¹Department of Microbiology and Immunology, East Carolina University School of Medicine, Greenville, NC, USA

²Member Leo Jenkins Cancer Center, East Carolina University School of Medicine, Greenville, NC, USA

³Escuela de Química y Farmacia, Facultad de Medicina, Universidad de Valparaíso, Valparaíso, Chile

⁴UCSF/Mt Zion Cancer Center, San Francisco, CA, USA

Correspondence to: J A McCubrey, Department of Microbiology and Immunology, East Carolina University School of Medicine, Brody Building 5N98C, Greenville, NC 27858, USA; Fax 252 816 3104

The Raf oncoprotein plays critical roles in the transmission of mitogenic signals from cytokine receptors to the nucleus. There are three Raf family members: A-Raf, B-Raf and Raf-1. Conditionally active forms of the Raf proteins were created by ligating N-terminal truncated activated forms to the estrogen-receptor (ER) hormone-binding domain resulting in -estradiol-inducible constructs. We introduced these chimeric Raf:ER oncoproteins into the murine FDC-P1 hematopoietic cell line. Two different types of cells were recovered after drug selection in medium containing either cytokine or -estradiol: (1) cytokine-dependent cells that expressed the Raf:ER oncoproteins; and (2) Raf-responsive cells that grew in response to the Raf:ER oncoprotein. Depending upon the particular Raf:ER oncoprotein, cytokine-dependent cells were recovered 10³ to 10⁵ times more frequently than Raf-responsive cells. To determine whether BCL2 could synergize with the Raf:ER oncoproteins and increase the frequency of cytokine-independent cells, cytokine-dependent Raf:ER-expressing cells were infected with either a BCL2 containing retrovirus or an empty retroviral vector. BCL2 overexpression, by itself, did not relieve cytokine dependency of the parental cell line. However, BCL2 overexpression increased the frequency of Raf-responsive cells approximately five- to 100-fold. Cytokine-dependent Raf:ER-infected cells entered the G₁ phase of the cell cycle after cytokine withdrawal and entered S phase only after cytokine addition. Raf-responsive Raf:ER cells entered the G₁phase of the cell cycle after estrogen deprivation and re-entered the cell cycle after addition of either IL-3 or the estrogen receptor antagonist tamoxifen which activates the Raf:ER constructs. Expression of the BCL2 oncoprotein often delayed the exit from the S and G₂/M phases demonstrating the protective effects BCL2 provided to these Raf and BCL2 infected cells. The Raf:ER cells expressed the Raf:ER proteins and downstream MEK and ERK activities after -estradiol treatment. Raf-responsive cells that were also infected with BCL2 expressed higher levels of BCL2 than the cells that were not infected with BCL2. Thus BCL2 can synergize with the activated Raf in the abrogation of cytokine dependency of certain hematopoietic cells. These cells will be useful in furthering our understanding of the roles of the Raf and BCL2 oncoproteins in hematopoietic cell growth, cell cycle progression and prevention of apoptosis. Leukemia (2000) 14, 1060-1079.

Raf; BCL2; apoptosis; cytokine dependency

Introduction

The proliferation of many hematopoietic precursor cells is promoted by interleukin-3 (IL-3), granulocyte/macrophage colony-stimulating factor (GM-CSF), as well as other growth factors.^{1,2,3,4,5,6,7,8,9,10} The murine IL-3/GM-CSF dependent FDC-P1 cell line was isolated from the bone marrow of a normal DBA/2 mouse and resembles cells with a CFU granulocyte/ macrophage morphology.⁹ The loss of cytokine dependency by hematopoietic cells may be an important factor in the development of leukemia.^1,2 Spontaneous factor-independent cells are rarely recovered from this cell line which makes it an attractive model system to analyze the effects various oncogenes have on signal transduction and leukemogenesis.^10,11,12,13

IL-3 and GM-CSF exert their biological activity by binding to the IL-3 and GM-CSF receptors respectively (IL-3R and GM-CSFR).^{1,2,3,4,5,6,7,8,14,15,16} These receptors activate a Janus (Jak2) protein tyrosine kinase which leads to the phosphorylation and dimerization of signal transducers and activators of transcription (STATs). In addition to Jak activation, receptor ligation promotes phosphorylation of the Shc protein. Shc then recruits the Grb2/Sos complex to the _c chain resulting in stimulation of Ras (see Figure 1). Activated Ras promotes the sequential activation of Raf, MEK and MAP (ERK1 and ERK2) kinases.¹ Certain cytokines and interferons, stimulate Jak1 activity which can activate the Raf kinase pathway indicating that there is cross-talk between these two major signaling pathways.^15,16,17,18 Moreover for optimal activation of STAT DNA binding activity, STAT dimers must be phosphorylated on key serine residues. This serine phosphorylation enhances STAT dimer nuclear translocation and is in some cases carried out by the MAP kinases ERK1 and ERK2.^15,16,17 Thus these signal transduction pathways are interconnected and often regulate each other.

There are three related raf genes in mammals: A-raf, B-raf and raf-1.^1,19 The Raf proteins have been dissected into three different functional domains; CR1, CR2, and CR3. The CR1 region has the binding site for an activated Ras protein.^1,18 The CR2 region negatively regulates the Raf kinase domain (CR3) which is located in the carboxyl-terminal half of the Raf protein. Raf proteins transmit their regulatory signals to MEK1, a dual specificity serine/threonine and tyrosine kinase which phosphorylates the downstream MAP kinases.^{1,20,21,22,23,24} The MAP kinases ERK1 and ERK2 can phosphorylate other kinases (eg p90^Rsk) or transcription factors (eg CREB, Elk) which enter the nucleus and regulate gene expression.^{25,26,27,28,29,30,31} N-terminal deleted forms of Raf and MEK1 proteins result in activated oncoproteins presumably due to the aberrant stimulation of downstream kinases, transcription factors and molecules involved in the prevention of apoptosis.^{32,33,34,35,36}

The efficiencies of hematopoietic cell transformation by activated Raf oncoproteins are much lower than that observed with v-Ha-Ras and other oncoproteins.^13,16,37 The nature of the inability of Raf to abrogate the cytokine dependency of most hematopoietic cells has not been elucidated. We are interested in the transforming capacity of Raf because it is located at pivotal positions in both signal transduction as well as apoptotic pathways.

The BCL2 gene has been described as a repressor of cell death.^{38,39,40,41,42,43,44,45,46,47} This gene was isolated from the chromosomal breakpoint of a t(14;18)-bearing follicular B cell lymphoma.^38,39 B cells bearing this translocation exhibit increased production of BCL2 protein and extended cell survival.^38,39 BCL2 has been shown to protect hematopoietic cell lines from apoptosis following growth factor withdrawal.⁴²

Cytokine stimulation has been documented to suppress apoptosis via different mechanisms. The carboxyl region of the IL-3/GM-CSF/IL-5 R _c chain has been associated with the prevention of apoptosis.^1,8,43 Stimulation of appropriate target cells by IL-3 leads to phosphorylation of Bad.^{1,15,48,49,50,51} Bad may become phosphorylated on either serine residues 112 or 136. Bad is phosphorylated on serine 112 via the MAPK or PI3K signal transduction pathways. The serine residue 136 may be reserved only for the phosphorylation by the PI3K pathway.^48,51 Unphosphorylated Bad normally forms heterodimers with anti-apoptotic factors, such as Bcl-X_L or BCL2. Phosphorylation of Bad releases Bad from Bcl-X_L and BCL2 thereby allowing Bcl-X_L and BCL2 to bind Bax, resulting in the prevention of apoptosis. The kinase(s), which directly phosphorylates Bad, is unknown, however, Raf, ERK2, p90^Rsk, Akt (PKB) and protein kinase A (PKA) can phosphorylate Bad.^{47,48,49,50,51} Phosphorylated Bad is bound and sequestered by the 14-3-3 family of proteins preventing its interaction with Bcl-X_L or BCL2 (see Figure 1). Recent evidence indicates that ERK2, can phosphorylate BCL2 on serine residue 70 and this results in an anti-apoptotic response.^{15,38,44,52,53} Furthermore, BCL2 is phosphorylated after treatment of cells with chemotherapeutic drugs such as taxol^{53,54,55,56,57,58,59} and retinoic acid. These BCL2 phosphorylation events are associated with cell death. The time frames of these phosphorylation events are different as cytokines stimulate BCL2 phosphorylation early after treatment, whereas treatment with microtubule destabilizing drugs such as taxol result in phosphorylation of BCL2 later in the G₂/M transition.^53,54,55,56 Thus regulation of BCL2 is complex with some phosphorylation events resulting in anti-apoptotic effects, whereas others are associated with cell death.

v-Raf has been shown to suppress apoptosis by IL-3 withdrawal in the IL-3-dependent 32D.3 cell line.^57,58 Another study in these cells demonstrated that BCL2 could be co-immunoprecipitated with v-Raf suggesting a functional link between the Raf kinase and the BCL2 protein in the suppression of apoptosis.⁵⁹ Bag-1, discovered for its ability to interact with BCL2, also forms complexes with Raf-1.⁶⁰ A complex comprised of BCL2, Raf-1, and Bag-1 proteins has been hypothesized either to activate a kinase on the mitochondrial membrane or to target other kinases to the mitochondrial membrane.⁶¹ Recently ERK2 has been shown to localize to the mitochondrial membrane.^44,61,62

The influence of Raf on apoptosis could lead to information of clinical value. Suppression of apoptosis in neoplastic cells is being considered as a means by which transformed cells avoid the activity of apoptosis-inducing chemotherapeutic regimens. BCL2 has also been shown to influence drug resistance and the interactions between Raf-1 and BCL2 implicate the Raf kinases in drug resistance. Thus, the study of signal transduction pathways involving Raf kinases are not only important for understanding normal cell signaling, but also may lead to development of treatments that circumvent drug resistance in leukemic cells.

Materials and methods

Cell lines and growth factors

Cells were maintained in a humidified 5% CO₂ incubator with Iscove's modified Eagle's medium (IMEM, Life Technologies, Gaithersburg, MD, USA) complemented with 5% fetal bovine serum (BCS) (Atlanta Biologicals, Atlanta, GA, USA). The IL-3/GM-CSF dependent murine myeloid cell line FDC-P1⁹ was cultured in this medium supplemented with 10% WEHI-3B(D^-) conditioned medium (WCM) as a source of IL-3. Estradiol-dependent FD/Raf:ER + BCL2 cells were grown in IMEM + 5% FCS + 1 M -estradiol (Sigma, St Louis, MO, USA).

In some cases Raf:ER-infected cells were also cultured with 4-OH tamoxifen, also dissolved in ethanol (4-HT, Sigma). 4-HT is an estrogen-receptor antagonist that activates the Raf:ER fusion proteins.^{32,33,34,35,36,37} Raf:ER-infected cells were also treated with the MEK1 inhibitor PD98059 (New England Biolabs (NEB), Beverly, MA, USA)^63,64 which was dissolved in dimethyl sulfoxide (DMSO, Sigma). Control cultures were set up with DMSO and either cytokine or -estradiol to measure the toxicity of the solvent. Cells were in some cases treated with the PKC-inducing agent, 10 to 20 nM phorbol-12-myristate-13-acetate (PMA, Sigma) which was dissolved in DMSO.

Retroviral infection of cells

Plasmid DNAs containing recombinant retroviruses were transfected into the retroviral packaging cell lines 2 or PA317 with lipofectin (Life Technologies) and retroviruses were passed sequentially from one cell line to the other to amplify their titers as described.^6,11,12,13 FDC-P1 cells were infected with viral stocks as described.^6,11,12,13 The A-Raf:ER, B-Raf:ER, Raf-1:ER, and a kinase-inactive Raf-1[301]:ER were contained in the pBP3puro retroviral vector which encodes resistance to puromycin.³⁴ Cells were also infected with a Raf-1:ER which was ligated to the green fluorescent protein (GFP).⁶⁵ Puro^r cells were isolated by culture in medium containing 1 g/ml puromycin (Sigma). Some FDC-P1 cells were infected with either LNL6, an empty retroviral vector containing neo^r,⁶⁶ or BCL2 contained in a neo^r vector.⁴² neo^r FDC-P1 cells were isolated by selection in medium containing G418 (2 mg/ml, Life Technologies). The nomenclature of the FDC-P1 cells is FD/Raf-1:ER, FD/A-Raf:ER, FD/B-Raf-1:ER, for cells infected with the Raf-1:ER, A-Raf:ER, or B-Raf:ER viruses, respectively. After the ER is the designation + BCL2 or + LNL6 for cells subsequently infected with the BCL2 and LNL6 retroviruses and whether the cells are IL-3 dependent (IL3) or estrogen-responsive (Est). Estrogen-responsive growth in these cases is proliferation promoted by the activated Raf:ER. C refers to a clone isolated by limiting dilution and pool refers to a pool of cells.

Assays of cell growth: cell growth curves, [3H]-thymidine incorporation and autocrine growth factor synthesis

Cells were washed three times with phosphate buffered saline (PBS) before being set up for growth curves or proliferation assays. Growth curves were performed in 5 ml of medium with the indicated supplements in T25 tissue culture flasks (Corning, Elmira, NY, USA). Cell proliferation assays were performed in 96 well flat bottom plates (Corning) with initially 10⁴ cells/well that were incubated for 1 day in the presence of the indicated supplements. The supplements could include different dilutions of -estradiol, tamoxifen, IL-3 or various concentrations of filter-sterilized cell supernants to determine whether or not the cells were producing autocrine growth factors. In some cases the supernatants were incubated with different concentrations of antibodies to IL-3 or GM-CSF (R&D Systems, Minneapolis, MN, USA) for 1 h before the indicator cells were added to determine which cytokine the cells were expressing. Also, in some cases the cytokine producing cells were incubated with the IL-3 and GM-CSF antibodies to determine the effects of these antibodies on cell growth. During the last 6 h of incubation, cellular proliferation was assayed by adding [³H]-thymidine (6.7 Ci/mmol, NEN, Boston, MA, USA) as described.^6,11,12,13

Analysis of cell cycle distribution

Cell pellets were prepared from cells isolated at the appropriate incubation times which were resuspended in PBS and then fixed in 70% ethanol overnight. The next day the cells were collected by centrifugation and washed with PBS containing 1% BSA. The cell pellets were resuspended in 0.5 ml of solution containing propidium iodide, Rnase and Triton X-100 and stained for 30 min in the dark. The cell suspension was passed through a 25-gauge needle before cell cycle analysis which were performed on a Becton Dickinson FACS analyzer. The percentage of cells in different phases of the cell cycle was statistically estimated using a ModFit 5.02 computer program.

Analysis of GFPRaf-1:ER protein expression

The expression of the GFP/Raf-1:ER protein was monitored by FACS analysis. The cells were washed once with PBS and then analyzed for fluorescence intensity on the Becton Dickinson FACS analyzer. Uninfected FDC-P1 and Raf-responsive FD/Raf-1:ER(Est)c2 cells were used as negative controls since they lacked the GFP moiety.

Polymerase chain reaction amplification of cytokine mRNA transcripts

Total cytoplasmic RNAs was prepared as described¹² and 1 g was included in a 20 l cDNA synthesis reaction containing: reverse transcriptase buffer, 1 mM of each dNTP, 20 g/ml oligo-dT and 20 units Mo-MuLV reverse transcriptase. After incubation at 42°C for 40 min, the reaction was terminated by addition of H₂O. For PCR amplification, 5 l of cDNA were included in a 50 l reaction mixture containing PCR buffer, dNTPs, 1-2 units Taq polymerase and 1 M of each oligonucleotide primer. The primers for murine IL-3 were: 5'AATCAGTGGCCGGGATACCC3' and 5'CGAAATCATCCAGATCTCG3' defining a 200 bp cDNA fragment which could readily be distinguished from a genomic IL-3 DNA fragment by size (>1 kb). The primers for murine GM-CSF were: 5'CCTGAGGAGGATGTGGCTGC3' and 5'CTGTCCAAGCTG AGTCAGCG3' defining a 601 bp fragment. The primers for murine ₂-microglobulin were 5'TTCTCTCACTGACCGGCCT G3' and 5'CAGTAGACGGTCTTGGGCTC3' defining a 308 bp fragment. Forty cycles of PCR were performed to detect cytokine cDNAs. The PCR products were electrophoresed on 1% agarose gels and visualized after ethidium bromide staining of the gel.

Preparation of cell extracts and analysis by Western blotting

Cells were washed twice with cold PBS and lysed on ice in gold lysis buffer (GLB) containing 20 mM Tris (pH 7.9), 137 mM NaCl, 5 mM Na₂EDTA, 1% (vol/vol) Triton X-100, 15% (vol/vol) glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 M pepstatin A, 1 mM sodium orthovanadate, 1 mM ethylene glycol-bis (-aminoethyl ether)- N,N,N',N'tetra-acetic acid (EGTA), 10 mM sodium fluoride, 1 mM tetrasodium PP_i, and 100 M -glycerophosphate. Insoluble material was removed by centrifugation at 15 000 g as described.^{16,32,33,34,35,36} All chemicals were purchased from Sigma unless otherwise indicated.

Cellular proteins were analyzed by electrophoresis through polyacrylamide SDS gels followed by Western immunoblotting on to polyvinylidene difluoride membranes (PVDF, Immobilon P; Millipore, Bedford, MA, USA). Western blots were incubated with the appropriate primary Ab at a dilution of 1:1000 to 1:2000 and then washed in Tris-buffered saline containing 0.5% (vol/vol) Nonidet P-40. Antigen-antibody complexes were visualized by using 1:10 000-diluted goat -rabbit antiserum, sheep -mouse antiserum, or protein A coupled to horseradish peroxidase as indicated in the enhanced chemiluminescence detection system (Amersham, Arlington Heights, IL, USA) as described.^{16,32,33,34,35,36} The blots were exposed to Kodak XAR5 X-ray film (Kodak, Rochester, NY, USA) for 30 to 60 s. The BCL2 antibody (Santa Cruz Biotechnology (SCB), Santa Cruz, CA, USA) was directed to the N-terminus of the human BCL2 protein and recognizes both the human and murine BCL2 protein.

Determination of Raf kinase activity

The cells were grown in phenol red free DMEM (Life Technologies) containing charcoal-stripped FCS and the indicated supplements. Immune complexes of Raf:ER proteins were prepared by incubating cell lysates with an -hbER Ab (SCB) and protein A-sepharose 4B beads. Immune complexes were collected, washed three times in lysis buffer, and the activities of Raf:ER were assessed by incubation for 30 min at 30°C in a reaction mix containing 25 mM HEPES (pH 7.4), 10 mM MgCl₂, 1 mM DTT, 50 M ATP and 10 M [-³²P]-ATP (3000 Ci/mmole, NEN) and 15 g of purified recombinant enzymatically inactive GST-MEK1 (Upstate Biotechnology (UPI), Lake Placid, NY, USA) as a substrate as described previously.^16,35,36 The reactions were analyzed by polyacrylamide gel electrophoresis and electrotransfer. The membranes were first exposed to X-ray film to quantitate GST-MEK1 phosphorylation and subsequently probed with an -hbER Ab to quantitate the amount of the Raf:ER kinases present in each immunoprecipitate.

Determination of MEK activity

MEK activity was measured by immunoprecipitating the MEK1 proteins overnight with an MEK1 Ab (SCB) and protein A-sepharose 4B beads as described.³⁵ The immunoprecipates were then used in a kinase reaction for 30 min at 30°C in a reaction mix containing: 100 mM HEPES pH 7.4, 25 mM MgCl₂, 1 mM DTT, 50 M ATP, 10 M [-³²P]-ATP and 2 g of bacterially expressed rp44, an enzymatically inactive form of MAP kinase (ERK2, UBI). The reaction mixtures were electrophoresed through a 10% polyacrylamide gel, transferred to PVDF and exposed to X-ray film.

Determination of p42/44MAPK (ERK1 and ERK2) activation and enzymatic activity

Cells were deprived of either IL-3 or -estradiol for 24 h in phenol-red free medium that contained 5% charcoal-stripped BCS. Then the cells were pulsed with cytokine, different concentrations of -estradiol or the positive control, 10 nM PMA, for varying periods of time. Cells were washed and resuspended in serum-free medium. One ml, containing 1.25 ´ 10⁶ cells, was added to microfuge tubes and placed at 37°C for at least 1 h prior to the start of the treatment. Cells were stimulated or mock-stimulated by the addition of PMA, -estradiol, IL-3, ethanol or DMSO to the tubes in a 10 l volume. Following stimulation for the indicated time, the tubes were centrifuged in a microcentrifuge for 30 s, the supernatants were removed, cell pellets were resuspended in 110 l of cold lysis buffer (25 mM Tris-HCl, pH 7.4; 50 mM NaCl; 0.5% sodium deoxycholate; 2% NP-40; 0.2% SDS; 1 mM phenymethylsulfonyl fluoride PMSF; 50 g/ml aprotinin, 50 M leupeptin; 0.5 mM Na₃VO₄) and placed on ice for 15 min. Lysates were centrifuged for 15 min at 14 000 g in an Eppendorf microcentrifuge, supernatants (98 l) were removed and mixed with 42 l of 3.3´ sample buffer (200 mM Tris-HCl, pH 6.8; 33% glycerol; 6.6% SDS; 16.6% -mercaptoethanol; 0.04% bromophenol blue). Samples were boiled (5 min) and frozen. Fifteen microliters of prepared samples were electrophoresed through a 10% SDS-PAGE gel and proteins were electrophoretically transferred to PVDF membranes. Membranes were incubated overnight at 4°C in blocking buffer (25 mM Tris-HCl, pH 8.0; 125 mM NaCl; 0.1% Tween 20; 1% BSA; 0.1% sodium azide). Membranes were then incubated for 2 h with the primary antibody diluted in blocking buffer (anti-active MAP kinase, 1:20 000 (Promega, Madison, WI, USA) or anti-p90^Rsk (Santa Cruz), 1:10 000). The blots were washed twice in TBST (25 mM Tris-HCl, pH 8.0; 125 mM NaCl; 0.025% Tween 20) and incubated with alkaline phosphatase (AP)-conjugated goat anti-rabbit Ig (Promega, 1:10 000 in TBST) for 1 h at room temperature. The blots were washed twice in TBST and developed with the colorogenic substrates BCIP and NBT (Promega protoblot AP system). The data shown are representative of at least two independently performed experiments.

ERK2 enzymatic activity was determined in the stimulated cells after immunoprecipitation with an ERK2 antibody (SCB) and an in vitro radioactive kinase assay was performed with 43 mM Hepes pH 7.4, 43 mM MgCl₂, 200 M ATP, 2.3 g myelin basic protein (MBP, Promega), 10 M [-³²P]-ATP and 0.5 mM DTT.³⁵

Results

Isolation of Raf-responsive cells after BCL2 infection of cytokine-dependent Raf:ER-infected cells

The overall design for the following experiments was to superinfect cytokine-dependent Raf:ER-infected FDC-P1 cells, with either an empty retroviral vector or a BCL2-containing retrovirus and determine whether BCL2 overexpression increased the frequency of isolation of Raf-responsive cells. As controls for the following experiments, the ability of a mock infection or infection with either BCL2 or the empty LNL6 retroviral vector to abrogate the cytokine dependency of FDC-P1 cells was determined. No viable cells were recovered after mock infection and culture in medium containing IL-3 + G418, -estradiol + G418, or G418 indicating that the selection conditions were sufficient to kill uninfected cells. Furthermore neither BCL2 over-expression nor infection with an empty retroviral vector abrogated the cytokine dependency of FDC-P1 cells (Table 1).

Previously we have demonstrated that after infection of FDC-P1 cells with retroviruses encoding A-Raf:ER, B-Raf:ER, and Raf-1:ER, cytokine-dependent puro^r cells were recovered.³⁷ The clones were determined to contain the Raf:ER proteins by Raf kinase and Western blot assays (Figure 2). The Raf:ER expression was induced by -estradiol. Therefore, expression of Raf:ER is increased in the presence of -estradiol possibly due to an increase in protein stability. An example of the Raf kinase assay and subsequent Western blot is presented in Figure 2. Endogenous Raf activity is induced by the addition of IL-3. However, the Raf assay that was used in these experiments was specifically designed to detect Raf activity associated with the Raf:ER proteins as an ER Ab protein is used to immunoprecipitate the Raf:ER proteins and then a Raf assay is preformed. Equal loading of the proteins was determined by electrophoresis of the same samples and performing Western blot analysis with antibodies to the MEK1 and ERK2 proteins.

The cytokine-dependent FD/A-Raf:ER(IL3)c2, FD/B-Raf:ER(IL3)c1 and FD/Raf-1:ER(IL3)c2 were subsequently infected with either a BCL2-containing retrovirus or the empty retroviral vector LNL6. Raf-responsive cells were isolated by plating the cells in the presence of G418 and -estradiol in the absence of exogenous cytokines (Table 1). More (four- to eight-fold) Raf-responsive cells were recovered after infection of FD/A-Raf:ER, FD/B-Raf:ER and FD/Raf-1:ER cells with the BCL2 than after infection with the LNL6 empty retroviral vector.

To determine whether similar results could be obtained by asking the question in reverse, cytokine-dependent FD/LNL6 and FD/BCL2 cells were infected with the Raf-1:ER retrovirus and the frequency of Raf-responsive cells determined. As controls, the abilities of either a mock infection or empty retroviral vector infection to relieve the cytokine dependency of FD/LNL6 or FD/BCL2 cells were determined. No Raf-1-responsive cells were isolated from either mock-infected or empty vector (pBP3puro) infected cells when the cells were plated in either IL-3 + puromycin or -estradiol + puromycin. Furthermore the ability of a kinase-inactive Raf-1[301]:ER mutant to abrogate the cytokine dependency of FD/LNL6 or FD/BCL2 cells was determined. Infection with this mutant Raf-1-containing retrovirus did not abrogate the cytokine dependency of either cell line. Although some Raf-responsive cells were isolated from Raf:ER-infected FDC-P1 cells, 3.8-fold more Raf-responsive cells were recovered from the BCL2 and Raf-1:ER-infected cells (Table 2).

Cytokine-dependent FD/LNL6 + FD/BCL2 cells were also infected with a GFPRaf-1:ER construct containing the green fluorescent protein (GFP) linked to the Raf-1:ER. After infection of FD/BCL2 cells with the GFPRaf-1:ER retrovirus, a similar frequency of Raf-responsive cells was recovered as after infection with the Raf-1:ER retrovirus (Table 2). The GFP tag in the GFPRaf-1:ER construct allows rapid examination of the GFPRaf-1:ER expression by FACS analysis in cells which are cytokine-dependent vs those that are Raf-responsive (Figure 3). As controls, the levels of background fluorescence were determined in cells which did not have the GFPRaf-1:ER (Figure 3a-d). These cells had a single peak of fluorescence as determined by FACS analysis which did not significantly change when the cells were cultured in IL-3 or IL-3 + -estradiol. The expression of the GFPRaf-1:ER protein was heterogenous in the cytokine-dependent FD/BCL2+GFPRaf-1:ER (IL3) pool, with about one-half of the cells positive for expression and one-half negative for GFPRaf-1:ER expression (Figure 3e). Culture of these cells for 24 h in the presence of IL-3 + -estradiol increased GFPRaf-1:ER expression as a more fluorescent shoulder was present (Figure 3f). In contrast, the Raf-responsive FD/BCL2+GFPRaf-1:ER(Est) clones were more fluorescent and homogenous than the cytokine-dependent cells. Two Raf-responsive clones are shown in Figure 3, similar results were observed with 12 other clones.

Frequency of isolation of Raf-responsive cells from pools of Raf:ER and BCL2-infected FDC-P1 cells

To determine the frequency of isolation of Raf-responsive cells from Raf:ER and BCL2 infected FDC-P1 cells, limiting dilution analyses were performed on the cells isolated in the presence of IL-3 + G418 (Figure 4). As controls, the abilities of FDC-P1 cells infected with the empty retroviral vector (Figure 4a) or the BCL2 (Figure 4b) viruses to give rise to factor-independent cells were determined. No cytokine-independent cells were isolated from these infected cells. In Figure 4c, e and g, limiting dilution analysis of Raf:ER and empty retroviral vector (LNL6) infected cells are presented. In Figure 4d, f and h, the limiting dilution analysis of Raf:ER and BCL2 infected cells are presented. Depending upon which Raf:ER oncoprotein the cells were infected with, BCL2 overexpression increased the frequency of isolation of FD/Raf:ER cells that grew in response to Raf:ER activation five- to 100-fold (Figure 4d, f and h) in comparison to cells that were infected with the empty retroviral vector (Figure 4c, e and g) (Table 3). The major reason why there were higher frequencies of Raf-responsive cells in these experiments as opposed to those presented in Tables 1 and 2 is that pools of 100% infected cells were examined, whereas in Tables 1 and 2, not all the cells were infected with either BCL2 or LNL6.

Thus, six different types of cells were isolated after these infection experiments: (1) cells which were infected with the empty retroviral vectors (LNL6 or pBP3puro) and remained cytokine-dependent; (2) cells which expressed just the BCL2 oncogene that remained cytokine-dependent; (3) cells which expressed the Raf:ER genes that remained cytokine-dependent; (4) cells which expressed the BCL2 and Raf:ER genes and remained cytokine-dependent; (5) cells which expressed the Raf:ER genes that grew in response to the Raf proteins in the absence of exogenous cytokines; and (6) cells which expressed the BCL2 and Raf:ER genes that grew in response to the Raf proteins in the absence of exogenous cytokines.

Growth properties of BCL2 and Raf:ER Infected FDC-P1 cells

To determine the growth properties of the cells which overexpressed the BCL2 and Raf:ER oncoproteins, growth curve experiments were performed. Cytokine-dependent Raf:ER-infected cells required IL-3 to proliferate (Figure 5a, d and g). Addition of -estradiol did not stimulate growth in these cells. In contrast, -estradiol did promote cell growth in the Raf-responsive FD/A-Raf:ER+LNL6(Est)c2 (Figure 5b), FD/B-Raf:ER+LNL6(Est)c1 (Figure 5e) and FDRaf-1:ER+LNL6(Est)c2 (Figure 5h) cells, however, these conditionally-transformed Raf-responsive cells did not grow in the absence of -estradiol or IL-3 (Figure 5b, e and h). Likewise, IL-3 and -estradiol stimulated the growth of the Raf-responsive FD/A-Raf:ER+BCL2(Est)c2 (Figure 5c), FD/B-Raf:ER+BCL2(Est)c1 (Figure 5f) and FD/Raf-1:ER+BCL2(Est)c2 (Figure 5i) cells.

Ability of -estradiol and IL-3 to stimulate [3H]-thymidine incorporation in BCL2 and Raf:ER-infected hematopoietic cells

[³H]-thymidine incorporation in response to -estradiol treatment is a measurement of the ability of the activated Raf kinases to promote DNA synthesis. As controls, the sensitivities of the empty vector infected cells (FD/LNL6, Figure 6a and b) and BCL2-infected cells (FD/BCL2, Figure 6c and d) to IL-3 and -estradiol were examined. These cells incorporated [³H]-thymidine in a dose-dependent fashion in response to IL-3 (Figure 6a and c) but significant [³H]-thymidine incorporation was not observed in response to -estradiol (Figure 6b and d). In contrast, the Raf-responsive cells incorporated [³H]-thymidine in a dose-dependent fashion in response to IL-3 (Figure 6e, g, i, k, m and o) and -estradiol (Figure 6f, h, j, l, n and p). The FD/A-Raf:ER+LNL6(Est)c2 and A-Raf:ER+ BCL2(Est)c2 cells showed a gradual increase in response to -estradiol starting at 3 nM and attaining maximal levels of [³H]-thymidine incorporation around 100 to 1000 nM (Figure 6f and h, and data not presented). In contrast, the Raf-responsive FD/B-Raf:ER+LNL6(Est)c1 (Figure 6j) and FD/B-Raf:ER+BCL2(Est)c1 (Figure 6l) cells incorporated significant levels of [³H]-thymidine around 30 nM and reached a plateau at 100 nM -estradiol (Figure 6j and l). Interestingly once the B-Raf:ER-infected cells had been maintained in -estradiol, they showed a decrease in their responsiveness to IL-3 as compared to other cell lines (Figure 6i and k). In contrast the FD/Raf-1:ER+LNL6(Est)c2 (Figure 6n) and FD/Raf-1:ER+BCL2(Est)c2 (Figure 6p) infected cells only incorporated [³H]-thymidine at higher doses of -estradiol (100 nM) (Figure 6n and p). Thus the different Raf genes influenced the sensitivities of the cells to -estradiol in different dose responses as measured by [³H]-thymidine incorporation.

Stimulation of [3H]-thymidine incorporation by anti-estrogens

To determine whether the estrogen-receptor antagonist 4-OH tamoxifen (4-HT) also promoted growth, similar dose-response experiments were performed (Figure 7). Raf-responsive FD/A-Raf:ER+BCL2(Est)c2 (Figure 7a), FD/B-Raf:ER+BCL2(Est)c1 (Figure 7b), and FD/Raf-1:ER+BCL2(Est)c2 (Figure 7c) cells incorporated [³H]-thymidine in response to 4-HT. FD/A-Raf:ER+BCL2(Est)c2 and FD/Raf-1:ER+BCL2(Est)c2 cells incorporated maximal amounts of [³H]-thymidine around 50 nM, whereas the FD/B-Raf:ER+BCL2(Est)c1 cells incorporated maximal levels at approximately 10-fold lower doses. These experiments eliminated the possibility that endogenous estrogen-receptors functioned either alone or in synergy with the activated Raf:ER proteins to stimulate the proliferation of the Raf-responsive FD/Raf:ER + BCL2 cells and again showed that the proliferation of B-Raf:ER-infected cells was more sensitive to Raf:ER activation.

Effects of Raf:ER and BCL2 expression on cell cycle progression

The Raf and downstream MEK and ERK activities have been shown to influence cell cycle progression by regulating the expression of the p21^Cip1 cell cycle inhibitory protein as well as growth stimulatory proteins.^1,16,65,67 Cell cycle progression can be examined by starving cytokine-dependent cells for IL-3 or in this case -estradiol, and then determining the extent of cell cycle progression after addition of IL-3 or tamoxifen. In these experiments, tamoxifen was used in place of -estradiol because tamoxifen is a pure estrogen receptor antagonist and therefore the effects of estrogen on cell cycle progression will not confuse the effects of Raf on cell cycle progression. Exponentially growing cells were collected, washed three times with PBS and cultured in phenol red free medium containing 5% FCS in the absence of IL-3 and -estradiol. -24 h indicates the start of the experiments. Then after 24 h of -estradiol-deprivation, IL-3, tamoxifen, or nothing was added to the cells and cell aliquots were removed every 12 h. A diagrammatic example of the effects of Raf-1:ER on cell cycle progression is presented in Figure 8.

We have chosen two examples, Raf-responsive FD/Raf-1:ER+LNL6(Est)c2 cells (Figure 8 upper panels) and Raf-responsive FD/Raf-1:ER+BCL2(Est)c2 cells (Figure 8 lower panels) to show how these two cell lines differ in the distribution of the cells in the various cell cycle phases. Cycling cells are shown in the Figure 8a (A and B). There are significant proportions of the cells in the G₁ phase (black histograms peaks), S phase (stripped histogram peaks) and G₂/M (open histogram peaks) phases, the actual percentages are presented in Figure 9. Cells deprived of -estradiol for 12 h (-12 h) are shown in Figure 8b (A and B). Cells deprived of -estradiol for 24 h are shown in the Figure 8c (A and B). The Raf-responsive FD/Raf-1:ER+LNL6(Est)c2 exit the S and G₂/M phases after 12 to 24 h of -estradiol deprivation (Figure 8A (b and c). In contrast, a significant proportion of the FD/Raf: ER+BCL2(Est)c2 cells remained in the S and G₂/M phases after 12 or 24 h of -estradiol withdrawal (Figure 8B (b and c). However, after culture of these cells for 48 or 72 h in the absence of -estradiol, the cells were predominantly in the G₁ phase (Figure 8B (d and g). Once either IL-3 or tamoxifen was added, both types of Raf-responsive cells entered the cell cycle (Figure 8A and B) (e, h, f, i).

The cell cycle progression profiles in the cytokine-dependent and Raf-responsive Raf and BCL2 infected FDC-P1 cells are presented graphically in Figure 9. These experiments represent the average of three to six time course experiments. Initially (designated -24h) approximately 25% of the IL-3-dependent and Raf-responsive cells were in the S phase of the cell cycle (Figure 9, A2, A5, A8, B2, B5, B8, C2, C5 and C8). This percentage of cells in S phase is a characteristic of FDC-P1 cells when they are proliferating. After 12 h of cytokine deprivation (designated -12), approximately 14% of the cytokine-dependent Raf:ER-infected cells were in the S phase of cell cycle (Figure 9A2, B2 and C2). In contrast, the percentage of cells in the G₁ phase increased from approximately 62 to 75% after 12 h of deprivation (Figure 9A1, B1 and C1).

The level of cytokine-dependent cells in S phase dropped to approximately 5% after 24 h of cytokine deprivation (Figure 9A2, B2 and C2). Likewise the percentage of cytokine-dependent cells in G₂/M decreased from approximately 10 to 5% (Figure 9A3, B3 and C3) and the percentage of cells in G₁ increased to approximately 90% (Figure 9A1, B1 and C1). Thus removal of cytokine for 24 h resulted in the cytokine-dependent cells exiting the S and G₂/M phases and entering the G₁ phase.

After 24 h of cytokine-deprivation, at T₀, the cells were treated with IL-3, tamoxifen or nothing and the percentages of cells in the different phases of the cell cycle were determined at 12 h intervals. The main difference between the cytokine-dependent FD/Raf:ER, Raf-responsive FD/Raf:ER + LNL6 and Raf-responsive FD/Raf:ER+BCL2 cells was that very few cytokine-dependent FD/A-Raf:ER, FD/B-Raf:ER or FD/Raf-1:ER cells entered the cell cycle after tamoxifen treatment whereas the Raf-responsive cells entered S phase after tamoxifen treatment (Figure 9A4, B4 and C4). It should be noted that a small percentage of the cytokine-dependent FD/Raf-1ER(IL3)c1 cells entered S phase (12%, Figure 9C2) and G₂/M phases (5%, Figure 9C3) after tamoxifen treatment.

The effects of estradiol deprivation were examined in the Raf-responsive cells. The Raf-responsive FD/A-Raf:ER+LNL6 (Est)c2, FD/B-Raf:ER+LNL6(Est)c1 or FD/Raf-1:ER+LNL6 (Est)c1 cells exited the cell cycle upon removal of tamoxifen. These cells entered the S phase (Figure 9A5, B5 or C5) and G₂/M phases (Figure 9A6, B6 or C6) of the cell cycle upon addition of either IL-3 or tamoxifen. Interestingly the FD/A-Raf:ER+LNL6(Est)c2 cells showed the most dramatic response to tamoxifen treatment in terms of entry into S phase (Figure 9A5) and exit from the G₁ phase (Figure 9A4) after tamoxifen treatment. The Raf-responsive FD/B-Raf:ER+LNL6(Est)c1 cells showed a more modest re-entry into the cell cycle after either IL-3 or tamoxifen treatment (Figure 9B5 and B6).

The cell cycle progression parameters were examined in the Raf-infected cells which also had BCL2. In the Raf-responsive FD/A-Raf:ER+BCL2(Est)c2 cells, the cells exited the S phase (Figure 9A8) and G₂/M (Figure 9A9) phase after -estradiol withdrawal for 24 h. The cells accumulated in the G₁ phase (Figure 9A7). The FD/A-Raf:ER+BCL2(Est)c2 re-entered the S and G₂/M phases approximately 24 h after addition of either tamoxifen or IL-3.

Some Raf-responsive FD/B-Raf:ER+BCL2 cells exited S phase 24 h after -estradiol deprivation (Figure 9B8), however, approximately 14% remained in S phase which was significantly higher than the percentage of cytokine-dependent or Raf-responsive FD/B-Raf:ER cells (Figure 9B2 and B5) which remained in S phase 24 h after either cytokine or -estradiol deprivation. After -estradiol withdrawal and subsequent treatment with either IL-3 or tamoxifen, the FD/B-Raf:ER+BCL2 cells showed less dramatic effects on cell cycle distribution than either the FD/A-Raf:ER+BCL2(Est)c2 or FD/Raf-1:ER+BCL2(Est)c1 cells.

Finally the cell cycle distribution was analyzed in the Raf-responsive FD/Raf-1:ER+BCL2(Est)c1 cells (Figure 9C7-C9). The percentages of cells in S phase decreased from approximately 25 to 8% after -estradiol withdrawal for 24 h. This was not as low a percentage as that observed in -estradiol-starved Raf-responsive FD/Raf-1:ER cells (Figure 9C5). After 36 to 72 h of -estradiol withdrawal, almost all the Raf-responsive FD/Raf-1:ER+BCL2 cells had exited the S and G₂/M phases of the cell cycle (Figure 9C8). The proportion of cells in G₁ increased upon cytokine withdrawal from 65 to 90%. Furthermore, these FD/Raf-1:ER+BCL2(Est) cells showed a gradual exit from S phase until after 24 h addition of IL-3 or tamoxifen. Addition of either tamoxifen or IL-3 to the FD/Raf:ER+BCL2 cells for 24 h resulted in the cells exiting G₁ and accumulating in the S and G₂/M phases of the cell cycle (Figure 9C7, C8 and C9).

Expression of the GFPRaf:ER protein in the Raf-responsive cells

We next wanted to determine what was happening to the level of the GFPRaf-1:ER protein when the cells were deprived of -estradiol and then subsequently treated with either -estradiol or IL-3. The expression of the GFPRaf:ER protein was compared in the Raf-responsive FD/LNL6+GFPRaf-1:ER(Est)c2 and FD/BCL2+GFPRaf-1(Est)c2 cells upon -estradiol deprivation and subsequent treatment with -estradiol and IL-3. Two different negative controls were used in these experiments, in Figure 10, top panel 1A, the autofluorescence of FD/LNL6 cells is shown, in Figure 10, bottom panel 1A, the autofluorescence of Raf-responsive FD/Raf-1:ER(Est)c2 cells is shown.

The fluorescence profiles of the FD/LNL6+GFPRaf-1:ER(Est)c2 and FD/BCL2+GFPRaf-1:ER(Est)c2 cells cultured in -estradiol continuously are shown in Figure 10, top and bottom panels 2E, 3E, 4E and 5E, respectively. When the FD/LNL6 +GFPRaf:ER cells were starved of -estradiol for 24 h, the mean fluorescence of the GFPRaf-1:ER protein decreased over four-fold (compare Figure 10, top panels 1B and 2E). In contrast, when the FD/BCL2+GFPRaf-1:ER(Est)c2 cells were deprived of -estradiol the change in GFPRaf-1:ER expression was less dramatic as less than a two-fold decrease in mean fluorescence was observed (Figure 10, bottom panels 1B and 2E).

The expression of the GFPRaf-1:ER protein was examined in the two different cell lines after treatment with IL-3 or -estradiol. IL-3 did not increase the level of GFPRaf-1:ER expression in either the FD/LNL6+GFPRaf-1:ER(Est)c2 (Figure 10, top panels 2B, 3B, 4B or 5B) or FD/BCL2+GFPRaf-1(Est)c2 (Figure 10, bottom panels 2B, 3B, 4B or 5B) cells, although the cells proliferated under these conditions. In contrast, addition of -estradiol increased the level of GFPRaf-1:ER expression in both types of cells after approximately 7 h of treatment (Figure 10, top and bottom panels 4C, 4D, 5C and 5D). Moreover, the cells which had BCL2 obtained a level of GFPRaf-1:ER expression similar to that observed in the cycling cells faster after -estradiol withdrawal and subsequent treatment with -estradiol (Figure 10, bottom panel 5D vs 2E) than the Raf-responsive cells which lacked BCL2 (Figure 10, top panel 5D vs 2E). Thus the recovery period for GFPRaf-1:ER expression was shorter in the cells which overexpressed BCL2.

Effects of Raf:ER on Raf and MEK activities

The GFP fluorescence studies indicated that the levels of the GFPRaf-1:ER protein were sensitive to -estradiol withdrawal but they did not determine what happened to the Raf enzymatic activity in these culture conditions. For the conditionally active Raf kinases to be effective, they should induce Raf as well as downstream MEK and ERK activities. To determine whether this was the case in the Raf:ER and BCL2 infected cells, Raf and downstream kinase assays were performed. The Raf assays were performed on Raf:ER immunprecipitated proteins recovered from the cells grown for 24 h under the different conditions (Figure 11).

When Raf or Raf+BCL2 infected cells were grown in the absence of -estradiol for 24 h (either no addition (-) or + IL-3) either no or very low levels of Raf kinase activity were detected. When FD/Raf:ER+BCL2(Est)c2 cells were cultured in IL-3, a very low level of Raf:ER kinase was detected. However when the cells were cultured with -estradiol (IL-3 and -estradiol or -estradiol alone) a significant increase in Raf activity was detected in all the Raf-responsive cell lines. For comparison purposes, the levels of Raf and MEK activities in Raf-responsive Raf:ER and Raf:ER+BCL2 infected cells are shown. The levels of Raf:ER proteins varied dramatically in the different culture conditions in the Raf-responsive cells which lacked exogenous BCL2 (Figure 11a, c, and e) whereas the Raf and BCL2 infected cells displayed similar levels of Raf:ER proteins under the different culture conditions (Figure 11b, d and f).

The levels of MEK1 activity were determined in the same samples. Higher levels of MEK1 activity were detected when the cells were treated with -estradiol (IL-3 + -estradiol or -estradiol alone) than when the cells were treated with IL-3. Thus the introduced Raf:ER proteins induced Raf and downstream MEK activity.

To determine the requirement of the MEK1 activity for the growth of the Raf and BCL2-infected cells, they were treated with the MEK inhibitor PD98059 and the effects on proliferation examined (Figure 12). The MEK1 inhibitor suppressed [³H]-thymidine incorporation in response to both -estradiol and IL-3 indicating that a functional MEK1 protein was necessary for growth. The MEK1 inhibitor suppressed [³H]-thymidine in response to -estradiol more than to IL-3 in FD/A-Raf:ER+BCL2(Est)c2 cells. This may indicate that IL-3 can induce additional growth related pathways which are MEK1 independent that also result in the stimulation of [³H]-thymidine incorporation. Similar results were observed with Raf-responsive cells which lacked exogenous BCL2.³⁷

The effects of MEK1 inhibitor PD98059 on MEK1 activity were determined in Raf-responsive FD/A-Raf:ER(Est)c2 and FD/A-Raf:ER+BCL2 cells (Figure 13). The MEK1 inhibitor did not reduce Raf activity in either FD/A-Raf:ER+LNL6(Est)c2 (Figure 13, lane 2) or FD/A-Raf:ER+BCL2(Est)c2 (Figure 13, lane 2). In contrast after treatment with the MEK1 inhibitor, the MEK activity was decreased in both cell lines.

Effects of IL-3 and -estradiol on ERK2 activation and activity in FD/Raf:ER+BCL2(Est) cells

To determine whether the activated Raf oncoproteins induced down-stream ERK2 activation, cells were deprived of -estradiol for 24 h and then treated with either IL-3 or -estradiol for ½ h (Figure 14a). Treatment of the cells with IL-3, -estradiol or the control phorbol esters resulted in activation of ERK1 and ERK2. To determine the requirement of functional MEK1 proteins in the activation of ERK1 and ERK2 in these cells, the cells were treated with the MEK1 inhibitor PD98059. Treatment with PD98059 eliminated Raf-induced ERK1 and ERK2 activation.

It should be noted that treatment with the MEK1 inhibitor did not suppress all of the PMA induced ERK1,2 activation. The PD98059 inhibition is not 100% complete, as the PMA induces such a strong signal that it is not completely blocked by the MEK1 inhibitor. In addition, PMA may phosphorylate ERK1,2 through a MEK1-independent mechanism. As a control the upper portion of the blots were probed with an p90^Rsk Ab. Relatively equal levels of p90^Rsk proteins were detected in the samples. Hyperphosphorylation of p90^Rsk was readily observed after PMA and -estradiol treatment indicating that Raf and PKC induced p90^Rsk phosphorylation. Treatment of the cells with the MEK1 inhibitor eliminated p90^Rsk hyperphosphorylation after -estradiol but not PMA treatment providing further evidence that PMA can induce a pathway which is independent of MEK1 activity.

To determine the effects of Raf on ERK2 kinase activity, an ERK kinase assay was performed with immunoprecipitated ERK2 and myelin basic protein as a substrate. Raf induced ERK2 kinase activity in the FD/Raf-1:ER+BCL2 infected cells. Interestingly Raf induced ERK2 activity earlier than after treatment with IL-3. Thus Raf activated downstream MEK, ERK2 and p90^Rsk and disruption of this pathway with the MEK1 inhibitor resulted in inhibition of cell growth.

Expression of the BCL2 oncoprotein in the FD/Raf:ER + BCL2 infected cells

It was important to confirm that the cells which were infected with the BCL2 retrovirus actually expressed the BCL2 oncoprotein. In the pool of cytokine-dependent FD/BCL2+ GFPRaf-1:ER cells, a low level of BCL2 expression was detected after immunoprecipitation and Western blot analysis with the BCL2 antibody (Figure 15a). These cytokine-dependent cells had not been selected for factor-independence. In contrast in the Raf-responsive FD/BCL2+Raf-1:ER cells there was a higher level of BCL2 expression suggesting that increased BCL2 expression was associated with abrogation of cytokine-dependency in these cells (Figure 15b). BCL2 expression was also examined in the different Raf-responsive FD/Raf:ER cells. Some Raf-responsive cells contained BCL2, whereas the FD/Raf-1:ER+LNL6(Est)c2 cell line did not. The level of BCL2 expression was lower in the FD/Raf-1:ER(Est)c1 cells (Figure 15c) than in the Raf-responsive cells which had BCL2 namely FD/A-Raf:ER+BCL2(Est)c2 (Figure 15d), FD/B-Raf:ER+BCL2(Est)c1 (Figure 15e) or FD/Raf-1:ER+BCL2(Est)c2 (Figure 15f). Thus the cells which were Raf-responsive and infected with BCL2 expressed more BCL2 protein than the cells that were either Raf-responsive and not infected with the BCL2 oncoprotein or the cytokine-dependent FD/BCL2+GFPRaf-1:ER pool.

Autocrine growth factor synthesis by BCL2 and Raf:ER infected cells

To determine whether there was a possible autocrine component to the growth induced by Raf:ER, the presence of autocrine growth factor synthesis was examined by RT-PCR (Figure 16). cDNAs encoding GM-CSF and IL-3 were detected in the control T cell thymoma EL4 and increased levels were detected when these cells were stimulated with the phorbol ester PMA for 24 h. cDNAs encoding IL-3 and GM-CSF were not detected in the parental FDC-P1 line or the cytokine-dependent Raf:ER-infected cells. cDNAs encoding GM-CSF but not IL-3 were detected in the Raf-responsive cells implicating a possible autocrine mechanism of transformation. These GM-CSF transcripts were detected in all Raf-responsive cells, whether or not the cells were infected with BCL2.

To determine whether these GM-CSF transcripts resulted in the expression of GM-CSF protein, the abilities of the supernatants prepared from the Raf-responsive Raf:ER+BCL2 infected cells to stimulate [³H]-thymidine incorporation in the parental cells were determined. -Estradiol by itself did not support [³H]-thymidine incorporation in FDC-P1 cells (Figure 17a). In contrast, supernatants prepared from FD/A-Raf:ER+BCL2(Est)c2 (Figure 17b), FD/B-Raf:ER+BCL2(Est)c1 (Figure 17c) and FD/Raf-1:ER+BCL2(Est)c2 (Figure 17d) stimulated [³H]-thymidine incorporation in FDC-P1 cells. Antibody neutralization experiments indicated that the cytokine responsible for stimulation of [³H]-thymidine incorporation was GM-CSF (data not shown). Thus the Raf-responsive cells synthesized an autocrine growth factor.

Discussion

After infection of hematopoietic cells with Raf:ER containing retroviruses, two different types of cells were isolated, cytokine-dependent or Raf-responsive, both types of cells expressed the Raf:ER oncoprotein. The cytokine-dependent Raf:ER cells died after IL-3 withdrawal for 24 to 48 h. -Estradiol would not replace this requirement of IL-3 for growth in the cytokine-dependent cells even though they expressed the Raf:ER encoded kinases. These results implied that another genetic event was required for the Raf-responsive growth of FDC-P1 cells. The growth of the Raf-responsive Raf:ER infected cells was conditional as removal of -estradiol eliminated growth and induced apoptosis.³⁷ This indicated that the activated Raf:ER oncoprotein was necessary for Raf-responsive growth. Moreover, the Raf-responsive cells exited the cell cycle after -estradiol withdrawal similar to the cytokine-dependent Raf:ER-infected cells after IL-3 withdrawal. However, Raf activation was able to promote entry into the S and G₂/M phases in the Raf-responsive but not in the cytokine-dependent cells.

We are interested in determining which genes may be capable of synergizing with Raf and result in the abrogation of cytokine dependency. One relatively simple and fast approach is to examine the ability of known oncogenes to increase the frequency of abrogation of cytokine dependency of Raf:ER infected cells. The BCL2 oncogene was an obvious choice since it has been shown to interact with Raf and is a target for downstream ERK2. Thus we determined the ability of BCL2 oncogene expression to synergize with Raf:ER expression and increase the frequency of Raf-responsive cells. While BCL2 expression by itself did not abrogate cytokine independence, BCL2 overexpression in conjunction with Raf:ER expression increased the frequency of cytokine-independent cells. Functional Raf expression was necessary as cytokine-independent cells were not recovered from FDC-P1 cells infected with BCL2 alone or FD/BCL2 subsequently infected with the kinase-inactive Raf-1[301]:ER mutant. Moreover higher levels of BCL2 expression were detected in the Raf and BCL2 infected cells which grew in response to Raf. Overexpression of BCL2 is probably not the reason why all the Raf-responsive cells grow in response to Raf as the FD/Raf-1:ER+LNL6(Est)c2 cells did not overexpress BCL2. However, that does not eliminate the importance of these results as we have observed that some MEK1-responsive cells do overexpress BCL2 (see accompanying article, this issue, pp 1080-1096). There may be other anti-apoptotic molecules (eg Bcl-X_L, A1) which are over-expressed in the Raf-responsive FD/Raf:ER+LNL6 cells. Alternatively certain pro-apoptotic molecules may be expressed at lower levels thus increasing the anti-apoptotic effects of BCL2.

A significant difference between the Raf-responsive FD/Raf-1:ER+LNL6(Est)c2 and FD/Raf-1:ER+BCL2(Est)c2 was that the Raf-responsive cells which contained exogenous BCL2 displayed a delayed exit from the S phase (Figure 9C8) and G₂/M (Figure 9C9) phases of the cell cycle when they were deprived of -estradiol. Thus BCL2 was impeding the exit of FD/Raf-1:ER+BCL2 cells from the cell cycle after -estradiol withdrawal. The combination of BCL2 overexpression and autocrine GM-CSF expression may keep a higher proportion of these cells in the S and G₂/M phases than those cells which lacked BCL2 overexpression. Another observation from these experiments was that activation of Raf in the Raf-responsive FD/A-Raf:ER(Est)c2 and FD/Raf-1:ER(Est)c2 cells often resulted in a more robust entry into S and exit from G₁ than after stimulation with IL-3 (Figure 9 panels A4, A5, A7, A8, C7 and C8). This may have resulted from the cells being selected to grow in response to Raf which can cause a down regulation of the IL-3 receptors (data not presented) (Figure 8). Moreover the change in the level of GFP/Raf-1:ER protein upon -estradiol withdrawal was decreased in the cells which had BCL2 in comparison to the cells which did not. Therefore BCL2 was exerting a protective effect on the cells.

Raf induced autocrine growth factor expression. The GM-CSF expression was due to Raf and not BCL2 expression as autocrine GM-CSF transcripts were observed in Raf-responsive cells which lacked BCL2 expression. Raf may be inducing the phosphorylation of downstream transcription factors which bind the GM-CSF promoter region and stimulate GM-CSF expression. A model for this stimulation of autocrine GM-CSF expression by the activation of the Raf/MEK/ERK pathway was presented in Figure 1.

These cells provide an interesting model to elucidate the effects of the different Raf kinases hematopoietic cell growth, prevention of apoptosis and malignant transformation. Identification of pathways which cross-talk with the Ras/Raf/MEK/MAPK pathway to promote cellular transformation and prevent apoptosis will aid in our understanding of the mechanisms responsible for tumor progression.

Acknowledgements

We appreciate the artwork done by Ms Catherine Spruill. This work was supported in part by a grant (R01CA51025) from the NCI and the North Carolina Biotechnology Center (9805-ARG-0006) to JAM. RAF was supported in part by grants from the American Cancer Society (IRG-97-149), American Heart Association (9930099N) and the North Carolina Biotechnology Center (9705-ARG-0009).

1 Blalock WL, Weinstein-Oppenheimer C, Chang F, Hoyle PE, Wang XY, Algate PA, Franklin RA, Oberhaus SM, Steelman LS, McCubrey JA. Signal transduction, cell cycle regulatory, and anti-apoptotic pathways regulated by IL-3 in hematopoietic cells: possible sites for intervention with anti-neoplastic drugs. Leukemia 1999; 13: 1109-1166, MEDLINE

2 Wang XY, Steelman LS, McCubrey JA. Abnormal activation of cytokine gene expression by intracisternal type A particle transposition: effects of mutations which result in autocrine growth factor stimulation and malignant transformation. Cytokin Cell Mol Ther 1997; 3: 3-19,

3 Lotem J, Sachs L. Control of apoptosis in hematopoiesis and leukemia by cytokines, tumor suppressor and oncogenes. Leukemia 1996; 10: 925-931, MEDLINE

4 Wang XY, McCubrey J. Regulation of interleukin 3 expression in normal and autocrine transformed hematopoietic cells. Int J Oncol 1997; 10: 989-1001,

5 Algate PA, Steelman LS, Mayo MW, Miyajima A, McCubrey JA. Regulation of the interleukin-3 (IL-3) receptor by IL-3 in the fetal liver-derived FL5.12 cell line. Blood 1994; 83: 2450-2468,

6 Steelman LS, Algate PA, Blalock WL, Wang XY, Prevost KD, Hoyle PE, McCubrey JA. Oncogenic effects of overexpression of the interleukin-3 receptor on hematopoietic cells. Leukemia 1996; 10: 528-542, MEDLINE

7 Miyajima A, Kitamura T, Harada N, Yokota T, Arai K-L. Cytokine receptors and signal transduction. Annu Rev Immunol 1992; 10: 295-331, MEDLINE

8 Kinoshita T, Yokata T, Arai K, Miyajima A. Suppression of apoptotic death in hematopoietic cells by signalling through the IL-3/GM-CSF receptors. EMBO J 1995; 14: 265-275,

9 Dexter TM, Garland J, Scott D, Scolnick E, Metcalf D. Growth factor-dependent hemopoietic precursor cell lines. J Exp Med 1980; 152: 1036-1047, MEDLINE

10 McKearn JP, McCubrey JA, Fagg B. Enrichment of hematopoietic precursor cells and cloning of multipotential B lymphocyte precursors. Proc Natl Acad Sci USA 1985; 82: 7414-7418, MEDLINE

11 McCubrey JA, Holland G, McKearn J, Risser R. Abrogation of factor-dependence in two IL-3-dependent cell lines can occur by two distinct mechanisms. Oncogene Res 1989; 4: 97-109, MEDLINE

12 McCubrey JA, Smith SR, Algate PA, deVente JE, White MK, Steelman LS. Retroviral infection can abrogate the factor-dependency of hematopoietic cells by autocrine and non-autocrine mechanisms depending on the presence of a functional viral oncogene. Oncogene 1993; 8: 2905-2915, MEDLINE

13 McCubrey JA, Steelman LS, Wang XY, Algate PA, Hoyle PE, White C, Davidian EW, Prevost KD, Robbins P, Mylott D, White MK. Differential effects of viral and cellular oncogenes on the growth factor-dependency of hematopoietic cells. Int J Oncol 1995; 7: 285-310,

14 Kitamura T, Sato M, Arai K-I, Miyajima A. Expression cloning of the human IL-3 receptor cDNA reveals a shared subunit of the human IL-3 and GM-CSF receptors. Cell 1991; 66: 1165-1174, MEDLINE

15 McCubrey JA, May WS, Duronio V, Mufson A. Serine/threonine phosphorylation in cytokine signal transduction. Leukemia 2000; 14: 9-21, MEDLINE

16 McCubrey JA, Steelman LS, Hoyle PA, Blalock WL, Weinstein-Oppenheimer CR, Franklin RA, Cherwinski H, Bosch E, McMahon M. Differential abilities of activated Raf oncoproteins to abrogate cytokine dependency, prevent apoptosis and induce autocrine growth factor synthesis in human hematopoietic cells. Leukemia 1998; 12: 1903-1929, MEDLINE

17 Sakatsume M, Stancato LF, David M, Silvennoinen O, Saharinen P, Pierce J, Larner AC, Finbloom DS. Interferon gamma activation of Raf-1 is Jak1-dependent and p21ras-independent. J Biol Chem 1998; 273: 3021-3026, Article MEDLINE

18 Stancato LF, Sakatsume M, David M, Dent P, Dong F, Petricoin EF, Krolewski JJ, Silvennoinen O, Saharinen P, Pierce J, Marshall CJ, Sturgill T, Finbloom DS, Larner AC. Beta interferon and oncostatin M activate Raf-1 and mitogen-activated protein kinase through a JAK1-dependent pathway. Mol Cell Biol, 1997; 17: 3833-3840, MEDLINE

19 Morrison D. Mechanisms regulating Raf-1 activity in signal transduction pathways. Mol Repro Dev 1995; 42: 507-514,

20 Dent P, Wu J, Romero G, Vincent LA, Castle D, Sturgill TW. Activation of the mitogen-activated protein kinase pathway in Triton X-100 disrupted NIH-3T3 cells by p21 ras and in vitro by plasma membranes from NIH3T3 cells. Mol Biol Cell 1993; 4: 83-493,

21 Dent P, Chow YH, Wu J, Morrison DK, Jove R, Sturgill TW. Expression, purification and characterization of recombinant mitogen-activated protein kinase kinases. Biochem J, 1994; 303: 105-112,

22 Dent P, Reardon DB, Morrison DK, Sturgill TW. Regulation of Raf-1 and Raf-1 mutants by Ras-dependent and Ras-independent mechanisms in vitro. Mol Cell Biol 1995; 15: 4125-4135, MEDLINE

23 Dent P, Jarvis WD, Birrer MJ, Fisher PB, Schmidt-Ullrich RK, Grant S. The roles of signaling by the p42/p44 mitogen-activated protein (MAP) kinase pathway; a potential route to radio- and chemo-sensitization of tumor cells resulting in the induction of apoptosis and loss of clonogenicity. Leukemia 1998; 12: 1843-1850, MEDLINE

24 Jelinek T, Dent P, Sturgill TW, Weber MJ. Ras-induced activation of Raf-1 is dependent on tyrosine phosphorylation. Mol Cell Biol 1996; 16: 1027-1034, MEDLINE

25 Sakamoto KM, Mignacca RC, Gasson JC. Signal transduction by granulocyte-macrophage colony-stimulating factor and interleukin-3 receptors. Recep Chann 1994; 2: 175-181,

26 Kwon EM, Sakamoto KM. The molecular mechanism of action of granulocyte-macrophage/colony-stimulating factor. J Invest Med 1996; 44: 442-446,

27 Lee HJ, Mignacca RC, Sakamoto KM. Transcriptional activation of egr-1 by granulocyte-macrophage colony-stimulating factor but not interleukin 3 requires phosphorylation of cAMP response element-binding protein (CREB) on serine 133. J Biol Chem 1995; 270: 15979-15983, MEDLINE

28 Wong A, Sakamoto KM. Granulocyte-macrophage colony-stimulating factor induces the transcriptional activation of egr-1 through a protein kinase A-independent signaling pathway. J Biol Chem 1995; 270: 30271-30273, MEDLINE

29 Watanabe S, Kubota H, Sakamoto KM, Arai K. Characterization of cis-acting sequences and trans-acting signals regulating early growth response 1 and c-fos promoters through the granulocyte-macrophage colony-stimulating factor receptor in BA/F3 cells. Blood 1997; 89: 1197-1206, MEDLINE

30 Mignacca RC, Lee HJ, Kwon EM, Sakamoto KM. Mechanism of transcriptional activation of the immediate early gene Egr-1 in response to PIXY321. Blood 1996; 88: 848-854, MEDLINE

31 Sakamoto KM, Fraser JK, Lee HJ, Lehman E, Gasson JC. Granulocyte-macrophage colony-stimultating factor and interleukin-3 signaling pathways converge on the CREB-binding site in the human egr-1 promoter. Mol Cell Biol 1994; 14: 5975-5985, MEDLINE

32 Samuels ML, Weber MJ, Bishop JM, McMahon M. Conditional transformation of cells and rapid activation of the mitogen-activated protein kinase cascade by an estradiol-dependent human raf-1 protein kinase. Mol Cell Biol 1993; 13: 6241-6252, MEDLINE

33 Pritchard CA, Samuels ML, Bosch E, McMahon M. Conditionally oncogenic forms of the A-raf and B-raf protein kinases display different biological and biochemical properties in NIH 3T3 cells. Mol Cell Biol 1995; 15: 9430-9442,

34 Bosch E, Cherwinski H, Peterson D, McMahon M. Mutations of critical amino acids affect the biological and biochemical properties of oncogenic A-Raf and Raf-1. Oncogene 1997; 11: 1021-1034,

35 Blalock WL, Steelman LS, Pearce M, Franklin RA, McCarthy SA, Cherwinski H, McMahon M, McCubrey JA. A conditionally active form of MEK1 results in autocrine transformation in human and mouse hematopoietic cells. Oncogene 2000; 19: 526-536, MEDLINE

36 McCubrey JA, Steelman LS, Moye PW, Hoyle PE, Weinstein-Oppenheimer C, Chang F, Pierce M, White MK, Franklin R, Blalock WL. Effects of deregulated Raf and MEK1 expression on the cytokine dependency of hematopoietic cells. Adv Enzyme Regl 2000; 40: (in press).

37 Hoyle PE, Moye PW, Steelman LS, Blalock WL, Franklin RA, Pearce M, Cherwinski H, Bosch H, McMahon M, McCubrey JA. Differential abilities of the Raf family of protein of kinases to abrogate cytokine dependency and prevent apoptosis in murine hematopoietic cells by a MEK1-dependent mechanism. Leukemia 2000; 14: 642-656, MEDLINE

38 Yang E, Korsmeyer SJ. Molecular thanatopsis: a discourse on the BCL2 family and cell death. Blood 1996; 88: 386-401, MEDLINE

39 Kroemer G. The proto-oncogene Bcl-2 and its role in regulating apoptosis. Nature Med 1997; 3: 614-620, MEDLINE

40 Yang E, Jiping Z, Jockel J, Boise LH, Thompson CB, Korsmeyer SJ. Bad, a heterodimeric partner for Bcl-X_L and Bcl-2, displaces Bax and promotes cell death. Cell 1995; 80: 285-291, MEDLINE

41 Young LS, Eliopoulos AG, Kerr DJ, Herod J, Hodgkins L, Krajewski S, Reed JC. The control of apoptosis and drug resistance in ovarian cancer: influence of p53 and Bcl-2. Oncogene 1995; 11: 1217-1228, MEDLINE

42 Núnez G, London L, Hockenbery D, Alexander M, McKearn JP, Korsmeyer SJ. Deregulated Bcl-2 gene expression selectively prolongs survival of growth factor-deprived hemopoietic cell lines. J Immnol 1990; 144: 3602-3610,

43 Sato N, Sakamaki K, Terada N, Arai K-I, Miyajima A. Signal transduction by the high-affinity GM-CSF receptor: two distinct cytoplasmic regions of the common subunit responsible for different signaling. EMBO J 1993; 12: 4181-4189, MEDLINE

44 Ruvolo PP, Deng X, Ito T, Carr BK, May WS. Ceramide induces Bcl2 dephosphorylation via a mechanism involving mitochondrial PP2A. J Biol Chem 1999; 274: 20296-20300, Article MEDLINE

45 Ruvolo PP, Deng X, Carr BK, May WS. A functional role for mitochondrial protein kinase C alpha in Bcl2 phosphorylation and suppression of apoptosis. J Biol Chem 1998; 273: 25436-25442, Article MEDLINE

46 Wang HG, Rapp UR, Reed JC. Bcl-2 targets the protein kinase Raf-1 to mitochondria. Cell 1996; 87: 629-638, MEDLINE

47 Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ. Serine phosphorylation of death agonist Bad in response to survival factor results in binding to 14-3-3 not Bcl-xL. Cell 1996; 87: 619-628, MEDLINE

48 del Peso L, Gonzalez-Garcia M, Page C, Herrera R, Nunez G. Interleukin-3-induced phosphorylation of BAD through the protein kinase Akt. Science 1997; 278: 687-689, MEDLINE

49 Ahmed NN, Grimes HL, Bellacosa A, Chan TO, Tsichlis PN. Transduction of interleukin-2 anti-apoptotic and proliferative signals via Akt protein kinase. Proc Natl Acad Sci USA 1997; 94: 3627-3632, Article MEDLINE

50 Srivastava RK, Srivastava AR, Korsmeyer SJ, Nesterova M, Cho-Chung YS, Longo DL. Involvement of microtubules in the regulation of Bcl2 phosphorylation and apoptosis through cyclic AMP-dependent protein kinase. Mol Cell Biol 1998; 18: 3509-3517, MEDLINE

51 Scheid MP, Duronio V. Dissociation of cytokine-induced phosphorylation of Bad and activation of PKB/akt: involvement of MEK upstream of Bad phosphorylation. Proc Natl Acad Sci USA 1998; 95: 7439-7444, MEDLINE

52 Haldar S, Basu A, Croce CM. Serine-70 is one of the critical sites for drug-induced Bcl2 phosphorylation in cancer cells. Cancer Res 1998; 58: 1609-1615, MEDLINE

53 Deng X, Ito T, Carr B, Mumby M, May WS. Reversible phosphorylation of Bcl-2 following IL-3 or byrostatin is mediated by interaction of PP2A. J Biol Chem 1998; 271: 34157-34163,

54 Blagosklonny MV, Schulte T, Nguyen P, Trepel J, Neckers LM. Taxol-induced apoptosis and phosphorylation of the Bcl-2 protein involves c-Raf-1 and represents a novel c-Raf-1 signal transduction pathway. Cancer Res 1996; 56: 1851-1854, MEDLINE

55 Blagosklonny MV, Chuman Y, Bergan RC, Fojo T. Mitogen-activated protein kinase pathway is dispensable for microtubule-active drug-induced Raf-1/Bcl-2 phosphorylation and apoptosis in leukemia cells. Leukemia 1999; 13: 1028-1036, MEDLINE

56 Hu ZB, Minden MD, McCulloch EA. Phosphorylation of Bcl-2 after exposure of human leukemic cells to retinoic acid. Blood 1998; 92: 1768-1775, MEDLINE

57 Rapp UR, Troppmair J, Carroll M, May WS. Role of raf-1 protein kinase in IL-3 and GM-CSF mediated signal transduction. Curr Top Microbiol Immunol 1990; 166: 129-139, MEDLINE

58 Cleveland JL, Troppmair J, Packham G, Askew DS, Lloyd P, Gonzalez-Garcia M, Nuñez G, Ihle JN, Rapp UR. v-raf suppresses apoptosis and promotes growth of interleukin-3-dependent myeloid cells. Oncogene 1994; 9: 2227-2233, MEDLINE

59 Wang HG, Miyashita T, Takayama S, Sato T, Torigoe T, Krajewski S, Tanaka S, Hovey L III, Troppmair J, Rapp UR, Reed JC. Apoptosis regulation by interaction of Bcl-2 and Raf-1 kinase. Oncogene 1994; 9: 2751-2756, MEDLINE

60 Wang HG, Takayama S, Rapp UR, Reed JC. Bcl-2 interacting protein, Bag-1 binds to and activates the kinase Raf-1. Proc Natl Acad Sci USA 1996; 93: 7063-7068, Article MEDLINE

61 Wang HG, Rapp UR, Reed JC. Bcl-2 targets the protein kinase Raf-1 to mitochondria. Cell 1996; 87: 629-638, MEDLINE

62 Srivastava RK, Mi QS, Hardwick JM, Longo DL. Deletion of the loop region of Bcl-2 completely blocks paclitaxel-induced apoptosis. Proc Nat Acad Sci USA 1999; 96: 3775-3780, MEDLINE

63 Dudley DT, Pung L, Decker SJ, Bridges AJ, Saltiel AR. A synthetic inhibitor of the mitogen-activated protein kinases cascade. Proc Natl Acad Sci USA 1995; 92: 7686-7689, MEDLINE

64 Pumiglia KM, Decker SJ. Cell cycle arrest mediated by the MEK/mitogen-activated protein kinase pathway. Proc Natl Acad Sci USA 1997; 94: 448-452, Article MEDLINE

65 Woods D, Parry D, Chewinski H, Bosch E, Lees E, McMahon M. Raf-induced proliferation or cell cycle arrest is determined by the level of Raf activity with arrest mediated by p21^Cip1. Mol Cell Biol 1997; 17: 5598-5611, MEDLINE

66 Miller AD, Rosman CJ. Improved retroviral vectors for gene transfer and expression. Biotechniques 1989; 7: 980-990, MEDLINE

67 Lloyd AC, Obermuller F, Staddon S, Barth CF, McMahon M, Land H. Cooperating oncogenes converge to regulate cyclin/cdx complexes. Genes Dev 1997; 11: 663-667, MEDLINE

Figure 1  Signal transduction and anti-apoptotic pathways induced by cytokines. The GM-CSF cytokine is indicated in red. Inactive signal transduction molecules are indicated in maroon, activated signal transduction molecules are indicated in green. The sites where the activated Raf:ER and MEK1:ER oncogenes affect signal transduction are indicated in red. The inactive GM-CSF receptor is indicated in maroon, the activated cytokine receptor is indicated in yellow. Pro-apoptotic molecules are indicated in brown, anti-apoptotic molecules are indicated in dark blue. Tyrosine phosphorylated proteins are indicated with a black P in a white circle. Serine/threonine phosphorylated proteins are indicated by a red P with a white circle around it. The 14-3-3 family of chaperonin proteins is indicated in black. The sites of interaction of the MEK1 inhibitor is indicated by a black oval and a minus sign. The sites where the Cis family of proteins (black oval) can inhibit cytokine-induced signal transduction are indicated by a minus sign. The sites of activation of the Src family of kinases are indicated in a yellow box and arrows. The interactions of kinases with anti-apoptotic molecules are indicated. The interactions, which result in the induction of apoptosis, are indicated with skull and cross bones. The interactions that result in the prevention of apoptosis are indicated with skull and cross bones with a red X through them. A similar scenario would occur after ligation of the IL-3 receptor with IL-3. The induction of autocrine GM-CSF transcription by the activated Raf and MEK oncoprotein is indicated.

Figure 2  Raf expression in cytokine-dependent FD/Raf:ER clones. A Raf kinase assay was performed to determine the level of Raf kinase activity in cytokine-dependent: FD/A-Raf:ER(IL3)c2, FD/B-Raf:ER(IL3)c1 and FD/Raf-1:ER(IL3)c2 cells. The cells were cultured in the absence of IL-3 (lane 1), in the presence of IL-3 (lane 2) or in the presence of IL-3 and -estradiol (lane 3) and cell extracts were prepared. Raf:ER proteins were immunoprecipitated with an ER Ab and then a Raf kinase assay was performed on the proteins. The levels of the Raf:ER proteins were subsequently determined by probing the filters containing the kinase reactions with an ER Ab. Uninfected FDC-P1 cells or FDC-P1 cells infected with empty retroviral vectors do not contain Raf:ER proteins or Raf:ER-induced Raf kinase activity (data not shown).

Figure 3  GFPRAF-1:ER expression in cytokine-dependent and Raf-responsive FD/BCL2+GFPRAF-1:ER cells. The expression of the GFPRaf-1:ER protein in cytokine-dependent and Raf-responsive cells was examined by FACS analysis. Cells were cultured in medium containing IL-3 (a, c, e, g and i) or medium containing IL-3 and -estradiol (b, d, f, h and j) for 24 h and then GFP expression was determined by FACS.

Figure 4  Isolation of Raf-responsive cells after infection of cytokine-dependent Raf:ER-infected FDC-P1 clones with a BCL2 containing retrovirus. Limiting dilution analysis of LNL6 and BCL2 infected cells: FDC-P1 (a and b), FD/A-Raf:ERc2 (c and d), FD/B-Raf:ERc1 (e and f), and FD/Raf-1:ERc1 (g and h). Growth in medium containing: IL-3 (), -estradiol () or no supplement (). Dotted line equals 37% of wells negative for growth from which the differential plating efficiency was estimated by Poisson distribution. Solid arrow on the X-axis indicates where the experimental data line intersects with the 37% negative for growth line and is an estimate of the cloning efficiency. These experiments were performed on three to five pools isolated from each type of Raf:ER-infected cells.

Figure 5  Growth of Raf:ER and BCL2 infected FDC-P1 cells. The growth properties of three different types of Raf:ER-infected cells are compared. Cytokine-dependent FD/A-Raf:ER c2 (a), FD/B-Raf:ER(IL3)c1 (d) and FD/Raf-1:ER(IL3)c2 (g) cells, Raf-responsive FD/A-Raf:ER+LNL6c2 (b), FD/B-Raf:ER+LNL6(Est)c1 (e), FD/Raf-1:ER+LNL6(Est)c2 (h) cells. Raf-responsive BCL2 infected FD/A-Raf:ER+BCL2(Est)c2 (c), FD/B-Raf:ER+BCL2(Est)c1 (f) and FD/Raf-1:ER+BCL2(Est)c2 (i) cells. Growth in medium containing: IL-3 (), -estradiol (), and no supplement (). The Raf-responsive cells also grew in response to tamoxifen which activates the chimeric Raf:ER construct. Similar results were observed with the respective BCL2+GFPRaf-1:ER infected cells. These results were repeated six times with at least two different clones from each type of Raf-infected cells and similar results were observed.

Figure 6  Stimulation of [³H]-thymidine incorporation by -estradiol and IL-3 in the Raf-responsive Raf:ER-infected cells. As controls, [³H]-thymidine incorporation was examined in cytokine-dependent FD/LNL6 (a and b) and FD/BCL2 (c and d) cells. The stimulation of the Raf-responsive cells by IL-3 is shown in (e, g, i, k, m and o). The stimulation of Raf responsive growth by -estradiol is shown in (f, h, j, l, n and p). Symbols: (), % IL-3 (% WEHI-3B conditioned supernatant); (), nM -estradiol.

Figure 7  Stimulation of [³H]-thymidine incorporation by 4-OH tamoxifen in the Raf-responsive RAF:ER-infected cells. , 4-hydroxy tamoxifen. (a) FD/A-Raf:ER+BCL2(Est)c2; (b) FD/B-Raf:ER+BCL2(Est)c1; and (c) FD/Raf-1:ER+BCL2(Est)c2.

Figure 8  Modfit histogram analysis of cell cycle distribution. Cell cycle progression in Raf-responsive FD/Raf-1:ER+LNL6(Est)c2 cells (A); cell cycle progression in Raf-responsive FD/Raf-1:ER+BCL2(Est)c2 cells (B). In these histograms, the Y axis (the cell number) was cut off at 500 in order to display the proportion of cells in S and G₂/M better. The actual percentages of the cells in the different phases were determined from the original data graphs and are presented in Figure 9. The (a) panels in A and B represent the start of the experiments (-24 h) when the cycling cells were collected washed with PBS and set up for the time course experiments. The (b) panels in A and B represent -12 h before the addition of cytokines or tamoxifen. The (c) panels represent T₀ = time of addition of cytokine or tamoxifen. The (d) panels represent cells cultured without IL-3 and tamoxifen for 48 h, the (e) panels represent cells cultured with IL-3 for 24 h, the (f) panels represent cells cultured with tamoxifen for 24 h, the (g) panels represent cells cultured without IL-3 and tamoxifen, the (h) panels represent cells cultured with IL-3 for 48 h, and the (i) panels represent cells cultured with tamoxifen for 48 h. The cells in panels d, e and f were all isolated at the same time. The cells in g, h, and i were all isolated at the same time. The percentage of cells in G₁, S and G₂/M is indicated underneath the growth conditions in each panel.

Figure 9  Effects of Raf and BCL2 on cell cycle progression in FDC-P1 cells. The percentages of cells in the G1, S and G2/M phases of the cell cycle of the indicated cell lines were statistically determined by the Modfit computer program after deprivation of cytokine or b-estradiol and then treatment with either IL-3 (l), tamoxifen (g)orno addition (p). (Top A panels) the three types of FD/DA-Raf:ER cells (panels A1-A9); (middle B panels) the three types of FD/DB-Raf:ER cells (panels B1-B9); and (bottom C panels) the three types of FD/DRaf:ER cells (panels C1-C9). -24 indicates the start of the experiments, 0 indicates when either IL-3 or tamoxifen were added. Panels A1-A3, cytokine-dependent FD/DA-Raf:ER(IL3)c2; panels A4-A6, Raf-responsive FD/DA-Raf:ER+LNL6(Est)c2; panels A7-A9, Raf-responsive and BCL2-infected FD/DA-Raf:ER+BCL2(Est)c2; panels B1-B3, cytokine-dependent FD/DB-Raf:ER+LNL6(IL3)c1; panels (B4-B6) FD/DB-Raf:ER+LNL6(Est)c1; panels B7-B9, Raf-responsive and BCL2 infected FD/DB-Raf:ER+BCL2(Est)c1; panels C1-C3, cytokine-dependent FD/DRaf-1:ER+LNL6(IL3)c1; panels C4-C6, Raf-responsive FD/DRaf-1:ER+LNL6(Est)c1; and panels C7-C9, Raf-responsive and BCL2-infected FD/DRaf-1:ER+BCL2(Est)c1 cells. These experiments were repeated three to six times and then averaged together. In those cases where the error bars are not visible, the standard deviations are contained within the size of the symbols.

Figure 10  GFPRaf-1:ER expression in Raf-responsive BCL2 and GFPRaf-1:ER infected FDC-P1 cells. The expression of the GFPRaf-1:ER protein after -estradiol deprivation (1B) and subsequent treatment with nothing (0), IL-3, IL-3 + -estradiol, -estradiol, or -estradiol continuously. GFP expression in FD/LNL6+GFPRaf-1:ER(Est)c2 (top panel), and FD/BCL2+GFPRaf-1:ER(Est)c2 (bottom panel). The negative control presented in top panel 1A is FD/LNL6 cells. The negative control presented in the bottom panel 1A is FD/Raf-1:ER(Est)c2.

Figure 11  Raf:ER induces RAF and MEK1 activities in Raf-responsive Raf:ER and BCL2 infected cells. (a) FD/A-Raf:ER+LNL6(Est)c2; (b) FD/A-Raf:ER+BCL2(Est)c2; (c) FD/B-Raf:ER+LNL6(Est)c1; (d) FD/B-Raf:ER+BCL2(Est)c1; (e) FD/Raf-1:ER+LNL6(Est)c2; and (f) FD/Raf-1:ER+BCL2(Est)c2. The cells were grown in phenol red free medium for 24 h in the indicated conditions. The Raf proteins were immunoprecipitated with an ER Ab, the MEK1 proteins were immunoprecipitated with an MEK1 Ab. The Raf and MEK assays were performed on immunoprecipitated proteins from the same samples, respectively.

Figure 12  Effects of a MEK1 inhibitor on [³H]-thymidine incorporation. The fold inhibition was determined by comparing the [³H]-thymidine incorporation observed when the cells were treated with the same concentration of DMSO (0.1%) which was used to dissolve PD98059. The concentration of PD98059 was 25 M. Three individual experiments for each cell line were averaged together.

Figure 13  Effects on the MEK1 inhibitor on Raf and MEK activities in Raf-responsive A-Raf:ER and BCL2 infected cells. Cell lysates were prepared from cells cultured in the presence of -estradiol and the MEK1 inhibitor PD98059 (25 M). The MEK assays were performed with the same samples. The Raf kinase immunoblots were subsequently probed with the ER Ab (bottom panel) to determine the levels of Raf:ER protein. The activated A-Raf:ER kinases induced Raf activity and downstream MEK activity. The MEK1 inhibitor inhibited MEK1 activity. Similar results were observed with B-Raf:ER + BCL2 and Raf1:ER+BCL2 estradiol-responsive cells.

Figure 14  Raf induces ERK and p90^Rsk activation in Raf-responsive RAF-ER and BCL2 infected cells. (a) FD/Raf-1:ER+LNL6 (Est)c2 cell lysates were prepared from -estradiol-starved cells that were stimulated with IL-3, -estradiol or PMA. As a loading control, the upper portions of the blots were probed with the -p90^Rsk Ab (bottom panel) to determine the levels of p90^Rsk protein. Aliquots of the same cells were also starved for -estradiol in the presence of 25 M PD98059 for 24 h. ERK activation was also examined in these cells. (b) ERK2 activity in FD/Raf-1:ER+BCL2(Est)c2 cells that had been starved of -estradiol for 24 h and then treated with either IL-3 or -estradiol for the indicated times.

Figure 15  Expression of the BCL2 oncoprotein in the Raf:ER and BCL2 infected cells. The expression of the BCL2 oncoprotein was determined by immunoprecipitating BCL2 with an BCL2 Ab and followed by Western blot analysis.

Figure 16  Expression of autocrine GM-CSF cDNAs in BCL2 and Raf:ER cells. RT-PCR analysis was performed on mRNA samples from the cells grown under the indicated conditions. The EL-4 thymoma was used as a control. Higher levels of GM-CSF cDNA transcripts were detected when EL-4 cells were stimulated with PMA than in unstimulated EL4 cells. Similar amounts of mRNAs were used in the RT-PCR analysis as determined by the relatively equal levels of ₂-microglobulin detected.

Figure 17  Stimulation of [³H]-thymidine incorporation by supernants prepared from Raf-responsive cells. [³H]-thymidine incorporation induced by supernatants prepared from: (a) -estradiol (negative control); (b) FD/A-Raf:ER + BCL2(Est)c2; (c) FD/B-Raf:ER + BCL2(Est)c1; and (d) FD/Raf-1:ER + Bcl2(Est)c2.

Table 1  Abrogation of cytokine dependency of Raf:ER clones after Bcl-2 infection

Table 2  Abrogation of cytokine dependency of FDC-P1 cells after infection with BCL2 and Raf:ER containing retrovirus

Table 3  Frequency of isolation of Raf-responsive cells after infection with empty retroviral vector or a BCL2 containing retrovirus