A human T-cell leukemia virus Tax variant incapable of activating NF-kappaB retains its immortalizing potential for primary T-lymphocytes.

The human T-cell leukemia virus type 1 (HTLV-1) transactivator (Tax) has been shown to interfere with regulated cellular proliferation. Many studies have focused on the ability of Tax to transform rodent fibroblasts; however, none has defined the molecular requirements for Tax transformation of human lymphoid cells. We show here that tax induces permanent growth of human primary T-lymphocytes by using a transformation/immortalization defective rhadinovirus vector. The cells phenotypically resemble HTLV-immortalized lymphocytes and contain episomally persisting recombinant rhadinoviral sequences, which stably express functional Tax protein. As Tax can activate major cellular signal transducing pathways including NF-kappaB and cAMP-responsive element binding protein (CREB), we asked for the relevance of these routes in the immortalization of human lymphocytes. By using Tax mutants that either activate exclusively CREB/activating transcription factor or are defective in activating this signaling pathway, we delineated that Tax can induce immortalization of primary human T-lymphocytes through a mechanism independent of NF-kappaB activation.

The human T-cell leukemia virus type 1 (HTLV-1) transactivator (Tax) has been shown to interfere with regulated cellular proliferation. Many studies have focused on the ability of Tax to transform rodent fibroblasts; however, none has defined the molecular requirements for Tax transformation of human lymphoid cells. We show here that tax induces permanent growth of human primary T-lymphocytes by using a transformation/immortalization defective rhadinovirus vector. The cells phenotypically resemble HTLV-immortalized lymphocytes and contain episomally persisting recombinant rhadinoviral sequences, which stably express functional Tax protein. As Tax can activate major cellular signal transducing pathways including NF-B and cAMP-responsive element binding protein (CREB), we asked for the relevance of these routes in the immortalization of human lymphocytes. By using Tax mutants that either activate exclusively CREB/activating transcription factor or are defective in activating this signaling pathway, we delineated that Tax can induce immortalization of primary human T-lymphocytes through a mechanism independent of NF-B activation.
HTLV-1, 1 a complex human retrovirus (1), is the etiological agent for adult T-cell leukemia (ATL) (2, 3) and a chronic neurodegenerative disease (4,5). The virus has little cytotoxic effect on T-lymphocytes, its target cells, but changes their growth properties. In contrast to normal T-cells, HTLV-1-infected lymphocytes can proliferate permanently without antigen stimulation (for review see Ref. 6). Cells freshly immortalized by HTLV-1 initially require exogenous interleukin (IL)-2 for proliferation (7)(8)(9); however, prolonged cultivation of these cells eventually results in IL-2 independent growth and loss of T-cell receptor and T-cell functions (7,8). The IL-2 independent growth can be caused by a constitutive stimulation of the IL-2 receptor-associated Jak-Stat pathway (10,11).
The HTLV-1-encoded transactivator protein p40 tax has many properties typical for oncoproteins. (i) It immortalizes primary rodent fibroblasts and cooperates with the ras oncogene in the transformation of these cells (12); in addition it transforms established rat fibroblast lines (13,14). (ii) It induces leukemia and neurofibromas in transgenic mice (15,16). (iii) It has a potential for immortalizing primary human T-lymphocytes (17)(18)(19)(20). Tax potently stimulates viral transcription and acts on an array of cellular promoters. Mechanistically, Tax can stimulate transcription either directly through an intrinsic activation domain (21) or through activation of signal transducing pathways such as NF-B. Despite much investigation, it currently still remains unclear whether NF-B activation by Tax is linked with its property to transform human lymphocytes.
To define the Tax-stimulated pathway, which mediates the immortalization of human T-cells, we expressed two types of Tax mutants in primary lymphocytes. One mutant activates CREB/ATF but not NF-B. The second mutant activates neither CREB nor NF-B. Here we report that the transduction of a Tax mutant that is active for CREB/ATF but inactive for NF-B resulted in permanent growth of human primary Tlymphocytes. These results, although not directly addressing the role of NF-B activation in lymphocyte growth, show that the NF-B activating function of Tax is dispensable for the immortalization of primary human T-cells.
Construction of Recombinant Rhadinoviruses and Transduction of Primary Human Lymphocytes-Recombinant viruses were generated as described previously (17,18). The DNAs of the Tax mutants M7 (27) and S258A (28) were assembled into the same expression cassette (pHISL) used previously for the construction of Tax wt -expressing rhadinoviruses (17,18). Upon transfection these open reading frames yielded the same amount of protein as wild type, if they were expressed under the control of the same promoter (data not shown and Ref. 27). This indicated similar stabilities of wild type and mutant Tax proteins. The phenotypes of the Tax mutants were verified in co-transfection experiments using LTR-CAT and NF-B CAT indicator plasmids. The expression cassettes were inserted into a recombination plasmid containing a neo r selection marker and flanking rhadinoviral sequences. These sequences permitted recombination with the rhadinovirus Herpesvirus saimiri A11S4 (29) and directed the Tax expression cassettes into the right junction of unique and repetitive sequences of its genome. The recombinant rhadinoviruses were purified by geneticin selection and subsequently by two plaque isolations. The presence of the mutations was verified in the purified recombinant viral stocks by Southern blot analysis and sequencing of PCR products. For immortalization assays, primary lymphocytes were isolated from cord blood by Ficoll-Hypaque density gradient centrifugation, stimulated by phytohemagglutinin and infected as described (18).
DNA and RNA Analyses-To detect episomally persisting recombinant rhadinovirus DNA, cells were lysed in situ on top of an agarose gel; the DNA was separated by electrophoresis into chromosomal, episomal, and linear fractions (18). The Southern blot was analyzed with a Taxspecific probe. To verify unrearranged tax sequences, total cellular DNA was subjected to PCR analyses using tax-specific primers (sense 5Ј-CAGCCCACTTCCCAGGGTTTGGAC-3Ј or 5Ј-CCAATCACTCATACA-ACCCCCA-3Ј and antisense 5Ј-GTGTGAGAGTAGAAATGAGGGGT-3Ј). These primers resulted in PCR products of 627 and 880 nucleotides. To confirm the presence of the mutation within tax, purified PCR products were sequenced. Cytokine RNAs were detected by standard reverse transcriptase-PCR using primers described in Platzer et al. (30). Tax-, IB-␣-, IL-1␣-, and IL-2R␣-RNA expression was detected by Northern blot analyses as described previously (18). Hybridization signals were quantified by phosphorimaging.

Immortalization of Cord Blood Cells by a Tax Variant Active for CREB/ATF and Deficient for NF-B
Signaling-To investigate whether the Tax-mediated activation of CREB/ATF or NF-B transcription is important for the immortalizing phenotype, we tested two distinctive Tax mutants using rhadinovirus transduction. The capacity of Tax to activate NF-B can be selectively abolished by point mutagenesis, without affecting its activation of the CREB/ATF path (27,31). This selectivity has been described for the Tax S258A mutant, which contains a single amino acid change at position 258 from serine to alanine. The S258A mutation within Tax abolishes NF-B activation. However, it leaves the protein activity on CREB/ATF-depend-ent transcription unaltered. Tax S258A was incapable of stimulating NF-B site-dependent transcription using indicator plasmids (23) (data not shown) but retains its stimulatory effect on the HTLV-1 LTR. The NF-B phenotype of Tax S258A has also been extensively characterized in a previous study; no NF-B activation above basal levels could be detected (31). A second Tax mutant M7 (27) is negative for both NF-B and CREB-dependent transcriptional stimulation. A rhadinovirus vector derived from the A11 strain of the New World primate agent H. saimiri, which has been clearly shown not to transform human lymphocytes (6,32), was used for the transduction of the tax mutants. This vector is a non-transforming apathogenic deletion variant of the A11 strain (17,18,29); it persists in human T-cells without detectable expression of herpesvirus genes (33). This ensures that biological effects elicited from this approach are a consequence of Tax expression.
The mutant Tax cDNA sequences were assembled into the same expression cassette used previously for the construction of Tax wt -expressing rhadinovirus recombinants and introduced into the viral genome by homologous recombination. The transforming potential of the recombinants was assessed in primary human lymphocytes. In five independent experiments using cord blood samples from four different individuals, Tax S258A resulted in immortalized lymphoid cultures (NATI). NATI cells were found to duplicate every 28 -59 h under optimal growth conditions. This does not significantly differ from TAXI-1/TRI-1 cells immortalized with wtTax (t 2 ϭ 26 -71 h). All of these cultures have been propagated for at least 16 months and some for more than 2 years. Although these show IL-2 growth dependence, control cells died within 10 -12 weeks even in the presence of IL-2. We also noted that in five out of five tries cultures infected with the vector or with recombinants carrying the Tax M7 failed to produce any immortalized cells. This finding suggests that Tax S258A but not Tax M7 was competent for the immortalization of primary human lymphocytes.
Cells Transduced with Tax Resemble HTLV-1-immortalized Cells-To verify if immortalization by Tax transduction may be a biologically important model for HTLV-1 transformation/ ATLL, the immortalized cells were characterized in detail. First the cells were checked for an absence of infectious H. saimiri. None of the human CBL-derived cell lines (NATI) produced infectious rhadinovirus, which was concluded from co-cultivation experiments with permissive OMK-637 cells. Futhermore it was confirmed that each human CBL line contained the expected tax mutant as a part of episomally persisting recombinant rhadinoviral DNA and not as linear DNA, which would be characteristic for virus particles. Additionally, PCR with two tax-specific primer pairs detected intact configuration of tax sequences. The preservation of the tax mutations 8 -12 months after transduction was verified by direct sequencing of PCR products. The expression of Tax could be quantitated by detection of specific transcripts in Northern blots (Fig.  1A) and of Tax protein in Western blots (Fig. 1B). The amounts of Tax protein in the cells immortalized by the mutant tax S258A were roughly equal to the amounts observed in cell lines immortalized previously by the transduction of the tax wt gene (not shown).
The human lymphocytes immortalized by tax were further characterized for surface expression of marker proteins (Table  I). Four cell lines expressing the tax S258A gene were phenotypically analyzed in detail. NATI-1-4 cells were compared with cells immortalized with tax wt (TAXI-1, Kati 40, and TRI-1) and with established HTLV-1-immortalized lymphocytes (MT-2, C91-PL, HuT-102, and Mondi). We found that all the cell lines transduced with tax S258A were T-cell lines expressing CD45 and CD2 but not CD19, a B-cell marker (Table I). These cells, like HTLV-immortalized cells, had no or diminished (compared with normal T-cells) CD3 surface expression (data not shown). All cell lines expressed either CD4 and/or CD8. The same subtypes of T-cells have been shown to be immortalized by HTLV-1 (9). Consistent with HTLV-1-transformed cells, all Tax-transduced cell lines had an activated phenotype. Qualitative reverse transcriptase-PCR demonstrated that the cells expressed RNAs coding for the inflammatory cytokines IL-1␣, IL-6, tumor necrosis factor-␣, and leukemia inhibitory factor but not IL-2 RNA (data not shown). In addition the cells expressed a series of activation-specific markers (CD25, CD69, CD95, and B7-1). No major quantitative alteration among the surface markers was detectable with the exception of CD25. The surface expression of this protein, the ␣ chain of the IL-2 receptor (IL-2R␣), was found significantly reduced in cells transduced by the mutant Tax protein compared with cells containing the wild type Tax (Fig. 2A). Taken together surface markers and cytokine expression profiles reveal significant phenotypic similarities between HTLV-1-immortalized cells and lymphocytes transduced by either Tax wt or its NF-B activation-deficient S258A derivative.
Reduced Expression of IL-2R␣ in Cells Immortalized with the Mutant Tax-Tax stimulates the IL-2R␣ promoter via NF-B (34). To determine whether the reduced surface expression of IL-2R␣ could be due to a loss of promoter transactivation by Tax, we investigated the IL-2R␣ transcription in cells without Tax, with Tax wt , and with mutant Tax. The mRNAs were detected by Northern blots and normalized to an internal standard mRNA (GAPDH). Fig. 2B shows the average from three independent experiments. As a result the content of IL-2R␣ transcripts was clearly reduced in cells containing the Tax mutant thus accounting for the diminished surface expression of CD25. These data suggested that the Tax S258A -expressing cells have lower NF-B activity compared with the wild type counterpart.
The remaining IL-2R␣ mRNA expression detected in the cells could be due to a recently identified upstream element of the human IL-2R␣ promoter (IL-2rE), which is not activated by NF-B (35). This element strongly responds to IL-2, which is a component of the growth medium used. To determine whether the IL-2rE contributes to the activity of the IL-2R␣ promoter CAT indicator plasmids were transfected. An IL-2rE containing promoter segment, which is not activated by NF-B, was found at least as active as the complete promoter in the Tax S258A -expressing NATI cells; in contrast the complete promoter was clearly more active in cells expressing Tax wt . This suggests that the IL2-R␣ promoter in Tax S258A -immortalized cells is stimulated by transcription factors different from NF-B and that in these cells the IL-2rE element accounts for most of the activity of the promoter. Taken together, these data suggested that lower NF-B activity in the Tax S258A -expressing cells caused reduced expression of IL2-R␣/CD25. Tax-immortalized CBLs, the state of NF-B activation was analyzed. This was measured using the expression level of RNAs initiated from NF-B site-dependent IL-1␣ promoter (36). The mRNA content of cells without Tax, with Tax wt , and with mutant Tax was quantitated by Northern blots, which were normalized to an internal standard mRNA (GAPDH). Fig.  3 shows the average from three independent experiments. A clear difference was seen between those cells expressing wild type and mutant Tax. IL-1␣ transcripts were strongly reduced in the lymphocytes transduced with mutant Tax S258A . These data suggest that the Tax S258A -expressing cells have lower NF-B activity compared with its wild type counterpart. This finding likely reflects an intrinsic property of Tax S258A since four out of four independently immortalized cell lines show the same phenotype.

Reduced NF-B Activity in T-cells Immortalized by Tax S258A -To check whether autocrine activation of NF-B exists in
As a further parameter of NF-B activity, the amount of IB-␣ mRNA was determined. This is a more direct measure of NF-B activity, since IB-␣ is directly involved in the feedback control of NF-B (37,38) and because the IB-␣ promoter is strongly induced by active NF-B. Fig. 4A shows that IB-␣ mRNA is highly expressed in the cells expressing Tax wt . By contrast, low levels of IB-␣ RNA are present in cell lines expressing the Tax S258A and in Tax-negative T-lymphocytic cell lines, consistent with low endogenous NF-B activity. We also determined the half-life time of the IB-␣ protein in cells expressing Tax S258A (NATI-2) and Tax wt . Cycloheximide was Immortalizing Potential of HTLV-1 Tax Variants added to block protein synthesis, and the amounts of IB-␣ after different time intervals were measured. Fig. 4B shows that IB-␣ protein is stable, with a half-life of 2-4 h in Tax S258A cells. The same stability (half-life, 2-4 h) was observed in an uninfected human T-cell line SS8-BPT (data not shown). In contrast, the IB-␣ protein has a shorter half-life in the presence of Tax wt (Fig. 3B). These data additionally corroborate that NF-B activity is not inappropriately elevated in cells expressing the Tax S258A .
As a third measure of the NF-B activation status, we analyzed the expression of transfected reporter genes. One indicator plasmid used contained four copies of an NF-B response element positioned upstream of a minimal promoter, and the second contained the promoter from the HTLV-1 LTR. The CAT assays were normalized by co-transfecting a constitutively active luciferase gene. We found that the activity of the HTLV promoter was higher compared with that of the NF-B test promoter in all cells immortalized by the tax wt and tax S258A genes and in the HTLV-1-transformed controls (Fig. 5A). Consistent with Tax-mediated activation of its own promoter, the activity of the HTLV promoter was much higher in these cells  (Tax S258A ) and TAXI-1 (Tax wt ) and MT-2 (HTLV-1) cells. B, content of IL-2R␣ mRNA in the indicated cells. Northern blots were quantitated by phosphorimaging and normalized to the GAPDH signal. The mean of two independent experiments is shown. PSL, phosphor-stimulated luminescence. C, the cells were transfected with CAT indicator constructs containing the unchanged IL2-R␣ promoter (black) and a deletion variant containing the IL-2 response element but not the NF-B response element (gray). A luciferase plasmid (pRSVluc) was used as an internal standard. The mean of two independent experiments is shown. Relative CAT activity normalized to transfections with the deleted promoter is shown. than in the Tax-negative control T-cells (Jurkat). Jurkat cells show even higher NF-B responses than HTLV-1 promoterresponsive expression (Fig. 5A). This confirms that the Tax wt and Tax S258A proteins were similarly functional in transactivating the HTLV-1 LTR. By contrast, whereas the NF-B indicator construct was strongly expressed in all cells containing Tax wt (23%, 35% of HTLV promoter activity) (Fig. 5B), its activity was low compared with the HTLV-1 CAT construct in Tax S258A cells (3-5% of HTLV promoter activity). This provides further support to the hypothesis that NF-B is not aberrantly activated in Tax S258A-immortalized cells and suggests that growth stimulation of these cells is more likely a consequence of NF-B independent activation.

DISCUSSION
How HTLV-1 Tax transforms cells is controversial. Although it is agreed that Tax can activate CREB/ATF and NF-B, there is disagreement about the pathway correlating with transformation (14,39). Thus when immortalized rodent cells were used each of these pathways have been linked to transformation. Because immortalized rodent cells are an artificial model for the development of adult T-cell leukemia, we asked what is required of Tax in the first step of transformation (i.e. immortalization) of primary human lymphocytes. The experiments described here indicate that a Tax mutant which activates CREB/ATF but cannot activate NF-B can immortalize human primary T-lymphocytes. We verified that the immortalizing phenotype was not due to fortuitous activation of NF-B since these cells were found to contain greatly reduced levels of NF-B activity. Our findings support an NF-B-independent pathway for Tax immortalization of human T-lymphocytes.
Although the NF-B phenotype of the mutant S258A was thoroughly established by two groups, it is formally not possible to exclude that some unmeasurable low effect on NF-B could have contributed to the T-cell immortalization. The residual IL-2R␣ transcription in the mutant immortalized cells does not indicate residual NF-B activity, since the promoter is activated by its IL-2 response elements. By using many independent criteria, we found that cells immortalized by the rhadinovirus tax hybrids resemble HTLV-1-immortalized lymphocytes. These cells express virtually identical cytokine patterns and have similar patterns of surface markers. Like HTLV-1transformed lymphocytes, Tax-immortalized lymphocytes can lose the CD-3 surface marker expression (8). This finding is unique to HTLV-1-associated immortalization since this does not occur with H. saimiri immortalized lymphocytes (32). Three facts support that Tax induced the immortalization of the primary T-cells and that rhadinovirus genes did not contribute to the immortalized phenotype. 1) The only transcription found in human T-cells originated from the genomic region, which is deleted in the vector strain used (40). 2) No herpes viral transcripts were found in T-cells infected and immortalized by the recombinants expressing the Tax wt gene (41).
3) The cells infected with vector expressing Tax S258A did not synthesize infectious herpesvirus particles. These findings, besides the observation that the transduction with the H. saimiri vector produces no immortalization, indicate that the cellular growth changes are a direct consequence of HTLV-Tax expression.
The human primary lymphocytes transduced by the Tax mutant S258A are immortalized since they have an extended life span, and they proliferate in the absence of antigen presenting cells. The cells resemble HTLV-1-infected cell cultures derived from patients both in surface marker expression and growth factor requirements. These findings underline the proximity of this model system to ATL lymphocytes. We are not persuaded that rodent fibroblast models for Tax transformation have similar physiological relevance. In this regard we note that in other systems Tax proteins of HTLV-1 and BLV have been shown to exert an immortalizing effect on primary rat fibroblasts (12,42) and transforming activity in established rat fibroblast cell lines. Thus Yamaoka and colleagues (14) showed that Tax, which is deficient in NF-B activation, does not transform the Rat-1 fibroblast line. Surprisingly, a different group showed that the Rat-2 cell line could be morphologically transformed by a Tax mutant, which was deficient for NF-B activation (39). Confusion about the impact of Tax mutations on fibroblast transformation probably stems from system to system variations (i.e. Tax mutants used) and cell type differences. These differences suggest cautiousness in extrapolating results about the role of Tax in transformation from rodent fibroblast cell line systems to human primary lymphocytes.
We believe that our current findings contribute toward the development of a relevant human primary T-cell system for studying HTLV-1-induced ATL. Unlike the findings in rodent cell lines, in primary human T-cells, NF-B activation by Tax is evidently not a critical component of immortalization. This observation agrees with the finding from Smith and Greene (39). However, our results do not exclude that in other settings activation of NF-B can be a potent modulator/inductor of cellular proliferation.
Acknowledgments-We thank the AIDS Research and Reference Reagent Program (Division of AIDS, NIAID, National Institutes of Health). We also thank Patrick Baeuerle, Brigitte Biesinger-Zwosta,

FIG. 5. Reduced NF-B activity in cells immortalized with the Tax mutant.
Tax S258A -transduced cell lines (NATI-1, NATI-2, and NATI-3) and controls (Jurkat, TRI-1, and MT-2) were transfected with CAT indicator plasmids. TRI-1 is a T-cell line immortalized by transduction of Tax wt ; MT-2 is an HTLV-infected T-cell line. The reporter plasmids contain either the HTLV-1 promoter or an artificial NFB-dependent promoter. A, the CAT extracts were standardized by luciferase activity derived from co-transfected pRSV-luc. B, CAT activities from three independent experiments were quantitated, and the mean was determined. To exclude variations resulting from different transfection efficiencies of the various cell lines, the activity of the NF-B test construct was normalized to the activity of the HTLV-1 promoter.