HTLV-2 Tax Immortalizes Human CD4⁺ Memory T Lymphocytes by Oncogenic Activation and Dysregulation of Autophagy

Background: Human T cell leukemia virus type 2 (HTLV-2) encodes a viral transactivator Tax.

Results: HTLV-2 Tax efficiently immortalizes human memory T lymphocytes with distinct T cell subtypes.

Conclusion: HTLV-2 Tax connects IκB kinase complex to autophagy pathways for promoting T cell survival and proliferation.

Significance: Learning the oncogenic potential of HTLV-2 Tax is critical for understanding HTLV-2 pathogenesis.

Abstract

Human T cell leukemia virus type 1 and type 2 (HTLV-1 and -2) are two closely related retroviruses with the former causing adult T cell leukemia. HTLV-2 infection is prevalent among intravenous drug users, and the viral genome encodes the viral transactivator Tax, which is highly homologous to the transforming protein Tax from HTLV-1. However, the link between HTLV-2 infection and leukemia has not been established. In the present study, we evaluated the activity of HTLV-2 Tax in promoting aberrant proliferation of human CD4 T lymphocytes. Tax2 efficiently immortalized CD4⁺ memory T lymphocytes with a CD3/TCRαβ/CD4/CD25/CD45RO/CD69 immunophenotype, promoted constitutive activation of PI3K/Akt, IκB kinase/NF-κB, mitogen-activated protein kinase, and STAT3, and it also increased the level of Mcl-1. Disruption of these oncogenic pathways led to growth retardation and apoptotic cell death of the Tax2-established T cell lines. We further found that Tax2 induced autophagy by interacting with the autophagy molecule complex containing Beclin1 and PI3K class III to form the LC3⁺ autophagosome. Tax2-mediated autophagy promoted survival and proliferation of the immortalized T cells. The present study demonstrated the oncogenic properties of Tax2 in human T cells and also implicated Tax2 in serving as a molecular tool to generate distinct T cell subtype lines.

Introduction

Adult T cell leukemia/lymphoma (ATL)² is caused by infection of CD4⁺ T lymphocytes with human T cell leukemia virus type 1 (HTLV-1) (1–3). This type of leukemia exhibits an aggressive clinical course without a cure. HTLV-1 infection is also etiologically linked to myelopathy/tropical spastic paraparesis (HAM/TSP) (4). Roughly 5% of HTLV-1 infected patients developed ATL with a long incubation period typically for >20 years, suggesting that genetic factors also play a role in the pathogenesis of ATL. In contrast, a closely related retrovirus, termed HTLV-2, is not currently linked to leukemia induction, although the viral genome was detected in some cases of hairy cell leukemia (5). Thus, HTLV-2 may lack certain abilities to transform virally infected T cells into malignant ones.

HTLV-1 and -2 share strikingly similar structures of the viral genomes, and both viruses encode highly homologous transforming proteins, Tax (6, 7). One of the oncogenic properties of Tax1 (HTLV-1 Tax) is the induction of persistent activation of NF-κB in host cells by stimulating both canonical and noncanonical pathways of NF-κB signaling (8–17). Tax2 (HTLV-2 Tax) only activates canonical NF-κB signaling (6, 7). The role of Tax2 in deregulating other oncogenic signaling pathways is poorly studied. Furthermore, the Tax1 transgenic mice developed ATL-like leukemia (18); however, the oncogenic activity of Tax2 in mouse model has not been reported.

Several studies suggested that Tax2 is less oncogenic than Tax1. Tax2 transformed rat fibroblast cells less efficiently than Tax1 (19). In addition, the transforming activity of Tax2 on CTLL-2 cell line, an IL-2-dependent murine T cell line, was weaker than Tax1 (20). Dissecting the protein structures of Tax indicated that although there is 75% amino acid sequence homology shared by both viral proteins, Tax2 lacks the leucine zipper-like region and the C-terminal PDZ binding motif on Tax1 (6, 7). When Tax2 was fused with the PDZ binding motif from Tax1, such chimeric protein induced NF-κB processing and resulted in an equal transforming efficiency to Tax1 in CTLL-2 cells (20). Intriguingly, both viruses immortalized primary human T cells at a comparable efficiency (6, 7). It is conceivable that the oncogenic potential of the Tax proteins is best determined in human primary T cells, the natural host cells for HTLV-1 or -2 infection, rather than in rodent cells that are highly susceptible to viral or cellular oncogene-mediated transformation. Indeed, devoid of other viral components, Tax1 appeared to be a weak oncoprotein as Tax1-mediated immortalization of primary human T cells was a rare incidence (21). The ability of Tax2 in immortalizing human T cells has not been previously reported. In the present study, we evaluated the activity of Tax2 in immortalizing human primary CD4 T cells and established four Tax2-immortalized T cell lines with unique properties of activated memory helper T cells. In addition to the oncogenic activations known in HTLV-1-transformed T cells, our study identified autophagy dysregulation by Tax2 as a novel survival mechanism in Tax2-immortalized T cells.

EXPERIMENTAL PROCEDURES

Lentivirus Vector, Viral Production, and Transduction of Primary CD4 T Cells

The tax gene from HTLV-2 was fused with enhanced green fluorescence protein (GFP), and the tax2-gfp fusion fragment was cloned into the lentivirus vector pLCEF8 (22), in which the human elongation factor 1 α promoter drives expression of Tax2-GFP. The procedure for lentiviral production and concentration was described previously (22). Human peripheral blood lymphocytes were isolated from healthy blood donors, and stimulated with PHA (1 μg/ml) for 24 h, followed by adding recombinant IL-2 (100 units/ml) (AIDS Reagent Program). The activated lymphocytes were cultured for 5–7 days, and the CD4⁺ cells were enriched with anti-CD4 magnetic beads (Invitrogen). These purified CD4 T cells were then transduced with the lentivirus carrying the tax2-gfp expression cassette. The transduced cells were cultured continuously in complete media containing 20% fetal bovine serum and 100 units/ml of recombinant IL-2. Lentivirus vector-based shRNAs specific for human Beclin1 were obtained from Open Biosystems, and IKK-specific shRNAs were described previously (22).

Cell Lines, Antibodies, and Chemicals

MT-2 and MoT cell lines were obtained from the AIDS Reagent Program (23), and the HT1080 line was from ATCC. Antibodies for pERK1/2, ERK1, pMEK1, MEK1, pAkt1, Akt1, and GST were purchased from Santa Cruz Biotechnology, and anti-Bcl-2, Bcl-x_L, Mcl-1, and pSTAT3 were from Cell Signaling. U0126, wortmanin, LY294002, BAY11-7082, 3-methyladenin, chloroquine, and bortezomib were purchased from Sigma.

Immunophenotype Analysis, Cell Proliferation Assay, and Human Telomerase Reverse Transcriptase Activity Assay

The immunophenotype of the Tax2-immortalized T cell line was determined with FACS. Cells were stained with allophycocyanin-conjugated antibodies, including anti-CD3, -CD4, -CD25, -TCRαβ, -CD45RO, and -CD69 (eBioscience), according to the manufacturer's instructions. The stained cells were subjected to FACS analysis. For IFNγ intracellular staining, TX2-1 and TX2-4 cells were incubated in phosphate-buffered saline containing 10 μg/ml brefeldin A (Sigma) for 4 h and were then stained with allophycocyanin-conjugated anti-IFNγ antibody after fixation and permeabilization using the intracellular staining kit from eBioscience, followed by FACS analysis. Cell proliferation assay was performed using tetrazolium compound-based CellTiter 96® AQueous One Solution cell proliferation (MTS) assay (Promega). Telomerase reverse transcriptase activity was measured using the TRAPEZE telomerase detection kit (Millipore).

Electrophoretic Mobility Gel Shift Assay (EMSA)

Nuclear extracts were prepared from various T cell lines using NE-PER nuclear and cytoplasmic extraction reagents (Pierce). The oligonucleotide was 5′-end-labeled with biotin (Integrated DNA Technologies) and annealed to its complementary strand. The binding activities were examined by EMSA using a light shift chemiluminescent EMSA kit (Pierce) following the protocol reported previously (22). The oligonucleotide probes are for STAT5 (5′-AGATTTCTAGGAATTCAATCC-3′), Oct-1 (5′-TGTCGAATGCAAATCACTAGAA-3′), STAT3 (5′-GATCCTTCTGGGAATTCCTAGATC-3′), and NF-κB (5′-GATCCGGCAGGGGAATCTCCCTCTC-3′).

Real-time Quantitative PCR

Total RNA was isolated using the RNeasy kit (Qiagen), and its concentration was determined using the NanoDrop1000 spectrophotometer (Thermo Scientific). Quality and integrity of total RNA was assessed on 1% formaldehyde-agarose gels. cDNA was synthesized using the Omniscript reverse transcriptase kit (Qiagen) following the manufacturer's recommended protocol. Template samples in triplicate were subjected to real-time quantitative PCR (Stratagene Mx3005P system) using Power SYBR Green (Applied Biosystems). The primer sequences are available upon request.

Plasmids, Immunoblot, Co-immunoprecipitation, and GST Pulldown

The plasmids for FLAG-BECN1, FLAG-UVRAG, FLAG-PI3KC3, p40phox-GFP, and GFP-LC3B were described previously (24). The co-immunoprecipitation and GST pulldown assays were performed using previously reported protocols (22).

Fluorescence Imaging and Autophagy Assay

To construct fluorescence protein-tagged proteins, mWasabi, encoding a monomeric green fluorescent protein, or mKate2, encoding a monomeric far red fluorescent protein, was amplified from pTEC15 or pTEC20 (kindly provided by Lalita Ramakrishnan, Addgene plasmid 30174 or 30179, respectively) and was fused into the N terminus of the BECN1 PCR fragment to generate mWasabi-BECN1 or mKate2-BECN1 respectively, which was cloned in the mammalian expression vector pEF2. The mKate2 PCR fragment was fused to the C terminus of Tax2 to generate Tax2-mKate2. Transient co-transfection was performed in HT1080 cells using FuGene HD transfection reagent (Roche Applied Science), and 48 h post-transfection, the cells were fixed in 4% formaldehyde-PBS and mounted with DAPI. Fluorescent images were taken using an OLYMPUS IX81 deconvolution microscope and analyzed using SlideBook software (version 5.0, Intelligent Imaging Innovations). For immunofluorescence staining, cells were fixed in 4% paraformaldehyde-PBS, blocked in 3% horse serum-PBS, stained with the indicated primary antibodies overnight at 4 °C, followed by incubation with fluorescent conjugated secondary antibodies, and then mounted with DAPI (Invitrogen).

RESULTS

Tax2 Immortalizes Human CD4⁺ Memory T Cells

Activated human primary lymphocytes grow in vitro typically for up to 4 weeks before reaching cell senescence. Extending growth beyond cell senescence is one of the essential steps toward immortalization and ultimately, oncogenic transformation if additional genetic alteration events occur. To investigate the ability of HTLV-2 Tax in immortalizing mature human CD4⁺ T cells, we isolated CD4⁺ T cell species from healthy blood donors, which were enriched using anti-CD4 antibody-coated magnetic beads, followed by lentivirus transduction of the tax2-gfp fusion gene. Roughly 40–60% of primary CD4⁺ T cells were transduced as evidenced by visualization of green fluorescence. Three weeks following transduction, nearly 100% of the transduced cells emitted a green fluorescence signal and exhibited clumpy growth patterns. The growth of the Tax2-GFP-expressing cells was strictly dependent on exogenous IL-2.

The tax2-gfp-transduced cells had a long term growth potential. In fact, continuous growth of these cells in vitro for more than 18 months was achieved without losing growth potential, indicating that these cells were immortalized. In contrast, GFP-expressing primary CD4 T cells ceased growth in less than 4 weeks following transduction. Analysis of the cell surface markers of four Tax2-GFP-expressing CD4⁺ T cell lines demonstrated a CD3⁺/CD4⁺/CD25⁺/TCRαβ⁺/CD69⁺/CD45RO⁺ immunophenotype (Fig. 1), indicating that these cells were activated memory T lymphocytes. Unlike HTLV-1-transformed T cells in which TCRαβ and CD3 were down-regulated (25), Tax2-immortalized T cells expressed normal levels of these cell surface molecules.

FIGURE 1.

Immunophenotype of Tax2-immortalized T cells. Expressions of cell surface molecules in four Tax2-established T cell lines, including TX2-1, TX2-2, TX2-3, and TX2-4, were analyzed with FACS using allophycocyanin-conjugated antibodies indicated in the figures.

Tax2-immortalized T Cell Lines Represent Unique Subsets of Helper T Cells

In determining expression patterns of selected genes in Tax2-immortalized T lymphocytes, we performed real-time quantitative PCR analysis. Compared with the normal CD4 T cell populations that were cultured at the identical conditions, we found that Tax2-immortalized T cells had differential gene expression profiles. TX2–3 cells expressed significantly high levels of IL-4, TGFβ1, and Fas, and TX2-4 cells showed high levels of IL-5, IL-15, IFNγ, and Fas (Fig. 2, A and B). TX2-2 cells produced an extremely high level of IL-10 (Fig. 2B). IL-13 and IL-2 were undetected or detected at very low levels in all Tax2-immortalized T cell lines (Fig. 2A), unlike HTLV-1-transformed T cells that expressed IL-13 (26, 27). Consistent with the results from real-time PCR, the intracellular FACS analysis showed that TX2-4 cells expressed a significantly higher level of IFNγ than TX2-1 cells, though a majority of cell populations from these two cell lines produced IFNγ (Fig. 2C). Because these T cell lines expressed comparable levels of Tax2-GFP (see Fig. 6C), the increased levels of cytokines in some T cell lines were unlikely caused by Tax2. Instead, these gene expression profiles implicated Tax2 in selectively immortalizing unique subsets of helper T cells. All Tax2-immortalized T cell lines expressed higher levels of Fas than normal CD4 T cells (Fig. 2B), similarly to HTLV-1-transformed T cells (28). These Tax2-immortalized T cells also produced PDGFRA but did not synthesize a detectable amount of its ligand, PDGF-BB (Fig. 2B), suggesting that the autocrine loop of the PDGFRA signaling is not the driving force for promoting proliferation of Tax2-expressing T cells.

FIGURE 2.

Selected gene expression profiles of Tax2-immortalized T cells. A and B, expression of two sets of selected genes was analyzed using real-time quantitative PCR. Relative expression level of a given gene to GAPDH was plotted. C, expression of IFNγ in TX2-1 and TX2-4 cells was analyzed by intracellular staining with anti-IFNγ-allophycocyanin and FACS. MMP, matrix metalloproteinase.

FIGURE 6.

Tax2 interacts with the autophagy molecule complex. A, GST pulldown assay to analyze the interaction of Tax2 with the autophagy molecules BECN1, PI3KC3, and UVRAG in transiently co-transfected 293 cells. B, reciprocal GST pulldown assay to detect Tax2-BECN1 interaction. C, co-immunoprecipitation of Tax2-GFP and BECN1 in four Tax2-immortalized T cell lines. D, subcellular localization of mKate2-BECN1, FLAG-PI3KC3, and GFP-IKKγ in HT1080 cells. E, dual fluorescence imaging to detect co-localization of Tax2 with BECN1, PI3KC3, or IKKγ in transfected HT1080 cells.

Tax2 Deregulates Oncogenic Signaling Molecules

We next examined oncogenic activation in Tax2-immortalized T cells. The expression patterns of molecules in T cell receptor signaling were first evaluated. As shown in Fig. 3A, Tax2-expressing T cell lines expressed proximal signaling molecules of TCR, including Lck, ZAP70, and LAT. In contrast, Lck and ZAP70 were undetected in HTLV-1-transformed MT-2 T cells. Analysis of the common oncogenic pathways showed that similar to MT-2 cells, the Tax2-immortalized T cell lines, TX2-1 and TX2-2, exhibited hyperphosphorylation of MEK1, ERK1/2, and Akt1 (Fig. 3B), implying constitutive activities of these kinases. The activities of the transcriptional factors, including STAT3 and NF-κB, were detected in both Tax2-expressing T cells and MT-2 cells (Fig. 3C), and the activity of STAT5 was only detected in Tax2-established, IL-2-dependent T cells (Fig. 3C). We next investigated additional prosurvival molecules, the Bcl-2 family proteins. A comparable level of Bcl-x_L was expressed in normal CD4 T cells, Tax2- T cells, and MT-2 cells (Fig. 3D). The level of Mcl-1 was increased in all Tax-expressing T cell lines (Fig. 3D). Interestingly, unlike MT-2 cells that produced a significantly high level of Bcl-2, TX2-1 and TX2-2 cells expressed almost undetectable level of Bcl-2 (Fig. 3D). The expression of Bax, a proapoptotic Bcl-2 family protein, was not altered in Tax-expressing T cells as compared with normal CD4 T cells (Fig. 3D). Expectedly, the tumor suppressor protein p53 was stabilized in Tax-expressing cells (Fig. 3D). Examination of TX2-3 and TX2-4 cells demonstrated constitutive activities of NF-κB, STAT3, and STAT5 (Fig. 3E), similar to that seen in TX2-1 and TX2-2 cells (Fig. 3C). The oncogenic activation in Tax2-immortalized T cell lines was apparently similar to that in HTLV-2-infected MoT cells that exhibited constitutive activities of STAT3 and NF-κB (Fig. 3F). Furthermore, the activity of telomerase reverse transcriptase was well maintained in all four Tax2-immortalized T cell lines after prolonged culture (Fig. 3G).

FIGURE 3.

Activation of oncogenic signaling in Tax2-immortalized T cells. A, total protein lysates from activated normal CD4 T cells from two healthy donors, Tax2-immortalized T cell lines TX2-1 and TX2-2 and the HTLV-1-transformed T cell line MT-2 were examined with immunoblot using antibodies reacting to PLCγ1, ZAP70, Lck, LAT, c-Cbl, and IKKα. B, detection of the phosphorylation status of ERK1/2, MEK1, and Akt1 in normal CD4 cells, TX2-1, and TX2-2 cells that were cultured in the presence of IL-2. MT-2 cells were used for control. C, activities of the transcriptional factors in Tax2-immortalized T cell lines TX2-1 and TX2-2 with EMSA. Jurkat and MT-2 cells were used for control. D, expression patterns of Bcl-2 family proteins and p53 in Tax2-immortalized T cells using immunoblot analysis. E, the activities of the transcriptional factors in Tax2-immortalized T cell lines TX2-3 and TX2-4 with EMSA. F, the activities of STAT3 and NF-κB in the HTLV-2-infected MoT cell line. Jurkat cell line serves as a negative control and MT-2 cell line for positive control. G, the telomerase reverse transcriptase assay in Tax2-immortalized T cell lines. AcPBLs were activated peripheral blood lymphocytes, and PosCtl was the positive control from the assay kit. Heat inactivation was applied to each cell line and was used as a control.

To determine the role of the oncogenic pathways in Tax2-mediated proliferation of primary T cells, we applied various chemical inhibitors to Tax2-immortalized T cells. As shown in Fig. 4, A and B, U0126, the inhibitor of MEK1, and LY294002, the inhibitor of PI3K, impaired phosphorylation of ERK1/2 and Akt1, respectively. Similarly, wortmanin, an inhibitor of PI3K, was shown to effectively inhibit Akt1 phosphorylation in TX2-1 cells (Fig. 4C). Bortezomib, an FDA-approved proteasome inhibitor, inhibited NF-κB activity in a dose-dependent manner, induced caspase activation as seen with cleaved forms of caspase-3, -7, and PARP and reduced cell viability of the Tax2-immortalized T cells (Fig. 4, D–F). The above chemical inhibitors and BAY11-7082, an inhibitor of IκB kinase, negatively affected viability of Tax2-immortalized T cells (Fig. 4G). Together, these results indicate that activation of multiple oncogenic signaling pathways is crucial for Tax2-mediated survival and proliferation of human primary memory T cells.

FIGURE 4.

Inhibition of oncogenic signaling results in growth arrest and apoptotic death of Tax2-immortalized T cells. The phosphorylation status of ERK1/2 and Akt1 in TX2-1 cells treated with U0126 (A), LY294002 (B), and wortmanin (C) at doses of 0.625, 1.25, 2.5, 5, and 10 μm for 2 h. D, TX2-1 cells were treated with dimethyl sulfoxide (DMSO) or bortezomib at 1.25, 2.5, 5, 10, or 20 nm. 16 h following the treatment, the nuclear extracts from these cells were prepared and were used for EMSA using an NF-κB probe. Oct-1 was used for nuclear extracts loading control. E, TX2-1 cells were treated similarly to D with dimethyl sulfoxide or bortezomib at 1.25, 2.5, 5, 10, 20, 40, or 80 nm for 16 h. The total protein lysates were examined with immunoblot for detecting cleaved forms of caspase (casp)-3, -7, and PARP. β-Actin was used for protein loading control. F, TX2-1 cells were treated with bortezomib at indicated doses for 48 h, and the cell viability was examined by MTT assay. G, TX2-1 cells were treated with dimethyl sulfoxide, U0126 (10 μm), LY294002 (10 μm), wortmanin (10 μm), or BAY11-7802 (10 μm) for 24, 48, or 72 h. Cell viability was determined with MTT assay.

Tax2 Deregulates Autophagy to Promote Survival of the Immortalized T Cells

Autophagy is a prosurvival cellular process for cells under various stresses such as nutrient deprivation and has been implicated to play a role in supporting T lymphocyte proliferation (29, 30). To determine whether autophagy plays an important role for the survival and proliferation of Tax2-immortalized T cells, we analyzed the autophagic process in these cells. One of the characteristic features of autophagy is generation of a lipidated form of LC3, LC3-II, which associates with autophagosome. In Tax2-expressing T cells, a significantly high level of LC3-II was detected (Fig. 5A). To determine whether autophagy provides a survival signal to Tax2-immortalized T cells, TX2-1 cells were treated with the autophagy inhibitor chloroquine or 3-methyladenin. Both chemicals inhibited growth of TX2-1 cells and induced cleavage of caspase-3/-7 (Fig. 5, B and C). Collectively, these results indicate that constitutive autophagy is necessary for promoting survival and proliferation of Tax2-immortalized T cells.

FIGURE 5.

Tax2 deregulates autophagy in promoting T cell survival. A, equal amounts of total protein lysates from four Tax2-established T cell lines were analyzed with immunoblot using antibodies for LC3 and β-actin. Non-Tax2-expressing Jurkat T cells were used for control. B, cell viability of TX2-1 cells treated with dimethyl sulfoxide (DMSO), 3-methyladenin (3-MA; 5 mm), or chloroquine (50 μm) at 24-, 48-, and 72-h time points. C, TX2-1 cells were treated with chloroquine at indicated doses for 24 h, and the total protein lysates were examined with immunoblot using antibodies detecting cleaved forms of caspase (casp)-3 and -7. D, GFP-LC3B fluorescence patterns in HT1080 cells transfected with vector, Tax2-HA, or the constitutively active form of IKKβ, FLAG-IKKβKA. E, co-transfection of p40phox-GFP with mKate2 or with Tax2-mKate2 in HT1080 cells. F, depletion of Beclin1 (BECN1) or IKKβ in HT1080 cells using lentivirus transduction of specific shRNAs. Total protein lysates from the modified cell lines were analyzed with anti-BECN1 or anti-IKKβ immunoblot. β-Actin was used as protein loading control. G, expression patterns of LC3-I and LC3-II in HT1080 cells transfected with vector, Tax2-HA, or FLAG-IKKβKA. H, expression patterns of LC3-I and LC3-II in Beclin1 knockdown (KD) HT1080 cells transfected with vector, Tax2-HA, or FLAG-IKKβKA. I, co-transfection of GFP-LC3B with vector or Tax2-HA in IKKβ-knockdown (KD) HT1080 cells. J, TX2-1 cells were transduced with lentiviruses expressing nonspecific shRNA (NS) or IKKα/β-specific shRNAs, and the transduced cells were selected with puromycin for 1 week and analyzed with immunoblots for IKKα/β, LC3-I/II, and actin. K, cell viability assay for TX2-1 cells transduced with nonspecific shRNA and IKKα/β shRNAs. p value < 0.05 for the time point at 72 h between nonspecific and IKKα/β shRNA-transduced cells. L, TX2-1 cells were transduced with lentiviruses expressing nonspecific shRNA (NS) or BECN1-specific shRNA, and the transduced cells were selected with puromycin and analyzed with immunoblots for BECN1, LC3-I/II, and actin.

PI3K/Akt1 signaling is known to inhibit autophagy by activating mammalian target of rapamycin (mTOR) complex, whereas IKK/NF-κB signaling has been linked to starvation- or rapamycin-induced autophagy (31, 32). Tax2 was shown to activate both signaling pathways in immortalized T cells (Fig. 3). To evaluate the role of Tax2 in induction of autophagy, we co-transfected Tax2 with GFP-LC3 to examine formation of LC3⁺ autophagosome foci. In the absence of Tax2, GFP-LC3 was evenly distributed in the transfected cells (Fig. 5D). Co-expression of Tax2 and GFP-LC3 led to formation of GFP-LC3 cytoplasmic foci reminiscent of autophagosomes (Fig. 5D). Similarly, the constitutive active form of IκB kinase β subunit, IKKβKA, promoted formation of the LC3⁺ autophagosomes (Fig. 5D). Furthermore, co-transfection of Tax2-mKate2, a far-red fluorescent protein (mKate2) tagged Tax2, and p40phox-GFP-induced aggregation of the p40phox (Fig. 5E), which indicated that the activity of PI3K class III (PI3KC3), one of the key mediators of autophagy pathway, was stimulated in Tax2-mediated autophagy.

To determine the role of IKK and Beclin1 (BECN1), an essential autophagy molecule, in Tax2-mediated autophagy, we generated BECN1- and IKKβ-deficient HT1080 cell lines by lentivirus transduction of specific shRNAs (Fig. 5F). As shown in Fig. 5G, strong accumulation of LC3-II was readily detected in Tax2- or IKKβKA-transfected cells. However, in BECN1-depleted cells, Tax2 or IKKβKA failed to induce accumulation of LC3-II (Fig. 5H). Similarly, in IKKβ-depleted cells, reduced numbers of LC3⁺ foci were seen upon Tax2 transfection (Fig. 5I). We further verified these findings in Tax2-immortalized T cells. Depletion of IKKα/β in TX2-1 cells resulted in a significant reduction of LC3-II (Fig. 5J) and reduced cell viability (Fig. 5K). In BECN1 knockdown, TX2-1 cells, LC3-II was also reduced compared with the control cells (Fig. 5L) but such reduction was less prominent than that in the IKKα/β-depleted cells (Fig. 5, J and L), suggesting that in addition to BECN1, additional autophagy molecules are possibly involved in Tax2-mediated autophagy. Consistent with this finding, BECN1 knockdown cells showed slight reduction of cell viability than the control cells (data not shown). Together, these results suggest that Tax2 deregulates autophagy by modulating a crosstalk between the IKK complex and the autophagy molecule complex.

Tax2 Connects IKK Complex to Autophagy Pathways

The underlying mechanism of Tax2 induction of autophagy was investigated to determine whether Tax2 physically interacts with key components of autophagy pathways. The autophagy molecule complex, which consists of BECN1, UVRAG, and PI3KC3, is essential for assembling autophagosomes (33). We found that Tax2 interacted with BECN1 and PI3KC3 in co-transfected cells but did not co-precipitate with UVRAG (Fig. 6A). In addition, Tax2 did not interact directly with other autophagy molecules such as LC3 (Fig. 6B). In Tax2-immortalized T cells, the interaction of Tax2 and BECN1 was also detected by co-immunoprecipitation (Fig. 6C). Next, we applied fluorescence imaging technique to visualize Tax2 interaction with autophagy molecules. BECN1 and PI3KC3 were evenly distributed in the cytoplasm, and IKKγ exhibited a cytoplasmic cluster pattern (Fig. 6D). Tax2 recruited BECN1, PI3KC3, and IKKγ into the cytoplasmic punctate foci (Fig. 6E). These results provided structural evidence that Tax2 connects the IKK complex to the autophagy pathways.

DISCUSSION

Our study demonstrates that Tax2 efficiently immortalizes human primary CD4⁺ memory T cells. Tax2 promotes survival and aberrant proliferation of primary T cells via several mechanisms. Similar to HTLV-1-transformed T cells, Tax2 constitutively activates oncogenic signaling pathways, including IKK/NF-κB, PI3K/Akt1, MAPK/ERK1/2, and STAT3, and induces expression of surviving factors such as Mcl-1. In addition, Tax2 deregulates autophagy by connecting the IKK complex to autophagy pathways. Our data show that the oncogenic activation and dysregulation of autophagy contribute significantly to Tax2 immortalization of human T cells.

In contrast to the previous report showing that Tax1 is a weak oncoprotein and immortalizes mature human T cells poorly (21), we found that Tax2 efficiently immortalizes primary T cells. All tax2-gfp transduced T cells were able to grow for more than 3 months, and four of 12 Tax2-GFP-established T cell lines have maintained growth potential for prolonged time. The differential immortalization abilities mediated by Tax1 and Tax2 remain to be solved. One possibility is a progressive loss of Tax1 expression in the late passages of the immortalized T cells. Unlike ATL leukemia cells in which Tax expression is typically lost, the growth of Tax-established primary T cells requires constitutive production of the Tax protein in order to promote T cell survival and proliferation. The activity of a heterologous viral promoter, such as cytomegalovirus promoter (CMV), which was previously utilized to establish Tax1-immortalized primary T cells (21), may be deleterious during prolonged cell growth due to a possible promoter silencing (34). In the Tax1-immortalized T cell line WT4, Tax1 was expressed at a much lower level than that in MT-2 cells in p40tax form (21). MT-2 cells express abundant amount of p68 Env-Tax fusion protein but produce extremely low level of p40Tax (35). The lower level of p40Tax in WT4 cells indicates that the Tax1 expression is barely maintained in most Tax1-established T cell lines. With use of human elongation factor promoter to drive the Tax2 expression in the present study, the expression of Tax2 in the immortalized T cell lines is sustainable for over 18 months in culture. To draw the conclusion that Tax2 performs better than Tax1 in immortalizing human T cells, the same technical approach must be employed. Nevertheless, our study validated the idea that immortalization of mature human T cells can be efficiently achieved by Tax2.

Our study also identifies autophagy as a novel survival mechanism in Tax2-immortalized T cells. Tax2 induces autophagy to promote T cell survival and proliferation by targeting the autophagy molecule complex containing BECN1 and PI3KC3. The role of autophagy in oncogenesis remains controversial. On one hand, autophagy molecules function as tumor suppressors as mice that lack expression of these genes are prone to tumor development (36, 37). On the other hand, vast amounts of evidence have shown that autophagy contributes to chemotherapy resistance duo to its cytoprotective function (38, 39). Although autophagy is a fundamental process for antiviral defense, viruses have developed strategies to subvert or use autophagy for their own benefit. Autophagic process is necessary for productive replication of some oncogenic viruses. For instance, autophagy is implicated in initiation of hepatitis C virus infection (40). Chloroquine, a chemical inhibitor of autophagy, or depletion of autophagy molecules represses hepatitis C virus replication (41, 42). The oncogenic X protein from hepatitis B virus sensitizes cells to starvation-induced autophagy by increasing BECN1 expression (43). Furthermore, latent membrane protein 1, an oncogene product from Epstein-Barr virus, induces early or late stage autophagy depending on its expression levels (44). Inhibition of autophagy in Epstein-Barr virus-infected cells suppresses transforming phenotypes resulting from accumulation of latent membrane protein 1. Not surprisingly, HTLV-2 Tax also deregulates cellular autophagic process for promoting aberrant proliferation of its host cells.

The underlying mechanism of Tax2 induction of autophagy appears to be related to the crosstalk between NF-κB signaling and autophagy pathways. Tax2 induces hyperactivation of IKK and NF-κB, and both factors are implicated in induction of autophagy (31, 32). NF-κB p65RelA induces autophagy by transcriptional activation of the beclin 1 gene (32). However, NF-κB has also been shown to inhibit autophagy in the context of TNFα-induced cell death (45). The activation of IKKα or IKKβ induced by TNFα has been implicated in the phosphorylation of tuberous sclerosis complex 1/2 (TSC1/2), resulting in activation of mTOR and inhibition of autophagy (46). Conversely, autophagic process has a negative impact on the activation of IKK and NF-κB. For instance, autophagy facilitates the degradation of the IKK complex and its upstream activator NF-κB-inducing kinase, contributing to the inhibition of NF-κB signaling (47). Furthermore, autophagic process depletes p62 (sequestosome1), an activator of the IKK complex, resulting in a diminished activity of NF-κB (48). Interestingly, IKK has been implicated to be the key regulator of autophagy induced by starvation or rapamycin stimulation that triggers IKK activation (31). Constitutively active IKKβ or myristoylated IKKγ induces autophagy, whereas depletion of IKKα or IKKβ prevents autophagy induction, suggesting that the IKK complex is central in connecting upstream NF-κB signaling to autophagy pathway. Our data clearly demonstrate that Tax2 recruits the IKK complex, together with the autophagy molecule complex containing BECN1 and PI3KC3, to form LC3⁺ autophagosome foci. This is executed by the interaction of Tax2 with the IKK complex, BECN1 and PI3KC3, resulting in an increased activity of PI3KC3 and assembly of autophagosomes. Inhibition of autophagy leads to apoptotic death of Tax2-immortalized T cells, supporting a critical role of Tax2-mediated autophagy in T cell immortalization. The present study provides insights into a possible therapeutic intervention with autophagy inhibitors in managing HTLV-2 infection.