Life, death, and tax: role of HTLV-I oncoprotein in genetic instability and cellular transformation.

Human T-cell leukemia virus type I (HTLV-I) causes adult T-cell leukemia (ATL) (1–3). The virus is also associated with a neuropathy/ myelopathy termed HTLV-associated myelopathy and tropical spastic paraparesis. ATL develops in 2–5% of HTLV-I-infected individuals after a long latent period, suggesting a multistage process of immortalization and transformation of T-lymphocytes. Extant data suggest that 8 discrete events likely occur serially in vivo before an HTLV-I-infected cell becomes immortalized and transformed (4). How HTLV-I infection progresses from clinical latency to T-cell malignancy is not well understood but involves the unique viral transactivator/oncoprotein, Tax (Fig. 1). Tax has been shown to be singly sufficient for immortalizing T-lymphocytes (5, 6) and transforming rat fibroblasts (7). Further, transgenic mice expressing Tax (driven by the HTLV-I long terminal repeat (LTR)) develop neurofibroma, a tumor of mesenchymal tissue (8). Finally, large granular lymphocytic leukemia has been found in mice transgenic for Tax expressed from the T-cell specific, granzyme B promoter (9). It is estimated that cells in the human body divide 10 times during a lifetime. To control and prevent errors in cell divisions, mammalian cells have evolved “gatekeepers” and “caretakers” to regulate the rate of cell growth and the fidelity by which cellular genetic information is transmitted to progenies (10). Gatekeepers monitor the net proliferative capacity of a cell, whereas caretakers act to eliminate DNA damages. Accordingly, one perspective is that transformation occurs when both gatekeeper and caretaker functions are abrogated. Using HTLV-I as a model, we review in a non-exhaustive fashion current thoughts on how Tax perturbs normal cellular regulation and engenders cellular transformation.

Human T-cell leukemia virus type I (HTLV-I) 1 causes adult T-cell leukemia (ATL) (1)(2)(3). The virus is also associated with a neuropathy/ myelopathy termed HTLV-associated myelopathy and tropical spastic paraparesis. ATL develops in 2-5% of HTLV-I-infected individuals after a long latent period, suggesting a multistage process of immortalization and transformation of T-lymphocytes. Extant data suggest that 8 discrete events likely occur serially in vivo before an HTLV-I-infected cell becomes immortalized and transformed (4). How HTLV-I infection progresses from clinical latency to T-cell malignancy is not well understood but involves the unique viral transactivator/oncoprotein, Tax (Fig. 1). Tax has been shown to be singly sufficient for immortalizing T-lymphocytes (5,6) and transforming rat fibroblasts (7). Further, transgenic mice expressing Tax (driven by the HTLV-I long terminal repeat (LTR)) develop neurofibroma, a tumor of mesenchymal tissue (8). Finally, large granular lymphocytic leukemia has been found in mice transgenic for Tax expressed from the T-cell specific, granzyme B promoter (9).
It is estimated that cells in the human body divide 10 16 times during a lifetime. To control and prevent errors in cell divisions, mammalian cells have evolved "gatekeepers" and "caretakers" to regulate the rate of cell growth and the fidelity by which cellular genetic information is transmitted to progenies (10). Gatekeepers monitor the net proliferative capacity of a cell, whereas caretakers act to eliminate DNA damages. Accordingly, one perspective is that transformation occurs when both gatekeeper and caretaker functions are abrogated. Using HTLV-I as a model, we review in a non-exhaustive fashion current thoughts on how Tax perturbs normal cellular regulation and engenders cellular transformation.

The Molecular Biology of HTLV-I
HTLV-I belongs to the Deltaretrovirus genera of the Orthoretrovirinae family. In vivo, the virus has a tropism for CD4ϩ T-cells (11) although CD8ϩ T-cells may also serve as a reservoir (12). HTLV-I infection is primarily transmitted via cell-cell contact (13,14). Recently, the human Glut1 glucose transporter has been identified as a receptor for infection by cell-free virus (15). The proviral genome of HTLV-I is roughly 9 kbp, and like other retroviruses, contains two LTRs flanking structural genes encoding Gag, Pol, and Env (Fig. 1). An additional region located between env and the 3Ј-LTR, known as the pX region, encodes accessory proteins. The pX region has four partially overlapping reading frames (ORF, Fig. 1), of which ORF IV encodes Tax.
Tax is predominantly a nuclear phosphoprotein (16), which can shuttle into the cytoplasm using a nuclear export signal (17). The mechanism of this shuttling is unclear; however, recent findings that Tax binds tristetrapolin (18) and that tristetrapolin associates with nucleoporin Nup214 (19) raise the possibility that tristetrapolin may serve as a possible nucleocytoplasmic transporter for Tax. Nevertheless, the primary nuclear activity of Tax is to modulate transcription from the HTLV-I LTR (20 -22) and cellular promoters including those for IL-2, IL-13, IL-15, IL-2R, c-Fos, and granulocyte macrophage colony-stimulating factor (23-30) among others. Indeed the breadth of Tax's transcriptional reprogramming of host cell genes was verified by DNA array studies which showed that of 2000 assayed genes the expression profiles of ϳ300 were significantly altered (31). Tax influences so many promoters through its capacity to act in four discrete signaling pathways: CREB/ATF (reviewed in Ref. 32); NF-B (reviewed in Ref. 33); AP-1 (34); and SRF (35). These Tax signaling cascades are discussed in greater detail elsewhere (36).

Tax and Cell Cycle Progression
In the course of transforming cells, viral oncoproteins such as E1A, HPV E7, and SV40 T Ag profoundly dysregulate cell cycle controls (37)(38)(39). Transition from one phase of the cell cycle to the next is normally governed by cyclin-dependent kinases (CDKs) partnered with cyclins. These CDK-cyclin complexes are in turn modulated by phosphorylation mediated through CDK-activating kinases and phosphatases, and through physical sequestration by CDK inhibitory proteins (reviewed in Refs. 40 and 41).
An important cell cycle control resides at the transition from G 1 to S, which is substantially governed by the retinoblastoma tumor suppressor (Rb) (42,43). At this juncture, D-and E-cyclins with partner CDKs (reviewed in Refs. 40, 41, and 44) converge to phosphorylate Rb. Hypophosphorylated Rb sequesters and inactivates E2F factors, which are needed for the expression of genes (such as dihydrofolate reductase, DNA polymerase ␣, and cyclins) that are critical for S phase events (reviewed in Ref. 45). Hyperphosphorylated Rb releases E2F, activates E2F-responsive genes, and secures the passage of cells from G 1 into S (45)(46)(47)(48). Thus, regulation of Rb phosphorylation by cyclin-Cdk and CDK inhibitory proteins such as p16 INK4a , p21 CIP1/WAF1 , and p27 Kip1 is a critical mechanism for influencing gatekeeper function (37).
The ability of Tax to shorten the length of G 1 and to accelerate cells into S (70) embodies a constitutive (i.e. DNA damage-independent) and a DNA damage-induced component. Thus, direct Tax binding of Cdk4 and its enhancement of CycD-Cdk4 activity (55) occur constitutively and are independent of any DNA damage-triggered events. At the same time, Tax can also subvert DNA damage-induced G 1 arrest enforced through p53 (71-74) (see more below). Currently, how Tax affects other phases of the cell cycle is less clear. Emerging findings suggest that this viral oncoprotein can also impair the DNA damage-induced checkpoint in G 2 /M transition (75,76).

Tax and Structurally Damaged Chromosomes
Cancer is a genetic disease. It is estimated that cancer cells can contain more than 100,000 discrete mutations (77). All cancers can be broadly divided into two groups (reviewed in Ref. 78): those arising from loss of DNA repair function (and therefore have structurally damaged chromosomes) and those with chromosomal instability (and therefore have polypoidy and/or aneuploidy). Clastogenic DNA damage is frequently found in HTLV-I-transformed cells (79) and cells transfected to express Tax (80) (Fig. 2A). Clastogenic changes (point mutations, deletions, substitutions, translocations) arise and persist when defects in DNA repair mechanisms co-exist in a cell with a loss in checkpoint functions that would normally eliminate damaged DNA.
All cells acquire DNA damage at a low frequency as they transit the cell cycle. Several mechanisms, including base excision repair (BER), nucleotide excision repair (NER), recombination, and direct repair of nicks by DNA ligation act to correct genetic mistakes. In 1990, the first clue that HTLV-I subverts cellular DNA repair came from the finding that Tax repressed the expression of DNA polymerase ␤, an enzyme involved in BER (81). Subsequently, reduced BER activity was confirmed in HTLV-I, HTLV-II, and bovine leukemia virus-transformed cells (82). Next, Tax was found to suppress the NER normally observed following UV irradiation of cells (83). NER requires DNA polymerases ␦ and ⑀ and uses proliferating cell nuclear antigen (PCNA) as a cofactor. Excessive PCNA can prompt DNA polymerase ␦ to synthesize inappropriately new DNA past template lesions, resulting in nucleotide misincorporation (84). Tax is believed to inhibit NER through its transcriptional up-regulation of PCNA (85); this inhibition of NER also depends, in part, on Tax's inactivation of p53 function (71)(72)(73)(74).
There is no evidence that Tax interferes with DNA ligation (86) or DNA recombination. However, recent data suggest that Tax represses the expression of human telomerase (hTert) (87). Repression of telomerase is significant because the telomeric repeats of chromosomes normally prevent aberrant end-to-end fusions (Fig. 2B) and protect the ends from degradation by exonucleases. Furthermore, de novo double-stranded breaks in chromosomes can also be stabilized by the transient addition of telomeric repeats (88 -90). Indeed, we have documented that Tax prevents such addition of telomeric repeats to new double-stranded breaks (91) and in this way potentially interferes with a protective mechanism used to prevent inappropriate breakages-fusions (Fig. 2B). The combined effects of Tax on BER, NER, DNA end stability, telomerase, and cell cycle progression create a setting in which repair of mistakes is compromised. These combined dysregulations might explain the observed 2.8-fold increase in genomic mutation frequency (92) in HTLV-I-infected cells.

Tax and Aneuploidy
The majority of cancers are aneuploid (93). In transformed cells, numerical chromosomal changes that include losses or gains of entire chromosomes (aneuploidy) generally co-exist with structural chromosomal damage. 2 Although controversial, increasingly aneuploidy is thought to be a cause, rather than a consequence, of transformation (95).
During normal mitosis, human diploid cells maintain euploidy by precisely partitioning 23 pairs of chromosomes from a mother cell to two daughter cells. convoluted, earning them the name of "flower" cells. This suggests that a cellular mechanism that guards against chromosomal missegregation in mitosis is also subverted by HTLV-I. The mitotic spindle assembly checkpoint (MSC) (96) is a key guardian of euploidy. Interestingly, when several ATL cell lines were tested ex vivo, all were found to be deficient in MSC function (97). A potential explanation for this loss arises from two findings: (a) Tax binds human Mad1 (98,99) and (b) Mad1 is an integral constituent of the MSC (96). That impairment of Mad1 function by Tax may contribute to ATL pathogenesis finds intriguing support in the clinical courses of non-HTLV-I acute myeloid leukemia (AML). In two large AML series (1213 and 1612 patients, respectively), loss of a single chromosome 7 (note that the gene for human Mad1 maps to chromosome 7 (100)) prognosticated an extremely poor outcome (101,102). In these two studies, whereas all AML patients had 5-year overall survival rates of 24 -44%, counterpart AML patients with monosomy 7 had survival rates of 0 -10%, respectively (101, 102). Other explanations not excluded, a tantalizing parallel between the two leukemias is that one (ATL) impairs Mad1 function through viral oncoprotein subversion whereas the other (AML) does so through physical loss of chromosome 7 (i.e. monosomy 7).
Is loss of MSC the sole reason for aneuploidy in ATL cells? The answer appears to be "no." Conceptually, one recognizes that loss of checkpoint can explain the tolerance of mistakes by cells, but checkpoint loss cannot create de novo mistakes. Recent studies suggest that Tax might directly trigger chromosomal separation errors in two ways. First, Tax can promote the unscheduled degradation of securin and cyclin B1 most likely through the premature activation of the CDC20-associated anaphase promoting complex (103), thereby leading to faulty mitosis. Second, like the human papilloma virus E7 oncoprotein (38), Tax can also induce aberrant centrosomal multiplication in G 1 . 3 Generating supernumerary centrosomes results in multipolar mitosis, which is another mechanism for creating aneuploidy (104).
Finally, there is a school of thought that suggests polyploidy as the precursor of aneuploidy (104). Relevant to this notion, we note that Tax expression does facilely create multinucleated (i.e. polyploid) cells (76,98). Add to this the fact that Tax can inactivate p53 and Rb (65,(71)(72)(73)(74), two factors essential to a G 1 tetraploid/ polypoid checkpoint (105), and one then can further envision how this might be yet another route traveled by HTLV-I/Tax/ATL cells toward aneuploidy (Fig. 3).

Proliferation versus Apoptosis
A long standing cancer paradox is that overexpression of oncogenes does not simply provide proliferative advantages to cells but frequently also triggers cells to undergo apoptosis. Findings from oncogenic transcription factors such as Myc, E1A, and E2F-1 show this duality to be the rule rather than the exception (reviewed in Ref. 106). Indeed, it is now apparent that oncogenic insults induce countervailing responses by the cell, which are reflected in cell cycle arrest and apoptosis. We reviewed, above, how Tax defeats cellular mechanisms for braking cell cycle progression. No cell cycle and/or genetic instability manifestations of Tax can confer selective growth advantage if cells fail to tolerate such phenotypic and genotypic changes and choose instead apoptotic death. Hence, disabling the cellular apoptotic response remains a requisite for transformation.
By definition, the clinical presentation of ATL implies that in a subpopulation of CD4ϩ T-cells, HTLV-I infection tips the balance between proliferation and apoptosis toward the former. Nevertheless, how HTLV-I Tax oncoprotein influences this choice is not fully understood. Many have examined the contribution of Tax to stressinduced apoptosis. Overall findings have been controversial and divergent. Some found that Tax protects cells from stress-induced cell cycle arrest or apoptosis (107)(108)(109), whereas others observed that Tax sensitizes cells to stress-induced apoptosis (110 -113). Likely, the decision between proliferation and death is influenced by the cellular environment, cell type genetic background, and multiple co-existing signaling events. Depending on context, which set of genes that Tax transcriptionally activates (31) and/or which cluster of gene products that Tax binds (94) will mean either the normal cellular response against oncogenic stress will either prevail (i.e. apoptosis) or be sub-verted (i.e. proliferation) by HTLV-I. A clear understanding of factors in addition to Tax that guide this choice for HTLV-I-infected T-cells will be a major topic for future research.

Concluding Comments
Over 20 million individuals globally are infected with HTLV-I. It is estimated that 2-5% of these carriers will develop ATL over their lifetime. The identification and isolation of HTLV-I 25 years ago have spurred intensive mechanistic investigations into ATL transformation. Using Tax as a model system, we have learned that viral means for transformation parallel similar mechanistic changes seen in spontaneously occurring cancers. A simplified sequence of events appears to be genetic damage initiated by oncogenic stimuli, followed by subversion of cellular checks allowing tolerance and fixation of changes into the genome, and finally selection over time for the correct mix of gene alterations that confer selective growth advantage. Clearly the process is complex and multifaceted. Fleshing out all the biological and molecular details to accompany this simplified framework will easily keep HTLV researchers busy for another 25 years.