Molecular Biology of HTLV-I: Recent Progress

Human T-cell lymphotrophic virus type I (HTLV-I) (1,2) is a causative agent of adult T-cell leukemia (ATL) (3) and HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) (4,5). In ATL cells, the sites of provirus integration were not common among patients, although they were clonal within individuals (3), so the viral transacting factor was implicated as a leukemogenic factor. Tax protein has been proposed as a candidate for this transacting factor (6), since Tax can immortalize human T cells (7), transform rodent fibroblastic cell lines (8), and induce tumors in its transgenic mice (9). Tax is also identified as a transacting activator of viral genome transcription (10-12) and thus is essential for viral replication. Tax also activates many specific cellular genes, including lymphokines (13-15), some of their receptors (13), and nuclear protooncogenes (16,17), so some of these genes have been thought to contribute to T-cell immortalization or transformation. We reported previously that Tax activates three enhancers (Fig. 1), including the 21-base-pair (bp) sequence of HTLV (18,19), the nuclear factor (NF)-KB-binding site of the interleukin (IL)-2 receptor α gene (20), and the serum-responsive element (SRE) of the c-fos and c-egr protooncogenes (21). The mechanism of the activation was shown to be the binding of Tax to transcription factors that bind to specific enhancers. These factors are the cAMP-responsive element binding (CREB) (22,23) or modulator (CREM) protein (22) that binds to the 21-bp sequence; NF-KB family proteins including p50 (24), p52 (25), p65 (26), and c-Rel (26) for the NF-KB-binding site; and the serum-responsive factor for the SRE (21). The complex of CREB and Tax on the specific enhancer DNA suggests an interesting mechanism. In normal activation of CRE, CREB protein is phosphorylated by protein kinase A, which responds to cAMP signaling (27,28), and only phosphorylated CREB protein can bind to CREB-binding protein (CBP) and activate the transcription (29). Therefore, phosphorylation of CREB protein at the specific site plays a key role in the activation of transcription. On the other hand, Tax binds to unphosphorylated CREB (23) and may mimic the function of CBP. This suggests that Tax binding to CREB could bypass the normal regulation by phosphorylation of CREB protein. Therefore, Tax can activate the 21-bp-dependent transcription in the absence of extracellular signals, which could result in the constitutive expression of genes directed by the 21-bp-like enhancers, including the HTLV-I genome. Tax BINDING TO IKBα AND NUCLEAR TRANSLOCATION OF NF-κB In the activation of NF-κBT we found an additional mechanism for the activation. In resting cells, NF-κB proteins are complexed with inhibitory proteins, including IκBα, IkBβ, or IκBγ, and are retarded in the cytoplasm (30). Signals for cellular proliferation or differentiation activate a kinase that phosphorylates IκB (31), although it is not identified yet. In the case of IκBα, the phosphorylation induces destabilization IκBα (32). The released NF-κB proteins are consequently translocated into the nucleus and activate transcription. Therefore, a serious question for us was how Tax can activate NF-κB in the nucleus without extracellular signals. The first interaction found between Tax and the NF-κB family was Tax binding to the NF-κB precursor p105 (33). Then we found that the Tax-binding site on p105 was the ankyrin motif (34), which is also conserved in IκB. Tax binding to IκB proteins was demonstrated using GST fusion proteins of IκBγ and IκBα in vitro (35). In vivo, NF-κB was localized in the cytoplasm, forming complex with IκBα in the presence of IKBα. Coexpression of Tax induced expression of NF-κB in the nucleus; however, a Tax mutant that does not bind to IκBα did not affect the cytoplasmic localization of NF-κB. We saw drastic reduction of the complex IκBα-NF-κB in these cells, clearly indicating that Tax binds to IκBα and results in nuclear translocation of c-Rel (Fig. 2). On the other hand, the effect of Tax on the stability of IκBα protein was also analyzed by Western blot after blocking the protein synthesis by cyclohexamide (35). When IκBα alone was expressed, IκBα was unstable and the protein was not detected after 45 min, but when NF-κB was coexpressed, IκBα protein was remarkably stabilized and a significant amount of IκBα protein was still detectable even after 8 h. Interestingly, when Tax was expressed, IκBα protein became unstable: namely, the stabilization effect of NF-κB was completely canceled by Tax and the stability became almost the same as that of free-form IκBα. As expected from these results, Tax alone did not affect the stability of IKB without NF-κB protein. Therefore, we concluded that Tax binds to IκBα and destabilizes the IκBα-c-Rel complex, resulting in the nuclear translocation of NF-κB. IκBα protein is originally unstable (32), so the complex between IκBα and Tax could not be detected. By this effect, Tax can bypass normal regulation through extracellular signals. Following this effect, Tax can then bind to NF-κB in the nucleus and potentiate the NF-κB activity in transcriptional initiation. Therefore, the effects of Tax on the NF-κB activation are twofold. On the other hand, NF-κB activates IκBα gene transcription (36), thus producing more IκBα (Fig. 2). This induction of IκBα expression makes NF-κB activation transient. In HTLV-I-infected cells, however, Tax destabilizes the IκBα-NF-κB complex and thus breaks the feedback control of NF-κB regulation. Therefore, if this is the case, increased expression of mRNA, but decreased expression of IκBα protein, is predicted in the presence of Tax. This was in fact demonstrated in HTLV-I-infected cell lines (36). Western blot analysis indicated a low expression of IκBα protein in HTLV-I-infected T-cell lines when compared with uninfected T-cell lines. In contrast to the protein, mRNA expression was greatly increased. This is good evidence that Tax is really functioning in the destabilization of the IκBα complex to maintain constitutive activation of NF-κB in infected T cells. Tax BINDING TO THE CELL CYCLE INHIBITOR p16INK4A The Tax-binding domain of the IκBα was identified as the ankyrin motif (34). Thus, it was postulated that other proteins carrying the ankyrin motif may be direct targets of Tax binding. Among many possible proteins, we were interested in a cell cycle inhibitor, p16INK4A, which has an ankyrin motif and binds to CDK4 kinase, keeping the kinase inactive (37)(Fig. 3). When cyclin D is available, the cyclin D forms a complex with CDK4 and then the CDK4 becomes active. The activated CDK4 phosphorylates the tumor-suppressor retinoblastoma (Rb) protein (38). Unphosphorylated or hypophosphorylated Rb protein binds to a transcription factor, E2F, which is kept inactive, but phosphorylation of Rb dissociates the complex and releases the active E2F (39). Consequently, the released E2F binds to the target-enhancer sequence on the DNA and activates various genes whose functions are required for cells to progress from the G1 to the S phase. Therefore, p16INK negatively regulates the upstream Rb pathway, which controls the G1 phase of the cell cycle. The binding of Tax to p16INK was shown in vitro by using GST-p16INK fusion protein and in vivo by immunoprecipitation. When CDK4 was expressed by transfection, significant kinase activity was demonstrated, but this activity was inhibited by additional expression of p16INK. Further expression of Tax recovered the kinase activity rather efficiently. These results clearly indicate that Tax binding to p16INK activates CDK4. The mechanism of CDK4 activation was shown by the Tax-induced reduction of the inactive complex of CDK4 and p16INK through binding to p16INK. Therefore, it was concluded that Tax binds to p16INK and suppresses its inhibitory function, thus eventually activating the CDK4 kinase. It was also demonstrated that Tax can compete with cell cycle arrest at the G1 phase induced by p16INK, inducing cell proliferation. These results clearly indicate that Tax binding to p16INK suppresses its inhibitory function and activates CDK4 kinase enabling complex formation with cyclin D and thus inducing cells to progress from the G1 to the S phase. Therefore, this functional inactivation of p16INK by Tax is expected to contribute to ATL development. p16INK is recognized as a tumor-suppressor gene, MTS1, and is frequently deleted from many human cancer cells (40). This strongly suggests that functional suppression of p16INK by Tax is taking place in infected T cells. Another aspect of interest is that Tax inactivates the signaling pathway that includes Rb. Many human tumors have mutations either in Rb or p16INK, indicating that this signaling pathway plays a critical role in maintaining cells in a normal state. Oncogenic proteins of DNA tumor viruses bind to Rb (41) and promote the cell growth. Therefore, our findings imply that HTLV-I Tax inactivates the Rb pathway complementary to DNA tumor viruses targeting different molecules. That such complementary inactivation was seen in spontaneous mutations or deletions of these two genes inhuman cancer cells strongly suggested that inactivation of p16INK by Tax protein might be critical in ATL development. POSSIBLE MECHANISMS OF T-CELL IMMORTALIZATION As just discussed, Tax activates transcription through at least three enhancers: the 21-bp enhancer, the NF-κB-binding site, and the SRE. Furthermore, Tax binds to p16INK, a cell cycle inhibitor, thereby suppressing its inhibitory function and enhancing cell proliferation. Apparently, there is no doubt that the interaction of Tax with p16INK is more directly linked to the promotion of cell growth and immortalization of infected T cells. Is the Tax-mediated transcriptional activation of specific genes linked to T-cell immortalization or transformation? If so, which pathway is responsible? Or do all of these pathways contribute to ATL development (Fig.2)? The pathway through the 21-bp enhancer is essential for the efficient replication of HTLV-I (10-12), but no cellular genes directed by the 21-bp or 21-bp-related enhancer have been identified. On the other hand the NF-κB pathway includes activation of many lymphokines and some of their receptors. Because these lymphokine signals are involved in proliferation and differentiation of T cells, it is rational to expect that Tax-mediated activation of this pathway may contribute directly or indirectly to the immortalization or transformation of HTLV-I-infected T cells. On the other hand, activation of the SRE pathway includes activation of immediate early nuclear protooncogenes, which are also expected to be involved in the promotion of cell growth in general. However, these are only predictions or expectations and not direct evidence that Tax-mediated activation of one or some of these genes is directly responsible for the induction of T-cell growth. Nevertheless, Smith and Greene (42) reported that a Tax mutant that activates the 21-bp pathway, but not the NF-κB pathway, transformed mouse fibroblasts in anchor-independent fashion, thus proposing that the NF-κB pathway is dispensable in the transformation capacity of Tax. In the Seventh Human Retrovirology Meeting in Paris, Grassmann and colleagues (43) reported that this was also true in immortalization of human T cells in vitro. Therefore, even in T-cell immortalization, activation of the 21-bp pathway rather than the NF-κB pathway, seems to be critical. However, a complementary experiment using a Tax mutant that is inactive in the 21-bp pathway, but active in the NF-κB pathway, would be essential to exclude the possibility that either pathway is sufficient for T-cell immortalization or transformation. The question of the identity of the target gene(s) of Tax-mediated activation for immortalization of T cells is the most interesting one in tumorigenesis. Another question regarding Tax function is whether Tax is required for maintenance of immortalized or transformed phenotypes. In this respect, Shmitt et al. (44) showed clearly that Tax was required for the immortalized phenotypes of T cells to suppress cell growth when Tax expression was canceled in T cells immortalized by the conditional expression system of the Tax gene. This was a rather expected, but very important demonstration of Tax function. However, this does not necessarily imply that Tax is required for the maintenance of the leukemic phenotypes of ATL cells. In vivo, most infected T cells, including leukemic and nonleukemic cells, do not express significant levels of Tax, so its expression can be detected only by reverse-transcription polymerase chain reaction (RT-PCR) of HTLV-I mRNA. More than 90% of infected cells in peripheral blood are negative by RT-PCR of Tax/Rex mRNA (45). Thus, it is likely that, in the absence of Tax, the ATL cells in the periphery maintain leukemic phenotypes. Of course, there are other arguments for example, (a) a trace amount of Tax might be sufficient for the phenotypes of leukemic cells, (b) the infected cells in the periphery do not divide and thus do not represent replicating leukemic cells, or (c) Tax might be expressed in other tissues than peripheral lymphocytes. Nevertheless, it might be proposed that Tax functions at the early stage before the leukemic transformation of infected T cells and promotes cell growth through transcriptional enhancement. Such an effect of Tax would be reversible, but the responsive gene(s) is not yet known. Apparently, more investigation is required: How many other enhancers are activated by Tax? What are they? How can we distinguish these activation pathways in T-cell immortalization or transformation? Are other cell cycle inhibitors direct targets of Tax? How about the other molecules involved in other checkpoints of cell cycle? Are all of these phenomena discovered in vitro taking place in vivo in HTLV-I-infected individuals or patients? If these are the rather early events in infected T cells during ATL progression, which are critical after Tax?FIG. 1: . Indirect binding of Tax to three different classes of transcriptional enhancers through specific transcription factors. These bindings enhance transcription, bypassing negative regulation in unstimulated cells.FIG. 2: . Tax binding to IκBα protein in cytoplasm, inducing nuclear translocation of NF-κB. Stimulating signals destabilize the IκBα-NF-κB complex, but Tax binding to IκBα in the complex destabilizes the complex without any extracellular signals.FIG. 3: . Tax binding to the cell cycle inhibitor p16INK, suppressing inhibitory activity. Tax binding to p16INK activates CDK4 and phosphorylates Rb. The phosphorylation destabilizes the Rb-E2F complex, releasing active transcription factor E2F to express factors required for G1 to S-phase transition.