Cellular and Viral Trans-Acting Factors Modulate N-Myc2 Promoter Activity in Woodchuck Liver Tumors

SHORT REPORT Cellular and viral trans-acting factors modulate N-myc2 promoter activity in woodchuck liver tumors

Marc Flajolet1, Anne Gegonne2, Jacques Ghysdael2, Pierre Tiollais1, Marie-Annick Buendia1 and GenevieÁ ve Fourel1,3

1UniteÂ de Recombinaison et Expression GeÂneÂtique (INSERM U163), Institut Pasteur, Paris and 2CNRS-URA 1443, Institut Curie, Centre Universitaire, BaÃtiment 110, Orsay, France

Activation of the N-myc2 oncogene by integration of of the N-myc gene. Like other myc oncogenes, N-myc2 woodchuck hepatitis virus (WHV) DNA is a central cooperates with activated H-ras in the transformation event in woodchuck liver oncogenesis. In this study, we of rat embryo ®broblasts (Fourel et al., 1990), and its have evaluated the in¯uence of several cellular and viral targeted expression in liver cells induces hepatocellular trans-acting factors and mediators of in¯ammation on N- carcinoma in transgenic mice (CA Renard and G myc2 promoter activity in hepatoma cell lines. Ets Fourel, unpublished results). In addition, forced oncoproteins, including Ets1, Ets2 and PEA3 eciently expression of N-myc2 in rodent hepatic cells triggers activated a chimeric N-myc2 promoter/luciferase repor- a strong proliferative response and an increased ter gene. By electrophoretic mobility shift assays, we propensity to undergo apoptosis, which is inhibited show that Ets1 and Ets2 proteins can eciently bind two by insulin-like growth factor II (Dandri et al., 1996; consensus Ets sites located within a 59 bp sequence Ueda and Ganem, 1996). While N-myc2 transcripts are upstream of the N-myc2 transcription start site. Site- undetectable in normal adult hepatocytes, they are directed mutagenesis of these Ets-binding motifs abol- abundantly produced in a majority of liver tumors ished transactivation of the N-myc2 promoter by Ets induced by woodchuck hepatitis virus (WHV) (Fourel proteins. Addition of interleukin-6 (IL-6) induced a weak et al., 1990). A prominent role of WHV DNA but reproducible activation of the N-myc2 promoter, integration in the process of N-myc2 activation has while IL-1 was ineective. We further show that the N- been demonstrated. In most woodchuck tumors, viral myc2 promoter can be transactivated by the hepadna- integration sites were found to be clustered either in the virus X protein, and that distal promoter sequences are vicinity of the N-myc2 coding region or at a 160 ± required for both IL-6 and X responsiveness. Similar 200 kb distance in the win locus (Fourel et al., 1990, eects of these factors were observed in the context of 1994). N-myc2 promoter activity was highly enhanced the N-myc2 promoter activated by WHV cis-regulatory in all cases, and the capacity of WHV enhancer elements. In view of the high-level expression of the N- elements to cis-activate the N-myc2 promoter was myc2 oncogene in most woodchuck liver tumors, the Ets demonstrated in transient transfection assays in oncoproteins, in¯ammation-associated cytokine IL-6 and hepatoma-derived cells (Fourel et al., 1996; Ueda et the viral X transactivator might play important roles in al., 1996a,b; Wei et al., 1992). At preneoplastic stages, hepadnavirus-associated tumorigenesis. N-myc2 expression is activated in precancerous lesions similar to altered hepatic foci occurring in rats exposed Keywords: N-myc2; woodchucks; IL-6 to carcinogenic regimens (Toshkov et al., 1990; Yang et al., 1993; reviewed in Fourel, 1994). At this stage however, it was not established whether N-myc2 transcription was activated through viral insertional The myc genes family contains three well-characterized mutagenesis or through independent, trans-acting proto-oncogenes c-myc,N-myc and L-myc, which share pathways. a common three exon structure and encode structurally In a previous report, a short N-myc2 promoter and functionally related proteins (reviewed in DePinho including a variant TATA box and potential binding et al., 1991). These genes play important roles in the sites for several transcription factors was mapped regulation of cellular proliferation, dierentiation and within sequences homologous to the 5' untranslated apoptosis. Oncogenic activation of myc genes by region of the second N-myc exon (Fourel et al., 1992). chromosomal translocation, gene ampli®cation, or In this study, we set out to investigate cellular and viral retroviral insertion is widespread in human and transcriptional trans-acting factors controlling the animal cancers (Marcu et al., 1992). activity of the woodchuck N-myc2 promoter in We recently described the presence of an additional hepatoma cells. gene of the myc family, termed N-myc2, in a subset of rodent species including woodchucks and squirrels Binding of Ets-1 and Ets-2 proteins at two dierent sites (Fourel et al., 1990; Transy et al., 1992). The in the minimal N-myc2 promoter intronless N-myc2 gene is a functional retrotransposon The ®nding of two sequences matching the consensus Ets-binding site (Wasylyk et al., 1993) in the N-myc2 Correspondence: M Flajolet promoter, as illustrated in Figure 1, led us to 3Present address: Ecole Normale SupeÂ rieure de Lyon, Lyon, France investigate whether Ets proteins can bind and activate Received 2 October 1996; revised 7 May 1997; accepted 7 May 1997 the N-myc2 promoter. Electrophoretic mobility shift Transactivation of the N-myc2promoter M Flajolet et al 1104

Figure 1 Nucleotide sequence of the N-myc2 promoter. Numbering starts at the major transcription initiation site, indicated by an arrow. Here is shown the 406 bp RsaI±BglII fragment (positions 7332/+74). The translation initiation site (position +49) is shown in bold character. The 5' ends of deletion mutants analysed in this study are indicated by dots and positions. Two sequences homologous to the Acute Phase Response Element (APRE) (Poli and Ciliberto, 1994), which is responsible for IL-6-responsiveness of a subset of IL-6-sensitive promoters, and putative Ets-binding sites (Ets-Pu1 and Ets-Pu2) are boxed. CAAT, Sp1, TATA and initiator (Inr) motifs are overlined. The boundaries of the N-myc2 domains homologous to N-myc exons 1 and 2 are indicated

assays (EMSA) were performed using whole cellular with cell extracts containing the Ets1 protein. As extracts from insect cells infected with Ac-NPV-Ets1 or shown in Figure 2, lane 14, the N-myc2.Pu1/Ets1 Ac-NPV-Ets2 recombinant baculoviruses as described interaction was detected as a major band with expected (Bosselut et al., 1990). Synthetic oligonucleotides used mobility in gel retardation assay. The unlabeled wild in this study were as follows: type N-myc2.Pu1 (lanes 15 and 16), N-myc2.Pu2 and

E74 oligonucleotides (not shown) competed for Ets1

U N-myc2.Pul/U61 to 37: protein binding, while the mutated oligo (lanes 17 and TCGGACCCTAGGGAAGGAAGCACCCCCC 18) was ineective. The abilities of N-myc2.Pu1 and N- TGGGATCCCTTCCTTCGTGGGGGGAGCC; myc2.Pu2 to bind Ets2 protein were similarly demon-

strated in gel retardation assays (results not shown).

U N-myc2.Pu2/U27 to 3: TCGGAAGAGCTGAGCGGAAAGAAGCTCT TTCTCGACTCGCCTTTCTTCGAGAAGCC; The N-myc2 promoter is eciently activated by Ets expression E74:TCGGGCTCGAGATAAACAGGAAGTGGTC Chimeric N-myc2/luciferase constructs were generated CGAGCTCTATTTGTCCTTCACCAGAGCC by inserting dierent fragments of the N-myc2 promoter in pSVOALD5', a plasmid carrying the These double-stranded oligomers and corresponding promoterless luciferase (LUC) gene and SV40 small-t mutated oligonucleotides (N-myc2.Pu1mut: AAGGA- intron/polyadenylation signal (De Wet et al., 1987). AGC?AACCAAGC and N-myc2.Pu2mut: GCGG- The full length (71100) N-myc2 promoter consists of a AAA?GCCCAAA) were labeled by ®ll-in with large 1174 bp HindIII ± BglII fragment, and the 7332 fragment of DNA polymerase I (Klenow). The E74 construct contains a 406 bp RsaI±BglII DNA frag- oligonucleotide corresponds to a strong binding motif ment as shown in Figure 1. The 759 and 737 of the drosophila ecdysone-inducible E74A protein, constructs were obtained after PCR ampli®cation of shown to be recognized by other ETS-domain proteins 133 bp and 111 bp N-myc2 sequences spanning the (Urness and Thummel, 1990). transcription initiation site. The DNA sequence of the The capacity of each of the two consensus Ets- fragments generated by PCR was con®rmed by binding motifs from N-myc2 promoter to compete for nucleotide sequencing. In all constructs, N-myc2 binding of Ets1 to the high-anity Ets-binding site E74 sequences terminated at the BglII site (position +74) was ®rst investigated. Incubation of the 32P-labeled E74 and included the ®rst 27 bp of the N-myc2 open probe with the Ets1-containing extracts yielded a major reading frame (ORF) fused in frame with the LUC DNA-protein complex (Figure 2, lanes 2 and 8) that ORF. did not appear using extracts from insect cells infected Hepatoma-derived Huh7 cells were transfected with with the empty vector (not shown). In competition expression plasmids directing the expression of either experiments, in which a 100- to 400-fold excess of the Ets1, Ets2 or PEA3 together with reporter constructs unlabeled N-myc2.Pu1 or N-myc2.Pu2 oligonucleotides carrying the LUC gene under control of the N-myc2 was included with the labeled E74 probe, both N-myc2 promoter. The pEts1 and pEts2 plasmids carrying Ets1 oligos competed eectively for the DNA-protein and Ets2 cDNAs linked to the SV40 early promoter complex (lanes 3, 4, 9 and 10); the mutated oligos have been described previously (Prosser et al., 1992), were ineective (lanes 5, 6, 11 and 12). Similarly, when and the plasmid carrying a PEA3 cDNA controlled by E74 was used as a probe, unlabeled N-myc2.Pu2 the SV40 early promoter was a gift from J Hassell. (Figure 2, lanes 21 and 22) as well as N-myc2.Pu1 Cells were transfected by the calcium phosphate oligonucleotides (data not shown) competed eectively coprecipitation method, using 15 mg of plasmid DNA for Ets2 protein binding, while a mutated oligo (lanes per 25 cm2 dishes and 2 mg of the b-galactosidase 23 and 24) had no eect. Direct assays were conducted expression vector pCH110 as an internal control for to determine whether the N-myc2.Pu1 and N-myc2.Pu2 transfection eciency. Luciferase activity was deter- oligonucleotides could also form speci®c complexes mined from cell lysates prepared 24 h or 48 h after Transactivation of the N-myc2promoter M Flajolet et al 1105

Proteins Ets1 Ets2

Probes E74 Nmyc2.Pul E74

Competitor Nmyc2.Pul Nmyc2.Pu2 Nmyc2.Pul Nmyc2.Pu2 wt mut wt mut wt mut wt mut 400 400 100 400 400 400 400 100 400 400 100 100 100 100 100 100 probe none none probe none probe none probe

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Figure 2 Analysis of Ets1 and Ets2 binding to two consensus purine boxes in the minimal N-myc2 promoter. Electrophoretic mobility shift assays were performed using extracts (10 mg) from insect cells infected with either Ac-NPV-Ets1 or Ac-NPV-Ets2 recombinant baculoviruses, and oligonucleotides E74 or N-myc2.Pu1 as 32P-labeled probes (100 fmol/assay) as indicated on the top. Assays contained either no competitor (lanes 2, 8, 14 and 20) or a 100- to 400-fold molar excess of the unlabeled oligonucleotides N- myc2.Pu1 or N-myc2.Pu2, as indicated above each lane. Wild type (wt) and mutated (mut) oligonucleotide sequences are described in the text. As controls, the labeled probes were analysed in the absence of lysate (lanes 1, 7, 13 and 19) or with lysates from insect cells infected with an empty baculovirus vector (Bosselut et al., 1990) (data not shown). The bound complexes were resolved in 5% polyacrylamide gels and visualized using a PhosphorImager (Molecular Dynamics) transfection as described previously (Wei et al., 1992). two dierent motifs within the minimal N-myc2 Overexpression of Ets1 or ETs2 resulted in strong promoter and eciently activate transcription driven activation of the full length N-myc2 promoter (Figure by this promoter in dierentiated hepatoma cell lines. 3, lanes 2 and 3), and the minimal (759) N-myc2 Residual Ets-transactivation upon mutation of both of promoter was similarly activated (Figure 3, lines 5, 6 and these sites may be due to the existence of unidenti®ed, 7). In contrast, transactivation was appreciably impaired degenerated sequences in the same region; alternatively, by deletion of sequences 759 nt to 737 nt that contain it may involve Ets proteins only as indirect eectors. Ets the 5' most purine box (Pu1) (lanes 10 and 11). family members are known to require cooperation with Mutations were introduced into the N-myc2Pu1 site other factors for their activity (Schneikert et al., 1992; (ACCAAG) and Pu2 site (CCCAAA) by using the Wasylyk et al., 1993). Notably, as in several cellular and enzymatic inverse PCR method (Stemmer and Morris, viral promoter/enhancer regions, N-myc2 promoter 1992). The response elicited by the Ets expression vectors sequences harbor a canonical Sp1 binding site located was signi®cantly decreased by mutating the Pu1 motif upstream of the 5' Purine box (see Figure 1). Additional and, to a lesser extent, the Pu2 motif (Figure 3, lines 14 ± studies will be required to determine whether synergistic 15 and 18 ± 19), and combination of Pu1 and Pu2 interactions between Ets and Sp1 contribute in N-myc2 mutations essentially abrogated activation of the promoter activation. On the other hand, an hepatocyte- minimal N-myc2 promoter by Ets-proteins (lanes 22 ± like cellular environment appeared to be necessary for N- 23). In parallel studies, the Herpes simplex virus myc2 responsiveness to Ets proteins, which may convey thymidine kinase (TK) promoter used as a negative the requirement for a liver-speci®c co-factor. control was found totally unresponsive (Figure 3, lanes 25 ± 28). Correct transcription initiation at the normal N- Interleukin IL-6 can increase N-myc2 promoter activity myc2 mRNA start site was con®rmed by RNase protection assays (data not shown). In cotransfections WHV infection induces a chronic in¯ammatory process with the PEA3 expression plasmid, activation of the N- in the liver of long term carrier woodchucks (Popper et myc2 promoter was also observed, although the levels al., 1987), and altered hepatic foci that overexpress the were generally twofold lower than with Ets1 or Ets2 N-myc2 gene were frequently detected in the vicinity of (Figure 3, lanes 4, 8, 12, 16 and 20). In contrast, pD318 ± portal tract in¯ammatory cell lesions, which are 363, which carries a Ets1 mutant harboring a deletion in capable of producing a variety of cytokines (Yang et the DNA-binding domain (Boulukos et al., 1989) had no al., 1993). Interleukin-6 (IL-6) is a multifunctional trans-acting activity (data not shown). Finally, transac- cytokine, which causes cell growth and dierentiation tivation could also be observed at similar levels in and induces the synthesis of acute phase proteins in HepG2 cells but not in CCL13 cells, an epithelial-like cell hepatocytes (Heinrich et al., 1990). Two families of line derived from human liver, suggesting the involve- transcription factors mediate IL-6 eects on hepatic ment of unidenti®ed, cell-speci®c cofactor(s). These data gene transcription. While C/EBP and NF-kB transcrip- taken together demonstrate that Ets oncoproteins bind tion factors are involved in activation of class I acute- Transactivation of the N-myc2promoter M Flajolet et al 1106

Figure 3 Transactivation of the N-myc2 promoter by Ets proteins. Left part: Schematic representation of the constructs bearing wild type (open box) or mutated (black box) N-myc2 promoter sequences linked to the LUC reporter gene. Cellular sequences ¯anking the 5' side of the N-myc2 retrotransposon are represented as a line, and regions of homology with N-myc exons 1 and 2 as boxes. The N-myc2 ORF (shaded) is fused in frame to the LUC coding region at position +74 (see Figure 1). Important motifs in N-myc2 minimal promoter sequences are shown besides each construct. Wild type or mutated Purine boxes 1 and 2 are indicated by an open or a black box, and the variant TATA motif (T) is shown as an open circle. TK: Herpes simplex virus thymidine kinase promoter. Right part: Plasmids were transfected into Huh7 cells or cotransfected with a plasmid expressing Ets1, Ets2 or PEA3 protein, and LUC activity was assayed 24 h after transfection. Two independent clones of each construct were analysed, and comparisons were carried out on identical plates prepared at the same time. The bar diagram depicts the mean (+standard error of the mean) relative LUC activities obtained in three independent transfection experiments. The activity of the full-length (71100) N- myc2 promoter construct was arbitrary set at one, and numbers indicate fold-activation

phase protein genes, for which maximal expression is data suggest that the two potential APRF sites located reached by a combination of IL-1 and IL-6, the acute- at positions 7108 to 7116 and 7281 to 7289 are phase response factor APRF/ IL6-REBP/ Stat3 either nonfunctional or not sucient on their own for mediates IL-6 responsiveness of class II genes, which IL-6 induction, and that the contribution of distal N- are responsive only to IL-6 and related cytokines. myc2 promoter sequences is required for IL-6 Examination of N-myc2 promoter sequences (see responsiveness. However, examination of the nucleo- Figure 1) revealed the presence of two putative IL-6 tide sequence of the full-length N-myc2 promoter responsive elements (Poli and Ciliberto, 1994), similar (unpublished data) did not provide evidence for the to those of type II acute-phase protein genes such as presence of additional consensus IL-6-responsive the b-®brinogen and a2 macroglobulin genes (Hocke et elements. An alternative hypothesis is that the APRF al., 1992). To investigate the responsiveness of the N- sites might be involved in combination with the myc2 promoter to IL-6, hepatoma-derived HepG2 cells upstream sequences, as described for IL-6-mediated known to express the high anity membrane receptor induction of the junB promoter (Nakajima et al., to IL-6 (a kind gift of G Courtois) were transfected 1993). In the same experimental conditions, the with dierent N-myc2 promoter/LUC reporter con- in¯ammation cytokine IL-1 had no eect on the structs, and recombinant human IL-6 (Boehringer activity of the N-myc2 and b-®brinogen promoters Mannheim) was added to the culture medium 24 h (Figure 4, lanes 3, 6 and 9), supporting the notion that after transfection at 60 U/ml. Cells were harvested 3 N-myc2, like the b-®brinogen gene, belongs to the days later. In the presence of IL-6, the N-myc2 group of hepatic genes that respond to IL-6 but not to promoter activity was reproducibly enhanced by 2.5- IL-1. fold (Figure 4, lane 2). This value is comparable to that obtained with pb-®b, a plasmid constructed by The hepadnaviral X transactivator activates the N-myc2 inserting the promoter of the IL-6-responsive human promoter b-®brinogen gene (Dalmon et al., 1993) at the HindIII site of pSVOALD5' (lane 8). In contrast, no activation The hepadnaviral X protein was shown to transactivate was observed upon deletion of N-myc2 promoter several genes implicated in cellular growth, such as c- sequences upstream of position 7332 (lane 5). These myc,c-fos and c-jun (reviewed in Rossner, 1992). This Transactivation of the N-myc2promoter M Flajolet et al 1107 led us to address whether the X protein is able to Taking into account the previously reported activate the N-myc2 promoter. CCL13 cells were co- transactivation of c-myc by the X protein (Balsano et transfected with pHBx2S, an expression plasmid al., 1991) and that of the classical N-myc gene (our carrying the HBV X gene under control of the SV40 unpublished results), each potential target gene for early promoter/enhancer (KekuleÂ et al., 1993) and WHV DNA integration in woodchuck tumors also LUC reporter constructs as described above. Cotrans- emerges as a target of X transactivation. The fection of low amounts (30 to 250 ng) of pHBx2S mechanisms of X-induced transcriptional activation resulted in activated transcription from both the N- are not clear, although there is accumulating evidence myc2 and SV40 promoters, but transactivation was less for both a nuclear eect directly on the transcription ecient using a higher amount of pHBx2S (Figure 5, machinery, and a cytoplasmic eect involving signal lanes 1 ± 6 and 9 ± 14). This may be attributed either to transduction pathways (reviewed in Yen, 1996). In our a squelching eect, or to the recently described system, the N-myc2 promoter was non-responsive to X cytotoxic eects of HBx overexpression (Oguey et al., overexpression in serum-starved cells, which suggests 1996). Deleting N-myc2 promoter sequences upstream the involvement of serum-sensitive intracellular media- of nt-322 abrogated transactivation (lanes 7 and 8) tors. Because the N-myc2 promoter is eciently indicating that sequences responsible for this activation activated by Ets proteins, which are subject to rapid are contributed by distal N-myc2 promoter sequences. and transient phosphorylation events in response to The presence of X protein in lysates from transiently mitogenic signals (Yang et al., 1996), we initially transfected cells was demonstrated by Western blotting envisioned that Ets proteins might be the terminal (data not shown). Interestingly, the N-myc2 promoter eectors of X-induced trans-activation of the N-myc2 proved to be unresponsive to X protein when recipient promoter. However, the ®nding that the minimal N- cells were serum-starved for 3 days before transfection myc2 promoter is responsive to Ets proteins but not to (data not shown). X expression argues against this hypothesis. Conver-

Figure 4 Dierential response of the N-myc2 promoter to the in¯ammatory cytokines IL-1 and IL-6. HepG2 cells were transfected with the 71100 and 7332 N-myc2/LUC constructs, in the presence or absence of IL-1 or IL-6 (60 U/ml). Cytokines were added into the medium 24 h after transfection, and cells were harvested after a 72 h incubation. The b-®b/LUC plasmid carrying sequences of the IL-6 responsive human b-®brinogen promoter was analysed in parallel. Legend is otherwise the same as in Figure 3

Figure 5 Transactivation of the N-myc2 promoter by the hepadnaviral X protein. Constructs schematized in the left part were transfected into CCL13 cells together with increasing amounts of the plasmid pHBx2S. LUC activity was assayed 48 h after transfection. SV40: SV40 early promoter, known to be activated by the hepadnaviral X protein. Data are presented as the mean (+s.e.m.) values obtained from at least ®ve independent experiments Transactivation of the N-myc2promoter M Flajolet et al 1108 sely, the requirement for upstream sequences ¯anking were obtained upon expression of either Ets1 or PEA3 the N-myc homology region for X transactivation as proteins (data not shown). Mutation of both the Pu1 well as for IL-6 responsiveness may suggest that this and Pu2 boxes reduced transactivation by Ets to region contains previously unrecognized regulatory approximately fourfold in the presence of juxtaposed elements. viral sequences (Figure 6, lanes 9 ± 10), a value similar to that obtained with the mutated N-myc2 promoter alone. These data indicate that only Ets-binding sites Cis- and trans-activating pathways can act in concert to located in the N-myc2 promoter and not in the viral trigger high level activity of the N-myc2 promoter sequences mediated activation by Ets proteins. In more than 80% of woodchuck liver tumors, WHV Integrated viral sequences in woodchuck tumors DNA integration either in the N-myc2 locus or in the may consist in subgenomic WHV fragments or in large, physically linked win locus is associated with enhanced rearranged sequences made of concatemerized frag- N-myc2 promoter activity. We have previously shown ments. We recently cloned a 2.85 kb long, highly that this promoter is activated in cis by WHV enhancer rearranged viral sequence inserted into the win locus in sequences, mainly by the We2 activator (Fourel et al., a woodchuck liver tumor. This sequence (WHVint) 1996; Wei et al., 1992). We therefore investigated the contains one copy of the enhancer 1 and ®ve copies of eects of trans-acting factors as described above in the the enhancer 2 (Flajolet et al., 1997). Introduction of context of the N-myc2 promoter placed under cis- WHVint upstream of the minimal N-myc2 promoter regulatory in¯uence of the viral enhancer. To this aim, resulted in a 45-fold induction of LUC activity (Figure the 1647 bp ApaI±BglII DNA fragment (WHVenh) 6, compare lanes 5 and 11). The enhancer activity of encompassing the viral enhancers 1 and 2 on the the complex integrant sequences was therefore eight WHV8 genome (Girones et al., 1989) was inserted times higher than that of the linear WHVenh fragment upstream of either the full length or the minimal N- (compare lanes 7 and 11). N-myc2 promoter under the myc2 promoter in a LUC vector. The resulting marked in¯uence of this sequence remained highly constructs were transfected in Huh7 cells. Insertion of responsive to Ets2, reaching a remarkably high (450- the viral sequences upstream of the 71100 N-myc2 fold) level of induction (lane 12). These data strongly promoter resulted in fourfold induction of LUC suggest that trans-activation by Ets proteins and cis- activity (Figure 6, lanes 1 and 3). Coexpression of activation by integrated viral sequences can act in Ets2 led to a further 15-fold activation, resulting in a concert to activate N-myc2 promoter to very high total 60-fold activation as compared to the native levels in woodchuck liver tumors. promoter activity (lane 4). Thus, the eects of The growing family of Ets transcription factors has combined cis-activation by the WHV enhancer and been implicated in the regulation of gene expression in trans-activation by Ets were clearly multiplicative. Each a variety of biological processes, including growth of these activating pathways proved even more eective control, transformation, T-cell activation, and devel- in combination with the minimal N-myc2 promoter, opmental programs in many organisms (reviewed in resulting in a ®nal 250-fold activation of transcription Wasylyk et al., 1993). Although individual members initiation over basal activity (lane 8). Similar results are dierentially expressed, they show similar binding

Figure 6 Cooperative eects of cis- and trans-activating pathways on N-myc2 promoter activity. Left part: WHV fragments inserted 5' of the 71100 or 759 N-myc2 promoter in LUC constructs are shown as boxes. WHVenh is a linear ApaI- BglII fragment of the WHV8 genome (position 887 ± 2534) (Girones et al., 1989); WHVint represents a highly rearranged, 2.85 kb long viral sequence integrated into the win locus in a woodchuck liver tumor. The WHV enhancers 1 and 2 are shown as stripped and black boxes. N-myc2 promoter doubly mutated at the Pu1 and Pu2 motifs is shown as a black box. The constructs were cotransfected with a Ets2 expression vector in Huh7 cells. Results are presented as the ratio of LUC activity for each construct versus the 71100 N-myc2 promoter construct and are means (+s.e.m.) of at least three independent transfections. Numbers refer to the fold-activation of LUC activity in the presence of co-expressed Ets2 protein Transactivation of the N-myc2promoter M Flajolet et al 1109 speci®cities and the identi®cation of the Ets-family majority of woodchuck hepatocellular carcinomas is members implicated in N-myc2 activation in wood- systematically associated with WHV DNA integration chuck liver is a dicult task. Ets2 expression has been into the N-myc2orwin loci. Thus, cis-activation by shown to increase during liver regeneration (Bhat et al., integrated viral enhancer elements and trans-activation 1987), and we observed that both Ets1 and Ets2 by cellular and viral factors may act in concert to drive mRNAs are produced at signi®cant levels in tumorous high level expression of the N-myc2 oncogene. and nontumorous liver tissues from WHV-infected However, recent evidence points to the N-myc2 gene woodchucks (our unpublished data). In this study, being constitutively repressed within chromatin in both Ets-1 and Ets-2 proteins were found to bind the normal woodchuck cells, although N-myc2 sequences two purine boxes in the N-myc2 promoter, and are not methylated (our unpublished results). In overexpression of Ets1, Ets2 or PEA3 was almost woodchuck hepatomas, integration of WHV DNA in equally eective in activating the N-myc2 promoter, the N-myc2 and win loci could thus play a dual role, as the fold-induction being high compared with other recently described for other enhancers (Walters et al., documented systems (Bosselut et al., 1990; Xin et al., 1996). It would ®rst alleviate a negative control 1992). A variety of intracellular pathways elicited in imposed on the promoter and confer accessibility to virus-infected liver and in hepatomas may lead to the the transcription machinery, and secondly, it would modulation of Ets-proteins activity, which may in turn directly activate transcription initiation. An array of contribute to N-myc2 activation. This hypothesis is trans-acting partners could then come into play supported by the recent ®nding that the activity of Ets2 sequentially, to eventually potentiate the oncogenic is regulated by the Ras-Raf-MEK-MAP kinase cascade activation of N-myc2. As a consequence, the trans- (McCarthy et al., 1997). activation pathways described in this study would not The eect of IL-6 on the activity of the N-myc2 lead to N-myc2 activation unless derepression occurs. promoter linked to the viral fragment WHVenh was Further analysis of the hepatic foci that develop in investigated in HepG2 cells. In these cells, we observed chronic WHV infections might allow to determine an eightfold activation of the N-myc2 promoter by the whether integration of viral DNA in the N-myc2orwin cis-acting eect of the WHV enhancer. Exposure of loci is responsible for upregulation of N-myc2 at this HepG2 cells to IL-6 resulted in further twofold early step of the transformation process. Another activation of LUC expression (data not shown). recently identi®ed factor in the progression of Collectively, the data presented here show that IL-6 woodchuck liver tumors is the inhibition of N-myc2- induces a moderate increase of the activity of the N-myc2 induced hepatocyte apoptosis by the insulin-like promoter, both in native con®guration and when linked growth factor II (IGF-II) (Dandri et al., 1996; Ueda to the WHV enhancer. Thus, in woodchuck liver, IL-6 and Ganem, 1996). Indeed, altered hepatic foci in might contribute in activating N-myc2 expression both at chronically infected livers usually express both N-myc2 preneoplastic and tumorous stages. It should not be and IGF-II, but N-myc2 expression is still low as inferred from the mild eect observed in our experi- compared to malignant tumors. This suggests that N- mental conditions that IL-6 could play only a minor myc2 superactivation triggers secondary steps that will role in N-myc2 activation, as the eects might be much eventually lead to malignant transformation. These more pronounced in vivo in the context of in¯amed liver. events, that may include alleviation of lymphokine/ During the multistep process of WHV-associated growth factor dependency, remain to be elucidated. liver carcinogenesis, the woodchuck N-myc2 gene is frequently activated in altered hepatic foci, suggesting that the N-myc2 oncoprotein may confer a selective advantage for malignant conversion at early stages of Acknowledgements the carcinogenic process (Yang et al., 1993). It is We are grateful to P Hofschneider and A Kekule for providing HBV-derived plasmids, and G Courtois and J tempting to speculate that in chronic viral infections, Hassell for the gift of plasmids pb-®b and pPEA3. We N-myc2 expression might be activated in trans by thank R Bosselut for stimulating discussions. This work intracellular signal transduction pathways triggered by was supported in part by a grant from the Association liver in¯ammation or by the expression of viral pour la Recherche sur le Cancer (contract 6550). MF was proteins such as the X transactivator. Then at supported by the MinisteÁ re de la Recherche et de tumorous stages, overexpression of N-myc2ina l'Enseignement SupeÂ rieur.

References

Balsano C, Avantaggiati ML, Natoli G, DeMarzio E, Will Dandri M, Schirmacher P and Rogler CE. (1996). J. Virol., H, Perricaudet M and Levrero M. (1991). Biochem. 70, 5246 ± 5254. Biophys. Res. Com., 176, 985 ± 992. DeWetJR,WoodKV,DeLucaM,HelinskiDRand Bhat NK, Fisher RJ, Fujiwara S, Ascione R and Papas TS. Subramani S. (1987). Mol. Cell. Biol., 7, 725 ± 737. (1987). Proc. Natl. Acad. Sci., 84, 3161 ± 3165. DePinho RA, Schreiber-Agus N and Alt FW. (1991). Adv. BosselutR,DuvallJF,GeÂgonne A, Bailly M, HeÂ mar A, Cancer Res., 57, 1 ± 45. Brady J and Ghysdael J. (1990). EMBO J., 9, 3137 ± 3144. Flajolet M, Fourel G, Wei Y, Transy C, Renard CA, Tiollais Boulukos KE, Pognonec P, Rabault B, Begue A and P and Buendia MA. (1997). Viral Hepatitis and Liver Ghysdael J. (1989). Mol. Cell. Biol., 9, 5718 ± 5721. Disease. Verme G, Rizzetto M and Bonino F. (eds) ARL: Dalmon J, Laurent M and Courtois G. (1993). Mol. Cell. New York; in press. Biol., 13, 1183 ± 1193. Transactivation of the N-myc2promoter M Flajolet et al 1110 Fourel G. (1994). Liver gene expression. Tronche F and Prosser HM, Wotton D, Gegonne A, Ghysdael J, Wang S, Yaniv M. (eds) RG Landes Company: Austin; pp 297 ± Speck NA and Owen MJ. (1992). Proc. Natl. Acad. Sci. 343. USA, 89, 9934 ± 9938. Fourel G, Couturier J, Wei Y, Apiou F, Tiollais P and Rossner M. (1992). J. Med. Virol., 36, 101 ± 107. Buendia MA. (1994). EMBO J., 13, 2526 ± 2534. Schneikert J, Lutz Y and Wasylyk B. (1992). Oncogene, 7, Fourel G, Ringeisen F, Flajolet M, Tronche F, Pontoglio M, 249 ± 256. Tiollais P and Buendia MA. (1996). J. Virol., 70, 8571 ± Stemmer WP and Morris SK. (1992). Biotechniques, 13, 8583. 214 ± 220. Fourel G, Transy C, Tennant BC and Buendia MA. (1992). Toshkov I, Hacker HJ, Roggendorf M and Bannasch P. Mol. Cell. Biol., 12, 5336 ± 5344. (1990). J. Cancer Res. Clin. Oncol., 116, 581 ± 590. Fourel G, TreÂ po C, Bougueleret L, Henglein B, Ponzetto A, Transy C, Fourel G, Robinson WS, Tiollais P, Marion PL Tiollais P and Buendia MA. (1990). Nature, 347, 294 ± 298. and Buendia MA. (1992). Proc.Natl.Acad.Sci.USA,89, Girones R, Cote PJ, Hornbuckle WE, Tennant BC, Gerin 3874 ± 3878. JL, Purcell RH and Miller RH. (1989). Proc. Natl. Acad. Ueda K and Ganem D. (1996). J. Virol., 70, 1375 ± 1383. Sci. USA, 86, 1846 ± 1849. Ueda K, Wei Y and Ganem D. (1996a). Virology, 217, 413 ± Heinrich PC, Castell JV and Andus T. (1990). Biochem. J., 417. 265, 621 ± 636. Ueda K, Wei Y and Ganem D. (1996b). J. Virol., 70, 4714 ± Hocke G, Barry D and Fey GH. (1992). Mol. Cell. Biol., 12, 4723. 2282 ± 2294. Urness LD and Thummel CS. (1990). Cell, 63, 47 ± 61. KekuleÂ AS, Lauer U, Weiss L, Luber B and Hofschneider Walters MC, Magis W, Fiering S, Eidemiller J, Scalzo D, PH. (1993). Nature, 361, 742 ± 745. Groudine M and Martin DIK. (1996). Genes Dev., 10, Marcu KB, Bossone SA and Patel AJ. (1992). Annu. Rev. 185 ± 195. Biochem., 61, 809 ± 860. Wasylyk B, Hahn SL and Giovane A. (1993). Eur. J. McCarthy SA, Chen D, Yang B-S, Garcia Ramirez J, Biochem., 211, 7±18. Cherwinsky H, Chen X-R, Klagsbrun M, Hauser CA, Wei Y, Fourel G, Ponzetto A, Silvestro M, Tiollais P and Ostrowski MC and McMahon M. (1997). Mol. Cell. Biol., Buendia MA. (1992). J. Virol., 66, 5265 ± 5276. 17, 2401 ± 2412. Xin J-H, Cowie A, Lachance P and Hassell JA. (1992). Genes Nakajima K, Kusafuka T, Takeda T, Fujitani Y, Nakae K Dev., 6, 481 ± 496. and Hirano T. (1993). Mol. Cell. Biol., 13, 3027 ± 3041. Yang B-S, Hauser CA, Henkel G, Colman MS, Van Beveren Oguey D, Dumenco LL, Pierce RH and Fausto N. (1996). C, Stacey KJ, Hume DA, Maki RA and Ostrowski MC. Hepatology, 24, 1024 ± 1033. (1996). Mol. Cell. Biol., 16, 538 ± 547. Poli V and Ciliberto G. (1994). Liver gene expression. Yang D, Alt E and Rogler CE. (1993). Cancer Res., 53, Tronche F and Yaniv M. (eds) RG Landes Company: 2020 ± 2027. Austin, pp 131 ± 151. Yen TSB. (1996). J. Biomed. Sci., 3, 20 ± 30. Popper H, Roth L, Purcell RH, Tennant BC and Gerin JL. (1987). Proc. Natl. Acad. Sci. USA, 84, 866 ± 870.