Tumour necrosis factor-a mediates tumour promotion via a ... - Nature

ª

Oncogene (2002) 21, 4728 – 4738 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc

Tumour necrosis factor-a mediates tumour promotion via a PKCa- and AP-1-dependent pathway Caroline H Arnott1, Kate A Scott1, Robert J Moore1, Alan Hewer2, David H Phillips2, Peter Parker3, Frances R Balkwill*,1 and David M Owens4 1

Cancer Research UK Translational Oncology Laboratory, Bart’s and The London School of Medicine and Dentistry, Queen Mary, University of London, John Vane Science Centre, Charterhouse Square, London EC1M 6BQ, UK; 2Institute of Cancer Research, Haddow Laboratories, Cotswold Road, Sutton, Surrey SM2 5NG, UK; 3Protein Phosphorylation Laboratory, Cancer Research UK, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK; 4Keratinocyte Laboratory, Cancer Research UK, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK

Tumour necrosis factor-a (TNF-a) deficient mice (TNFa7/7 mice) are resistant to skin carcinogenesis. Cellular signalling via the transcription factor complex AP-1 is thought to play a key role in tumour promotion. The induction of a specific subset of AP-1 responsive genes thought to be important for tumour development, namely GM – CSF, MMP-9 and MMP-3, was suppressed in TNF-a7/7 compared to wild-type mouse skin in response to the tumour promotor TPA. The differential induction of these genes correlated with a temporal shift in AP-1 activation and c-Jun expression in TNF-a7/7 compared to wild-type epidermis. The major receptor for TPAinduced signalling in basal keratinocytes, PKCa, was also differentially regulated in wild-type compared with TNFa7/7 epidermis. A marked delay in TPA-induced intracellular translocation and downregulation of PKCa was observed in TNF-a7/7 epidermis, which correlated with the deregulated TPA-induced AP-1 activation and cJun expression. The frequency of DNA adduct formation and c-Ha-ras mutations was the same in wild-type and TNF-a7/7 epidermis after DMBA treatment, suggesting that TNF-a was not involved in tumour initiation. These data suggest that the pro-inflammatory cytokine TNF-a is a critical mediator of tumour promotion, acting via a PKCa- and AP-1-dependent pathway. This may be one mechanism by which chronic inflammation increases susceptibility to cancer. Oncogene (2002) 21, 4728 – 4738. doi:10.1038/sj.onc. 1205588 Keywords: inflammation; skin; keratinocyte; matrix metalloproteinase; carcinogenesis

Introduction The pro-inflammatory cytokine tumour necrosis factora (TNF-a) is recognized as a key mediator of

*Correspondence: FR Balkwill; E-mail: [email protected] Received 14 December 2001; revised 8 April 2002; accepted 15 April 2002

inflammation with actions directed towards both tissue destruction and recovery from damage (Balkwill and Mantovani, 2001; Balkwill, 2002). The biological activities of TNF-a are mediated by two structurally related but functionally distinct receptors, designated the p55 and p75 TNF-a receptors. It has become apparent that a complex array of signalling events are initiated in response to TNF-a receptor activation, giving rise to the pleiotropic effects of TNF-a on cells (reviewed by Baud and Karin, 2001; Locksley et al., 2001). The cytokine TNF-a has paradoxical roles in the evolution and treatment of malignant disease. High dose local administration of TNF-a selectively destroys tumour blood vessels and has powerful anti-cancer actions (Lejeune et al., 1998). However, TNF-a appears to act as an endogenous tumour promoter when chronically produced, contributing to the tissue remodelling and stromal development necessary for tumour growth and spread (reviewed by Balkwill and Mantovani, 2001). Maintenance of the oncogenic phenotype is dependent on the accumulation of genetic changes, as well as epigenetic events such as the induction of specific promotion-relevant effector genes. Certain target genes of the AP-1 transcription factor complex are thought to mediate in neoplastic transformation, although the identity of these genes remains largely unknown (Vogt, 2001). AP-1 is known to be involved in TNF-a receptor signalling, enabling TNF-a to alter the expression of many genes (see Baud and Karin, 2001), some of which may be important in tumour development (Balkwill and Mantovani, 2001). Moreover, AP-1 has been described as a key player in epidermal tumour promotion and progression by 12-O-tetradecanoylphorbol-13-acetate (TPA), a mimic of diacylglycerol (Dekker and Parker, 1994), in studies using cell culture models (Bernstein and Colburn, 1989; Ben-Ari et al., 1992; Dong et al., 1994, 1997; Domann et al., 1994) and transgenic mice (Young et al., 1999; Zhong et al., 2001). The AP-1 complex is composed of heterodimers of Jun (c-Jun, JunB and JunD) and Fos (c-Fos, Fos B, Fra1 and Fra2) family members or homodimers of Jun-Jun. In particular, c-Jun has been implicated in

TNF-a modulates PKCa and AP-1 signalling CH Arnott et al

4729

events leading to tumour development (Ben-Ari et al., 1992; Dong et al., 1997; Young et al., 1999; Li et al., 2000; Behrens et al., 2000), whereas c-Fos appears to be involved in the malignant conversion of benign papillomas (Saez et al., 1995). The AP-1 transcription factor complex is a nuclear target of PKC-mediated signal transduction in skin (Angel et al., 2001). PKC isoforms have been identified as major cellular receptors for TPA (Nishizuka, 1984; Kikkawa et al., 1983) and as such TPA is thought to exert many of its cellular effects through interaction with PKC. Mouse epidermis has been shown to express 6 PKC isoforms (PKCa d, e, Z, m, and z). Of these, PKCa is thought to be of functional relevance since it has been implicated in events leading to keratinocyte differentiation, epidermal tumour promotion and cutaneous inflammation (Hansen et al., 1990; Mills et al., 1992; Dlugosz et al., 1994; Lee et al., 1997; Wang and Smart, 1999). Interestingly, several recent reports have suggested that TNF-a can act as a physiological effector of PKCa activation (Chen et al., 2000, 2001; Lee et al., 2000). Recent work in our laboratory and others has implicated endogenous TNF-a as a tumour promotor both in vitro (Komori et al., 1993; Suganuma et al., 1996) and in vivo (Robertson et al., 1996; Moore et al., 1999; Suganuma et al., 1999). Although some links have been made between TNF-a and the activation of PKCa and AP-1, it remains unclear whether these events are linked to TNF-a-mediated tumour promotion by TPA. The current investigation was designed to identify the signalling molecules acting downstream of TNF-a in mouse skin tumour promotion. Data obtained in this study suggest that temporal differences in the activation of key TPA-responsive signalling molecules are responsible for the resistance of TNF-a7/7 mice to skin tumour development. Our results show for the first time that the pro-inflammatory cytokine TNF-a is a critical mediator of tumour promotion via a PKCa- and AP-1-dependent mechanism. Results TNF-a is not involved in initiation of epidermal keratinocytes by DMBA TNF-a may affect the metabolic activation of initiating carcinogens, such as DMBA, by keratinocytes since TNF-a regulated expression of the cytochrome P450 family members has been shown (Piscaglia et al., 1999). Therefore, TNF-a may modulate the frequency or site of DMBA-induced DNA-adduct formation (Pazzaglia et al., 2001), which normally causes an activating mutation in c-Ha-ras (Balmain et al., 1984). Analysis of DMBA-treated epidermis revealed the presence of three major adducts in wild-type (wt) and TNF-a7/7 samples, indicating that at least three distinct DMBA – DNA binding complexes were formed. The total adduct levels were not significantly different in wt and TNF-a7/7 epidermis (Figure 1a(i);

P=0.4456). Moreover, individual levels of the three main adducts were not significantly different in wt compared with TNF-a7/7 epidermis (Figure 1a(ii); adduct 1, P=0.5957; adduct 2, P=0.2059; adduct 3, P=0.1781), indicating that site-specific DMBA – DNA binding was the same in wt and TNF-a7/7 epidermis. DNA-adducts were not detectable in acetone treated epidermis from wt and TNF-a7/7 mice (data not shown). This indicates that putative regulation of cytochrome P450 by TNF-a is not critical for DMBA – DNA-adduct formation in skin and thus does not alter the frequency of mutations formed in skin after DMBA treatment. To investigate this further, papillomas resulting from DMBA/TPA-treatment of wt and TNF-a7/7 mouse skin were analysed for the presence of the characteristic c-Ha-ras gene mutation (A?T transversion in codon 61), thought to be the key event in tumour initiation (Balmain et al., 1984). Mutational analysis of the c-Ha-ras gene by PCR is possible since an XbaI restriction site is introduced into the c-Ha-ras PCR product if the codon 61 (A?T) mutation is present (Chakravarti et al., 1998). All papillomas tested from both wt and TNF-a7/7 mice contained the codon 61 mutation in c-Ha-ras (Figure 1b), suggesting that TNF-a does not affect the frequency of codon 61 c-Ha-ras mutations by DMBA. Collectively, these data imply that TNF-a is not involved in the initiation of epidermal keratinocytes by DMBA. TNF-a plays a role in tumour promotion Work in this laboratory has shown that TNF-a7/7 mouse skin is resistant to two stage carcinogenesis using TPA as a tumour promotor (Moore et al., 1999). TNF-a7/7 mice developed only 5 – 10% the number of tumours developed by wt mice using DMBA/TPA (Moore et al., 1999). In addition, TPA treatment of wt mouse skin induces TNF-a expression in the basal layer of the epidermis but not the dermis (Moore et al., 1999), suggesting a role for TNF-a in keratinocyte tumour promotion. To better understand the role of TNF-a in tumour promotion, wt and TNF-a7/7 mice with DMBAinitiated skin were subjected to equivalent tumour promoting doses of TPA and the non-phorbol ester type promoting agent, mirex (Moser et al., 1992). TNFa7/7 mice were not resistant to tumour promotion by mirex (Figure 2). Papillomas were first observed 10 weeks after the start of mirex treatment and by 21 weeks 81.8% (9 of 11) of these mice had developed an average of 2.27 tumours per mouse (Figure 2). At 21 weeks, tumour multiplicity was not significantly different between mirex treated wt and TNF-a7/7 mice (Figure 2, P=0.65), providing further evidence that TNF-a does not play a role in tumour initiation by DMBA. In contrast, only 8.3% (1 of 12) of TNFa7/7 mice had developed tumours 21 weeks after the start of promotion with TPA giving a tumour multiplicity of 0.33 tumours per mouse. The difference in tumour multiplicity between mirex and TPA in TNFOncogene


4730

Figure 1 Levels of DMBA-induced DNA-adduct formation and c-Ha-ras mutations are similar in wt and TNF-a7/7 epidermis. DNA was obtained and adduct analysis or c-Ha-ras mutation analysis was performed (Materials and methods). (a) Total adduct levels observed in DMBA-treated wt and TNF-a7/7 epidermis. In wt and TNF-a7/7, n=5. Error bars represent mean+s.d. (b) Levels of each of the three main adducts found in DMBA-treated wt and TNF-a7/7 epidermis. In wt and TNF-a7/7, n=5. Error bars represent mean+s.d. (c) c-Ha-ras mutations in papillomas from wt and TNF-a7/7 mice is indicated by the presence of an XbaI restriction site in c-Ha-ras PCR products

a7/7 mice was significant at 21 weeks (P=0.0051). As previously reported, tumour multiplicity after TPA treatment was significantly different between wt and TNF-a7/7 mice from week 12 onwards (P50.0001; Moore et al., 1999). These results indicate that the tumour promoting effects of TNF-a are specific for phorbol ester-related events. A specific subset of TPA-responsive genes are differentially expressed in wt and TNF-a7/7 epidermis

Figure 2 TNF-a7/7 mice are resistant to papilloma formation by TPA but not by mirex. DMBA-initiated wt and TNF-a7/7 mouse skin was subjected to either TPA or mirex tumour promotion as described in Materials and methods. Papilloma formation was scored once a week (all groups, n=12) Oncogene

The expression of a number of genes recently implicated in tumour development was compared in wt and TNF-a7/7 epidermis with a view to determining potential candidate effectors of TNF-a signalling. Granulocyte/macrophage-colony stimulating factor (GM – CSF) production is enhanced by TNF-a (see Carballo and Blackshear, 2001) and GM – CSF overexpression increases the incidence of skin tumours after DMBA/TPA treatment (Mann et al., 2001). Whole cell extracts from wt and TNF-a7/7 epidermis were assayed for GM – CSF expression by ELISA. Acetone


4731

treated epidermis from both wt and TNF-a7/7 mice expressed low levels of GM – CSF (55 pg/mg total protein; data not shown). TPA-treatment induced a substantial increase in epidermal GM – CSF expression in wt epidermis, which was maximal 6 h after four applications (Figure 3a; &50-fold). However, the level of GM – CSF induction in TPA-treated TNF-a7/7 epidermis was markedly lower than in wt epidermis (Figure 3a; 10-fold in TNF-a7/7 compared with 50fold in wt), suggesting that TNF-a is an important

regulator of GM – CSF induction in epidermal keratinocytes. The matrix metalloproteinases (MMPs) have been strongly implicated in all stages of tumour progression (Chambers and Matrisian, 1997). Exogenous expression of MMPs in cell lines and in transgenic mice results in an enhancement of tumour growth, invasion and metastasis (Hulboy et al., 2001). MMP-9 and MMP-3 expression in epidermis is regulated by TNF-a (MacNaul et al., 1990; Han et al., 2001) and TPA (MacNaul et al., 1990; Hanemaaijer et al., 1993). MMP-9 protein expression was not detected in acetone treated epidermis from wt and TNF-a7/7 mice (Figure 3b, lanes 1,2). TPA induced the expression of MMP-9 in both wt and TNF-a7/7 epidermis. However, MMP9 levels were markedly lower in TNF-a7/7 than wt epidermis at 6 h (Figure 3b, lane 4 compared to lane 3; TNF-a7/7 66% of wt) and 24 h (Figure 3b, lane 6 compared to lane 5; TNF-a7/7 42% of wt). This observation was confirmed using real-time RT – PCR of epidermal mRNA (data not shown). TPA also stimulated the expression of MMP-3 in both wt and TNF-a7/7 epidermis when measured by real-time RT – PCR of epidermal RNA. Relative levels of MMP-3 expression were significantly higher in wt than TNF-a7/7 epidermis at both 8 h (2.4-fold, P=0.0015) and 24 h (3.5-fold, P=0.0201) after TPA treatment (Figure 3c). Thus, TNF-a appears to play a role in the induction of certain MMPs during tumour promotion. AP-1 activity is differentially regulated in wt and TNF-a7/7 epidermis

Figure 3 Several TPA-responsive genes are expressed to a lesser extent in TNF-a7/7 epidermis compared to wt epidermis. Shaved dorsal mouse skin was treated with either one or four applications of TPA (4 mg; Material and methods). (a) Whole cell epidermal extracts were made 6 h after four applications of TPA or acetone. Samples were assayed for GM – CSF by ELISA (Materials and methods). Graph illustrates expression levels above the acetone control. Results are expressed as mean+s.e., n=2. (b) Equal amounts of whole cell epidermal extract from skin treated with four applications of acetone/TPA were analysed for MMP-9 expression by Zymography. Results are representative of two independent experiments. (c) mRNA was extracted from wt and TNFa7/7 skin at times indicated after one application of TPA and subjected to real-time RT – PCR for MMP-3. Error bars represent mean+s.d., n=3

Since the proteins identified as being differentially expressed in wt and TNF-a7/7 epidermis are all reported to be products of AP-1-responsive genes (reviewed by Angel et al., 2001), the regulation of AP-1 activity was examined in wt and TNF-a7/7 epidermis. AP-1 DNA-binding activity in nuclear extracts from wt and TNF-a7/7 epidermis was measured by EMSA. There was a low level of AP-1 activity in acetonetreated epidermis in both wt and TNF-a7/7 mice (Figure 4a(i), lanes 1,2). Increased AP-1 DNA-binding activity was observed in both wt and TNF-a7/7 epidermis 24 h after TPA treatment, but this TPAstimulated activity was 2.4-fold greater in wt epidermis than TNF-a7/7 epidermis (Figure 4a(i), lane 3 compared to lane 4 and b(i)). EMSA was performed to determine the Jun family members present within the AP-1 complex. Supershifted AP-1 complexes were observed with anti-c-Jun and anti-JunB antibodies in TPA-treated epidermis (Figure 4a(ii), lanes 5 – 8). Importantly, the wt AP-1 complex contained twofold more c-Jun than TNF-a7/7 AP-1 (Figure 4a(ii), lane 5 compared to lane 6 and b(ii)). In addition, levels of JunB in the AP-1 complex were twofold higher in wt than TNF-a7/7 epidermis (Figure 4a(ii), lane 7 compared to lane 8 and b(ii)). Supershifted complexes were not detected with the anti-JunD antibody (Figure 4a(ii), lanes 9,10). The differences in AP-1 DNA Oncogene


4732

Figure 4 AP-1 DNA-binding activity and c-Jun expression is differentially regulated in wt and TNF-a7/7 epidermis. Shaved dorsal mouse skin was treated with one application of TPA (4 mg; Material and methods) and harvested at the time points shown. (a) EMSA for AP-1 was performed on nuclear epidermal extracts; c-Jun, JunB and JunD antibodies were used for supershift assays; Cold AP-1 oligonucleotide was used as a negative control. Two exposures of the same EMSA are shown: (i) illustrates TPA stimulated AP-1 activity; (ii) illustrates supershift complexes. Results are representative of multiple replicate samples. (b) Quantitation of EMSA shown in (a). Graphs illustrate (i) AP-1 activity and (ii) supershift observed with anti-c-Jun and anti-JunB antibodies, where 100% was taken to be the level in the wt condition in each case. (c) Equal quantities of epidermal nuclear extract were subjected to Western blotting against an anti-c-Jun antibody (top panel); Equal amounts of whole cell epidermal homogenate were immunoblotted against anti-phospho-Ser73-c-Jun antibody (bottom panel). (d) Quantitation of Western blots shown in (c) of nuclear cJun and phospho-Ser73-c-Jun expression (fold increase relative to acetone). Results are expressed as mean+s.e., n=2

binding activity and supershifted complexes containing c-Jun and JunB were observed in multiple replicate epidermal nuclear extracts (data not shown). These data suggest that AP-1 DNA-binding activity is differentially regulated in wt and TNF-a7/7 epidermis. Of the many proteins that constitute AP-1 transcription factor complexes, c-Jun in particular has been implicated in events leading to tumour promotion (Ben-Ari et al., 1992; Dong et al., 1997; Young et al., 1999; Li et al., 2000; Behrens et al., 2000). To investigate whether the differences in AP-1 activation observed in wt and TNF-a7/7 epidermis correlated with altered c-Jun expression, epidermal nuclear lysates were subjected to Western blotting against an anti-cJun antibody. Low levels of c-Jun protein were detected in nuclear extracts from acetone treated wt or TNF-a7/7 epidermis (Figure 4c, top panel, lanes 1,2). TPA treatment induced a substantial increase in levels of nuclear c-Jun after 24 h in wt epidermis Oncogene

(Figure 4c, top panel, lane 3 compared with lane 1, 2.7fold above acetone; Figure 4d) whereas c-Jun induction was markedly lower in TNF-a7/7 epidermis (Figure 4c, top panel, lane 4 compared with lane 2, 1.5-fold above acetone; Figure 4d). c-Jun transcriptional activation is regulated by a variety of post-translational modifications, the most important of which is thought to be the phosphorylation of Ser63 and Ser73 within the N-terminus of c-Jun (see Vogt, 2001). This is necessary for skin tumour development (Behrens et al., 2000). Thus, the phosphorylation status of c-Jun protein was examined in wt and TNF-a7/7 epidermis. Whole cell epidermal lysates were probed with antibodies against phospho-Ser63-cJun and phospho-Ser73-c-Jun. Levels of phosphoSer63-c-Jun in these samples were not within the limit of detection for the antibody (data not shown). However, measurable levels of phospho-Ser73-c-Jun were observed in TPA-treated wt and TNF-a7/7


4733

epidermis, but levels were markedly lower in TNF-a7/7 epidermis (Figure 4c, lower panel, lane 4 compared with lane 3 and d). To investigate whether the c-Jun kinases (JNKs) were differentially regulated in wt and TNF-a7/7 epidermis, the same samples were probed with an anti-phospho-p46/p54 JNK antibody. Similar levels of phospho-JNK were observed between wt and TNF-a7/7 samples (data not shown), indicating that the higher level of phospho-Ser73-c-Jun in wt compared to TNF-a7/7 epidermis reflects changes in nuclear c-Jun protein expression rather than elevated JNK activity. Thus, TNF-a appears to mediate in TPA-induced tumour promotion by modulating the expression and phosphorylation of c-Jun, which in turn causes an alteration in AP-1 activity. This differential AP-1 regulation may account for the differential expression of MMP-9, MMP-3 and GM – CSF observed between wt and TNF-a7/7 mice. TNF-a regulates TPA-mediated tumour promotion through a PKCa-dependent pathway PKC is the proximal target of TPA-mediated signal transduction (Dekker and Parker, 1994) and AP-1 is a known target of PKC in skin (Angel et al., 2001). Therefore, PKC activation was examined in wt and TNF-a7/7 epidermis. Treatment of mouse skin with TPA is known to result in an initial stimulation of PKC activity followed by downregulation of the protein, which is maintained for several days (Fournier and Murray, 1987; Hansen et al., 1990). It is thought that downregulation of PKC is a critical event in tumour promotion (Hansen et al., 1990; Mills et al., 1992). PKC activation and downregulation can be detected by measuring intracellular translocation from the cytosol to the membrane or cytoskeleton (Dempsey et al., 2000). In the present study, downregulation of the PKCa isoform was investigated since it is the major PKC isoform in basal keratinocytes (Wang and Smart, 1999; Jansen et al., 2001). PKCa levels in acetone treated epidermis were the same in wt and TNF-a7/7 mice, suggesting that TNFa does not regulate basal levels of PKCa protein in skin. TPA-treatment resulted in the loss of a majority of cytosolic PKCa in wt epidermis within 24 h (Figure 5, lane 3; 87% decrease relative to acetone control). However, in TNF-a7/7 epidermis a substantial level of PKCa remained in the cytosolic fraction at this time (Figure 5, lane 4; 43% remaining relative to acetone control), indicative of a lesser degree of PKCa activation. This correlated with the lower level of AP-1 activity seen in TNF-a7/7 epidermis relative to wt epidermis 24 h after TPA treatment. To determine whether the decreased PKCa activity in TNF-a7/7 epidermis was sustained after 24 h, cytosolic and particulate levels of PKCa were examined 48, 72 and 96 h after TPA treatment. Cytosolic PKCa levels had started to recover in wt epidermis 48 h after TPA treatment (Figure 5, lane 5; levels were 46% of acetone control). In contrast, a further loss of cytosolic PKCa

Figure 5 TNF-a regulates TPA-mediated PKCa downregulation. Shaved dorsal mouse skin was treated with one application of TPA (4 mg; Material and methods) and harvested at the time points shown. Epidermal cytosolic and particulate homogenates were subjected to Western blotting against an anti-PKCa antibody (lanes 1 – 10) or anti-phospho-Serine/Threonine Phenylalanine antibody (lanes 11,12). Results are representative of three independent experiments

in TNF-a7/7 epidermis was observed relative to the level present at 24 h (Figure 5, lane 6; 12.3% compared with 43% respectively). This suggested the existence of a delay in PKCa activation in TNF-a7/7 epidermis. PKCa expression increased back to control levels at a similar rate in wt and TNF-a7/7 epidermis 72 and 96 h after TPA-treatment (Figure 5, lanes 7 – 10). This time course of recovery is in agreement with data published previously (Fournier and Murray, 1987). It was unclear whether the difference in PKCa protein levels in wt and TNF-a7/7 epidermis both 24 and 48 h after TPA treatment was due to delayed translocation and downregulation of PKCa or due to more rapid de novo synthesis of PKCa. Real time RT – PCR for PKCa at each of these time points after TPA treatment revealed no major differences in PKCa mRNA expression in wt and TNF-a7/7 epidermis (data not shown), suggesting that TNF-a does not regulate de novo expression of PKCa. Multisite phosphorylation of PKCa is an essential event for its translocation and subsequent activation (reviewed by Dempsey et al., 2000). Importantly, phosphorylation of Thr638 in PKCa (an autophosphorylation site which regulates the duration of activation of the enzyme; see Bornancin and Parker, 1996) was not markedly different in acetone treated skin from wt and TNF-a7/7 mice (Figure 5, lanes 11,12), indicating that TNF-a does not regulate PKCa activity through altering its phosphorylation status. The temporal shift in PKCa activation is reflected in the kinetics of AP-1 activation To investigate whether the observed delay in signalling through PKCa in TNF-a7/7 epidermis correlated with a temporal shift in AP-1 activation, AP-1 DNAbinding activity and nuclear c-Jun expression were measured in wt and TNF-a7/7 epidermis 48 and 72 h after TPA treatment. EMSA of nuclear extracts revealed that a higher level of AP-1 activity was present in TNF-a7/7 epidermis relative to wt epidermis 48 h after TPA-treatment (Figure 6a, lane 2 compared Oncogene


4734

with lane 1). However, AP-1 activity had declined substantially and was not markedly different in wt and TNF-a7/7 epidermis 72 h after TPA treatment (Figure 6a, lanes 3,4). The differential AP-1 activity seen in wt and TNF-a7/7 epidermis 48 h after TPA treatment appeared to be reflected by the pattern of nuclear c-Jun expression, since levels of c-Jun expression were higher in TNF-a7/7 epidermis compared to wt epidermis (Figure 6b, lane 2 compared with lane 1 and c). The level of nuclear c-Jun expression in wt epidermis had declined substantially between 24 and 48 h after TPA treatment (Figure 6c). In contrast, nuclear c-Jun expression in TNF-a7/7 epidermis was higher at 48 h relative to 24 h (Figure 6c). Expression of nuclear cJun had declined to basal levels in both wt and TNFa7/7 epidermis by 72 h after TPA-treatment (Figure 6b, lanes 3,4 and c).

Figure 6 The TPA-induced increase in AP-1 DNA-binding activity and c-Jun expression is delayed in TNF-a7/7 epidermis compared to wt epidermis. Shaved dorsal mouse skin was treated with one application of TPA (4 mg; Material and methods) and harvested at the time points shown. (a) EMSA for AP-1 was performed on nuclear epidermal extracts. (b) Equal quantities of epidermal nuclear homogenate were subjected to Western blotting against an anti-c-Jun antibody. (c) Quantitation of nuclear c-Jun expression at a variety of time points after TPA treatment as measured by Western blotting. Results are expressed as mean+s.e., n=2 Oncogene

These results suggest that the kinetics of TPAstimulated AP-1 activation and c-Jun expression are slower in TNF-a7/7 mice than wt mice. Importantly, this delay in AP-1 activation appears to correlate directly with the delay in PKCa activation seen in TNF-a7/7 mice. The possibility of a direct link between TNF-aregulated PKCa and TPA-induced AP-1 activation was investigated. For this, PKCa protein was depleted from epidermis to differing extents and the effect on TPAinduced c-Jun expression and phosphorylation was measured. TNF-a7/7 and wt mice were re-treated with acetone or TPA 48 h after the initial TPA treatment (a time point at which markedly different levels of PKCa were present; Figure 5) and levels of epidermal c-Jun and phospho-Ser73-c-Jun were measured 2 h later (a time point where some degree of c-Jun and phosphoSer73-c-Jun is normally present). TNF-a7/7 mice treated with acetone 48 h after the initial TPA treatment displayed higher levels of epidermal c-Jun expression compared to wt mice (Figure 7, upper panel, lane 2 compared with lane 1). This is in agreement with earlier observations (see Figure 6b, lanes 1,2). Wt mice re-treated with TPA 48 h after the initial TPA treatment showed a marked induction of epidermal c-Jun expression after 2 h, when compared to the acetone control (Figure 7, upper panel, lane 3 compared with lane 1). In contrast, re-stimulation of TNF-a7/7 epidermis with TPA caused a minimal increase in c-Jun expression after 2 h (Figure 7, upper panel, lane 4 compared with lane 2). In addition, the level of phospho – Ser73-c-Jun was higher in wt epidermis than TNF-a7/7 epidermis 2 h after the second application of TPA (Figure 7 lower panel, lane 3 compared with lane 4). This result was important since it demonstrated that TNF-a-mediated regulation of PKCa has the capacity to compromise AP-1 activity. Taken together, these data suggest the involvement of TNF-a in the regulation of TPA-stimulated PKCa activity. A lack of TNF-a appears to cause a temporal

Figure 7 PKCa depletion alters TPA-induced c-Jun expression and phosphorylation. Shaved dorsal mouse skin was treated with one application of TPA (4 mg; Material and methods). A second application of TPA (as indicated) was applied to mouse skin 48 h after the initial TPA treatment. Whole cell epidermal extracts were made from skin taken 2 h later and were normalized for protein concentration. Western blots were probed with anti-cJun and anti-phospho-Ser73-c-Jun antibodies


4735

disruption in epidermal PKCa activation that is reflected by differential kinetics of AP-1 activation. This asynchronous signalling phenomenon appears to be translated into changes in the gene expression profile of wt and TNF-a7/7 epidermis. The alteration in expression of TPA-target genes may be responsible for resistance of TNF-a7/7 mice to carcinogenesis. Discussion Carcinogenesis in humans and laboratory animals is a complex process comprising at least three stages, initiation, promotion and progression. The current study demonstrates that TNF-a does not play a role in initiation but can mediate tumour promotion by modulating the expression of a specific subset of AP1-responsive genes. The lack of induction of these AP1-target genes in response to TPA appeared to be due to a temporal delay in the activation of PKCa and AP1 transcription factor complexes. Spatial and temporal regulation of signal transduction is essential for determining the speed and precision by which signalling events occur (see Schechtman and Mochly-Rosen, 2001). This regulation also provides an effective way by which different outcomes can be obtained from the activation of signalling pathways involving the same effector molecules. Results from the present work demonstrate how temporal changes in the activity of certain signal transduction pathways may be as critical for tumour formation as constitutive activation or repression. PKC activation is an event capable of regulating a variety of biological responses ranging from cell growth to cell death. In this study, the normal response of PKCa to TPA (i.e. translocation and downregulation) appeared to be substantially delayed in TNF-a7/7 epidermis when compared to wt epidermis, although the overall extent of PKCa activation by TPA appeared to be similar. In agreement with this observation, stimulation of PKCa translocation to the membrane by TNF-a has been demonstrated in L929 cells (Lee et al., 2000). Regulation of PKCa is known to reflect the existence of a complex array of cross-talk between multiple signalling pathways (Lee et al., 2000). At present, the precise mechanism by which deficiency of TNF-a affects TPA-stimulated PKCa activation remains unclear. However, it is most likely due to metabolic changes within the epidermis (e.g. metabolism of TPA) or regulation of the cellular machinery involved in translocation, since TNF-a deficiency does not affect PKCa phosphorylation status or de novo synthesis of PKCa. c-Jun (as opposed to other AP-1 component proteins) has been described as a major player in the induction of AP-1 transcriptional activity required for TPA-mediated tumour promotion (Ben-Ari et al., 1992; Dong et al., 1994; Young et al., 1999; Behrens et al., 2000). Moreover, impaired tumour formation has been reported in mice expressing a c-Jun mutant which lacks

the phosphoacceptor sites for JNK (i.e. Ser63 and Ser73; Behrens et al., 2000). In the current study, a delay in both TPA-stimulated AP-1 activation and cJun expression was observed in TNF-a7/7 epidermis, which was concomitant with the temporal shift in PKCa activation detected. In addition, the level of TPA-stimulated phospho-Ser73-c-Jun was markedly reduced in TNF-a7/7 compared to wt epidermis. PKCa appeared to act as a regulator of AP-1 activity in this model, since lower levels of PKCa in skin correlated with a lesser degree of TPA-stimulated c-Jun phosphorylation. There is now much evidence from loss of function approaches in mice that different AP-1 complexes have specific and non-overlapping cellular functions (see Szabowski et al., 2000 and references therein). This has been shown for c-Jun and JunB, which are reported to act antagonistically in certain cell types to control cell transformation, differentiation and expression of AP-1 target genes (see Angel et al., 2001). Results from the present study suggested that both c-Jun and JunB levels were reduced in the AP-1 complex in TNF-a7/7 compared to wt epidermis. Given the reported antagonistic action of c-Jun and JunB, this could be envisaged to cancel out any differential effect of AP-1 signalling in wt and TNF-a7/7 epidermis. However, this seems unlikely since a study examining the distribution of Fos and Jun family members in skin has shown that c-Jun and JunB are expressed within different cellular layers of the epidermis (Rutberg et al., 1996). Thus, specific AP-1 complexes with potentially different target-genes appear to be differentially regulated in a time-dependent fashion in TPA-treated wt and TNF-a7/7 epidermis. The altered activation kinetics of PKCa and AP-1 in TNF-a7/7 epidermis (rather than a change in extent of activation) appeared to be sufficient to cause an alteration in the gene expression profile in wt and TNF-a7/7 skin. Indeed, a number of known AP-1target genes (GM – CSF, MMP-9 and MMP-3) were induced to a lesser extent in TNF-a7/7 epidermis compared to wt epidermis after TPA-treatment. Very few genes known to be regulated by Jun have been shown to induce a change in oncogenic phenotype when they are up- or down-regulated by themselves (Vogt, 2001). Thus, reduced GM – CSF expression in TNF-a7/7 epidermis is especially interesting since GM – CSF has been shown to act as an endogenous tumour promotor when overexpressed in mouse skin (Mann et al., 2001), through a c-Jun dependent mechanism (Szabowski et al., 2000). Moreover, altered MMP-3 and MMP-9 expression between wt and TNFa7/7 mice is important since exogenous expression of MMPs in cell lines and transgenic mice has been shown to result in an enhancement of tumour growth, invasion and metastasis (Hulboy et al., 2001). In the present study, the expression of many other genes (including a number of cytokines, cytokine receptors, chemokines and cyclo-oxygenase-2) was examined in wt and TNF-a7/7 epidermis and these did not appear to be differentially regulated (data not shown). Thus, it Oncogene


4736

is likely that the expression of a specific subset of TPAresponsive genes is essential for endogenous TNF-a to act as a tumour promotor. The non-phorbol ester type tumour promotor mirex is a potent tumour promotor in mouse skin that does not stimulate PKC activity (Moser et al., 1992). Since TNF-a7/7 mice were not resistant to tumour promotion by mirex, this further substantiates a role for PKCa in TNFa-mediated tumour promotion. In summary, these data strongly suggest that temporal differences in the activation of key TPAresponsive signalling molecules are responsible for the resistance of TNF-a7/7 mice to skin tumour development. Our results show for the first time that the proinflammatory cytokine TNF-a is a critical mediator of tumour promotion via a PKCa- and AP-1-dependent pathway. This may be one mechanism by which chronic inflammation increases susceptibility to cancer. Procedures that modulate endogenous levels of TNF-a (or its effector molecules) may have potential utility in cancer therapy. Materials and methods Mice All mice were housed in negative-pressure isolation at the ICRF containment facility. Mice homozygous for the mutant TNF-a knockout allele (TNF-a7/7 mice) and wild-type (wt) mice were maintained on a BALB/c background. In each experiment, wt and TNF-a7/7 female mice were age matched to within 3 days.

Preparation of skin homogenates Mice were killed at specified times after treatment (stated in Figure legends) and skin was removed, spread on card and snap frozen in liquid nitrogen. Skin from two mice was used to make each sample. Homogenates were prepared as described below. Whole cell epidermal homogenates The epidermis was scraped from the dermis using a surgical scalpel and placed in ice-cold cell lysis buffer (1% (v/v) Triton-X100, 0.5% (w/v) sodium deoxycholate and 0.1% (w/v) SDS in PBS). Samples were homogenized using an Ultra-Turrax T25 homogenizer and incubated on ice for 10 min. Undigested tissue was removed by centrifugation at 15 000 6 g for 10 min and the supernatant was stored at 7708C. Preparation of cytosolic and nuclear epidermal homogenates

A 1 – 2 cm2 area of dorsal skin was clipped with electric clippers when mice reached 6 weeks of age. Mice were initiated with a topical application of 9,10-dimethyl-1,2benzanthracene (DMBA; 25 mg in 100 ml acetone; Sigma) at 7 weeks of age. After 1 week, mice received 12-OTetradecanoylphorbol 13-acetate (TPA; 4 mg in 100 ml acetone; Sigma) twice weekly for 15 weeks or mirex (200 nmol in acetone; Fischer Chemicals) 3 times weekly for 20 weeks. Control mice were painted with acetone alone. Tumours were recorded weekly and were defined as raised lesions of a minimum diameter of 1 mm that had been present for at least 2 weeks (Bogovski, 1994). For statistical analysis the Mann-Whitney (two-tailed U-test) was used.

Cytosolic and nuclear fractions were prepared as described by Andrews and Faller (1991). Briefly, epidermis was scraped from the dermis using a surgical scalpel and homogenized in ice-cold cytosolic lysis buffer (10 mM HEPES – KOH (pH 7.9; 48C), 10 mM KCl, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 0.5 mM dithiothreitol (DTT), 1 mM b-glycerophosphate, 1 mM sodium orthovanadate, 25 mg/ml aprotinin, 50 mg/ml leupeptin, 25 mg/ml pepstatin A, 1 mM phenylmethylsulfonylfluoride (PMSF)) using an Ultra-Turrax T8 homogeniser. Homogenates were centrifuged (30 000 6 g; 15 min; 48C) and the supernatant (cytosolic fraction) stored at 7708C. The pellet was resuspended in ice-cold nuclear lysis buffer (20 mM HEPES – KOH (pH 7.9; 48C), 25% (v/v) glycerol, 420 mM NaCl, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 0.5 mM DTT, 1 mM b-glycerophosphate, 1 mM sodium orthovanadate, 25 mg/ml aprotinin, 50 mg/ml leupeptin, 25 mg/ml pepstatin A, 1 mM PMSF) and incubated on ice for 20 min for high-salt extraction. Cellular debris was removed by centrifugation (14 000 r.p.m., 15 min, 48C) and the supernatant (nuclear fraction) stored at 7708C.

DNA adduct analysis in skin

Preparation of cytosolic and particulate epidermal homogenates

Tumour induction in mouse skin

Mice were initiated with DMBA or treated with acetone at 7 weeks of age. Mice were killed 24 h after treatment and skins were snap frozen in liquid nitrogen. Epidermis was scraped from the dermis and epidermal DNA was extracted by standard procedures. The presence of DNA adducts was determined by 32 P-post-labelling analysis using the nuclease P1 enrichment procedure described by Reddy and Randerath (1986). For statistical analysis an unpaired two-tailed t-test was used.

Oncogene

overnight), phenol/chloroform extraction and ethanol precipitation. For each tumour, PCR amplification of exons 1 and 2 of mouse c-Ha-ras was performed using primers MRF and MRR as described by Chakravarti et al. (1998). The PCR products were subjected to agarose gel electrophoresis and excised bands were purified using a QIAquick gel extraction kit (Qiagen Inc). A restriction digestion with XbaI (New England Biolabs) was then performed to determine the presence of an A?T transversion in codon 61 of c-Ha-ras.

Cytosolic and particulate homogenates were obtained from intact epidermis as described by Goodell et al. (1996). Estimation of protein concentration For cytosolic, nuclear and particulate extracts the Bio-Rad protein assay, based on the method of Bradford (1976), was used. For whole cell homogenates, protein concentration was determined using the BCA protein assay kit (Sigma, UK).

Analysis of c-Ha-ras codon 61 mutation in papillomas

SDS – PAGE and Western blotting

Genomic DNA was isolated from frozen papillomas. Briefly, tissue was subjected to proteinase K digestion (1 mg/ml, 558C

Normalized epidermal homogenates were prepared by boiling with one quarter volume of 56 concentrated Laemmli


4737

sample buffer (1008C, 5 min). Proteins were resolved on a 10% denaturing gel (Dual Mini Slab kit; Atto) using 16 running buffer (25 mM Tris; 192 mM glycine, 0.1% (w/v) SDS). See-Blue markers (Invitrogen Ltd, UK) were used to determine protein size. Protein was transferred to nitrocellulose (Hybond-ECL, Amersham, UK). Blots were blocked and antibody incubations performed according to the manufacturers protocol. Proteins were visualised using ECLTM reagent (Amersham, UK). Antibodies The anti-PKCa polyclonal antibody was purchased from Sigma, UK. Anti-c-Jun (sc-1694), anti-JunB (sc-46-G) and anti-JunD (sc-74-G) were purchased from SantaCruz Biotechnology, Inc. The anti-phospho-Ser73-c-Jun (Cat. No. 9162) and anti-Phospho-Serine/Threonine Phenylalanine (Cat. No. 9631) antibodies were purchased from New England Biolabs, MA, USA. Assay of AP-1 DNA-binding activity in vivo The AP-1 binding oligonucleotide probe, 5’-CGC TTG ATG AGT CAG CCG GAA-3’ was end-labelled with [g-32P]ATP according to standard procedures. For electrophoretic mobility shift assays binding reactions were performed in a total volume of 20 ml in 40 mM HEPES pH 7.9, 2 mM magnesium chloride, 2 mM DTT, 100 mM NaCl, 50 mg/ml polydeoxyinosine-deoxycytidine (Pharmacia). 0.2 mCi of 32Plabelled oligonucleotide and 3 mg of nuclear extract was used in each reaction. Reactions were performed for 30 min at room temperature. Duplicate samples were set up containing 506 excess of cold oligonucleotide as a competitor. The DNA-protein complexes were resolved on a 4% acrylamide gel. DNA binding was visualized by autoradiography. Measurement of MMP-3 mRNA by real time RT – PCR Skin was treated and mice killed at specified times (as described in figure legends). The dorsal shaved skin was removed and agitated in warm water (608C) for 5 s. Skin was then spread on card and the epidermis scraped from the dermis using a spatula. Epidermis from two mice was pooled for each sample. RNA was extracted and purified from epidermal samples using solution D as described previously (Chomczynski and Sacchi, 1987). RNA was DNase treated with 10U DNase (Pharmacia) following manufacturer’s instructions. DNase treated RNA (2 mg) was reverse transcribed with M-MLV reverse transcriptase (Promega) according to manufacturer’s instructions and diluted to 100 ml with nucleotide-free water. Primers and probe for mouse MMP-3 were designed using Primer Express 1.5a (PE Applied Biosystems, Warrington, UK) (Forward 5’-TCCTGATGTTGGTGGCTTCA-3’; Reverse 5’-TCCTGTAGGTGATGTGGGATTTC-3’ and probe 5’-CCTTCCCAGGTTCGCCAAAATGGA-3’). Multiplex real-time RT – PCR analysis was performed using MMP-3 (FAM) and predeveloped 18s rRNA (VIC) with the ABI PRISM 7700

Sequence Detection System instrument and software (PE Applied Biosystems). PCR was carried out with the TaqMan Universal PCR Master Mix (PE Applied Biosystems) using 2.5 ml of cDNA in a 25 ml final reaction mixture. The cycling conditions were an incubation at 508C for 2 min, followed by 10 min at 958C and 40 cycles of 15 s at 958C, and 1 min at 608C. Experiments were performed in triplicate for each sample. MMP-3 was normalised (DCt) to 18s by subtracting the cycle threshold (DCt) value of 18s from the DCt value of MMP-3. Average DCt values for TPA treated TNF-a7/7 epidermis were compared with wt DCt i.e. DDCt=DCt [WT] DCt [TNF-a7/7]. Fold difference compared with TNF-a7/7 was calculated by 2-DDCt. A paired two-tailed t-test was used for statistical analysis. Quantitation of blots To determine the intensity of bands on Western blots and autoradiographs, they were scanned using an Epson scanner and appropriate software. The image generated was then quantified using NIH image software. Protein determination by ELISA Protein levels of GM – CSF in whole cell epidermal homogenates were determined by ELISA (R&D systems). Samples were assayed at 1 mg/ml total protein. Cytokine levels were corrected to pg/mg total epidermal protein. Zymography for MMP-9 activity Zymography was carried out as described previously (Leber and Balkwill, 1997). Briefly, 10 mg epidermal whole cell protein was loaded into each well with non-reducing sample buffer. Electrophoresis was performed with a 11% polyacrylamide gel co-polymerized with 1.2 mg/ml gelatin at 220 V for 3.5 h. To remove the SDS and allow the MMPs to renature, the gel was incubated in 2.5% (v/v) Triton-X-100 for 1 h at room temperature on an orbital shaker. The gel was subsequently incubated in low salt collagenase buffer (50 mM Tris-HCl pH 7.6, 0.2 M NaCl, 5 mM CaCl2, 0.02% (v/v) Brij-35) at 378C overnight. To reveal the lysis zones, gels were stained in stain/destain solution for 2 h. Protein identity was confirmed using an MMP-9 standard (R&D systems, Oxon, UK).

Acknowledgements The authors would like to thank George Kollias and Manolis Pasparakis for originally supplying us with TNFa7/7 mice; Nick East and Hazel Holdsworth for technical assistance; Stephen Robinson and Lala Khalique for help with Real-time RT – PCR; and Fiona Watt for advice with the manuscript. DM Owens was supported by the National Research Service Award CA75638 from the National Cancer Institute. Alan Hewer and David Phillips were supported by the Cancer Research Campaign.

References Andrews NC and Faller DV. (1991). Nucleic Acids Res., 19, 2499. Angel P, Szabowski A and Schorpp-Kistner M. (2001). Oncogene, 20, 2413 – 2423. Balkwill F. (2002). Cytokine Growth Factor Rev., 13, 135 – 141.

Balkwill F and Mantovani A. (2001). Lancet, 357, 539 – 545. Balmain A, Ramsden M, Bowden GT and Smith J. (1984). Nature, 307, 658 – 660. Baud V and Karin M. (2001). Trends Cell Biol., 11, 372 – 377.

Oncogene


4738

Behrens A, Jochum W, Sibilia M and Wagner EF. (2000). Oncogene, 19, 2657 – 2663. Ben-Ari ET, Bernstein LR and Colburn NH. (1992). Mol. Carcinog., 5, 62 – 74. Bernstein LR and Colburn NH. (1989). Science, 244, 566 – 569. Bogovski P. (1994). IARC Sci. Publ., 111, 1 – 45. Bornancin F and Parker PJ. (1996). Curr. Biol., 6, 1114 – 1123. Bradford MM. (1976). Anal. Biochem., 72, 248 – 254. Carballo E and Blackshear PJ. (2001). Blood, 98, 2389 – 2395. Chakravarti D, Mailander P, Franzen J, Higginbotham S, Cavalieri EL and Rogan EG. (1998). Oncogene, 16, 3203 – 3210. Chambers AF and Matrisian LM. (1997). J. Natl. Cancer Inst., 89, 1260 – 1270. Chen C, Chou C, Sun Y and Huang W. (2001). Cell Signal, 13, 543 – 553. Chen CC, Sun YT, Chen JJ and Chiu KT. (2000). J. Immunol., 165, 2719 – 2728. Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162, 156 – 159. Dekker LV and Parker PJ. (1994). Trends Biochem. Sci., 19, 73 – 77. Dempsey EC, Newton AC, Mochly-Rosen D, Fields AP, Reyland ME, Insel PA and Messing RO. (2000). Am. J. Physiol. Lung Cell. Mol. Physiol., 279, L429 – L438. Dlugosz AA, Cheng C, Williams EK, Dharia AG, Denning MF and Yuspa SH. (1994). Cancer Res., 54, 6413 – 6420. Domann FE, Levy JP, Birrer MJ and Bowden GT. (1994). Cell Growth Differ., 5, 9 – 16. Dong Z, Birrer MJ, Watts RG, Matrisian LM and Colburn NH. (1994). Proc. Natl. Acad. Sci. USA, 91, 609 – 613. Dong Z, Crawford HC, Lavrovsky V, Taub D, Watts R, Matrisian LM and Colburn NH. (1997). Mol. Carcinog., 19, 204 – 212. Fournier A and Murray AW. (1987). Nature, 330, 767 – 769. Goodell AL, Oh HS, Meyer SA and Smart RC. (1996). J. Invest. Dermatol., 107, 354 – 359. Han YP, Tuan TL, Hughes M, Wu H and Garner WL. (2001). J. Biol. Chem., 276, 22341 – 22350. Hanemaaijer R, Koolwijk P, le Clercq L, de Vree WJ and van Hinsbergh VW. (1993). Biochem. J., 296, 803 – 809. Hansen LA, Monteiro-Riviere NA and Smart RC. (1990). Cancer Res., 50, 5740 – 5745. Hulboy DL, Matrisian LM and Crawford HC. (2001). Mol. Cell Biol., 21, 5478 – 5487. Jansen AP, Dreckschmidt NE, Verwiebe EG, Wheeler DL, Oberley TD and Verma AK. (2001). Int. J. Cancer, 93, 635 – 643. Kikkawa U, Takai Y, Tanaka Y, Miyake R and Nishizuka Y. (1983). J. Biol. Chem., 258, 11442 – 11445. Komori A, Yatsunami J, Suganuma M, Okabe S, Abe S, Sakai A, Sasaki K and Fujiki H. (1993). Cancer Res., 53, 1982 – 1985. Leber TM and Balkwill FR. (1997). Anal. Biochem., 249, 24 – 28.

Oncogene

Lee JY, Hannun YA and Obeid LM. (2000). J. Biol. Chem., 275, 29290 – 29298. Lee YS, Dlugosz AA, McKay R, Dean NM and Yuspa SH. (1997). Mol. Carcinog., 18, 44 – 53. Lejeune FJ, Ruegg C and Lienard D. (1998). Curr. Opin. Immunol., 10, 573 – 580. Li JJ, Cao Y, Young MR and Colburn NH. (2000). Mol. Carcinog., 29, 159 – 169. Locksley RM, Killeen N and Lenardo MJ. (2001). Cell, 104, 487 – 501. MacNaul KL, Chartrain N, Lark M, Tocci MJ and Hutchinson NI. (1990). J. Biol. Chem., 265, 17238 – 17245. Mann A, Breuhahn K, Schirmacher P, Wilhelmi A, Beyer C, Rosenau A, Ozbek S, Rose-John S and Blessing M. (2001). Cancer Res., 61, 2311 – 2319. Mills KJ, Bocckino SB, Burns DJ, Loomis CR and Smart RC. (1992). Carcinogenesis, 13, 1113 – 1120. Moore RJ, Owens DM, Stamp G, Arnott C, Burke F, East N, Holdsworth H, Turner L, Rollins B, Pasparakis M, Kollias G and Balkwill F. (1999). Nat. Med., 5, 828 – 831. Moser GJ, Meyer SA and Smart RC. (1992). Cancer Res. 631 – 636. Nishizuka Y. (1984). Nature, 308, 693 – 698. Pazzaglia S, Mancuso M, Primerano B, Rebessi S, Biozzi G, Covelli V and Saran A. (2001). Mol. Carcinog., 30, 111 – 118. Piscaglia F, Knittel T, Kobold D, Barnikol-Watanabe S, Di Rocco P and Ramadori G. (1999). Biochem. Pharmacol., 58, 157 – 165. Reddy MV and Randerath K. (1986). Carcinogenesis, 7, 1543 – 1551. Robertson FM, Ross MS, Tober KL, Long BW and Oberyszyn TM. (1996). Carcinogenesis, 17, 1719 – 1728. Rutberg SE, Saez E, Glick A, Dlugosz AA, Spiegelman BM and Yuspa SH. (1996). Oncogene, 13, 167 – 176. Saez E, Rutberg SE, Mueller E, Oppenheim H, Smoluk J, Yuspa SH and Spiegelman BM. (1995). Cell, 82, 721 – 732. Schechtman D and Mochly-Rosen D. (2001). Oncogene, 20, 6339 – 6347. Suganuma M, Okabe S, Marino MW, Sakai A, Sueoka E and Fujiki H. (1999). Cancer Res., 59, 4516 – 4518. Suganuma M, Okabe S, Sueoka E, Iida N, Komori A, Kim SJ and Fujiki H. (1996). Cancer Res., 56, 3711 – 3715. Szabowski A, Maas-Szabowski N, Andrecht S, Kolbus A, Schorpp-Kistner M, Fusenig NE and Angel P. (2000). Cell, 103, 745 – 755. Vogt PK. (2001). Oncogene, 20, 2365 – 2377. Wang HQ and Smart RC. (1999). J. Cell Sci., 112, 3497 – 3506. Young MR, Li JJ, Rincon M, Flavell RA, Sathyanarayana BK, Hunziker R and Colburn N. (1999). Proc. Natl. Acad. Sci. USA, 96, 9827 – 9832. Zhong S, Quealy JA, Bode AM, Nomura M, Kaji A, Ma WY and Dong Z. (2001). Cancer Res., 61, 4084 – 4091.