Estrogen Modulates NFκB Signaling by Enhancing IκBα Levels and Blocking p65 Binding at the Promoters of Inflammatory Genes via Estrogen Receptor-β

Background NFκB signaling is critical for expression of genes involved in the vascular injury response. We have shown that estrogen (17β-estradiol, E2) inhibits expression of these genes in an estrogen receptor (ER)-dependent manner in injured rat carotid arteries and in tumor necrosis factor (TNF)-α treated rat aortic smooth muscle cells (RASMCs). This study tested whether E2 inhibits NFκB signaling in RASMCs and defined the mechanisms. Methodology/Principal Findings TNF-α treated RASMCs demonstrated rapid degradation of IκBα (10–30 min), followed by dramatic increases in IκBα mRNA and protein synthesis (40–60 min). E2 enhanced TNF-α induced IκBα synthesis without affecting IκBα degradation. Chromatin immunoprecipitation (ChIP) assays revealed that E2 pretreatment both enhanced TNF-α induced binding of NFκB p65 to the IκBα promoter and suppressed TNF-α induced binding of NFκB p65 to and reduced the levels of acetylated histone 3 at promoters of monocyte chemotactic protein (MCP)-1 and cytokine-induced neutrophil chemoattractant (CINC)-2β genes. ChIP analyses also demonstrated that ERβ can be recruited to the promoters of MCP-1 and CINC-2β during co-treatment with TNF-α and E2. Conclusions These data demonstrate that E2 inhibits inflammation in RASMCs by two distinct mechanisms: promoting new synthesis of IκBα, thus accelerating a negative feedback loop in NFκB signaling, and directly inhibiting binding of NFκB to the promoters of inflammatory genes. This first demonstration of multifaceted modulation of NFκB signaling by E2 may represent a novel mechanism by which E2 protects the vasculature against inflammatory injury.

In the setting of vascular injury, TNF-a activates NFkB, a transcription factor that mediates the immediate-early inflammatory response [17][18][19][20]. Although numerous NFkB proteins exist, the most common NFkB heterodimer contains p65 and p50. Each of the NFkB proteins contains an N-terminal Rel homology domain (RHD), which is important for DNA binding, dimerization, inhibitor association and nuclear localization [21,22]. In most cells, NFkB is bound to and inhibited by IkBa, which reduces the ability of NFkB to bind DNA [23]. In response to TNF-a, interleukin-1b (IL-1b), or other stimuli, the inhibitor of NFkB kinase (IKK) complex is activated and phosphorylates IkBa, which targets it for degradation by the proteasome. This effectively liberates NFkB, which then translocates into the nucleus where it binds to cognate DNA response elements found within the promoters of target genes to induce their expression. NFkB activation is critical for the expression of a variety of genes, including IkBa and those involved in vascular inflammation, e.g. cytokine-induced neutrophil chemoattractant (CINC)-2b and monocyte chemotactic protein (MCP)-1 [24][25][26]. Previously, we have shown that expression of MCP-1 and CINC-2b is inhibited by E2 in an ER dependent manner in balloon injured carotid arteries of rats and in RASMCs in vitro [9,16]. However, at present, it is not clear exactly how E2 inhibits NFkB mediated expression of these genes in SMCs. The current study tested directly the hypothesis that E2, in an ER dependent manner, modulates the inflammatory response to TNF-a stimulation in isolated RASMCs in vitro by interfering with NFkB signaling and defined the precise sites of molecular merging of E2 and NFkB signaling cascades that are responsible for this effect.

E2 does not Prevent IkBa Phosphorylation and
Degradation, but does Enhance IkBa mRNA and Protein Levels in TNF-a treated RASMCs Consistent with previous observations that IkBa processing is a target for E2/ER signaling [27][28][29], we tested the hypothesis that E2 inhibits cytokine-induced IkBa phosphorylation and degradation in RASMCs, thus attenuating NFkB signaling. Quiescent RASMCs were incubated with E2 or vehicle for 24 hrs, followed by TNF-a for 10, 20, 30, 40, 50 and 60 mins. Total protein was extracted and the levels of total and phospho-IkBa were assessed using Western blot analyses. RASMCs treated with TNF-a for 10 min demonstrated increased levels of phospho-IkBa, with rapid degradation of IkBa between 10-30 min ( Figure 1A), followed by a dramatic recovery at 60 min. Levels of phospho-IkBa were not reduced by pretreatment with E2 ( Figure 1). Although IkBa was degraded in the presence of E2 and TNF-a between 10-30 min, the total levels of IkBa were elevated compared to those in the presence of TNF-a alone between 30-60 min ( Figure 1A). These results were analyzed by densitometry and are presented in Figure 1B. Because E2 does not prevent TNF-a induced IkBa degradation, these data suggest that E2 may attenuate NFkB signaling by inducing new IkBa mRNA synthesis.
To evaluate the effects of E2 on TNF-a induced IkBa mRNA levels, RASMCs were treated as described above and IkBa levels were analyzed using real-time RT-PCR analyses. The levels of IkBa mRNA were increased by TNF-a stimulation between 30-60 min (Figure 2), and were further enhanced by E2. These findings suggest that E2 can reduce NFkB activity by increasing the expression of IkBa mRNA and protein.

ERb Activation Enhances IkBa mRNA Expression and Restoration of IkBa Protein in TNF-a treated RASMCs
We have previously shown that in vitro, E2 inhibits TNFa induced inflammatory mediator expression in RASMCs in an ERb-dependent manner [16]. To test whether the effects of E2 on TNF-a-induced IkBa expression are also mediated by ERb, RASMCS were pretreated with the selective ERb agonist diarylpropiolnitrile (DPN), the selective ERa antagonist methylpiperidinopyrazole (MPP) alone or in combination with E2, E2 alone or vehicle for 24 hrs, followed by TNF-a for an additional 45 or 60 min and subjected to Western blot analysis for IkBa protein and real-time RT-PCR analysis for IkBa mRNA, respectively. These time points were chosen because they capture the recovery phase of IkBa resynthesis following TNF-a induced phosphorylation and degradation (Figures 1 and 2).
At 45 min post TNF-a treatment, IkBa protein levels were significantly lower in TNF-a treated RASMCs than in vehicle-treated control cells ( Figure 3A, lane 2), indicating that IkBa protein expression had not completely recovered to vehicle control levels (lane 1) at this time point Pretreatment with E2 or DPN for 24 hr significantly accelerated the recovery of IkBa protein levels in TNF-a-treated cells (lanes 3, 4). In contrast, pretreatment with the ER a agonist propylpyrazole triol (PPT) did not alter the inhibitory effect of TNF-a on IkBa protein levels ( Figure 3B, lane 4). In addition, the stimulatory effect of E2 on IkBa protein levels in TNF-a-treated cells was abolished by pretreatment with tetrahydrochrysene-R,R,-enantiomer (R,R-THC, an agonist on ERa and an antagonist on ERb) 1 hr prior of E2 ( Figure 3C

E2, Through ERb, Recruits NFkB p65 to the IkBa Promoter
To understand the molecular mechanisms by which E2 might enhance IkBa mRNA synthesis, Chromatin Immunoprecipitation (ChIP) analyses were performed. Quiescent cells were pretreated with E2, DPN or vehicle for 24 hrs and then treated with TNFa for 1 hr. In vehicle treated cells, ChIP assays revealed that NFkB p65 was not detected at the IkBa promoter ( Figure 5, lane 1). Treatment with TNF-a, E2 or DPN alone (lanes 2, 3 and 5) resulted in recruitment of p65 (4 to 9 fold) to the IkBa promoter compared to vehicle control. When cells were pretreated with E2 or DPN and then challenged with TNF-a (lanes 4 and 6), the levels of p65 at the IkBa promoter were not altered significantly in response to additional TNF-a compared to the levels in the presence of E2 or DPN alone. In addition, pretreatment with the ERb antagonist R,R-THC blocked E2 induced recruitment of p65 to the IkBa promoter in TNF-a-treated cells (lane 8), indicating ERb dependency of the effect.
ChIP analyses with anti-ERb antibody were performed to test whether ERb was recruited to the IkBa promoter. In the vehicle treated cells ( Figure 5B, lane 1), ERb was detectable at the IkBa promoter. TNF-a treatment did not alter the binding of ERb at the IkBa promoter (lane 2). In the E2 alone or E2+TNF-a treated cells, ERb level was increased 2-fold at the IkBa promoter (lanes 3 and 4). E2 induced-recruitment of ERb to the IkBa promoter was abolished by pretreatment with the ERb antagonist R,R-THC (lane 5). In contrast, ERa was not detected at the IkBa promoter in response to E2 alone or coincides with the increased level of p65 at the IkBa promoter in the presence of E2 or E2+TNF-a (Data not shown).
Histones are acetylated at promoters that are undergoing active transcription [30]. The binding of acetylated histone at the promoter of a gene indicates that the gene is actively transcribing. ChIP assays determined that the levels of AcH4 at the IkBa promoter increased 5-fold in response to TNF-a treatment compared to vehicle ( Figure 5C, lanes 1 and 2). E2 alone had no effect on binding of AcH4 to the IkBa promoter (lane 3). In the presence of E2+TNF-a, the levels of AcH4 at the IkBa promoter increased significantly (7-fold) compared to vehicle treatment (lane 4). The level of AcH4 at the IkBa promoter in the presence of E2+TNF-a was higher (about 40%) than the level in the presence of TNF-a alone, but the difference was not statistically significant. In cells pretreated of R,R-THC prior to E2+TNF-a (lane 6), the level of AcH4 at the IkBa promoter was not significantly different from the levels in E2+TNF-a treated cells. Together, these data suggest that treatment with E2, combined with TNF-a, significantly enhanced the transcriptional activity of the IkBa gene through an effect on ERb.  ChIP assays determined that NFkB p65 was present at the MCP-1 and CINC-2b promoters at low levels in the absence of TNF-a or E2 ( Figure 6A) and that these levels were not affected by addition of E2 alone. At 1 hr post TNF-a treatment, the levels of NFkB p65 at these promoters were increased (14-and 21-fold), and these levels were reduced nearly to the control levels in the presence of pretreatment with E2, suggesting that E2 inhibits the ability of NFkB p65 to bind the promoters of these genes.
In the absence of TNF-a or E2 ( Figure 6B), or in the presence of E2 alone or TNF-a alone, ERb was barely detected at the MCP-1 or CINC-2b promoters. However, in the presence of E2+TNF-a, ERb was detected at the MCP-1 and CINC-2b promoters. These data suggest that in the presence of E2+TNF-a, ERb is recruited to these promoters and that the presence of ERb coincides with the reduced levels of NFkB p65.
ChIP assays determined that the MCP-1 and CINC-2b promoters harbored moderate levels of AcH3 in the absence of any stimuli ( Figure 6C,), and that these levels were reduced in the presence of E2 alone. TNF-a treatment increased the levels of AcH3 at both promoters (5 and 3 fold, respectively) and these levels were diminished in the presence of E2, indicating that these genes have reduced transcriptional activity in the presence of E2. Together, these data indicate that these genes are inhibited by E2 in both basal and induced states. In the basal state, E2 reduces levels of AcH3. In the induced (by TNF-a) state, E2 reduces the levels of p65 and AcH3.

E2, Through ERb, Inhibits MCP-1 and CINC-2b mRNA Expression in TNF-a treated RASMCs
To test whether E2 inhibits TNF-a-induced MCP-1 and CINC-2b mRNA expression and to assess the ER subtype dependence of the E2 effect, RASMCS were pretreated with E2, the selective ERb agonist DPN, the selective ERa antagonist MPP alone or the selective ERb antagonist R,R-THC alone in combination with E2, or vehicle for 1 hr and subjected to real time RT-PCR analysis for MCP-1 and CINC-2b mRNA, respectively. Quantitative real time RT-PCR analysis showed that TNF-a stimulated expression of MCP-1 and CINC-2b significantly compared to the vehicle control ( Figure 7). Pretreatment with E2 or DPN significantly inhibited expression of MCP-1 and CINC-2b in cells treated with TNF-a. In contrast, R,R-THC, but not MPP antagonized the inhibitory effects of E2 on MCP-1 and CINC-2b mRNA expression in TNF-a-treated cells. E2, DPN, MPP or R,R-THC alone did not alter MCP-1 and CINC-2b mRNA in RASMCs in the absence of TNF-a treatment. Together, findings suggest that the E2 mediated anti-inflammatory effect in TNF-a treated RASMCs is mediated by ERb, and not ERa.

Discussion
The multifaceted crosstalk between NFkB signaling and the ERs has been well documented [31]. In numerous models, E2 and ERs have been shown to increase levels of IkBa and reduce levels of phosphorylated IkBa [28,[32][33][34]. Moreover, both ERa and ERb reportedly inhibit NFkB activity in an E2 dependent manner in a variety of cell types [31,[35][36][37][38][39][40][41][42], and molecular studies have mapped the minimal domains of ERa necessary for these effects to the ligand binding domain (LBD), hinge domain and DNA binding domain (DBD) [43,44]. In vitro, ERa binds to NFkB p65, p50 and c-Rel [43,45]; ERb inhibits the DNA binding ability of NFkB p50, c-Rel and NFkB p65/p50 dimers [36,43,46], and both ERs can prevent NFkB from binding to the IL-6 promoter [43,46,47]. However, at present, there is a paucity of data to clarify the role of E2 and/or ERs in regulating the activity of NFkB in vascular cells.
Previously, we demonstrated that isolated RASMCs express high levels of inflammatory mediators, including the neutrophiland monocyte-selective chemokines CINC-2b and MCP-1, when stimulated by TNF-a and that E2 inhibits this process and reduces the neutrophil chemotactic activity of media conditioned by TNFa treated RASMCs via an ERb-dependent mechanism [16]. Herein we extend our studies in order to elucidate the molecular mechanisms by which E2 and ERb negatively regulate the NFkB signaling pathway in RASMCs. Specifically, this study demonstrates for the first time the multifaceted effects of E2 in negatively modulating events in the NFkB pathway in a vascular cell type. We show that E2 neither inhibits the production of TNF-a by RASMCs (See Text S1 and Figure S1), nor blocks the nuclear translocation of NFkB p65 ( Figure S2). Further, we demonstrate that both ERa and ERb proteins are expressed in our RASMCs in an E2 and TNF-a independent manner ( Figure S3). We demonstrate that E2, via ERb, attenuates signaling through the NFkB signaling pathway via a novel bimodal mechanism. First, E2 selectively enhance NFkB p65 binding to the IkBa promoter in order to stimulate the expression of IkBa, a direct inhibitor of NFkB activation. Second, E2 reduces the ability of NFkB p65 to bind to the promoters of pro-inflammatory genes such as MCP-1 and CINC-2b, thereby inhibiting their transcriptional activity, indicated by the binding of AcH3 to the promoters, and mRNA expression. These findings support the intriguing hypothesis that E2, via ERb, selectively modulates the nuclear activity of NFkB p65 to ensure that NFkB signaling is dampened by heightened IkBa levels, as well as by reducing the binding of nuclear NFkB p65 to the promoters of genes that mediate the inflammatory response.
IkBa is the one of the best documented inhibitors and transcriptional targets of NFkB. Through its ability to interact with NFkB proteins, IkBa masks the DBD of NFkB in order to maintain NFkB inactive in the cytoplasm until such time that NFkB is activated. While NFkB is initially activated through proteasomal-mediated degradation of IkBa, NFkB signaling is ultimately terminated through NFkB mediated resynthesis of IkBa, which re-establishes the inactive cytoplasmic pool of NFkB/ IkBa complexes [48,49]. Studies of the murine IkBa promoter identified six NFkB and NFkB-like response elements that are highly conserved in sequence, orientation and position within the genomes of humans and pigs [48]. Although the IkBa promoter appears to be devoid of NFkB proteins in the basal state, the IkBa promoter is bound and activated by NFkB proteins within minutes of NFkB activation [50,51].
Our studies demonstrate that neither DPN nor E2 when administered alone stimulated IkBa mRNA expression in RASMCs despite substantial recruitment of NFkB p65 at the IkBa promoter. Furthermore, E2 alone -induced recruitment of and AcH4 (C) to the IkBa promoter. Cells were pretreated with/ without E2 (10 27 M) or DPN (10 27 M) for 24 hrs and then stimulated with TNF-a (1 ng/mL) for 1 hr. THC (10 26 M) was given to cells at 1 h before E2 treatment in some experiments. ChIP samples were prepared as described in the text and analyzed using antibodies specific for p65, ERb or AcH4. The immunoprecipitated DNA fragments and input DNA were analyzed by real-time PCR. The y axis shows values were normalized to input DNA with values for vehicle treatment defined as 1. The numbers represent the mean6SEM from three experiments repeated in duplicate. *p,0.05 vs. Vehicle-treated RASMCs; #p,0.05 vs. TNF-a-treated RASMCs. doi:10.1371/journal.pone.0036890.g005 Figure 6. ChIP assays of binding of NFkB p65, ERb and AcH3 to the MCP-1 and CINC-2b promoters. Cells were pretreated without or with E2 for 24 hrs, then stimulated with TNF-a (1 ng/mL) for 1 hr. ChIP samples were prepared as described in the text and analyzed using antibodies specific for p65, ERb or AcH3. NFkB p65 was not accompanied by recruitment of AcH4 at the IkBa promoter, indicating that the increased p65 binding was insufficient to increase IkBa gene transcription. This finding suggests that other unidentified cofactors are required for NFkB p65-induced transcription of the IkBa gene under these conditions. However, when cells were pretreated with E2 or DPN and then challenged with TNF-a, both E2 and DPN further enhanced the TNF-a-induced increases in IkBa mRNA expression and protein levels, suggesting the possibility that TNFa may have recruited cofactors needed for IkBa gene transcription. The binding of ERb, but not ERa at the IkBa promoter was increased by E2 treatment. The ERb antagonist R,R-THC blocked the enhancement effects of E2 on IkBa gene transcription (p65 and AcH4 binding) and expression (mRNA and protein), suggesting that E2 may inhibit NFkB signaling by specifically targeting and enhancing events at the IkBa promoter, perhaps in a manner dependent on ERb. Curiously, using a computer program that analyzes promoters for putative transcription factor binding sites, we failed to identify any potential ER binding elements (ERE) within the IkBa promoter. These data suggest that ERb may not interact directly with the IkBa promoter to promote the binding of NFkB p65 to the promoter, but instead may work through recruitment of cofactors that enhance both binding of NFkB p65 to the promoter and transcription of the IkBa gene. Future studies will address how ERb is required for E2 mediated NFkB recruitment to and enhanced transcription of the IkBa gene.
In addition, we have observed that NFkB p65 is rapidly recruited to the MCP-1 and CINC-2b promoters in the presence of TNF-a. Under these conditions, ERb is absent from these promoters, and transcriptional activity of these genes is significantly increased compared to vehicle treatment, as indicated by AcH3 binding on these promoters and mRNA expression of these genes. In response to E2 pretreatment, binding of NFkB p65 to these promoters is greatly reduced and binding of ERb is greatly increased, transcriptional activity of these genes is significantly reduced, as indicated by decreased binding of AcH3 on these promoters and mRNA expression of these genes. At present, we can not definitively state why binding of ERb and NFkB p65 at the MCP-1 and CINC-2b promoters is mutually exclusive. Using computer programs designed to identify putative ERE, we could not identify any EREs within either the MCP-1 or CINC-2b promoters. Thus, these findings suggest that the presence of ERb at these promoters may occur through the use of an element that remains to be identified, or that ERb interacts with these promoters indirectly, i.e., through another DNA-binding protein (cofactor). Our future studies are attempting to address this question.
In summary, this study has elucidated a novel bimodal mechanism by which E2 inhibits NFkB signaling and thereby the inflammatory response to TNF-a in RASMCs. E2 both 1) enhances expression of IkBa, a direct inhibitor of NFkB activation, thus accelerating a negative feedback loop in NFkB signaling, and 2) directly inhibits binding of NFkB p65 to the promoters of inflammatory genes, including MCP-1 and CINC-2b, thereby inhibiting their expression. The findings that, in the presence of E2+TNF-a, ERb is recruited and the binding of NFkB is reduced at the MCP-1 and CINC-2b promoters, suggest that the ability of selective ERb activation to inhibit expression of inflammatory mediators in activated RASMCs may be related, in part, to interference with the DNA binding ability of NFkB p65 by ERb.

Cell Culture
Primary cultures of RASMCs were derived from 10-week-old female Sprague-Dawley rats (Charles River), as previously described [16,52]. 100908574. Cells were cultured in complete medium containing phenol red-free DMEM (Gibco) supplemented with 10% (vol/ vol) FBS, 4 mmol/L L-glutamine, 100 U/mL penicillin, and 100 mg/ml streptomycin. RASMCs were pre-treated with E2 (10 27 M) or vehicle (ethanol at a final concentration ,0.01%) for 24 hrs in all experiments. Cells were used within 5 passages and were identified as RASMCs by their characteristic morphology and positive immunostaining for a-smooth muscle actin (a-SMA, clone 1A4, DAKO). RASMCs pre-treated with or without E2 for 24 hours were then incubated with TNF-a (1 ng/ mL) for various time periods from 10 min to 6 hrs. To assess the ER dependence of the E2 effect on IkBa expression, cells were pretreated with the selective ERb agonist DPN (10 -7 M) or the selective ERa agonist PPT (10 -7 M) (Tocris Cookson, Ellisville, MO) for 24 hrs and then incubated with 1 ng/ml TNF-a for an additional 45 or 60 min. Another set of cells from the above experiments were exposed to the selective ERa antagonist MPP (10 26 M) or the selective ERb antagonist R,R-THC (10 26 M) (Tocris Cookson, Ellisville, MO) for 1 hr before the E2 (10 -7 M) pretreatment.

Real-time Quantitative RT-PCR Analyses
Real-time quantitative RT-PCR analysis was performed as described before [9,10,16]. Total RNA was extracted from cells using TRIzol (Invitrogen, Carlsbad, CA), and treated with DNAase I to remove genomic DNA. The protein-and DNAfree RNA was reverse transcribed to cDNA and analyzed using the SYBR Green RT-PCR kit (Applied Biosystems, Foster City, CA) and specific primers: IkBa forward, 59-CAGCAGACTCCACTC-CACTT-39 and IkBa reverse, 59-GAGAGGGGTATTTCCTC-GAA-39. MCP-1 forward 59-ATGCAGGTCTCTGTCACGCT -39 and MCP-1 reverse, 59-GGTGCTGAAGTCCTTAGGGT-39; CINC-2b forward 59-TCAGGGACTGTTGTGG -39 and CINC-2b reverse, 59-TGACTTCTGTCTGGGTG-39. cDNA was amplified by PCR in the iCycler for 40 cycles and relative RNA levels were calculated using the iCycler software. Samples were compared by the relative (comparative) Ct method. Fold induction or repression was measured relative to controls and calculated after adjusting for 18 s RNA (endogenous control) using 2 2DDCt , where D Ct = Ct interested gene -Ct 18 s RNA and DDCt = DCt treatment -DCt vehicle control.

Immunoblot Analyses
Quiescent RASMCs were incubated with E2 or vehicle for 24 hrs, followed by TNF-a for 10, 20, 30, 40, 50 and 60 min. Total protein was extracted and total and phospho-IkBa levels were assessed using Western blot analysis with selective anti-IkBa (Santa Cruz) and anti-phospho-IkBa (Cell Signaling) antibodies. Expression of ERa and ERb protein was assessed using Western blot analysis with selective anti-ERa (Santa Cruz HC-20) and anti-ERb (Millipore 07-359) antibodies. Protein loading was assessed by stripping the membranes and reprobing with anti-b-actin antibody (Sigma).

Chromatin Immunoprecipitation Assays
RASMCs were pretreated with E2 (10 27 M) or vehicle for 24 hrs and then treated with TNF-a (1 ng/mL) or vehicle for 1 hr. Cells were fixed with formaldehyde and subjected to chromatin immunoprecipitation (ChIP) analyses as previously described [53][54][55]. Briefly, cells were fixed with formaldehyde for 15 min and nuclei purified, then passed through a 22-gauge needle three times and sonicated to an average size of 500-1000 bps. Protein-DNA complexes were immunoprecipitated (IP) using 5 mg of antibodies selective for NFkB p65 (Abcam), ERb (Millipore), AcH3 or AcH4 (Upstate Signaling Solutions). The immune complexes were adsorbed with protein A beads or protein G beads blocked with bovine serum albumin and salmon sperm DNA (Upstate Signaling Solutions). Immunoprecipitants were washed, eluted and crosslinks were reversed overnight. The next day, samples were digested with Proteinase K and clarified by phenol:chloroform:isoamyl alcohol extraction. DNA was purified using mini spin columns and IP and non-IP DNA (Input) was analyzed by real time PCR using specific primers: IkBa forward, 59 AAGTCGTCGGTGGGAAAC 39 and IkBa reverse, 59 CCTGAGTGGCTGGAAAGT 39 that amplify 2405 to 2280 in the rat IkBa gene promoter; MCP-1 forward 59 GCACTTACTCAGCAGATTC 39 and reverse, 59 GCCTCAGCCTTTTATTGT 39 that amplify 2208 to 291 in the rat MCP-1 gene promoter; forward 59 CAAACGAG-GACTGGGTAG 39 and reverse, 59 GACTTAGGTGCAGG-GACT 39 that amplify 2346 to 2541 in the rat CINC-2b gene promoter. Results are representative of three experiments.

Statistical Analysis
Data are expressed as mean6SEM. Statistical analysis was performed with one-way ANOVA or Student's t test, as appropriate. Values of P,0.05 were considered significant.

Supporting Information
Figure S1 Co-treatment with E2 and TNF-a does not stimulate TNF-a expression in RASMCs. Cells were grown to subconfluence (<95%) in 6-well plates, deprived of serum for 24 hrs, pretreated with 10 27 M E2 or vehicle for 24 hrs, then treated with TNF-a (1 ng/mL) for the periods indicated. Conditioned media was collected. Data, expressed as means6SEM, are from a double sandwich ELISA assay. (TIF) Figure S2 Representative micrographs of RASMCs pretreated with E2 (10 27 M) or vehicle for 24 hrs before incubated with TNF-a (1 ng/mL) for 30 min. Cells were analyzed using anti-NFkB p65 antibody (A1,B1,C1,D1) and nuclei were stained with DAPI (A2,B2,C2,D2). Merged images are shown in the panel A3,B3,C3,D3. E. Bar graph demonstrating the percentage of cells with NFkB p65 nuclear translocation after TNF-a6E2 treatment for 0, 15, 30 and 60 min. Results are mean6SE from 3 slides/group; a total of .200 cells were counted/group). *P,0.05 compared with vehicle control group. (TIF) Figure S3 Representative Western blots of ERa and ERb in E26TNF-a treated RASMCs. Cells were pretreated with E2 (10 2 7 M) or vehicle for 24 h, and then treated with TNF-a (1 ng/ ml) for an additional 6 hrs. Blots was reprobed with antibody against b-actin for input loading. (TIF) Text S1 Detailed protocol. (DOC)