Bromodomain and Extraterminal (BET) Protein Inhibition Suppresses Human T Cell Leukemia Virus 1 (HTLV-1) Tax Protein-mediated Tumorigenesis by Inhibiting Nuclear Factor κB (NF-κB) Signaling*

Background: The HTLV-1 oncoprotein Tax induces sustained NF-κB activation for cell survival and proliferation gene expression. Results: Recruitment of Brd4 to acetylated RelA is essential for Tax-mediated NF-κB target gene expression and cell proliferation. Conclusion: Blocking the binding of Brd4 to acetylated RelA by JQ1 suppresses Tax-mediated NF-κB activation and tumorigenesis. Significance: Inhibiting Brd4 could be a potential therapeutic strategy for diseases associated with HTLV-1 infection. The etiology of human T cell leukemia virus 1 (HTLV-1)-mediated adult T cell leukemia is associated with the ability of viral oncoprotein Tax to induce sustained NF-κB activation and the expression of many NF-κB target genes. Acetylation of the RelA subunit of NF-κB and the subsequent recruitment of bromodomain-containing factor Brd4 are important for the expression of NF-κB target genes in response to various stimuli. However, their contributions to Tax-mediated NF-κB target gene expression and tumorigenesis remain unclear. Here we report that Tax induced the acetylation of lysine 310 of RelA and the binding of Brd4 to acetylated RelA to facilitate Tax-mediated transcriptional activation of NF-κB. Depletion of Brd4 down-regulated Tax-mediated NF-κB target gene expression and cell proliferation. Inhibiting the interaction of Brd4 and acetylated RelA with the bromodomain extraterminal protein inhibitor JQ1 suppressed the proliferation of Tax-expressing rat fibroblasts and Tax-positive HTLV-1-infected cells and Tax-mediated cell transformation and tumorigenesis. Moreover, JQ1 attenuated the Tax-mediated transcriptional activation of NF-κB, triggering the polyubiquitination and proteasome-mediated degradation of constitutively active nuclear RelA. Our results identify Brd4 as a key regulator for Tax-mediated NF-κB gene expression and suggest that targeting epigenetic regulators such as Brd4 with the bromodomain extraterminal protein inhibitor might be a potential therapeutic strategy for cancers and other diseases associated with HTLV-1 infection.


The etiology of human T cell leukemia virus 1 (HTLV-1)-mediated adult T cell leukemia is associated with the ability of viral oncoprotein Tax to induce sustained NF-B activation and the expression of many NF-B target genes. Acetylation of the
Human T cell leukemia virus 1 (HTLV-1) 2 is the etiological factor of adult T cell leukemia/lymphoma (ATL) (1). The HTLV-1 pathogenesis has been largely attributed to its encoded 40-kDa regulatory protein Tax, which is critical to the viral life cycle and plays a central role in the early stage of ATL leukemogenesis (2)(3)(4). Tax immortalizes human primary T cells in the presence of IL-2 and is able to transform rodent fibroblasts (2,3). In addition, Tax-transformed lymphoid cells and fibroblasts form tumors in immunodeficient mice (5,6). Moreover, Tax is directly required for establishment and maintenance of the transformed phenotype because the HTLV-1 genome without Tax loses its original transforming ability (7). Furthermore, Tax-transgenic mice developed tumors and several inflammatory diseases (6).
The oncogenic properties of Tax are largely due to its abilities to regulate cellular transcription factors, including cAMP response element-binding protein (CREB) and NF-B (3,8,9). NF-B plays a key role in regulating immune and inflammatory responses, apoptosis, cell proliferation and differentiation, and tumorigenesis (10). Tax regulates the expression of a variety of NF-B target genes, including cytokines, growth factors, cytokine and growth factor receptors, and antiapoptotic factors. The expression of these target genes contributes to Tax-mediated tumorigenesis (8,11). Unlike the transient activation of NF-B by proinflammatory cytokines, one hallmark of HTLV-1-infected T cell lines is the constitutive nuclear localization of NF-B, which critically contributes to T cell transformation both by promoting T cell proliferation and inhibiting T cell apoptosis (5,12). The constitutively active NF-B is also essential for Tax-induced transformation of rat fibroblasts (13,14). * This work is supported, in whole or in part, by National Institutes of Health Grants DK-085158, DK-093865 (to L. F. C.), and CA116616 (to G. T. X.). This work was also supported by funds provided by the University of Illinois at Urbana-Champaign (to L. F. C.). 1 To whom correspondence should be address: Dept Additionally, transgenic mice expressing a Tax mutant defective for NF-B activation fail to develop a lethal cutaneous disease that resembles a skin disease occurring during the preleukemic stage in HTLV-1-infected patients (15). Tax utilizes multiple tactics to persistently activate NF-B. Tax associates with TAK1 and IKK␥ to activate IKK␤/⌱⌲⌲2, which then phosphorylates IB␣ and triggers its ubiquitination and degradation, leading to the rapid nuclear translocation of the p50 and RelA heterodimer to activate the expression of its target genes (5,10). Tax also associates with TAX1BP1, a component for the active A20-E3 ligase complex, to terminate the negative feedback regulation of NF-B activity by A20 (16). In addition to the cytoplasmic events stimulating the nuclear translocation of NF-B, Tax also regulates the nuclear activity of NF-B. Tax activates IKK␣/⌱⌲⌲1 to phosphorylate RelA at serine 536, enhancing Tax-induced transcriptional activity of NF-B (17). Phosphorylation of serine 536 has been shown to stimulate the binding of RelA to histone acetyltransferases such as p300/CREB-binding protein, which, in turn, mediates the acetylation of RelA (18). Acetylation of RelA at lysine 310 recruits bromodomain-containing factor Brd4 to activate the positive transcription elongation factor b complex and RNA polymerase II for the transcription of NF-B target genes (19,20). Hyperacetylated RelA has also been found in many cancer cells, and acetylation accounts for constitutively active NF-B in cancer cells by prolonging NF-B nuclear retention in tumors (21,22). However, whether Tax induces the acetylation of RelA and the contribution of acetylation of RelA and the subsequent binding of Brd4 to acetylated RelA in Tax-mediated sustained NF-B activation and tumorigenesis remain largely unknown.
Brd4 belongs to the bromodomain extraterminal (BET) family and has been shown to play important roles in viral segregation, cell cycle progression, gene transcription, and tumorigenesis (23,24). By binding to acetylated histone H3 or H4, Brd4 regulates gene transcription by recruiting different transcription components such as Mediator and positive transcription elongation factor b (23,24). Aberrant chromosomal translocation of Brd4 to nuclear protein in testis (NUT) causes a rare and aggressive human squamous carcinoma named NUT midline carcinoma (25). Brd4 has also been shown to be essential for the maintenance of acute myeloid leukemia (26). Interestingly, small molecules targeting bromodomains of Brd4 possess strong antitumor activities (26 -28). These small molecules prevent the interaction between the BET bromodomains and acetylated lysine peptides (29). One of these BET protein inhibitors, JQ1, displays a selective and potent antiproliferative effect in NUT midline carcinoma cells in vitro (26 -28, 30). In addition, JQ1 promotes tumor regression in patient-derived xenografts in vivo and is highly effective in a number of hematological malignancies, including acute myeloid leukemia and multiple myeloma (26 -28, 30).
In an effort to determine the potential role of acetylated RelA and the subsequent Brd4 recruitment in Tax-mediated tumorigenesis, we found that Brd4 facilitated Tax-mediated NF-B target gene expression and cancer formation. Blockage of the interaction between Brd4 and RelA with JQ1 effectively inhibited the proliferation of Tax-positive HTLV-1-infected cells and Tax-mediated tumorigenesis by inducing the ubiquitina-tion and degradation of constitutively active nuclear NF-B. Our results suggest possible therapeutic approaches for the treatment of NF-B-driven cancer by targeting the interaction between NF-B and Brd4.
Antibodies-Antibodies against RelA, histone H3, HDAC1, tubulin, and ubiquitin were from Santa Cruz Biotechnology. Antibodies against c-Rel and p50 were from Cell Signaling Technology. Antibody against acetylated lysine 310 of RelA was from Abcam, and antibody against Brd4 was from Bethyl Laboratories. Anti-T7 antibody and antibody-conjugated agarose beads were from EMD.
Transient Transfection, Luciferase Reporter, and Immunoprecipitation Assay-HeLa cells were transfected with FuGENE6 (Promega). HEK293T cells were transfected using the calcium phosphate transfection method with various plasmids and luciferase reporters. Firefly and Renilla luciferase activities were measured with the Dual-Luciferase assay system from Promega. Immunoprecipitation and immunoblotting were performed as described previously (19).
Quantitative Real-time PCR Analysis-HeLa or C8166 cells were treated with DMSO or JQ1 for 24 h, and the total RNA was extracted using an RNeasy Mini kit (Qiagen). Complementary DNA was synthesized with an Omniscript RT kit (Qiagen). Quantitative real-time PCR was performed using a Qiagen SYBR Green PCR kit with a 7300 real-time PCR system (ABI). PCR primers for various target genes were purchased from Qiagen. Samples were normalized using the housekeeping gene ␤-actin or GAPDH.
Proliferation Assay-Cell proliferation was determined using a CellTiter 96 Aqueous One Solution kit (Promega). Briefly, cells were plated at a density of 1000 cells/well (for Rat-1-Tax cells) or 2000 cells/well (for HTLV-1-infected cells) in a 96-well plate and treated with DMSO or JQ1 for various times. CellTiter 96 Aqueous solution was added to the cells and incubated for 1-4 h. The absorbance was measured at 490 nm with a microplate reader.
Anchorage-independent Colony Formation Assay-Rat-1-Tax cells were seeded at a density of 1.5 ϫ 10 4 cells/well in complete medium containing 0.3% Difco noble agar (BD Biosciences) on a precoated 6-well plate with 0.6% agar in complete medium. Complete medium containing DMSO or JQ1 was added to the cells twice a week. Colony growth was scored after 25 days of cell incubation.
Cell Cycle Analysis-C8166 cells were treated with either DMSO or JQ1 (2 M) for 24 h before cell cycle analysis. Cells were washed with PBS twice and fixed with 70% ethanol at Ϫ20°C overnight. The cells were pelleted, washed with PBS, and stained at 37°C for 20 min with propidium iodide staining buffer (0.2 mg/ml RNase A, 0.02 mg/ml propidium iodide, and 0.1% (v/v) Triton X-100 in PBS). Samples were subjected to flow cytometry by a BD Biosciences LSR II, and the data were analyzed using FCS Express 4 (De Novo Software).
Annexin V/Propidium Iodide Staining-C8166 cells were treated with DMSO or JQ1 (2 M) for 24 h. The apoptotic cells were measured using a FITC Annexin V Apoptosis Detection Kit I (BD Pharmingen) following the protocols of the manufacturer. Cells were analyzed using a BD Biosciences LSR II. The results were generated using FCS Express 4 (De Novo Software).
In Vivo Tumorigenicity Assays-Six-week-old nude mice (Harlan Laboratories) were implanted subcutaneously with Rat-1-Tax cells (1 ϫ 10 6 ) for 10 days. Mice bearing Rat-1-Tax xenografts were assigned into two groups. Four mice were treated with control vehicles, and six mice were treated with JQ1 daily for 19 days with a dosage of 50 mg per kg of animal weight (mpk). Body weight and tumor volume were measured twice a week. The tumor volume was calculated from caliper measurements with the formula W 2 ϫ L ϫ 0.52. Mice were killed and dissected for tumor evaluation at day 19.

RESULTS
Tax Induces the Binding of Brd4 to Acetylated RelA at Lysine 310-Emerging evidence suggests that acetylation of RelA is critical for the transcriptional activation of NF-B in response to various stimuli, including proinflammatory cytokines and bacterial infection (31). Furthermore, acetylation of lysine 310 contributes to maintain constitutive NF-B activity in tumors (21). Because HTLV-1 Tax induces sustained NF-B activation and NF-B target gene expression (32), we first investigated whether Tax stimulated the acetylation of RelA. When Tax was coexpressed with RelA in HEK293T cells, we observed that Tax enhanced the acetylated lysine 310 of RelA (Fig. 1A). In contrast, Tax failed to enhance the acetylation-deficient RelA-K310R mutant, where lysine 310 was mutated to arginine (Arg) (Fig. 1A), suggesting that Tax stimulates the acetylation of RelA at lysine 310. We next investigated whether Tax-induced acetylation of lysine 310 is important for Tax-mediated NF-B target gene expression. We stably expressed Tax in RelA-deficient mouse embryonic fibroblasts (MEFs) reconstituted with WT RelA or RelA-K310R and examined Tax-induced NF-B target gene expression in these cells. In WT RelA-reconstituted MEFs, Tax induced the expression of IL-1␣ and IL-6 ( Fig. 1B). However, the expression of these NF-B target genes was compromised in the RelA-K310R-reconstituted MEFs (Fig. 1B). The reduced expression of IL-1␣ and IL-6 likely reflects the reduced NF-B transactivation in these cells because Tax-induced 5X-B-luciferase reporter activity was similarly reduced in RelA-K310R-reconstituted cells compared with WT-RelA-reconstituted cells (Fig. 1C). These data suggest that acetylation of lysine 310 is important for Tax-mediated NF-B target gene expression.
We next determined whether Tax induced the interaction between acetylated RelA and Brd4 because our recent studies demonstrated that acetylation of lysine 310 of RelA recruits Brd4 to facilitate the transcriptional activation of NF-B (19). When the interaction of Brd4 and RelA was examined by cotransfecting HEK293T cells with the expression vectors for Brd4 and RelA, we found that Brd4 slightly coimmunoprecipitated RelA in the absence of Tax (Fig. 1D, lane 3). Coexpression of Tax, which induces the acetylation of RelA (Fig. 1A), significantly enhanced the coimmunoprecipitated RelA and acetylated RelA by Brd4 (D, lane 4). Conversely, Tax barely induced the interaction between Brd4 and RelA-K310R, which cannot be acetylated (Fig. 1D, lane 6), indicating that Tax stimulates the interaction of Brd4 and acetylated RelA. The two bromodomains of Brd4 are essential for this interaction because the bromodomain deletion mutant of Brd4 failed to coimmunoprecipitate RelA, even in the presence of Tax (Fig. 1E).
Recent studies demonstrate that the small-molecule JQ1 specifically binds to the asparagines of the bromodomains of Brd4 and prevents the binding of acetylated histone peptides to the binding pocket (27). Structural analysis of the bromodomains of Brd4 also reveals that the highly conserved asparagines are critically involved in the direct interaction with acetylated lysine 310 of RelA (33,34), suggesting that JQ1 might interfere with the binding of acetylated RelA to the bromodomains of Brd4. When the effect of JQ1 on Tax-induced interaction between Brd4 and acetylated RelA was assessed, we found that the interaction was reduced in the presence of JQ1 in a dosedependent fashion (Fig. 1F, lanes 5 and 6). Next, we investigated the effect of JQ1 on the physical interaction of endogenous RelA and Brd4 in Tax-expressing HTLV-1-infected cells. C8166 cells were treated with JQ1 for 24 h, and the nuclear extracts from C8166 cells were immunoprecipitated with IgG or anti-Brd4 antibodies and immunoblotted for the associated RelA. RelA was coimmunoprecipitated with Brd4 when anti-Brd4 antibodies, but not the IgG control, were used to immunoprecipitate Brd4 in the absence of JQ1 (Fig. 1G), indicating that endogenous Brd4 interacts with RelA in the nucleus of C8166 cells. However, the treatment of cells with JQ1 significantly reduced the interaction of endogenous RelA and Brd4 (Fig. 1G). Together, these data suggest that JQ1 interferes with the interaction between RelA and Brd4.
Brd4 Is Required for Tax-mediated NF-B Target Gene Expression and Cell Proliferation-Because JQ1 effectively inhibits the functions of BET family proteins, including Brd4 (27,35), and Brd4 facilitates the transcription activation of NF-B target genes (19,34), we next accessed the possibility that Brd4 might be involved in Tax-mediated NF-B target gene expression. We first examined the effect of Brd4 on Taxinduced NF-B activation using different B-luciferase reporters, including 5XB-Luc, IL-8-Luc, and E-selectin-Luc. Taxmediated transcriptional activation of NF-B was enhanced with cotransfected Brd4 in three different reporters ( Fig. 2A). In contrast, the bromodomain deletion mutants of Brd4 failed to coactivate NF-B activity ( Fig. 2A). These results suggest that Brd4 functions as a coactivator for Tax-mediated NF-B activation and that the bromodomains of Brd4 are essential for the coactivation function of Brd4.
To further demonstrate that Brd4 is a coactivator for Tax  Reconstituted WT or K310R RelA-deficient MEFs stably expressing Tax were transiently transfected with the 5XB reporter. Luciferase (Luc) activities were measured 36 h after transfection, and relative NF-B activity was calculated between reconstituted WT or K310R RelA-deficient MEFs. D, Tax induces the binding of Brd4 to acetylated RelA. HEK293T cells were transfected with different plasmids as indicated. Whole-cell extracts were subjected to immunoprecipitation with anti-Myc antibodies, followed by immunoblotting the Brd4 immunoprecipitates with anti-T7 or anti-acetylated K310 RelA antibodies. E, the bromodomains of Brd4 are important for the Tax-induced interaction between RelA and Brd4. HEK293T cells were transfected with different plasmids as indicated. Whole-cell extracts were subjected to immunoprecipitation with anti-Brd4 antibodies, followed by immunoblotting the Brd4 immunoprecipitates with anti-T7 antibodies. F, JQ1 attenuates the Tax-induced interaction between acetylated RelA and Brd4. HEK293T cells were transfected with different plasmids as indicated. Whole-cell extracts were incubated with increasing doses of JQ1 and subjected to immunoprecipitation and immunoblotting as in D. G, JQ1 blocks the interaction of endogenous RelA and Brd4. C8166 cells were treated with vehicle or 10 M JQ1 for 24 h. Nuclear extracts (NE) were prepared and subjected to immunoprecipitation with IgG or anti-Brd4 antibodies. Brd4-associated RelA was detected by immunoblotting the immunoprecipitates with anti-RelA antibodies. DECEMBER 13, 2013 • VOLUME 288 • NUMBER 50 NF-B target genes, E-selectin and IL-13, in Jurkat-Tet-o-Tax cells expressing control shRNA (Fig. 2B). However, in Brd4 shRNA-expressing Jurkat-Tet-o-Tax cells, Tax-induced mRNA expression of both E-selectin and IL-13 was compromised, although the expression of Tax was not affected (Fig. 2B). Depletion of Brd4 also reduced the expression of NF-B target genes in Rat-1-Tax cells (Fig. 2C). These data suggest that Brd4 is essential for Tax-induced NF-B target gene expression. To further confirm this, we knocked down the expression of Brd4 in Tax-positive C8166 cells using shRNA and measured Taxinduced NF-B target gene expression. Depletion of Brd4 in C8166 cells reduced Tax-induced expression of CSF2, IL-13, and OX40 mRNA (Fig. 2C). A similar reduced expression of NF-B target genes was also found in Brd4 depleted MT4 cells (data not shown). These data indicate that Brd4 is essential for Tax-mediated NF-B target gene expression. Finally, we examined the effect of Brd4 knockdown on the proliferation of Tax-positive HTLV-1-infected cells. We generated a Dox-inducible Brd4 shRNA C8166 cell line. When the Brd4 shRNA was induced in the presence of Dox, the proliferation of C8166 was impaired by the depletion of Brd4 (Fig. 2E). Together, these data suggest that Brd4 is essential for Tax-mediated NF-B target gene expression and cell proliferation.

BET Inhibition Suppresses NF-B Activation in ATL
JQ1 Inhibits the Proliferation of Tax-positive HTLV-1-infected Cells-Because JQ1 effectively inhibits Tax-induced binding of Brd4 to acetylated RelA and because Brd4 is essential for Tax-mediated NF-B activation and cell proliferation (Figs. 1 and 2), we next explore the potential effect of JQ1 on Tax-mediated tumorigenesis. We first examined the effect of JQ1 on the proliferation of Tax-transformed rat fibroblast Rat-1 cells (Rat-1-Tax). The growth of Rat-1-Tax cells was suppressed with different concentrations of JQ1, ranging from 0.1-2 M (Fig. 3A). 0.5 M JQ1 effectively blocked the growth of Rat-1-Tax cells (Fig. 3A).
Next we examined the effect of JQ1 on the proliferation of Tax-positive HTLV-1-infected MT4, C8166, and SLB1 cells. Human PBMCs were used as control cells for non-HTLV-1infected cells. We treated PBMCs and HTLV-1-infected cells with various concentrations of JQ1 for 72 h and measured cell   proliferation. JQ1 had little effect on the proliferation of PBMCs (Fig. 3B). However, the growth of all three Tax-positive HTLV-1-infected cells was decreased with different degrees of sensitivity after 72 h of JQ1 treatment (Fig. 3B). MT4 and SLB1 cells were more sensitive to JQ1 than C8166 cells.
To further evaluate the effect of JQ1 on the proliferation of Tax-positive HTLV-1 cells, we treated C8166 and MT4 cells with different concentrations of JQ1 for a longer period of time and measured the half-maximal inhibitory concentration (IC 50 ) of JQ1 on these cells. C8166 cells were less sensitive to JQ1 before 3 days. The sensitivity of C8166 cells to JQ1 was increased significantly after 3 days, with an estimated IC 50 value of around 0.5 M after 5 days of treatment (Fig. 3C, left panel). MT4 cells were more sensitive to JQ1, with an estimated IC 50 value of less than 0.1 M after 5 days of treatment (Fig. 3C, right  panel). These data further demonstrate that JQ1 effectively inhibits the proliferation of Tax-positive HTLV-1-infected cells.
Next we examined the effect of JQ1 on cell cycle progression and apoptosis in HTLV-1-infected cells. When treated with JQ1 for 24 h, C8166 cells displayed significant changes in cell cycle progression, with a reduced proportion of cells in S phase and a concurrently increased proportion of cells in G 0 /G 1 phase (Fig. 3D). In addition to cell cycle arrest, the number of C8166 cells undergoing apoptosis also increased after 24-h JQ1 treatment (Fig. 3E). Collectively, these data indicate that JQ1 inhibits proliferation, blocks cell cycle progression, and induces apoptosis in HTLV-1-infected cells.
JQ1 Suppresses Tax-mediated Tumorigenesis-We next sought to investigate the effect of JQ1 on Tax-mediated tumorigenesis using Rat-1-Tax cells, which have been frequently used to study HTLV-1 Tax-mediated leukemogenesis (2). In line with the data from cell proliferation (Fig. 3A), JQ1 inhibited the potential of Rat-1-Tax cells to form colonies in an anchorageindependent growth assay (Fig. 4A). Both the size and number of the colonies formed were significantly reduced by JQ1 in a dose-dependent fashion (Fig. 4, A and B), suggesting that JQ1 suppresses the transformation potential of Rat-1-Tax cells.
We then determined the effect of JQ1 on tumor formation and growth of Rat-1-Tax cells in a mouse xenograft model. Tumors were formed 2 weeks after subcutaneous inoculation of nude mice with Rat-1-Tax cells. Mice bearing tumors were then randomized to receive vehicle (n ϭ 4) or JQ1 treatment (n ϭ 6) daily for 19 days. Compared with mice treated with control vehicles, mice treated with JQ1 produced smaller tumors in nude mice (Fig. 4C). The volume of tumors from JQ1-treated animals was markedly reduced, ϳ69% smaller, compared with vehicle-treated mice (Fig. 4C). The average weight of the tumors formed in JQ1-treated mice was lower than in those treated with control vehicles (Fig. 4D). We also noticed that JQ1 was well tolerated at the given dose, without any signs of toxicity or body weight loss to mice. These results support the notion that JQ1 inhibits the oncogenic activity of HTLV-1 Tax.
JQ1 Down-regulates HTLV-1 Tax-mediated NF-B Target Gene Expression-Constitutive activation of NF-B has been shown to be essential for the transformation of rat fibroblasts by Tax (14). Because JQ1 inhibits proliferation and tumorigenesis of Rat-1-Tax cells (Figs. 3 and 4), we next examined the effect of JQ1 on Tax-induced activation of NF-B. HEK293T cells were transfected with the Tax expression vector and a reporter plasmid containing five copies of the B binding sites (5X-B-Luc). As expected, Tax activated 5XB luciferase reporter activity (Fig. 5A). However, treatment of cells with JQ1 attenuated Taxinduced NF-B activation (Fig. 5A, left panel). Similarly, JQ1 inhibited the Tax-induced activation of the IL-8-luciferase reporter, which contains the B response element from the promoter of the IL-8 gene, in a dose-dependent manner (Fig.  5A, right panel). These data suggest that JQ1 inhibits the Taxinduced transcriptional activation of NF-B.
We next assessed the effect of JQ1 on Tax-mediated expression of NF-B target genes. When HeLa cells were transfected with the Tax expression vector followed by the treatment of JQ1, we found that Tax-induced expression of a subset of NF-B target genes, including IL-2Ra, IL-6, and CSF2, was down-regulated by JQ1 (Fig. 5B). When Rat-1-Tax cells were treated with JQ1, the expression of a subset of NF-B target genes, including IL-2Ra, IL-6, and TNF-␣, was similarly downregulated (Fig. 5C). These data indicate that JQ1 inhibits the transcriptional activation of NF-B induced by Tax.
To further confirm the inhibitory effect of JQ1 on Tax-mediated NF-B target gene expression, we examined the expression of Tax  cells, JQ1 potently suppressed the transcription expression of a subset of NF-B target genes in C8166 cells. These genes include cytokines (IL-8, IL-6, IL-13, and CSF2) and growth factor (TGF-␤) and cytokine receptors (IL-2R␣ and OX40) (Fig.  5D). JQ1 had the strongest inhibitory effect on the expression of IL-2R␣ and CSF2 (Fig. 5D). In contrast, JQ1 had little effect on the expression of NFKB2 and BIRC5 (Fig. 5D), which have also been reported to be Tax-induced NF-B target genes (8). Collectively, these data demonstrate that JQ1 selectively inhibits a subset of Tax-mediated NF-B target gene expression. RelA decreased with the increased concentration of JQ1 (Fig.  6A). However, the levels of cytoplasmic RelA were not altered by JQ1 (Fig. 6A). The effect of JQ1 on nuclear RelA seems to be specific because both the nuclear and the cytoplasmic levels of other two NF-B family proteins, p50 and c-Rel, remained unchanged (Fig. 6A). Similarly, JQ1 decreased the nuclear levels of RelA, but not c-Rel, and p50 in MT4 cells (Fig. 6A). These data indicate that JQ1 regulates the levels of nuclear, but not the cytoplasmic portion, of RelA in Tax-positive HTLV-1-infected cells.
We next explored the possibility that the reduced nuclear RelA might result from proteasome-mediated degradation. We added the proteasome inhibitor MG-132 into JQ1-treated C8166 cells and found that MG-132 reversed the decreased nuclear levels of RelA (Fig. 6B). These data suggest that the reduced expression of RelA in the nucleus is due to proteasome-mediated degradation. Consistently, the transcription of RelA was not affected by JQ1 because RelA mRNA remained at a similar level with or without JQ1 treatment (Fig. 6C).
Because ubiquitination is a prerequisite event for proteasome-mediated protein degradation, we next determined whether JQ1 promoted the ubiquitination of nuclear RelA.
Nuclear RelA was moderately ubiquitinated in the absence of JQ1 in C8166 cells (Fig. 6D). However, JQ1 treatment enhanced the ubiquitination of nuclear RelA (Fig. 6D), indicating that JQ1 promotes RelA ubiquitination. Consistent with the finding that nuclear c-Rel levels were not affected by JQ1 (Fig. 6A), ubiquitination of c-Rel was not affected by JQ1 (E).

DISCUSSION
Reversible acetylation of RelA at lysine 310 controls the functions of NF-B under physiopathological conditions by facilitating the expression of NF-B-mediated inflammatory genes and maintaining the sustained NF-B activation in cancers (21, 31, 36 -38). In this study, we explored the potential role of acetylation of RelA and the subsequent binding of Brd4 to acetylated RelA in Tax-mediated NF-B activation and tumorigenesis. We found that binding of Brd4 to acetylated RelA facilitated Tax-mediated NF-B target gene expression and was involved in the proliferation of Tax-positive HTLV-1-infected cells. Blocking the interaction of Brd4 and the acetylated RelA with the BET protein inhibitor JQ1 suppressed Tax-mediated NF-B activation and tumorigenesis. We have shown previously that Brd4 binds to acetylated lysine 310 of RelA, facilitating the transcription of a subset of NF-B target genes in response to proinflammatory cytokines (19). The binding of Brd4 to acetylated lysine 310 of RelA appears to be similarly important for Tax-induced transcriptional activation of NF-B because Brd4 coactivates Tax-induced NF-B activation in a bromodomain-dependent manner (Fig. 2), and depletion of Brd4 or mutation of lysine 310 impairs Tax-induced NF-B target gene expression (Figs. 1B and 2). It has to be noted that the mutation of lysine 310 to arginine only partially reduced the Tax-mediated transcriptional activity of NF-B (Fig. 1, B and C), raising the possibility that other modifications of lysine 310 could also be involved in Tax-mediated NF-B activation because lysine can be modified by methylation and ubiquitination, which might counteract the effect of acetylation. It is also possible that the acetylation of other lysines of RelA also contributes to Tax-mediated transcriptional activation of NF-B. By preventing the Tax-induced binding of Brd4 to acetylated RelA (Fig. 1), JQ1 down-regulates Tax-induced NF-B target gene expression (Fig. 5). Additionally, by disrupting the interaction between Brd4 and RelA, JQ1 induces the ubiquitination and degradation of RelA (Fig. 6). As such, the decreased nuclear RelA also accounts for the downregulation of NF-B target gene expression (Fig. 5).
JQ1 suppresses the proliferation, blocks cell cycle progression, and induces apoptosis of Tax-expressing cells (Fig. 3). Additionally, JQ1 suppresses the tumorigenesis of Tax-expressing Rat-1 cells (Fig. 4). These effects are likely due to the down-regulation of certain NF-B target genes because NF-B regulates the expression of many genes involved in cell growth, the cell cycle, and apoptosis in HTLV-1-infected cells (8). It has been well documented that the ability of Tax to transform rat fibroblasts and human T lymphocytes is largely associated with its ability to activate the expression of NF-B-dependent genes (10,40). JQ1 inhibits the expression of cytokines (IL-6, IL-8, and CSF2), growth factor (TGF-␤), and cytokine receptors (IL-2R␣ and OX40) (Fig. 5). All of these have been shown to be Taxinduced NF-B target genes and have been suggested to contribute to the proliferation of HTLV-1-infected cells (8). Of note, not all Tax-induced NF-B target genes are inhibited by JQ1. For example, the expression of NFKB2 and BIRC5 are not affected by JQ1 (Fig. 5), indicating that JQ1 selectively inhibits Tax-mediated NF-B target gene expression. This selectivity might reflect the differential requirement of Brd4 for NF-B target gene expression (31).
Interestingly, of all the identified Tax-induced NF-B target genes inhibited by JQ1, we found that IL-2R␣ showed the greatest sensitivity to JQ1, whereas other target genes displayed various levels of sensitivity (Fig. 5). 0.1 M of JQ1 was sufficient to abolish Tax-induced expression of IL-2R␣ in HeLa cells and the expression of endogenous IL-2R␣ in HTLV-1-infected cells (Fig. 5, B and D). IL-2R␣ is known to be constitutively expressed in the malignant T cells of patients with HTLV-1-induced ATL but not expressed in resting normal T cells, representing a hallmark of HTLV-1-infecetd cells (41). The expression of IL-2R␣ is regulated by Tax-induced activation of NF-B, and IL-2R␣ plays a pivotal role in the proliferation of HTLV-1-transformed cells (42)(43)(44). Therefore, the markedly down-regulated IL-2R␣ by JQ1 might account largely for the suppressed proliferation of Tax-transformed rat cells and Tax-positive HTLV-1-infected cells in response to JQ1 treatment (Fig. 4). JQ1 has been shown to down-regulate a variety of cellular proteins, including c-Myc, IL-7R, and FOSL1, to suppress the proliferation of a variety of cancer cells (27,45,46). The anti-tumor activity of JQ1 in different cancers might result from the inhibition of one specific protein or in combination with the inhibition of several other proteins. The suppressed HTLV-1 cell proliferation by JQ1 might also reflect a combinational effect from the down-regulation of multiple genes (Fig. 5).
Recent studies reveal that JQ1 possesses antitumor activity in a broad spectrum of cancers, including hematological malignancies and solid tumors (26 -28, 30, 46, 47). JQ1 suppresses the proliferation of these cancer cells by down-regulating the expression of different cellular genes involved in cell proliferation. For example, JQ1 down-regulates the transcription of the oncogene c-Myc, the first identified JQ1 target (26 -28, 30). However, not all c-Myc-overexpressing cancer cell lines are sensitive to JQ1, indicating that additional targets, rather than c-Myc, are responsible for the antitumor activity of JQ1 (30). In addition to c-Myc, JQ1 inhibits the transcription of IL-7R to suppress the JAK-STAT signaling-mediated proliferation of B cell acute lymphoblastic leukemia cells (45). JQ1 has also been shown to suppress lung cancer cell proliferation by down-regulating FOSL1 expression (46). Of all these cellular targets identified so far, JQ1 acts at the transcription level by depleting the Brd4 from the promoters of these genes (26 -28, 30, 45, 46). In this study, we found that JQ1 inhibits the binding of Brd4 to acetylated RelA (Fig. 1, F and G), but the transcription of RelA was not affected (Fig. 6C). Instead, JQ1 induces the ubiquitination and degradation of RelA (Fig. 6). It appears that JQ1 could regulate cellular protein functions by various mechanisms either at the transcription level or at the cellular protein level. It remains unclear how JQ1 stimulates the ubiquitination and degradation of RelA. One possibility is that the removal of Brd4 from acetylated RelA would expose the acetylated lysine, which is then subject to deacetylation, resulting in increased ubiquitination and degradation of RelA. In supporting this, acetylation of lysine 310 and the binding of Brd4 to acetylated lysine 310 have been shown to inhibit RelA ubiquitination (34,48).
ATL has a poor prognosis largely because of limited and ineffective treatment options. Many conventional and combinatorial anticancer therapies have been used in clinical trials with limited degrees of success, and novel approaches targeting the constitutively active IL-2R␣ and NF-B have been explored (49,50). Small molecules or agents that inhibit the NF-B signaling pathway have been identified as potent anticancer agents for a variety of cancers, including ATL (51,52). The majority of these NF-B inhibitors target the proximal upstream signaling molecules and prevent the nuclear translocation of NF-B (51, 53). Our results revel that JQ1 represents a novel type of NF-B inhibitor that specifically targets the epigenetic regulators, such as Brd4, triggering the ubiquitination and degradation of nuclear NF-B. Targeting epigenetic regulators has emerged as a new therapeutic strategy for cancer therapy, and JQ1 has been demonstrated to be one of the effective agents (39). The prominent effect of JQ1 on the expression of IL-2R␣ and the prolif-eration of HTLV-1-infected cells (Figs. 3 and 5) raises an intriguing possibility that JQ1, alone or in combination with other therapies such as antibody therapy, might provide an alternative therapeutic application against ATL. In this study, we focused primarily on Tax-expressing HTLV-1-infected cells. It has to be noted that Tax is essential for the early stage of transformation of HTLV-1-infected T cells and that Tax expression is lost in about 60% of all ATLs during the late stages of leukemogenesis (9,10). However, Tax-negative HTLV-1 infected cells also contain constitutively active NF-B (9, 10). We found that JQ1 also suppresses the proliferation of some Tax-negative HTLV-1 cells (data not shown), suggesting that JQ1 might be a broadly applicable therapy in the prevention and treatment of HTLV-1-induced ATL.