Pim-2 is a transcriptionally regulated oncogenic kinase that promotes cell survival in response to a wide variety of proliferative signals. Deregulation of Pim-2 expression has been documented in several human malignancies, including leukemia, lymphoma, and multiple myeloma. Here, we show that the ability of Pim-2 to promote survival of cells is dependent on nuclear factor (NF)-κB activation. Pim-2 activates NF-κB–dependent gene expression by inducing phosphorylation of the oncogenic serine/threonine kinase Cot, leading to both augmentation of IκB kinase activity and a shift in nuclear NF-κB from predominantly p50 homodimers to p50/p65 heterodimers. Blockade of NF-κB function eliminates Pim-2–mediated survival in both cell lines and primary cells, and both Cot phosphorylation and expression are required for the prosurvival effects of Pim-2. Although Pim-2 cooperates with Myc to promote growth factor-independent cell proliferation, this feature is abrogated by NF-κB blockade. The ability of Pim-2 to serve as an oncogene in vivo depends on sustained NF-κB activity. Thus, the transcriptional induction of Pim-2 initiates a novel NF-κB activation pathway that regulates cell survival.
Cancer is the end result of the accumulation of genetic mutations that function cooperatively to promote neoplastic transformation. These lesions provide many advantages to transformed cells, including an increased ability to grow and proliferate relative to neighboring cells and an insensitivity to programmed cell death or apoptosis (1) . One of the most extensively studied oncogenes is the transcription factor Myc. Although Myc deregulation in isolation promotes growth, it also sensitizes cells to apoptosis. Thus, the combination of Myc expression and the deregulation of antiapoptotic oncogenes can lead to rapid tumor onset (2) .
One of the best examples of cooperation with Myc includes the Pim (proviral integration of Moloney virus) family of oncogenic serine/threonine kinases. Pim-1, Pim-2, and Pim-3 were identified as common sites of retroviral insertion in the lymphoid tumors that arose in wild-type and Myc transgenic animals after viral infection (3) . Of particular note is Pim-2, the expression of which is regulated by a wide variety of proliferative signals (4) . In humans, Pim-2 is overexpressed in multiple myeloma, leukemia, lymphoma, and prostate cancer (5, 6, 7, 8, 9) . Although 10% to 20% of mice expressing either a Pim-2 or Myc transgene develop lymphoma at 6 months of age, mice expressing both transgenes display profoundly accelerated tumor development, with all animals dying before or shortly after birth from aggressive leukemias (10) . On its own, Pim-2 is a potent regulator of cell survival in response to growth factor signaling (4) . Ectopic Pim-2 expression promotes cell survival in a manner comparable with expression of activated Akt or overexpression of Bcl-XL, although it appears to function independently of these proteins. Little is known about the pathways through which Pim-2 promotes cell growth and survival or what its molecular targets may be.
The nuclear factor (NF)-κB family of transcription factors has a well-established role in the regulation of cell survival in the immune system. The classical NF-κB pathway consists of transcription factor homodimers or heterodimers that are sequestered in the cytoplasm by direct binding to IκB proteins (11) . Phosphorylation of IκB by IκB kinases results in the release of NF-κB dimers, which transit to the nucleus and activate transcription. IκB kinases are activated by a large number of cellular signaling molecules, which are themselves affected by upstream regulation, such as cytokine stimulation. Transcriptional targets of the active NF-κB dimers include many genes associated with cell survival, including XIAP, cellular inhibitors of apoptosis (c-IAPs), FLIP, A1, Bcl-XL, and Bcl-2 (12) . The NF-κB pathway appears to be deregulated in a variety of tumors, with the sustained activity of NF-κB leading to apoptotic resistance in tumor cells (13) .
Cytokine stimulation of lymphocytes leads to the activation of intracellular kinases, such as Pim-2 and Akt (14) . Activation of Akt downstream of such signals has been proposed to promote cell growth and survival in part through NF-κB activation (15) . However, Akt activation alone does not account for the increases in NF-κB activity measured after stimulation of lymphocytes or myeloid cells, suggesting that another cytokine-induced intracellular kinase may play a role in these responses (16 , 17) .
Here, we report that ectopic expression of Pim-2 in factor-dependent cell lines leads to sustained NF-κB activity independent of growth factor availability. Constitutive Pim-2 expression correlated with NF-κB transcriptional activity and the expression of multiple NF-κB–regulated genes. Introduction of a dominant repressor IκB deletion mutant (IκBΔN) abrogated the effects of Pim-2 on cell viability. Pim-2 activated NF-κB by inducing phosphorylation of Cot, and both Cot expression and phosphorylation were required for Pim-2–mediated survival. Phosphorylation of Cot led to enhanced IκB kinase activity and reduced expression of p50/p105. These effects were abrogated by introduction of a Pim-2 phosphorylation site mutant of Cot. Transfection of cells with both Pim-2 and Myc resulted in factor-independent proliferation in tissue culture and tumor formation in mice, both of which were inhibited by an IκBΔN transgene.
MATERIALS AND METHODS
The full-length, wild-type, and kinase-dead murine Pim-2 transgenes were expressed in the pEF6-TOPO mammalian expression vector (Invitrogen, Carlsbad, CA) as described previously (4) . Murine and human FLAG-tagged Cot were generated by reverse transcription-polymerase chain reaction from FL5.12 and Jurkat cell total RNA using forward primer 5′-CCACCATGGACTACAAGGACGACGACGACAAGGAGTACATGAGCACTGGAAG-3′ and reverse primer 5′- TCAGCCGTATTCCAGGGTTGGTGGCCCA-3′for the murine gene and 5′-TCAGCCATATTCAAGCGTTGGTGGTCCCCG-3′ for the human gene. FLAG-tagged Cot was expressed in the pEF6-TOPO vector. Cot phosphorylation site mutants were generated using the Quikchange site-directed mutagenesis protocol (Stratagene, La Jolla, CA). Primers used to generate Cot mutants are as follows: (a) Cot S400A (DN Cot), 5′-CAGCCACGCTGTCAGGCTCTGGACTCTGCCCTC-3′ and 5′-GAGGGCAGAGTCCAGAGCCTGACAGCGTGGCTG-3′; (b) Cot S413A, 5′-CGCAAGAGGCTGCTGGCTAGGAAGGAGCTGGAAC-3′ and 5′-GTTCCAGCT CCTTCCTAGCCAGCAGCCTCTTGCG-3′; and (c) Cot S443A, 5′-CTCAAGAGGCAACAGGCTCTCTAC ATCGACCTC-3′ and 5′-GAGGTCGATGTAGAGAGCGCGTTGCCTCTTGAG-3′. A dominant trans-repressor FLAG-tagged deletion mutant of IκB (IκBΔN) with amino acids 1 to 36 removed was obtained (18) and cloned into the EcoRI site of pEF-6-Myc (Invitrogen). A p65 overexpression plasmid was obtained as a generous gift from M. May (University of Pennsylvania, Philadelphia, PA). RNA interference constructs were generated as described previously using a plasmid-based expression system in which short hairpin RNAs were expressed downstream of a U6 promoter in a vector derived from pBabe-Puro (4) . Targeted sequences in murine Cot were 5′-GGGCGATACTGTCCATCTCTT-3′ and 5′-GGCCTTTGGAAAAGTATACTT-3′. pSFFV-Bcl-XL was obtained as described previously (19) . Murine hemagglutinin (HA)-tagged c-Myc was cloned from FL5.12 cell total RNA by reverse transcription-polymerase chain reaction using the forward primer 5′-CCACCATGTACCCATACGATGTTCCAGATTACGCTCTTCCCCTCAACGTGAACTTCACCAACAG-3′ and reverse primer 5′-TTATGCACCAGAGTTTCGAAGCTGTTC-3′. HA–c-Myc was cloned into the EcoRI site of pSFFV and pLPC. Transient transfections in FL5.12 cells were performed using the Nucleofector system (Amaxa, Gaithersburg, MD) or by standard electroporation (950 μF capacitance, 250 V). LipofectAMINE 2000 (Invitrogen) was used for transfection of 293 cells.
Cell Lines and Culture.
All cell culture experiments were performed in triplicate. For standard growth and transient transfection experiments, FL5.12 cells were grown in standard medium [RPMI 1640 (Invitrogen) + 10% fetal calf serum (FCS; Gemini, Woodland, CA) and 400 pg/mL recombinant interleukin (IL)-3 (PharMingen, San Diego, CA)]. In transient transfection experiments, cells were withdrawn from IL-3 24 hours after transfection. Primary mouse T cells were purified from homogenized spleens and lymph nodes using the StemStep T-Cell Purification kit (Stem Cell Technologies, Vancouver, BC, Canada). Primary T cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) + 10% FCS with HEPES, l-glutamine, non-essential amino acids (NEAA), and penicillin/streptomycin (Penn/Strep) 293 cells were grown in DMEM + 10% FCS. Cell viability was assessed by exclusion of 1 μg/mL propidium iodide (Molecular Probes, Eugene, OR) using a FACScalibur flow cytometer (Becton Dickinson, San Jose, CA).
Western Blots and Antibodies.
Whole cell extracts were prepared in either radioimmunoprecipitation assay buffer or PBS containing 1% Nonidet P-40 supplemented with protease inhibitors (Roche, Indianapolis, IN) and phosphatase inhibitors (Sigma, St. Louis, MO). For Western blots, 50 μg of protein were resolved on a 4% to 12% or 10% NuPage Bis-Tris polyacrylamide gel and transferred to nitrocellulose as directed (Invitrogen), blocked in PBS containing 10% milk and 0.2% Tween 20, and then incubated in primary antibody in 5% milk overnight at 4°C with subsequent incubation with a species-appropriate horseradish peroxidase (HRP)-conjugated secondary antibody for 1 hour. Anti–Pim-2 1D12, rabbit anti-Cot M20, goat antiactin C-11, rabbit anti-p65 H286, goat anti-p50 C-19, goat anti–Cox-2 C-20, and goat anti-IκB kinase γ M-18 antibodies and mouse and goat HRP-conjugated IgG secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-IκB, rabbit anti–phospho-IκB, rabbit anti–phospho-extracellular signal-regulated kinase (ERK), rabbit anti-ERK, rabbit anti–phospho-359/363 p90 ribosomal S6 kinase (RSK), rabbit anti–phospho-380 RSK, rabbit anti-RSK antibodies, and antirabbit HRP-conjugated secondary were purchased from Cell Signaling Technologies (Beverley, MA). Rabbit anti–c-IAP2 was purchased from R&D Systems, Inc. (Minneapolis, MN). HRP-conjugated anti-FLAG M2 was purchased from Sigma. Rabbit anti-p50 was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Rat anti-HA was purchased from Roche. Bcl-XL was detected with an antibody created previously (4) . Enhanced Chemiluminescence-Plus (Amersham) was used to visualize target proteins.
A Pim-2 transgenic mouse was generated by oocyte injection of a wild-type Pim-2 transgene clone into pLck.E2, a transgenic expression vector containing the Lck promoter, a human growth hormone minigene containing splice donor and acceptor sites and a poly(A) sequence and the CD2 enhancer. Lck-FLAG-IκBΔN mice were obtained (20) and bred to Pim-2 transgenic mice. Oncogenesis experiments were performed in Avertin-sedated BALBc/Nu mice that received intravenous injection with 100 × 106 cells.
Electrophoretic Mobility Shift Assays.
Nuclear extracts were generated as described previously (21) . Ten million cells were incubated in 10 mmol/L HEPES and 10 mmol/L MgCl2 on ice for 15 minutes and then pelleted before resuspension in 20 mmol/L HEPES, 25% glycerol, 430 mmol/L NaCl, 1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, and 0.1 mmol/L KCl and incubation for 30 minutes on ice and subsequent centrifugation. Ten micrograms of the supernatant were then used as nuclear extract and added to a reaction tube containing poly(deoxyinosinic-deoxycytidylic acid), dithiothreitol, Nonidet P-40, and a T4 PNK end-labeled NF-κB gel shift oligomer, which was generated by incubating 3.5 pmol of double-stranded oligomer (Promega, Madison, WI) with 20 μCi of γ-ATP at 37°C for 30 minutes. Appropriate p50 or p65 (Santa Cruz Biotechnology) supershift antibodies were added, and the reaction was allowed to progress for 30 minutes at room temperature. Products were then separated on a 5% acrylamide gel. NF-κB–dependent transcription was assayed by a NF-κB luciferase assay, as described previously (22) .
In vitro kinase assays were performed using recombinant Pim-2 as described previously (4) . FLAG-tagged, wild-type, and Cot alanine substitution mutants were generated by transient transfection of 293 cells and subsequent immunoprecipitation using anti-FLAG M2-agarose beads (Sigma) per the supplier’s protocol. One milligram of total cell lysate was used for immunoprecipitation, with 10% reserved after three bead washes for a Western blot. Kinase reactions were resolved on 10% gels and then dried before exposure. In vivo kinase assays were performed using 293 cells transfected with both FLAG-Cot and either wild-type or kinase-dead Pim-2. After 48 hours, transfected cells were washed twice and then incubated in phosphate-free DMEM for 30 minutes. Five millicuries of [32P]Pi were then added to cells for 2 hours, and lysates were then prepared for FLAG immunoprecipitation. After immunoprecipitation, 32P-labeled proteins were resolved by SDS-PAGE electrophoresis. IκB kinase kinase assays were performed in vitro using recombinant IκB (Biosource International, Wetlake Village, CA) as a substrate.
Ectopic Pim-2 Drives Nuclear Factor-κB Activity.
Stable IL-3–dependent FL5.12 lymphoid cell lines expressing murine full-length Pim-2 or empty vector control were generated as described previously (4) . Previous work has suggested that Pim-2–mediated survival is dependent on ongoing transcription and translation, but the mechanism by which Pim-2 stimulates transcription has not been defined. One transcription factor linked to cell survival is NF-κB. The presence of p50/p65 heterodimers has been shown to correlate with NF-κB transcriptional activity and the expression of NF-κB–regulated prosurvival genes (23) . To investigate the possibility that Pim-2 induces NF-κB transcriptional activity, NF-κB gel shift analysis of cells transfected with either a Pim-2 transgene or an empty vector control was performed. Ectopic Pim-2 correlated with a greater proportion of oligonucleotide-bound p50/p65 heterodimers relative to p50/p50 homodimers as compared with control cells both in the presence of IL-3 and after IL-3 withdrawal (Fig. 1A) ⇓ . In control cells, the p50/p50 homodimer predominated. Because the observed differences in gel shift profiles might be the result of alterations in the expression level of particular NF-κB subunits, immunoblots were performed on lysates from these cells. A decrease in the levels of p50/p105 expression was reproducibly observed in the Pim-2–expressing cells, whereas the level of p65 was not affected by Pim-2 expression (Fig. 1B) ⇓ .
Given that ectopic Pim-2 maintained p50/p65 NF-κB heterodimers independent of growth factor availability, we next sought to determine whether Pim-2 expression correlated with increased NF-κB transcriptional activity. NF-κB transcriptional targets, such as the c-IAPs and the antiapoptotic oncogene Bcl-XL, are known to be key regulators of cell survival (24) . To assess whether the expression of NF-κB target genes could be correlated with Pim-2 expression, immunoblots were performed in Pim-2–expressing and control cells cultured in the presence of IL-3 or after factor withdrawal. The expression levels of c-IAP2, an inhibitor of proapoptotic caspases, Bcl-XL, and Cox-2, correlated closely with the NF-κB activity associated with Pim-2 expression (Fig. 1B) ⇓ .
Because Pim-2 increased the expression of multiple NF-κB target genes, we next sought to confirm that Pim-2 directly affects NF-κB–dependent transcriptional activity. Pim-2–expressing cells were transiently transfected with a NF-κB luciferase reporter plasmid containing six NF-κB binding sites from the murine IL-6 promoter, and luciferase activity was measured. Pim-2–expressing cells grown in the presence of IL-3 exhibited a >5-fold increase in NF-κB activity as compared with controls (5.29 ± 0.79 versus 1 ± 0.2 relative units, respectively). In the absence of growth factor, a ∼20-fold increase was observed relative to control (3.27 ± 0.58 versus 0.17 ± 0.03 units, respectively; Fig. 1C ⇓ ).
Pim-2 Expression Enhances IκB Phosphorylation and Degradation.
One possible way that Pim-2 may increase NF-κB activity is via the phosphorylation-dependent inactivation of IκB. Analysis of the phosphorylation state of serines 32 and 36 on IκB, sites critical for the release of the NF-κB transcriptional complex (12) , showed that Pim-2–expressing cells contained more phosphorylated IκB when cultured in both the presence or absence of IL-3 as compared with control cells (Fig. 2A) ⇓ . Because the phosphorylation of IκB is associated with its subsequent proteasomal degradation, we also examined the expression level of total IκB (Fig. 2A) ⇓ . Ectopic Pim-2 correlated with a decrease in total IκB relative to control, suggesting that Pim-2 promotes both phosphorylation and degradation of IκB.
Because previous work has shown that Pim-2 functions to prevent apoptosis in response to growth factor withdrawal (4) , we next sought to test whether this survival activity was dependent on NF-κB activity. NF-κB activity can be manipulated in cells through introduction of ectopic p65 or a dominant repressor form of IκB (IκBΔN), in which deletion of its NH2-terminal 36 amino acids prevents the release of cytoplasmic NF-κB (25) . Transient transfection of p65 enhanced the IL-3–independent survival of FL5.12 cells by 2-fold when compared with controls (54.3 ± 0.7% versus 29.1 ± 1%) at 24 hours but had no significant effect on the viability of Pim-2–expressing cells. Similarly, transient introduction of IκBΔN in the context of ectopic Pim-2 had no effect on the growth rate or viability of Pim-2 or control cells in the presence of IL-3. However, Pim-2–dependent protection from IL-3 withdrawal-induced apoptosis was abrogated by expression of IκBΔN (Fig. 2B and C) ⇓ . Like the empty vector control, all Pim-2–expressing cells transfected with IκBΔN were dead after 48 hours of IL-3 withdrawal.
A Pim-2 Transgene Promotes Nuclear Factor-κB–dependent Cell Survival In vivo.
To complement the above-mentioned cell line data, transgenic mice expressing full-length Pim-2 downstream of a T-cell–specific Lck promoter were generated, as well as compound Lck-Pim-2 + Lck-IκBΔN transgenics. Thymocytes from Pim-2 transgenic animals showed no change in CD4 or CD8 ratios as compared with controls (data not shown). However, the Lck-Pim-2 transgenic animals consistently had larger thymuses as compared with littermate controls, showing an increase of 2- to 3-fold in total thymocyte number (Fig. 3A) ⇓ . This was absent in compound Pim-2 + IκBΔN transgenic mice (Fig. 3A) ⇓ . To assess whether the relative degree of NF-κB activity correlated with Pim-2 and IκBΔN expression, immunoblots were performed on thymic lysates for the NF-κB target genes Cox-2, c-IAP2, and Bcl-XL. As in FL5.12 cells, expression of these NF-κB target genes was consistently increased in Pim-2 transgenic mice as compared with wild-type IκBΔN or Pim-2 + IκBΔN compound transgenics (Fig. 3B) ⇓ .
Because Pim-2 expression correlated with increased cell number and NF-κB activity in thymocytes, we next sought to determine whether Pim-2 expression might regulate T-cell survival. T cells purified from Pim-2 transgenic animals exhibited enhanced survival in culture as compared with littermate controls, with over twice as many live T cells present after 96 hours of culture (Fig. 3C) ⇓ . As demonstrated in vitro, Pim-2 promoted cell-autonomous survival in a manner dependent on NF-κB in peripheral T cells because Pim-2 + IκBΔN transgenic mice exhibited no increase in T-cell survival when compared with littermate controls (Fig. 3C) ⇓ . In contrast, the IκBΔN transgene had no generalized effect on thymocyte selection or T-cell survival on its own (Fig. 3C) ⇓ .
Cot Is Required for Pim-2 Activation of Nuclear Factor-κB.
We next sought to determine the mechanism by which Pim-2 might be regulating NF-κB activity. Although increased IκB phosphorylation correlates with Pim-2 expression in the absence of IL-3, kinase assays indicated that recombinant Pim-2 did not directly phosphorylate IκB in vitro (data not shown). One possibility is that Pim-2 promotes the phosphorylation of IκB indirectly, perhaps by affecting the activity of other serine/threonine kinases. The IκB kinase activator Cot was found by sequence analysis to contain multiple Pim consensus phosphorylation sites (26) , including serines 400 and 413.
Pim-2 kinase assays were performed with FLAG-tagged, wild-type Cot as substrate, as well as Cot mutants in which serine 400, serine 413, or both were replaced with alanines (S400A, S413A, and S400A/S413A, respectively). As a specificity control, serine 443 was separately mutated to alanine. Pim-2 induced phosphorylation of wild-type Cot in vitro (Fig. 4A) ⇓ . Phosphorylation of the mutant Cot substrates S400A, S413A, and S400A/S413A was reduced as compared with the wild-type protein (Fig. 4A) ⇓ . In contrast, an S to A substitution at residue 443 did not affect Pim-2–dependent phosphorylation (Fig. 4A) ⇓ . As an additional control, we found that kinase-dead Pim-2 did not induce phosphorylation of wild-type Cot (Fig. 4A) ⇓ . In addition, wild-type Cot did not display significant autophosphorylation but was able to carry out the phosphorylation of other proteins that contaminated the immunoprecipitate of FLAG-Cot from 293 cells. To confirm the Pim-2–dependent induction of Cot phosphorylation, we used human 293T cells, which do not express either endogenous Pim-2 or Cot (data not shown). 293T cells were transfected with FLAG-Cot and wild-type or kinase-dead Pim-2 murine transgenes and subsequently labeled with [32P]Pi, followed by Cot immunoprecipitation. An increase in Cot phosphorylation was observed with expression of kinase-active Pim-2 as compared with kinase-dead Pim-2 (Fig. 4B) ⇓ .
Cot has been reported to influence NF-κB activation indirectly through the ERK/mitogen-activated protein kinase (MAPK) pathway and directly through the IκB kinase-dependent NF-κB pathway (27 , 28) . MAPK activation can lead to indirect activation of NF-κB by stimulating the kinase RSK, which in turn can phosphorylate IκB (29) . Therefore, we sought to determine whether expression of Pim-2 results in sustained MAPK activity in the absence of growth factor. Immunoblots performed to assess the phosphorylation of ERK1/2 and p90RSK revealed that ectopic Pim-2 expression did not affect the phosphorylation of these proteins in the presence or absence of growth factor (Fig. 4C) ⇓ .
Because Pim-2 expression induced Cot phosphorylation and activated NF-κB, we sought to test whether Cot phosphorylation contributes to NF-κB activation through an effect on IκB kinase activity. Previous studies have reported that ectopic expression of Cot can increase IκB kinase activity and that a serine-to-alanine mutation at residue 400 of Cot renders Cot dominant negative (DN) with respect to its ability to activate NF-κB, but it retains the ability to activate MAPK (30) . IκB kinase kinase assays were performed using extracts from Pim-2–expressing cells transiently transfected with DN Cot or a control vector. Phosphorylation of IκB was detected in extracts from Pim-2–expressing cells grown in the presence or absence of IL-3 (Fig. 4D) ⇓ . Although the DN Cot transgene had little effect on IκB kinase activity in the presence of growth factor, the coexpression of Pim-2 and DN Cot was associated with a marked decline in IκB kinase activity (Fig. 4D) ⇓ .
To test whether expression of the DN Cot transgene would affect Pim-2–dependent survival, Pim-2–expressing cells were transfected with DN Cot, and IL-3 was withdrawn. Like IκBΔN, DN Cot expression abrogated the ability of Pim-2 to promote survival after IL-3 withdrawal (Fig. 5A) ⇓ . All Pim-2 cells transfected with DN Cot were dead after 48 hours of IL-3 deprivation.
Several studies have demonstrated that the kinase activity of Cot is regulated by the p105 form of NF-κB1 (31 , 32) . Cot has been shown to directly interact with p105, and Cot induction leads to both NF-κB activation and p105 degradation (30 , 33) . Because Pim-2 expression led to both an increase in IκB kinase activity and a decrease in p50/p105 protein levels, we sought to determine whether the observed decline in p50/p105 levels was dependent on the ability of Pim-2 to induce Cot phosphorylation. Immunoblots performed on Pim-2–expressing and control cells grown in the presence and absence of IL-3 after transfection with DN Cot or empty vector showed that Pim-2 expression led to reduced p50/p105 levels and that DN Cot reversed this effect (Fig. 5A) ⇓ .
In a complementary approach, RNA interference was used to decrease the endogenous expression of Cot in FL5.12 cells to determine whether Cot expression was required for Pim-2–mediated cell survival. Two distinct short hairpin RNAs were generated, each targeting a specific sequence in murine Cot. Both RNA hairpins were stably expressed and resulted in reduced expression of the endogenous Cot protein (Fig. 5B ⇓ , right panel). Knockdown of Cot had no effect on the viability of cells grown in IL-3 (data not shown). In IL-3-withdrawn, Pim-2–expressing cells, the repression of endogenous Cot was associated with a significant reduction in viability when compared with the Pim-2–expressing controls (Fig. 5B ⇓ , left panel).
Pim-2 Cooperation with Myc Depends on Sustained Nuclear Factor-κB Activation.
Pim-2 is an oncogene that synergizes with Myc to promote rapid leukemia formation in double transgenic mice (10) . One of the hallmarks of cellular transformation is the acquisition of growth factor-independent proliferation. In support of this idea, FL5.12 cells expressing both Pim-2 and Myc transgenes became IL-3 independent (Fig. 6) ⇓ , whereas cells expressing either transgene alone failed to accumulate in the absence of IL-3 (data not shown). Given that Pim-2 appeared to promote survival through NF-κB activation, we next sought to determine whether NF-κB activation might be a critical component of Pim-2–mediated oncogenesis. To distinguish Pim-2–specific effects on this phenotype, Pim-2 + Myc-expressing cells were compared with cells expressing another antiapoptotic oncogene that synergizes with Myc to promote IL-3 independence, Bcl-XL. Both Myc + Bcl-XL- and Myc + Pim-2–expressing cells accumulated at comparable rates in the absence of IL-3. However, the IL-3–independent accumulation of the Myc + Pim-2 cells was dramatically suppressed by the introduction of the IκBΔN transgene, whereas the growth of Myc + Bcl-XL cells was unaffected (Fig. 6) ⇓ . In addition, apoptotic cells were present in the Myc + Pim-2 cultures 24 hours after IκBΔN transfection but were not observed in the Myc + Bcl-XL cells after transfection. These results again suggest that IκBΔN does not have a generalized effect on cell growth and is instead a specific inhibitor of Pim-2–dependent growth and survival.
Another hallmark of transformation is the autonomous growth of cells in vivo. Athymic nude mice received injection with Myc + Pim-2 FL5.12 cells stably transfected with either IκBΔN or a control green fluorescent protein plasmid. When maintained in IL-3, these lines showed comparable rates of proliferation. Nude mice that received injection with the Myc + Pim-2 line all died in ≤4 weeks, manifesting significant tumor burdens after dissection as indicated by grossly enlarged spleens and lymph nodes (Fig. 7A and C) ⇓ . In comparison, 90% of the mice that received injection with Myc + Pim-2 + IκBΔN line were alive at 4 weeks. No animals given cells expressing either Pim-2 or Myc alone died during the 4 weeks after injection. All moribund mice that received injection with Pim-2 + Myc cells exhibited pronounced hepatosplenomegaly and large lymph nodes compared with controls sacrificed at the end of the experiment. Spleen mass was increased ∼5-fold in mice given Pim-2 + Myc cells as compared with those given Pim-2 alone or Pim-2 + Myc + IκBΔN (Fig. 7B and C) ⇓ . In addition, the ability of IκBΔN to suppress the Pim-2 + Myc phenotype correlated with its ability to suppress the expression of c-IAP2 (Fig. 7C) ⇓ .
Pim-2 functions as an inhibitor of apoptosis that is transcriptionally regulated by a variety of proliferative signals. Pim-2 expression maintains high levels of NF-κB activity and NF-κB–dependent gene expression in the absence of growth factor stimulation, and NF-κB activation by Pim-2 is required for its antiapoptotic function. Pim-2 expression correlates with an increase in the level of phosphorylated IκB, DNA binding of p50/p65 heterodimers, and a decrease in cellular levels of p50/p105. Mice expressing a T-cell–specific Pim-2 transgene exhibit increased thymocyte numbers and increased T-cell survival, two features that are absent in the presence of an IκBΔN transgene. Pim-2 activates NF-κB by inducing phosphorylation of Cot and increasing cellular IκB kinase activity. Both Cot phosphorylation and expression are required for Pim-2–mediated cell survival. Pim-2 and Myc cooperate to form tumors in vivo, and this cooperation requires sustained NF-κB activity.
Recent genetic studies support the conclusion that Pim-2 and NF-κB have comparable antiapoptotic effects. Random insertional mutagenesis performed on Pim-2–null/Eμ-Μyc transgenic mice indicated that deregulated expression of Tpl2 (tumor progression locus 2) could substitute for Pim-2 in cooperative oncogenesis with Myc (3) . Tpl2 was independently identified as c-Cot (cancer Osaka thyroid) based on being cloned from human thyroid carcinoma cell lines that express a transforming truncated form of Cot, which renders it constitutively active (27 , 34 , 35) . Truncated Cot has reduced binding to its cellular inhibitor p105 (31 , 36) .
The cellular Cot protein is a serine/threonine kinase that has been linked to both MAPK and NF-κB signaling downstream of surface receptor signaling in a number of different systems (28 , 37, 38, 39, 40) . It is unclear how oncogenic Cot activates NF-κB, although it has been suggested that Cot may directly bind to and activate NF-κB complexes or that it may act more indirectly by activating IκB kinase complexes through its actions on NIK or MAPK effectors (30 , 33) . In addition, recent studies have also suggested that the p105 protein can inhibit Cot MAPK activity through direct binding (32) . We did observe a decrease in cellular p50/p105 levels that correlated with Pim-2 expression and Cot phosphorylation; however, Pim-2 does not appear to activate Cot solely by reducing p50/p105 expression because DN Cot reduced Pim-2–mediated down-regulation of p50/p105. Instead, Pim-2 induction of Cot phosphorylation appears to be required for both IκB kinase activation and p50/p105 down-regulation. Together, these data support the hypothesis that Cot and p105 mutually regulate each other and that the regulation of this interaction determines whether Cot mediates activation of the MAPK/ERK pathway or the IκB kinase/NF-κB pathway (32) .
The precise mechanisms by which Cot is activated are largely unknown. Direct phosphorylation of Cot has been reported (30) . Here, we present evidence that Pim-2 induces Cot phosphorylation at multiple sites in vitro and in vivo. However, the results do not rule out the possibility that another kinase/phosphatase that coassociates with Cot acts as an intermediary between Cot and Pim-2. Nevertheless, the results demonstrate that Cot-dependent activation of NF-κB is required for Pim-2–mediated survival in both cell lines and primary cells. Furthermore, the data suggest that Cot-dependent activation of NF-κB can occur via the transcriptional induction of Pim-2 rather than as a direct result of a receptor-initiated kinase cascade.
Pim-2 expression maintains cell-autonomous NF-κB activation and target gene transcription. Because these Pim-2–dependent functions are required for Pim-2–mediated cell survival, NF-κB activation appears to be a critical mediator of the antiapoptotic function of Pim-2. Because Pim-2 expression is regulated by cytokine stimulation in a wide variety of cell types, this study identifies a novel link between cytokine receptors and NF-κB activation in nontransformed cells. In the case of neoplastic cells, the continued activation of NF-κB by Pim-2 would be pathological and confer a distinct survival advantage to the tumor, allowing it to persist in a microenvironment where growth-promoting cytokines are limiting. Pim-2 overexpression has been documented in several human cancer types, many of which are malignancies that have been successfully treated with pharmacological agents that function by targeting NF-κB, e.g., the therapeutic inhibition of NF-κB has been reported using proteasome inhibitors as chemotherapeutic agents (41 , 42) . On the basis of the results reported here, it will be of interest to test whether proteasome inhibitors are effective against tumors that overexpress Pim-2.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Craig B. Thompson, Abramson Family Cancer Research Institute, Room 451 BRB II/III, Philadelphia, PA 19104-6160. Phone: 215-746-5515; Fax: 215-746-5511; E-mail:
- Received July 1, 2004.
- Revision received September 9, 2004.
- Accepted September 13, 2004.
- ©2004 American Association for Cancer Research.