Raf-induced Transformation Requires an Interleukin 1 Autocrine Loop

It is known that human cancer reflects the loss of a proper balance between cell death and cell proliferation such that cancer cells often show a resistance to apoptosis and can proliferate under conditions in which normal cells cannot. It is thought that these phenotypes are caused by the accumulation of mutations in a variety of genes whose products normally play a role in the biochemical pathways that regulate cell death and cell proliferation. One of the key pathways controlling cellular proliferation identified to date is the MAPK²pathway. This pathway is known to be regulated by Raf, which phosphorylates MEK1 and MEK2, which then directly phosphorylate ERK1 and ERK2 (1, 2, 3). The ERKs then phosphorylate a variety of proteins, both in the cytosol and in the nucleus, which can regulate cell proliferation. Although the MAPK pathway appears to be critical for cell proliferation, we have found that the activation of MAPK by a constitutively active form of MEK is not sufficient to transform NIH 3T3 cells.³However, activated forms of both Ras and Raf are capable of transforming NIH 3T3 cells. This suggests that both Ras and Raf are capable of activating at least one pathway required for transformation that MEK cannot activate.

Besides MEK, other targets of Raf that have been identified previously include Cdc25A, pp70 S6K, and NF-κB, all of which are known to play key roles in cell proliferation (4, 5, 6, 7). However, with the exception of MEK, the relevance of these in Raf-induced transformation have not yet been explored. In this study, the dependence of Raf-induced transformation on NF-κB activation was explored using a dominant-negative form of IκB. NF-κB is a ubiquitously expressed transcription factor that consists of heterodimeric and homodimeric complexes (8). These complexes are sequestered in the cytoplasm by IκB, which is thought to function by masking the nuclear localization signal of NF-κB (9, 10). A pair of related kinases, IKKα and IKKβ, has been identified recently that phosphorylate IκBα on Ser-32 and Ser-36, thereby targeting IκB for ubiquitination and its subsequent degradation by the 26S proteasome, thus allowing NF-κB to enter the nucleus(11, 12, 13, 14, 15, 16). NF-κB has been shown to play a role in the regulation of a variety of genes, including those encoding various cytokines, cell adhesion molecules, antiapoptotic genes, and many others (8, 17).

One gene in particular, which is induced by NF-κB and could provide an explanation for the need for NF-κB in Raf-induced transformation,is IL-1β, as this factor is mitogenic for fibroblasts as well as a variety of other cell types (18, 19, 20). The IL-1 family consists of three known members: IL-1α, IL-1β, and IL-1Ra. Both IL-1α and β can bind to the p80 IL-1 receptor as agonists, whereas IL-1Ra functions as a pure receptor antagonist (19). Both IL-1α and IL-1β are synthesized as precursors that require cleavage to generate the mature forms (19). Caspase-1 (IL-1βconverting enzyme) has been shown to be the protease responsible for cleaving the precursor form of IL-1β, whereas calpain has been identified as a protease capable of cleaving the precursor form of IL-1α (21). Both the immature and the mature form of IL-1α can activate the IL-1 receptor, whereas only the mature form of IL-1β can activate the receptor (19, 22).

Although the nature of the signal transduction pathways triggered by activation of the IL-1R are ill-defined at this point and have not been completely delineated, the activation of NF-κB, c-Jun NH₂-terminal kinase, p38, PI3K, and sphingomyelinase have all been demonstrated (19, 23, 24, 25). The activation of NF-κB by IL-1 appears to involve the binding of IRAK to the IL-1R accessory protein MyD88 with the subsequent appearance of highly phosphorylated forms of IRAK (26, 27). IRAK then appears to associate with TRAF6, which itself binds to NIK, a kinase that activates IKKα, thereby bringing about the activation of NF-κB by IL-1 (28). The role of this pathway in transformation mediated by Raf has not been examined previously.

NIH 3T3 cells were cultured in DMEM (Life Technologies, Inc.)supplemented with 10% calf serum (Life Technologies, Inc.) and maintained in a humidified chamber with 5% CO₂at 37°C. Stable transfections of NIH 3T3 cells were done using the calcium phosphate method. Briefly, 1 × 10⁶ cells were plated in 100-mm dishes (Nunc) the day before transfection and fed with 10% calf serum 4–6 h before the transfection. NIH 3T3 genomic DNA was used in all transfections as carrier DNA. The precipitate was left on the cells for 16 h,followed by feeding the cells with fresh 10% calf serum. The cells were then split 1:4 on the following day into selective media: two plates with 10% calf serum supplemented with G418 (Life Technologies,Inc.) to a final concentration of 400 μg/ml for measuring transfection efficiency and analyzing protein expression and into two plates with 5% calf serum for focus formation assays. Transient transfections were done using Lipofectamine (Life Technologies, Inc.)according to the manufacturer’s instructions. Cells were serum-deprived in 0.5% calf serum for 48 h before lysis in 1×reporter lysis buffer (Promega). Luciferase assays were then performed according to the manufacturer’s instructions (Promega).

Murine IL-1Ra was removed as an EcoRI fragment from pBlueScript and cloned into the EcoRI site of pcI-neo(Promega) and checked for proper orientation. IL-1Ra was then removed from pcI-neo with NheI and NotI and cloned into pCEP4 (Invitrogen). pCEP4 CrmA was generated by removing the coding sequence from pcDNA3 with HindIII and XhoI. All other plasmids have been described previously (29).

To examine the levels of phosphorylated and total ERK,transiently transfected NIH 3T3 cells were lysed in boiling 1% SDS and 10 mm Tris (pH 7.5) and sonicated to reduce viscosity. Protein concentrations were measured with the D_Cprotein assay kit from Bio-Rad. Ten μg of protein were then subjected to SDS-PAGE electrophoresis, transferred to an Immobilon-P membrane(Millipore), blocked with 5% nonfat powdered milk, and then probed with anti-active ERK (Promega) and anti-ERK1 (Santa Cruz), which cross-reacts with ERK2. ECL (Amersham) was then performed according to the manufacturer’s protocol. To examine expression of RasV12, RafBXB,IκBΔN, and TRAF6Δ, pools of G418-resistant colonies of NIH 3T3 cells were lysed in boiling 1% SDS and 10 mm Tris (pH 7.5)and sonicated to reduce viscosity. Ten μg of protein were then subjected to SDS-PAGE electrophoresis, transferred to an Immobilon-P membrane (Millipore), blocked with 5% nonfat powdered milk, and then probed with the appropriate antibodies. Secondary antibodies coupled to horseradish peroxidase were purchased from Cappel. ECL (Amersham) was then performed according to the manufacturer’s protocol.

We sought to confirm previous work that had shown that Ras and Raf were capable of activating NF-κB-dependent gene expression before investigating the possible requirement for this activation in Ras- and Raf-induced transformation. Expression constructs of RasV12 (a constitutively active mutant of human H-Ras), RafBXB (human Raf-1 in which amino acids 26–302 have been deleted), and MEKΔ (a constitutively active form of human MEK containing S218E and S222D mutations and a deletion of amino acids 32–51) were transfected into NIH 3T3 cells along with a reporter construct, pBIIXluc, which contains two NF-κB binding sites from the immunoglobulin κ enhancer upstream of a minimal promoter and the luciferase gene (4, 29, 30, 31). As shown in Fig. 1, both RasV12 and RafBXB were capable of stimulating NF-κB-dependent gene expression by 7.1- and 6.3-fold, respectively. MEKΔ, however,did not induce NF-κB-dependent gene expression, suggesting that the activation of NF-κB by Ras and Raf was not mediated by the MAPK pathway. As shown in Fig. 1 B, RasV12, RafBXB, and MEKΔwere all equally capable of activating MAPK, as shown through the use of an antibody that specifically recognizes the activated (dually phosphorylated) forms of ERK1 and ERK2.

The specificity of the reporter construct was documented by cotransfecting RafBXB along with a dominant-negative form of IκB,IκBΔN, which lacks amino acids 1–36 (32). As shown in Fig. 2, dominant-negative IκB completely blocked the up-regulation of luciferase activity induced by RafBXB, confirming that expression of the reporter construct occurred through the activation of NF-κB.

Having confirmed that Raf can stimulate NF-κB, we next sought to determine whether Raf-induced transformation required NF-κB activation. To examine this possibility, NIH 3T3 cells were transfected with expression constructs for RasV12 or RafBXB alone or with an expression construct for dominant-negative IκBΔN. As shown in Figs. 3 and 4, RafBXB-induced but not RasV12-induced transformation was blocked by IκBΔN. To rule out the possibility that the decrease in the number of foci seen with RafBXB and IκBΔN related to decreased expression of RafBXB, Western blots were carried out on pooled lysates of G418-resistant colonies of cells. As shown in Fig. 4,C,RafBXB expression was similar. In addition, the number of G418-resistant colonies obtained from transfections with RafBXB alone or with IκBΔN were similar, ruling out the possibility that IκBΔN was toxic to the cells. We also determined whether IκBΔN expression affected the ability of RafBXB to activate MAPK and found that there was no effect (data not shown). Representative plates showing RasV12- and RafBXB-induced foci are shown in Fig. 3.

Because the activation of NF-κB by RafBXB was required for transformation, we sought to identify a possible mechanism accounting for the NF-κB requirement. Because NF-κB is known to regulate the expression of IL-1 and IL-1 is known to be mitogenic for fibroblasts,it seemed possible that an IL-1 autocrine loop could play a role in RafBXB-induced transformation. To examine the potential involvement of IL-1 in Raf-induced transformation, we coexpressed RasV12 or RafBXB alone or with an expression construct for CrmA, a cowpox viral protein that inhibits IL-1β converting enzyme (33). As shown in Fig. 5, RafBXB-induced but not RasV12-induced transformation was blocked by CrmA. To examine the involvement of IL-1 in RafBXB-induced transformation in greater detail, we expressed RasV12 or RafBXB alone and with an expression construct for IL-1Ra, a protein that binds to the IL-1 receptor and competes with IL-1α and IL-1β for binding to the receptor but fails to initiate a signal (19, 34). As shown in Fig. 5, IL-1Ra blocked RafBXB-induced but not RasV12-induced transformation. To examine a potential requirement for a signaling pathway downstream of the IL-1R that could account for the requirement for IL-1 in RafBXB-induced transformation, we coexpressed RafBXB or RasV12 with a dominant-negative form of TRAF6, TRAF6Δ, which consists of amino acids 289–522 (28). As shown in Fig. 5, TRAF6Δblocked RafBXB but not RasV12-induced transformation.

To examine the role of IL-1 in RafBXB-induced NF-κB activation and transformation, we considered the possibility that Raf might require IL-1 for the activation of NF-κB. To examine this, RafBXB was expressed alone or along with CrmA, IL-1Ra, and TRAF6Δ, and the impact on the activity of the NF-κB reporter construct was measured. As shown in Fig. 6, RafBXB induced activation of NF-κB-dependent gene expression was not affected by coexpression of CrmA, IL-1Ra, or TRAF6Δ, thereby demonstrating that Raf-induced activation of NF-κB was independent of IL-1.

We have demonstrated in this study that RafBXB-induced transformation of NIH 3T3 cells required activation of NF-κB and an IL-1 autocrine loop. Because inhibitors of the IL-1 pathway did not alter Raf-induced NF-κB activation, it is likely that Raf induces NF-κB that is upstream of IL-1 production in the cascade of events leading to transformation.

The role of IL-1 in transformation was clearly documented by overexpressing IL-1Ra or CrmA. On the basis of the fact that CrmA blocks Raf-induced transformation, it is most likely that IL-1β is the species of IL-1 involved, but we cannot rule out the possibility that IL-1α is also involved. Although other autocrine growth factors have been found previously to be induced by Raf, this is, to our knowledge, the first identification of an autocrine growth factor identified that is actually required for Raf-induced transformation. In contrast, we found that RasV12-induced transformation of NIH 3T3 cells did not require NF-κB activity or an IL-1 autocrine loop. This result may reflect the multiple effector pathways downstream of Ras (29, 35, 36). These may function to transform cells independently of an IL-1 autocrine loop. However, it is also possible that Ras only interacts with and activates such a hypothetical effector pathway because RasV12 was dramatically overexpressed. Consistent with this possibility is the recent finding that activation of endogenous Ras by growth factors led to very little stimulation of PI3K activity, despite the fact that Ras can directly interact with and stimulate the activity of p110, the catalytic subunit of PI3K (36, 37). In contrast, it appears that the overexpression of oncogenic Ras is capable of robust stimulation of PI3K, suggesting that artificially high levels of Ras allow it to stimulate signaling pathways that even an endogenous oncogenic form of Ras could not activate(37). In addition, it has been shown that the transformation of cultured cells in vitro requires 100-fold overexpression of RasV12 as compared with expression levels seen in human tumors expressing RasV12 (38). An analysis of cancer cell lines that express RasV12 from the endogenous promoter will ultimately serve as a model to determine whether the transformation of cells by RasV12 also requires an IL-1 autocrine loop. Finally, it should be noted that Ras-independent Raf activation has been established to occur in response to a number of different stimuli(38, 39, 40, 41). The possibility that transformation of cells by such Raf-dependent, Ras-independent pathways may involve an IL-1 autocrine loop requires consideration.

Although it has been reported previously that Raf can directly phosphorylate IκB and thereby activate NF-κB, this appears unlikely to occur in vivo (42, 43). Currently, the mechanism by which Raf activates NF-κB is not known, but based on the finding that a constitutively active form of MEK was incapable of activating NF-κB in the current studies, it would appear that an alternative effector of Raf may be involved in the activation of NF-κB. Although the activated form of MEK used in this study lacks a part of the nuclear exclusion sequence, this form of MEK localizes equally between cytosol and nucleus (44). Therefore, it is unlikely that MEKΔ failed to activate NF-κB,because it is was restricted to the nucleus with no access to the cytosol, where NF-κB activation primarily takes place. Other investigators have found that Raf-induced NF-κB activation can occur through the activation of the MAPK pathway with a subsequent autocrine growth factor activating the p38 pathway, resulting in the activation of NF-κB (45). In 293 cells, Raf-induced activation of NF-κB has been found to occur though the production of an autocrine factor that activates the c-Jun NH₂-terminal kinase pathway and causes the activation of NF-κB (46). Whatever the mechanism by which Raf induces the activation of NF-κB,it is significant that this activation is in fact required for Raf to induce transformation of NIH 3T3 cells. It is intriguing to speculate that Raf, a member of the MAPKKK family which also contains NIK and MEKK, two kinases that directly phosphorylate IKKα and IKKβ, might itself be directly regulating one or more of the IKK enzymes (47, 48). In support of the possibility that Raf might physiologically regulate the activity of NF-κB rather than just when overexpressed is the finding that Ras and Raf have both been implicated in the activation of NF-κB in response to both insulin and hypoxic stress (49, 50).

Although Raf has been reported previously to activate pathways other than the MAPK pathway, there has been doubt about whether these events were physiologically relevant because of the difficulty in some cases of reproducing the result or of showing an actual effect on a substrate subsequent to its phosphorylation by Raf. This study confirms that Raf can activate NF-κB and that this is an event required for Raf-induced transformation. We cannot rule out that biochemical events other than the activation of MAPK and NF-κB are also required for transformation. Thus far, we have not seen cooperation between MEK and other oncogenes that can activate NF-κB, such as MEKK, Rac, Rho, or Cdc42 (data not shown). This suggests that pathways downstream of Raf in addition to MAPK and NF-κB are required for it to induce transformation. Further work is clearly needed to delineate the mechanisms by which activated Raf induces transformation of NIH 3T3 as well as other cells. However, the current results further suggest that the ability of some genes to cooperate with RafBXB to transform cells may not be simply activating parallel pathways required for transformation but might cooperate with RafBXB at the level of inducing IL-1 or substituting for or mimicking signaling pathways from the IL-1 receptor. Further work will be required to examine this possibility.

The physiological relevance of the current findings is suggested by the finding that NF-κB has been found to be constitutively active in a variety of tumor cells, although the reason(s) for this in some tumors is unknown (51). It is intriguing to speculate that the constitutive activation of Raf by upstream signaling molecules may account for this phenotype, implying biological relevance of the results in NIH 3T3 cells for tumor cells in vivo.

Although further work in other model systems is required before the results of this study can be generalized, it is noteworthy that some tumors appear to use IL-1 as a necessary autocrine growth factor in vivo (52, 53, 54). Although the nature of the mutations in such cells has not been fully documented, it is possible that they have mutations resulting in the constitutive activation of Raf. It has also been shown, for example, that several cell lines derived from small cell lung carcinoma patient samples show overexpression of Raf at the mRNA and protein level, which results in a high level of Raf activity (55). Although the mechanism underlying the elevated Raf kinase activity is unknown, it is worth pointing out that Ras genes are rarely mutated in small cell lung carcinoma (56), thus raising the possibility that activation of Raf might occur independently of Ras activation and that the mechanism of transformation of these tumor cells might involve IL-1 Overall, the current study suggests the possibility that IL-1Ra, or other methods to interfere with IL-1 function, could block the growth of tumor cells in vivo.

We thank Drs. John Abrams, Dean Ballard, Melanie H. Cobb, Sankar Ghosh, David V. Goeddel, and Charles J. Sherr for providing critical reagents. We also thank David V. Goeddel for critical reading of the manuscript.