IFN Regulatory Factor 4 Participates in the Human T Cell Lymphotropic Virus Type I-Mediated Activation of the IL-15 Receptor α Promoter1

IL-15Rα mRNA and protein levels are increased in human T cell lymphotropic virus type-I (HTLV-I)-associated adult T cell leukemia. Previously, we demonstrated that IL-15Rα expression was activated by HTLV-I Tax, in part, through the action of NF-κB. However, there appeared to be additional motifs within the IL-15Rα promoter that were responsive to HTLV-I Tax. In this study, we demonstrated that IL-15Rα mRNA expression was activated in human monocytes by IFN treatment, suggesting a role for IFN regulatory factors (IRFs) in IL-15Rα transcription. In addition, an IRF element within the Tax-responsive element of the IL-15Rα promoter was necessary for maximal Tax-induced activation of this promoter. Furthermore, we demonstrated that IRF-4, a transcription factor known to be elevated in HTLV-I-infected cells, activated the IL-15Rα promoter. Inhibition of IRF-4 action lead to reduced Tax-induced activation of the IL-15Rα promoter, while inhibition of both IRF-4 and NF-κB severely inhibited the Tax-induced activation of this promoter. These findings suggest a role for both NF-κB and IRF-4 in the transcriptional regulation of IL-15Rα by HTLV-I Tax. It is possible that the HTLV-I Tax-mediated induction of IL-15Rα and IL-15 may lead to an autocrine cytokine-mediated stimulatory loop leading to the proliferation of HTLV-I infected cells. This loop of proliferation may facilitate viral propagation and play a role in HTLV-I-mediated disease progression.

I nterleukin-15 is a member of the 4␣-helix bundle cytokine family that shares similar properties with IL-2 (1, 2). These functional similarities can be explained in part by the use of common receptors. IL-2 and IL-15 share the IL-2R␤ and the common ␥ receptor subunits (1)(2)(3)(4), yet each cytokine has its own distinct receptor, namely IL-2R␣ (5) and IL-15R␣ (6), respectively. IL-2 and IL-15 use the ␤ and common ␥ chains to initiate similar signal transduction pathways. This common signaling pathway contributes to the shared functions of IL-2 and IL-15 in both T and NK cells.
There are expression and functional differences between IL-2 and IL-15. IL-2 mRNA is largely restricted to lymphoid tissues, yet IL-15 has a widespread mRNA expression in many cells and tissues including T cells, B cells, liver, and skeletal muscle (7). IL-15 also has activities that are not shared with IL-2. For example, addition of IL-15 to a myoblast cell line affected skeletal muscle fiber hypertrophy, suggesting that IL-15 may be an anabolic agent that increases skeletal muscle mass (8). IL-15 also plays a major role in the development, survival, and activation of NK cells (9 -12). IL-2 and IL-15 also have profoundly different effects on activation-induced cell death. IL-2 activates self-reactive T cell suicide and thus, plays a role in peripheral tolerance (13)(14)(15). IL-2 is also important in the inhibition of CD8 ϩ memory T cell maintenance (16). In contrast, IL-15 has an anti-apoptotic effect on T and B cells (17), inhibits IL-2-induced activation-induced cell death (18), and is critical for the survival of CD8 ϩ memory cells (16,19). IL-15 also stimulates the proliferation of mast cells (20).
The effects of IL-15 can be explained in part not only by its widespread mRNA expression, but also in part by the expression of its distinct receptor IL-15R␣. IL-15R␣ is a 58 -60-kDa type I transmembrane protein that does not belong to the cytokine receptor family (6). IL-15R␣-mediated functions were demonstrated in IL-15R␣ null (IL-15R␣ Ϫ/Ϫ ) mice (21). These mice are deficient in NK cells, CD8 ϩ lymphocytes, and TCR ␥␦ intraepithelial lymphocytes. In addition, IL-15R␣ knockout mice exhibit marked lymphopenia due to decreased homing of lymphocytes to peripheral lymph nodes. Furthermore, IL-15 knockout mice exhibited marked reductions in the number of thymic and peripheral NK cells, CD8 ϩ lymphocytes, and populations of intraepithelial lymphocytes (22). These findings suggest that both IL-15 and its binding receptor are necessary for the development of NK and some T cells.
IL-15 has been shown to be elevated in a number of diseases including rheumatoid arthritis, inflammatory bowel disease, human T cell lymphotropic virus type-I (HTLV-I) 3 -associated adult T cell leukemia (ATL) (23), and HTLV-I-associated myelopathy/ tropical spastic paraparesis (HAM/TSP) (24). In addition, IL-15R␣ levels are also elevated in the T cells of ATL patients (25). Increased levels of IL-15 and IL-15R␣ in HTLV-I-associated diseases are induced by the HTLV-I Tax protein. Tax is expressed from the pX sequence within the HTLV-I proviral genome (26) and is responsible for the transactivation of the HTLV-I long terminal repeat (LTR) (27). In addition, HTLV-I Tax activates a variety of cellular host genes including IL-2 (28,29), IL-2R␣ (29,30), IL-15 (23), and IL-15R␣ (25). Activation of cellular genes by HTLV-I Tax is mediated by a number of cis-acting DNA elements including cAMP responsive element (26), serum responsive elements (31), and NF-B motifs (32). Many of these genes are ILs or growth factors that may aid the virus in its propagation.
HTLV-I infection also plays a role in the expression of IFN regulatory factor (IRF)-4. IRFs are activated upon viral infection or IFN activation and act as transcription factors (33). IRF-4 was originally cloned as ICSAT, an IFN consensus binding protein that was overexpressed in ATL cell lines (34). This finding suggested that IRF-4 was activated by HTLV-I infection. Additional studies revealed that IRF-4 mRNA expression was elevated in Jurkat cells following transient expression of the HTLV-I Tax protein (34), implying that Tax was involved in the transactivation of the IRF-4 gene. Furthermore, IRF-4 is constitutively expressed in HTLV-Iinfected cell lines and the IRF-4 promoter is activated by Tax expression (35). In addition, the peripheral blood cells of ATL patients have elevated levels of IRF-4 protein. 4 Up-regulation of IRF-4 by HTLV-I Tax suggests that IRF-4 may be involved in the transformation of T cells by HTLV-I (35).
We have previously demonstrated that IL-15R␣ is activated by HTLV-I Tax in part through the action of NF-B (25). In this study, we delineate additional Tax-responsive elements within the IL-15R␣ promoter. Specifically, we demonstrated that an IRF element (IRF-E) was necessary for Tax-induced activation of the IL-15R␣ promoter. In addition, we showed that IRF-4 was capable of activating the IL-15R␣ promoter and that inhibition of IRF-4 action reduced the Tax-induced activation of the IL-15R␣ promoter. These findings suggest that HTLV-I Tax is capable of activating IL-15R␣ in part through the action of IRF-4.

Analysis of alternative Tax-responsive elements within the IL-15R␣ promoter
To delineate additional Tax-activated sites within the Tax responsive element of the IL-15R␣ promoter (Del.1/pGL3pro), deletion constructs were made that included or excluded the NF-B site (Fig. 1). Plasmid 209 contained the NF-B site (bases Ϫ970 to Ϫ1061), while plasmid 207 contained the remaining bases in the Del.1/pGL3pro construct (bases Ϫ844 to Ϫ971), excluding the NF-B site and all bases 5Ј to it. Sense and anti-sense primers of the promoter inserts were prepared with KpnI and XhoI restriction sites at the 5Ј and 3Ј ends, respectively. To anneal the primers, 10 g of each primer was added to 10 mM NaCl and placed in a 95°C heat block. The heat block was immediately placed at room temperature and the primers were allowed to anneal gradually. Once the heat block attained room temperature, the annealed primers were ethanol precipitated. The precipitated DNA was then ligated to the pGL3 promoter vector using the XhoI and KpnI sites overhangs found on the annealed primers (see primer sequences below) and the resulting product was transformed into One Shot (Invitrogen, San Diego, CA) competent cells. The primer sequences for plasmid 207 were as follows: 5Ј-C(XhoI)ATTGTTAATTTTTAAATTT GATCTTATCATCTTGAGATTTATTATTAAGTATTGAGAGAAATT GAATAATGTGGGATTTCCCCAGTTGGAC(KpnI)-3Ј (sense) and 5Ј-TC GAG(XhoI)TCCAACTGGGGAAATCCCACATTATTCAATTTCTCTC AATACTTAATAATAAATCTCAAGATGATAAGATCAAATTTAAAA ATTAACAATGGTAC(KpnI)-3Ј (antisense). The primer sequences for plasmid 209 were as follows: 5Ј-C(XhoI)GTAAGGGTGGCTATTCC TGCTTAGAAAAAAAGAATGGACCTTGTGTGGTCACTGCCGGATG GTAGGTTCATTATTCTTTCTACTTTTTTTTTTTAAATTTAAGAGAC AGGAACTGGCTGTGTTGCCCAC(KpnI)-3Ј (sense) and 5Ј-TCGAG (XhoI)TGGGCAACACAGCCAGTTCCTGTCTCTTAAATTTAAAAAA AAAAAGTAGAAAGAATAATGAACCTACCATCCGGCAGTGACCACA CAAGGTCCATTCTTTTTTTCTAAGCAGGAATAGCCACCCTTACG GTAC(KpnI)-3Ј (antisense). The HTLV-I LTR/pGL3 luciferase construct was a kind gift from J. Brady (National Cancer Institute, National Institutes of Health, Bethesda, MD). Plasmids were tested for luciferase activity in Jurkat cells using transient transfection assays. Four million cells containing 100 ng of luciferase construct and 2 g Tax/pBC72 or Tax M22/pBC72 expression plasmids were electroporated at 280 V and 950 F. Electroporated cells were transferred into 6-well plates and luciferase activity was determined at 24 h posttransfection. All results are reported as the fold induction over that of the pGL3 promoter alone and represent an average of three independent experiments. Error bars represent the SD of the fold induction of samples.

Detection of IL-15R␣ mRNA by Northern blot analysis
To analyze the possible role of IFNs in the regulation of IL-15R␣, humanelutriated monocytes were treated under various conditions. A total of 3 ϫ 10 7 freshly elutriated monocytes were treated with media, PHA (1 g/ml; Sigma-Aldrich, St. Louis, MO), LPS (1 g/ml; Calbiochem, La Jolla, CA), IFN-␣ (1000 U/ml; Biosource International, Camarillo, CA), IFN-␤ (1000 U/ml; Biosource International), or IFN-␥ (1000 U/ml; R&D Systems, Minneapolis, MN) for 6.5 h at 37°C. Cells were also treated with combinations of IFN and LPS. Immediately following treatment, RNA was extracted from each sample using the Purescript RNA isolation kit (Gentra, Minneappolis, MN). A total of 10 g of total RNA was run on a formaldyhyde containing agarose gel and transferred to nitrocellulose. The blot was probed with a labeled IL-15R␣, IL-15 probes, and subsequently with an ␤-actin probe to analyze mRNA loading. Fold induction of IL-15R␣ expression was analyzed using phosphoimager quantitation. Samples were normalized to ␤-actin levels. Fold induction of IL-15R␣ was calculated based on the media alone control.

Analysis of IL-15R␣ activation by IRFs
Cotransfection studies were performed using 50 ng of the full-length IL-15R␣ (IL-15R␣/pGL3) reporter construct and 2 g of IRF-1/pACT (36), IRF-3 (5D)/CMVBL (37), IRF-4/pCDNA3 (38), and IRF-7/pFlag-CMV-2 (39) expression plasmids. COS-7 cells were transfected using the DEAE Dextran method in 6-well dishes. Briefly, 2.5 ϫ 10 5 cells were seeded per well and incubated overnight. Cells were washed with PBS and transfection mixtures (300 l PBS containing DNA and 300 g DEAE Dextran) were added for 30 min at 37°C, gently rocking every 5 min. Following the incubation, 1.5 ml DMEM media containing 80 M chloroquine was added to each well and plates were incubated at 37°C for 2.5 h. Chloroquine was aspirated from the wells and 1.5 ml of DMEM containing 10% DMSO was added to each well for 2.5 min. Media was aspirated from the wells and 4 ml of DMEM was added to each well. Cells were incubated at 37°C and luciferase activity was measured at 24 h posttransfection. Assays were performed in triplicate and error bars represent the SD of the fold induction of samples over that of the negative control (pGL3 basic).
In addition, the effect of IRF-4 inhibition on Tax activation of the IL-15R␣ promoter was analyzed in cotransfection studies in COS-7 cells. A total of 50 ng of the IL-15R␣/pGL3 reporter construct, 100 ng Tax/pBC72 (a kind gift from J. Brady), and 2 g of FKBP52/pFLAG (38) and/or 2 g super dominant IB␣/pCDNA3 were cotransfected using the DEAE Dextran method described above. Luciferase assays were performed as previously described.

Western blot analysis of IRFs
Transfected COS-7 cells were trysinized from plates and divided in half for luciferase activity or radioimmunoprecipitation buffer lysis. Thirty micrograms of each sample were run on a 4 -12% NuPage SDS polyacrylamide gel (Invitrogen) and transferred to polyvinylidene difluoride membrane. Membranes were blocked with Super Block (Pierce, Rockford, IL) overnight and probed with anti-IRF-1, anti-IRF-3, anti-IRF-4, and anti-IRF-7 (all from Santa Cruz Biotechnology, Santa Cruz, CA). Reprobing with anti-vinculin (Sigma-Aldrich) served as a loading control.

Analysis of transcription factor consensus sequences using EMSA
We performed EMSA using whole-cell extracts from HTLV-I-infected MT-2 cells. Cells were lysed in modified lysis buffer (50 mM Tris-HCl, 100 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 30 mM sodium pyrophosphate, 0.5% Nonidet P-40, complete protease inhibitor and diisopropylfluorophosphate) and incubated on ice for 15 min. Lysates were centrifuged at 42,000 rpm for 30 min at 4°C. The probes used in this assay were double stranded 32 P-labeled oligonucleotides encompassing the IL-15R␣ IRF-E (IL-15R␣ IRF-E) motif (GGTAGGTTCATTATTCTTTCTACTTTTTT TTTTT) or a consensus IRF-E motif (cIRF-E)(CTAGCGAGAAATA AAAGGAAGTGAAACCAAGT) from the B element (35). Extracts were then mixed with radiolabeled probes, unlabeled competitor probes, poly (dI-dC), Buffer D (20 mM HEPES (pH 7.9), 20% v/v glycerol, 100 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, and 0.5 mM PMSF) and BSA at room temperature for 45 min. Probes for competition included the IL-15R␣ IRF-E, the cIRF-E, and a nonspecific consensus NF-B probe (25). An Ab to IRF-4 (Santa Cruz Biotechnology) was added to the cell extracts for 30 min on ice. Extracts were then mixed with the IL-15R␣ IRF-E radiolabeled probe, poly(dI-dC), and BSA at room temperature for 30 min. Samples were loaded onto an acrylamide (30%)/bis-acrylamide (0.8%) gel and subjected to electrophoresis at 150 V for the initial 10 min followed by 120 V for the remainder of the run. Gels were dried and exposed to Kodak MS film (Kodak, Rochester, NY).

Additional Tax-responsive elements within the IL-15R␣ promoter
In previous studies, we demonstrated that the IL-15R␣ promoter was induced by HTLV-I Tax expression (25). This promoter activation was mediated in part by an NF-B motif within a 220 bp Tax-responsive element of the promoter. In addition, mutation of the NF-B motif within the Tax-responsive region (Del.1/ pGL3pro) as well as mutation of the NF-B motif within the fulllength promoter (IL-15R␣ pro/pGL3) dramatically reduced Tax activation of the IL-15R␣ promoter. In this study, to further address the importance of NF-B in the Tax-induced activation of the IL-15R␣ promoter, we demonstrated that a mutant form of Tax that does not activate NF-B (M22) also eliminates Del.1/ pGL3pro promoter activity (data not shown). In addition, a reporter construct bearing the HTLV-I LTR was activated by M22, thus demonstrating the necessity of NF-B activation for IL-15R␣, but not the HTLV-I LTR. Previous studies showed that the HTLV-I LTR is not significantly activated by NF-B (40). Furthermore, the level of activation of the Del.1/pGL3pro was comparable to that of the HTLV-I LTR with wild-type Tax. This finding further demonstrates that IL-15R␣ promoter activity needs NF-B.
Cotransfection studies previously performed (25) with the Del.1/pGL3pro, Tax, and super dominant IB␣ showed that IB␣ could inhibit the Tax-induced activation of the IL-15R␣ promoter mediated by NF-B. However, we did not demonstrate complete inhibition of Tax activation using this NF-B inhibitor. This finding also suggested that additional sites within the first 200 bp of the IL-15R␣ promoter were activated by Tax. To delineate these sites, we made deletion constructs within the 200 bp region and analyzed each for Tax responsiveness.
Del.1/pGL3pro was first divided into two regions, one containing the NF-B site (plasmid 209) and the other containing the 3Ј end of this fragment (plasmid 207) (Fig. 1A). Each plasmid was analyzed for promoter activity in the presence or absence of wildtype Tax expression (Fig. 1B) in Jurkat cells. As expected, plasmid 209 was responsive to Tax expression. Interestingly, plasmid 207 was also activated by Tax expression. This finding indicated the presence of additional Tax-responsive elements within the IL-15R␣ promoter.
Based on this observation, plasmid 207 was divided into four segments and each segment was cloned into the pGL3 promoter reporter construct (Fig. 1A). These plasmids (254, 255, 256, and 257) were then used in cotransfection studies in the presence or absence of Tax expression in Jurkat cells. As shown in Fig. 1B, reporter activities of plasmids 255 and 257 were not enhanced by Tax expression at the same level as plasmid 207. However, plasmids 254 and 256 were activated by Tax expression at a level greater than that of plasmid 207. This suggested that motifs located in these regions of the promoter were important for the additional Tax-induced activation of plasmid 207. Furthermore, increased Tax-induced activation of these promoter regions suggested that negative regulatory elements found in the sequence of 207 were eliminated in plasmids 254 and 256, thus increasing Tax-induced activation.
Analysis of the DNA sequence in these regions of the IL-15R␣ promoter (GenBank accession no. AF283296) showed the presence of two putative IRF-Es. This was interesting in that the IL-15 promoter was activated in part by NF-B, IRF-1, and IRF-3 transcription factors (23,41). Knowing that IL-15 and IL-15R␣ were activated similarly by NF-B, it was possible that IL-15R␣ was also similarly regulated by IRFs.

IL-15R␣ mRNA expression in human moncytes was increased upon treatment with IFNs and LPS
To examine whether IL-15R␣ expression was influenced by IRFs, we first analyzed IL-15R␣ mRNA expression in elutriated monocytes following treatment with IFNs. Monocytes were used because they are known to up-regulate both IL-15 and IL-15R␣ upon treatment with IFNs (6,42). IRFs are transcription factors induced upon viral infection and/or upon treatment with IFNs. If IRFs were responsible for IL-15R␣ activation, treatment with IFNs would likely activate IL-15R␣ mRNA expression.
Elutriated human monocytes were treated with PHA, LPS, IFN-␣, IFN-␤, IFN-␥, or LPS ϩ IFN-␣, -␤, or -␥ for 6.5 h. Elutriated monocytes were used to obtain a homogeneous population of monocytes capable of responding to IFN stimulation. Following treatment, total RNA was isolated from each condition and analyzed in a Northern blot assay for IL-15R␣, IL-15, and ␤-actin expression. Blots were analyzed by a phosphoimager and IL-15R␣ induction was measured after normalization for ␤-actin loading. As shown in Fig. 2, PHA had no effect on IL-15R␣ expression, yet LPS induced IL-15R␣ expression 12.  (41). These findings demonstrated that IFNs were capable of activating IL-15R␣ mRNA expression. In turn, this suggested that IRFs played a role in the transcriptional regulation of the gene.

Mutations in the IRF-E within the IL-15R␣ promoter lead to reduced Tax-induced activation
Mutations within the IRF-Es of the IL-15R␣ promoter were made to analyze their effect on Tax-induced activation of these regions of the promoter in cotransfection studies using Jurkat cells. First, mutations were made in plasmids 254 and 256, the smallest regions of the IL-15R␣ promoter that were induced by Tax expression. All mutations were made within the putative IRF-Es at sites consistent with consensus IRF-E motifs. As shown in Fig. 3, mutation of three bases within the putative IRF-E at positions Ϫ943 to Ϫ945 (TTT3 CCC) using site-directed mutagenesis (plasmid 261) reduced the Tax-induced activation of this site 56%. This suggested that the putative IRF-E at this site was important for Tax-induced activation. In addition, mutations of plasmid 256 were made at positions Ϫ894 to Ϫ896 (TTT3 CCC) using sitedirected mutagenesis (plasmid 262). Tax-induced activation of this plasmid was inhibited 87% when compared with that of the wildtype 256 plasmid (Fig. 3). Interestingly, plasmids 254 and 256 were not activated by the NF-B inactive mutant of Tax, M22 (data not shown). Although this result was surprising, it was shown previously by Grumont and Gerondakis (43) that IRFs could be activated by NF-B. Therefore, in the context of plasmids 254 and 256, NF-B might activate the IRF that acts on the IL-15R␣ IRF-E sites (see Discussion). These observations indicated that the IRF-E within these isolated regions of the promoter and were essential for the Tax responsiveness of the wild-type promoter.
Although we demonstrated an effect on Tax activation by mutating the individual IRF-Es found within plasmids 254 and 256, we also wanted to examine their effect on the full-length IL-15R␣ promoter in Jurkat cells to determine whether they were important  for the Tax-induced activation in the presence of the NF-B site. This was important because previous studies showed that an NF-B site located within the first 200 bp of the promoter was important for Tax activation (25). We also analyzed the effects of individual mutations of the NF-B site and both of the IRF-Es on the Tax-induced activation of the full-length promoter (Fig. 4). Mutation of the first IRF-E site in plasmid 263 had no effect on the Tax-induced activation of the IL-15R␣ promoter. This indicated that this IRF-E site played no essential role in the Tax activation of the full-length promoter. This also suggested that in isolation, this element could be used by Tax, yet in the context of the full-length promoter, Tax worked through alternate motifs. Mutation of the second IRF-E site in plasmid 270 showed a 41% decrease in Tax responsiveness. This decrease suggested that this region of the promoter contributed to the Tax activation of the full-length pro-moter. Finally, mutation of the NF-B site (plasmid 230) again contributed to a severe inhibition (75%) of Tax activation. These findings demonstrated that the second IRF-E site and the NF-B site were important in the Tax activation of the IL-15R␣ promoter.
We also analyzed the Tax-induced activation of plasmids containing double and triple mutations. Plasmid 272 was mutated at the first IRF-E and the NF-B site, while plasmid 271 was mutated at the second IRF-E and the NF-B site. Both of these plasmids had a reduction in Tax responsiveness, by 56 and 78%, respectively. Mutation of the first IRF-E alone in plasmid 263 had no effect on the Tax activation of the full-length promoter; therefore, this analysis indicated that the NF-B site appeared to be the dominant factor in the Tax activation of the promoter. In addition, a double mutation of both IRF-E sites showed a 50% decrease in Tax-induced activation similar to that of the mutation of the second IRF-E site alone (plasmid 270). Furthermore, a triple mutation of both IRF-E sites and the NF-B site also reduced the Taxinduced activation of the promoter by 70%. In fact, the activation level of plasmids 271 and 273 in which IRF-E sites and the NF-B motif were mutated were similar to that of plasmid 230 in which the NF-B site alone was mutated, again indicating that the NF-B motif was the dominant site of Tax-induced activation. Taken together, these findings suggested that the second IRF-E site was important for the Tax-induced response of the IL-15R␣ promoter. Furthermore, these findings showed that activation of NF-B is the crucial element necessary for Tax-induced activation of the IL-15R␣ promoter. This suggested that the IRF-E sites within the promoter were activated by Tax; however, the NF-B motif played a dominant role in the Tax-activation of the IL-15R␣ promoter.

IRF-4 is involved in the activation of the IL-15R␣ promoter
Mutational analysis of deletion constructs and the full-length promoter demonstrated that an IRF-E site was important for the transactivation of the IL-15R␣ promoter (IL-15R␣pro/pGL3). We next analyzed the promoter activity of the IL-15R␣ reporter construct in cotransfection assays of COS-7 cells using expression plasmids for various IRFs. We used COS-7 cells instead of Jurkat cells as shown above due to the higher transfection efficiency of COS-7 cells. Detectable levels of IRFs were seen in these cells as demonstrated by Western blot analysis (see below). In this study, cells were cotransfected with IRF-1 and IRF-3 (5D), a constitutively active form of IRF-3, both of which were previously shown to activate the IL-15 promoter (41). In the same experiment, we also cotransfected IRF-7, an IRF that is restricted to lymphoid cells, and IRF-4, an IRF expressed in HTLV-I infected cells. Although  IRF-1, IRF-3 (5D), and IRF-7 were capable of activating the IL-15R␣ promoter 12, 10-, and 30-fold, respectively, IRF-4 expression demonstrated the strongest activation with an 185-fold increase over the pGL3 basic construct (Fig. 5A). Expression patterns of the various IRFs are demonstrated in Fig. 5B by Western blot analysis to ensure comparable levels were expressed to activate the IRF-E within the IL-15R␣ promoter.
IRF-4 was isolated from an HTLV-I-infected ATL cell line (34). In addition, Yamagata et al. (34) showed that transient expression of Tax in Jurkat cells activated IRF-4 mRNA expression. Furthermore, Sharma et al. (35) showed that the IRF-4 promoter was activated by Tax expression in reporter assays. We showed that IL-15R␣ mRNA is also activated by Tax expression (25); therefore, IRF-4 was a good candidate for the activation of IL-15R␣ by Tax through the IRF-E site defined above. As seen in Fig. 5A, coexpression of IRF-4 greatly enhanced the activity of the fulllength IL-15R␣ promoter construct. This finding suggested that IRF-4 in HTLV-I-infected T cells is capable of activating the IL-15R␣ promoter.

An inhibitor of IRF-4 reduced the Tax-induced activation of the IL-15R␣ promoter
Initial studies demonstrated that HTLV-I Tax activated the IL-15R␣ promoter via an NF-B site (25). The super dominant IB␣ molecule inhibited Tax activation of the IL-15R␣ promoter; however, this inhibition was not complete. As demonstrated above, an IRF-E also played a role in the Tax-induced activation of the promoter. In addition, we demonstrated that IRF-4 activated the IL-15R␣ promoter in the absence of Tax expression. To examine the role of IRF-4 in the Tax-induced activation of the promoter, we performed transient coexpression assays in COS-7 cells with Tax and FKBP52, an inhibitor of IRF-4. FKBP52 inhibits IRF-4 DNA binding through its peptidyl-propyl isomerase activity (38). Furthermore, FKBP52 inhibits the action of IRF-4, but does not inhibit the action of NF-B. As shown in Fig. 6, Tax activated the IL-15R␣ promoter 83.2-fold over that of the pGL3 vector alone. The Tax-induced activation of the IL-15R␣ promoter was again inhibited by a super dominant IB␣ expression plasmid in a transient transfection assay. The expression of super dominant-IB␣ reduced the fold activation of the IL-15R␣ promoter by Tax to 2.8-fold over the pGL3 basic construct. In addition, the IRF-4 inhibitor FKBP52 also inhibited the Tax-induced activation of the promoter to 7.1-fold over the pGL3 basic plasmid. Tax activation of the IL-15R␣ promoter was also inhibited by the coexpression of both super dominant IB␣ and FKBP52 to 1.6-fold over that of pGL3 vector alone. These findings showed that both NF-B and IRF-4 played roles in the Tax-activation of the IL-15R␣ promoter.
In addition, these findings demonstrated that both NF-B and IRF-4 are necessary for the complete activation of the IL-15R␣ promoter by HTLV-I Tax.

The functional IRF-E within the IL-15R␣ promoter bound proteins in the lysates of the HTLV-I-infected T cell line, MT-2
To determine whether the IRF-E motif in the IL-15R␣ promoter was capable of binding IRF-4 proteins, we performed EMSA analysis using the lysates from the HTLV-I-infected T cell line MT-2. MT-2 cells were chosen because they express high levels of IRF-4 protein. 4 As shown in Fig. 7, the IL-15R␣ IRF-E motif exhibited binding in the absence of cold competitive probes (Fig. 7, lane 1). This binding was specific because cold IL-15R␣ IRF-E oligonucleotides competed out the binding of the labeled probe in a dosedependent manner (Fig. 7, lanes 2-5). In addition, the binding of the IL-15R␣ IRF-E was also competed out with the addition of a cold IRF-4 consensus probe (cIRF-E; Fig. 7, lanes 6 -9). This finding suggested that IRF-4 was involved in the binding seen in Fig.  7 (lane 1). Furthermore, cold consensus NF-B probe was used as a negative binding control. This probe did not compete for the binding of the labeled IL-15R␣ IRF-E probe (Fig. 7, lanes 10 -13), suggesting that the binding seen in Fig. 7 (lane 1) is specific for IRF-4. Furthermore, addition of an anti-IRF-4 Ab completely abrogated the binding of the labeled probe. This finding suggested that the interaction of IRF-4 with the IL-15R␣ IRF-E site was inhibited by the addition of Abs to IRF-4. Taken together, these findings suggested that cellular proteins are capable of binding the IL-15R␣ IRF-E, and that this binding is mediated by IRF-4.

Discussion
Previous studies showed that IL-15R␣ is elevated in HTLV-I-infected T cell lines and in the T cells of patients with ATL (25). Furthermore, we demonstrated that HTLV-I Tax activated IL-15R␣ expression through the action of NF-B. Although NF-B appeared to play a major role in the Tax-induced activity of the IL-15R␣ promoter, additional elements appeared to be involved. Expression of a super dominant IB␣ expression plasmid did not completely inhibit the Tax-induced activation of the promoter (25); therefore, additional transcription factors were implicated. To examine this issue, deletion constructs were made within the Tax responsive region of the promoter and tested for Tax inducible activity (Fig. 1). Interestingly, the Tax-induced activation was not limited to the constructs bearing the functional NF-B site. In fact, Tax responsiveness also localized to two putative IRF-Es.
IFNs are cytokines that are activated in response to viral pathogens. IFN-␣ and IFN-␤ are activated in many cell types upon viral infection, while IFN-␥ is produced in activated T cells and NK cells. Human monocytes treated with IFN-␥ and LPS have increased levels of IL-15 mRNA (42), suggesting that effector molecules downstream of IFN-␥ regulate IL-15 expression. In addition, IL-15R␣ mRNA levels were increased in human monocytes following treatment with IFN-␣, IFN-␤, and IFN-␥ (Fig. 2). Activation of genes by IFNs or viral infection is mediated by downstream transcription factors termed IRFs. These transcription factors in turn activate IFN-responsive genes. The IL-15 promoter is activated by IRF-1 and IRF-3 (41).
We asked which IRF was responsible for the Tax activation of the IL-15R␣ promoter. The initial study that characterized IRF-4 demonstrated that this factor is elevated in ATL patient cells (34). In addition, HTLV-I Tax activates the IRF-4 promoter in cotransfection studies (35). These findings suggest a role for IRF-4 in HTLV-I-associated disorders; however, they do not demonstrate the downstream targets of IRF-4 activation.
In this study, we showed through mutational analysis and cotransfection studies that the IRF-E sites within the IL-15R␣ promoter were necessary for its maximal Tax-induced activation (Figs. 3 and 4). The IRF-E site within the IL-15R␣ promoter was essential for maximal Tax activation; however, it is important to note that the NF-B site is the dominant factor involved in Tax activation. For example, IRF-E sites within the IL-15R␣ promoter were not activated by a Tax mutant deficient in NF-B activation. This result suggested that NF-B was involved in the activation of IRF-4 by HTLV-I Tax. Grumont and Gerondakis (43) previously demonstrated that IRF-4 expression was activated by NF-B in lymphoid cells. These findings taken together suggest that NF-B is essential for maximal Tax-induced activation of the IL-15R␣ promoter.
Furthermore, we demonstrated that IRF-4 activated the IL-15R␣ promoter in the absence of Tax (Fig. 5). This finding suggests that the Tax-responsive IRF-E site within the IL-15R␣ promoter is responsive to IRF-4. We also demonstrated that inhibition of IRF-4 by FKBP52 severely inhibited Tax-induced activation of the IL-15R␣ promoter (Fig. 6). FKBP52 exhibits peptidyl-propyl isomerase activity and interferes with the binding of IRF-4 to its DNA binding site. FKBP52 does not inhibit NF-B activity (38); therefore, the reduced Tax activation seen in this experiment is contributed to IRF-4 inhibition. Furthermore, we demonstrated that the IRF-E within the IL-15R␣ promoter was capable of binding IRF-4 proteins (Fig. 7). These findings suggested that IRF-4 activated the transcription of IL-15R␣ under the influence of Tax. Taken in concert with HTLV-I Tax activation of IL-15, activation of IL-15R␣ by IRF-4 and NF-B by HTLV-I Tax (Fig. 8) could represent an activation of a host immune response to retroviralinduced proliferation of HTLV-I-infected cells.

Model for IL-15R␣ in HTLV-I-associated diseases
IL-15R␣ expression is regulated by both NF-B and IRF-4 in HTLV-I-infected cells. Studies by Azimi et al. (23,41) showed that IL-15 is also activated by both NF-B and IRFs. HTLV-I Tax activates both IL-15 and IL-15R␣ levels, suggesting that activation of these genes is important for viral propagation. Patients with the HTLV-I-associated neurological disorder HAM/TSP maintain spontaneous proliferation of ex vivo PBMC cultures in the absence of exogenous cytokines or growth factors. Based on these data, it is possible that both IL-15 and IL-15R␣ participate in an autocrine/ paracrine loop of proliferation in HAM/TSP T cell cultures much like that demonstrated previously for IL-2 and IL-2R␣ (29). Therefore, we propose a model in which IL-2, IL-15, and their binding receptors are transcriptionally regulated by HTLV-I Tax. Upon HTLV-I infection, Tax induces the transcription of these cytokine systems via downstream transcription factors such as NF-B and FIGURE 7. The IRF-E site that is responsible for Tax-induced activation of the IL-15R␣ promoter functions like that of a consensus IRF-4 binding motif. EMSA analysis was performed on the Tax-responsive IRF-E site in the presence or absence of cold competition probes. Binding was inhibited in the presence of the specific IL-15R␣ IRF-E and cIRF-E probes while the nonspecific consensus NF-B probe had no effect on binding. The addition of an anti-IRF-4 Ab abolished IL-15R␣ IRF-E binding, suggesting that IRF-4 specifically binds to this site. IRFs. IL-15R␣ is transcriptionally regulated by these virally induced factors and in turn can participate in the spontaneous proliferation of HAM/TSP or ATL T cells. This production of cytokines and their receptors could lead to an autocrine/paracrine loop of spontaneous proliferation of T cells. Understanding the mechanisms behind the regulation of these cytokine systems could lead a to combinatorial approach directed against both IL-2 and IL-15 or their receptors for the treatment of patients with HTLV-I-associated diseases.