NF-κB represses retinoic acid receptor–mediated GPRC5A transactivation in lung epithelial cells to promote neoplasia

Chronic inflammation is associated with lung tumorigenesis, in which NF-κB–mediated epigenetic regulation plays a critical role. Lung tumor suppressor G protein–coupled receptor, family C, member 5A (GPRC5A), is repressed in most non–small cell lung cancer (NSCLC); however, the mechanisms remain unclear. Here, we show that NF-κB acts as a transcriptional repressor in suppression of GPRC5A. NF-κB induced GPRC5A repression both in vitro and in vivo. Intriguingly, transactivation of NF-κB downstream targets was not required, but the transactivation domain of RelA/p65 was required for GPRC5A repression. NF-κB did not bind to any potential cis-element in the GPRC5A promoter. Instead, p65 was complexed with retinoic acid receptor α/β (RARα/β) and recruited to the RA response element site at the GPRC5A promoter, resulting in disrupted RNA polymerase II complexing and suppressed transcription. Notably, phosphorylation on serine 276 of p65 was required for interaction with RARα/β and repression of GPRC5A. Moreover, NF-κB–mediated epigenetic repression was through suppression of acetylated histone H3K9 (H3K9ac), but not DNA methylation of the CpG islands, at the GPRC5A promoter. Consistently, a histone deacetylase inhibitor, but not DNA methylation inhibitor, restored GPRC5A expression in NSCLC cells. Thus, NF-κB induces transcriptional repression of GPRC5A via a complex with RARα/β and mediates epigenetic repression via suppression of H3K9ac.


Introduction
Tumorigenesis, a dedifferentiation process, is associated with chronic inflammation (1-3). However, the molecular mechanism that switches the expression profile from homeostasis to dedifferentiation remains elusive. Lung is an organ that expose to various environmental insults, such as bacteria and virus infection, air pollution or cigarette smoking. These risk factors trigger chronic inflammation which release reactive oxygen and nitrogen species (ROS and RNS), causing DNA damage and mutations (4). Pro-inflammatory cytokines from microenvironment can activate the NF-κB signaling pathway, a pivotal transcriptional factor for induction of multiple downstream genes, in target cells.
These downstream NF-κB targets include molecules involved in anti-apoptosis, cell survival, proliferation, angiogenesis, and immune-evasion. Up to now, nearly all NF-κB-mediated functions are attributed to its trans-activation activities, or activation of NF-κB downstream genes. However, little is known about the role of NF-κB in repression of the genes for homeostasis or differentiation.
These observations suggest that GPRC5A is a biomarker of differentiation in lung and oral epithelial cells, whereas repression of GPRC5A promotes dedifferentiation or pro-oncogenic process for neoplasia. However, the underlying mechanism of GPRC5A repression remains elusive.
In this study, we showed that NF-κB acts as a transcription repressor to repress GPRC5A expression. We showed that NF-κB, via interaction with retinoic acid receptor (RAR), inhibits RARmediated transcription of GPRC5A, switching the gene expression from homeostasis to dedifferentiation in lung epithelial cells for neoplasia.

Results
Active NF-κB is strongly correlated with GPRC5A repression in human NSCLC and COPD samples. GPRC5A is predominately expressed in lung epithelial cells, but its expression is often loss in lung cancer. To determine the relationship between NF-κB activation and GPRC5A expression, we measured the levels of GPRC5A expression and active p65 (nuclear p65) in samples via immunohistochemistry (IHC) staining. The results showed that GPRC5A expression was significantly repressed, whereas active p65 was greatly increased, in most of NSCLC, including squamous cell carcinoma (SCC) and adenocarcinoma (ADC), and all COPD tissues, compared to that in adjacent normal (AN) lung tissues ( Figure 1A and B). Pearson correlation analysis showed that GPRC5A protein level and active p65 were inversely correlated in normal lung, NSCLC and COPD tissues ( Figure 1C). Further analysis of TCGA database showed that the mRNA level of TNFα, a key cytokine responsible for NF-κB activation, was inversely correlated with GPRC5A expression in Lung adenocarcinoma (LUAD) ( Figure 1D). Taken together, these results suggest that activated NF-κB is strongly correlated with GPRC5A repression.

Inflammatory stimuli repress GPRC5A expression both in vitro and in vivo.
To determine the role of inflammatory signaling on GPRC5A expression, we examined the effects of TNFα in small airway epithelial cells (SAEC). Immunoblot assay showed that, all trans-retinoic acid (ATRA) induced GPRC5A expression, confirming that GPRC5A is a RA target gene (5,16,17) (Figure 2A).
Interestingly, treatment with TNFα as well as cigarette smoking condense (CSC) suppressed ATRAinduced GPRC5A expression. Similarly, TNFα treatment inhibited GPRC5A expression in several NSCLC cell lines (Calu-1, H322 and H292G cells) ( Figure 2B). Q-PCR analysis showed that the mRNA level of GPRC5A was significantly suppressed after 6 hours of TNFα treatment ( Figure 2C). In comparison, GPRC5A protein level started to decrease after 12 hours of TNFα treatment in Calu-1 cells ( Figure 2D). These results suggest that TNFα represses GPRC5A at transcriptional level.
Next, we examined the effect of inflammation on GPRC5A repression in vivo using an NF-κBdriven luciferase (NF-κB-luc) mouse model. After aero-exposure of these mice to LPS (aero-LPS) for 30 min, luciferin was injected intraperitoneally (i.p.) to visualize NF-κB activation. The images showed that aero-LPS exposure induced intensive bioluminescence in the lungs of these mice ( Figure   2E, top panel), quantitation of relative bioluminescence intensity confirmed NF-κB activation in these mice ( Figure 2E, bottom panel). When lung tissues from these mice were analyzed for Gprc5a expression, immunoblot showed that Gprc5a was significantly suppressed following aero-LPS exposure for 4-8 days ( Figure 2F). Consistently, Q-PCR analysis showed that Gprc5a mRNA was greatly suppressed in mouse lung tissues treated with aero-LPS for 4 day compared to untreated ones ( Figure 2G), indicating that NF-κB-mediated repression of Gprc5a occurs at the transcription level.
Taken together, inflammatory stimuli, TNFα used in vitro or LPS used in vivo, significantly suppress GPRC5A expression at the transcriptional level both in vitro and in vivo.
The trans-activation domain of p65 is essential for NF-κB-mediated GPRC5A repression. NF-κB is a major intracellular mediator of inflammatory signaling elicited from TNFα. To determine whether NF-κB is the mediator responsible for TNFα-induced GPRC5A repression, we examined the role of TNFα in cancer cells treated with either RelA/p65 (subunit of NF-κB) knockdown by small interfering RNA (siRNA) or overexpression of dominant negative inhibitor IκBα S32/36A mutant (IκBα-AA). The immunoblot assay showed that p65-knockdown abolished TNFα-induced GPRC5A repression in Calu-1 cells ( Figure 3A). Similarly, over-expression of IκBα-AA blocked TNFα-induced GPRC5A repression ( Figure 3B). This suggests that TNFα-induced repression of GPRC5A is through NF-κB.
Post-translation modifications, including phosphorylation, are known to play an important role in regulating the trans-activation activities of RelA/p65 (18)(19)(20). Previously, it was shown that phosphorylation of serine-276 is critical for p65-mediated repression of BRMS1 (breast cancer metastasis suppressor 1) (21). We asked if serine-276 is essential for repression of GPRC5A. To determine the role of S276 of p65, we established inducible clones, expressing either wild type (WT) p65 or p65-S276A mutant in Calu-1 cells. Immunoblot analysis showed that WT-p65 suppressed GPRC5A expression whereas p65-S276A did not ( Figure 3E). Consistently, Q-PCR analysis showed that expression of WT-p65 repressed GPRC5A at mRNA level, whereas p65-S276A mutant did not ( Figure 3F). Taken together, the S276 site in the RHD domain of p65 is critical for NF-κB-mediated repression of GPRC5A.
NF-κB-mediated GPRC5A repression is independent of its trans-activation functions. NF-κB is known to act as a transcription activator. One possibility is that NF-κB-mediated repression is through trans-activation of its downstream target genes, which indirectly repress GPRC5A expression. To test this possibility, we examined the effects of protein translation inhibitor cycloheximide (CHX) on NF-κB-induced trans-activation and trans-repression. Immunoblot analysis showed that TNFα induced IκBα degradation at 30 minutes in Calu-1 cells ( Figure 4A), via an ubiquitination pathway (22,23); The IκBα level was restored at 60 minutes through de novo protein synthesis ( Figure 4A) since IκBα is a well-known NF-κB target gene (22). Although CHX treatment blocked IκBα expression at 60 minutes ( Figure 4A), it did not block the IκBα mRNA level ( Figure 4B), indicating that CHX treatment indeed blocked newly protein synthesis. Importantly, CHX treatment, which blocks protein synthesis, did not alter TNFα-induced GPRC5A repression at both the mRNA ( Figure 4C) and protein levels ( Figure 4D). These suggest that trans-activation of NF-κB downstream target genes is not required for TNFα-induced GPRC5A repression. In addition, treatment with proteasome inhibitor MG132 and autophagy inhibitor chloroquine (CQ) had no effect on GPRC5A level following TNFα treatment or p65 induction (Supplementary Figure 1), supporting that NF-κB-mediated GPRC5A repression is regulated at the transcriptional level. Taken together, NF-κB represses GPRC5A expression at transcriptional level, which is independent of trans-activation of NF-κB downstream target genes.
Next, we asked whether NF-κB acts via direct binding to the potential NF-κB response cis-element (NFRE) in the GPRC5A promoter. By screening potential NFRE-like sequences in GPRC5A promoter via software, we found that there are three potential NFRE sites in the promoter region. To determine whether these sites, designated as A, B, C ( Figure 4E, upper), play a role in NF-κB-mediated transcriptional GPRC5A repression, we examined the effect of p65 on GPRC5A promoter-driven luciferase reporters (GPRC5A-luc) with various mutations on these sites. Transfection of HEK293T cells with various constructs showed that p65 repressed the luciferase activity from GPRC5A-luc.
Importantly, mutation of any of three potential sites or different combination did not affect p65mediated repression of GPRC5A-luc reporters ( Figure 4E, below). This suggests that, none of three potential NF-κB-binding cis-elements is required for NF-κB-mediated repression. Instead, we found that the construct GPRC5A-luc-DR5mut, which has mutated retinoic acid response element (RARE) in the promoter of GPRC5A (24), completely lost luciferase activity regardless of the expression of NF-κB p65 subunit ( Figure 4F). This suggests that RARE at the GPRC5A promoter is critical for regulating its expression. Thus, it raises an interesting question if NF-κB-mediated repression of GPRC5A is through disrupting the transcription complex at RARE in its promoter.
NF-κB represses GPRC5A via inhibiting RA signaling. Since GPRC5A is a RA target gene, we asked whether NF-κB represses GPRC5A expression via inhibiting RA-mediated transcription. First, we examined the role of the RA signaling pathway in GPRC5A expression. All-trans retinoic acid (ATRA) treatment induced both GPRC5A and RARβ in NSCLC cell line H157 by RT-PCR analysis; however, the induction was eliminated when RARβ is knockdown by siRNA ( Figure 5A). These suggest that RARβ is essential for GPRC5A expression. Consistently, immunoblot analysis showed that, ATRA could induce GPRC5A in H157-V (vector) cells, but this induction was eliminated in H157-ASβ cells, in which RARβ was blocked by RARβ antisense (ASβ) RNA (25) ( Figure 5B). The inhibition was at mRNA level since induction of GPRC5A mRNA was eliminated in H157-ASβ ( Figure 5C). These suggest that RARβ and RA signaling are essential for GPRC5A expression.
Next, we examined the effect of TNFα on ATRA-induced GPRC5A expression. Immunoblot showed that ATRA treatment induced GPRC5A expression in Calu-1 cells ( Figure 5D). Importantly, TNFα treatment significantly inhibited GPRC5A expression at both basal and RA-induced levels ( Figure 5D). Consistently, RT-PCR analysis confirmed that TNFα-mediated GPRC5A repression was at mRNA level ( Figure 5E). Similar effects were observed in H157 cells, in which GPRC5A at both basal and ATRA-induced levels were significantly repressed by TNFα ( Figure 5F). Consistently, cotransfection of p65 suppressed both basal and ATRA-induced luciferase activities in GPRC5A promoter driven-luc reporter in HEK293T cells ( Figure 5G). Taken together, RA-induced GPRC5A is repressed by NF-κB signaling.

RelA/p65 interacted with RARα/β and is recruited to the RARE at the GPRC5A promoter.
GPRC5A is a target gene of retinoic acid (5,16,17); ATRA treatment induces the association of RARα/β to RARE at the GPRC5A promoter (17). To determine the mechanism of NF-κB-mediated GPRC5A repression, we examined the binding of NF-κB at the GPRC5A promoter following TNFα treatment via chromatin immunoprecipitation (ChIP) assay. For comparison, we first examined the recruitment of p65 to IκBα promoter. ChIP analysis showed that p65 was recruited to IκBα promoter after 30 minutes of TNFα treatment, which was followed by enhanced recruitment of RNA polymerase II at 60 and 150 minutes ( Figure 6A), indicating that NF-κB induces the assembly of transcription machinery at the IκBα promoter. Next, we examined the effect of NF-κB association at the GPRC5A promoter. Although TNFα treatment induced p65 recruitment to RARE, it significantly suppressed the recruitment of RNA polymerase II to RARE at the GPRC5A promoter at 150-minute interval compared to those at 0 and 30 minutes ( Figure 6B). In comparison, the binding of RARα and RXRα to GPRC5A promoter was not changed ( Figure 6B). This suggests that TNFα treatment disrupts the assembly of RNA Pol II complex after p65 is recruited to RARE at the GPRC5A promoter.
Since the cis-element RARE is recognized by the trans-element RAR/RXR complex, it raises a question if recruitment of NF-κB to RARE is through the RAR/RXR complex. Next, we examined the interaction between RARα and p65 via immunoprecipitation (IP) assay. IP-immunoblot analysis showed that IP Flag-tagged RARα pull-downed myc-tagged p65 ( Figure 6C). This suggests that p65 can interact with RARα. Similarly, IP Flag-tagged RARβ2 and RARβ4 also pull-downed myc-tagged p65 ( Figure 6D), indicating that p65 can also interact with RARβ. Thus, NF-κB can interact with the RARα/β complex. Because p65-S276A loss the ability to repress GPRC5A, we then asked if p65-S276A is able to interact with RAR complex. IP-immunoblot analysis showed that, while IP Flag-RARα pull-downed GFP-p65, it failed to pull-down GFP-p65-S276A ( Figure 6E). Consistently, immunofluorescent (IF) staining analysis showed that GFP-p65 (green) was highly co-localized with RARα-F (red) in transfected Calu-1 cells; whereas GFP-p65-S276A (green) was mainly located in cytoplasm, not co-localized with RARα-F (red) ( Figure 6F). These suggest that the S276 of p65 is essential for interaction with RARα. Taken together, these results suggest that p65, through interaction with RARα/β, is recruited to RARE at the GPRC5A promoter, leading to disrupted the assembly of RNA polymerase II, resulting in suppressed transcription. DNA methylation and histone post-translation modification are the major mechanisms involved in epigenetic silence of TSGs (27,28). To determine the potential role of DNA methylation in GPRC5A repression, we examined the status of DNA methylation in two CpG islands at the GPRC5A promoter in lung cancer tissues and adjacent normal tissues via bisulfite sequencing PCR (BSP). We found that there was no significant difference in GPRC5A promoter methylation between lung cancer and adjacent normal tissues (Supplementary Figure 2). Moreover, we found that overexpression of p65 did not significantly alter the DNA methylation status in Calu-1 cells (Supplementary Figure 3). These suggest that DNA methylation is unlikely to play an important role in GPRC5A repression in NSCLCs.
Next, we asked if post-translation modification of histones plays a role in GPRC5A repression and whether it is induced by NF-κB. ChIP assay showed that TNFα treatment significantly suppressed RNA polymerase II assembly in GPRC5A promoter ( Figure 7B); noticeably, acetylated H3K9 (H3K9ac), an active marker of gene expression, was also significantly suppressed in GPRC5A promoter following TNFα treatment. In comparison, H3K27ac at the GPRC5A promoter was not changed following TNFα treatment ( Figure 7B). To extend the analysis further, we also examined the effects of NF-κB mutants in histone modification. Over-expression of WT-p65, but not S276A-p65, significantly inhibited the recruitment of RNA polymerase II at the GPRC5A promoter in Calu-1 cells ( Figure 7C); importantly, expression of WT p65 inhibited H3K9ac, whereas expression of p65-S276A did not ( Figure 7D). These suggest that NF-κB inhibits both RNA polymerase II assembly and H3K9ac at the GPRC5A promoter. Thus, NF-κB-induced suppression of H3K9ac is involved in epigenetic repression of GPRC5A.

Discussion
In this study, we showed that NF-κB can function as a transcriptional repressor on GPRC5A expression via a trans-activation independent mechanism. NF-κB-mediated GPRC5A repression is through interaction with RARα/β, resulting in the suppression of RNA polymerase II recruitment in RARE at the GPRC5A promoter. Moreover, NF-κB inhibits histone H3K9ac, rather than induces DNA hyper-methylation, at the GPRC5A promoter. Consistently, SAHA treatment largely restores GPRC5A expression in NSCLC cells, whereas treatment with 5-Aza-dc has little effect. Because NF-κB is a well-known transcriptional activator (30). However, little is known about its role as a transcriptional repressor. Previously, it was reported that RelA/p65 can act as a transcriptional repressor to repress metastasis suppressor gene BRMS1 (breast cancer metastasis suppressor 1).
Mechanistically, p65 binds the specific κB binding site at the BRMS1 promoter and recruit DNMT-1 (DNA (cytosine-5)-methyltransferase 1), which mediates DNA methylation in CpG islands located at the promoter region, resulting in epigenetic repression (21). Nevertheless, the mechanism of NF-κBmediated BRMS1 repression is via induction of DNA methylation, which is distinct from the repression of GPRC5A via histone deacetylation found in this study.
Retinoic acid (RA) signaling plays an important role in the regulation of cell proliferation, differentiation and homeostasis. RA exerts its role by interacting with nuclear retinoic acid receptors (RAR) and retinoic X receptors (RXRs). GPRC5A is a RA target. Thus, the status of RARs are crucial for GPRC5A expression. Of note, RARβ expression is often lost or reduced in a large percentage of lung cancer (31,32). For example, methylation of the RARβ2 promoter is found in 40% of NSCLCs. NF-κB has been found to crosstalk with other signaling pathways, including EGFR (33), p53 (34), HIF1α (35), glucocorticoid signaling (36). It was shown that vitamin A deficiency enhances inflammatory response or NF-κB activation in vivo, whereas administration of ATRA reduces inflammatory response or NF-κB activation in animal models (37). These suggest that vitamin A or RA signaling inhibits NF-κB signaling. Moreover, the p50 and p65 components of NF-κB were shown to bind to retinoid X receptor (RXR) in a ligand-independent manner, and RXR can inhibit NF-κB activation in a ligand-dependent manner (38). These suggest that there is a crosstalk between NF-κB and RA signaling. Retinoids play a fundamental role in development and homeostasis (39), whereas disruption of the RA signaling pathway is implicated in neoplasia or cancers (40) (41). Conversely, activation of RA signaling is used as a strategy for cancer prevention and therapy (42). GPRC5A was originally cloned as retinoic acid-inducible gene 1 (RAIG1) (5). A RA response element (RARE) at the GPRC5A promoter is critical for its expression (17). Thus, GPRC5A functions as a key mediator of RA signaling for differentiation or homeostasis in lung tissue. Presumably, the mutually inhibitory effects between NF-κB and RARα/β signaling represent a molecular switch at the crossroad of cell fate determination between differentiation and dedifferentiation.
Previously, we showed that Gprc5a-ko leads to increased NF-κB activation in lung epithelium, which is associated with lung tumorigenesis in mouse model. Gprc5a-ko mice are susceptible to pulmonary inflammation and LPS-induced acute lung injury (10)(11). However, the regulatory role of Gprc5a-deletion on NF-κB activation in lung epithelial cells appears to be indirect since the effects are stronger in vivo than in vitro. In addition, the effect of GPRC5A-knockdown on the NF-κB activation in human lung epithelial cells is not as strong as that in mouse lung epithelial cells. It is likely that multiple mechanisms are involved in the cross-talk between GPRC5A and NF-κB.
Taken together, we find that NF-κB functions as a transcriptional repressor to suppress GPRC5A expression. NF-κB interacts with RARα/β via its transactivation domain, which disrupts the assembly of RAR-mediated RNA polymerase II complex at the GPRC5A promoter, leading to its transcriptional repression. Concurrently, NF-κB induces epigenetic repression via H3K9ac histone deacetylation at the GPRC5A promoter ( Figure 8). Importantly, this epigenetic mechanism of GPRC5A repression is prevalent in NSCLCs. We propose that NF-κB-mediated GPRC5A repression contributes to dedifferentiation or neoplasia in lung epithelial cells. Animal experiment. NF-κB-driven luciferase transgenic mice were administrated with LPS through atomization inhalation (500g in 5 ml for each cage per day). D-luciferin was injected intraperitoneally 30 minutes after LPS treatment; another two hours later, fluorescence imaging was performed to verify the activation of NF-κB. Mice were sacrificed at day 4, 6, 8, and lungs were analyzed via western blot and Q-PCR assays to examine the protein and mRNA levels of Gprc5a.

Cells
Luciferase Reporter Assay. Cells were grown to 50% confluence in 24-well plate and then cotransfected with reporter gene constructs (GPRC5A-luc or NF-κB-luc) and pcDNA3.1-p65 and its mutants. Plasmid pRL-TK (Promega) was used as internal control in all transfection assays. All transfections were done with Lipofectamine 2000 (Invitrogen) according to the instruction of manufacture (43). Cell extracts were prepared 48 hours after transfection, and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI). All experiments were performed three times in triplicate.
Immunoprecipitation assay and Western blot analysis. Cells or tissue were lysed with RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.5% deoxycholate, 50 mM Tris (pH 8.0), 25 mM NaF, 2 mM Na3VO4, 5 mM PMSF and 2 mg/ml of aprotinin) (44). Immunoprecipitation and Western blot were performed as described previously (45,46). Briefly, 2 g Flag antibody were used for 500g total whole cell lysate for each immunoprecipitation. For western blot, whole cell lysate containing 30 g whole protein was mixed with 2x SDS-PAGE reduced loading buffer and boiled at 95℃ for 5 min for per gel well. Protein samples were separated by SDS-PAGE and transferred onto nitrocellulose membrane for western blot. Non-specific binding to antibodies were blocked with 5% non-fat milk at room temperature for 1h. Primary antibodies were incubated at 4℃ overnight. To remove the residual primary antibody, membranes were washed three times in TBST (0.05% Tween-20) while agitating, 5 minutes for per wash. Second antibody conjugated with HRP were incubated at room temperature for 1 hour and washed three times in TBST while agitating. ECL kit (Millipore) was used to detection. Immunohistochemistry staining. Tissue samples from human lung cancer, chronic obstructive pulmonary disease (COPD) and adjacent normal tissues were stained with anti-GPRC5A (1:200 dilution) and anti-p65 NLS (1:100 dilution) antibodies, and each sample was scored by an H-score method that combines the values of immune-reaction intensity and the percentage of cell staining as described previously (47). Pearson correlation analysis was used to analyze the relationship between GPRC5A expression and activated p65; statistical significance was defined as p < 0.05.  was treated with TNFα (10 ng/ml) for 24 hours and the protein level of GPRC5A was determined by western blotting. (C, D) Calu-1 cells were treated with TNFα (10 ng/ml) for indicated times. The levels of mRNA (C) and protein (D) of GPRC5A was determined by quantitative PCR and western blotting, respectively. Data are presented as the mean ± SD. (E) Represent data of luciferase activity of NF-κB driven luciferase transgenic mice treated with placebo or LPS (500 μg/5 ml) for 30 min through inhalation. (F) Mice treated with LPS through inhalation for various days and the mouse lung tissue was analyzed for Gprc5a expression via western blotting, data are presented as the mean ± SD. (G) Mouse was treated with LPS through inhalation for 4 days, and lung tissue was collected, the mRNA level of Gprc5a was determined by quantitative PCR. All data are presented as mean ± SD from three independent experiments with duplicates and analyzed by 2-tailed Student's t-test; ns, not significant; *p<0.05; **p< 0.01; ***p<0.001.

Figure 3. Transcription activation domain of NF-κB is required for GPRC5A repression. (A)
Small interfering RNA (siRNA) targeting p65 and scramble control siRNA were transfected to Calu-1 cells, then treated with or without TNFα. The protein levels of GPRC5A and p65 were determined by western blotting. (B) Calu-1 cells were transfected with plasmid overexpressing IκBα AA mutant (S32A, S36A) or vector control, then treated with or without TNFα, GPRC5A and IκBα protein levels were determined with specific antibodies through western blotting. (C) Schematic representation of wild type and truncation mutant of RelA/p65. (D) Calu-1 transfectants harboring vector control or inducible expression of full-length and truncated p65 were established; cells were treated with doxycycline (300 ng/ml) for 24 hours, and GPRC5A protein levels were analyzed by western blotting. These blots were run in separated gels performed parallelly with equal loading (please see Uncropped/Unedited gel document). (E, F) Calu-1 cells with inducible expression wild-type p65 and S276A mutant were treated with doxycycline (300 ng/ml) for 24 hours, GPRC5A protein and mRNA levels were determined by western blotting (E) and quantitative PCR (F) respectively. Data are presented as mean ± SD from three independent experiments with duplicates and analyzed by 2-tailed Student's t-test, *p<0.05; **p<0.01; ***p< 0.001. Calu-1 cells were treated with TNFα and CHX separately or in combination, IκBα protein and mRNA levels were determined by western blotting and quantified with image J (A) and quantitative PCR (B) respectively. Data are presented as the mean ± SD. (C, D) Calu-1 cells were treated with TNFα and CHX separately or in combination, GPRC5A mRNA and protein levels were determined by quantitative PCR (C) and western blotting and quantified with image J (D) respectively. Data are presented as the mean ± SD. (E) Three NF-κB binding sites (designated as A, B and C) on GPRC5A promoter-luc plasmid were mutated individually or in combination. The repression effect of p65 was determined by luciferase assay. (F) RAR response element (RARE) at the GPRC5A promoter-luc plasmid was mutated and the repression effect of p65 was determined by luciferase assay. Data are presented as mean ± SD from three independent experiments with duplicates and analyzed by 2-tailed Student's t-test, *p<0.05; **p<0.01; ***p<0.001. Calu-1 cells was treated with TNFα and ATRA separately or in combination, and the GPRC5A protein and mRNA levels were determined by western blotting (D) and RT-PCR and quantified by image J (E). (F) H157 cells was treated with TNFα and ATRA separately or in combination. GPRC5A mRNA levels were determined by RT-PCR and quantified by image J. (G) 293T cells was transfected with GPRC5A-luc and p65 plasmids and treated with or without ATRA. The p65 repression effect was determined by luciferase assay. Data are presented as mean ± SD from three independent experiments with duplicates and analyzed by 2-tailed Student's t-test, *p<0.05; **p<0.01; ***p<0.001.  GPRC5A protein expression level in multiple human NSCLC cell lines and normal human bronchial epithelial cell line (16HBE) were analyzed by western blotting. (B) Calu-1 cells treated with or without TNFα (10 ng/ml) for 12 hours, binding of RNA polymerase II and the histone modification at the GPRC5A promoter were analyzed by ChIP using specific antibody. (C, D) Calu-1 cells with inducible expressing of WT and S276A mutant p65 treated with doxycycline (300 ng/ml) for 12 hours, the change of RNA polymerase II binding (C) and histone modification (H3K9ac) at the GPRC5A promoter (D) were analyzed by ChIP. Input as positive control and normal IgG (N IgG) as negative control. (E-H) A549, H1975 and Calu-1 cells were treated with 5-Aza-dc (1 μM, 4 days) or SAHA (2.5 μM, 24 hours) individually or in combination, GPRC5A protein (E-G) and mRNA levels (H) were analyzed via western blotting and quantitative PCR. Data are presented as the mean ± SD. (I) Calu-1 cells were treated with or without SAHA (2.5 μM) for 12 hours, RNA polymerase II binding at the GPRC5A promoter was analyzed by ChIP. (J) Normal human bronchial epithelial cell line (16HBE) and multiple human NSCLC cell lines were treated with or without SAHA (2.5 μM, 24 hours), GPRC5A protein levels were analyzed by western blotting. (K) Calu-1 cells were pretreated with DMSO (as vehicle control), 5-Aza-dc (1 μM, 3 days) or SAHA (2.5 μM, 3 hours) followed by TNFα (10ng/ml) treatment for additional 24 hours. GPRC5A protein levels were analyzed by western blotting. All data are presented as mean ± SD from three independent experiments with duplicates and analyzed by 2-tailed Student's t-test, *p<0.05; **p<0.01; ***p<0.001.