Báo cáo Y học: N-myc oncogene overexpression down-regulates leukemia inhibitory factor in neuroblastoma

doi:10.1046/j.1432-1033.2002.03066.x

Eur. J. Biochem. 269, 3732–3741 (2002) (cid:1) FEBS 2002

N-myconcogene overexpression down-regulates leukemia inhibitory factor in neuroblastoma

Elissavet Hatzi1, Carol Murphy1,2, Andreas Zoephel3, Horst Ahorn,3 Ulrike Tontsch3, Ana-Maria Bamberger4, Keiko Yamauchi-Takihara5, Lothar Schweigerer6 and Theodore Fotsis1 1Laboratory of Biological Chemistry, Medical School, University of Ioannina, Greece; 2Biomedical Research Institute, Ioannina, Greece; 3Boehringer Ingelheim Austria GmbH, Vienna, Austria; 4Institute of Pathology, Department of Gynecophathology, University Hospital Hamburg Eppendorf, Hamburg, Germany; 5Department of Molecular Medicine, Osaka University Graduate School of Medicine, Suita, Japan; 6Abt. Ha¨matologie, Onkologie und Endokrinologie, Universita¨ts-Kinderklinik Essen, Germany

contiguous N-terminal N-Myc domain. STAT3, a down- stream signal transducer, was essential for LIF activity as infection with adenoviruses expressing a phosphorylation- deficient STAT3 mutant rendered endothelial cells insensi- tive to the antiproliferative action of LIF. LIF did not influence neuroblastoma cell proliferation suggesting that, at least in the context of neuroblastoma, LIF is involved in paracrine rather than autocrine interactions. Our data shed light on the mechanisms by which N-myc oncogene ampli- fication enhances the malignant phenotype in neuroblas- toma.

Keywords: N-myc; LIF; STAT3; endothelial cell; neuro- blastoma.

Amplification of N-myc oncogene is a frequent event in advanced stages of human neuroblastoma and correlates with poor prognosis and enhanced neovascularization. Angiogenesis is an indispensable prerequisite for the pro- gression and metastasis of solid malignancies, which is modulated by tumor suppressors and oncogenes. We have addressed the possibility that N-myc oncogene might regu- late angiogenesis in neuroblastoma. Here, we report that experimental N-Myc overexpression results in down-regu- lation of leukemia inhibitory factor (LIF), a modulator of endothelial cell proliferation. Reporter assays using the LIF promoter and a series of N-Myc mutants clearly demon- strated that down-regulation of the LIF promoter was independent of Myc/Max interaction and required a

The N-myc proto-oncogene encodes a 64-kDa nucleopro- tein (N-Myc) which associates with a 21- to 22-kDa Max protein to form N-Myc/Max heterodimers [1]. These dimers can bind to the E-box consensus sequence (CACGTG) in the promoter regions of target genes [2], including alpha prothymosin and ornithine decarboxylase, eventually inducing their up-regulation [3]. The physiological functions of N-Myc have remained elusive although there is evidence for developmentally important activities [4–6]. In the neural crest, enhanced N-Myc expression may facilitate prolifera- tion of immature neuronal precursor cells at the expense of differentiation [7,8].

N-myc is implicated in the pathogenesis of neural crest- derived tumors including neuroblastoma [9], the most frequent solid malignancy of infants. Amplification of N-myc oncogene is a frequent event in advanced stages (III and IV) of human neuroblastoma [10] and correlates with poor prognosis [11]. In fact, N-myc amplification is nearly exclusively observed in neuroblastoma. N-myc amplifica- tion results in high N-Myc protein levels that could perturb

the finely tuned interplay of N-Myc and Max and eventually induce abnormal expression patterns of target genes [1]. However, few N-Myc target genes have been identified so far [1]. A number of data indicate that some of the target genes might modulate the cardiovascular system. Indeed, whereas N-myc knockout mice die in utero [12,13], com- pound heterozygotes suffer from serious heart defects [6]. Moreover, N-myc may also regulate the growth of tumor vessels as neuroblastoma with N-myc amplification exhibit enhanced neovascularization [14] suggesting that N-Myc oncogene could stimulate tumor angiogenesis and thereby enhance neuroblastoma progression. Indeed, stable trans- fection and 100-fold overexpression of N-Myc in a neuro- blastoma cell line (SH-EP) resulted in an enhanced malignant phenotype of the transfectants (WAC2) and the ability to form well vascularized tumors in nude mice [15]. As tumor angiogenesis derives from an imbalance between angiogenic factors and inhibitors [16,17], it is conceivable that the genetic changes of cancer could initiate angiogenesis by disturbing this balance in the tumor vicinity. Indeed, normal p53 regulates the expression of the angio- genesis inhibitors thrombospondin [18] and glioma-derived angiogenesis inhibitory factor [19]. Also, activation of oncogenes, such as ras, has been shown to cause both up-regulation of the expression of angiogenesis stimulators, such as basic fibroblast growth factor (bFGF) or vascular endothelial growth factor (VEGF), and down-regulation of angiogenesis inhibitors such as tissue inhibitor of matrix metalloproteinases (TIMP) and thrombospondin [20].

Towards the aim of identifying N-myc-regulated mole- cules modulating neuroblastoma tumor angiogenesis, we

Correspondence to T. Fotsis, Laboratory of Biological Chemistry, Medical School, University of Ioannina, 45110 Ioannina, Greece. Tel.: + 30 65197560, Fax: + 30 65197868, E-mail: thfotsis@cc.uoi.gr Abbreviations: LIF, leukemia inhibitory factor; bFGF, basic fibroblast growth factor; VEGF, vascular endothelial growth factor; TIMP, tissue inhibitor of matrix metalloproteinases. (Received 20 February 2002, revised 30 May 2002, accepted 21 June 2002)

N-myc oncogene down-regulates LIF (Eur. J. Biochem. 269) 3733

(cid:1) FEBS 2002

After exchange of detergent from the denaturing SDS to the nondenaturing Chaps, the polyacrylamide tubes were cut in 2.5 mm slices and a small piece of each slice was allowed to diffuse, at 4 (cid:3)C overnight, in 700 lL of endothelial cell medium. The diffusates and/or their dilutions were tested for inhibitory activity on endothelial cell proliferation.

have previously screened conditioned media from SH-EP control transfectants (SH-EP007) and WAC2 cells for the presence of inhibitors or stimulators of endothelial cell proliferation [21]. We were able to demonstrate that three endothelial cell proliferation inhibitors present in SH-EP007 supernatants (SI.1, SI.2, and SI.3) were completely down- regulated in WAC2 cells [21]. In a further study, we have identified SI.3 as being activin A and documented antian- giogenic properties for this TGF-family member [22]. The present study deals with the structural and functional characterization of SI.1.

M A T E R I A L S A N D M E T H O D S

Cell culture and cell proliferation assays

The slices corresponding to the active fractions were then lyophilized and the bound proteins were allowed to difuse to tissue culture tested water. Following concentration, they dissolved in 0.25% trifluoroacetic acid/distilled H2O and loaded onto a Bakerbond Wide-Pore C18 RP-HPLC) column (4.6 · 250 mm; Malinckrodt Baker, Philipsburg, NJ, USA) equilibrated with 0.1% trifluoroacetic acid. Bound material was eluted with a linear gradient from 0 to 30% acetonitrile in 15 min and 30–60% acetonitrile in 45 min at a flow rate of 1 mLÆmin)1. Fractions of 1 mL were collected, lyophilized and dissolved in tissue culture water for activity evaluation on the proliferation of endothelial cells and amino-acid sequencing.

Mass fingerprint analysis and microsequencing

Dishes, media and recombinant growth factors have been described previously [21,23]. Cells were cultured as described [15,23,24]. Collection of conditioned medium from SH-EP007 cells was carried out as mentioned before [21]. Cell proliferation assays using BBCE cells were previously described [21,23]. Briefly, cells were seeded (day 0) in 12-well tissue culture plates at a density of 1250 cellsÆcm)2 (5000 cells per well) and the following day (day 1), wells received 10 lL of the fractions to be tested and 2.5 ngÆmL)1 bFGF. This treatment was repeated after two days (day 3). On day 5 or 6, cells in duplicate wells were trypsinized and counted using a Coulter particle counter. SHEP-007 and WAC2 cells were seeded at a density of 2500 cellsÆcm)2 (10 000 cells per well). For growth curves of WAC2 stable transfectants, cells were seeded in 12-well tissue culture plates at a density of 2500 cellsÆcm)2 (10 000 cells per well) and counted daily using a Coulter particle counter. Goat anti-human polyclonal LIF neutralizing antibody was obtained from R & D systems.

Purification of SI.1

The active fractions from the HPLC step were electropho- resed on a 12.5% SDS-polyacrylamide gel under nonreduc- ing conditions and stained with Coomassie blue R250. Protein bands were excised from the stained gel, destained, reduced and alkylated with iodoacetamide, and digested with trypsin. The resulting peptides were subjected to mass spectroscopy analysis using a MALDI-TOF instrument (Voyager DE-STR, PerSeptive Biosystems). Peptide samples were prepared using dihydroxybenzoic acid as matrix. From the calibrated MS peptide mass map a peak table list of the peptide-mass fingerprint were designed omitting signals observed in the chemical background spectrum. The peak table list served as input data and the MS-FIT software program (K. Clauser & P. Baker, available from http:// prospector.ucsf.edu/ucsfhtml4.0/msfit.htm) was used for searching the SWISS Prot and NCBI protein databases for sequence similarities and thus suggesting the identity of the protein. Parameters were set for mass tolerance at 50 p.p.m., minimum number of peptides required at 4, molecular mass of proteins from 1000 to 10 000 Da, protein pI from 3 to 10, cysteines modified as amidomthylated. MS-FIT search results were checked regarding the MOWSE score, molcular mass (Da), pI, species and % masses matched. Edman sequencing of the RP-HPLC separated tryptic peptides was performed using an 494 cLC ABI-PerkinElmer apparatus.

Transfections and LIF promoter reporter assays

Concentration of conditioned medium (47 L), acidification and SP-Sepharose Fast Flow chromatography were carried out as described previously [21]. Fractions containing the SI.1 were subjected to concavalin A affinity chromatogra- phy. A concavalin A–Sepharose column was used (11-mL bed volume, Pharmacia) equilibrated in 50 mM Tris/HCl, pH 7.2 containing 500 mM NaCl, 0.1% Chaps, 1 mM CaCl2 and 1 mM MnCl2. Samples were applied in equil- ibration buffer, and elution was carried out using: (a) 10 bed volumes of equilibration buffer; (b) a linear gradient consisting of 5 bed volumes of equilibration buffer and 5 bed volumes of equilibration buffer containing 500 mM a- methyl mannopyranoside; and (c) 5 bed volumes of equilibration buffer containing 500 mM a-methyl manno- pyranoside. A flow rate of 12 cmÆh)1 was used and fractions of 2.5 mL were collected.

The active fractions were desalted,

Full-length N-myc (pcDNA3-N-myc) was generated by PCR amplification using pN-myc as template (a gift from M. Schwab, German Cancer Research Center, Heidelberg, Germany) [24]. All deletion mutants of N-myc were gener- ated from the above construct by ligation of PCR products. All constructs were sequenced. Transient transfections were performed using lipofectamine-plus transfection reagent (Invitrogen) according to the manufacturer’s instructions.

A reporter luciferase construct containing the 666 bp human LIF promoter fragment (phLIF-Luc) was used [25]. SH-EP007 cells were plated at a density of 5 · 105 cells per well in a six-well plate were transfected with 0.4 lL of the phLIF-Luc construct, alone or in combination with 0.4 lg of

lyophilized and subjected to a preparative SDS/PAGE electrophoresis in tubes (150 · 5 mm) containing 80 · 5 mm of resolving (12%) and 30 · 5 mm of stacking (4%) polyacrylamide gel. The tubes and buffers were according to Laemmli [21a] and electrophoresis was carried out with buffer recirculation at 2 mA per tube (1 W maximun power). Following electro- phoresis the tubes were washed for 30 min in 1% Chaps in NaCl/Pi followed by a further 30 min wash in NaCl/Pi.

3734 E. Hatzi et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

subjected to concanvalin A–Sepharose affinity chromatog- raphy (Fig. 1A) and preparative SDS/PAGE tube electro- phoresis (Fig. 1B). Following exchange to a nondenaturing detergent, the activity was evaluated and the active fractions were subjected to a final C18 RP-HPLC chromatography (Fig. 1C). The main SDS/PAGE band exhibited a large size distribution (40–60 kDa) as a result of heterogeneity and glycosylation (Fig. 1D). Sequence analysis of five different segments of the main band revealed that one of the candidate proteins was LIF, a cytokine that has been previously shown to exhibit inhibitory activity on endothe- lial cells (Table 1). The rest of the sequenced peptides belonged to proteins that appeared irrelevant concerning inhibition of endothelial cell proliferation (Table 1). Con- firmation of SI.1 identity as LIF was further obtained by excising the corresponding SDS/PAGE gel segment, to that containing the sequenced LIF peptides, from another lane of the final SDS/PAGE gel. Following renaturation, this gel segment showed both inhibitory activity on endothelial cells (Fig. 2A) and the presence of LIF by Western blot analysis with a specific LIF antibody (Fig. 2B). More importantly, neutralizing LIF antibodies abolished the antiproliferative SI.1 activity on endothelial cells (Fig. 2C).

wild-type or mutated N-myc expression vectors for 28 h. In each transfection a constant total DNA concentration was used. Cells were collected and analyzed for luciferase activity using a kit according to the manufacturer (Promega, Madison, WI, USA) and a luminometer (EG & G Junior). As internal control of transfection efficiency, 0.2 lg of b-galactosidase expression vector (CMV-b-gal) was cotrans- fected and the enzymatic activity was measured. The expression and nuclear localization of all constructs was determined by Western blot using a mouse antihuman N-myc antibody (0.3 lgÆmL)1, Cymbus Biotechnology Ltd). The full-length cDNA encoding human LIF, kindly donated by Y. Jacques (INSERM U211, Nantes, France), was subcloned into the XhoI site of a CMV promoter construct. WAC2 cells (70% confluent) were transfected with effectene (Qiagen) using 1 lg of the LIF construct and 0.1 lg of a construct containing puromycin resistance gene (Clontech). Clones were selected in medium containing 0.7 lgÆmL)1 puromycin and expression of LIF was evalu- ated by Western blot analysis of supernatants using a goat anti-LIF Ig (R & D). Several independently derived clones were obtained for each construct. Control transfectants were generated using the empty CMV promoter construct.

LIF expression is down-regulated by N-Myc

Phosphorylation of STAT3 protein in BBCE cells

BBCE cells were seeded in 12-well plates (400 000 cells per well), cultured for 2 days in DMEM containing 0.5% (v/v) newborn calf serum and STAT3 and phospho-STAT3 (Tyr705) were detected as described in a kit purchased from New England Biolabs.

Preparation of adenoviruses expressing STAT3 and [3H]thymidine incorporation assays

Recombinant adenoviruses harbouring wild-type (AD/WT) and dominant negative (Y705F) (AD/DN) Stat3 cDNAs as well as a control adenovirus carrying only the vector (AD) were prepared as described previously [26]. Adenoviruses were amplified in 293 cells. The efficiency of infection and the expression of the proteins were monitored by immuno- fluorescence (Zeiss, Axiovert S100 microscope) and Western blot analysis.

Detection and isolation of LIF in SH-EP007 cells (vs. WAC2) suggested that N-Myc down-regulated LIF. To substantiate N-Myc-induced down-regulation of LIF, we used promoter–reporter assays. Upon transfection of SH-EP007 cells with a vector containing wild-type N-myc cDNA (Fig. 3A), LIF promoter activity was suppressed substantially (Fig. 3B). In order to identify the domain(s) responsible for this inhibition, several deletions within the N-myc cDNA were generated and tested (Fig. 3A). All mutant N-Myc proteins were equally well expressed (data not shown). Transfection of N-myc mutants lacking the DNA-binding [d(381–395)] or the helix/loop/helix leucine zipper [d(350–464)] regions inhibited LIF promoter-luc activity, in a manner similar to the wild type N-myc. In contrast, the N-terminal part of N-myc seemed to be important for the inhibition of the LIF promoter transcrip- tion as the mutants d(1–134) and d(1–300) had completely lost this ability. As deletion of either the MbI [d(20–90)] or the MbII [d(96–140)] domains, together with flanking sequences, did not result in derepression of the inhibitory activity (Fig. 3B), the results indicate that a contiguous N-terminal N-Myc domain is essential for suppression of the LIF promoter transcription. These data indicate an N-myc-specific role in inhibiting LIF promoter activity.

LIF inhibits the proliferation of endothelial cells but not that of WAC2 neuroblastoma cells

BBCE cells grown to subconfluency were infected with recombinant adenoviruses at a multiplicity of infection of 5 : 1 for 2 h in full medium. The medium was renewed for another 6 h allowing the constructs to be expressed and then various concentrations of recombinant LIF (R & D Systems) were added for an additional day. Then, 1 lCiÆmL)1 of [3H]methyl-thymidine (ICN) was added to each of the wells for the last 4 h of the incubation. Culture medium was removed and the cells were fixed with ice-cold 10% trichloroacetic acid for 20 min at 4 (cid:3)C, washed three times with water and solubilized in 0.1 M NaOH overnight at 4 (cid:3)C. The radioactivity was counted in a beta liquid scintillation counter (LKB).

R E S U L T S

SI.1 is identified as LIF

The flow-through of the initial cation exchange chroma- tography step containing SI.1 activity [21] was sequentially

We had identified LIF due to its ability to inhibit an important step of angiogenesis, i.e. vascular endothelial cell proliferation. In agreement with this finding, in addition to bovine brain capillary endothelial cells, recombinant human LIF (rhLIF) strongly inhibited the bFGF-stimulated pro- liferation of aorta (BAE) and human dermal microvascular endothelial cells (Fig. 4A, and data not shown). However, endothelial cells from other tissue origin, such as those from adrenal cortex (ACE), were weakly inhibited (Fig. 4A). At

N-myc oncogene down-regulates LIF (Eur. J. Biochem. 269) 3735

(cid:1) FEBS 2002

Fig. 1. Purification and isolation of SI.1. (A) The flow through fractions of the previous cation exchange chromatography step, containing SI.1, were ultrafiltrated, lyophilized, dissolved in equilibration buffer and applied to a concanvalin A–Sepharose column. Bound material was eluted with linear gradients of a-methyl-mannopyranoside as indicated (–). Aliquots of the fractions were tested for protein content (s) and activity on endothelial (BBCE) cell proliferation (d). Ten microliters of the fractions were added every other day with 2.5 ngÆmL)1 bFGF. The results were expressed as percentage of control (cells receiving buffer only). (B) The active fractions from concanvalin A–Sepharose chromatography were concentrated and diafiltrated using ultrafiltration and were subjected to SDS/PAGE in tubes as described under Material and methods. Following electrophoresis, the tubes were washed with NaCl/Pi/1% Chaps followed by a further 30-min wash in NaCl/Pi. The 12% polyacrylamide tubes were cut into 2.5-mm slices and a small piece of each slice was allowed to diffuse, overnight at 4 (cid:3)C with gentle agitation, in 700 lL of endothelial cell medium. The diffusates were tested for inhibitory activity on endothelial cell proliferation (d). (C) The proteins of the active gel pieces were eluted in dH2O and lyophilized for application onto a C18 HPLC column. The application was performed in 0.1% TFA/dH20 and elution was carried out with a linear gradient from 0 to 30% acetonitrile in 10 min and 30–60% acetonitrile in 45 min. A flow rate of 1 mLÆmin)1 was used and fractions of 1 mL were collected, lyophilized and resuspended into 100 lL culture-tested water for activity evaluation. Ten microliters of 100-fold dilutions of the fractions were used for BBCE proliferation test (d). (D) The fractions corresponding to the active peak of the HPLC column in Fig. 1C were pooled, lyophilized, and separated on a 12.5% SDS-polyacrylamide gel under reducing conditions. The bands were excised from Coomassie blue stained gel and subjected to in-gel digestion followed by mass fingerprint analysis as described in material and methods. Left lane, pooled HPLC fractions; right lane, molecular markers in kDa.

Table 1. Proteins identified by mass analysis and Edman sequencing of tryptic peptides. The position of peptide amino acid in protein was determined by Edman sequencing.

Band No. number Protein position SWISS-PROT accession Amino-acid

1, minor Metabotropic glutamate Q14416 20–34 receptor 2

2, major

3, major

29–43 323–337 87–107 148–157 205–213 23–37, 61–80, 193–202, 202–107 4, major 5, minor Signal recognation particle Alpha antitrypsin Carboxypeptidase H Prec Carboxypeptidase H Prec Alpha antitrypsin Leukemia inhibitory factor Keratin I Keratin II P13624 P34955 P16870 P16870 P34955 P15018 P13645 P04264

3736 E. Hatzi et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

80 min. Simultaneous administration of angiogenic factors such as bFGF did not alter the phosphorylation pattern of STAT3 (data not shown).

We wished to determine whether the STAT3 pathway was necessary and crucial for the LIF-induced inhibition of vascular endothelial cell proliferation. To that aim, we infected BBCE cells with recombinant adenoviruses con- taining various stat3 forms and investigated the ability of rhLIF to inhibit bFGF-induced proliferation of the infected cells. rhLIF was able to inhibit proliferation of BBCE cells infected with the empty vector controls (AD) or with the vector containing the wild-type stat3 (AD/WT) (Fig. 5B). In contrast, rhLIF was unable to inhibit proliferation of BBCE cells infected with the dominant-negative stat3 mutant (AD/ DN) (Fig. 5B). Thus, the STAT3 pathway is important for mediating the inhibitory signals of LIF regarding endothe- lial cell proliferation.

D I S C U S S I O N

for the first

time,

We have previously shown that neuroblastoma cell super- natants contained three endothelial cell proliferation inhib- itors (SI.1, SI.2 and SI.3), the expression of which were dramatically down-regulated upon N-Myc overexpression [21]. In the present study, SI.1 activity was isolated from 47 L of SH-EP007 supernatants by a series of chromatographic steps using inhibition of endothelial cell proliferation as a marker for biological activity. The main SDS/PAGE band exhibited a large size distribution (40–60 kDa) as a result of heterogeneity and glycosylation. Sequence analysis of tryptic peptides combined with detection with specific antibodies identified SI.1 as LIF. A conclusion further supported by inhibition of the SI.1 activity by specific anti- LIF neutralizing antibodies. Identification of SI.1 as LIF strongly implied that N-Myc overexpression down-regulat- ed LIF expression in neuroblastoma. Indeed, reporter assay experiments revealed, that N-Myc overexpression dramatically down-regulated LIF promoter transcription. LIF has never been reported to belong to a set of genes shown to be regulated by the myc gene family members using various assays [1,27], including crosslinking [28], and cDNA microarrays [29].

the concentrations used, rhLIF was not cytotoxic, as seen by microscopic evaluation and by the fact that cell densities never fell below those present at seeding. rhLIF also inhibited basal proliferation of BBCE cells (data not shown and Fig. 5). The inhibitory effect of LIF on endothelial cell proliferation did not, however, exclude a direct autocrine inhibitory effect of LIF on neuroblastoma cell proliferation. Towards this end, rhLIF had no effect on the proliferation of SH-EP007 and WAC2 neuroblastoma cells, even at very high concentrations (Fig. 4B). The same results were obtained when we transfected WAC2 neuroblastoma cells with a vector containing the human LIF cDNA under the control of a CMV promoter (not down-regulated by N-Myc) compared to cells transfected with the empty vector. The proliferation potential of the LIF-expressing clones was either similar (c6) or higher (c11, c13) than the control clones (cve-1 and cve-2) (Fig. 4C). Indeed, in the case of c11 and c13 the proliferation rate was similar to that of the parental WAC2 cells and clearly distinct from that of the low N-myc expressing SH-EP007 cells (Fig. 4C). Thus, forced overexpression of LIF in WAC2 cells did not inhibit their in vitro proliferation.

Myc/Max interactions play an important role in Myc- induced transcriptional regulation. Indeed, the C-terminus of N-Myc contains the HLH and Zip domains, which mediate heterodimerization with Max [2,30,31], and the BR domain, which is required for binding of the N-Myc/ Max heterodimers to consensus sites known as E boxes [30,32,33]. Our data suggest that down-regulation of LIF is independent of N-Myc/Max interaction and DNA binding as neither the HLH-Zip nor the BR deletion released the transcriptional repression. This excludes an indirect repres- sion of LIF gene via some of the known N-Myc/Max regulated genes [1].

LIF inhibits endothelial cell proliferation via the STAT3 pathway

Our data, however, reserve an important role for the N-terminal domain of N-Myc in down-regulation of LIF. Indeed, deletion of amino acids 1–134 abolished almost all of the N-Myc repressing activity. The N-terminal domain contains Myc boxI (MbI) and Myc boxII (MbII), two highly conserved domains of the myc family that are considered important sites for protein interactions. In most transcriptional assays, MbII domain is associated with repression rather than transactivation [34,35]. Indeed, MbII

LIF can mediate its effects by binding to specific cell surface receptors with subsequent phosphorylation of the transcrip- tion factor STAT3 at Tyr705. When rhLIF was added to BBCE cells, it was, in fact, able to phosphorylate STAT3 in a time-dependent manner (Fig. 5A). Phosphorylation of STAT3 Tyr705 was observed at 2.5 min post induction, reached a maximum by 20 min and declined slowly over

Fig. 2. Identification of SI.1 as LIF. A small fraction of the sample used for the sequencing of SI.1 was subjected to the same SDS/PAGE electrophoresis under nonreducing conditions and treated as in Fig. 1B. The gel pieces were analyzed for cell growth inhibition and correlation with LIF immunoreactivity. (A) Part of the eluates of the gel pieces (presented in this figure the gel pieces 7, 13, 17) was exami- nated for their ability to inhibit bFGF-stimulated proliferation of BBCE cells. (B) Aliquots of the same eluates were analyzed by Western blot using a goat anti-LIF Ig (1 lgÆmL)1). The bound antibodies were detected with horseradish peroxidase conjugated anti-goat IgG fol- lowed by ECL detection system. (C) Neutralization of SI.1 activity. BBCE cells were treated either with rhLIF (2.5 ngÆmL)1) or aliquots of the gel piece 13 alone or in combination with antibodies against LIF (anti-LIF, 1 lgÆmL)1) as described in material and methods. The results were expressed as percent of control (cells received only bFGF).

N-myc oncogene down-regulates LIF (Eur. J. Biochem. 269) 3737

(cid:1) FEBS 2002

Identification of LIF down-regulation by the N-myc oncogene in neuroblastoma provides novel and important information regarding the progression mechanisms of this tumor to more malignant stages. Indeed, it has been reported that LIF induces differentiation of NBFL neuro- blastoma cell line [39] and Myc proteins are known to repress the expression of several differentiation-associated genes, including C/EBP regulated genes, mim-1 and lyso- zyme as well as c/EBP itself [34,40]. It is interesting to note that Myc-mediated repression, rather than transactivation, has been suggested to correlate with transformation [41–43]. To find further links between LIF and neuroblastoma progression, and because we have purified LIF as an inhibitor of endothelial cell proliferation, it was important to see whether LIF had any effect on neuroblastoma cell proliferation itself. Treatment of WAC2 and SH-EP007 cells with recombinant LIF did not result in inhibition of

domain has been shown to repress promoters with Inr elements [36] and interact with TRRAP [37] a protein that is part of the SAGA transcriptional regulatory complex [38]. Moreover, an important repression element lies between MbI and the transcription-activating domain upstream of MbII. A second repression element overlaps MbII itself [1]. In this study, deletion of either of the MbI and MbII domains, together with flanking sequences, did not release the N-myc mutants from the suppressive effect on LIF promoter. It appears therefore that a contiguous N-terminal N-Myc is required for repression. Indeed, Bin1, a protein recruiting repressive activity to Myc-responsive promoters and antagonizing the oncogenic activity of Myc, require both MbI and MbII domains. Also, TRRAP requires sequences of Myc N-terminal domain in addition to MbII and does not bind Myc when MbII upstream deletion (amino acids 1–110) is imposed [37].

Fig. 3. Regulation of LIF promoter transcription by N-myc. (A) Schematic diagram of N-myc protein and mutants thereof demonstrating important regions with abbreviations in parenthesis: MYC-boxes I and II (MBI, MBII), acidic region at the exon 2–3 border (Ex2/3), nuclear localization region (N), basic region (BR), helixl-loop-helix2 (Hl-L-H2) and leucine zipper region (Zip). Lines instead of boxes indicate deletions. (B) Effect of N-myc and its mutants on the transcriptional activation of the LIF promoter in SH-EP007 cells. Cells at 60–70% confluence were transfected by lipofectamine plus reagent with 0.4 lg phLIF-luc alone or with 0.4 lg of each N-myc constructs. A b-galactosidase encoding plasmid was cotransfected in order to normalize luciferase activity. 24 h after transfection, cells were Iysed and transcriptional activation was measured as described in Material and methods. The graph represents relative expression of luciferase activity where one arbitrary unit represents the expression of luciferase in the presence of wild-type N-myc. Each luciferase transfection experiment was carried out in triplicate and repeated at least three times. Mock transfected cells served as control. (C) Western blot analysis of the N-myc constructs expressed in the above transfection experiment. A monoclonal anti-(N-Myc) Ig (0.3 lgÆmL)1) and a second anti-mouse peroxidase conjugated antibody were used for the detection. Arrowheads indicate the N-myc protein and mutants thereof.

3738 E. Hatzi et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

their proliferation. Also, re-introduction, by stable trans- fection, of LIF into WAC2 cells under the control of a CMV promoter (not affected by N-Myc) resulted in clones that showed either no difference or even slightly enhanced growth characteristics as compared to control transfectants. This result suggested that LIF inhibits preferentially endo- thelial cell proliferation. Perhaps inhibition of angiogenesis is a contributing mechanism through which this cytokine could be involved in N-Myc-induced progression of neuro- blastoma malignant phenotype. Thus, N-myc amplification by down-regulating LIF could tilt the balance of angio- modulators in favor of the stimulators and thus participate in increasing neovascularization in N-myc-amplified neuro- blastoma.

and interleukin-18,

interleukin-12,

differential actions, depending upon tissue of origin or the differentiation stage of endothelial cells. Thus, LIF stimu- lates the growth of adrenal cortex capillary and embryonic endothelial cells but inhibits proliferation of aortic endo- thelial cells [45–47]. Also in our hands, LIF inhibited proliferation of brain, aortal, and dermal endothelial cells whereas it had a weak effect on adrenal cortex and umbilical vein endothelial cells (Fig. 4A and data not shown). Regardless of whether the effect of LIF on proliferation is inhibitory or stimulatory, activation of STAT3 by the LIF/ gp130 receptors appears to play an important role. Indeed, activation of STAT3 can either induce cell proliferation via cyclin D1 overexpression [48] or cause growth arrest by enhancing expression of p21 or p27 cell cycle inhibitors [49,50]. In all cases, dominant negative mutants of STAT3 reverse the effects on proliferation [48–50]. In the present study, infection of BBCE cells with adenoviruses expressing negative mutant STAT3 (Y705F) effectively reversed the inhibitory effect of LIF on BBCE proliferation in the presence or in the absence of bFGF (Fig. 4B and data not shown). This result indicates an important role for STAT3

Several cytokines have been reported to participate in the regulation of the angiogenic switch. Whereas some of these cytokines, such as interleukin-8 and interleukin-4, stimulate angiogenesis, the majority of the others, including platelet factor 4, interferon-c-inducing protein 10, IFN-2a, inter- leukin-10, inhibit neovascularization [44]. LIF has been shown to exert

Fig. 4. Effect of rhLIF on endothelial and neuroblastoma cell proliferation. (A) rhLIF inhibits endothelial cell proliferation. Dose–response of endothelial cell inhibition by recombinant human LIF. BBCE (d), ACE (j), BAE (m) cells were cultured in the presence of bFGF (2.5 ngÆmL)1) and the indicated concentrations of rhLIF. Cell numbers were determined in triplicates and varied by less than 10% of the mean. Values are expressed as percent of controls (cells receiving only bFGF). (B,C) rhLIF had no effect on the proliferation of neuroblastoma cells. (B) SH-EP007 (j) and WAC (d) neuroblastoma cells were cultured in the presence of different concentrations of rhLIF. Cell numbers were determined in triplicates and varied by less than 10% of the mean. Values are expressed as percent of controls (cells receiving no rhLIF). (C) Growth rates of LlF- transfected human neuroblastoma cells in vitro. SH-EP007 (j), WAC2 cells (d) and LIF transfected WAC2 cell clones c6 (n), c11 (h), c13 (.) and theirs controls cve-l (s) and cve-2 (m) were seeded at densities of 10 000 cellsÆcm)2 and counted daily. Cell numbers were detennined at the indicated times and represent the means of triplicate determinations. The results were presented as fold proliferation.

N-myc oncogene down-regulates LIF (Eur. J. Biochem. 269) 3739

(cid:1) FEBS 2002

BMP2 or TGF-b, respectively [54,55]. It is tempting to speculate that perhaps the Stat3 and Smad2/3 pathways cooperate in suppressing angiogenesis in early stages of neuroblastoma. Upon N-Myc amplification, simultaneous down-regulation of LIF and activin A triggers the angio- genic switch by abrogating both the Stat3 and Smad2/3 pathways. N-Myc appears to achieve these effects via two different domains as repression of activin A promoter is Myc/Mac dependent [22] whereas that of the LIF promoter is Myc/Max independent and requires a contiguous N-terminal domain.

In conclusion, this work sheds light on the molecular mechanisms contributing to the enhancing effect of N-myc amplification on malignant transformation. In particular, it suggests a repressive role of N-Myc on LIF expression, an inhibitor of endothelial cell proliferation. The result indi- cates a possible role of N-Myc in tilting the balance between stimulators and inhibitors of angiogenesis towards the former, at least in some neuroblastoma tumors. STAT3, a downstream transducer of signals from LIF receptors, appear to play a critical role regarding the inhibitory effect of LIF on endothelial cell proliferation.

A C K N O W L E D G E M E N T S

The skillful technical support of Lambrini Kyrkou is grarefully acknowledged. This work was supported by Boehringer Ingelheim Austria GmbH, Vienna, Austria, Deutsche Forshungsgemeinschaft and a grant from the Joint Research and Technology Program between Greece (General Secretariat of Research and Technology) and Germany (T. F. and L. S.). E. H. was supported by a postdoctoral grant from the State Award Foundation of Greece.

R E F E R E N C E S

1. Facchini, L.M. & Penn, L.Z. (1998) The molecular role of Myc in growth and transformation: recent discoveries lead to new insights. FASEB J. 12, 633–651.

2. Wenzel, A., Cziepluch, C., Hamann, U., Schurmann, J. & Schwab, M. (1991) The N-myc oncoprotein is associated in vivo with the phosphoprotein Max (p20/22) in human neuroblastoma cells. EMBO J. 10, 3703–3712.

Fig. 5. LIF activates STAT3 in endothelial cells. (A) LIF phosphory- lates STAT3 at Tyr705 in BBCE cells. Time course analysis of LIF induced STAT3 phosphorylation was performed by treated the cells with 20 ngÆmL)1 rhLIF for the indicated time. The activation was tested by Western blots of cell Iysates with antibodies against phos- pho705-STAT3 (phospho-STAT3) or against total STAT3 (STAT3) as a loading control. (B) LIF induced inhibition of BBCE cells is STAT3 dependent. 80% confluent BBCE cells were infected either with adenovirus carrying wild-type STAT3 (AD/WT) or adenovirus carrying the dominant negative STAT3 (AD/DN) as it was described under Material and methods. Adenovirus vector itself used as control (AD). Then the cells were treated with LIF (60 ngÆmL)1) for 24 h. For the last 4 h, 1 lCiÆmL)1 of [3H]methyl-thymidine was added. The radioactivity were counted in a beta liquid scintillation counter. 3. Lutz, W., Stohr, M., Schurmann, J., Wenzel, A., Lohr, A. & Schwab, M. (1996) Conditional expression of N-myc in human neuroblastoma cells increases expression of alpha-prothymosin and ornithine decarboxylase and accelerates progression into S-phase early after mitogenic stimulation of quiescent cells. Oncogene 13, 803–812.

as

previously

reported

for

4. Mugrauer, G., Alt, F.W. & Ekblom, P. (1988) N-myc proto- oncogene expression during organogenesis in the developing mouse as revealed by in situ hybridization. J. Cell Biol. 107, 1325–1335.

5. Giroux, S. & Charron, J. (1998) Defective development of the embryonic liver in N-myc-deficient mice. Dev. Biol. 195, 16–28. 6. Moens, C.B., Stanton, B.R., Parada, L.F. & Rossant, J. (1993) Defects in heart and lung development in compound hetero- zygotes for two different targeted mutations at the N-myc locus. Development. 119, 485–499.

7. Wakamatsu, Y., Watanabe, Y., Nakamura, H. & Kondoh, H. (1997) Regulation of the neural crest cell fate by N-myc: pro- motion of ventral migration and neuronal differentiation. Development 124, 1953–1962.

in modulating angiogenic responses of endothelial cells. Some recent studies have shown that both bFGF [51] and VEGF [52,53] can induce STAT3 phosphorylation and this is an excellent point for crosstalk. However, we were unable to observe phosphorylation of STAT3 from either angio- genic factor. Perhaps, cyclical, recurrent or delayed phos- phorylation bFGF phosphorylation of STAT3 might have been the reason [51]. We have previously shown that activin A, a member of the TGF-b superfamily, is also a target gene down-regulated by N-Myc in neuroblastoma [22]. Moreover, we have demonstrated that activin A inhibits neovascularization in the chorioallantoic membrane [22]. Interestingly, a complex formation between Stat3 and Smad1 or Smad3, bridged by p300, is involved in the cooperative signaling of LIF and

8. Thiele, C.J., Reynolds, C.P. & Israel, M.A. (1985) Decreased expression of N-myc precedes retinoic acid-induced morpholo-

3740 E. Hatzi et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

gical differentiation of human neuroblastoma. Nature 313, 404– 406. transduces hypertrophic signals via STAT3 in cardiac myocytes. Circulation. 98, 346–352. 27. Grandori, C. & Eisenman, R.N. (1997) Myc target genes. Trends Biochem. Sci. 22, 177–181.

28. Mac, S.M., D’Cunha, C.A. & Farnham, P.J. (2000) Direct recruitment of N-myc to target gene promoters. Mol Carcinog. 29, 76–86.

9. Weiss, W.A., Aldape, K., Mohapatra, G., Feuerstein, B.G. & Bishop, J.M. (1997) Targeted expression of MYCN causes neuroblastoma in transgenic mice. EMBO J. 16, 2985–2995. 10. Brodeur, G.M., Seeger, R.C., Schwab, M., Varmus, H.E. & Bishop, J.M. (1984) Amplification of MYCN in untreated human neuroblastomas correlates with advanced disease stage. Science 224, 1121–1124.

29. Guo, Q.M., Malek, R.L., Kim, S., Chiao, C., He, M., Ruffy, M., Sanka, K., Lee, N.H., Dang, C.V. & Liu, E.T. (2000) Identifi- cation of c-myc responsive genes using rat cDNA microarray. Cancer Res. 60, 5922–5928.

30. Blackwood, E.M. & Eisenman, R.N. (1991) Max: a helix-loop- helix zipper protein that forms a sequence-specific DNA-binding complex with Myc. Science 251, 1211–1217.

11. Grady-Leopardi, E., Schwab, M., Ablin, A. & Rosenau, W. (1986) Detection of NMYC oncogene expression in human neuroblastoma by in situ hybridization and blot analysis: Relationship to clinical outcome. Cancer Res. 46, 3196–3199. 12. Sawai, S., Shimono, A., Hanaoka, K. & Kondoh, H. (1991) Embryonic lethality resulting from disruption of both N-myc alleles in mouse zygotes. New Biol. 3, 861–869. 31. Blackwood, E.M., Kretzner, L. & Eisenman, R.N. (1992) Myc and Max function as a nucleoprotein complex. Curr. Opin. Genet Dev. 2, 227–235.

13. Stanton, B.R. & Parada, L.F. (1992) The N-myc proto-onco- gene: developmental expression and in vivo site-directed muta- genesis. Brain Pathol. 2, 71–83. 32. Prendergast, G.C. & Ziff, E.B. (1991) Methylation-sensitive sequence-specific DNA binding by the c-myc basic region. Sci- ence 251, 186–189.

amplification, outcome poor and in 33. Kerkhoff, E., Bister, K. & Klempnauer, K.H. (1991) Sequence- specific DNA binding by Myc proteins. Proc. Natl Acad. Sci. USA 88, 4323–4327. 14. Meitar, D., Crawford, S.E., Radmaker, A.W. & Cohn, S.L. (1996) Tumor angiogenesis correlates with metastatic disease, human N-myc neuroblastoma. J. Clin. Oncol. 14, 405–414.

34. Li, L.H., Nerlov, C., Prendergast, G., MacGregor, D. & Ziff, E.B. (1994) c-myc represses transcription in vivo by a novel me- chanism dependent on the initiator element and Myc box II. EMBO J. 13, 4070–4079. 15. Schweigerer, L., Breit, S., Wenzel, A., Tsunamoto, K., Ludwig, R. & Schwab, M. (1990) Augmented MYCN expression advances the malignant phenotype of human neuroblastoma cells: Evidence for induction of autocrine growth activity. Cancer Res. 50, 4411–4416.

16. Liotta, L.A., Steeg, P.S. & Stetler-Stevenson, W.G. (1991) Can- cer metastasis and angiogenesis: An imbalance of positive and negative regulation. Cell 64, 327–336. 35. Penn, L.J., Laufer, E.M. & Land, H. (1990) C-myc: evidence for multiple regulatory functions. Semin. Cancer Biol. 1, 69–80. 36. Lee, L.A., Dolde, C., Barrett, J., Wu, C.S. & Dang, C.V. (1996) A link between c-myc-mediated transcriptional repression and neoplastic transformation. J. Clin. Invest. 97, 1687–1695.

17. Hanahan, D. & Folkman, J. (1996) Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86, 353–364.

37. McMahon, S.B., Van Buskirk, H.A., Dugan, K.A., Copeland, T.D. & Cole, M.D. (1998) The novel ATM-related protein TRRAP is an essential cofactor for the c-Myc and E2F onco- proteins. Cell 94, 363–374. 18. Dameron, K.M., Volpert, O.V., Tainsky, M.A. & Bouck, N. (1994) Control of angiogenesis in fibroblasts by p53 regulation of thrombospondin-1. Science. 265, 1582–1584.

19. Van Meir, E.G., Polverini, P.J., Chazin, V.R., Su Huang, H.-J., De Tribolet, N. & Cavenee, W.K. (1994) Release of an inhibitor of angiogenesis upon induction of wild type p53 expression in glioblastoma cells. Nat. Genet. 8, 171–176. 20. Bouck, N., Stellmach, V. & Hsu, S.C. (1996) How tumors 38. Grant, P.A., Duggan, L., Cote, J., Roberts, S.M., Brownell, J.E., Candau, R., Ohba, R., Owen-Hughes, T., Allis, C.D., Winston, F., Berger, S.L. & Workman, J.L. (1997) Yeast Gcn5 functions in two multisubunit complexes to acetylate nucleosomal histones: characterization of an Ada complex and the SAGA (Spt/Ada) complex. Genes Dev. 11, 1640–1650. become angiogenic. Adv. Cancer Res. 69, 135–174.

39. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S.J., Yasukawa, K., Kishimoto, T., Anderson, D.J. et al. (1992) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gp130. Cell 69, 1121–1132.

40. Mink, S., Mutschler, B., Weiskirchen, R., Bister, K. & Klemp- nauer, K.H. (1996) A novel function for Myc: inhibition of C/EBP-dependent gene activation. Proc. Natl Acad. Sci. USA 93, 6635–6640.

21. Fotsis, T., Breit, S., Lutz, W., Rossler, J., Hatzi, E., Schwab, M. & Schweigerer, L. (1999) Down-regulation of endothelial cell growth inhibitors by enhanced MYCN oncogene expression in human neuroblastoma cells. Eur. J. Biochem. 263, 757–764. 21a. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage Tu. Nature 227, 680–685. 22. Breit, S., Ashman, K., Wilting, J., Rossler, J., Hatzi, E., Fotsis, T. & Schweigerer, L. (2000) The N-myc oncogene in human neu- roblastoma cells: down-regulation of an angiogenesis inhibitor identified as activin A. Cancer Res. 60, 4596–4601.

41. Lee, L.A. & Dang, C.V. (1997) C-myc transrepression and cell transformation. Curr. Top Microbiol Immunol. 224, 131–135. 42. Spotts, G.D., Patel, S.V., Xiao, Q. & Hann, S.R. (1997) Identi- fication of downstream-initiated c-myc proteins which are dominant-negative inhibitors of transactivation by full-length c-myc proteins. Mol. Cell. Biol. 17, 1459–1468. 23. Fotsis, T., Pepper, M.S., Aktas, E., Breit, S., Rasku, S., Adler- creutz, H., Wahala, K., Montesano, R. & Schweigerer, L. (1997) Flavonoids, dietary-derived inhibitors of cell proliferation and in vitro angiogenesis. Cancer Res. 57, 2916–2921.

24. Hiller, S., Breit, S., Wang, Z.Q., Wagner, E.F. & Schwab, M. (1991) Localization of regulatory elements controlling human MYCN expression. Oncogene 6, 969–977. 43. Xiao, Q., Claassen, G., Shi, J., Adachi, S., Sedivy, J. & Hann, S.R. (1998) Transactivation-defective c-mycS retains the ability to regulate proliferation and apoptosis. Genes Dev. 12, 3803– 3808.

44. Cao, R., Farnebo, J., Kurimoto, M. & Cao, Y. (1999) Inter- leukin-18 acts as an angiogenesis and tumor suppressor. FASEB J. 13, 2195–2202.

45. Ferrara, N., Winer, J. & Henzel, W.J. (1992) Pituitary follicular cells secrete an inhibitor of aortic endothelial cell growth: 25. Bamberger, A.M., Erdmann, I., Bamberger, C.M., Jenatschke, S.S. & Schulte, H.M. (1997) Transcriptional regulation of the human (cid:2)leukemia inhibitory factor(cid:3) gene: modulation by gluco- corticoids and estradiol. Mol. Cell. Endocrinol. 127, 71–79. 26. Kunisada, K., Tone, E., Fujio, Y., Matsui, H., Yamauchi- Takihara, K. & Kishimoto, T. (1998) Activation of gp130

N-myc oncogene down-regulates LIF (Eur. J. Biochem. 269) 3741

(cid:1) FEBS 2002

identification as leukemia inhibitory factor. Proc. Natl Acad. Sci. USA 89, 698–702. 51. Megeney, L.A., Perry, R.L., LeCouter, J.E. & Rudnicki, M.A. (1996) bFGF and LIF signaling activates STAT3 in proliferating myoblasts. Dev. Genet. 19, 139–145.

46. Pepper, M.S., Ferrara, N., Orci, L. & Montesano, R. (1995) Leukemia inhibitory factor (LIF) inhibits angiogenesis in vitro. J. Cell Sci. 108, 73–83. 52. Bartoli, M., Gu, X., Tsai, N.T., Venema, R.C., Brooks, S.E., (2000) Vascular endo- in aortic Marrero, M.B. & Caldwell, R.B. thelial growth factor activates STAT proteins endothelial cells. J. Biol. Chem. 275, 33189–33192. 47. Gendron, R.L., Tsai, F.Y., Paradis, H. & Arceci, R.J. (1996) Induction of embryonic vasculogenesis by bFGF and LIF in vitro and in vivo. Dev. Biol. 177, 332–346.

48. Bromberg, J.F., Horvath, C.M., Besser, D., Lathem, W.W. & Darnell, J.E. Jr (1998) Stat3 activation is required for cellular transformation by v-src. Mol. Cell. Biol. 18, 2553–2558. 53. Korpelainen, E.I., Karkkainen, M., Gunji, Y., Vikkula, M. & Alitalo, K. (1999) Endothelial receptor tyrosine kinases activate the STAT signaling pathway: mutant Tie-2 causing venous malformations signals a distinct STAT activation response. Oncogene 18, 1–8.

54. Nakashima, K., Yanagisawa, M., Arakawa, H., Kimura, N., Hisatsune, T., Kawabata, M., Miyazono, K. & Taga, T. (1999) Synergistic signaling in fetal brain by STAT3-Smad1 complex bridged by p300. Science 284, 479–482. 49. Bellido, T., O’Brien, C.A., Roberson, P.K. & Manolagas, S.C. (1998) Transcriptional activation of the p21 (WAF1,CIP1,SDI1) gene by interleukin-6 type cytokines. A prerequisite for their pro- differentiating and anti-apoptotic effects on human osteoblastic cells. J. Biol. Chem. 273, 21137–21144.

55. Yamamoto, T.S., Takagi, C., Hyodo, A.C. & Ueno, N. (2001) Suppression of head formation by Xmsx-1 through the inhibition of intracellular nodal signaling. Development 128, 2769–2779.

50. Kortylewski, M., Heinrich, P.C., Mackiewicz, A., Schnierts- hauer, U., Klingmuller, U., Nakajima, K., Hirano, T., Horn, F. & Behrmann, I. (1999) Interleukin-6 and oncostatin M-induced growth inhibition of human A375 melanoma cells is STAT- dependent and involves upregulation of the cyclin-dependent kinase inhibitor p27/Kip1. Oncogene 18, 3742–3753.