Synergistic Regulation of the Human Interleukin-12 p40 Promoter by NFκB and Ets Transcription Factors in Epstein-Barr Virus-transformed B Cells and Macrophages*

Monocytes/macrophages produce interleukin-12 (IL-12) in response to pathogenic stimulation, whereas most Epstein-Barr virus-transformed (EBV+) B cells constitutively secrete IL-12. The molecular mechanism regulating the constitutive IL-12 gene expression in EBV+ B cells has not been addressed. In this study, using the EBV+ B cell line RPMI-8866, we localized to the human IL-12 p40 promoter two essentialcis elements, the NFκB site and the Ets site. The NFκB site was shown to interact with members of the NFκB family: p50 and c-Rel. The Ets site constitutively bound a multi-component Ets-2-containing complex. While the NFκB and Ets sites appear equally critical for inducible p40 promoter activity in macrophage cell lines, NFκB plays a more dominant role in the constitutive p40 promoter activity in EBV+ B cells. Transient expression of Ets-2 and c-Rel in B, T, and monocytic cell lines synergistically activated the IL-12 p40 promoter, apparently overcoming the requirement for cell type- or stimulant-specific transcription factors. These data provide new evidence that full activation of the human IL-12 p40 promoter may result primarily from the interplay between NFκB and Ets family members.

Interleukin 12 (IL-12), 1 a heterodimeric cytokine composed of two disulfide-linked subunits of 35 (p35) and 40 (p40) kDa, was originally identified in the supernatant fluid of Epstein-Barr virus-transformed (EBV ϩ ) human B-cell lines (BCL) (1,2). IL-12 exerts multiple effects, including: (a) induction of cytokine production, particularly interferon-␥ (IFN-␥) by T and NK cells, (b) induction of proliferation in T and NK cells and enhancement of their cytotoxic activity, and (c) induction of T helper-type 1 (Th1) responses and inhibition of Th2 responses (3). The finding that IL-12 p40 knockout mice have a severely depressed Th1 response (4) supports the role of IL-12 in Th1 cell differentiation.
IL-12 is produced by phagocytic cells and other antigenpresenting cells in response to bacteria, bacterial products, intracellular pathogens, and viruses (1,2,(5)(6)(7)(8)(9)(10)(11). Normal B cells appear to be relatively poor producers of IL-12 (6). Most EBV ϩ BCL and EBV ϩ lymphomas produce IL-12, with the highest levels observed in EBV ϩ BCL derived from AIDSassociated lymphomas (9,12). SCID mice injected with human lymphocytes from EBV-seropositive donors developed EBV ϩ human B cell lymphoma secreting in vivo high levels of human IL-12, 2 suggesting that a similar production of IL-12 during initial proliferation of EBV ϩ B cell in patients may affect the reactivity of their immune cells against the infected cells. Although any extrapolation to the in vivo situation from data observed with established cell lines should be taken with caution, it is tempting to speculate that, in healthy individuals, IL-12-producing EBV ϩ cells would be easily rejected by immune response and thus only cells that have lost the ability to produce IL-12 would be able to give rise to Burkitt's lymphomas; in immunodeficient AIDS patients, these protective mechanisms would be inefficient and their IL-12-producing EBVtransformed cells could give rise to lymphomas in a relatively high proportion of patients.
The key role of IL-12 in inflammation and in the immune response, and the importance of this cytokine in anti-tumor resistance, have raised considerable interest in the mechanisms of IL-12 gene transcription. We (13) have shown previously that, in lipopolysaccharide (LPS)-and IFN-␥-treated human monocytes, the expression of IL-12 p40 and p35 is primarily regulated at the transcriptional level. In a luciferase reporter gene construct, a 3300-bp genomic DNA fragment including the upstream sequences of the human p40 gene largely recapitulates its cell type specificity and transcriptional regulation (13). The cis elements in the p40 promoter region responsible for the inducible activation of the genes have been studied extensively in phagocytic cells, but no information is available for the elements controlling the constitutive expression of the gene in EBV ϩ BCL. We have previously localized a DNA sequence spanning nucleotides Ϫ292 to Ϫ196 (relative to the transcription start site) responsible for induced promoter activity (13). The core element at position Ϫ212 binds a series of IFN-␥-and LPS-induced nuclear proteins (termed the F1 complex) including Ets-2, IFN-regulatory factor-1 (IRF-1), c-Rel, and Ets-related factors (14). A downstream NFB site between bp Ϫ117 and Ϫ107 was also characterized as an LPS response element in the murine macrophage J774 cell line (15). Recently, Plevy et al. (16) identified a third cis element located at Ϫ96 and Ϫ88 (downstream of the NFB site) of the murine IL-12 p40 promoter, also conserved in humans, that binds members of the C/EBP family of transcription factors in activated murine macrophage cell line. The C/EBP element exhibits functional synergy with the upstream NFB site, although no physical interactions between C/EBP and Rel proteins were observed (16). Together, these results suggest that in myeloid cells a complex interplay exists between multiple inducible transcription factors contributing to the regulation of IL-12 p40 gene activation.
In this study, we focused upon the role of cis-acting regulatory elements in the transcriptional activation of the human IL-12 p40 gene in EBV ϩ BCL. In these cells, we showed constitutive binding of multiple transcription factors to the NFB and Ets sites of the p40 promoter. We found that cotransfection of NFB and Ets-2 transcription factors strongly and specifically synergized in the induction of p40 promoter activity in EBV ϩ BCL as well as in both inducible IL-12-producing macrophage cell lines and IL-12-nonproducing EBV Ϫ and T cell lines.
Plasmids-A 3300-bp fragment of the human IL-12 p40 promoter was cloned into the luciferase reporter construct pXP2 (13) at the PstI site. All deletion mutant constructs were generated by PCR and fully sequenced for verification as described previously (13). NFB expression vectors were a gift from Dr. K. M. Murphy (Washington University School of Medicine, St. Louis, MO). The CMV-based murine Ets-2 expression vector was a gift from Dr. R. Maki. The TNF-␣ promoter was obtained from a TNF cosmid library (ATCC, catalog no. 57590) by PCR amplification of a 1294-bp fragment. The sequence of the 5Ј primer used is 5Ј-TCTCTGAAATGCTGTCTGCTTGT-3Ј and the 3Ј primer used is 5Ј-CCTCTTAGCTGGTCCTCTGC-3Ј. The PCR product was cloned into pXP2 expression vector. All plasmids were banded twice in cesium chloride and verified by restriction mapping.
Transient Transfection-DNA transfection experiments were performed by electroporation as described (13). Briefly, 0.4 ml of the cell suspension was mixed with 60 g of DNA (10 g of p40-luc/3.3 kb plus 3 g of pCMV-␤-galactosidase (transfection efficiency control) and 47 g of BS/KSIIϩ) and electroporated in 0.45-cm electroporation cuvettes (Gene Pulser, Bio-Rad) at 960 microfarads and 250 V for BJAB and Jurkat cells, 300 V for RPMI-8866 cells, and 350 V for RAW264.7 cells. Cells were harvested 24 h post-transfection, pelleted by centrifugation, and resuspended in 100 l of lysis buffer (Promega, Madison, WI). Lysates were used for both luciferase and ␤-galactosidase assay (13).
For cotransfection studies, pRSV-␤-galactosidase was used instead of pCMV-␤-galactosidase, and the molar ratio between the reporter plasmid (luciferase p40 promoter constructs) and expression vector was 1:2.5 for Rel family members and 1:1 for the Ets-2.
Luciferase activity was corrected for transfection efficiency by normalizing to the measured ␤-galactosidase value.

Constitutive Transcription of the Human IL-12 p40
Gene in EBV-transformed B Cells-Based on our previous observation that EBV ϩ BCL constitutively produce IL-12 protein (1, 6, 9), we tested whether the p40 and p35 genes were also constitutively transcribed in nuclear run-on assays performed with unstimulated EBV ϩ BCL (Fig. 1A). Nascent p40 transcripts were constitutively present in RPMI-8866 (Fig. 1A) and in all other EBV ϩ BCL tested (PA682BM-2 and AS283A), and expression was not further enhanced by TPA treatment (data not shown). Jurkat T cells and BJAB EBV Ϫ BCL, which do not express p40 mRNA or protein, were used as negative controls (Fig. 1A). Transcription of the IL-12 p40 gene, unlike that of the [␣-32 P]UTPlabeled nuclear RNA was hybridized to PCR-amplified p40, p35, TNF-␣, and ␤-actin cDNAs (500 ng/slot) immobilized on nylon membranes. Results are from one of two experiments with superimposable results. B, transfection of EBV ϩ , EBV Ϫ , and T cell line. RPMI-8866, BJAB, and Jurkat were transfected by electroporation with the IL-12 p40 promoter-luciferase construct (Ϫ3300/ϩ108) or the promoterless control plasmid (pXP2). After 24 h, luciferase activity was determined by normalizing the relative light units from each transfectant for transfection efficiency, using ␤-galactosidase activity determined in the same cell extract. Results are mean (Ϯ S.E.) from four independent experiments. TNF-␣ gene, was not detectable in unstimulated cells or in the EBV Ϫ BCL or T cell lines stimulated with TPA alone or TPA in combination with ␣-CD3, respectively. However, TNF-␣ transcription was shown to be further up-regulated by TPA treatment of all cells examined. Nascent p35 transcripts were present in all cell lines and under all conditions tested. Thus, the IL-12 p40 gene appears to be constitutively transcribed in EBV ϩ but not in EBV Ϫ BCL or T cells. We then analyzed the IL-12 p40 promoter activity in the same EBV ϩ , EBV Ϫ and T cell lines using a full-length promoter construct linked to a luciferase reporter gene (Ϫ3300/ϩ108) in a transient transfection assay. The promoter construct was active in EBV ϩ BCL, but poorly or not active at all in EBV Ϫ and T cell lines (Fig. 1B), confirming that the construct contains adequate sequence information for appropriate cell type-specific expression (13).
Important cis-Regulatory Elements Located in the Ϫ222/ ϩ108 bp Region of the IL-12 p40 Promoter-To delineate the cis-acting elements regulating the constitutive IL-12 p40 expression in EBV ϩ B cells, we first examined the activity of the full-length promoter (Ϫ3300/ϩ108) and a series of nested 5Ј deletion constructs (Fig. 2) in EBV ϩ BCL. Luciferase activity of the Ϫ3300/ϩ108 construct was 65-fold that of the parental, promoterless pXP2 control vector (Fig. 3). Promoter activity was progressively decreased when the full-length promoter was deleted from the 5Ј end at position Ϫ471 (32-fold over pXP2 control vector), Ϫ292 (38-fold), Ϫ265 (28-fold), and Ϫ243 (29fold). Surprisingly, the activity of Ϫ222 bp plasmid was only slightly reduced compared with the full-length promoter. An additional 100-bp deletion to Ϫposition 122 abolished half of the full-length promoter activity. Truncation of the promoter to Ϫ28 bp (containing the TATA box) markedly diminished the constitutive promoter expression to a level only 10-fold greater than the promoterless control plasmid pXP2. These data suggest that the region between nucleotide positions Ϫ222 and Ϫ28 is most critical for constitutive promoter activity in the EBV ϩ BCL.
Functional Role of the NFB and Ets Sites-Since the Ϫ222/ ϩ108 promoter region was shown to exhibit significant promoter activity, we analyzed the activity of this construct in further detail. To evaluate the functional role of the putative NFB half-site at Ϫ117 and the Ets site at Ϫ212, we prepared a series of additional promoters with defined mutations. The Ϫ222NFBm construct has a substitution of the core Ϫ109 CC Ϫ108 with Ϫ109 GG Ϫ108 . The Ϫ222ets-2⌬ construct has a 5-bp deletion that eliminates the Ϫ211 TTTCC Ϫ207 sequence (13). The same mutations were made in the context of the Ϫ3300/ϩ108 full-length promoter and were designated respectively, Ϫ3300NFBm and Ϫ3300ets-2⌬. Analysis of the activity of these constructs in transiently transfected EBV ϩ BCL (RPMI-8866) (Fig. 4A) revealed a dramatic decrease in the luciferase activity of promoter constructs mutated at the NFB site (Ϫ3300NFBm and Ϫ222NFBm) (80 -88% less than their corresponding wild-type constructs), but only a 50% decrease for the Ets-2 deletion mutants. In contrast, analysis of these constructs in the macrophage cell line RAW264.7 (Fig. 4B) demonstrated that the mutations of both the NFB and Ets sites almost completely abolished the inducibility of the p40 promoter by IFN-␥ and LPS (reduced to 12-20% of the Ϫ3300/ ϩ108 wild-type promoter activity), the latter result being consistent with our previous finding that the Ets site plays an important role in this system (13). Together, these results suggest that the B cells may use the same cis elements and transcription factors as do macrophages, but the relative contribution of each factor may vary. Characterization of the DNA-Protein Complexes Associated with the NFB Site-To establish the relationship between promoter activity and specific DNA-protein interactions at the NFB site (Ϫ117/Ϫ107) in EBV ϩ BCL, EMSA were performed using a 37-mer oligonucleotide probe (Ϫ135/Ϫ99) encompassing the putative NFB and PU.1 sites. Three DNA-protein complexes were observed via EMSA (Fig. 5A); two of them indicate specific binding to nuclear proteins since they were competed by an excess of the unlabeled oligonucleotide corresponding to the probe Ϫ135/Ϫ99, but not by an unrelated (NS) oligonucleotide competitor. Competition of the slower mobility complex was observed using a 25-mer oligo (Ϫ123/Ϫ99) containing the NFB site, but not the putative upstream PU.1 site located at Ϫ127 of the promoter. Use of an oligonucleotide with recognized binding activity for members of the NFB/Rel family (22) as competitor resulted in complete disappearance of the top band, indicating that this DNA-protein complex likely involves members of the NFB/Rel family. Further characterization of the NFB protein members bound to the Ϫ135 to Ϫ99 promoter region (Fig. 5B) revealed a slowly migrating supershifted complex in the presence of NFB anti-p50 and anti-c-Rel antibodies. The appearance of this supershifted band was accompanied by a partial disappearance of the upper complex. Thus, p50 and c-Rel-containing complexes constitutively bind to a transcriptionally active 6B site in the IL-12 p40 gene in EBV ϩ BCL. Anti-PU.1 antibody decreased the intensity of the faster-migrating band, suggesting the presence of PU.1 in this complex. No supershifted bands were detected with anti-c-Fos used as a control.
We then examined the ability of nuclear extract from IFN-␥and LPS-treated RAW264.7 macrophages to bind the probe Ϫ135/Ϫ99. In unstimulated cells, a series of three rapidly migrating complexes were observed (group III, Fig. 6A). However, only the slowest migrating complex of the three was consistently seen, suggesting that the faster migrating complexes may be proteolytic products of the upper complex. IFN-␥ treatment did not yield any additional binding, in contrast to an earlier report that IFN-␥ treatment of the murine macrophage cell line J774 resulted in augmented NFB binding to this region of the murine p40 promoter (15). LPS, however, induced two major complexes (I and II). Pretreatment with IFN-␥ did not affect the binding activity of these two complexes. Competitive EMSA to assess the specificity of the three complexes showed that each of the three groups of complexes was abolished specifically by an excess of nonradiolabeled oligomer Ϫ135/Ϫ99, but not by an unrelated oligonucleotide (NS). The smaller oligonucleotide Ϫ123/Ϫ99 and an NFB consensus oligonucleotide completely blocked the formation of complexes I and II (Fig. 6B), but had little effect on complex III. An anti-p50 antibody inhibited the formation of complexes I and II (Fig.  6C). Neither two distinct antibodies against the amino terminus (N) or the carboxyl terminus (C) of RelA (p65), nor an anti-c-Rel antibody affected the formation of the doublet band, although the combination of an anti-RelA and anti-c-Rel partially reduced the formation of complex I. Therefore, complex I appears to include mixed heterodimers p50/c-Rel and p50/RelA (p65) as well as c-Rel/RelA, while complex II is composed predominantly of p50. The faster-migrating complex (complex III) was decreased following IFN-␥ and LPS stimulation (Fig. 6A) and was competed by an anti-PU.1 antibody (Fig. 6C), thus confirming the presence of PU.1 in complex III.

Differences in Ets Site Binding Activities between B and Macrophage Cell
Lines-We previously demonstrated the functional predominance of the Ets site at Ϫ212/Ϫ207 in macrophage cells and the characteristics of the nuclear complexes formed around this site in response to stimulation (13,14). To extend this analysis to EBV ϩ BCL with respect to differences in the relative importance of the Ets site between EBV ϩ and macrophage cells, we performed EMSAs using the same probe (spanning Ϫ292 to Ϫ196) as in our previous studies (14). Nuclear extract from IFN-␥-and LPS-stimulated macrophage (RAW264.7) cells formed the previously characterized complex F1, whereas extract from EBV ϩ BCL (RPMI-8866) formed a slightly faster migrating complex, designated F1a. F1a was efficiently competed by a 50-fold molar excess of the Ϫ292/ Ϫ196 DNA (Fig. 7A). Furthermore, F1a was supershifted by an anti-Ets-2 and an anti-c-Rel affinity-purified polyclonal antibody (Fig. 7B), as observed previously for the F1 complex (14). However, formation of the F1a complex was blocked by an anti-IRF-2 antibody and not, as in the case of the F1 complex, by an anti-IRF-1. These data suggest that Ets-2, c-Rel, and IRF-2, but not IRF-1, may be constituents of the F1a complex in EBV ϩ BCL.
Transactivation of the IL-12 p40 Promoter by NFB and Ets-2-To assess the ability of several NFB proteins and the Ets-2 protein to activate the p40 promoter, we transfected EBV ϩ BCL with the luciferase-based p40 reporter construct (Ϫ3300/ϩ108) together with an expression vector for either NFB RelA (p65), NFB c-Rel, NFB p50, or Ets-2. No transactivation of the Ϫ3300/ϩ108 construct was detected by c-Rel, RelA, and p50, alone or any combination. Ets-2 alone was able to activate the promoter 4-fold, similarly to its actions in macrophages (13). Interestingly, co-expression of Ets-2 and c-Rel, but not RelA or p50, synergistically enhanced the p40 promoter activity (Fig. 8A). The decreased activity seen in Ets-2 and p50 cotransfection may have been due to the formation of the p50/p50 homodimer, which acts as a repressor, and may be a default method of regulation, as we have observed high levels of endogenous p50 produced by EBV ϩ BCL (data not shown). Transactivation by Ets-2 and c-Rel was dependent upon the integrity of their respective sites, since a 5-base deletion of the Ets site (Ϫ3300ets2⌬) or a CC 3 GG substitution in the NFB core sequence (Ϫ3300NFBm) in the context of the 3.3-kb promoter resulted in the failure of the promoter to respond to Ets-2 and c-Rel (Fig. 8B). In addition, mutation at the NFB site abolished the transactivating effect of exogenous Ets-2 (Fig. 8B). This effect, seen with the wild type construct, may have resulted from a moderate level of synergy between exogenous Ets-2 and constitutively present endogenous NFB in EBV ϩ BCL.
Cell Type and Promoter Specificity of the Synergistic Activation by Ets-2 and c-Rel-To determine whether the same combination of expression vectors for the NFB and Ets family could transactivate the p40 promoter in uninduced macrophage cells and in Jurkat T cells (which do not produce IL-12), c-Rel and Ets-2 expression vectors were transiently cotransfected in these cells. In both systems, the expression vectors strongly induced the activity of the full-length p40 promoter reporter construct (Ϫ3300/ϩ108) (Fig. 9). Mutations of the NFB or the Ets site dramatically decreased the level of transactivation by the effector constructs (Fig. 9, A and C). The remaining activity observed in macrophage cells using the Ϫ3300NFBm reporter construct, and in T cells using the Ϫ3300ets2⌬, might reflect FIG. 8. Rel proteins synergize with Ets-2 in mediating activation of the p40 promoter in EBV ؉ BCL. A, EBV ϩ BCL, and RPMI-8866 were transiently cotransfected with different combinations of RelA (p65), c-Rel, and Ets-2 and the luciferase-based promoter (Ϫ3300/ ϩ108). B, cotransfection was performed with the Ets-2 and c-Rel expression vectors and either the wild-type or the mutated Ϫ3300/ϩ108 constructs (Ϫ3300NFBm and Ϫ3300ets2⌬). After 24 h, cells were assayed for luciferase expression. Luciferase counts were normalized using ␤-galactosidase activity determined in the same cell extract (normalized luciferase counts). The empty expression vectors for all transcription factors were tested and produced background levels of luciferase activity (data not shown). Results are mean (Ϯ S.E.) from three independent experiments. the presence of additional noncanonical NFB or Ets-2 sites, or a cooperativity with other endogenous transcription factors. Thus, the synergy between co-transfected NFB and Ets-2 transcription factors in IL-12 p40 gene regulation is observed regardless of whether the endogenous gene is expressed in different cell types. This synergy is not observed in macrophage cells transfected with a luciferase construct under the control of the TNF-␣ promoter (Fig. 9B), also known to contain functional NFB and Ets sites (23)(24)(25). In addition, overexpression of the two transcription factors failed to transactivate the thymidine kinase minimal promoter, lacking NFB and Ets sites (data not shown). Taken together, these results demonstrate that the functional synergy between NFB and Ets-2 is promoter-specific. DISCUSSION The present study was performed to delineate the molecular mechanisms by which EBV transformation of B lymphocytes results in the activation of IL-12 gene expression. We demonstrate that two regulatory elements, Ets at Ϫ212/Ϫ207 and NFB at Ϫ117/Ϫ107, play a role in the high level expression of the human IL-12 p40 promoter in EBV ϩ BCL and in its inducible expression in macrophages. In EBV ϩ BCL, activity of a 3300-bp p40 promoter parallels transcription of the endogenous p40 gene, whereas no promoter activity was detectable in EBV Ϫ BCL or in Jurkat T cells in which the endogenous p40 gene is inactive. In EBV ϩ BCL, a segment of the p40 promoter containing 222 bp upstream and 108 bp downstream of the transcription start site (Ϫ222/ϩ108) retains most of the activity of the full-length promoter, indicating that constitutive transcription from the p40 promoter is regulated by transcription factors acting on this region. This basal promoter activity is dependent mostly on the integrity of the NFB site (Ϫ117/ Ϫ107), as demonstrated by mutation analysis. A DNA probe encompassing this site binds the p50 and c-Rel NFB family members, consistent with previous reports that NFB proteins in mature B cell lines are constitutively present in the nucleus (26 -28). Similar experiments on activated macrophages revealed that the NFB site, bound by p50/RelA, p50/c-Rel, and RelA/c-Rel heterodimers, is also critical for p40 promoter activity. When this site is mutated, macrophage cells completely lose their ability to activate the p40 promoter after IFN-␥ and LPS stimulation. The data provided here are corroborated by the studies of Murphy et al. (15) on the murine p40 promoter in activated macrophages. Our studies indicate that NFB transcription factors are essential for constitutive IL-12 p40 gene expression in EBV ϩ BCL and for p40 gene inducibility in macrophage cells. However, in both B and macrophage cells, p50, c-Rel, and RelA individually are unable to transactivate the IL-12 p40 promoter, suggesting the requirement for additional signals. Ets-2 appeared to be a likely candidate as the IL-12 p40 promoter contains a functional Ets motif, TTTCCT or AGGAAA (Ϫ212 to Ϫ207) essential for promoter activity in stimulated macrophage cells (13). Complete activation of the IL-12 p40 promoter in Ets-2 transiently transfected macrophages has been shown to require IFN-␥ or LPS stimulation (13), suggesting that, although vital, Ets-2 alone is not sufficient to achieve optimal p40 promoter activation. Here, we clearly demonstrate, in both B and unstimulated macrophage cells, that Ets-2 exerts a strong enhancing effect on p40 promoter activity only in the presence of the NFB c-Rel. Ets-2 and c-Rel are sufficient for potent transactivation of the p40 promoter even in T and EBV Ϫ BCL (data not shown), where the p40 promoter is normally inactive. The specificity of this response is demonstrated by mutation or deletion of the NFB or Ets binding sites resulting in greatly diminished promoter activation. Cooperativity between these transcription factors seems to be specific for the IL-12 p40 promoter, since no synergistic activation was observed for the TNF-␣ promoter, also known to contain NFB and Ets sites (23)(24)(25).
Physical interactions between NFB and Ets-like protein have been reported to regulate expression of the IL-2 receptor ␣ gene (29) and the activation of the human immunodeficiency virus enhancer (30). The Ets domain was shown to be necessary and sufficient to mediate this interaction, suggesting that the highly conserved Ets domain may act broadly to facilitate multiple protein-protein interactions. Indeed, it is now well established that Ets family proteins can activate transcription in conjunction with other proteins (31)(32)(33)(34)(35). Here, we have shown an association between Ets-2 and c-Rel in the F1a complex in EBV ϩ BCL similar to the F1 complex of activated macrophages (14). We are currently investigating whether the F1a and NFB complexes physically interact at the p40 promoter.
FIG. 9. C-Rel and Ets-2 synergize in macrophage and T cell line. A, macrophage cells, RAW264.7, were transiently transfected with the p40 promoter construct (Ϫ3300/ϩ108) or its mutant forms (Ϫ3300NFBm and Ϫ3300ets2⌬), together with c-Rel and/or Ets-2 expression vectors, without further stimulation. B, transient cotransfections were performed as in A, except that the reporter used was under the control of the TNF-␣ promoter. C, transient cotransfections were performed in T cell line, Jurkat, with the IL-12 p40 promoter constructs (Ϫ3300/ ϩ108). After 24 h, cells were assayed for luciferase expression. Luciferase counts were normalized using ␤-galactosidase activity determined in the same cell extract (normalized luciferase counts). The empty expression vectors for all transcription factors were tested and produced background levels of luciferase activity (data not shown). Results are mean (Ϯ S.E.) from three independent experiments.
The composition of the F1a complex differs from the previously characterized F1 of activated macrophages (14), in that the former contains the IRF-2 protein but not IRF-1. IRF-1deficient mice fail to produce IL-12, resulting in a severely compromised Th1 immune response in vivo and in vitro (36,37) strongly implicating a role for IRF-1 in IL-12 activation. IRF-1 and IRF-2 are highly homologous and bind to common DNA sequence elements (38,39) with similar affinities. As a consequence of mutating the Ets site, we demonstrated here, in EBV ϩ BCL, only a partial reduction in the constitutive p40 promoter activity. In contrast, mutation of the same site in macrophages resulted in total abrogation of inducible p40 promoter activity. In other promoters, such as that of IFN-␤, it has been shown that IRF-2 antagonizes the activating function of IRF-1 (40). Therefore, the possibility that IRF-2 binding to the Ets site in EBV ϩ BCL may function as a repressor of IRF-1 activity by preventing its binding; thus, decreasing the relative participation of this site in the constitutive activation of the p40 promoter in EBV ϩ BCL should be considered. However, the ability of IRF-2 to act as a suppressor is by no means absolute, as both IRF-1 and IRF-2 have been shown to up-regulate expression of two genes, the histone H4 gene FO108 and the EBNA 1 gene (41,42).
We provided evidence of a pivotal role of the NFB site in constitutive IL-12 p40 gene expression in EBV ϩ BCL and a functional cooperativity between NFB and Ets-2 transcription factors, critical for the constitutive and inducible activation of the IL-12 p40 gene. Parallel to our findings, it was recently demonstrated that a synergy exists between C/EBP and NFB in human and mouse p40 promoter regulation (16), but whether C/EBP plays a role in constitutive p40 gene expression in EBV ϩ BCL remains to be determined. Together, this information suggests that the expression of IL-12 p40 gene requires the coordinated interactions between multiple regulatory proteins. The B cell lines used in our studies expressed the EBV latent-infection membrane protein 1 (LMP1) capable of activating the NFB transcription factors (43)(44)(45)(46). However, transient or stable transfection of LMP1 in EBV Ϫ B cells was insufficient to transactivate the p40 promoter. 3 This suggests that it is likely that EBV employs more elaborate mechanisms stimulating multiple transcription factors critical for the activation of the IL-12 p40 gene.