May 9, 1996 - We then investigated whether inhibition of wild-type. Ras with RaplA ...... analysis of the Kirsten-ras-revertant-1 gene: potentiation of its tumor.
The EMBO Journal vol.15 no.17 pp.4497-4505, 1996
Ras signaling in tumor necrosis factor-induced apoptosis
Jonathan C.Trent,Il, David J.McConkey1, Susan M.Loughlin, Matthew T.Harbison1, Antonio Fernandez and Honnavara N.Ananthaswamy2 Department of Immunology and 'Department of Cell Biology, The University of Texas M.D.Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA
2Corresponding author
Thmor necrosis factor (TNF) exerts cytotoxicity on many types of tumor cells but not on normal cells. The molecular events leading to cell death triggered by TNF are still poorly understood. Our previous studies have shown that enforced expression of an activated H-ras oncogene converted non-tumorigenic, TNFresistant C3H 1OT1/2 fibroblasts into tumorigenic cells that also became very sensitive to TNF-induced apoptosis. This finding suggested that Ras activation may play a role in TNF-induced apoptosis. In this study we investigated whether Ras activation is an obligatory step in TNF-induced apoptosis. Introduction of two different molecular antagonists of Ras, the raplA tumor suppressor gene or the dominant-negative rasN17 gene, into H-ras-transformed 10TEJ cells inhibited TNF-induced apoptosis. Similar results were obtained with L929 cells, a fibroblast cell line sensitive to TNF-induced apoptosis, which does not have a ras mutation. While Ras is constitutively activated in TNFsensitive 10TEJ cells, TNF treatment increased Rasbound GTP in TNF-sensitive L929 cells but not in TNFresistant 1OT1/2 cells. Moreover, RasN17 expression blocked TNF-induced Ras-GTP formation in L929 cells. These results demonstrate that Ras activation is required for TNF-induced apoptosis in mouse fibroblasts. Keywords: apoptosis/endogenous endonuclease/Ras activation/tumor necrosis factor
Introduction Tumor necrosis factor (TNF) is a multifunctional cytokine that is produced by a variety of cell types including activated macrophages and lymphocytes. In addition to causing hemorrhagic necrosis of tumors, TNF stimulates proliferation of normal cells; exerts cytolytic or cytostatic activity against tumor cells; and causes inflammatory, antiviral and other immunoregulatory effects (for reviews see Beutler and Cerami, 1986; Jaatela, 1991; Sidhu and Bollon, 1993; Beyaert and Fiers, 1994). While many tumor cells are sensitive to TNF-induced cell death in vitro, most normal fibroblasts are resistant (Sugarman et al., 1985). The cellular and molecular bases for this difference © Oxford University Press
in susceptibility to TNF-induced cell death have been elusive. Recent studies have shown that TNF sensitivity of cells can be regulated by certain oncogenes. For instance, overexpression of HER2/erbB2 (Hudziak et al., 1988) and v-abl (Suen et al., 1990) oncogenes confer TNF resistance, whereas enforced expression of the adenovirus EIA (Chen et al., 1987) or H-ras (Fernandez et al., 1994a, 1995) oncogenes confer TNF sensitivity. Our previous studies have shown that transfection with an activated H-ras oncogene converts the non-tumorigenic, TNFresistant C3H IOT1/2 fibroblasts into tumorigenic cells that also became extremely sensitive to TNF-induced cell death (Fernandez et al., 1994a, 1995). The differential sensitivity of normal I OT 1/2 and H-ras-transformed 1 OTEJ cells to TNF-induced cell death did not correlate with the number of TNF receptors on their cell surface (Fernandez et al., 1994a). TNF-resistant 10T1/2 cells, however, expressed higher levels of TNF-induced bcl-2, c-myc and MnSOD mRNA, genes known to control apoptosis, than the TNF-sensitive 1OTEJ cells (Fernandez et al., 1994a). H-ras-transformed 1OTEJ cells underwent apoptosis when exposed to TNF (Fernandez et al., 1994a, 1995). Although the precise biochemical events leading to TNF-elicited cell death are poorly understood, TNF binding to its cell surface receptor activates complex signal transduction pathways, including activation of protein kinases A and C, adenyl cyclase, G-proteins and NF-KB (Beyaert and Fiers, 1994; Fernandez and Ananthaswamy, 1994; Goossens et al., 1995). Recent studies indicate that certain molecules activated by Ras-dependent growth factor stimulation may also be involved in TNF signaling. For instance, several studies have demonstrated that the Map kinases Erk- 1 and Erk-2 are rapidly phosphorylated and activated in response to TNF treatment (Van Lint, et al., 1992; Vietor et al., 1993; Rafiee et al., 1995). Yao et al. (1995) demonstrated that ceramide-activated protein kinase forms a complex with Raf- 1 in HL-60 cells, and upon treatment with TNF and ceramide analogues, phosphorylates and activates Raf-1. Gulbins et al. (1995) demonstrated that engagement of Fas/APO- 1, another member of the TNF-R superfamily, also leads to Ras activation via ceramide. Additionally, Sluss et al. (1994) found that TNF treatment activates JNK-1 and JNK-2 leading to serine phosphorylation of the Jun transcription factor and enhanced AP-1 activity. These studies suggest that molecules activated by Ras in growth factor signaling may also be activated by TNF. Our finding that the expression of constitutively activated H-ras oncogene sensitized IOT1/2 mouse fibroblasts to TNF-induced apoptosis, combined with the involvement of Ras effector molecules in TNF signaling, suggested to us that TNF-mediated activation of Ras might be required for apoptosis. To investigate this possibility, we introduced two independent molecular antagonists of Ras, the raplA 4497
J.C.Trent,ll et aL
100
B
100
10 0
0
C)
1
10 0
1
10
100
0
TNF, ng/mI
1
10
100
TNF, ng/mI
I-
100 *0
Fig. 1. Western blot analysis of cell lysates from parental and raplAor rasN17-gene-transfected cell lines. (A) Expression of RaplA in IOTEJ (lane 1), lOTEJ-Hygro (lane 2) and two raplA transfectants (l0TEJ-R9 and 10TEJ-R12; lanes 3 and 4). (B) Expression of RaplA in L929 (lane 1), L929-Neo (lane 2) and two raplA transfectants (L929-R4 and L929-R5; lanes 3 and 4). Expression of RasN17 protein in rasN17 gene-transfected IOTEJ (C) and L929 (D) cells. Cells were treated with 100 lM Zn2+ for 0 h (lane 1), 9 h (lane 2) or 18 h (lane 3) to induce expression of RasN17. Arrows indicate RaplA and RasN17 protein bands.
C.)
1
tumor suppressor gene (Pizon et al., 1988; Kitayama et al., 1989, 1990; Quilliam et al., 1990; Cook et al., 1993) or the dominant-negative rasN17 gene (Feig and Cooper, 1988; Stacey et al., 1991) into TNF-sensitive IOTEJ and L929 cell lines and determined whether inhibition of Ras suppresses TNF-induced apoptosis. Although RaplA protein does not interfere with the formation of active Ras-GTP, it is a potent downstream inhibitor of Ras (Kitayama et al., 1989; Cook et al., 1993). Preliminary work done in our laboratory indicated that the raplA gene not only suppressed the transformed phenotype of mutant K-ras but also the transformed phenotype caused by mutant H-ras (unpublished data). On the other hand, the dominant-negative RasNl7 protein acts as a competitive antagonist upstream of Ras by preventing formation of the active Ras-GTP molecule (Feig and Cooper, 1988; Stacey et al., 1991).
Results Expression of Rap lA and RasN17 in 10TEJ and L929 cells We introduced either the raplA tumor suppressor gene (Pizon et al., 1988; Kitayama et al., 1989, 1990; Quilliam et al., 1990; Cook et al., 1993) or the dominant-negative rasN17 gene (Feig and Cooper, 1988; Stacey et al., 1991), two well-known molecular antagonists of Ras, into 1OTEJ and L929 cell lines by DNA-mediated gene transfer. While 1OTEJ cells contain a transfected mutant H-ras oncogene, the L929 cells do not contain activating mutations at codons 12, 13 or 61 of H-, K- and N-ras genes (unpublished data). Expression of RaplA and RasNl7 proteins in transfected cells was confirmed by immunoblot analysis. Transfectant clones 10TEJ-R9 and L929-R4 had 3.5- and 5-fold increased RaplA protein expression over that detected in untransfected 1OTEJ and L929 cell lines, respectively (Figure IA and B). Because constitutive expression of the rasN17 gene
4498
10
TNF, ng/ml
100
10
0
1 10 TNF, ng/ml
100
Fig. 2. TNF sensitivity of parental cell lines and raplA or rasN17 transfectants. (A) IOTEJ (a) and IOTEJ-Hygro (A), lOTEJ-R7 (El), IOTEJ-R9 (0), IOTEJ-R12 (A). (B) L929 (0), L929-Neo (A), L929R2 (O), L929-R4 (0), L929-R5 (A). lOTEJrasN17 transfectants (C) and L929rasN17 transfectants (D) treated without Zn2+ (0) and with 100 tM Zn> for 9 h (0). Data are mean + SE of three expenments, each performed in triplicate.
was cytotoxic, we transfected cells with a rasN17 construct under the control of a Zn2+-inducible promoter. The immunoreactive p21 was detected by utilization of an anti-pan Ras antibody that would react with endogenous Ras proteins as well as the Zn2+-driven RasNl7 protein. Thus, it was critical to look for time-dependent increase in p21 expression. As seen in Figure IC and D, the levels of immunoreactive RasN17 rise quantitatively following treatment with Zn2+ in both 1OTEJ and L929 transfectants. Quantification by densitometric scanning revealed a 5- to 7-fold increase in p21 at 9 h and a 7- to 9-fold increase in p21 at 18 h after Zn2+ treatment (data not shown). The stable expression of RaplA and inducible expression of RasNl7 allowed experimental analysis of Ras inhibition in TNF-sensitive cell lines.
Suppression of TNF-induced cell death by Ras inhibitors in 10TEJ and L929 cells Next, we investigated whether inhibition of Ras suppressed TNF-induced cell death. Clones expressing RaplA or RasNl7 proteins were treated with 0, 1, 10 or 100 ng/ml TNF for 24 h and cell survival was then determined by the MTT assay. The results shown in Figure 2A indicate that the 1OTEJ and vector control (lOTEJ-Hygro) are very sensitive to the cytotoxic effects of TNF, whereas the raplA transfectants, IOTEJ-R7, 1OTEJ-R9 and 1OTEJR12, are more resistant. In fact, at a dose of 10 ng/ml, the 1OTEJ-R9 transfectant is 5-fold more resistant to TNF-induced cell death. Analogous to RaplA studies,
ri.-1. --__2. :z.ms- .5_g,:_X
Ras activation in TNF-induced apoptosis
expression of RasNl7 in IOTEJ cells also induced partial resistance to TNF-induced cell death. Induction of RasN 17 afforded a 3-fold reduction in cytotoxicity in transfectants treated with Zn>2 for 9 h before TNF exposure (Figure 2C). However, longer pretreatment with Zn>2 resulted in an overall decrease in viability, which is most likely due to Zn>2 toxicity (data not shown). Nonetheless, these results support the data obtained with the rap 1A transfectants and further suggest that inhibition of Ras suppresses TNF-induced cell death in IOTEJ cells containing a constitutively activated mutant form of Ras. The data also support a role for Ras signaling in TNF-induced cell death. We then investigated whether inhibition of wild-type Ras with RaplA and RasNl7 in L929 cells also blocks TNF-induced cell death. The results shown in Figure 2B reveal that the parental L929 cells and L929 cells transfected with the vector alone are extremely sensitive to TNF-induced cell death, whereas L929 clones expressing the Rap 1 A protein are 3-fold more resistant to TNFinduced cell death. This resistance appears to be greatest at the 10 ng/ml dose of TNF. Similarly, L929 cells that were induced to express the dominant-negative RasN17 were found to be more resistant to TNF-induced cell death than were the uninduced transfectants (Figure 2D). However, the protection afforded by expression of RasN 17 was less pronounced than the protection afforded by the RaplA protein. At a dose of 10 ng/ml, the induced RasN17 L929 cells were >2-fold more resistant to TNF-induced cell death than the uninduced RasN17 L929 cells.
Inhibition of TNF-induced apoptosis by Ras inhibitors in 10TEJ and L929 cells Since our previous work showed that ras-transformed IOTEJ cells displayed the biochemical and morphological features of apoptosis when treated with TNF (Fernandez et al., 1994a, 1995), we investigated whether inhibition of Ras by RaplA and RasN17 blocks TNF-induced cell death by apoptosis. To investigate this, cells were treated with cytotoxic doses of TNF for 12 to 16 h, after which DNA fragmentation was assessed by agarose gel electrophoresis. The results shown in Figure 3A and B reveal that parental IOTEJ and L929 cells and their respective vector controls show marked DNA fragmentation in response to TNF treatment, whereas the raplA transfectants of IOTEJ and L929 display nearly complete abrogation of DNA fragmentation. We also found that expression of RasN17 in IOTEJ and L929 cells also suppressed TNF-induced DNA fragmentation, although suppression of apoptosis was not as complete with RasN17 as with Rap IA (data not shown). Thus, in order to detect more subtle differences, we used a quantitative diphenylamine (DPA) assay for measuring DNA fragmentation. This assay revealed slight but consistent differences in DNA fragmentation between the RasN 17-induced and uninduced transfectants. Uninduced transfectants of 1OTEJ displayed -1.5-fold more DNA fragmentation at 1 and 10 ng/ml TNF (Figure 3C). This attenuated suppression of apoptosis may be due to the effects of Zn2+ or perhaps by a low level of baseline RasN17 produced by the uninduced cells. At doses of 1 and 10 ng/ml TNF, the L929 RasN17-induced transfectants showed -2.5-fold less DNA fragmentation than did the uninduced cells (Figure 3D). Importantly, the dominant-negative RasN17 does not
A
...
t.
w
40 -
a
B
^- gS': r, ...
40 -
C
0
D
C)
E0, 0 Uz c
0 c)
O0.
_
0
0.1
1
I
I 0~
~ 0
~
I ~~
i
TN,ng;
.m
10
TNF. ng/ml
Fig. 3. TNF-induced DNA fragmentation in parental and ralp lA- or rasNl7-transfected lOTEJ and L929 cells. (A) IOTEJ (lane 1), lOTEJHygro (lane 2), lOTEJ-R7 (lane 3), 1OTEJ-R9 (lane 4) and lOTEJ-R12 (lane 5). (B) L929 (lane 1), L929-Neo (lane 2), L929-R2 (lane 3), L929-R4 (lane 4) and L929-R5 (lane 5). TNF-induced DNA fragmentation in rasNl7-transfected IOTEJ (C) and L929 (D) cells pre-treated with 100 MM Zn>+ for 9 h (solid bar) or without Zn'+ (hatched bar). Quantitative DNA fragmentation was assessed by the DPA method. Results shown are mean values of two separate experiments.
completely inhibit constitutively activated Ras, but it can prevent activation of wild-type Ras. This may explain the greater inhibition of apoptosis exerted by RasN 17 in L929 cells than that observed in IOTEJ cells.
Ras activation in TNF-induced apoptosis The inhibition of TNF-induced apoptosis by Ras inhibitors implies that Ras activation is regulated by TNF. To investigate this possibility directly we measured Ras activation by quantification of Ras-bound GTP in TNFtreated IOTI/2, IOTEJ and L929 cell lines. Cells were treated for 5 min with 0-100 ng/ml TNF, Ras was immunoprecipitated, guanine nucleotides eluted and separated by thin-layer chromatography. We found that untreated lOT1/2 cells contained a very low percentage of Ras-bound GTP (Figure 4), and that this level did not increase after treatment with TNF. As expected, the 1OTEJ cells containing a constitutively activated form of Ras expressed very high levels of Ras-bound GTP that was not detectably altered by treatment with TNF (Figure 4). Notably, in striking contrast to IOTl/2 and IOTEJ cell lines, TNF treatment increased the GTP-bound form of Ras in uninduced L929rasN 17 transfectants (Figure 5A, right panel). In fact, at 10 ng/ml TNF, -30% of Ras was bound to GTP in uninduced L929rasN17 transfectants
4499
J.C.Trent,II et al
t * * ^ * *:k, *;* A..
UPt)F1
:Fii& V-1 e>.
*.. -r.
V
Fig. 6. Activation of Ras by TNF in parental L929 and raplA transfectants. Untreated and TNF-treated (10 ng/ml) cells were analyzed for Ras-bound GDP and GTP as described in Materials and methods. Lanes I and 3, L929; lanes 2 and 4, L929-Neo; lanes 5 and 7, L929-R4; lanes 6 and 8, L929-R5. 4
t.-11
Fig. 4. Ras-bound GDP and GTP in untreated and TNF-treated 1OTI/2 and IOTEJ cells. Cells, untreated (lane 1) or treated with ng/ml (lane 2), 10 ng/ml (lane 3) or 100 ng/ml (lane 4) of TNF, were analyzed for Ras-bound GDP or GTP as described in Materials and methods.
5. '
J
if*
Fig. 7. Ca2+-induced DNA fragmentation in isolated nuclei from parental and raplA-transfected lOTEJ and L929 cells. Nuclei were isolated by detergent lysis of cells and centrifugation through a glycerol cushion. 5X 106 nuclei were incubated in the absence or presence of 5 mM CaCl, for I h at 37'C, lysed in a hypotonic lysis buffer and centrifuged for 20 min at 12 000 g. DNA fragments in the supernatant were resolved by 1.5% agarose gel electrophoresis as described before. (A) DNA fragmentation in the absence or presence of Ca'+ in nuclei from IOTEJ (lanes 1 and 2), l0TEJ-R9 (lanes 3 and 4) and l0TEJ-R12 (lanes 5 and 6). (B) DNA fragmentation in the absence or presence of Ca'+ in nuclei from L929 (lanes 1 and 2) and L929-R4 (lanes 3 and 4). Lane marked as M represents molecular weight marker (100 bp ladder).
treatment. Similarly, RaplA expression in 1OTEJ cells also did not affect Ras-GTP levels in untreated or TNF treated lOTEJ cells (data not shown). These results are in accord with previous reports that Rap A protein does not
interfere with the formation of active Ras-GTP (Kitayama et al., 1989, 1990; Cook et al., 1993).
Activation of endogenous endonuclease activity from isolated nuclei Fig. 5. Activation of Ras by TNF in rasN 17-transfected L929 cells under uninducing (-Zn2+) and inducing (+ 100 aM Zn2+ for 9 h) conditions. (A) Autoradiography showing Ras-bound GTP and GDP in rasNl7-transfected L929 cells. (B) Densitometric scanning of autoradiographs shown in (A). Scanning was performed using a densitometer (Molecular Dynamics, Sunnyvale, CA).
(Figure SB). More importantly, induction of dominantnegative RasN17 protein expression by Zn2+ treatment in L929rasNl7 transfectants completely abrogated TNFinduced Ras activation (Figure SA, left panel and SB). Analogous to uninduced L929rasN 17 transfectants, parental L929 and neo-transfected cells also showed increased Ras-bound GTP following TNF treatment (Figure 6). In addition, the two RapIA-expressing L929 transfectants also exhibited increased Ras-bound GTP following TNF 4500
One of the best-characterized biochemical events in apoptosis involves the activation of an endogenous nuclear endonuclease, leading to the step-wise cleavage of host chromatin into large (50-300 kb) and subsequently into oligosome-length DNA fragments (Wyllie, 1980; Filipski et al., 1990; Cohen et al., 1994). Since our results indicated that Ras activation is an obligatory step in TNF-induced apoptosis of 1OTEJ and L929 cells, we investigated whether inhibition of Ras had any effect on Ca2+-dependent endogenous endonuclease activity. As a marker for intrinsic apoptosis sensitivity, we isolated nuclei from 1 OTEJ (Figure 7A) and L929 (Figure 7B) cells along with raplA transfectants and determined Ca2+-dependent endogenous endonuclease activity by analysis of DNA fragmentation. Interestingly, we found that inhibition of Ras by RaplA made little impact on the ability of Ca>+ to activate the endogenous endonuclease in isolated nuclei.
Ras activation in TNF-induced apoptosis
1001
A
B
ioo1
CO) 0 1o
in
D 0.5 1.0 1.5 2.0
c
0
0.5 1.0 1.5 2.0
Doxorubicin, rig/mI
(U 100
C
0
10 I 0 100 200 300 400
101 0 100 200 300 400
Thapsigargin, nM Fig. 8. Sensitivity of parental and raplA transfected IOTEJ and L929 doxorubicin and thapsigargin. Cells were treated with doxorubicin (A and B) or thapsigargin (C and D) for 24 h and cell viability was determined by the MTT assay. (A and C) IOTEJ (0), IOTEJ-Hygro (A) and lOTEJ-R9 (0). (B and D) L929 (0), L929-Neo (A), L929-R4 (0). to
As seen in Figure 7, Ca2+-treated nuclei display a vigorous induction of DNA fragmentation in parental cells and rapIA transfectants alike. On the other hand, in the absence of Ca2+ there was no DNA fragmentation. The presence of the Ras inhibitors seems to have no effect on the activity of the endogenous endonuclease which is activated by Ca2>. Importantly, this contrasts with the effects of Bcl-2, which we have shown can completely inhibit endonuclease activity in nuclei from bcl-2 transfectants of IOTEJ cells (Fernandez et al., 1995). It is still possible that subtle differences in endonuclease activity may not be detected in this assay, or that this concentration of Ca2+ may overwhelm an inhibition set forth by RaplA. Nonetheless, inhibition of the Ras pathway by RaplA does not appear to alter the intrinsic activity of the endogenous endonuclease. These results are in general agreement with the notion that activation of endogenous endonuclease responsible for DNA ladder formation may not be required or a necessary event in apoptosis.
Effect of Ras inhibition on cell death induced by other cytotoxic drugs To determine whether Ras signaling is specific to TNFinduced cell death or is also involved in cell death induced by other agents, we tested the sensitivity of IOTEJ, L929 and their respective raplA transfectants to doxorubicinor thapsigargin-induced cell death. The results indicate that raplA-transfected lOTEJ and L929 cells were as sensitive to doxorubicin and thapsigargin as the parental and vector-transfected cells (Figure 8). Doxorubicin is an anti-tumor antibiotic isolated from Streptomyces spp. that exerts its effects via interaction with topoisomerase II, disruption of replication and transcription by intercalation, single- and double-stranded DNA scission, and by generation of free radicals. In light of its mechanisms of action,
it is perhaps not surprising that doxorubicin cytotoxicity was not affected by raplA. Thapsigargin, on the other hand, is an inhibitor of the ATP-dependent Ca2+ transporter found on the endoplasmic reticulum (ER), promoting ER Ca2+ release and an increase in the cytosolic Ca2+ concentration. The Ca2+ then, presumably, activates endogenous endonuclease and DNA fragmentation proceeds. If this is the case then direct activation of apoptosis by Ca2+ would bypass Ras signaling pathways. This is consistent with what we see when 1OTEJ, L929 and their raplA transfectants are treated with thapsigargin. Rasindependent cell death by thapsigargin suggests that the endogenous endonuclease is present and inducible by elevated intracellular Ca>. In summary, inhibition of the Ras pathway by Rap 1 A does not provide protection against the cytotoxic effects of doxorubicin or thapsigargin. Thus, these agents exert their cytotoxicity independent of Ras signaling or downstream from Ras activation, and the effects of the Ras inhibitors in these systems are selective for TNF-induced cell death.
Discussion The finding that TNF-resistant C3HlOTI/2 fibroblasts became sensitive to TNF-induced cell death when transformed by an activated Ha-ras oncogene (Fernandez et al., 1994a) suggested that Ras signaling might be involved in TNF-induced apoptosis. To test this hypothesis formally, we introduced raplA and rasN17 genes, two independent molecular inhibitors of Ras, into TNF-sensitive IOTEJ and L929 cells and determined whether inhibition of Ras suppresses TNF-induced cell death. We found that enforced expression of RaplA or RasN17 inhibited TNFinduced apoptosis in lOTEJ cells, which contains a mutant form of H-Ras. More interestingly, we found that these same inhibitors also suppressed TNF-induced apoptosis in L929 cells, which do not express constitutively activated Ras. This result implies that activation of wild-type Ras may be involved in the transmission of the TNF cytotoxic signal. While RasN17 expression inhibited cell death, it was not quite as effective as Rap 1 A. This incomplete suppression of TNF-induced apoptosis may be explained by several possibilities. First, the concentration of Zn2+ required to induce expression of RasN17 was itself cytotoxic upon prolonged exposure of cells. This possibility is supported by the finding that treatment of untransfected lOTEJ and L929 cells with 100 tM Zn>+ for 24 h resulted in 12-17% cytotoxicity (data not shown). Second, the inducible rasN17 plasmid may be 'leaky', producing a baseline level of RasN 17 that makes differences in cell death less obvious. This is an unlikely explanation because the uninduced transfectants did not express the RasN 17 protein and exhibited the same sensitivity to TNF as the parental IOTEJ and L929 cells. Third, overexpression of the dominant-negative RasN17 itself may be cytotoxic to the cell and mask effects of interfering with the TNF pathway. Although it is difficult to discern which of these possibilities is correct, the results obtained with the RasN17 cells corroborate those obtained with the RaplA transfectants. The involvement of Ras signaling in TNF-induced apoptosis was confirmed by the analysis of Ras-bound GDP and GTP. As expected, TNF-sensitive IOTEJ cells 4501
J.C.Trent,ll et aL
constitutively expressed high amounts of Ras-bound GTP, and TNF treatment did not further increase this level, most likely because the Ras-GTP level in these cells is already at a maximum. On the other hand, the TNFresistant 1OTI/2 cell line constitutively expressed very low levels of Ras-bound GTP, yet TNF treatment did not further increase the Ras-GTP level. Interestingly, TNFsensitive L929 cells expressed constitutively very low levels of Ras-GTP, and TNF treatment increased Rasbound GTP -6-fold. Furthermore, induction of dominantnegative RasN17 protein expression by Zn2+ treatment in L929 cells completely abrogated the TNF-induced RasGTP level. In contrast to RasN17, RaplA expression did not affect TNF-mediated Ras activation in L929 or 1OTEJ cells. Although previous studies have shown that stable expression of Rap 1A in Rat- 1 fibroblasts partially blocked activation of MAP kinases (Cook et al., 1993), it is not known whether rap1A-transfected L929 or 1OTEJ cells are defective in MAP kinase or other signaling molecules that are downstream of Ras. Nonetheless, these results demonstrate that Ras activation is required for TNFinduced apoptosis. A second conclusion that we can draw from these studies is that Ras signaling may not be involved in global resistance to all mechanisms of cell death. This conclusion is based on the finding that the Ca2+-mobilizing agent thapsigargin and the anti-neoplastic agent doxorubicin display equivalent cytotoxicity toward 1OTEJ, L929 and their respective raplA transfectants. This suggests that the cytotoxicity of doxorubicin and thapsigargin are not dependent on the presence of activated Ras but instead exert their influence in an independent manner. Alternatively, it is plausible that doxorubicin and thapsigargin may activate signaling molecules distal to the point of RaplA intervention in the Ras pathway. This is re-emphasized by the finding that the Ras inhibitors used had no affect on intrinsic Ca2+-dependent endogenous endonuclease activity detected in isolated nuclei. The latter observation suggests that inhibition of Ras suppresses a pro-apoptotic signal to the endonuclease. Interestingly, 1OTEJ cells transfected with Bcl-2 undergo cell death but fail to activate the endonuclease by either treatment of whole cells with TNF or by treatment of isolated nuclei with Ca2+ (Fernandez et al., 1995). This reflects an inherent impairment of endonuclease activity since, in the presence of Bcl-2, Ca2+ does not cause DNA fragmentation, an important feature of apoptosis, in isolated nuclei. Thus, Bcl-2 appears to be interacting in the apoptotic pathway but not the proliferative pathway. Taken together, these results support the notion that activation of endogenous endonuclease responsible for DNA ladder formation may not be required or a necessary event in apoptosis. Nonetheless, based on these findings we propose a model for the involvement of Ras signaling in TNFinduced apoptosis (Figure 9). Consider the TNF-resistant lOT1/2 cell line, in which Ras is neither constitutively activated by mutation nor activated by TNF stimulation. In this cell line, failure to activate Ras correlates with TNF resistance. Thus, while TNF is delivering an apoptotic signal, it is ineffective in the absence of activated Ras. It is interesting to speculate that an inhibitory molecule, located between the TNF receptor and Ras, may be preventing Ras activation by TNF treatment, or that
4502
Apoptosis Fig. 9. A model of Ras signaling in TNF-induced apoptosis. Ras may play a central role in both proliferation and apoptosis. Both TNF and members of the TNF-R family, such as Fas/APO-1, generate ceramide (Kim et al., 1991; Dbaibo et al., 1993; Obeid et al., 1993; Gulbins et al. 1995), which in turn promotes accumulation of GTP-bound Ras (Gulbins et al., 1995). The consequences of Ras activation, proliferation or apoptosis, may depend on the presence or absence of additional signals. Sole activation of Ras leads to proliferation, as with 1OTEJ cells. Activation of Ras in conjunction with additional signals provided by TNF results in apoptosis, as with 1OTEJ and L929 cells. Absence (in 10T1/2 cells) or inhibition (in raplA and rasN17 transfectants) of Ras activation blocks apoptotic signals. Finally, Bcl-2 can block TNF-induced apoptosis in IOTEJ cells expressing activated Ras (Fernandez et al., 1995).
activation of Ras by a mitogen such as EGF or PDGF might cooperate with the apoptotic signal delivered by TNF, resulting in enhanced cell death of 10T1/2 cells. On the other hand, the TNF-sensitive 1OTEJ cells express a constitutively activated Ras protein, locked into the GTPbound form, and are sensitive to the apoptotic effects of TNF. Thus, in 1OTEJ cells, the presence of activated Ras correlates with the TNF-sensitive phenotype. Inhibition of Ras by RaplA or RasN17 removes the signal necessary to cooperate with other signals triggered by TNF for the induction of apoptosis. Transfection of activated oncogenes in the Ras signaling pathway such as constitutively activated Raf-1, Map kinase kinase, or Map kinase into 10T1/2 cells would allow better characterization of the signals generated by Ras activation. For instance, if an activated Raf-I sensitizes 10TI/2 cells to TNF-induced apoptosis but Map kinase kinase does not, the proapoptotic signal generated by Ras activation is a novel pathway involving Raf-1. Few studies have shown that Ras acts as an antiapoptotic agent (Wyllie et al., 1987; Arends et al., 1993). Overexpression of c-Ha-ras in rodent fibroblasts resulted in a slight increase in in vitro growth rate and apoptosis in response to serum withdrawal. On the other hand, cells expressing mutant Ras proliferated more rapidly in vitro but exhibited decreased apoptosis when serum was removed from culture conditions, leading to the conclusion that mutant Ras can interfere with apoptosis (Wyllie et al.,
Ras activation in TNF-induced apoptosis
1987; Arends
et al., 1993). These findings are in contrast the results obtained by Tanaka et al. (1994) in which embryonic fibroblasts underwent apoptosis upon expression of activated Ras. In addition, Gulbins et al. (1995) demonstrated that engagement of Fas/APO- 1, another member of the TNF-R superfamily, also leads to Ras activation via ceramide. More recently, Yao et al. (1995) demonstrated that ceramide-activated protein kinase forms a complex with the Ras target, Raf-1 in HL-60 cells, and upon treatment with TNF and ceramide analogues, phosphorylates and activates Raf-1. In addition, preliminary studies of Yao et al. (1995) implicate Ras protein in TNF-induced MAP kinase activation. Our studies, on the other hand, show conclusively that Ras, whose role in proliferation is well established, is activated in the transduction of the TNF-induced apoptotic signal. However, it should be noted that Ras activation, in and of itself, does not result in fulminant apoptosis, as the IOTEJ cells expressing constitutively activated Ras display no obvious increase in spontaneous cell death. In addition, we have found that 1 OTEJ cells expressing the mutant Ras are more resistant to serum withdrawal-induced cell death than the parental 10TI/2 cells containing the wild-type Ras (S.M.Loughlin et al., in preparation). These results are in accord with the results of Arends et al. (1993). However, co-expression of RaplA in IOTEJ cells reduced their growth rate in medium containing either 10% or 0.0 1% serum, as compared with IOTEJ cells. These results are analogous to the previous reports which showed that expression of RaplA in NIH3T3 containing a mutant Ras caused suppression of growth as well as reversion of the transformed phenotype (Kitayama et al., 1989, 1990). Interestingly, even though IOTEJ cells are quite resistant to serum withdrawal-induced cell death, they still are highly sensitive to TNF-induced cell death under conditions of serum withdrawal as compared with the parental 10TI/2 cells or the raplA-transfected IOTEJ cells (S.M.Loughlin et al., in preparation). Therefore, it is interesting to postulate that TNF activates a Ras-independent parallel pathway that cooperates with Ras to send an apoptotic signal that results in activation of Ca2'-dependent endogenous endonuclease. We speculate that this signal alone does not activate the endonuclease, since 10TI/2 fibroblasts are resistant to TNF and their isolated nuclei do not undergo apoptosis following activation of Ca2+dependent endogenous endonuclease activity (Fernandez et al., 1995). It is possible, however, that concomitant signaling via activated Ras and a TNF apoptotic signal act in concert to activate the endonuclease and other molecules involved in apoptosis. Identification of TNFinduced signals that cooperate with Ras should provide new insights into the molecular mechanisms involved in TNF-induced apoptosis. The p55 TNF-R does not possess an intrinsic kinase domain (Tartaglia et al., 1991). It does, however, contain a motif known as the 'death domain' that is homologous with a cytosolic region of Fas/APO-1 (Tartaglia et al., 1993). Several studies have shown that TNF and members of the TNF-R family generate production of ceramide (Kim et al., 1991; Dbaibo et al., 1993; Obeid et al., 1993), and Gulbins et al. (1995) have shown that both Fas and exogenous ceramide promote accumulation of GTP-bound Ras. Thus, we speculate that the death domain may to
be responsible for activation of sphingomyelinase and generation of ceramide with subsequent activation of Ras. Preliminary experiments have shown that RaplA- or RasN17-expressing cells are more resistant to ceramideinduced cell death (unpublished data). This finding suggests that these agents share a common pathway that involves Ras signal transduction (Figure 9). Finally, the presence of activated Ras appears to confer two distinct phenotypes in C3HlOTl/2 cells, tumorigenicity and TNF sensitivity (Fernandez et al., 1994a). Interestingly, these two phenotypes are distinct since TNFresistant cells isolated from TNF-sensitive IOTEJ cells are still tumorigenic (Fernandez et al., 1994b). This phenotypic dichotomy in the Ras pathway appears to be distal to Ras, i.e. inhibition by RaplA and RasN17 appears to suppress both pathways. This result argues that there is a site downstream of Ras where transformation is separable from apoptosis. Alternatively, selection in TNF may produce a TNF-resistant phenotype that is independent of the Ras pathway. Nonetheless, our results indicate that inhibition of Ras by either Rap 1 A or by RasN 17 in 1 OTEJ and L929 cells suppresses the apoptotic response to TNF.
Materials and methods Cells C3Hl0Tl/2 (Reznikoff et al., 1973), 1OTEJ (Fernandez et al.. 1994a) and L929 (Xie et al., 1993) cell lines were grown at 37°C in a humidified 10%7 CO, incubator in Dulbecco's modified Eagle's medium (DMEM) (Grand Island Biological Company. Grand Island. NY) supplemented with 10%c bovine calf serum (BCS) (Hyclone Laboratories. Inc., Logan. UT). 100 U/ml penicillin and 100 pg/ml streptomycin. Stock cultures were maintained under subconfluent conditions by serial transfer at weekly intervals.
Plasmids The pZipr-aplA-nteo plasmid was obtained from Dr L.Quilliam (Scripps Cancer Center. La Jolla, CA). This is an 11.6 kb plasmid containing the 1.4 kb human raplA cDNA flanked by murine leukemia virus LTRs. -apl A was cloned into the BamiiHI site of pZIP-nieo. The rasN17 plasmid was obtained from Dr M.Fernandez-Sarabia (Onyx pharmaceuticals, CA). Plasmid DNAs were isolated and purified on cesium chloride gradients.
DNA transfection L929 cells (3X 105/60 mm dish) were transfected with 10 .g of plasmid DNA by CaPO4 precipitation with 20 peg of calf thymus DNA as a carrier. At 24 h after transfection. cells were trypsinized and transferred to three 100-mm dishes. After another 24 h, cells were refed with media containing 400 jg/ml G418. Since 1OTEJ cells already contained the nieo gene (Fernandez et al., 1994a). they were co-transfected with 10 jug of rap-1A plasmid DNA, I peg of pSV2-hygro and 20 p.g of carrier calf thymus DNA. Selection was in medium containing 300 jeg/ml of hygromycin instead of G418. Fresh selection media were added every 3-4 days. After 2 weeks' growth in G418 or hygromycin medium. individual drug-resistant colonies were isolated using sterilized glass cloning cylinders, expanded in culture and stored in liquid nitrogen.
MTT assay Cells were plated at a density of 5 x 103 in triplicate in 96-well microtiter plates in 100 VI of DMEM and incubated overnight at 37°C. The medium was removed and replaced with fresh medium containing different concentrations of human recombinant TNF-a. Cell lines transfected with -asN 17 were incubated in the absence or presence of 100 pM ZnCl, for 4.5-18 h before and during TNF treatment. The plates were incubated for an additional 12-24 h. and cell viability was determined by the MTT assay (Green et al.. 1984). Briefly. 10 pi of an MTT solution of (5 mg/ml) was added to the wells to a final concentration of 0.5 mg/ml. The plates were incubated at 37C for 2 h. after which time the medium was removed. dimethylsulfoxide was added (100 pl/well), and the plates were kept on an orbital shaker for 10 min. The optical density (OD)
4503
J.C.Trent,il et al. was measured at 570 nm using an ELISA reader. The percentage of cells surviving was calculated as follows: in presence of TNF X 100 % cell viability = OD in the absence of TNF
Analysis of DNA fragmentation Cells (lX 106) were plated in DMEM containing 10% BCS in triplicate T-25 flasks. Approximately 24 h later, TNF (50 ng/ml) was added to two flasks while the other flask served as a control. After incubation for 24 h, cells were harvested and washed once with PBS. Cell pellets were resuspended in 0.5 ml of lysis buffer (10 mM Tris, 1 mM EDTA, 0.2% Triton X-100, pH 7.5), vortexed and incubated for 20 min on ice. After centrifugation at 14 000 g for 10 min, DNA was precipitated from the supernatants by the addition of 5.0 M NaCl (to a final concentration 0.5 M) and 0.5 volumes of 100% isopropanol. After storage at -20°C. the samples were centrifuged at 14 000 g for 10 min. The pellets were resuspended in 50 Id of TE buffer containing proteinase K (300 pg/ml) and RNase A (100 pg/ml) and incubated for 30 min at 50°C. The samples were electrophoresed through a 1.5% agarose gel in 0.6x TBE buffer at 150 V for 2 h. The DNA bands were visualized by staining with ethidium bromide and photographed under UV light using a
transilluminator. The DPA assay was modified from Burton (1956). Cells were treated with TNF as above and lysed on ice with a hypotonic lysis solution (25 mM EDTA, 25 mM Tris, pH 7.5 and 0.5% Triton X-100). After microcentrifugation for 15 min at 14 000 g, supernatants were collected and DNA precipitated with 5% trichloroacetic acid (TCA) at 4°C overnight. The pellets were collected by microcentrifugation at 14 000 g for 5 min, washed once with 5% TCA and resuspended in 150 pd 5% TCA. Aliquots were transferred in duplicate and 75 pd of DPA (15 mg/ml diphenylamine, 100% acetic acid, 1.5% sulfuric acid, 0.5% acetaldehyde) reagent was added. After incubation for 16 h at 30°C, absorption was measured at 600 nm in a spectrophotometer. Percent DNA fragmentation was calculated as the ratio of absorption value of supernatant divided by the absorption values of pellet plus supernatant multiplied by 100.
Ras GTP/GDP binding This assay was adapted from Adari et al. (1988). Briefly, 1 X 106 cells were plated in 100-mm dishes and allowed to reach 75% confluency, after which the dishes were washed three times with HEPES-buffered saline (150 mM NaCl, 20 mM HEPES, pH 7.4) and incubated for 3-6 h in 8 ml phosphate-free DMEM containing 5% dialyzed FCS (Hyclone Laboratories, Inc., Logan, UT) and 50 ,tCi/ml of [32P]orthophosphate (NEX-053, NEN Dupont, Boston, MA). The medium was then decanted and washed twice with cold cell wash buffer. Cells were then lysed on the dish by scraping with cold lysis buffer, transferred to an Eppendorf tube and vortexed for 10 s. Cell debris was removed by centrifugation at 14 000 g for 10 min. 3 .tg of Y13-259 anti-pan Ras antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the supernatant and incubated overnight at 4°C. 25 ,tl of packed protein A/ G-agarose beads were then added to each sample and rotated overnight at 40C. Beads were collected by centrifugation at 14 000 g for 2 min and washed eight times with IP wash buffer (50 mM HEPES, pH 7.4, 0.5 M NaCI, 0.1% Triton X-100, 0.005% SDS, 5 mM MgCl2. The washed beads were resuspended in 25 pg of elution buffer (0.2% SDS, 5 mM DTT, 1 mM GDP, 1 mM GTP, 2 mM EDTA), heated at 68°C for 15 min and 5.0 ,l separated on a PEI-cellulose thin-layer chromatography plate (EM Science, Gibbstown, NJ). Separation buffer was I M LiCl, pH 3.4. Results were calculated from the following equation based on densitometric scanned values: GTP/[(GDPX 1.5) + GTP]X 100. The factor of 1.5 is used to correct for the incorporation of three [32P]orthophosphate molecules into GTP and two into GDP.
Western analysis Cells (6X 106) were lysed by 30 passages through a 22-G needle in 0.3 ml of NP-40 buffer containing aprotinin, leupeptin and PMSF. The supernatants were found to contain 2-6 mg/ml crude protein by the Bradford assay (Bio-Rad Laboratories, Hercules, CA). 10 pg of crude protein was separated on a 12% SDS-PAGE mini-gel at 150 V for 1.5 h. The gel and the nitrocellulose (0.22 ,um; Schleicher and Schuell, Keene, NH) were equilibrated for 20 min in transfer buffer (glycine, SDS, 20% methanol) and transferred to nitrocellulose by semi-dry electrophoretic transfer (Bio-Rad). Detection was performed by enhanced chemiluminescence (ECL, Amersham, Arlington Heights, IL) using the manufacturer's protocol, blocking solution and secondary antibody. The primary antibodies used were polyclonal rabbit anti-Rap-lA peptide (residues 131-
4504
140) (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and mouse monoclonal antibody (mAb), pan-Ras (Ab-4, Oncogene Sciences, Uniondale, NY) (for detecting RasN17).
Endogenous endonuclease activity in isolated nuclei Nuclei were isolated and endogenous endonuclease activity was assessed as described previously (McConkey et al., 1989; Fernandez et al., 1995). Cells (5 X 106) were cultured as above to 80% confluency, at which time they were scraped from tissue culture plates, centrifuged and resuspended in 3 ml SMTG buffer (150 mM NaCl, 1.5 mM MgCl,, 10 mM Tris, pH 7.4, 3% glycerol) on ice. Next, 3 ml of lysis buffer (50% SMTG buffer with 5 mM DTT, 100 nM PMSF, 100 ,ug/ml leupeptin and 0.1% NP-40) was added and the cells allowed to lyse on ice for 10 min. The samples were then gently loaded onto a glycerol cushion (25% glycerol, 10 mM Tris, pH 7.4, 1.5 mM MgCl,) and spun at 1250 r.p.m. for 20 min at 4°C. The nuclei were resuspended in TKM buffer (25 mM Tris, pH 7.5, 150 mM KCI, 5 mM MgCl2) and either treated or untreated for 1 h with 0.5 mM CaCl, at 37°C. DNA fragments were isolated and resolved by 1.5% agarose gel electrophoresis as described above.
Acknowledgements We thank R.Palacios, L.Owen-Schaub and G.Gallick for critical reading of the manuscript, and L.Quilliam and M.Fernandez-Sabaria for providing raplA and rasN17 plasmids, respectively. Supported in part by NIH grants CA46523 (H.N.A) and CA16672 (D.J.M). J.C.T. was supported by NIH supplemental grant ROI-CA-46523 and by the University of Texas Health Science Center MD/Ph.D Program.
References Adari,H., Lowy,D.R., Willumsen,B.M., Der,C.J. and McCormick,F. (1988) Guanosine triphosphatase activating protein (GAP) interacts with p21 ras effector binding domain. Science, 250, 518-521. Arends,M.J., McGregor,A.H., Toft,N.J., Brown,E.J.H. and Wyllie,A.H. (1993) Susceptibility to apoptosis is differentially regulated by c-myc and mutated H-ras oncogenes and is associated with endonuclease availability. Br J. Cancer, 68, 1127-1133. Beutler,B. and Cerami,A. (1986) Cachectin and tumour necrosis factor as two sides of the same biological coin. Nature, 320, 584-588. Beyaert,R. and Fiers,W. (1994) Molecular Mechanisms of tumor necrosis factor-induced cytotoxicity: what we do understand and what we do not. FEBS Lett., 340, 9-16. Burton,K. (1956) A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid. Biochein. J., 62, 315-323. Chen,M.J., Holskin,B., Stickler,J., Gorniak,J., Clark,M.A., Johnson,P.J., Mitcho,M. and Shalloway,D. (1987) Induction by EIA oncogene expression of cellular susceptibility to lysis by TNF. Natulre, 330, 581-583. Cohen,G.M., Sun,X.M., Fearnhead,H., MacFarlane,M., Brown,D.G., Showden,R.T. and Dinsdale,D. (1994) Formation of large molecular weight fragments of DNA is a key committed step of apoptosis in thymocytes. J. Ihnmuntitnol., 153, 507-516. Cook,S.J., Rubinfeld,B., Albert,I. and McCormick,F. (1993) RapV12 antagonizes Ras-dependent activation of ERK1 and ERK2 by LPA and EGF in Rat-l fibroblasts. EMBO J., 12, 3475-3485. Dbaibo,G., Obeid,L. and Hannun,Y. (1993) Tumor necrosis factor-alpha (TNF-alpha) signal transduction through ceramide. J. Biol. Chem., 268, 17762-17766. Feig,L.A. and Cooper,G.M. (1988) Inhibition of 3T3 cell proliferation by a mutant Ras protein with preferential affinity for GDP. Mol. Cell. Biol., 8, 3235-3243. Fernandez,A. and Ananthaswamy,H.N. (1994) Molecular basis for tumor necrosis factor-induced apoptosis. Cancer- Blull., 46, 153-160. Fernandez,A., Marin,M., McDonnell,T. and Ananthaswamy,H.N. (1994a) Differential sensitivity of normal and Ha-ras-transformed C3H mouse embryo fibroblasts to tumor necrosis factor: induction of bcl-2, c-rnvc, and manganese superoxide dismutase in resistant cells. Oncogene, 9, 2009-2017. Fernandez,A., Chen,P.W. and Ananthaswamy,H.N. (1994b) rcs-induced neoplastic transformation and sensitivity to tumor necrosis factor. Aniticancer Res.. 14, 2649-2652. Fernandez,A., Fosdick,L.J., Marin,M.C., Diaz,C., McDonnell,T.J., Ananthaswamy,H.N. and McConkey,D. (1995) Differential regulation
Ras activation in TNF-induced apoptosis of endogenous endonuclease activation in isolated murine fibroblast nuclei by ras and bcl-2. Onicogene. 10, 769-774. Filipski.J., Leblanc,J., Youdale,T., Sikorska,M. and Walker,P.R. (1990) Periodicity of DNA folding in higher order chromatin structures. EMBO J.. 9. 1319-1327. Goossens.V.. Grooten.J.. De Vos,K. and Fiers,W. (1995) Direct evidence for tumor necrosis factor-induced mitochondrial reactive oxygen intermediates and their involvement in cytotoxicity. Proc. Natl Acad. Sci. USA. 92, 8115-8119. Green,L.M.. Reade.J.L. and Ware,C.F. (1984) Rapid colorimetric assay: application to the quantitation of cytotoxic and growth inhibitory lymphokines. J. Inmuniol. Methodls. 70, 257-268. Gulbins,E. et al. (1995) FAS-induced apoptosis is mediated via a ceramide-initiated RAS signaling pathway. Immiuiiilnitv, 2, 341-351. Hudziak.R.M., Lewis,G.D., Shalaby.M.R., Eessalu,T.E.. Aggarwal,B.B., Ullrich.A. and Shepard.H.M. (1988) Amplified expression of the HER2/ERBB2 oncogene induces resistance to tumor necrosis factorax in NIH3T3 cells. Proc. Natl Acad. Sci. USA. 85, 5102-5106. Jaatela,M. (1991) Biology of disease: biological activities and mechanism of action of tumor necrosis factor-oc/cachectin. Lab. Invest., 64. 724-742. Kim,M.Y.. Linardic,C., Obeid,L. and Hannun,Y. (1991) TNF-alpha and interferon-gamma signaling through ceramide. J. Biol. Chemn.. 266, 484-489. Kitayama.H., Sugimoto,Y., Matsuzaki,T., Ikawa,Y. and Noda,M. (1989) A a-cis-related gene with transformation suppressor activity. Cell. 56, 77-84. Kitayama,H., Matsuzaki,T., Ikawa,Y. and Noda,M. (1990) Genetic analysis of the Kirsten-ras-revertant-1 gene: potentiation of its tumor suppressor activity by specific point mutations. Proc. Natl Acad. Sci. USA, 87. 4284-4288. McConkey,D.J., Hartzell,P., Jondal,M. and Orrenius,S. (1989) Inhibition of DNA fragmentation in thymocytes and isolated thymocyte nuclei by agents that stimulate protein kinase C. J. Biol. Chemi.. 264, 13399-13402. Obeid,L., Linardic.C.. Karolak,L. and Hannun,Y. (1993) Programmed cell death induced by ceramide. Science, 259. 1769-1771. Pizon,V.. Chardin,P., Lerosey,l.. Oloffsson,B. and Tavitian,A. (1988) Human cDNAs rapl1 and rap2 homologues to the Drosophila gene Dras3 encode proteins closely related to ras in the 'effector' region. Onicogente. 3. 201-204. Quilliam.L.A.. Der.C.J.. Clark.R.. O'Rourke.E.C.. Zhang.K.. McCormick,F. and Bokoch,G.M. (1990) Biochemical characterization of baculovirus-expressed rap 1 A/K-rev- I and its regulation by GTPaseactivating proteins. Mol. Cell. Biol.. 10, 2901-2908. Rafiee.P.. Lee.J.. Leung.C.-C. and Raffin.T. (1995) TNF-alpha induces tyrosine phosphorylation of mitogen-activated protein kinase in adherent human neutrophils. J. Immunollt10o.. 154, 4785-4792. Reznikoff,C.A., Brankow,D.W. and Heidelberger,C. (1973) Establishment and characteristics of a cloned line of C3H mouse embryo cells sensitive to post confluence inhibition of division. Cancer Res.. 33. 3231-3238. Sidhu,R. and Bollon,A. (1993) Tumor necrosis factor activities and cancer therapy - a perspective. Plharm1tacol. Tlher, 57, 79-128. Sluss,H.K., Barrett.T., Derijard,B. and Davis.R. (1994) Signal transduction by tumor necrosis factor mediated by JNK protein kinases. Mol. Cell. Biol., 14, 8376-8384. Stacey,D.W.. Feig.L.A. and Gibbs.J.B. (1991) Dominant inhibitory Ras mutants selectively inhibit the activity of either cellular or oncogenic Ras. Mol. Cell. Biol.. 11, 4053-4064. Suen,T.C., Rodriguez.R.U., Hung,M.C. and Klostergaard,J. (1990) Vabl confers resistance and growth advantage to TNF-a in NIH3T3 cells. Lvminplhokine Res.. 9. 15-25.
Van Lint,J., Agostinis,P., Vandevoorde.V., Haegeman,G.. Fiers.W.. Merlevede,W. and Vandenheede.J. (1992) Tumor necrosis factor stimulates multiple serine/threonine protein kinases in Swiss 3T3 and L929 cells. J. Biol. Clem.., 267, 25916-25921. Vietor,I., Schwenger.P.. Li,W.. Schlessinger.J. and Vilcek,J. (1993) Tumor necrosis factor-induced activation and increased tyrosine phosphorylation of mitogen-activated protein kinase in human fibroblasts. J. Biol. Chenm., 268, 18994-18999. Wyllie,A.H. (1980) Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Natulre, 284. 555-556. Wyllie,A.H.. Rose,K.A., Morris,R.G., Steel,C.M.. Foster.E. and Spandidos,D.A. (1987) Rodent fibroblast tumors expressing human inyc and ras genes: growth, metastasis and endogenous oncogene expression. Br: J. Cancer, 56, 251-259. Xie,K., Huang.S.. Dong,Z. and Fidler,I.J. (1993) Cytokine-induced apoptosis in transformed murine fibroblasts involves synthesis of endogenous nitric oxide. Itot. J. Oncol., 3, 1043-1048. Yao,B., Zhang,Y., Delikat,S., Mathias,S., Basu,S. and Kolesnick,R. (1995) Phosphorylation of Raf by ceramide-activated protein kinase. Nature, 378, 307-310.
Receiv!ed on Decemiiber 14, 1995; revised oni May 9, 1996
Sugarman,B.J.. Aggarwal.B.B., Hass,P.E., Figari,I.S., Palladino,M.A. and Shepard.H.M. (1985) Recombinant human tumor necrosis factor alpha: effects on proliferation of normal and transformed cells in litro. Science. 230. 943-945.
Tanaka,N., Ishihara,M.. Kitagawa.M., Harada,H., Kimura.T., Matsuyama,T., Lamphier,M.S., Mak.T.W. and Taniguchi.T. (1994) Cellular commitment to oncogene-induced transformation or apoptosis is dependent on the transcription factor IRF- 1. Cell. 77, 829-839. Tartaglia,L., Weber.R.. Figari.I., Reynolds.C.. Palladino.M. and Goeddel.D. (1991) The two different receptors for tumor necrosis factor mediate distinct cellular responses. Proc. Naitl Ac-ad. Sci. USA. 88, 9292-9296. Tartaglia,L.. Ayres,T., Wong,G. and Goeddel,D. (1993) A novel domain wAithin the 55 kDa TNF receptor signals cell death. Cell. 74, 845-853.
4505
