Positive and Negative Regulation of Type I Interferons by the Human T Cell Leukemia Virus Antisense Protein HBZ

ABSTRACT

The pathogenesis of human T cell leukemia virus type 1 (HTLV-1) is strongly linked to the viral regulatory proteins Tax1 and HBZ, whose opposing functions contribute to the clinical outcome of infection. Type I interferons alpha and beta (IFN-α and IFN-β) are key cytokines involved in innate immunity, and IFN-α, in combination with other antivirals, is extensively used in the treatment of HTLV-1 infection. The relationship between HTLV-1 and IFN signaling is unclear, and to date the effect of HBZ on this pathway has not been examined. Here we report that HBZ significantly enhances interferon regulatory factor 7 (IRF7)-induced IFN-α- and IFN-stimulated response element (ISRE) promoter activities and IFN-α production and can counteract the inhibitory effect of Tax1. In contrast to this, we show that HBZ and Tax1 cooperate to inhibit the induction of IFN-β and ISRE promoters by IRF3 and IFN-β production. In addition, we reveal that HBZ enhances ISRE activation by IFN-α. We further show that HBZ enhances IRF7 and suppresses IRF3 activation by TBK1 and IKKε. We demonstrate that HBZ has no effect on virus-induced nuclear accumulation of IRF3, suggesting that it may inhibit IRF3 activity at a transcriptional level. We show that HBZ physically interacts with IRF7 and IKKε but not with IRF3 or TBK1. Overall, our findings suggest that both HBZ and Tax1 are negative regulators of immediate early IFN-β innate immune responses, while HBZ but not Tax1 positively regulates the induction of IFN-α and downstream IFN-α signaling.

IMPORTANCE Type I interferons are powerful antiviral cytokines and are used extensively in the treatment of HTLV-1-induced adult T cell leukemia (ATL). To date, the relationship between HTLV-1 and the IFN pathway is poorly understood, and studies so far have focused on Tax1. Our study is unique in that it examined the effect of HBZ, alone or in combination with Tax1, on type I IFN signaling. This is important because HBZ is frequently the only viral protein expressed in infected cells, particularly at later stages of infection. A better understanding of the how HBZ regulates IFN signaling may lead to the development of therapeutics that can modify such responses and improve the clinical outcome for infected individuals.

INTRODUCTION

Human T cell leukemia virus type 1 (HTLV-1) is the causative agent of adult T cell leukemia (ATL), an aggressive malignancy of CD4⁺ T lymphocytes, and a chronic inflammatory neurodegenerative disorder termed HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) (1–3). The majority of HTLV-1-infected individuals remain asymptomatic throughout their lives, with only 2 to 5% going on to develop either ATL or HAM/TSP. HTLV-1 infection is governed by viral regulatory proteins, the most widely studied of which are Tax1 and HBZ (4, 5). Tax1 is a pivotal transcriptional regulator of both viral and cellular gene expression, which is linked to their ability to modulate several cellular signaling pathways, including CREB, NF-κB, and AP1. The molecular functions of HBZ are still being discovered, but it is well established that it plays a key role in the pathogenesis of both HAM/TSP and ATL. Based on the fact that it is expressed from the antisense strand of the HTLV-1, HBZ is very often the only viral gene expressed at later stages of infection and is linked to ATL development (6) and the severity of HAM/TSP (7). In cells that express both Tax1 and HBZ, HBZ can counteract Tax-mediated activation of key cellular pathways, including CREB, the canonical NF-κB pathway, and AP1 pathways, all of which are involved in viral gene expression, transformation, and cell growth (8–12). Downregulation of Tax-mediated NF-κB activation by HBZ has been shown to relieve cellular senescence and promote viral latency (13). Moreover, HBZ promotes T cell proliferation (14, 15) and induces T cell lymphomas and chronic inflammation in transgenic animals (16).

Type I interferons alpha and beta (IFN-α and IFN-β) play key roles in both innate and adaptive immunity (17). Viral infection activates cellular signaling pathways through various pattern recognition receptors (PRRs) comprising the cytosolic RIG-like helicases (RHLs), cGAS, and DAI receptors together with endosomal or cell surface Toll-like receptors (TLRs). These receptors are activated by a variety of microbial nucleic acids and by-products, leading to the induction of type I IFNs. Activation of the cytosolic receptors and TLR3 and TLR4 by infection ultimately results in the activation of the kinases TBK1 and IKKε, making them a focal point in the IFN signaling pathway. Activated TBK1 and IKKε phosphorylate interferon regulatory factors 3 and 7 (IRF3 and IRF7), resulting in their nuclear translocation and activation of IFN production (18, 19). IRF3 mainly regulates IFN-β expression, whereas IRF7 predominantly activates IFN-α promoters (20, 21). Nuclear IRF3 associates with CBP/P300 (22) and together with p50/p65 NF-κB subunits and ATF-2/c-Jun stimulates IFN-β transcription (23). Secreted IFN binds IFN receptors in an autocrine and paracrine manner, resulting in the activation of the JAK/STAT pathway and the expression of a multitude of IFN-stimulated genes (ISGs), including that for IRF7. Phosphorylated IRF7 participates with IRF3 in the amplification of IFN responses (24). As opposed to IRF3, which is ubiquitously expressed, IRF7 is expressed at low levels in most cells except for plasmacytoid dendritic cells (pDCs), which express high levels of latent IRF7 (25–27). As a result, the induction of IFN-α by IRF7 in pDCs is independent of IFN-β/IFN receptor feedback, making these cells highly responsive to infection and major producers of IFN-α in vivo.

The relationship between HTLV-1 infection and innate immune responses is poorly understood. However, it is now well established that in addition to CD4⁺ cells, HTLV-1 also infects innate immune cells such as dendritic cells (DCs) and monocytes/macrophages (28–33). Myeloid dendritic cells (myDCs) and pDCs can be infected ex vivo by HTLV-1 and transmit infection to autologous activated CD4⁺ T cells, suggesting that they may be involved in spreading the virus in vivo (28–30). DCs isolated from asymptomatic individuals and HAM/TSP and ATL patients are infected by HTLV-1 (28, 31) but are reported to have impaired IFN-α production capacity (31). Stimulation of DCs with cell-free HTLV-1 virions results in the potent induction of IFN-α through TLR7 signaling and the upregulation of cell surface expression of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), transforming them into IFN-producing killer pDCs (34). Recent studies have shown that infected monocytes contribute up to 15% of the viral load in HAM/TSP patients, suggesting an important role for these cells in viral persistence and pathogenesis in vivo (33). However, ex vivo infection of primary monocytes using purified HTLV-1 was reported to result in nonproductive infection and apoptosis due to the activity of the host restriction factor SAMHD1 and activation of innate immune responses (35). The role of HTLV-1 regulatory proteins in PRR signaling is somewhat unclear and in some cases contradictory. The HTLV-1 p30 protein was reported to downregulate the surface expression of TLR4, resulting in a decrease in proinflammatory cytokine levels and increased expression of the anti-inflammatory cytokine interleukin 10 (IL-10) (36). Moreover, p30 was found to inhibit TLR3 but not TLR7/8 induction of ISGs in ex vivo monocytes and monocyte-derived dendritic cells, suggesting that at early stages of infection, HTLV-1 dampens IFN responses, which may account for the establishment and persistence of infection (37). Some studies have shown that IRF3 is constitutively activated in Tax1-positive HTLV-1-transformed cell lines (38) and enhances IFN-β promoter activity by forming a complex with the IRF3 kinases TBK1, IKKε, and TRAF3 (39). In contrast to this, another study has shown that the induction of the suppressor of cytokine signaling 1 (SOCS1) by HTLV infection results in the proteosomal degradation of IRF3 and hence evasion of innate immune responses by HTLV-1 (40). Moreover, recent studies have shown that Tax1 potently inhibits RIG-I, MDA-5, and TRIF-mediated IFN signaling (41), leading to the suggestion that Tax1 is involved in innate immune evasion. In support of this view, Tax1 was reported in another study to suppress the induction of IFN-β by binding TBK1, IKKε, and IRF3 and inhibiting the ability of TBKI to phosphorylate IRF3 (42).

The aim of the present study was to investigate physical and functional interactions between the HTLV-1 regulatory proteins HBZ and Tax1 and the induction of type I IFNs by IRF3 or IRF7. Our data show that HBZ and Tax1 synergistically inhibit IRF3-induced IFN-β and IFN-stimulated response element (ISRE) promoter activities, while HBZ but not Tax1 augments the induction of IFN-α and promotes downstream IFN signaling.

RESULTS

HBZ enhances IRF7 but inhibits IRF3 induction of type I IFNs.

Interferon expression is governed by IRF7 and IRF3 (17). To date, the effect of HBZ on the transcriptional activities of IRF3 and IRF7 has not been investigated. To address this issue, luciferase assays were performed on lysates from 293T cells transfected with IFN-α or IFN-β luciferase (IFN-α-Luc and IFN-β-Luc) reporter plasmids together with expression plasmids encoding Flag-IRF7 or IRF3 and increasing concentrations of Flag-HBZ. Under these conditions, HBZ significantly enhanced IFN-α promoter activity in a dose-dependent manner (Fig. 1A). Conversely, HBZ potently inhibited IFN-β activation by IRF3, even at the lowest concentration of IRF3 expression plasmid tested (Fig. 1B). The binding of type I IFNs to cell surface IFN receptors leads to the activation of hundreds of IFN-stimulated genes (ISGs) due to the presence of IFN-stimulated response elements (ISREs) in their promoters (17). To extend our existing data and determine the effect of HBZ on downstream IFN signaling, 293T cells were transfected with an ISRE-Luc reporter plasmid together with expression plasmids encoding Flag-IRF7, IRF3, or Flag-HBZ or empty plasmid as indicated below. Under these conditions, we observed that HBZ enhanced IRF7 and inhibited IRF3 activation of ISRE promoter activity (Fig. 1C and D). Similarly, HBZ enhanced the induction of IFN-α by IRF7 and inhibited the induction of IFN-β by IRF3 in Jurkat cells (Fig. 1E and IF). IRF7-induced ISRE promoter activity was enhanced by HBZ in Jurkat cells (Fig. 1G), but in contrast to its inhibitory effect on ISRE activation by IRF3 in 293T cells, HBZ enhanced the induction of ISRE promoter activity by IRF3 in Jurkat cells (Fig. 1H).

FIG 1.

Effect of HBZ on IRF3 and IRF7 activities. (A to D) 293T cells were cotransfected with plasmids encoding Flag-IRF7 or IRF3 as indicated, together with an IFN-α-Luc (A) or IFN-β-Luc (B) or ISRE-Luc reporter plasmid (C and D) and either 2 μg (C and D) or increasing concentrations of a Flag-HBZ plasmid (250 ng to 4 μg [A and B]). Cells were lysed 24 h posttransfection, and relative luciferase values were obtained using the luciferase reporter assay system and normalized to protein concentrations. Values are expressed as fold luciferase activation relative to that of the control, which was assigned the arbitrary value of 1. (E to H) Jurkat cells (1 × 10⁶) were transfected with plasmids encoding Flag-IRF7 or IRF3 as indicated, together with an IFN-α-Luc (E), IFN-β-Luc (F), or ISRE-Luc reporter plasmid (G and H) and 4 μg of a plasmid encoding Flag-HBZ. Cells were lysed 48 h posttransfection, and luciferase values were determined as described above. Data are representative of those from at least three independent experiments. Error bars represent standard deviations. Single, double, and triple asterisks indicate statistical significance at P values of ≤0.05, ≤0.01, and ≤0.001, respectively, obtained using a two-tailed Student t test. Western blot analysis was performed on lysates using the indicated antibodies. NS, no statistical significance.

Sendai virus (SeV) is a well-established activator of IRF3 and IRF7 and consequently type I IFN expression (43). We next determined the effect of HBZ on the transcriptional activities of IRF3 and IRF7 following SeV infection. As expected, SeV infection of 293T cells resulted in the induction of IFN-α and IFN-β promoters compared to basal levels (Fig. 2A and B). Consistent with the results described above, we observed that HBZ expression significantly enhanced SeV-induced IRF7-driven IFN-α promoter activity, while it diminished IFN-β activation by IRF3 following infection. Based on previous studies showing that Tax1 can either enhance (39) or suppress (41, 42) IFN signaling, we additionally sought to determine the overall impact of HBZ and Tax1 on IFN-α and IFN-β production following SeV infection. Consistent with its effects at a transcriptional level, HBZ promoted the production of IFN-α by IRF7 and counteracted the inhibitory effect of Tax1 (Fig. 2C). Moreover, HBZ and Tax1 individually and collectively inhibited IFN-β production following viral infection (Fig. 2D).

FIG 2.

Regulation of IFN signaling by HBZ induced by SeV infection. (A and B) HBZ enhances IRF7 (A) but suppresses IRF3 activation (B) of IFN signaling following SeV infection of cells. 293T cells were cotransfected with Flag-IRF7 or IRF3 and Flag-HBZ expression plasmids together with IFN-α-Luc (A) or IFN-β-Luc (B) reporter plasmids as indicated. Six hours posttransfection, cells were infected or not with SeV (180 HA units) and incubated overnight. Cells were lysed and luciferase activities were determined and normalized as described above. (C and D) 293T cells were cotransfected with the indicated expression and reporter plasmids and infected with SeV as described above. Cell culture supernatants were assayed for IFN-α production (C) or IFN-β production (D). (E) HBZ enhances the induction of ISRE promoter activity by IFN-α. 293T cells grown on 12-well plates were transfected with an ISRE-Luc reporter plasmid together with increasing concentrations of a plasmid encoding Flag-HBZ. Six hours posttransfection, cells were incubated or not with human IFN-α (100 U/ml) and incubated overnight. Cells were lysed and luciferase activities were determined as described above. Luciferase values are expressed as fold luciferase activation relative to that of the control, which was assigned the arbitrary value of 1. Data are representative of those from at least three independent experiments. Error bars represent standard deviations. Single, double, and triple asterisks indicate statistical significance at P values of ≤0.05, ≤0.01, and ≤0.001, respectively, obtained using a two-tailed Student t test. Western blot analysis was performed on lysates using the indicated antibodies.

IFN-α in combination with zidovudine and other therapeutics is used as a standard therapy for ATL (44). Based on the fact that the HBZ gene is very often the only viral gene expressed at later stages of infection and is linked to ATL development (6) and the severity of HAM/TSP (7), we sought to determine the effect of HBZ on ISRE activation by IFN-α. As expected, IFN-α treatment of 293T cells resulted in a significant induction of ISRE-Luc activity (Fig. 2E). HBZ expression significantly enhanced IFN-α-induced ISRE activity levels in a dose-dependent manner. Overall, these results suggest that HBZ affects both the induction of type I IFNs and downstream IFN signaling.

Identification of the domain in HBZ involved in the regulation of IRF3 and IRF7.

HBZ contains distinct functional domains, including an N-terminal activation domain (AD), a central domain, and a C-terminal basic ZIP domain (45). Interactions between these domains and various cellular factors, such as c-Jun, JunB, JunD, and CREB-2 (9–11, 46), dictate the functionality of HBZ. To identify the functional domain in HBZ involved in its ability to regulate the activities of IRF3 or IRF7, we cotransfected 293T cells with either IFN-β-Luc or IFN-α-Luc reporter constructs together with expression plasmids encoding IRF3 or IRF7, respectively, and wild-type HBZ, HBZΔAD, HBZΔCD, and HBZΔbZIP and determined IFN-α and IFN-β promoter activity. The loss of each of the functional domains of HBZ appeared to reduce its ability to activate IRF7, with the loss of the bZIP domain having the most pronounced effect (Fig. 3A). In contrast to this, no specific HBZ domain appeared to be involved in the inhibition of IRF3-mediated IFN-β activation (Fig. 3B).

FIG 3.

The HBZ bZIP domain is involved in positively regulating IRF7 activity, while no specific HBZ domain is involved in the negatively regulating IRF3 activity. 293T cells were cotransfected with Flag-IRF7 or pcDNA IRF3 expression plasmids and IFN-α-Luc (A) or IFN-β-Luc (B) reporter plasmids together with expression plasmids for WT Flag-HBZ or Flag-HBZ with deletions in the activation domain (ΔAD), central domain (ΔCD), or basic Zip (ΔbZip) domain as indicated. Cells were lysed 24 h posttransfection, and relative luciferase values were obtained using the luciferase reporter assay system. Luciferase values were normalized to protein concentrations and are expressed as fold luciferase activation relative to that of the control, which was assigned the arbitrary value of 1. Single, double, and triple asterisks indicate statistical significance at P values of ≤0.05, ≤0.01, and ≤0.001, respectively, obtained using a two-tailed Student t test. Protein expression levels were analyzed by Western blotting. Relevant protein bands are indicated by arrows. Tubulin was used as a loading control.

Effect of HBZ and Tax1 on IKKε and TBK1 activity.

In an effort to delineate the mechanisms involved in the regulation of IRF3 and IRF7 activities by HBZ, we sought to determine the effect of HBZ on the activities of the upstream IRF kinases, IKKε and TBK1. Based on previous studies showing that Tax1 can either enhance (39) or suppress (41, 42) the activities of TBK1 and IKKε, we additionally sought to determine the overall impact of HBZ and Tax1 on the activities of both kinases. For this purpose, 293T cells were transfected with expression plasmids encoding Flag-IKKε or Flag-TBK1, Flag-IRF7, and HIS-Tax1 together with either IFN-α-Luc or IFN-β-Luc reporter constructs. As expected, IKKε and TBK1 significantly enhanced IFN-α-Luc activation by IRF7 (Fig. 4A and C) and HBZ further enhanced activation levels, 3.5- and 5-fold, respectively (Fig. 4B and D), indicating that HBZ enhances IRF7 activation by both of these kinases. In contrast to this, we show that Tax1 potently inhibits activation of IRF7 by both IKKε and TBK1. Interestingly, we observed that HBZ could almost completely abolish the inhibitory effect of Tax1 on TBK1 (Fig. 4D) and partially counteract the inhibitory effect of Tax1 on IKKε, restoring activity levels to those obtained in the absence of HBZ (Fig. 4B). These results suggest that HBZ and Tax1 have opposite effects on IRF7 activation by IKKε and TBK1. Consistent with previous studies, we show that IKKε and TBK1 strongly induced IFN-β promoter activity (Fig. 5A and C). In contrast to their opposing effects on IRF7 activation by IKKε and TBK1, HBZ and Tax1 individually and synergistically suppressed the induction of IFN-β promoter activity by both of these kinases (Fig. 5B and D). In summary, HBZ upregulates IKKε and TBK1 activation of IFN-α but downregulates the induction of IFN-β by these kinases.

FIG 4.

HBZ enhances activation of IRF7 by IKKε and TBK1 and counteracts the inhibitory effect of Tax1. 293T cells were cotransfected with Flag-IRF7, Flag-HBZ, pCAGGS Tax1-HIS, and Flag-IKKε (A and B) or Flag-TBK1 (C and D) expression plasmids as indicated, together with an IFN-α-Luc reporter plasmid. Cells were lysed 24 h posttransfection, and relative luciferase values were obtained using the luciferase reporter assay system. Luciferase values were normalized to protein concentrations and are expressed as fold luciferase activation relative to that of the control, which was assigned the arbitrary value of 1. Single, double, and triple asterisks indicate statistical significance at P values of ≤0.05, ≤0.01, and ≤0.001, respectively, obtained using a two-tailed Student t test. Protein expression levels were analyzed by Western blotting. Tubulin was used as a loading control.

FIG 5.

HBZ and Tax1 syngeristically inhibit the induction of IFN-β by IKKε and TBK1. 293T cells were cotransfected with Flag-HBZ, pCAGGS Tax1-HIS, and Flag-IKKε (A and B) or Flag-TBK1 (C and D) expression plasmids as indicated, together with an IFN-β-Luc reporter plasmid. Cells were lysed 24 h posttransfection, and relative luciferase values were obtained using the luciferase reporter assay system. Luciferase values were normalized to protein concentrations and are expressed as fold luciferase activation relative to that of the control, which was assigned the arbitrary value of 1. Single, double, and triple asterisks indicate statistical significance at P values of ≤0.05, ≤0.01, and ≤0.001, respectively, obtained using a two-tailed Student t test. Protein expression levels were analyzed by Western blotting. Tubulin was used as a loading control.

HBZ associates with IRF7 and IKKε but not with IRF3 or TBK1.

Based on our data showing that HBZ regulates the activities of TBK1 and IKKε, we next sought to determine whether the mechanisms involved entailed physical interactions between these proteins. Initially, coimmunoprecipitation assays were performed on lysates from cells expressing Flag-IRF7 or IRF3 together with HIS-HBZ or empty plasmid. Surprisingly, we found that HBZ specifically interacts with IRF7 (Fig. 6A) but not with IRF3 in uninfected (Fig. 6B) or in SeV-infected (Fig. 6C) cells. To further characterize the interaction between IRF7 and HBZ, we sought to map the domains involved. IRF7 contains several important functional domains, including an amino-terminal DNA binding domain (amino acids [aa] 1 to 146), a constitutive activation domain (aa 151 to 246), a virus activation domain (aa 278 to 305), and an inhibitory domain spanning aa 305 to 466 (20) (Fig. 6D). To determine the domains in IRF7 involved in interactions with HBZ, we performed coimmunoprecipitation experiments with HBZ and a variety of IRF7 truncated proteins. This analysis revealed a specific interaction between HBZ and the C-terminal half of IRF7 (Fig. 6E) and in particular a domain encompassing amino acids 283 to 466 (Fig. 6F). No interaction was observed between HBZ and the N-terminal half of IRF7 (IRF7 aa 1 to 283) (Fig. 6E). These data demonstrate that HBZ interacts with a region of IRF7 encompassing the inhibitory domain which normally masks the N-terminal DNA binding and transactivation domains, thereby preventing the downstream activation of IRF7 in uninfected cells (20). Further coimmunoprecipitation assays on lysates from cells expressing Flag-TBK1 or Flag-IKKε together with HIS-HBZ showed a specific interaction between HBZ and IKKε but not TBK1 (Fig. 6G). Hence, HBZ forms a complex with IRF7 and IKKε but not IRF3 or TBK1.

FIG 6.

HBZ interacts with IRF7 and IKKε but not with IRF3 or TBK1. (A and B) 293T cells were cotransfected with HIS-HBZ and either Flag-IRF7 or IRF3 expression plasmids. Transfected cells were lysed after 24 h, and coimmunoprecipitations were performed using anti-Flag M2 resin. Immunoblot analysis of immunoprecipitates (IP) and whole-cell lysates (WCL) was performed using the indicated antibodies. (C) 293T cells were cotransfected with HIS-HBZ and either Flag-IRF7 or Flag-IRF3 expression plasmids. The cells were infected with SeV, and coimmunoprecipitations were performed as described above. (D) Schematic representation of the functional domain in IRF7 (20): DNA binding domain (DBD), constitutive activation domain (CAD), virus-activated domain (VAD), inhibitory domain (ID), and signal response domain (SRD). (E and F) 293T cells were cotransfected with HIS-HBZ and either Flag-IRF7 (wild type) or Flag-IRF7 1-283, Flag-IRF7 283-503, or Flag-IRF7 283-466 expression plasmids. Transfected cells were lysed after 24 h, and coimmunoprecipitations were performed using anti-Flag M2 resin. Immunoblot analysis of immunoprecipitates and whole-cell lysates was performed using the indicated antibodies. (G) 293T cells were cotransfected with HIS-HBZ and either Flag-TBK1 or Flag-IKKε or Flag empty expression plasmids. Coimmunoprecipitation and Western blotting were carried out as described above. Flag-TBK1 and Flag-IKKε are indicated by arrows.

HBZ inhibits the induction of IFN-β by IRF3 in the nucleus.

Following infection, IRF3 is phosphorylated and is transferred to the nucleus, where it interacts with CBP/p300 and activates IFN-β expression (22). Based on our data showing that HBZ potently inhibits IRF3-mediated IFN-β activation, we sought to determine whether HBZ affects the nuclear accumulation of IRF3 following viral infection. To examine this, immunofluorescence analysis was performed on HeLa cells transfected with green fluorescent protein (GFP)-HBZ and subsequently infected with SeV (Fig. 7). As expected, endogenous IRF3 was localized in the cytoplasm of uninfected cells and GFP-HBZ was exclusively nuclear. Following infection, IRF3 relocalized to the nucleus as expected, and we observed no discernible difference in the levels of nuclear IRF3 in the presence and absence of HBZ (Fig. 7B). In order to strengthen these data, we sought to determine the effect of HBZ on the levels of phosphorylated IRF3 in the nucleus following viral infection. To this end, we performed nuclear/cytoplasmic fractionation of cells transfected or not with an expression plasmid encoding HBZ and subsequently infected with SeV. Western blot analysis of fractions showed that HBZ was exclusively nuclear, as expected, and we observed similar levels of P-IRF3 in HBZ-expressing and HBZ-negative cells (Fig. 7C). This strongly suggests that HBZ does not block steps, such phosphorylation and dimerization, involved in the nuclear translocation of IRF3 but instead may inhibit the activity of IRF3 at a transcriptional level.

FIG 7.

HBZ does not affect the nuclear accumulation of IRF3 induced by SeV infection. HeLa cells grown on two-well chamber slides were transfected with a GFP-HBZ expression plasmid and mock infected (A) or infected with SeV (B) overnight. Endogenous IRF3 was detected using anti-IRF3 followed by anti-rabbit antibody conjugated with Alexa Fluor 594. Nuclei were stained using DAPI. Immunofluorescence images were obtained using a Zeiss Axiolmager MI fluorescence microscope. (C) 293T cells were cotransfected with FLAG-HBZ or empty plasmid. Six hours posttransfection, cells were infected with SeV (180 HA units) and incubated overnight. Nuclear/cytoplasmic fractionation of cells was performed, and equal concentrations of proteins were analyzed by Western blotting using the indicated antibodies. Histone deacetylase 1 (HDAC1) was used as a nuclear loading control.

DISCUSSION

The induction of type I IFNs is a fundamental cellular response to combat infection. Stimulation of pattern recognition receptors (PRRs) by invading pathogens activates signaling pathways that converge at IRF3 and IRF7, resulting in IFN production and the expression of IFN-stimulated genes (ISGs) and host antiviral responses (17). Conversely, prolonged and sustained stimulation of IFN production due to chronic infection is linked to persistent immune activation and inflammation that can contribute to viral pathogenesis (47, 48). The impact of HTLV-1 infection on IFN signaling is poorly understood and in some cases contradictory. Studies so far have focused on Tax1 and show that Tax1 can either activate (39) or suppress (41, 42) the IFN pathway. The role of the HTLV-1 regulatory protein HBZ alone or in combination with Tax1 on IRF3/IRF7 activation and IFN signaling has so far not been investigated. Our data show that HBZ can significantly enhance IRF7-induced IFN-α and ISRE promoter activity and can counteract the inhibitory effect of Tax1 on IFN-α. We further show that HBZ and Tax1 synergistically inhibit the induction of IFN-β and ISRE promoters and IFN-β production. Regulation of type I IFN by HBZ was associated with its ability to either up- or downregulate TBK1 and IKKε activation of IRF7 and IRF3, respectively. We further show HBZ does not inhibit the nuclear accumulation of IRF3, suggesting that it inhibits the activity of IRF3 at a transcriptional level. We show that HBZ physically interacts with IRF7 and IKKε but not with IRF3 or TBK1. Overall our data suggest that HTLV-1 regulates IFN responses at multiple levels, which may play a role in not only immune evasion and persistence but may also contribute to abnormal IFN signaling and viral pathogenesis.

The type I IFN response is bimodal in that IFN-β initially produced and secreted by infected cells acts in an autocrine and paracrine manner to induce numerous ISGs, including that for IRF7, which, in turn, activates IFN-α and amplifies host antiviral responses (17). IRF3 governs IFN-β expression and is phosphorylated by IKKε and TBK1 following infection, resulting in its transfer to the nucleus, where it interacts with CBP/p300 and activates IFN-β expression (22). Our data clearly demonstrates that HBZ and Tax1 individually and synergistically suppress both virus-induced and IKKε- and TBK1-activated IFN-β expression and hence the early wave of IFN responses. Even though previous studies have not examined the effect of HBZ on type I IFN signaling, our data on Tax1 are consistent with some previous studies showing that Tax 1 inhibits the induction of IFN-β by both TBK1 and IKKε, which was linked to its ability to interact with TBK1 and inhibit IRF3 phosphorylation (42). Moreover, Tax1 has been shown to interact with TRIF, RIG-1, and RIP1 to potently inhibit the induction of IFN-β (41). This study suggested that since TRIF is upstream of TBK1 and IKKε, the observed effects on IFN-β might be mediated through these kinases. Our data contradict other studies which show that Tax1 upregulates the induction of IFN-β by interacting with TBK1 and IKKε (39) or through TAK1 and TBK1 (38). It is possible that differences in the integrity of cell lines, experimental procedures, or expression constructs used might account for the conflicting observations. The underlying mechanisms involved in HBZ suppression of IFN-β activation by IRF3 are unclear. Our failure to detect a physical interaction between IRF3 and HBZ in either SeV-infected or uninfected cells shows that a physical interaction between these two proteins is not involved. In addition, we could not identify a domain in HBZ involved in IFN-β inhibition, suggesting that multiple factors may be involved. It is also possible that HBZ RNA rather than protein is involved in regulating IFN-β, as HBZ RNA has previously been shown to regulate genes involved in several cellular processes, including cell survival and proliferation (49). Our finding that HBZ has no effect on the nuclear translocation or phosphorylation status of IRF3 following SeV infection suggests that HBZ may inhibit IRF3 at a transcriptional level, possibly by interfering with its interaction with the coactivators CBP and p300 (Fig. 8). HBZ has previously been shown to repress viral gene expression by binding CREB-2, p300, and CBP, resulting in diminished levels of these coactivators at the viral promoter and hence lower levels of promoter activity (10, 11). A similar mechanism may be involved in HBZ-mediated inhibition of IRF3 activity, but this requires further investigation. Overall, our data suggest that HTLV-1, through HBZ alone or in combination with Tax1, effectively evades the early IFN-β responses that rely on IRF3, which plausibly leads to enhanced survival and persistence within the host.

FIG 8.

Schematic diagram of hypothetical model of IRF3 and IRF7 regulation by HBZ. HBZ has no effect on the nuclear accumulation of phosphorylated IRF3 and may suppress IFN-β expression by competing with IRF3 for CBP/p300 (left). The binding of HBZ to a region of IRF7 encompassing the inhibitory region may enhance the transcriptional activity of IRF7 (right).

IRF7 is regarded as a master regulator of all IFN-dependent immune responses based on studies showing that IRF7-deficient mice are highly susceptible to viral infection as a result of reduced levels of serum IFN (50). As opposed to IRF3, which is ubiquitously expressed, most cells do not express IRF7 without prior stimulation with IFN (20). Hence, IRF7 expression relies on IFN-β expression in most cells. Our finding that both Tax1 and HBZ potently inhibit the induction of IFN-β expression suggests that they may also negatively influence the induction of IFN-α by limiting the expression of IRF7. However, some cells, such as plasmacytoid dendritic cells (pDCs), uniquely express latent IRF7, making them ideally suited to respond rapidly to PRR activation in the absence of prior IFN stimulation (25–27). pDCs can therefore produce large quantities of IFN-α, resulting in ISG expression and the induction of an antiviral state. HTLV-1 has been previously shown to infect dendritic cells, and pDCs exposed to cell free HTLV-1 are reported to produce high levels of IFN-α (28–32). However, other studies showed that the number of pDCs is decreased in ATL patients and the remainder has impaired IFN-α production capacities (31). Nevertheless, our finding that HBZ enhances the induction of IFN-α by IRF7 suggests that HBZ may contribute to the overproduction of IFN-α by infected DCs. Even though the mechanisms involved in HBZ-mediated IRF7 activation have not been determined, our finding that HBZ binds the inhibitory domain of IRF7 may be relevant. Previous studies have shown that alteration or removal of this domain significantly enhances the activity of IRF7 (20). This raises the possibility that by binding this domain, HBZ relieves its inhibitory effect on the transcriptional activity of IRF7, but this requires further investigation. In addition, we show that HBZ enhances ISRE promoter activation in response to IRF7 or exogenous IFN-α, suggesting that it enhances not only the expression of IFN-α but also downstream signaling. Amplification of such responses by HBZ could be both beneficial and detrimental to infected hosts. On the one hand, enhanced IFN-α and ISG expression may promote the induction of a potent antiviral state in vivo and hence limit the spread of the infection. Conversely, sustained stimulation of IFN-α by HBZ may contribute to the development of an inflammatory state in vivo that promotes the development of pathology. A proinflammatory role for HBZ is supported by studies showing that HBZ transgenic mice develop systemic inflammatory disease (16). Moreover, the severity of the inflammatory disorder HAM/TSP is linked to HBZ expression (7) and the overexpression of a distinct subset of ISGs (51). Our data suggest that modulation of the IFN-α pathway by HBZ could contribute to the induction of a heightened inflammatory state.

In conclusion, the results presented here show that HBZ and Tax1 dampen early IFN-β responses, while HBZ but not Tax1 augments the expression of IFN-α and downstream IFN signaling, which may contribute to immunopathogenesis.

MATERIALS AND METHODS

Cell culture and reagents.

293T and HeLa cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco, Life Technologies) supplemented with 10% fetal bovine serum (FBS; Gibco, Life Technologies). Sendai virus (SeV; Cantell strain) was purchased from Charles River Laboratories. Cells grown in 12-well dishes were infected with 182 hemagglutinin (HA) units of virus per well. Following a 1-h incubation at 37°C, cells were further incubated in serum-free medium containing 0.005% l-1-tosylamide-2-phenyl chloromethyl ketone (TPCK)–trypsin for a further 24 h before further analysis. Human IFN-α (11100-1) was purchased from PBL InterferonSource. Levels of secreted IFN-α and IFN-β from transfected cells were determined using enzyme-linked immunosorbent assays (ELISAs) from PBL Assay Science according to the manufacturer's instructions.

Plasmid constructs.

The expression plasmids encoding Flag-HBZ and HIS-Tax1 were previously described (52, 53). Plasmids encoding Flag-IRF7 (wild type), Flag-IRF7 aa 1 to 283 (Flag-IRF7 1-283), Flag-IRF7 283-503, and Flag-IRF7 283-466, together with the luciferase reporter plasmids for IFN-α and IFN-β, were gifts from Fanxiu Zhu, Florida State University (54). The pcDNA HBZ plasmid was obtained from M. Matsuoka, Kyoto University, Japan (16), and the pEGFP-HBZ-Sp1 plasmid was a kind gift from Jean-Michel Mesnard, Université de Montpellier, France (45). Plasmids encoding the kinases Flag-TBK1 and Flag-IKKε were kindly provided by M. Schröder, National University of Ireland Maynooth (43). The plasmid pcDNA-IRF3-FL, encoding IRF3, was purchased from Addgene.

Transfections.

293T cells were transiently transfected using Lipofectamine 2000 (Life Technologies) according to the manufacturer's guidelines. Transfections of HeLa cells were performed using Turbofect (Thermo Scientific) by following the manufacturer's instructions. The overall DNA concentration was normalized for all transfections using the relevant parent plasmid. Jurkat cells were transfected using Xfect transfection reagent from Clontech according to the manufacturer's instructions.

Coimmunoprecipitations.

293T cells were transfected with the relevant amounts of expression constructs using Lipofectamine 2000 (Life Technologies) according to the manufacturer's guidelines. Transfected cells were incubated for 24 h and lysed in a buffer containing 1× TBS (50 mM Tris-HCl [pH 7.5], 150 mM NaCl), 5 mM EDTA, and 1% Triton X-100 supplemented with protease inhibitors (complete protease inhibitor cocktail, EDTA free; Roche) and phosphatase inhibitors (PhosSTOP, Roche) or RIPA buffer containing 50 mM Tris-HCl (pH 8), 150 mM NaCl, 1% NP-40, 0.1% SDS, and 0.5% sodium deoxycholate supplemented with protease and phosphatase inhibitors. Cellular lysates were subjected to coimmunoprecipitation with an anti-Flag M2 resin (Sigma-Aldrich) overnight at 4°C. The beads were then washed extensively in the relevant lysis buffers. Coimmunoprecipitations were analyzed by Western blotting using appropriate antibodies.

Immunofluorescence.

HeLa cells were seeded onto chamber slides and transiently transfected with the indicated expression vectors using Turbofect transfection reagent (Thermo Scientific) according to the manufacturer's protocol. Transfected cells were incubated for 24 h. All cells were fixed using 4% paraformaldehyde for 15 min at room temperature (RT) and permeabilized using 0.5% Triton X-100–phosphate-buffered saline (PBS) for 5 min at RT. All slides were blocked in 0.02% Tween 20–5% FBS–10% goat serum in PBS for 1 h at room temperature. Endogenous IRF3 was detected using anti-IRF3 (Sc9082) incubated for 2 h at room temperature or overnight at 4°C, followed by Alexa Fluor 594-conjugated goat anti-rabbit IgG for 1 h at room temperature. Nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich), and slides were mounted in ProLong Gold antifade mountant (Life Technologies). Images were obtained by use of a Zeiss AxioImager MI fluorescence microscope and Axiocam HR camera.

Western blotting.

Cellular lysates were subjected to SDS-PAGE and analyzed by Western blotting using standard procedures. Membranes were probed using the SNAP i.d. system (Merck Millipore), using antibodies against Flag (Sigma; F7425), tubulin (Abcam; Ab7291), and Sendai virus (MBL; PD029) together with antibodies against IRF7 (Santa Cruz; Sc9083), IRF3 (Santa Cruz; Sc9082), and phosphorylated-IRF3 (Cell Signaling Technology; 29047). Anti-mouse (NA9310V) and anti-rabbit (NA9340V) secondary antibodies were from GE Healthcare.

Luciferase reporter gene assays.

293T were transfected with either IFN-α-Luc or IFN-β-Luc and different combinations of expression vectors using Lipofectamine 2000 (Life Technologies). Transfected cells were lysed in cell culture lysis reagent, and luciferase assays were performed according to the manufacturer's instructions (Promega). Luciferase values were normalized for all samples with the respective protein concentrations obtained by performing a bicinchoninic acid (BCA) assay (Pierce BCA protein assay kit; Thermo Scientific).