1 Introduction

Human T cell leukemia virus type 1 (HTLV-1) was the first retrovirus to be identified as a causative agent of a cancer in humans. In addition to cancer, adult T-cell leukemia (ATL), HTLV-1 also causes inflammatory diseases, including HTLV-1-associated myelopathy (HAM)/tropical spastic paraparesis (TSP), and HTLV-1 associated uveitis [1, 2]. HTLV-1 transmits to uninfected cells through cell conjugation, as cell-free virions are not efficient in transmission [1]. HTLV-1 increases its chance of transmission by the increase of infected cells, rather than viral replication. Subsequently, HTLV-1-encoded products can induce cellular transformation. In addition to essential retroviral components, such as long terminal repeats (LTR), gag, pol and env, HTLV-1 provirus has a unique region between env and the 3′LTR; and this region is named pX [3]. The pX region encodes viral regulatory and accessory proteins Tax, Rex, p8, p12, p13, p30, p21, and HTLV-1 bZIP factor (HBZ) which are implicated in viral infectivity and the proliferation of infected cells [3–6]. Tax is recognized as a potent oncoprotein, since it immortalizes human primary T cells by itself, and Tax transgenic mice form tumors [7–15]. Nevertheless, tax transcripts are detected in only ~40% of ATL cases [16, 17]. Recently, a viral factor, HBZ, has been shown to have an oncogenic effect in vivo [18]. Expression of HBZ is conserved in all ATL cells, strongly suggesting that it contributes to leukemogenesis.

2 New insights into the machinery of HTLV-1 infection

It is well known that living infected cells are required for the infection of HTLV-1; and this fact indicates that the mechanism of HTLV-1 transmission is quite different from that of another human retrovirus, human immunodeficiency virus 1 (HIV-1). Novel findings on the machinery used by HTLV-1 in transmission are summarized below.

2.1 Receptors

Since HTLV-1 can infect many types of cells, its receptor is thought to be a commonly expressed molecule [20, 21]. So far, three molecules, a glucose transporter, GLUT1, heparan sulfate proteoglycan (HSPG), and neuropilin-1, are thought to be important for the interaction between the HTLV-1 envelope and the cell membrane, and for entry of the virus to cells. It has been suggested that the virus may first contact HSPG and then form complexes with neuropilin-1, followed by GLUT1 association on the cell surface prior to membrane fusion and entry into the cell [22–24].

2.2 Mechanisms of cell-to-cell transmission (Fig. 1)

HTLV-1 is mainly transmitted by cell-to-cell contact, as cell-free HTLV-1 viral particles are not efficient at infection of the target cells [1, 19]. To date, several distinct mechanisms of cell-to-cell transmission of the virus have been proposed. In 2003, Igakura et al. [25] showed that HTLV-1-infected T cells attach to uninfected cells and form a virological synapse (VS), composed of viral and cellular molecules, at the point of contact. Through this structure, HTLV-1 transmits to the target cells from the donor cells. More recently, another piece of the cell-to-cell infection machinery was demonstrated; after viral budding, HTLV-1-infected cells keep viral particles on their surface, trapped in extracellular viral assemblies composed of collagen, agrin, and linker proteins [26]. When HTLV-1-infected cells covered with these viral assemblies attach to uninfected cells, the extracellular component containing the viral particles is rapidly transferred to the surface of the target cells, resulting in infection.

Fig. 1
figure 1

Mechanisms of cell-to-cell transmission of HTLV-1. a HTLV-1 virions are caught in biofilm-like extracellular assemblies surrounding infected cells, and are rapidly transferred to the surface of target cells after conjugation, resulting in infection. b An HTLV-1-infected T cell and an uninfected cell form a virological synapse (VS) at the point of contact. HTLV-1 transmits to the target cell through the VS. c HTLV-1 viral particles are transferred through conduits to target cells. HTLV-1 p8 protein induces conduit formation

Full size image

A third new mechanism of viral transfer has also been recently demonstrated. HTLV-1 encodes a protein, p8, in its pX region. The p8 protein is generated by processing of p12I [4]. By interacting with lymphocyte function-associated antigen-1 (LFA-1) and intercellular adhesion molecule 1 (ICAM-1), p8 enhances T-cell conjugation [27]. Moreover, p8 induces conduit formation among T cells and increases viral transmission through these conduits. Since the replication level of HTLV-1 in infected cells is generally low, HTLV-1 may combine multiple strategies to establish efficient viral transmission directly between cells.

2.3 Integration

After incorporation into target cells, HTLV-1 genomic RNA is reverse transcribed to generate provirus, which is integrated into cellular genomic DNA. It has been shown that integration occurs at random sites in the host genome, whereas proviruses in leukemic cells tend to be integrated near the transcriptional start sites of cellular genes, indicating positive selection of cells with this feature during leukemogenesis [28]. It has long been known that defective proviruses lacking the 5′LTR, or aberrant proviruses containing nonsense mutations in the tax gene, are occasionally observed in leukemic cells [29, 30]. Detailed analysis of the sequences of these proviruses reveals two important facts. First, deletion of the 5′LTR in the proviruses occurred before integration in some cases [31]. Since the 5′LTR is the promoter of viral genes encoded in the plus strand of the provirus, cells infected with these proviruses cannot express plus-strand genes, even at the beginning of infection. In contrast, the 3′LTR is maintained in all ATL cases [17]. Second, APOBEC3G generates point mutations in proviral DNA during reverse transcription, resulting in induction of nonsense mutations in all viral genes except for HBZ [32]. Both genetic modifications in the provirus can occur before integration, meaning that some ATL cells are derived from HTLV-1-infected cells that originally lack Tax expression, but possess HBZ.

3 HTLV-1 bZIP factor

HBZ is encoded in the minus strand of the HTLV-1 provirus [5], and constitutively transcribed from the 3′LTR [6]. HBZ is expressed in all ATL cases, whereas transcription of the tax gene is frequently inactivated by epigenetic modifications or deletion of the 5′LTR [16, 17, 31, 33]. Sequencing analysis of whole HTLV-1 provirus in 60 ATL cases revealed that only the HBZ coding sequence is preserved in all cases, despite the fact that there are many nonsense mutations or deletions in other regulatory and accessory genes [32]. These findings suggest that HBZ is a critical factor in leukemogenesis.

3.1 HBZ is pathogenic in vivo (Table 1)

We generated transgenic mice expressing HBZ (HBZ-Tg) in CD4+ T cells [6], and recently reported that there are many similarities in symptoms and immunological features between HBZ-Tg mice and HTLV-1-infected individuals [18].

Table 1 Phenotypic and immunological findings in HBZ transgenic mouse
Full size table

3.1.1 Development of T-cell lymphomas

It is known that ectopic expression of HBZ in human T cells supports cell proliferation [6, 34]. We have observed that CD4+ thymocytes from HBZ-Tg mice are more sensitive to stimulation with anti-CD3 antibody and IL-2 than those of non-Tg mice ex vivo [6]. We also found that in vivo proliferation of CD4+ splenocytes in HBZ-Tg was higher than that in non-Tg mice [18]. Most importantly, ~40% of HBZ-Tg mice developed T-cell lymphomas after a long latent period. All lymphomas in HBZ-Tg were CD3+CD4+, and their monoclonal proliferation was proven by the T-cell receptor (TCR) gene rearrangement. Interestingly, most of the lymphomas in HBZ-Tg mice expressed Foxp3, a master molecule of regulatory T cells (Tregs), although the percentage of Foxp3+ cells in each tumor was variable. This heterogeneity of Foxp3 expression in HBZ-Tg lymphomas is also observed in lymphoma tissues from human ATL patients [35]. The development of T-cell lymphomas in HBZ-Tg mice shows oncogenic potential of HBZ, indicating the presence of a cell transformation pathway common to HBZ-Tg mice and ATL cases.

3.1.2 Inflammatory complications

Most HBZ-Tg mice developed skin inflammation by 18 weeks after birth [18]. Histological examination showed infiltration of CD3+CD4+ T cells into the dermis and epidermis in the lesions. Similar infiltration of lymphocytes was observed in the alveolar septa in the lung. The incidence of these inflammatory diseases correlated with the expression level of HBZ, suggesting a role of HBZ on their pathogenesis. Similar inflammation in skin and lung is known to develop in HTLV-1 carriers [36, 37].

3.1.3 Increased regulatory T cells and effector/memory T cells

It is known that the HTLV-1 provirus is mainly detected in effector/memory CD4+ T cells [38, 39] and Tregs [40]. The number of effector/memory cells increases in HTLV-1 carriers, and correlates with HTLV-1 provirus load [39]. Tregs are CD4+CD25+ T cells, and are also increased in HTLV-1-infected individuals [40]. In HBZ-Tg mice, both CD44highCD62Llow effector/memory CD4+ T cells and CD4+Foxp3+ T cells are increased compared with non-transgenic littermates [18].

3.2 Molecular functions of HBZ (Fig. 2)

3.2.1 Inhibition of the cyclic-AMP responsive element binding protein (CREB) pathway

HBZ was originally identified by yeast two hybrid screening as an interactant of CREB-2 protein in HTLV-1-infected cells [5]. It has been shown that HBZ protein has a basic leucine zipper (bZIP) motif, and forms heterodimers with CREB family proteins [41, 42]. HTLV-1 LTRs contain three 21 bp repeats called Tax-responsive elements (TREs) and CREB recognizes these sites [43]. Tax, by binding to CREB, activates viral transcription by recruiting transcriptional cofactors such as CBP/p300 to the 5’LTR. Meanwhile, HBZ interacts with CREB and CBP/p300, and inhibits Tax-induced viral transcription by dissociating CREB from TREs and inhibiting the binding between Tax and CBP/p300 [44]. Similarly, HBZ perturbs the effect of CREB proteins upon cellular gene transcription. Recently, we reported that HBZ interacts with activating transcription factor 3 (ATF3) and interferes with the activation of p53 by ATF3, suggesting an anti-apoptotic effect of HBZ [42].

Fig. 2
figure 2

Function of HBZ RNA and protein. Left location of the coding regions of HBZ and Tax in HTLV-1 provirus. Right functions of HBZ RNA and protein in the literature

Full size image

3.2.2 Modification of AP-1 activity

Activator protein-1 (AP-1) is a transcription factor complex formed by heterodimers of cellular Jun and Fos proteins. Some AP-1 components, such as c-Jun, JunB, and JunD, have a bZIP domain, and HBZ protein heterodimerizes with them via this motif [45–49]. HBZ suppresses the activity of c-Jun and JunB. There are several mechanisms by which HBZ inhibits c-Jun; HBZ suppresses the DNA-binding activity of c-Jun and tethers it to the proteasome, resulting in ubiquitin-independent degradation of c-Jun [45, 46, 49]. In contrast, HBZ enhances the transcriptional activity of JunD. HBZ–JunD binding induces transcription of telomerase reverse transcriptase (hTERT), a catalytic subunit of telomerase [48]. Various malignant tumors, including ATL, are known to express high levels of hTERT, and this telomerase activation is thought to be associated with cellular transformation. Induction of hTERT by HBZ may be implicated in the leukemogenesis of ATL.

3.2.3 Regulation of Foxp3 expression and function

The number of FoxP3+ Tregs is increased in the thymus and spleen of HBZ-Tg mice, and most primary lymphoma tissues developed in HBZ-Tg mice express Foxp3 to varying degrees. Indeed, HBZ directly activates the Foxp3 promoter and induces its transcription [18]. On the other hand, HBZ interferes with the Foxp3 function; association between Foxp3 and NFAT is critical for the transcription of Treg-related genes such as CTLA-4 and GITR, but HBZ physically interacts with both Foxp3 and NFAT, and suppresses the function of Tregs [18], suggesting that HBZ can expand functionally impaired Treg cells and lead them to transformation in vivo.

3.2.4 HBZ inhibits the canonical NF-κB pathway

HBZ suppresses Tax-induced NF-κB activation through inactivation of an NF-κB transcription factor, p65/RelA [50]. HBZ inactivates p65/Rel by two distinct mechanisms: first, it inhibits the DNA binding ability of p65 through physical interaction; and second, it induces p65 degradation by elevating the expression of PDLIM2, an E3 ubiquitin ligase for p65. Importantly, HBZ inhibits the canonical NF-κB pathway, but not the non-canonical pathway, resulting in perturbation of the regulation of NF-κB activities.

3.2.5 HBZ RNA supports T-cell proliferation

Ectopic expression of HBZ enhances cellular growth of T cells, and knocking down HBZ in ATL cell lines attenuates their proliferation, indicating that HBZ is crucial in the continuous expansion of ATL cells [6]. It is suggested that HBZ RNA promotes cell proliferation by forming secondary stem-loop structures, like those formed by Epstein-Barr virus non-coding RNAs, EBERs [51]. HBZ RNA activates transcription of E2F1 and its target genes, and increases G1/S transition, but further studies will be required to elucidate the biological properties of HBZ RNA in more detail.

4 Tax

Tax is thought to be a potent oncoprotein, as it transforms rodent cells and immortalizes human primary T cells by itself [7–9]. Importantly, Tax transgenic mice develop spontaneous tumors and inflammation [10–15, 52, 53]. Tax enhances viral replication through transactivation of the viral promoter, the 5’LTR, and its pleiotropic functions support cellular proliferation, inhibit apoptosis, impair cell cycle checkpoints, and induce DNA damages [1]. Thus, Tax is thought to play an important role in the leukemogenesis of ATL.

4.1 Activation of the NF-κB pathway

NF-κB is a major survival pathway engaged by HTLV-1. Tax was shown to bind IKKγ, and to activate both the canonical and non-canonical pathways [54].

4.2 Cell cycle progression

Tax also induces significant mitogenic activity, especially at the G1–S-phase transition, by provoking upregulation of G1 D cyclins, activation of cyclin-dependent kinases (CDKs), and downregulation of CDK inhibitors (CKIs) [55].

4.3 Induction of aneuploidy

It has been reported that Tax can induce aneuploidy by several mechanisms [56]. Tax induces multipolar mitoses through interaction with cellular TAX1BP2 and RANBP1 proteins. Tax also impairs the mitotic spindle assembly checkpoint (SAC). Tax can bind to one of the SAC proteins, MAD1 and inactivate its function, thus causing a loss of SAC activity.

4.4 Induction of DNA damage and impairment of DNA repair

Tax can induce direct DNA damage through increased reactive oxygen species [57]. In addition, Tax inactivates p53, CHK1 and CHK2 kinases, and perturbs DNA repair by suppression of base excision repair (BER), and nucleotide excision repair (NER), resulting in accumulation of DNA damage [1].

4.5 Tax-induced cellular transformation in vivo (Table 2)

Several transgenic mice have been generated to analyze Tax function in vivo. In these models, Tax induces neoplasms such as neurofibroma, mesenchymal tumor, large granular lymphocytic leukemia, and pre-T-cell leukemia, and also inflammatory diseases like exocrinopathy, arthritis, and dermatitis in vivo [10–15, 52, 53]. Transgenic mice expressing Tax from the Lck proximal promoter were shown to develop thymus-derived immature T-cell leukemia characterized by tumor cells with hyperlobulated nuclei, immunodeficiency, and constitutive NF-κB activation; these findings resemble features of ATL [14]. In addition, Tax expression through a Lck distal promoter was shown to induce mature T-cell leukemia/lymphoma in mice [15]. Recently, it has been demonstrated that immune activation enhances Tax expression in the CD4+ T cells of HTLV-1 LTR-Tax transgenic mice, leading to immortalization of these cells [58]. When not immune activated, this strain does not develop any T-cell-associated diseases [11], suggesting that immune activation supports Tax-induced oncogenesis. Another group showed that inflammatory signals, such as TCR stimulation, accelerate tumor promotion by Tax in mice [59]. The link between immune stimulation and oncoproteins such as Tax may be important in the oncogenic process in vivo.

Table 2 Phenotypes of Tax transgenic mice
Full size table

Cancer stem cell (CSC) theory proposes that even a small number of CSCs can generate a tumor, due to their self-renewal properties and potent proliferative potential, and tumors of various tissue types are thought to be initiated from CSCs. It has been reported that Tax-expressing or HTLV-1-infected human hematopoietic stem cells can develop CD4+ T-cell lymphomas after transplantation to immunodeficient mice [60]. Tax-expressing CSCs were also identified in Lck proximal promoter-Tax transgenic mice [61]. These findings suggest the possibility that Tax can target somatic stem cells and utilize their proliferative properties for transformation.

5 Concluding remarks

Intensive studies on Tax since the discovery of HTLV-1 have revealed some molecular strategies used by HTLV-1 for viral replication and cellular transformation [62]. However, the precise mechanisms of viral transmission and leukemogenesis have yet to be clarified. Emerging evidence is highlighting the previously unknown mechanisms of viral pathogenesis. Further studies will be needed to develop new treatment and prophylaxis strategies based on the growing knowledge of HTLV-1 molecular biology.