. Figure 5. IKKq expression in androgen-independent prostate cancer cells. A. IKKq expression in prostate cancer cell lines was measured following treatment with TNF-a (10 ng/mL) for 8 h. Western blot analysis of IKKq and TBK-1 in total extracts. Equal amounts of protein from the four different cell lines were loaded on 7.5% SDS-PAGE. B. Longer exposure of A. C. NF-nB transcriptional activity was measured by luciferase assay following treatment with TNF-a (10 ng/mL) for 8 h. Cells were grown to 80% confluency and cotransfected with pCMV-Renilla and 3nB-conAFirefly. Thiry-six hours later, cells were treated during 8 h with TNF-a (10 ng/mL) and subsequently assayed for luciferase activity. Transfection efficiency was normalized to that of Renilla luciferase. Fold induction of NF-nB activity. NF-nB Regulation of IKKq Expression in Prostate Cancer Mol Cancer Res 2007;5(1). January 2007 91 Research. on October 18, 2017. © 2007 American Association for Cancer mcr.aacrjournals.org Downloaded from using LipofectAMINE reagent. NF-nB activity was measured using the 3nB-conA-Firefly plasmid (400 ng/well of 3nBconA-Firefly; ref. 16). The total amount of plasmid DNA was adjusted to 500 ng/well by addition of pCMV-Renilla plasmid (Promega, Madison, WI), which codes for the Renilla luciferase gene under the control of the cytomegalovirus promoter. After 24 h, medium was replaced with RPMI 1640 plus 10% FCS containing TNF-a (10 ng/mL). Cells were collected after 24 h of incubation using the lysis buffer provided in the luciferase kit (Promega). Luciferase activities were measured using the DualLuciferase Assay System (Promega) with the aid of a multiplate luminometer (BMG Labtechnologies, Inc., Durham, NC). Luciferase activities were normalized using the Renilla activity of the samples as measured by the multiplate luminometer. All transfection experiments were carried out in duplicate and repeated at least thrice.