Human T-cell leukemia virus type 1 (HTLV-1) was the first retrovirus proven to be associated with human disease. The infection of HTLV-1 causes adult T-cell leukemia (ATL) and chronic inflammatory diseases including HTLV-1-associated myelopathy (HAM)/tropical spastic paraparesis (TSP; Uchiyama et al., 1977; Poiesz et al., 1980). HTLV-1 belongs to deltaretrovirus family, which include HTLV-2, HTLV-3, and HTLV-4; the simian T-cell lymphotropic viruses (STLV-1, STLV-2, and STLV-3) and the bovine leukemia virus (BLV; Matsuoka and Jeang, 2007). Like other retroviruses, the HTLV-1 proviral genome has structural genes, gag, pol, and env, flanked by long terminal repeat (LTR) at both ends (Figure 1; Matsuoka and Jeang, 2007). The characteristic of HTLV-1 proviral genome is the presence of pX region between env and 3′ LTR, and encoded several accessory genes, which include tax, rex, p12, p21, p30, p13, and HTLV-1 bZIP factor (HBZ; Gaudray et al., 2002; Nicot et al., 2005). HBZ was first identified as an antisense viral protein that associated with cAMP-response element binding protein-2 (CREB-2) in HTLV-1-infected cells (Gaudray et al., 2002). We found that the expression of HBZ is conserved in all ATL cells (Satou et al., 2006), and non-sense mutations of all HTLV-1 genes except the HBZ gene were generated byAPOBEC3G before integration of the provirus (Fan et al., 2010). In addition, the HBZ gene promotes the proliferation of ATL cells (Satou et al., 2006). Thus, the HBZ gene is believed to be essential for leukemogenesis. In this review, we discuss the updated understanding of HTLV-1 and focus on the roles of HBZ in HTLV-1 oncogenesis.

The existence of a viral transcript encoded by the minus strand of the HTLV-1 provirus was reported in 1989 (Larocca et al., 1989). Thereafter, a viral protein that binds to CREB-2 was identified by yeast 2-hybrid screening system, and termed as HTLV-1 bZIP factor (HBZ; Gaudray et al., 2002). The 5′ RACE experiment identified the transcription start sites of HBZ were located in the U3 and R regions of the HTLV-1 3′ LTR (Satou et al., 2006). 5′ LTR of HTLV-1 was frequently lost and methylated in ATL cases, while 3′ LTR was not deleted and remained unmethylated (Matsuoka and Jeang, 2007). It may be one of the reasons why HBZ is consistently expressed in all ATL cells. Two transcripts of the HBZ gene have been revealed: a spliced form (sHBZ) and an unspliced form (usHBZ; Figure 1). The promoter of the spliced and unspliced HBZ transcripts was TATA-less and contained initiators and downstream promoter elements (Yoshida et al., 2008). Detailed analyses of the promoter of sHBZ gene showed that Sp1 sites were critical for its transcription. U3 region of LTR in the both end of HTLV-1 proviral genome contains Tax-response element (TRE), which acts as an enhancer of viral transcription. Tax and CREB can activate transcription of the HBZ gene through 3′ LTR, meanwhile, Tax regulated antisense promoter activity is influenced by the integration site. However, Tax induced 3′ LTR activation is not as significant as it did on the transcription of the HTLV-1 plus strand (Landry et al., 2009).

The spliced and unspliced HBZ genes are translated into a polypeptide of 206 and 209 amino acids, respectively. Both forms of HBZ contain three domains: N-terminal activation domain (AD), central domain (CD), and basic ZIP domain (bZIP) in the C-terminal (Figure 2). There are two LXXLL-like motifs located within the N-terminal AD domain of HBZ. Both are important for the binding to p300/CBP (Clerc et al., 2008). The HBZ protein is localized in the nucleus with a speckled pattern. Three nuclear localization signals (NLSs) were identified to two regions in the CD of HBZ and a basic region of the bZIP domain (Hivin et al., 2005). Moreover, the integrity of HBZ protein sequence was necessary for its accumulation in nuclear speckles. HBZ specifically distributed in the nucleus to heterochromatin, while it is not associated with heterochromatin (Hivin et al., 2005).

The only difference between the protein form of sHBZ and usHBZ are seven amino acids in their N-terminal, while it causes significant difference between these two proteins. The half-life of usHBZ protein is much shorter than that of sHBZ (Yoshida et al., 2008). In ATL cell, only sHBZ protein could be detected, since sHBZ gene is more predominant than usHBZ gene (Usui et al., 2008).

The transcription of HTLV-1 viral gene is the symbol of HTLV-1 infection. Among all of the HTLV transcripts, the tax gene is thought to play critical role in leukemogenesis because of its pleiotropic functions. However, about 60% of ATL cases lost the tax expression by three mechanisms: (1) deletion of HTLV-1 5′ LTR, (2) hypemethylation of 5′ LTR, and (3) non-sense mutation, deletion, and insertion of tax gene (Matsuoka and Jeang, 2007). Moreover, the loss of Tax is believed to occur during the process of oncogenesis. It indicates that Tax expression is not always necessary for ATL. Unlike 5′ LTR, HTLV-1 3′ LTR is intact and remains unmethylated. The HBZ gene expression was conserved in all ATL cases. In addition, expression of the HBZ gene could be detected not only in ATL cells but also in HTLV-1-infected carriers, furthermore, its expression are well correlated with proviral load. It has been reported the correlation between the level of the HBZ gene and severity of HAM/TSP, indicating that HBZ might contribute to the development of HAM/TSP except for ATL (Saito et al., 2009).

The two forms of the HBZ gene are not equally expressed in ATL cells. The expression level of the sHBZ gene was found to be four times higher than that of the usHBZ gene transcript in HTLV-1 carriers and ATL cases.

In PBMCs from HTLV-1-infected rabbits revealed that the tax gene was expressed at the highest levels immediately after infection and progressively decreased, eventually stabilizing at low levels. However, HBZ was detectable at low levels early after infection, and slowly increased and stabilized. At 8 weeks post infection, the HBZ mRNA was expressed at the highest concentration of all mRNA measured (an average of ninefold more than tax mRNA; Li et al., 2009). Additionally, Suemori et al. (2009) reported that the expression levels of HBZ mRNA and HBZ protein in HTLV-1-infected cells were not parallel, suggest that the post transcriptional regulation of HBZ gene may differ among different cell types. However, other group detected the consistently high level of HBZ mRNA and protein in ATL- and HTLV-1-associated cell lines (Satou et al., 2006; Zhao et al., 2011). The physiological level of HBZ could regulate the cellular signaling pathway, which is critical for T-cell differentiation. The discrepancy between two independent studies may due to the different experiment condition and different HBZ antibody they used.

Accumulating evidence showed that Tax and HBZ are two important factors for ATL. The crossover and the interaction between these two proteins are important for the HTLV-1 oncogenesis. Tax could transform rodent cells and immortalizes human primary T cells, moreover, it enhances viral replication, supports cellular proliferation, inhibits apoptosis, impairs cell-cycle checkpoints, and induces DNA damages. Although Tax has pleiotropic functions as mentioned above, ~60% of all ATL cases did not express Tax. Moreover, abortive genetic changes, including deletions, insertions, and non-sense mutations, were frequent in all viral genes include the tax gene. Non-sense mutations in the tax gene were also observed in proviruses from asymptomatic carriers. Contrary to the tax gene, HBZ is consistently expressed in all ATL cells and remains intact through the whole period of ATL development. Tax plays a central role in the immortalization of T-lymphocytes in cell culture, on the contrary, HBZ was dispensable for viral-induced immortalization of primary human T cells (Arnold et al., 2006). Tax is the major target of cytotoxic T-lymphocytes (CTLs) in vivo (Kannagi et al., 1991), while HBZ is not a target antigen for HTLV-1-specific CTLs (Suemori et al., 2009). Since HBZ promote the T-cell proliferation and induced lymphoma and chronic inflammation which is similar to those in HTLV-1-infected individuals, we believe HBZ play more important role in the maintenance of HTLV-1-induced transformation.

Interestingly, Tax and HBZ have opposite effect on the regulation of cellular signaling pathways (Table 1). Tax activates 5′ LTR transcription; enhances AP-1 and classic NF-κB signaling, and suppresses TGF-β pathway. Contrary, HBZ suppressed Tax-mediated HTLV-1 viral transcription; inhibit activated AP-1 and classical NF-κB signaling. However, HBZ enhances TGF-β pathway. HTLV-1 may take advantage of this opposite function of Tax and HBZ on regulating signaling pathway to allowing the better survive of HTLV-1-infected cells and escape the immune surveillance.

There are four kinds of HTLVs have been identified, and their oncogenic property is different. HTLV-1 mainly infected CD4⁺ T cells. However, HTLV-2 tends to target CD8⁺ T cells, and induces cancer only in rare cases. Unlike Tax1, Tax2 which encodes by HTLV-2 does not have PDZ domain-binding motif (PBM). Since PBM is important for the transforming activity, such differences between Tax1 and Tax2 may responsible for the different pathogenicities of HTLV-1 and HTLV-2 (Hirata et al., 2004). It was reported that HTLV-2 produced an antisense transcript termed antisense protein (ASP) of HTLV-2 (APH-2; Figure 3; Halin et al., 2009). Like the sHBZ gene, APH-2 mRNA is transcribed from the 3′ LTR, and spliced and polyadenylated. APH-2 does not contain bZIP domain but it still could interact with CREB and suppresses Tax2-dependent LTR activation. In addition, APH-2 did not colocalize with the nucleolus, although in non-T and T cells, it did have nuclear localization. Importantly, APH-2 was correlated with proviral load, while it does not promote cell proliferation and does not cause lymphocytosis.

Recent reports showed that HTLV-3 and HTLV-4 produce a spliced and polyadenylated antisense transcript (Barbeau and Mesnard, 2011). These two proteins were named as APH-3 and APH-4, respectively, since both proteins lack consensus bZIP domain (Figure 3). But, they retained the capacity to inhibit LTR activation that mediated by their respective Tax proteins. This study demonstrated a nuclear-restricted pattern for APH-4 while APH-3 was localized both in the cytoplasm and in the nucleus.

A naturally occurring antisense RNA was identified in the cell lines infected with HIV-1, and encodes an ASP with an apparent molecular mass of 19 kDa (Figure 3; Vanhee-Brossollet et al., 1995; Clerc et al., 2011). Antibodies against this protein have been detected in the sera of HIV-positive individuals. However, it remains to be elucidated the function of this ASP in vivo.

Intensive studies on the leukemogenesis of ATL since the discovery of HTLV-1 have revealed some molecular mechanisms by which viral proteins induce viral replication and cellular transformation. However, the precise mechanisms of oncogenesis by HTLV-1 remain to be clarified. Recent progresses on the researches of the HBZ gene give new ideas into the mechanisms of oncogenesis. HBZ is the only viral gene that is expressed in all ATL cases.Except the HBZ gene, non-sense mutations of other viral genes are detected before integration of provirus. Importantly, HBZ maintains the cell survival when Tax is silenced. These data collectively indicate that HBZ protein has predominate function on the oncogenesis of HTLV-1. Further studies are needed to clarify the oncogenic function of HBZ and develop therapy targeted against HBZ based on the growing knowledge of HTLV-1 molecular biology.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.