HTLV-1 infection can remain latent for years before full disease onset, suggesting that there are sophisticated mechanisms regulating the on/off switching of viral protein expression, but also an increase in function of genes related to the T cell receptor/NF-kB and signalling related to immune surveillance such as HLA and FAS [
46]. The outcome and progression of infection is regulated by Tax, whose expression is also critical for stimulating the cytotoxic T-lymphocytes immune response which is thought to play a role in the viability of ATL long-term survivors [
47]. Loss or reduction of Tax expression in immortalised cells protects them from immune response attack and renders them vulnerable to survival, expansion, and proliferation through a process of continuous transient expression [
48]. However, Tax is not the only responsible for the various changes observed in viral expression and cell fate during infection. As already indicated in
Section 3, HBZ also plays an important role by counteracting Tax to some extent and, on the other hand, integrating the process of virus-induced leukemogenesis [
49]. In fact, HBZ does not directly affect HTLV-1-mediated immortalization: rather, it regulates the establishment and maintenance of chronic infection [
50]. Expression of hbz has been shown to upregulate JunD abundance and HBZ heterodimerizes with JunD, which recruits the transcription factor Sp1 to the 3’ LTR of the provirus to enhance its activity [
51]. Moreover, HBZ is indirectly involved in leukemogenesis by increasing the expression of two oncogenic miRNA, miR-17 and miR-21. These miRNA in turn downregulate the expression of the single-stranded DNA-binding protein hSSB2, thus promoting genomic instability [
52]. Another indication of the possible involvement of HBZ in neoplastic transformation is the detection of its translocation to the nucleus in cells from ATL patients [
53]. Other accessory proteins play a selective role in HTLV-1 leukemogenesis. The p30 counteracts viral transformation by inhibiting the export of tax/rex mRNA from the nucleus [
54], while p13 is involved in the transformation by increasing the intracellular content of reactive oxygen species [
55]. On the other hand, p12 enhances STAT5 activation in transduced peripheral blood mononuclear cells, allowing them to proliferate even in the presence of a low IL -2 concentration [
56]. Thus, a number of different mechanisms underlie HTLV-1 latency/leukemogenesis involving both viral and cellular signals, although the major players are Tax and HBZ proteins. Apparently, Tax is involved in the initiation of immortalization/transformation, whereas HBZ is required for the maintenance of the transformed stage. Nevertheless, the differential low expression of Tax and HBZ at different stages of infection appears to favour the establishment of a dynamic, rather than static, state of latency and persistence. HBZ has been shown to inhibit viral sense transcription and favour the entry of HTLV-1 provirus into the latency phase [
57]. Ex vivo experiments have shown that after HTLV-1 provirus integration into the cell, its replication essentially consists of a series of mechanisms based on successive phases of burst and reactivation alternating with viral latency and controlled by Tax, with the expression of pro- and anti-apoptotic genes playing a central role over cell cycle gene expression [
58]. In addition, the role of Tax in the switch between life and death leading to the selection of immortalized clones as the final phase of the HTLV-1 transformation process has been investigated by experimental in vitro models (see below). However, an important point in this context is that seemingly conflicting results suggest that the role of Tax in controlling apoptosis is not clear. Indeed, several studies have supported a preventive effect of HTLV-1 gene expression against apoptosis in infected cells and have shown a reduced susceptibility to apoptosis induced by anti-Fas [
59] and TNFα [
60]. This resistance to apoptosis has been associated with activation of NF-κB via the Tax protein in HTLV-1 infected cells [
45], or with Tax-induced repression of pro-apoptotic genes [
61], or expression of anti-apoptotic genes [
62]. However, despite the evidence for the prevention of cell death during HTLV-1 infection, other studies reported that the viral protein Tax not only exerts some anti-apoptotic effects but is also responsible for promoting apoptosis in HTLV-1-infected cells under certain experimental conditions, as also mentioned in the previous paragraph. Clear evidence for this phenomenon was first provided by a study reporting that the expression of Tax in an inducible system induced rather than inhibited cell death [
63]. The same and other authors even provided mechanistic details on apoptosis induction by Tax expression, showing that Tax-triggered death was dependent on ICEprotease [
64], related to upregulation of the Fas ligand (FasL) gene [
65], and presumably dependent on nuclear expression of the CREB-binding protein (CBP)/p300-binding domain of the HTLV-1transactivator Tax [
66]. In another study, Tax was shown to induce cell death through NF-κB-mediated activation of tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) [
67]. Taken together, these data fit the apparent paradox of a dual role of Tax for apoptotic cell death in the early phase preceding the completion of the transformation process triggered by HTLV-1. To investigate this apparent paradox, some of us pioneered the experimental model of HTLV-1 infection in vitro several years ago. Our model consisted of long-term culture of PBMCs from healthy donors after exposure to irradiated HTLV-1-infected cells as viral cell donor [
68]. We felt that such a model was the best strategy to recapitulate in vitro what occurs naturally in vivo in HTLV-1-infected patients during the long period of immortalization and transformation that could lead to ATL. Indeed, the results of this study have shown that in vitro HTLV-1 infection of mononuclear cells induces a high input of cell proliferation during the first weeks with concomitant high susceptibility to regulated cell death (RCD) through apoptosis and a progressive decrease in cell death rates during the phase of long-term culture towards immortalization [
68]. Moreover, in situ hybridization showed that the cells undergoing apoptosis were indeed infected and had HTLV-1 proviral DNA, rather than residual uninfected cells [
68]. Although this event was observed in a mass population, we hypothesized that some infected cells died by apoptosis because pro-apoptotic signals prevailed in response to infection triggered by viral gene expression, whereas the infected cells that survived were protected by the prevalent expression of anti-apoptotic genes, which were themselves triggered by viral proteins. Moreover, based on our results, we suggested for the first time that induction of massive apoptosis in response to infection might act as a selection pressure for the emergence of anti-apoptotic, well-endowed infected clones prone to immortalization [
68]. A more recent study performed at the single cell level in the MT -1 ATL cell line attempted to further elucidate the same paradox of Tax's opposing effects on cell death by demonstrating a situation in which Tax expression affects susceptibility to apoptosis. The results showed that a balance between antiapoptotic and proapoptotic genes depended on the on/off of Tax expression in the cells used in this study. High expression of Tax was preferentially associated with an antiapoptotic gene expression scenario, whereas low or absent expression of Tax was associated with greater susceptibility to apoptosis. This switch was continuous in culture and due to the coexistence of different expanding clonesnt expanding clones [
69]. To explain the dual role of Tax in ATL, in which cells progress in the cell cycle despite low Tax expression, the results of another study hypothesized that ATL cells acquire genetic/epigenetic alterations during the transformation process that can bypass the Tax/NF-kB-dependent induction of senescence [
70]. However, it should be considered that the above studies were performed on ATL cells in which the HTLV-1 transformation process had already been completed. Thus, the role of Tax in these cells could explain what happens in the cells of ATL patients in an advanced phase of the disease to maintain the leukemic state and not describe the role of Tax in the long premalignant phase of HTLV-1 infection, characterized by oligo-/polyclonal expansion of non-malignant HTLV-1-infected cells, which precedes overt ATL in patients and actually involves HTLV-1-driven oncogenesis. A recent study investigated the association between early transcription of the positive viral strand, viral burst,and expression of pro- and anti-apoptotic genes in two naturally infected T cell clones from patients transduced with a Tax-responsive timer protein [
58]. The results showed that anti-apoptotic genes were expressed during the early positive-strand virus burst, followed by a phase in which proapoptotic genes outlasted the virus burst [
58]. Another study of naturally HTLV-1-infected patient Tax-expressing T cell clones showed that high Tax expression occurred during the burst phase, immediately followed by a phase of Tax expression heterogeneity associated with poor proliferation, slow cell cycle, and high susceptibility to apoptosis [
71]. Conversely, Tax-expressing clones showed long-term increases in proliferation and decreases in apoptosis [
71]. Interestingly, apoptosis may play a role in the awakening of the virus from latency. It has been provocatively hypothesized that in cells that become apoptotic RCD in response to infection, apoptosis itself may awaken viral latency and promote viral replication through direct upregulation of caspase 9, which, in addition to its proapoptotic effect, may form a complex with Sp1-p53 and activate viral LTR [
72]. On the other hand, viral reactivation from latency could also be triggered by metabolic changes such as hypoxia, which has been shown to increase transcription of the HTLV-1 proviral plus strand; conversely, inhibition of glycolysis and mitochondrial electron transport chain hinders transcription of the proviral plus strand [
73]. Interestingly, the results of these recent detailed studies appear to be essentially consistent with the observations reported above. Overall, the results of these studies suggest that both the apoptotic RCD signaling pathway and the metabolic pathway in HTLV-1 infection may represent potential targets for the development of molecules to reverse latency.