12-O-tetradecanoylphorbol-13-acetate-induced Apoptosis Is Mediated by Tumor Necrosis Factor α in Human Monocytic U937 Cells*

12-O-tetradecanoylphorbol-13-acetate (TPA), a phorbol ester that is known as a tumor promoter, induces differentiation of myeloid cells and suppresses their proliferation. We studied the regulation of apoptosis by TPA in human monocytic cell line U937 cells that lack p53. Untreated U937 cells constitutively underwent apoptosis, and TPA enhanced apoptosis in these cells. Further studies showed that TPA increased production of tumor necrosis factor-α (TNFα) in U937 cells, and exogenously added TNFα induced apoptosis. Moreover, the induction of apoptosis by TPA was blocked by anti-TNFα antibody. Similar results were obtained in the myeloblastic cell line KY821 cells. We also found that the induction of apoptosis by TPA was increased in cells overexpressed with TNF receptor 1 but not in control cells. Furthermore, TPA failed to induce the production of TNFα and apoptosis in cells with either their protein kinase C or mitogen-activated protein kinase pathway blocked. Our results indicate that TPA induces apoptosis, at least in part, through a pathway that requires endogenous production of TNFα in U937 cells. Our data also suggest that the induction of apoptosis by TPA occurs through activation of protein kinase C and mitogen-activated protein kinase and TNFα is an autocrine-stimulating factor for the induction of apoptosis in these cells.


12-O-tetradecanoylphorbol-13-acetate (TPA), a phorbol ester that is known as a tumor promoter, induces differentiation of myeloid cells and suppresses their
proliferation. We studied the regulation of apoptosis by TPA in human monocytic cell line U937 cells that lack p53. Untreated U937 cells constitutively underwent apoptosis, and TPA enhanced apoptosis in these cells. Further studies showed that TPA increased production of tumor necrosis factor-␣ (TNF␣) in U937 cells, and exogenously added TNF␣ induced apoptosis. Moreover, the induction of apoptosis by TPA was blocked by anti-TNF␣ antibody. Similar results were obtained in the myeloblastic cell line KY821 cells. We also found that the induction of apoptosis by TPA was increased in cells overexpressed with TNF receptor 1 but not in control cells. Furthermore, TPA failed to induce the production of TNF␣ and apoptosis in cells with either their protein kinase C or mitogen-activated protein kinase pathway blocked. Our results indicate that TPA induces apoptosis, at least in part, through a pathway that requires endogenous production of TNF␣ in U937 cells. Our data also suggest that the induction of apoptosis by TPA occurs through activation of protein kinase C and mitogen-activated protein kinase and TNF␣ is an autocrinestimulating factor for the induction of apoptosis in these cells.
Apoptosis, a morphologically distinguished form of programmed cell death, plays important roles not only during development and tissue homeostasis but also in the pathogenesis of a variety of diseases including cancer, autoimmune disease, viral infection, and neurodegenerative disorder (1)(2)(3)(4)(5)(6). Moreover, various chemicals and drugs for treatment of cancers are known to destroy tumor cells through apoptosis (7,8). The precise mechanisms that control apoptosis have not been elucidated; it appears that this form of cell death is regulated by a genetic program involving both inducers and repressors (6). The tumor suppressor p53 is one of the key proteins affecting the fate of cells exposed to external insults; activation of p53 leads to apoptosis or cell cycle arrest. Inactivation of p53 by either a mutation of the gene or interaction with oncogenic viral or cellular proteins is a common event in the development of malignancies (9). On the other hand, recent studies have found pathways of apoptosis independent of p53 (10).
Protein kinase C (PKC) 1 plays important roles in physiological events such as cell differentiation and proliferation (11). PKC phosphorylates and activates Raf-1, which leads to the activation of mitogen-activated protein kinase (MAPK) or extracellular signal-regulated protein kinase (ERK) (12). By contrast, studies have shown that proliferation of cells is also inhibited by PKC activation; phorbol esters that can activate PKC have been shown to inhibit the phosphorylation of retinoblastoma susceptibility gene product (Rb) (13). 12-O-tetradecanoylphorbol-13-acetate (TPA) is a novel activator of PKC; previous studies (14 -16) have reported that TPA induces apoptosis in MCF-7 breast cancer cells, U937 cells, and Jurkat T-lymphoma cells. TPA is also known to modulate apoptosis induced by various agents (17,18). However, the effects of TPA on induction of apoptosis seem to depend upon the type of cells; treatment of cells with inhibitors of PKC resulted in apoptosis in lymphocytes and HL60 cells, and TPA blocked apoptosis induced by VP-16 or interleukin withdrawal (18 -21). Studies have suggested that this different response to TPA in apoptosis is, in part, due to the existence of multiple PKC isotypes (22,23).
Tumor necrosis factor-␣ (TNF␣) is a pleiotropic cytokine that was originally identified on the basis of its ability to induce hemorrhagic necrosis in murine transplantable tumors and of its direct cytotoxic effects against transformed cells lines in vivo (24 -26). TNF␣ plays a physiological role in oncogeny and in the control of major inflammatory and immune reactions affecting both myelocytic and nonmyelocytic cells (27). It is primarily produced by activated macrophages and exerts its effect through two distinct cell surface receptors, denoted TNFR1 and TNFR2. These two TNF␣ receptors are present on all nucleated cell types (28,29). TNF␣ preferentially kills transformed cells in vivo although it is not cytotoxic to normal diploid cells (26,28,29). The mechanisms underlying the cytotoxic effects of TNF␣ are not completely understood. However, it has been shown that cytotoxicity of TNF␣ is mediated through mainly TNFR1, which has a 70-amino acid region in the cytoplasmic domain termed the "death domain" (30). Further studies have found that inhibition of DNA topoisomerase, production of reactive oxygen intermediates in the mitochondria, and induction of proteases appears to be involved (31)(32)(33). It has been reported that the cytotoxic effects of TNF␣ on neoplastic cells are synergistically enhanced by TPA in vitro although the molecular mechanisms underlying this synergism are not known (34).
In this study, we examined the cytotoxic effects of TPA on a human monocytic cell line, U937, and a human myeloblastic cell line, KY821. Our results demonstrate that TPA induces apoptosis through TNF␣ production in U937 and KY821 cells.

MATERIALS AND METHODS
Cells and Cell Culture-U937, derived from a diffuse human histiocytic lymphoma (American Type Tissue Culture Collection, Manassas, VA) and KY821, derived from human myeloblastic leukemia cells (kindly provided by Dr. K. Kishi, Niigata University School of Medicine, Niigata, Japan) (35) were cultured in ␣MEM (Cosmo Bio Co. Ltd., Tokyo, Japan). The medium was supplemented with 7% fetal calf serum (Mitsubishi Kasei Co, Tokyo, Japan). The cells were incubated at 37°C in a humidified atmosphere with 5% CO 2 .
Reagents-A monoclonal TNF␣ antibody and a neutralizing polyclonal rabbit antibody against human TNF␣ were purchased from Genzyme Co. (Cambridge, MA). 10 l of this antibody neutralizes Ͼ100 units/ml TNF␣ in L929 cell cytotoxic assay. The neutralizing antibody against IL-1␤ was a polyclonal rabbit anti-human IL-1␤ antibody and was kindly provided by Dr. T. Nishida (Otsuka Pharmaceutical Co. Ltd. Tokushima, Japan). This antibody neutralized more than 100 units of IL-1␤ at a final dilution of 1:100 in the thymocyte co-stimulating assay. The p53 monoclocal antibody PAb1801 (Ab-2) was purchased from Oncogene Science (Cambridge. MA). The polyclonal antibody against human Fas was from Calbiochem. TPA was purchased from Sigma. A selective PKC inhibitor, GF-109203X (36), and a selective inhibitor of MAPK kinase (MEK), PD-98059 (37), were purchased from RBI (Natick, MA) and BIOMOL (Plymouth Meeting, PA), respectively.
Analysis of DNA Fragmentation-DNA was isolated from cells treated with or without TPA as described previously (38). Cells were washed with phosphate-buffered saline without Ca 2ϩ or Mg 2ϩ (PBS Ϫ ) twice and then lysed in buffer containing 0.1 M Tris-HCl, pH 7.4, 50 mM NaCl, 10 mM EDTA, 0.2% SDS, and incubated with 1 mg/ml proteinase K (Sigma) overnight at 37°C. After extraction by phenol/chloroform and precipitation in ethanol, the precipitates were treated with RNase A (Sigma, 56 g/ml) at 37°C for 2 h in TE buffer (10 mM Tris-NaCl, pH 8.0, 1 mM EDTA). The DNA was extracted with phenol/chloroform and resuspended in TE buffer. The same amount of DNA for each sample was loaded and electrophoresed in a 2% agarose gel with TAN buffer (1.6 M Tris-HCl, pH 7.2, 0.8 M NaOAc, 40 mM EDTA) and the gel was stained with ethidium bromide.
Quantitative Analysis of Apoptosis-Cells were harvested, then fixed with 1% glutaraldehyde in PBS Ϫ at room temperature overnight. Samples were washed with PBS Ϫ and stained with 8 g/ml of DNA-binding dye Hoechst 33258 (39). Cells were examined under an Axioplan microscope (Carl Zeiss, Oberkochen, Germany) and cells characterized by nuclear fragmentation and chromatin condensation were defined as apoptotic cells.
DNA Probes-Human TNF␣ cDNA probe (0.8 kb, EcoRI) and IL-1␤ cDNA were from pSPl 42-2 and PA 26, respectively (43,44), and the p53 cDNA was purified from the pR4 -2 plasmid (0.5 kb, NcoI). For TNFR1 probe, the two 20-mer oligodeoxynucleotide, the forward primer (5Ј-TCCCCTCCCACCTTCTCTCC-3Ј) and the reverse primer (5Ј-GC-CCACCAGCCCACTCTTCC-3Ј) were synthesized according to the nucleotide sequence of human TNFR1 DNA (45). The 333-base pair DNA fragment corresponding to exon 1 of the human TNFR1 gene was amplified from genomic DNA of human granulocytes by polymerase chain reaction in a Thermal Cycler MP (Takara, Tokyo, Japan) using EXTaq DNA-polymerase (Takara). The resulting polymerase chain reaction product was subcloned into pCR 2.1 vector (Invitrogen, Inc., Carlsbad, CA). DNA sequence analysis of the cloned DNA confirmed that it contained the complete sequence for exon 1 of human TNFR1 DNA (45). For Northern blotting, a 0.3-kb DNA fragment was excised by digesting the exon 1 fragment with EcoRI. These probes were 32 Plabeled by the random priming method (46). The specific activity was approximately 2 ϫ 10 8 cpm/g of DNA.
Construction of Expression Vectors-A plasmid bearing the complete coding region of human 55-kDa TNF receptor (TNFR1 in pUC19) was provided by Dr. H. Loetscher (F. Hoffmann-La Roche Ltd., Basel, Switzerland). An expression vector of human TNFR1 was constructed as described previously (48). Briefly, TNFR1 cDNA was cloned in the pRC/RSV eukaryotic expression vector (Invitrogen, Carlsbad, CA) that contains Rous sarcoma virus long terminal repeat (RSV-LTR) and the neomycin resistance gene (Neo) expressed under the SV40 promoter. For control, N-terminal deletion mutant of TNFR1 was generated by digestion of a TNFR1-expression vector with HindIII, and subsequent self-ligation (Fig. 5A). In this vector, approximately 240 amino acids were removed from the N-terminal end, resulting in lack of extracellular domain and transmembrane domain sequences.
DNA Transfection and Selection of Transformed Cells-2 g of plasmid DNA was introduced into 5 ϫ 10 5 of U937 cells using FuGENE6 TM , lipofection reagent (Roche Molecular Biochemicals). 48 h after exposure to DNA, cells were cultured on two-layer soft agar in selection medium supplemented with 0.5 mg/ml of G418 for 4 weeks. Single colonies of G418-resistant cells were isolated, cultured in complete medium supplemented with 0.5 mg/ml of G418 and then screened for levels of the expression of TNFR1 mRNA by Northern blotting using a TNFR1 cDNA probe as indicated in Fig. 5A. These transformants were maintained in complete medium containing 0.3 mg/ml of G418.

TPA Induces Apoptosis in U937
Cells-Cells were treated with various concentrations of TPA. After 8 h, DNA was extracted from cells and electrophoresed in 2% agarose gel. Analysis of genomic DNA revealed a fragmentation pattern, characteristic of apoptosis, upon treatment with as low a concentration as 0.02 nM of TPA and distinct nucleosome ladders at 0.2 or 2 nM of TPA (Fig. 1, left panel).
To determine the kinetics of apoptosis by TPA, cells were treated with 2 nM of TPA and harvested sequentially at several different time points. DNA fragmentation was observed within 2 h after treatment with TPA and maximum induction of apoptosis occurred at 8 h (Fig. 1, right panel).
Induction of TNF␣ mRNA Expression by TPA in U937 and KY821 Cells-To study the effect of TPA on TNF␣ expression, we performed Northern blot analysis of total RNA using 32 Plabeled TNF␣ cDNA probe (Fig. 2). U937 and KY821 cells were treated with TPA at a concentration of 2 nM and harvested sequentially at different time points. The time course study showed that TNF mRNA expression was induced after 1 h with a peak at 4 h or 8 h after treatment in U937 cells (Fig. 2, left  panel). On the other hand, TPA slightly increased the levels of TNFR1 mRNA at detectable levels. TPA also increased levels of IL-1␤ RNA in these cells; an increase was detected at 4 h after exposure to TPA, and thereafter the induction of IL-1␤ mRNA occurred almost in a time-dependent fashion.
We also studied the levels of these mRNAs in TPA-treated KY821 cells (Fig. 2, right panel). The results obtained were similar to those in U937, although levels of TNFR1 were markedly increased and an increased level of IL-1␤ mRNA was observed at 1 h in KY821 cells.
TPA Induces TNF␣ Production in U937 Cells-We determined whether TPA affected translation of TNF␣ in U937 cells (Fig. 3A). The U937 cells were cultured for 8 h with TPA at different concentrations, and the levels of TNF␣ in the culture supernatants were determined by Western blotting using a TNF␣ polyclonal antibody (Fig. 3A). Supernatants from untreated U937 cells did not contain TNF␣ protein at a detectable level. An increased level of TNF␣ was observed at a concentration of 0.02 nM, and TPA stimulated the production in a dosedependent manner. The level of TNF␣ in cells treated with 20 nM TPA was approximately 10 times that of cells treated with 0.02 nM TPA. Studies of TNF␣ production using enzyme-linked immunosorbent assay were also performed. The results obtained were almost identical with those described above (data not shown). In parallel, cells were lysed and determined for expression of Fas or p53 (Fig. 3B). However, TPA did not affect expression of Fas or induce production of p53 in these cells. Northern blot analysis also showed that p53 expression was not detected in U937 cells, and TPA did not induce p53 in these cells (Fig. 3C).
TPA Induces Apoptosis through Production of TNF␣-To investigate the mechanisms of apoptosis induced by TPA, U937 cells were preincubated with a neutralizing antibody against human TNF␣ for 30 min, which neutralizes 1,000 units/ml of TNF␣ (Fig. 4A). Then 2 nM TPA was added, and the cells were cultured in the presence of anti-TNF␣ antibody. After 8 h, cells were harvested, and quantitative analysis of apoptosis by Hoechst staining was performed using cells cultured in medium alone as a control. Neutralization of the endogenous TNF␣ with anti-TNF␣ antibody almost completely blocked the induction of TPA-induced apoptosis. In parallel, cells were cultured with TNF␣ for 8 h. Exogenously added TNF␣ induced apoptosis in a dose-dependent fashion (Fig. 4B). TPA also induced expression of IL-1␤ in U937 cells as shown in Fig. 2. However, pretreatment of the cells with anti-IL-1␤ antibody for 30 min failed to inhibit the induction of apoptosis by TPA in these cells (Fig. 4C). Treatment of KY821 cells with anti-TNF␣ antibody also inhibited apoptosis induced by TPA (Fig. 4D).
Apoptosis Induced by TPA in TNFR1 Gene-transfected U937 Cells-To further clarify the role of TNF␣ in apoptosis induced by TPA, TNFR1 gene was stably transfected into U937 cells (Fig. 5A). The level of TNFR1 expression was determined by Northern blot analysis. We obtained several clones with different levels of TNFR1 mRNA. Fig. 5B shows one of these clones, and this clone had a 6-fold higher level of TNFR1 mRNA as compared with that of untreated U937 cells or control-transfected cells. We studied apoptosis in cells overexpressed with TNFR1. Control-transfected cells and cells transfected with TNFR1 were exposed to 2 nM of TPA for 8 h; quantitative analysis of apoptosis was performed by Hoechst staining ( Table  I). Transfection of cells with TNFR1 resulted in induction of apoptosis as compared with untreated cells (p Ͻ 0.05). Furthermore, cells overexpressed with TNFR1 underwent apoptosis following exposure to TPA more extensively as compared with that in control-transfected cells.
Role of the PKC and MEK Pathways in Induction of Apoptosis by TPA-Many extracellular stimuli including growth factors and stresses result in activation of phosphorylation cascades using MAPK. To examine the role of MAPK cascade in the production of TNF␣ and the induction of apoptosis by TPA, we used a specific inhibitor of MAPK kinase (also known as MEK), PD-98059. This compound is specific for MEK; it inhibits both the phosphorylation and activation MAPKs with no inhibitory activity against a number of other serine/threonine and tyrosine kinases (37). To examine the ability of the compound to inhibit MEK kinase activity induced by TPA, then cultured with 20 nM of TPA for 8 h (Fig. 6A). Activation of MEK kinase by TPA was completely blocked by pretreatment with these compounds. We next determined whether these compounds inhibit the production of TNF␣ (Fig. 6B) and apoptosis (Fig. 6C) induced by TPA. Pretreatment with PD-98059 attenuated by more than 90% the production of TNF␣ induced by TPA. Treatment with GF-109203X almost completely inhib-ited the TPA-induced production of TNF␣ in U937 cells. In parallel, we studied apoptosis in these cells. Cells were harvested and analyzed for apoptosis by Hoechst staining. Preincubation of cells with either GF-109203X or PD-98059 almost completely inhibited the induction of apoptosis by TPA. DISCUSSION The development and maintenance of tissues are achieved by several dynamically regulated processes that include cell proliferation, differentiation, and programmed cell death (51,52). When cells are exposed to external insults, negative regulation of the cell cycle or apoptosis occurs; cells which fail to repair DNA are eliminated through apoptosis (53,54). In this study, we have investigated the mechanisms of apoptosis by a potent stimulator of PKC, TPA in the human monocytic cell line U937 and the myeloblastic cell line KY821. Our results show that TPA induces apoptosis, at least in part, through a pathway which requires endogenous production of TNF␣ in these cells.
Our study showed that TPA stimulated production of TNF␣ in these cells, and pretreatment of cells with anti-TNF␣ neutralizing antibody abrogated apoptosis induced by TPA. Moreover, exogenously added TNF␣ also induced apoptosis in U937 cells. TPA also increased levels of IL-1␤ mRNA in U937cells. IL-1␤ is a cytokine, which is also produced by cells of monocytic lineage, and TNF␣ and IL-1␤ share a number of biological properties although they are molecularly distinct cytokines (55). However, treatment with anti-IL-1␤ antibody did not affect apoptosis induced by TPA. TNF␣ induces apoptosis through binding to TNFR1 (56). TNFR1 belongs to the same receptor family as Fas (APO/CD95) which is known to trigger apoptosis in a number of different cell types upon ligand binding (57). Studies have shown that both receptors contain an approximately 70-amino acid domain in the cytoplasmic region known as the "death domain," which is required for the cytotoxic effects (58,59). In this study, TPA increased levels of TNFR1 mRNA in U937 and KY821 cells. Moreover, the overexpression of TNFR1 enhanced the apoptosis induced by TPA in U937 cells whereas TPA did not affect levels of Fas. The cytotoxic effect of TNF␣ on neoplastic cells has been shown also to be synergistically enhanced by TPA in vitro (34). Taken together, these results suggest that endogenous production of TNF␣ plays a crucial role in apoptosis of these cells.
Death domain receptors including TNFR1 are constitutively expressed in various types of cells; these receptors are maintained in inactive state in the absence of ligands (60,61). Prior studies (60) have suggested that it is difficult to generate stable cell lines overexpressing death domain receptors because the overexpression leads to receptor aggregation and constitutive apoptosis. We stably transfected U937 cells with a human TNFR1 expression vector and obtained several clones with different levels of TNFR1 mRNA. The most highly TNFR1overexpressing U937 cells showed an approximately 6-fold TNFR1 level, as compared with that of control-transfected cells, with a slightly but significantly elevated basal level of apoptosis. A recent study has found a protein that protects against ligand-independent signaling by TNFR1 and other   (60). The mechanism responsible for the overexpression of TNFR1 in our U937 cells remains to be elucidated but may involve the expression of silence of death domain at a high level in these U937 cells.
Recent studies have demonstrated that an activation of PKCmodulated apoptosis though results are contradictory; TPA induces apoptosis in certain cells including U937 cells but inhibits apoptosis induced by various agents including ceramide (18,(21)(22)(23)(62)(63)(64). Other studies have also reported that activation of PKC by TPA prevented apoptosis induced by TNF␣ in U937 cells (65,66). These results suggest that whether apoptosis is induced in response to TPA depends on the type of cell. In this study, TPA induced apoptosis in U937 cells; these results were consistent with previous studies (15,23). We also showed that TPA stimulated MAPK activity and inhibition of PKC or MAPK pathway almost completely prevented the production of TNF␣ and the induction of apoptosis by TPA. Activation of the MAPK signaling pathway via PKC is an important mechanism for several biological events such as apoptosis, and PKC regulates the MAPK pathway, alone or with other mechanisms (67,68). Our results indicate that activation of the PKC and MAPK-signaling cascade leads to the accumulation of TNF␣, which is required for the induction of apoptosis by TPA. However, much evidence accumulated suggests that TPA and TNF␣ exert opposing effects on the induction of apoptosis in U937 cells (60).
TNF␣ stimulates sphingomyelin degradation with the subsequent formation of ceramide, and ceramide induces apoptosis in leukemia cells (69 -75). Moreover, there are reports that PKC activation is antagonistic to ceramide-induced apoptosis (64,75,76). Also TPA has been shown to activate sphingosine kinase, leading to increased concentrations of sphingosine-1phosphate (SPP), which is a metabolite of ceramide and suppresses ceramide-mediated apoptosis (65,77). Furthermore, it has been demonstrated that blocking the MAPK pathway prevented the cytoprotective effect of SPP on TNF␣-induced apoptosis in U937 cells (65). SPP stimulates MAPK activity and activation of MAPK promotes cell survival (77,78). Taken together, these studies suggest that activation of MAPK may stimulate different pathways responsible for the opposing effects on apoptosis and cell survival.
Leukemia cell lines including U937 and HL60 promyelomatic cells consist of phenotypically and biologically heterogeneous populations of cells that arrest at different stages of differentiation (79). A recent report has shown that certain PKC isotypes were down-regulated in differentiated cells (23). Although we provided evidence that TPA induced apoptosis through production of TNF␣ in U937 cells, the mechanism for the discrepancy between studies by other groups and us remains unclear. Studies suggest that the balance between the intracellular levels of ceramide and SPP is critical for the fate of cells (65,66). Considering these studies, the ceramide/SPP rheostat may be characteristic of cell type and, also, the differentiation-stage. Furthermore, the levels of products of PKC activation, TNF␣, and SPP may be varied in each U937 cell line, leading to opposing effects on programmed cell death.
Studies have shown that U937 cells have no wild type p53 transcripts or protein; a point mutation that converts G into A at the first base of intron 5 results in a partial deletion of 46 bases from the p53 mRNA (80,81). Our Western blot detected no p53 protein in U937 cells, and TPA did not induce p53. TNF␣-induced apoptosis can occur in cells deficient for p53 or with mutated p53 (82). Thus, TNF␣ may be of critical importance for apoptosis in U937 cells that lack functional p53. Furthermore, in this study, untreated U937 cells constitutively underwent apoptosis to an extent. Moreover, the constitutive apoptosis was augmented by the overexpression of TNFR1 in these cells. Whereas the TNF␣ protein in the supernatant of untreated cells was below the level of undetectability, enzymelinked immunosorbent assay showed that cell lysates of untreated U937 cells contained TNF␣ (0.39 g/mg protein and 20 pg/10 6 cells). Endogenously produced TNF␣ may induce apo-FIG. 6. Role of PKC or MEK pathway in apoptosis induced by TPA. Panel A, inhibitory effect of PD-98059 or GF-109203X on MEK1/2 kinase activity induced by TPA in U937 cells. U937 cells were incubated with either 100 M PD-98059 or 10 M GF-109203X for 1 h, and then cultured with 20 nM of TPA for 8 h. Cell extracts were immunoprecipitated with a phosopho-MEK1/2 antibody, and MEK was assayed with a recombinant p42 MAPK (ERK2) protein as a substrate. Kinase reactions products were separated on a 12% SDS-polyacrylamide gel, transferred to a polyvinylidene fluoride membrane, and then visualized by immunoblotting with a phospho-p44/42 MAPK (ERK1/ERK2) antibody. Panel B, effect of inhibition of PKC or MEK on production of TNF␣. After cell lysis, 20 g of whole cell protein was electrophoresed in a 12% SDS-polyacrylamide gel and transferred to a polyvinylidene fluoride membrane. Western blotting was performed using anti-TNF␣ antibody. Panel C, apoptosis induced by TPA in cells with MEK or PKC pathway blocked. In parallel, apoptosis was determined by Hoechst staining. The data are mean Ϯ S.D. from the results of three experiments. ptosis, in part, through autocrine loops in the endoplasmic reticulum.