HTLV-1 bZIP Factor Suppresses Apoptosis by Attenuating the Function of FoxO3a and Altering Its Localization

Human T-cell leukemia virus type 1 (HTLV-1) is estimated to infect 10 to 20 million people in the world (1). This virus causes not only a neoplastic disease of CD4⁺ T cells, adult T-cell leukemia (ATL), but also chronic inflammatory diseases of the central nervous system, lung, or skin (2). HTLV-1 can be transmitted efficiently in a cell-to-cell fashion (3, 4), whereas free virus shows poor infectivity (5, 6), and virions are not detected in infected individuals. To increase the number of infected cells and facilitate transmission, HTLV-1 increases its copy number primarily by triggering the proliferation of infected cells, replicating within the host genome instead of undergoing viral replication (7, 8). Thus, HTLV-1 promotes proliferation and suppresses apoptosis of infected cells via complex interactions of viral proteins with host factors.

Among the viral genes encoded in HTLV-1, the tax gene has been extensively studied. Tax can activate various signal pathways like NF-κB, AP-1, and SRF (9). However, Tax expression is frequently undetectable in ATL cases. Importantly, the nonsense mutations in the tax gene are often observed in not only ATL cases but also infected cells of asymptomatic HTLV-1 carriers (10). These findings suggest that other mechanisms suppress the apoptosis of HTLV-1–infected cells in the absence of Tax expression (2). We have reported that the HTLV-1 bZIP factor (HBZ) gene is expressed in all ATL cases (11). Furthermore, HBZ promotes the proliferation of T cells and induces development of T-cell lymphomas and inflammatory diseases in transgenic mice (12). Therefore, we speculated that HBZ might also influence apoptosis.

There are two major pathways for apoptosis: the extrinsic and intrinsic apoptotic pathways, which are mediated by Fas and Bim, respectively (13). ATL cells are known to express high levels of Fas antigen, and are susceptible to Fas-mediated signaling (14). However, FasL expression is suppressed in ATL cells by silencing of the early growth response 3 (EGR3) gene transcription, a phenomenon that enables ATL cells to escape activation-induced cell death (15). In addition, Tax increases expression of c-FLIP, which confers resistance to Fas-mediated apoptosis (16, 17). Furthermore, activation of NF-κB by Tax also enables HTLV-1–infected cells to be resistant to apoptosis (18). To date, the effects of HTLV-1 infection on Bim-mediated apoptosis remain unknown.

In this study, we analyzed transcriptional changes induced by HBZ expression in T cells, and found that transcription of a proapoptotic gene, Bim, was hindered by HBZ. This suppression led to decreased activation-induced cell death. We found that HBZ suppressed Bim transcription by targeting FoxO3a, a critical transcription factor for the Bim and FasL gene. In some ATL cell lines and ATL cases, the Bim gene transcription was also silenced by epigenetic mechanisms, but this phenomenon seemed to be secondary to HBZ-mediated suppression of transcription. Thus, it is suggested that HBZ suppresses both intrinsic and extrinsic apoptotic pathways and contributes to the proliferation of ATL cells.

HTLV-1 immortalized cell lines (MT-4), ATL cell lines (ED, TL-Om1, and MT-1), T-cell lines not infected with HTLV-1 (Jurkat, SupT1, and CCRF-CEM) were cultured in RPMI 1640 medium supplemented with 10% FBS and antibiotics at 37°C under a 5% CO₂ atmosphere. Jurkat cells stably expressing spliced form of HBZ (sHBZ), Jurkat-HBZ, were maintained as described previously (19). To construct CCRF-CEM cells stably expressing HBZ, CEM-HBZ, the coding sequence of HBZ was subcloned into pME18Sneo vector and then the expression vector or its empty vector were transfected into CCRF-CEM cells by using Neon (Invitrogen) according to the manufacturer's instructions. Stable transfectants were selected in G418 (1 mg/mL). 293T cells were cultured in Dulbecco's Modified Eagle Medium supplemented with 10% FBS and antibiotics and when 293FT cells were cultured, 500 μg/mL G418 was added. Fas-blocking antibody was purchased from Alexis.

This study was conducted according to the principles expressed in the Declaration of Helsinki. The study was approved by the Institutional Review Board of Kyoto University (G204). All patients provided written informed consent for the collection of samples and subsequent analysis.

Wild-type form of FoxO3a was generated by PCR amplification using Jurkat cDNA library and constitutively active form of FoxO3a (FoxO3aAAA) was also generated by PCR amplification with mutated primers (20). These PCR fragments were then subcloned into pCMV-Tag2B vector and pIRES-hrGFP-1a (Stratagene). The vectors encoding the myc-His-tagged form of HBZ and its mutants used in this study have been described previously (19, 21). We modified pLKO.1-EGFP vector for delivery of anti-FoxO3 short hairpin RNAs (shRNA) to Jurkat, Jurkat-control, and Jurkat-HBZ. shRNA sequence used was 5′-GCACAACCTGTCCTGCATAG-3′. The 6xDBE-Luc construct that contains six FOXO-binding sites known as DAF-16 binding elements (DBE) was kindly provided by Dr. Furuyama (Kagawa Prefectural University of Health Sciences, Kagawa, Japan) and the backbone of this vector was pGL3-basic (Promega; ref. 22).

Jurkat cells were transfected with 0.2 μg/well of luciferase reporter plasmid, 1 ng/well of Renilla luciferase control vector (phRL-TK), 0.2 μg/well of FoxO3aAAA expression plasmid or its empty vector, and 0.6 μg/well of HBZ expression plasmid or its empty vector with caspase inhibitor Z-VAD-FMK (MBL). Plasmids were transfected using Neon (Invitrogen) according to the manufacturer's instructions. After 24 hours, cells were collected and luciferase activities were measured using the Dual-Luciferase Reporter Assay (Promega). Relative luciferase activity was calculated as the ratio of firefly to Renilla luciferase activity. Three independent experiments, each with triplicate transfections, were performed and typical results are shown.

Jurkat-control and Jurkat-HBZ were stimulated with phorbol myristate acetate (PMA; 50 ng/mL) and ionomycin (Io; 1 μg/mL) for 9 hours. After the stimulation, cells were collected and total RNA was isolated using TRIzol Reagent (Invitrogen) according to the manufacturer's instructions. We then digested DNA using deoxyribonuclease I (Invitrogen) and cleaned up RNA using RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. We then synthesized cDNA and performed microarray processing according to the GeneChip Expression Analysis Technical Manual (Affymetrix). All data were analyzed by using GeneSpring GX (Agilent Technologies). The microarray data related to this article have been submitted to the Gene Expression Omnibus under the accession number GSE48029.

293FT cells were transfected with expression vectors using Lipofectamine LTX (Invitrogen) or TransIT (TaKaRa). Twenty-four hours after transfection, cells were reseeded on the poly-L-lysine–coated glass (Matsunami Glass Ind., Ltd.) or poly-D-lysine (Sigma)–coated glass. Twenty-four hours after the reseeding, cells were fixed with 4% paraformaldehyde for 15 minutes and permeabilized with 0.2% Triton X-100 for 15 minutes, and blocked by incubation in 5% BSA/PBS for 30 minutes. For immunostaining, the cells were incubated with anti-Foxo3a, anti-p-Foxo3a (Cell Signaling Technology), Cy3-conjugated anti-c-Myc (Sigma) or biotinylated anti-FLAG (Sigma) antibodies for 1 hour or in case of observation of endogenous expression, cells were incubated overnight at 4°C. Primary antibodies were visualized by incubating the cells with AlexaFluor 488–conjugated goat anti-rabbit immunoglobulin G (IgG) antibody (Invitrogen) or AlexaFluor 488–conjugated streptavidin (Invitrogen). Nuclei were stained and mounted with ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). To concentrate nonadherent cells onto a microscope slide, CytoFuge (StatSpin) was used. Fixation and blocking were performed as described earlier.

Apoptotic cells were routinely identified by Annexin V-APC (eBioscience) or phycoerythrin (PE) or fluorescein isothiocyanate (FITC; BioVision) -staining according to the manufacturer's instructions and analyzed with a flow cytometer (BD FACSCanto II; BD Biosciences). Data files were analyzed by using FlowJo software (TreeStar).

Total RNA was isolated for the analysis using TRIzol reagent. RNA was treated with DNase I to eliminate the genomic DNA. Reverse transcription was performed using random primer and SuperScript III Reverse Transcriptase (Invitrogen). CD25⁻CD4⁺ cells from healthy donor were obtained by using human CD4 T Lymphocyte Enrichment kit (BD Pharmingen). Then, cells were stimulated with PMA/Io for 9 hours, RNA was isolated, and reverse transcription was performed as described earlier. cDNA products were analyzed by real-time PCR using the Taqman Universal PCR Master Mix (PE Applied Biosystems) or FastStart Universal SYBR Green Master (Roche) and Applied Biosystems StepOnePlus Real-Time PCR System according to the manufacturer's instructions. Specific primers and Taqman probes for the Bim gene, FasL gene, and GAPDH internal control gene were purchased from Applied Biosystems. Primer sequences for the HBZ gene and GAPDH gene used for the evaluation of the knockdown efficiency in MT-1 cells have been described previously (11, 23). Primer sequences for the HBZ gene used for another experiment to evaluate the HBZ expression in Jurkat-HBZ, MT-1, TL-Om1, and ED cells were 5′-ATGGCGGCCTCAGGGCTGTT-3′ and 5′-GCGGCTTTCCTCTTCTAAGG-3′. Primer sequences for the FoxO3a gene used were 5′-ACAAACGGCTCACTCTGTCCCAG-3′ and 5′-AGCTCTTGCCAGTTCCCTCATTCTG-3′. All amplifications were conducted in triplicates. The relative quantification was calculated according to the method described in Applied Biosystems ABI prism 7700 SDS User Bulletin #2.

Chromatin immunoprecipitation (ChIP) assay was performed according to the protocol recommended by Millipore. Cells were fixed with 1% formaldehyde for 10 minutes at room temperature, washed twice with ice-cold PBS, treated with SDS-lysis buffer (1% SDS, 50 mmol/L EDTA, and 200 mmol/L Tris-HCl) for 10 minutes on ice and then sonicated. Thereafter, the DNA/protein complexes were immunoprecipitated with antibodies specific for acetylated-Histone H3, acetylated-Histone H4, dimethylated-Histone H3 (Lys4), RNA polymerase II clone CTD4H8 (Millipore), trimethylated-Histone H3 (Lys27), anti-trimethyl-Histone H3 (Lys9) antibodies (Cell Signaling Technology), or normal rabbit IgG (Santa Cruz Biotechnology) overnight at 4°C. Immune complexes were collected with salmon sperm DNA-protein A and G Sepharose slurry, washed, and eluted with freshly prepared elution buffer (1% SDS, 100 mmol/L NaHCO₃). Protein-DNA complexes were de-cross-linked at 65°C for 4 hours. DNA was purified and subjected to real-time PCR for quantification of the target fragments. Sequences for the primer set are described previously (24, 25). For the evaluation of binding of FoxO3a to the FOXO-binding sites, 293T cells were transfected with 5 μg of 6xDBE-Luc construct, 5 μg of FoxO3aAAA expression plasmid together with or without 5 μg of HBZ plasmid using TransIT in 10-cm dishes. Anti-FLAG (Sigma) antibody was used for the immunoprecipitation. Primers used were 5′-AGTGCAGGTGCCAGAACATT-3′ and 5′-GCCTTATGCAGTTGCTCTCC-3′, which were constructed inside of the pGL3-basic vector. For the evaluation of the DNA-binding capacity of FoxO3a with or without HBZ, expression vectors for the HA-tagged FoxO3a and Flag-tagged HBZ were transiently cotransfected into 293T cells using the TransIT reagent. Twenty-four hours after the transfection, cells were collected and chromatin immunoprecipitation assay was performed as described earlier. For the immunoprecipitation, anti-HA (Sigma) antibody was used. Primers used for Bim gene promoter were 5′-CCACCACTTGATTCTTGCAG-3′ and 5′-TCCAGCGCTAGTCTTCCTTC-3′, which were constructed to contain the FOXO-binding sequence located in intron1. Primers used for FasL gene promoter were 5′-ACGATAGCACCACTGCACTCC-3′ and 5′-GGCTGCAAACCAGTGGAAC-3′, which were also constructed to contain the three FOXO-binding sequences.

Individual PCRs were carried out in triplicate to control for PCR variation and mean C_t values were collected. Fold difference of the antibody-bound fraction (IP) versus a fixed amount of input (In) was calculated as

Then, the fold difference value for a target antibody (t) was subtracted by the nonspecific value derived from mouse or rabbit IgG (t₀):

Sodium bisulfite treatment of genomic DNA was performed as described previously (26). DNA regions were amplified using bisulfite-treated genomic DNA by nested PCR. To amplify promoter region (promoter 1) of Bim, primers used in the first PCR were 5′-TTTAGAGGGAGGAGAGTTTAAAG-3′ and 5′-CCCTACAACCCAACTCTAACTA-3′. Primers for the second PCR were 5′-AGGGTATAGTGAGAGCGTAGG-3′ and 5′-CAACTCTAACTAACGACCCC-3′. For promoter, two primers used in the first PCR were 5′-GTGTGATTGTTTTTTGAGGG-3′ and 5′-AAAATACCCCCAAACAAAATAC-3′. Primers for the second PCR were 5′-GCGGATTTAGTTGTAGATTTTG-3′ and 5′-ACTCTTTACCCAAAACAAACTTC-3′. PCR products were purified, cloned into pGEM-T Easy vector (Promega), and sequenced using the ABI PRISM 3130 Genetic Analyzer. For CpG methylation analysis, Web-based bisulfite sequencing analysis tool called QUMA (quantification tool for methylation analysis) was used (27).

Expression vectors for the relevant genes were transiently cotransfected into 293T cells using the TransIT reagent. Forty-eight hours later, cells were collected and coimmunoprecipitation assays were performed as described previously (28). For the analysis of the p-FoxO3a localization, nuclear and cytoplasmic proteins were extracted using Nuclear Complex Co-IP Kit (Active Motif). The proteins were subjected to SDS-PAGE analysis followed by immunoblotting with various antibodies. Antibodies used were anti-p-FoxO3a, anti-α-tubulin (Sigma), anti-FLAG, anti-HA (Sigma), and anti-His (Marine Biological Laboratory).

Lentiviral vector expressing shRNA against HBZ was constructed and recombinant lentivirus was infected as described previously (11). When more than 90% of cells expressed enhanced green fluorescent protein (EGFP), the HBZ and Bim gene expressions were analyzed by real-time PCR.

To determine the effects of HBZ on gene expression, we first performed microarray analysis. Jurkat cells with or without expression of spliced form of HBZ (Jurkat-HBZ and Jurkat-control, respectively) were stimulated with PMA and Io for 9 hours. Gene expression profiles were then analyzed by DNA microarray. Table 1 shows the apoptosis-associated genes that were downregulated or upregulated in stimulated Jurkat-HBZ cells. Transcription of the Bim gene was prominently downregulated in HBZ-expressing Jurkat cells. To confirm the effect of HBZ on the Bim gene expression, we evaluated Bim mRNA levels in Jurkat-control and Jurkat-HBZ cells with or without PMA/Io stimulation using real-time PCR. As reported in the previous studies showing that treatment by PMA/Io or other stimulators induced Bim expression (29, 30), the Bim mRNA level of stimulated Jurkat-control cells was three-times higher than that of unstimulated cells, but that of Jurkat-HBZ cells did not change after stimulation (Fig. 1A). Similarly, increased Bim transcription by stimulation was also inhibited by HBZ in CCRF-CEM cells (Fig. 1A).

It has been reported that Bim plays an important role in activation-induced cell death and T-cell homeostasis (31). Because the earlier data demonstrated that HBZ inhibits stimulation-induced Bim expression, we next investigated whether HBZ inhibits apoptosis in response to PMA/Io stimulation. To test this, Jurkat-control and Jurkat-HBZ were each incubated with or without PMA/Io for 9 hours, and then apoptosis was measured using Annexin V. The percentages of apoptotic cells in Jurkat-control and Jurkat-HBZ were 40.2% and 15% respectively, indicating that HBZ suppressed activation-induced apoptosis (Fig. 1B). We also treated cells with doxorubicin and found that HBZ slightly inhibited doxorubicin-induced apoptosis (Supplementary Fig. S1). Fas-mediated apoptotic pathway might be involved in antiapoptotic effect by HBZ. To assess the effect of Fas-mediated signaling on the activation-induced apoptosis, cells were also treated with or without Fas-blocking antibody (0.5 μg/mL) 30 minutes before the PMA/Io stimulation. The percentage of apoptotic cells without Fas-blocking antibody in Jurkat-control and Jurkat-HBZ were 36.9% and 22.4%, respectively. When cells were treated with Fas-blocking antibody, the percentage of apoptotic cells reduced and those were 24% and 13.2% in Jurkat-control and Jurkat-HBZ, respectively (Fig. 1C). Thus, Fas-blocking antibody partially inhibited apoptosis in Jurkat-HBZ, which indicates that Fas-mediated signals are also implicated in activation-induced cell death. Indeed, we found that the transcription level of FasL was suppressed in stimulated Jurkat-HBZ and CEM-HBZ cells compared with Jurkat-control and CEM-control cells (Fig. 1D), suggesting that downregulation of FasL by HBZ was also associated with inhibition of apoptosis.

We analyzed how HBZ suppresses the expression of Bim and FasL. It has been reported that a forkhead factor, FoxO3a, and p73 are important for the transcription of Bim and FasL (32, 33). FoxO3a and other FOXO family members are phosphorylated by protein kinases such as Akt or SGK on highly conserved serine and threonine residues (especially Thr32, Ser253, and Ser315 in FoxO3a), resulting in impaired DNA-binding activity and increased binding to the chaperone protein, 14-3-3 (20, 34, 35). Newly formed 14-3-3-FOXO complexes are then exported from the nucleus, thereby inhibiting FOXO-dependent transcription of key target genes such as Bim, FasL, and TRAIL (36).

First, we investigated whether FoxO3a is implicated for the activation induced cell death. As shown in Supplementary Fig. S2, the knockdown of FoxO3a resulted in the decreased apoptotic rate in Jurkat-control cells (P < 0.05). Furthermore, inhibition of Foxo3a did not influence activation-induced cell death in Jurkat-HBZ cells, suggesting the inhibitory effect of HBZ on Foxo3a function. To investigate whether HBZ affects FoxO3a function, Jurkat cells were transiently transfected with a plasmid expressing FoxO3aAAA, the constitutive active mutant of FoxO3a, which is no longer phosphorylated by Akt and is localized in the nucleus. The FoxO3aAAA was expressed together with hrGFP using an internal ribosome entry site (IRES; FoxO3aAAA-IREShrGFP). Jurkat cells were transiently transfected with full-length HBZ or its mutants. HBZ has three domains, an activation domain (AD), a central domain (CD), and a basic leucine zipper domain (bZIP; ref. 12). In this study, the deletion mutants (HBZ-ΔAD, HBZ-ΔbZIP, and HBZ-ΔCD) were used. The percentage of FoxO3aAAA induced apoptotic cells in the absence of HBZ was 69.6% while it was suppressed by HBZ (40.6%; P < 0.001; Fig. 2A). We also found that an HBZ mutant without activation domain lacks the activity to inhibit FoxO3aAAA-induced apoptosis (Fig. 2A), indicating the significance of activation domain in suppression of FoxO3a-mediated apoptosis. It has been reported that LXXLL motif in FoxO3a binds to its coactivator CBP/p300 (37). Similarly, HBZ has LXXLL-like motifs located in the NH₂-terminal region, which bind to KIX domain of CBP/p300 (38). We speculated that the LXXLL-like motifs of HBZ might affect FoxO3aAAA function through KIX domain of CBP/p300. An HBZ mutant, which has substitutions in 27th and 28th residues (LL to AA) of LXXLL-like motif, lack the activity to suppress FoxO3aAAA-mediated apoptosis (Fig. 2B), indicating that LXXLL-like motif of HBZ is critical for suppression of FoxO3a-mediated apoptosis.

Next, we analyzed the effect of HBZ on a FoxO3a responsive reporter. As shown in Fig. 2C, HBZ suppressed FoxO3a-mediated transcriptional activity (P < 0.01). To check whether HBZ inhibits DNA binding of FoxO3a, 293T cells were transiently transfected with FoxO3aAAA and FoxO3a reporter, 6xDBE-Luc, together with or without HBZ. The interaction of FoxO3aAAA to FOXO-binding sites was analyzed by ChIP assay. As shown in Figure 2D, the interaction of FoxO3aAAA to the FOXO-binding sites was interfered by HBZ, suggesting that HBZ inhibits FoxO3a-mediated apoptosis through suppression of the DNA binding of FoxO3a. To clarify the mechanism of HBZ-mediated FoxO3a inhibition, we examined interaction between HBZ and FoxO3a by the immunoprecipitation assay. It showed that HBZ interacted with FoxO3a (Fig. 2E and F). Experiments with FoxO3a deletion mutant revealed that HBZ interacted with the forkhead domain of FoxO3a (Fig. 2E). Analysis using HBZ deletion mutants showed that the central domain of HBZ interacted with FoxO3a (Fig. 2F).

Next, we investigated the effect of HBZ on FoxO3a localization by confocal microscopy. We cotransfected 293FT cells with a plasmid expressing human wild-type FoxO3a (FoxO3aWT) protein and an HBZ-expressing plasmid. Consistent with previous reports, FoxO3a remained mainly in cytoplasm when cells were cotransfected with empty vector (Fig. 3A; refs. 20, 34). However, when it was expressed along with HBZ, FoxO3a was localized in both nucleus and cytoplasm (Fig. 3A). To determine whether mislocalized FoxO3a is phosphorylated (pFoxO3a) or not, we used anti-pFoxO3a antibody. Figure 3B and C demonstrated that nuclear-localized FoxO3a was phosphorylated in HBZ-expressing cells. Thereafter, we analyzed the localization of endogenous FoxO3a in HeLa and an ATL cell line, MT-1. Although pFoxO3a was localized widely both in cytoplasm and nucleus in HeLa cells, most pFoxO3a was localized in the nucleus in MT-1 (Fig. 3D), suggesting that endogenous HBZ inhibits the extranuclear translocation of pFoxO3a in this cell line. From the study of crystal structure of the human FoxO3a-DBD/DNA complex, it has been reported that phosphorylation at Ser253 causes a decrease on the DNA-binding ability (39). Abnormal localization of phosphorylated FoxO3a by HBZ might interfere the function of unphosphorylated FoxO3a in the nucleus. The abnormal localization of pFoxO3a prompted us to investigate whether HBZ bound to 14-3-3 along with FoxO3a, as 14-3-3 is a chaperon protein involved in nuclear–cytoplasm shuttling of FOXO family. As shown in Figure 3E, HBZ, FoxO3a, and 14-3-3 form a ternary complex. However, the binding of FoxO3a and 14-3-3 was not affected by HBZ (result of IP with anti-FLAG antibody and detected with anti-HA antibody).

As another possible mechanism for downregulation of Bim and FasL, we compared the transcription level of p73 in Jurkat cells with and without HBZ expression. Activation of HBZ-expressing cells reduced transcription of p73, but the expression level of p73 was variable among ATL cell lines (Supplementary Fig. S3A and S3B). We conclude that p73 is not responsible for suppression of Bim expression in ATL cells.

HBZ has been shown to suppress Bim expression through two different mechanisms as revealed in this study. To analyze Bim expression in ATL cells, we studied Bim mRNA levels in non-ATL cell lines and ATL cell lines with or without PMA/Io stimulation, and found that the Bim gene transcript was upregulated in Jurkat and CCRF-CEM cells, but not in SupT1 after activation. However, Bim transcripts were not increased in all stimulated ATL cell lines (Fig. 4A). The Bim gene transcript was also downregulated in primary ATL cells (Fig. 4B) compared with resting peripheral blood mononuclear cells (PBMC) and phytohemagglutinin (PHA)-stimulated T cells. We also stimulated primary ATL cells and normal CD4⁺ T cells with PMA/Io. The Bim gene transcription was quite low in primary ATL sample compared with normal CD4⁺ T cells even though the cells were stimulated with PMA/Io (Fig. 4C). To confirm HBZ expression in representative ATL cell lines, we quantified the level of the HBZ mRNA transcription in Jurkat-HBZ, CEM-HBZ, MT-1, ED, and TL-Om1 by real-time PCR and confirmed that HBZ is expressed in these ATL cell lines (Supplementary Fig. S4). Microarray data, obtained from Gene Expression Omnibus (GEO), show that both Bim and FasL transcription levels are lower in ATL cases than healthy donors (accession number: GSE33615; Supplementary Fig. S5), supporting our data that Bim expression was suppressed in ATL cells.

Because the Bim gene transcription was severely suppressed in ATL cells, we investigated the epigenetic status (DNA methylation and histone modification) of the promoter region of the Bim gene in ATL cells. A previous study showed that the 0.8 kb region immediately upstream of exon 1 contains the important elements for the control of Bim expression (promoter 1). The Bim promoter does not contain a TATA or CAAT box and has the characteristics of a “TATA-less” promoter (40). In addition, the alternative promoter has been reported to exist in intron 1 (promoter 2; refs. 41, 42). These two promoter regions are highly GC-rich and contain the binding sites for several transcription factors, including FoxO3a. To determine whether CpG sites in these Bim gene promoter regions are methylated in ATL cell lines, their methylation status was analyzed by bisulfite-mediated methylcytosine mapping (Supplementary Fig. S6A and S6B). The promoter 1 of Bim was hypermethylated in two ATL cell lines (ED and TL-Om1) and ATL case 1, whereas this region was not so methylated in MT-1 cells and two ATL cases. On the other hand, the promoter 2 was heavily methylated in two ATL cell lines (TL-Om1 and MT-1) and ATL case 1 and partially methylated in Jurkat cells (Supplementary Fig. S6B). These results suggest that in some cases, heavily methylated CpG sites of promoter 1 and 2 are associated with silencing of Bim transcription but these methylations can not account for suppressed Bim expression in all ATL cell lines and ATL cases.

Therefore, we next focused on the histone modification in the promoter region of Bim. It is well known that deacetylation of the histones are also common features of cancer, which results in transcriptional silencing of tumor suppressor genes (43). First, we analyzed the histone H3 and H4 acetylation and H3K4 trimethylation, which are all permissive marks (44), in promoter 1 of Jurkat, MT-1, and TL-Om1 cells. Contrary to our speculation, neither H3, H4 acetylation, nor H3K4 trimethylation differed between MT-1 and Jurkat cells (Supplementary Fig. S7). We next analyzed the histone modification status in promoter 2. As shown in Figure 5A, MT-1 and TL-Om1 cells exhibited decreased level of histone H3 acetylation and H3K4 trimethylation but not histone H4 acetylation. Because methylation of DNA is often preceded by dimethylation of H3K9 or trimethylation of H3K27 (both repressive marks) in oncogenesis (44), we asked whether there were differences in these epigenetic chromatin marks on the Bim gene promoter in ATL cell lines. TL-Om1 cells exhibited upregulated level of H3K9 dimethylation and H3K27 trimethylation compared with Jurkat cells (Fig. 5B and C), whereas MT-1 exhibited a little upregulated level of H3K27 trimethylation (Fig. 5C) in the promoter 2. These data suggest that histone modifications of promoter 2 are critical for the suppressed Bim gene transcription. We also performed ChIP analysis using anti-RNA polymerase II antibody (Fig. 5D) and revealed that Pol II binding was decreased in MT-1 and TL-Om1 cells, confirming suppressed transcription of the Bim gene. To further investigate the mechanisms involved in FoxO3a-mediated Bim gene transcription in the promoter 2, we transfected HA-tagged FoxO3a expression vector together with or without a HBZ expression vector into 293T cells and immunoprecipitated with anti-HA antibody. Then, the DNA-binding capacity of FoxO3a was quantified by real-time PCR. Figure 5E shows that HBZ attenuated the DNA-binding capacity of FoxO3a in the promoter 2 of Bim (i) and FasL promoter (ii), suggesting that the suppressed binding of FoxO3a to the promoter regions leads to inhibition of the Bim and FasL genes transcription by HBZ.

Next, we treated MT-1 cells with trichostatin A (TSA), a cell-permeable chemical inhibitor of class I/II histone deacetylases (HDAC). Treatment of TSA resulted in a clear upregulation of acetylation of histone H3 (Fig. 5F) followed by Bim expression both at the mRNA (Fig. 5G) and protein levels (Fig. 5H), indicating that histone modification is associated with suppressed Bim transcription in MT-1. We also performed ChIP assay using Jurkat-control and Jurkat-HBZ cells, which were stimulated with PMA and ionomycin for 9 hours, and found that acetylation of histone H3 decreased in Jurkat-HBZ cells (Fig. 5I), suggesting that HBZ is implicated in histone deacetylation in T cells. To verify whether HBZ inhibits transcription of the Bim gene, we suppressed the HBZ gene transcription by shRNA as reported previously (11). Efficiencies of lentivirus vector transduction, which were determined by EGFP expression, were 90.5% and 90.3% for control MT-1 cells and HBZ-knockdown MT-1 cells, respectively. Suppressed HBZ expression led to increase the Bim gene transcription (Fig. 5J and K), indicating that HBZ expression is linked to suppression of Bim expression in ATL cells.

Human immunodeficiency virus type 1 (HIV-1) replicates vigorously and the generated virus infects target cells in vivo. Unlike HIV-1, HTLV-1 induces proliferation to increase the number of infected cells, as this virus is transmitted primarily by cell-to-cell contact (5). Therefore, HTLV-1–encoded proteins promote proliferation of infected cells and inhibit their apoptosis, resulting in an increased number of infected cells in vivo (2). In this study, we show that HBZ inhibits both the intrinsic and extrinsic apoptotic pathways via targeting FoxO3a, which leads to suppressed transcriptions of Bim and FasL. We demonstrated two mechanisms for perturbation of FoxO3a by HBZ: interaction of HBZ with FoxO3a and interference of nuclear export of phosphorylated FoxO3a. HBZ suppresses DNA-binding ability of active form of FoxO3a through interaction between central domain of HBZ and forkhead domain of FoxO3a. In addition, LXXLL-like motif of HBZ is implicated in inhibition of FoxO3a-mediated apoptosis, suggesting that HBZ interferes in interaction of CBP/p300 and FoxO3a. Furthermore, HBZ retains inactive form of FoxO3a in the nucleus through interaction with 14-3-3, leading to transcriptional repression of the target genes. Interestingly, accumulation of phosphorylated form of FoxO3a in the nucleus has been observed in HIV Vpr-expressing cells, which might be implicated in HIV-mediated resistance against insulin (28). Thus, FoxO3a is a target of both human retroviruses.

In this study, we showed that central domain of HBZ interacts with FoxO3a while LXXLL-like motif in activation domain of HBZ is responsible for suppressed apoptosis. LXXLL-like motif of HBZ has been reported to interact with KIX domain of p300 (38). The central domain of HBZ interacts with the forkhead domain of FoxO3a, which binds to the target sequence (35). This is the mechanism how HBZ inhibits DNA binding of FoxO3a. However, inhibitory effect of HBZ on apoptosis largely depends on LXXLL-like motif of activation domain (Fig. 2A and B). FoxO3a is also reported to interact with KIX domain of CBP/p300 (37). Forkhead domain of FoxO3a intramolecularly interacts with its conserved regions (CR) 3, and binding of forkhead domain to DNA releases CR3, allowing it to bind KIX of CBP/p300 (45). These findings suggest that HBZ interferes in the complex interaction between FoxO3a and CBP/p300, which is likely important to induce apoptosis.

It has been reported that Bim has a tumor-suppressor function in various cancers. Hemizygous loss of the Bim gene promoted development of B-cell leukemia in Myc-transgenic mice in which c-myc expression was driven by the immunoglobulin gene intron-enhancer (46). Insulin-like growth factor 1 (IGF-1), an important growth factor for myeloma cells, has been reported to suppress Bim expression by epigenetic and posttranslational mechanisms (25). In Epstein–Barr virus–infected B cells, Bim transcription is silenced by DNA methylation of the Bim gene promoter (47). Thus, impaired expression of Bim is associated with the various cancers, including the virus-related malignancies. FoxO3a is also a target of oncogenesis. BCR-ABL induces phosphorylation of FoxO3a, which leads to suppressed expression of Bim in Ph1⁺ chronic myelogenous leukemia cells (32). In breast cancer, IκB kinase interacts with, phosphorylates FoxO3a, which causes proteolysis of FoxO3a (48). In this study, we revealed that HBZ hinders nuclear export of phosphorylated FoxO3a, and impairs function of FoxO3a likely through interaction of FoxO3a and p300. Thus, suppressed Bim and FasL expression through inhibition of FoxO3a by HBZ is a new mechanism for oncogenesis.

Besides FoxO3a perturbation by HBZ, we also have identified the epigenetic aberrations in the promoter region of the Bim gene in ATL cells, and found that Bim expression is suppressed by DNA methylation and histone modification. ATL cell lines exhibited upregulated level of H3K27 trimethylation in the promoter regions of Bim. It has been reported that enhancer of zeste (EZH) 2, a methyltransferase and component of the polycomb repressive complex 2, expression is increased in ATL cell lines (42). Because EZH2 plays an essential role in the epigenetic maintenance of H3K27 trimethylation, upregulated H3K27 trimethylation of the Bim gene promoter might be associated with increased expression of EZH2 in ATL cells. In addition, HBZ seems to be associated with histone deacetylation in MT-1 cells. According to the previous studies, it is known that both HBZ and FoxO3a bind to the histone acetyltransferase p300/CBP through the LXXLL motif (38). In this study, we found that the same motif is important for FoxO3a suppression and resulting inhibition of apoptosis. It is likely that HBZ decreases histone acetylation level on Bim promoter through the interaction with FoxO3a and dissociation of p300/CBP from the promoter. In addition to histone modifications, hypermethylation of CpGs in Bim promoter was observed in some ATL cells. These epigenetic aberrations likely occur as the secondary changes following long-time silencing of Bim by HBZ, although further investigations will be required.

In this study, we demonstrated that HBZ suppresses activation-induced apoptosis by downregulation of proapoptotic genes, Bim and FasL. HBZ perturbs the function of FoxO3a by interaction, and induces epigenetic aberrations in the promoter region of the Bim gene. It has been shown that HBZ induces not only cancer but also inflammation in vivo. Because inflammatory diseases are essentially caused by failure to negatively regulate unnecessary immune responses by apoptosis, suppression of apoptosis by HBZ might be associated with HTLV-1–induced inflammation as well. Collectively, HBZ-mediated inhibition of apoptosis is likely implicated in both neoplastic and inflammatory diseases caused by HTLV-1.

No potential conflicts of interest were disclosed.

Conception and design: A. Tanaka-Nakanishi, J. Yasunaga, M. Matsuoka

Development of methodology: A. Tanaka-Nakanishi, M. Matsuoka

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Tanaka-Nakanishi, K. Takai

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Tanaka-Nakanishi, J. Yasunaga, K. Takai, M. Matsuoka

Writing, review, and/or revision of the manuscript: A. Tanaka-Nakanishi, J. Yasunaga, M. Matsuoka

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Matsuoka

Study supervision: M. Matsuoka

The authors thank T. Furuyama (Kagawa Prefectural University of Health Science) for the 6xDBE-Luc plasmid DNA, P. Bouillet for valuable comments on this study, and L. Kingsbury for proofreading of this manuscript.

This study was supported by a Grant-in-aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan to M. Matsuoka (MEXT grant number 221S0001), a grant from Japan Leukemia Research Fund to M. Matsuoka, and a grant from the Takeda Science Foundation to J. Yasunaga.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.