ERK5 Activates NF-κB in Leukemic T Cells and Is Essential for Their Growth In Vivo

Abs and reagents

The anti-human CD3, anti-mouse CD3, and anti-mouse CD28 mAbs were purified, as described previously (25, 26). The anti-CD28, CD69, and anti-protein kinase C (PKC)θ mAbs were from BD Biosciences. The rabbit polyclonal Abs against ERK5 were from Cell Signaling Technology or have been described (6). The p65 and p50 rabbit polyclonal Abs were from Santa Cruz Biotechnology. The rat monoclonal anti-hemagglutinin (HA) (3F10) was from Roche. Donkey anti-rabbit or sheep anti-mouse IgG Abs were obtained from Amersham, and the goat anti-hamster IgG was from Pierce Biotechnology. Annexin V-PE was from BD Pharmingen. The anti-Fas Ab was from Upstate Biotechnology.

Plasmids

The expression vectors for ERK5 and ERK5 mutant (T218/Y220F) resistant to MEK5 activation (ERK5 AEF), constitutively active MEK1 (MEK1SS3) and MEK5 (S313D/T317D, termed MEK5DD), the IκB kinase (IKK) superrepressor (27), along with the c-jun and the TRE reporter (9), have been described. The expression vectors for ERK2, β-galactosidase, AP-1 and NF-AT reporters, and IL-2 promoter have been previously used (28, 29). The conventional NF-κB reporter is based on a NF-κB binding site in the Igκ L chain promoter (30). pRc-hcRel consists of pRcCMV with c-Rel cDNA inserted in the HindIII-XbaI site (31). The 3× κB thymidine kinase luc reporter plasmid contains a trimer of the NF-κB-binding motif of the H-2k gene upstream of the thymidine kinase minimal promoter and the luciferase reporter gene (32). The pcDNA3p65 and pcDNA3p50 expression plasmids were provided by J. Alcamí (Hospital 12 de Octubre, Madrid, Spain). The GAL4-p65 and 5×Gal4 Luc plasmids were provided by L. Schmitz (Genzentrum Ludwig-Maximilians-Universität, Munich, Germany) (33). Primers containing 19 bases of ERK5 coding sequence (AGCTGCCCTGCTCAAGTCT) or a corresponding scrambled sequence were ligated into pRetroSUPER vector to generate the small hairpin RNA ERK5 (shERK5) or control expression plasmid (11). The same primers were subcloned into pSIREN-RetroQ-puro (BD Biosciences) and used to create the EL-4 stable cell lines.

Cell culture and transfection

The cell lines EL-4, Jurkat T, and Jurkat T stably expressing IκB-DN (dominant negative) (34) were grown in RPMI 1640 (Invitrogen Life Technologies) supplemented with 10% FBS, 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 1× MEM nonessential amino acid solution (Invitrogen Life Technologies), 50 μM 2-ME, and 100 U/ml each of penicillin G and streptomycin. Jurkat cells in logarithmic growth phase were transfected with the indicated amounts of plasmid by electroporation, as described (26, 28). In each experiment, cells were transfected with the same total amount of DNA by supplementing expression vector DNA with empty vector. Cells were incubated for 10 min at room temperature with the DNA mix and electroporated at 260 mV, 960 μF in 400 μl of RPMI 1640. GFP was detected by excitation with 488 Blue Laser and emission FL1 530/30 by FACS.

T cell activation

Plates were coated with 5 μg/ml of each anti-human or anti-mouse CD3 and CD28 Abs in 50 mM Tris (pH 9), left 1 h at 37°C, and at least 4 h at 4°C. After three washes with PBS, Jurkat T cells were incubated on these plates at the indicated times. Plates were placed on ice, and directly lysed with 1× SDS-PAGE sample buffer. PMA, ionomycin, etoposide, taxol, cycloheximide (CHX), and TNF-α were added to the cell suspension for the indicated times. Cells were placed on ice, washed once with cold PBS, and processed.

Immunoprecipitation of HA-tagged proteins

A total of 2 × 10⁷ transfected Jurkat Tag cells was washed twice with cold PBS, lysed in 300 μl of immunoprecipitation buffer (0.5% Nonidet P-40, 50 mM Tris-HCl (pH 8), 5 mM MgCl₂, 0.5 mM EDTA, 1 mM DTT, and 10% glycerol), and incubated on ice for 10 min. The lysate was centrifuged at 14,000 rpm for 10 min at 4°C. The supernatant was precleared by incubation 30 min at 4°C with protein G agarose beads and incubated 90 min at room temperature with 1 μg of anti-HA (3F10) Ab. Twenty microliters of a 50% slurry of protein G agarose beads was added to the reaction and incubated for additional 30 min. Beads were washed three times with immunoprecipitation buffer and resuspended in 30 μl of SDS-PAGE sample buffer, denatured, separated by SDS-PAGE, and immunoblotted, as described below.

Reporter assays

In all experiments, cells were transfected with a β-galactosidase reporter plasmid (28). Transfected cells (1 × 10⁶) were harvested after 2 days and washed twice with PBS. Cells were lysed in 100 μl of luciferase lysis buffer (Promega), and luciferase assays (40 μl) were performed, according to the manufacturer’s instructions (Promega), using a Berthold luminometer (9). For β-galactosidase assays, 40 μl of lysates was added to 200 μl of β-galactosidase assay buffer (50 mM phosphate buffer (pH 7.4), 200 μg of O-nitrophenyl-β-galactopyranoside, 1 mM MgCl₂, and 50 mM 2-ME), and the absorbance was measured at 400 nm. The results are expressed as luciferase units normalized to the corresponding β-galactosidase activity. The expression level of the transfected proteins was routinely control by immunoblot analysis.

IL-2 measurement

Cells were cultured at 1 × 10⁶ cells/ml in RPMI 1640 and stimulated as indicated with anti-CD3 Abs or PMA plus ionomycin. Supernatants were harvested, and IL-2 was measured using detection kit (BD Pharmingen). The cytokines’ values were expressed as picograms per milliliter.

Immunoblotting

Cells were washed with PBS and lysed in SDS-PAGE sample buffer. Proteins were separated by SDS-PAGE on minigels and processed for immunoblot analysis, as described previously (28).

Cell cycle

Cells were washed and resuspended in 900 μl of propidium iodide staining solution (50 μg/ml propidium iodide, 0.1% trisodium citrate dihydrate, 0.1 mg/ml RNase A, and 0.1% Triton X-100). After an overnight incubation, the cell cycle profile was determined with a FACSCalibur flow cytometer (BD Biosciences).

Subcellular fractionation

For preparation of nuclear extracts, cells were washed twice in cold PBS, resuspended in 400 μl of 10 mM HEPES (pH 7.4), 10 mM KCl, 2 mM MgCl₂, 0.1 mM EDTA, and protease inhibitors (Roche), and incubated on ice for 10 min. Nonidet P-40 was added to the cells to a final concentration of 0.5%, followed by mixing and centrifugation at 1,000 × g for 20 s at 4°C to collect nuclei. Nuclear extracts were prepared by resuspending the pellets in 50 μl of 50 mM HEPES (pH 7.8), 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 10% glycerol, and protease inhibitors (Roche). Proteins were extracted by agitation for 20 min at 4°C, the insoluble fraction was removed by centrifugation at 13,000 × g for 10 min, and the soluble proteins were analyzed by immunoblotting.

Generation of stable cell lines

A total of 5 × 10⁶ EL-4 cells was cultured at 1.5 × 10⁶ cells/ml. One milliliter of supernatant from 293T cells expressing the retroviral vector pSIREN for shERK5 or control plasmid was added 1 h after plating. The same protocol was repeated 1 day later. Three days later, cells were cultured with 2.5 μg/ml puromycin (Sigma-Aldrich). After 1 wk of selection, surviving cells were isolated and kept on selection medium until used.

Cell survival and apoptosis assays

Jurkat Tag cells were transfected with pSUPER-GFP or pSUPER-GFP-siERK5. Two days later, GFP⁺ cells were sorted and left in culture to recover for 1 additional day. These cells or EL-4 cells were treated with the indicated apoptotic agents. Where appropriate, PMA (100 ng/ml) was added to one-half of the cells 20 min before anti-Fas treatment. Cell survival was analyzed by forward/side scatter (FSC/SSC) or by annexin V staining, following the manufacturer’s instructions (BD Pharmingen), as previously described (35).

CD69 and CD25 expression

Jurkat Tag cells were transfected with pSUPER-GFP or pSUPER-GFP-siERK5. Two days later, GFP⁺ cells were sorted and left in culture to recover for 1 additional day. CD69 expression was measured essentially as previously described (29). Briefly, cells were transfected with the indicated vectors and harvested 2 days later. GFP⁺ cells were purified by FACS, washed in RPMI 1640, and suspended in 500 μl of RPMI 1640 plus 20 μl of PE-conjugated anti-human CD69 mAb for 15 min, as suggested by the manufacturer, washed twice on RPMI 1640, and analyzed by flow cytometry (FACScan; BD Biosciences) after gating for GFP-positive cells.

Tumor progression in vivo

EL-4 cells were washed in PBS and resuspended to 1.5 × 10⁶ cells/ml. A total of 2.5 × 10⁵ cells was injected s.c. in the flank of C57B6 mice. Tumor development was analyzed every 2 days. Tumors were measured, and the volume was calculated following the formula: V = length × (weight)²/2. All the experiments involving animals were performed according to the guidelines and regulations of the Centre Nationale de la Recherche Scientifique. J. Garaude, S. Cherni, R. Hipskind, and M. Villalba have the official degree for animal experimentation delivered by the Ministere de la Recherche (France).

Statistical analysis

The statistical analysis of the difference between means of paired samples was performed using the paired t test. The results are given as the confidence interval (p).

Inhibition of ERK5 expression induces apoptosis in Jurkat cells

The ERK5 cascade has been linked to several different types of cancer. To examine its role in leukemogenesis, we reduced ERK5 levels in leukemic Jurkat T cells by electroporation of a bicistronic vector that expresses shERK5 (11) and GFP. As a control, we used a scrambled version of this sequence. Two days after transfection, >50% of the living cells expressed GFP (Fig. 1⇓A, first lane). These cells were purified by FACS and placed in culture for 1 day. At this point, >90% of the cells were GFP⁺ (Fig. 1⇓A, second lane), and the expression of ERK5 was largely reduced, unlike that of ERK2 (Fig. 1⇓B). At later times, the percentage of GFP⁺ Jurkat cells transfected wtih shERK5 was substantially lower than in control shRNA-transfected cells (13.6 vs 42.7% at 4 days, and 7.2 vs 23.7% at 6 days for Jurkat-shERK5 or Jurkat-control-shRNA cells, respectively). These data suggested that expression of shERK5 inhibited cell proliferation, induced cell death, or both. In the next set of experiments, we used cells 1 day after cell sorting, i.e., 3 days after transfection.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Inhibition of ERK5-induced apoptosis in Jurkat cells. A, Twenty million Jurkat cells were transfected with 10 μg of pSUPER-GFP containing a shERK5 sequence or a control sequence. Two days after transfection, cells were analyzed for GFP expression (first lane) and purified by cell sorting. These purified cells were analyzed for GFP expression at different days. GFP was detected using flow cytometry by excitation with a 488-nm Blue Laser and by emission FL1 530/30. B, An aliquot of purified GFP⁺ cells (1 day after purification) was subjected to Western blotting to analyze the expression of ERK5 and ERK2. C, 5 × 10⁵ GFP⁺ Jurkat cells were stimulated with PMA (100 ng/ml) 20 min before Fas-Ab (5 ng/ml CH11) and harvested 24 h later. The FSC and SSC profiles were analyzed by flow cytometry. Average percentages of alive cells from three independent experiments are presented with SD in the right-side graphic. D, Jurkat cells treated and stimulated as in C were stained with annexin V-PE, and the positive cells were analyzed by flow cytometry for increase in fluorescence intensity (FL-2 height). Average percentages of apoptotic cells from three independent experiments are presented with SD in the right-side graphic. E, Jurkat T cells were treated with PMA (100 ng/ml) for 20 h, and cell survival and apoptosis were analyzed as in C and D.

We tested the effect of shERK5 on various parameters of cell viability. First, we examined cell morphology by flow cytometric analysis of cell size (FSC) and granularity (SSC). Dying, or apoptotic, cells are smaller and more granular than viable cells. Dot blots from one experiment are shown in Fig. 1⇑C, and the average of three independent experiments is presented in the graphic. In the absence of stimulation, shERK5-transfected cells showed a lower rate of survival (76%) compared with control-transfected cells (89%). In early stages of apoptosis, different changes occur at the cell membrane. One involves the translocation of phosphatidylserine to the outer side of the plasma membrane, which can be detected with Annexin V. In agreement with the results of Fig. 1⇑C, shERK5-transfected cells showed increased staining by Annexin V and therefore a higher percentage of apoptotic cells (17%) than control-transfected cells (5%). This suggests that ERK5 played a role in the survival of leukemic T cells and protected them from spontaneous apoptosis.

A classical mechanism to induce apoptosis in leukemic T cell is the engagement of death receptors, of which Fas is well characterized (36). shERK5-transfected cells were more sensitive to anti-Fas treatment, developing characteristic features of apoptosis like higher granularity and Annexin V staining (Fig. 1⇑, C and D). Therefore, blocking ERK5 expression facilitated Fas-induced apoptosis.

Treatment with the phorbol ester PMA blocks Fas-induced apoptosis in different cell types. In T cells, activation of PKCθ by PMA inhibits anti-Fas-induced cell death (35). Interestingly, shERK5-treated cells were less effectively protected from this type of apoptosis by PMA than control cells (Fig. 1⇑, C and D), suggesting that the intracellular pathways involved in protecting leukemic T cells were also impaired in shERK5-treated cells. In fact, PMA alone induced apoptosis in a small, but consistent number of shERK5-transfected cells (Fig. 1⇑E).

ERK5 knockdown renders leukemic Jurkat T cells more sensitive to TNF-α

The transcription factor NF-κB plays an antiapoptotic role in different cell lines and mediates, in part, the protective effect of phorbol esters on Fas-induced apoptosis (34, 37). Interestingly, PMA becomes proapoptotic if NF-κB activation is blocked in Jurkat T cells (34), suggesting that blocking ERK5 activity would have a similar proapoptotic effect in leukemic cells as blocking NF-κB activation. TNF-α only induces apoptosis in Jurkat cells that fail to activate NF-κB (38). Moreover, p65^−/− mice die from the increased sensitivity of hepatocytes to TNF-α, an effect that is reversed by inactivation of the type I TNF receptor (39). This protective effect of NF-κB requires gene expression, because compounds that block transcription or translation sensitize cells to TNF-α-induced apoptosis. Therefore, we investigated whether shERK5 sensitized Jurkat T cells to TNF-α. Because shERK5 induced apoptosis per se, we calculated the specific apoptosis induced by the different treatments by subtracting the percentage of apoptotic cells under resting conditions from that observed after treatment. In Jurkat cells transfected with the control scrambled shRNA, TNF-α or the protein synthesis inhibitor CHX induced apoptosis in 10% of the cells (Fig. 2⇓A), whereas the two together induced >60% apoptosis. In contrast, TNF-α alone induced apoptosis in >50% of the shERK5-transfected cells. These data suggested that ERK5 was implicated in NF-κB activation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

ERK5 expression is essential for NF-κB-induced survival. A, shERK5 sensitizes Jurkat cells to TNF-α. A half million GFP⁺ Jurkat cells were transfected as in Fig. 1⇑, sorted at day 2, and treated the next day with TNF-α (100 ng/ml), CHX (2 μg/ml), or a combination of both. Twenty-four hours later, cells were stained with annexin V-PE and analyzed by flow cytometry for increase in fluorescence intensity (FL-2 height). Mean percentages ± SEM of specific apoptosis from three independent experiments are presented in the right-side graphic, Student’s t test: ∗, p < 0.01; compare with control cells. The specific apoptosis was calculated by subtracting the percentage of apoptosis under resting conditions from the percentage of apoptosis after treatment. These values were 6 and 17% for control or shERK5 cells, respectively. B, Inhibition of ERK5 expression does not cooperate with taxol or etoposide. A total of 5 × 10⁵ GFP⁺ Jurkat cells was treated for 24 h with anti-Fas (5 ng/ml), etoposide (0.5 μM), or taxol (0.3 μM), and analyzed as in A.

Interestingly, CHX alone induced a similar rate of specific apoptosis in control and shERK5 cells. Because CHX blocks protein expression, we suspected that ERK5-induced NF-κB activation did not play a major role in CHX-induced apoptosis. Therefore, we tested two anticancer drugs whose proapoptotic effects were largely independent of NF-κB activation: etoposide, an antineoplastic agent that targets the DNA-unwinding enzyme topoisomerase II (40, 41), and taxol, which targets microtubules and causes mitotic arrest leading to apoptosis (42). Consistent with the idea that shERK5 targets NF-κB activation, shERK5 did not sensitize Jurkat cells to apoptosis induced by these two compounds (Fig. 2⇑B). In contrast, it sensitizes Jurkat cells to Fas engagement, which cooperates with NF-κB inhibition to induce apoptosis (Figs. 1⇑ and 2⇑B) (43, 44, 45). Importantly, this effect was also observed in the mouse leukemic EL-4 T cell line (see Fig. 8⇓).

ERK5 knockdown does not affect cell cycle progression or activation of leukemic T cells

The role of ERK5 in tumorigenesis could also be linked to cell proliferation. shERK5 expression in leukemic cells did not alter the cell cycle (Fig. 3⇓A). Activation of T cells with PMA/ionomycin induced cell cycle arrest at G_O/G₁, which was unaffected by ERK5 knockdown. Activation-induced IL-2 production was not inhibited in shERK5-expressing cells (Fig. 3⇓B); similarly, these cells showed no changes in other activation markers like IL-2Rα (CD25) or CD69 (Fig. 3⇓C). These results suggested that ERK5 did not play a central role in T cell activation and excluded that defects in IL-2 production or in the up-regulation of its receptor could underlie the decrease in survival of shERK5-transfected cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

The ERK5 pathway is not necessary for Jurkat T cell activation and proliferation. A, 10⁶ Jurkat cells transfected as in Fig. 1⇑ were activated with PMA (100 ng/ml) plus ionomycin (1 μg/ml) during 24 h, and the cell cycle was analyzed by propidium iodide staining and FACS analysis. B, The level of IL-2 expression in cell treated with CD3 (plate-bound), PMA (100 ng/ml), or ionomycin (1 μg/ml) during 24 h was analyzed by ELISA. C, Jurkat cells were treated as in A, and the levels of expression of CD69 and CD25 were analyzed by FACS. The percentage of positive cells for each marker is depicted in the graphic. The percentage of cells reacting with the control IgG (data not shown) was almost identical with that labeled by CD25 in resting cells. The experiments were repeated three times with similar results.

Therefore, ERK5 plays a significant role in leukemic T cell survival, but not in leukemic T cell activation or proliferation, suggesting that shERK5 did not block all T cell functions. Nevertheless, we cannot formally exclude that a low residual level of ERK5 is sufficient to mediate T cell activation and proliferation, but not T cell survival.

The ERK5 pathway activates NF-κB, but not AP-1 or NF-AT, in leukemic T cells

To test the possibility that the ERK5 pathway activates NF-κB in Jurkat cells, we used an ERK5 expression vector together with one encoding a constitutively active mutant of the upstream kinase MEK5 (MEK5DD) that heavily phosphorylates and activates ERK5 in Jurkat cells (Fig. 4⇓A) (1). Neither vector alone activated a NF-κB reporter gene in Jurkat cells. However, the combination of MEK5DD and ERK5 induced NF-κB trans activation in resting, PMA-, or PMA/ionomycin-stimulated cells (Fig. 4⇓B). Similar results were obtained with the mouse leukemic T cell line EL-4, although the induction was lower (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

The ERK5 module targets NF-κB, but not AP-1 or NF-AT, in Jurkat cells. B and C, 10⁷ Jurkat-TAg cells were cotransfected with 3 μg of each ERK5 and/or MEK5DD, 2 μg of the NF-κB or TRE luciferase reporters, and 1 μg of β-galactosidase expression vector. Control cells represent cells transfected with equal amounts of an empty vector. Two days later, equal numbers of cells were left unstimulated or stimulated for 4 h with PMA (100 ng/ml) or PMA plus ionomycin (1 μg/ml), and luciferase and β-galactosidase activities were evaluated. A, 10⁶ cells were analyzed by immunoblotting after anti-HA immunoprecipitation. The data represent means ± SD, n = 8, Student’s t test: ∗∗, p < 0.0005; compare with control cells.

Next, we examined whether ERK5 activated other transcription factors in Jurkat T cells. In fact, we have shown previously that selective activation of the ERK5 pathway targets the c-jun promoter in leukemic T cells (1). AP-1 is composed of heterodimers of the Fos and Jun families of transcription factors. We cotransfected ERK5 and/or MEK5DD and investigated the effect of these plasmids on a PMA response element (TRE)-luciferase reporter gene (46) that is driven by four canonical AP-1 binding sites (Fig. 4⇑C). The expression of either kinase alone showed no effect in resting or stimulated cells. Cotransfection of both kinases induced a 2-fold increase in resting cells, but had no significant effect in stimulated cells. Similar observations were obtained in Jurkat cells using a reporter driven by the collagenase promoter (47) that contains a single AP-1 response element (data not shown). ERK5 also failed to activate a NF-AT/AP-1 reporter derived from the IL-2 promoter (data not shown). These results show that ERK5 plays at best a minor role in AP-1 activation in T cells, probably through induction of c-Jun expression, suggesting that the antiapoptotic effect of ERK5 involves NF-κB activation.

ERK5 activates NF-κB by a noncanonical pathway

NF-κB is retained in the cytoplasm by the IκB family of proteins. Upon TCR stimulation, IκB proteins are phosphorylated by one of a number of IKKs, which leads to their ubiquitinylation and degradation. This frees NF-κB to translocate to the nucleus and activate its target genes. An IκB superrepressor (IκB-DN) that cannot be phosphorylated, and therefore is not degraded, retains NF-κB proteins in the cytosol and completely blocks their induction via IKK (27). NF-κB induction by PMA requires IKK activation, a process that was effectively blocked by IκB-DN (Fig. 5⇓A). Interestingly, IκB-DN reduced ERK5 + MEK5DD-induced NF-κB activation by only 30% in resting cells (Fig. 5⇓A), whereas the mutant completely abolished cooperation between these kinases and either PMA or PMA plus ionomycin. We obtained similar results in Jurkat cells that constitutively express IκB-DN (data not shown) (34). In T cells, PMA or anti-CD3 Abs activate PKCθ and thereby NF-κB. Conversely, inhibition of PKCθ function by the DN PKCθ-K409R mutant blocks PMA- or anti-CD3-induced NF-κB activation (30), but not that induced by ERK5 (data not shown). Therefore, ERK5 mainly activated NF-κB through a nonconventional pathway that cooperated with the canonical one to superinduce this transcription factor somewhere downstream of IκB.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

The ERK5 module activates NF-κB by a nonconventional mechanism. A, 10⁷ Jurkat-TAg cells were cotransfected with 3 μg of each ERK5 and MEK5, 6 μg of IκB superrepressor (IκB-DN), 2 μg of the NF-κB luciferase reporter, and 1 μg of β-galactosidase expression vector. Equal numbers of cells were left unstimulated or stimulated for 4 h with PMA (100 ng/ml) or PMA plus ionomycin (1 μg/ml). The data represent means ± SD, n = 4, Student’s t test: ∗, p < 0.005; compare with ERK5/MEK5-transfected cells. B, 10⁷ Jurkat-TAg cells were cotransfected with 3 μg of ERK5, MEK5, or ERK5 AEF expression vectors and 2 μg of the NF-κB luciferase reporter along with 1 μg of β-galactosidase expression vector. Equal numbers of cells were left unstimulated or stimulated for 4 h with PMA (100 ng/ml), anti-CD3 plus anti-CD28 Abs (1 μg/ml each), or TNF-α (20 ng/ml), and luciferase and β-galactosidase activities were evaluated. The data represent means ± SD, n = 4, Student’s t test: ∗, p < 0.01; compare with control-transfected cells. In the insert, an aliquot of the transfected cell lysate was separated on SDS-PAGE and analyzed by Western blotting with an anti-ERK5 Ab.

Next, we tested whether ERK5 cooperated with other signals that induce the canonical pathway. In T cells, PMA and CD3 engagement induce a similar cascade of intracellular events, but TNF-α induces a different intracellular signaling cascade before activating the IKK complex (48). The ERK5 pathway not only cooperated with PMA, but also anti-CD3 Abs and TNF-α (Fig. 5⇑B). This effect required MEK5-induced ERK5 phosphorylation, because a mutant of ERK5 that cannot be activated by MEK5 (ERK5 AEF; Fig. 5⇑B insert and Ref. 9) had no effect in resting cells and failed to cooperate with any treatment (Fig. 5⇑B). Therefore, the ERK5 cascade targets a pathway, activated by either TNF and/or the TCR, which lies downstream of IKKs.

The ERK5 pathway predominantly targets p65

Members of the NF-κB family bind as homo- or heterodimers to NF-κB regulatory elements, with p50/p65 representing the prototypical NF-κB complex and c-Rel playing a central role in T cell activation by TCR/CD28 costimulation. Although p65 and c-Rel contain transcriptional activation domains, p50 homodimers act as repressors (49). In Jurkat T cells, only overexpression of p65 activated our NF-κB reporter and cooperated with PMA above control levels (Fig. 6⇓A). Moreover, cotransfected ERK5 + MEK5DD cooperated with p65, but not with p50 or c-Rel, to activate NF-κB in resting or PMA-stimulated Jurkat cells (Fig. 6⇓A). c-Rel was active in our system, as it strongly induced a reporter controlled by a CD28 response element (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The ERK5 module targets p65 in Jurkat T cells. A and B, 10⁷ Jurkat-TAg cells were cotransfected with 3 μg of ERK5 and MEK5DD expression vectors and/or 1 μg of the expression vectors for p65, p50, and c-Rel, together with 2 μg of the corresponding luciferase reporters and 1 μg of the β-galactosidase expression vector. Two days later, equal numbers of cells were either left unstimulated or stimulated for 4 h with PMA (100 ng/ml), and luciferase and β-galactosidase activities were evaluated. The data represent means ± SD, n = 4, Student’s t test: ∗∗, p < 0.01; compare with p65-transfected cells.

We next investigated the effect of the ERK5 cascade on the 3× κB thymidine kinase reporter that is predominantly targeted by p65 (Fig. 6⇑B). Cotransfection of ERK5 + MEK5DD significantly activated this reporter in resting or PMA-stimulated cells, and synergized with p65 to activate this reporter in resting or PMA-stimulated cells (Fig. 6⇑B).

ERK5 controls p65 localization in the nucleus

The p65 protein shuttles in and out of the nucleus, where it exerts its biological activity (50). Proteins that induce p65 nuclear translocation or block its export to the cytosol will increase its transcriptional activity. To investigate whether ERK5 controlled p65 localization, we first blocked ERK5 expression with shERK5 (Fig. 7⇓A) and stimulated the cells to induce p65 nuclear translocation (Fig. 7⇓B). In the control scrambled shRNA-transfected cells, most of p65 localized in the nucleus after 30 min of stimulation, and half of the protein was still in the nucleus 2 h postinduction. In Jurkat cells with decreased ERK5 levels, the nuclear fraction of p65 was reduced after 30 min of stimulation, and was equal to resting cells after 120 min (Fig. 7⇓B). JunB and ERK2 showed the expected localization, indicating that the different fractions were largely pure.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

ERK5 controlled p65 nuclear localization. A and B, 6 × 10⁷ Jurkat cells were transfected with shERK5 or with a scramble sequence (control). Three days later, GFP⁺ cells were sorted by FACS (Fig. 1⇑), and stimulated for different times with PMA (100 ng/ml) plus ionomycin (1 μg/ml). After subcellular fractionation, the presence of the different proteins in the nuclear and cytosolic fractions was analyzed by immunoblotting. C, 10⁸ million Jurkat-TAg cells were cotransfected with 2 μg of GFP and an empty vector or ERK5 plus MEK5DD expression vectors. Two days later, GFP⁺ cells were sorted and equal numbers of cells were left unstimulated or, as a positive control, stimulated for 10 min with 20 ng/ml TNF-α. Cells were fractionated, and only the nuclear fractions were analyzed for enrichment of p65 or p50 by Western blotting. D, 10⁷ Jurkat-TAg cells were cotransfected with 1 μg of ERK5 and MEK5DD expression vectors and/or 0.5 μg of the expression vectors for Gal-DBD or Gal4-p65, together with 1.5 μg of the Gal4-luciferase reporter plasmid and 1 μg of the β-galactosidase control vector. Two days later, equal numbers of cells were either left unstimulated or stimulated for 4 h with PMA (100 ng/ml) and PMA plus ionomycin (1 μg/ml), and luciferase and β-galactosidase activities were evaluated. The data represent means ± SD, n = 4, Student’s t test: ∗, p < 0.02; ∗∗, p < 0.01; ∗∗∗, p < 0.005; compare with Gal4-p65-transfected cells.

We next selectively activated the ERK5 pathway by transfection of MEK5DD and ERK5 and investigated p65 localization. As a positive control, we used TNF-α (Fig. 7⇑C). Constitutive activation of the ERK5 pathway led to nuclear localization of p65 before stimulation without affecting that of p50 (Fig. 7⇑C). As above, the pattern of JunB and ERK2 localization confirmed the purity of the different fractions (data not shown).

ERK5 induces the trans activation activity of p65

To investigate whether the intranuclear localization of p65 correlated with an increase in its transcriptional activity, we used an expression vector (Gal4 DBD-p65) that contains the DNA binding domain of the yeast GAL4 transcription factor fused to the trans activation domain of p65. The activity of this fusion protein is not blocked by IκB association or activated by its degradation. This construct, or the Gal4 DBD alone, was transfected into Jurkat T cells along with a 5×Gal4 Luc reporter gene, in which luciferase expression is controlled by five binding sites for Gal4. Cotransfection of MEK5DD + ERK5 increased trans activation by Gal4 DBD-p65 by >4-fold (Fig. 7⇑D), and was more efficient than PMA or PMA/ionomycin. Moreover, MEK5DD + ERK5 increased the trans-activating activity of p65 induced by either treatment. Taken together, our results suggest a role for ERK5 in potentiating trans activation by p65.

ERK5 is essential for tumor progression of leukemic T cells

The mouse leukemic T cell line EL-4 has been extensively used to generate tumors in C57B6 recipient mice (51, 52). We infected EL-4 cells with a viral vector expressing shERK5 and the puromycin resistance gene to obtain a stable cell line. Although the control shRNA viral vector rapidly generated puromycin-resistant cells, the shERK5 vector gave rise to very few puromycin-resistant cells. We finally obtained a cell line, called EL-4-shERK, that showed decreased ERK5 expression (Fig. 8⇓A), increased spontaneous apoptosis relative to control cells (Fig. 8⇓C), and grew in vitro at a comparable rate to control cells. shERK5-infected cells showed impaired PMA-induced p65 accumulation in the nucleus (Fig. 8⇓B) and were more sensitive to Fas ligand-induced apoptosis (Fig. 8⇓C). All of these results were in agreement with our results in Jurkat T cells, and confirmed our hypothesis that ERK5 protects leukemic T cells from death receptor-induced apoptosis. Similarly to Jurkat cells (Fig. 2⇑), taxol, CHX, and etoposide induced comparable levels of apoptosis in EL-4-control and EL-4-shERK5 cells (data not shown). Interestingly, ERK5 down-regulation in EL-4 stable cell lines, but not in transient transfected Jurkat cells, led to an increase in the ERK2 protein. We confirmed these results by preparing a second stable shERK5-expressing EL-4 cell line (shERK5 B) (Fig. 8⇓A, right panel). Whereas ERK2 overexpression in these cells might compensate for the decrease of ERK5 in certain cellular functions, it does not account for the increase in apoptosis or the decrease in p65 accumulation in the nucleus (Fig. 8⇓, B and C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

ERK5 down-regulation in EL-4 cells inhibits p65 accumulation and facilitates death receptor-induced apoptosis. A, 10⁶ control or shERK5-expressing cells were subjected to immunoblotting to analyze the expression of ERK5, ERK2, and actin. Jurkat cells were transiently transfected, and EL-4 cells were stable transfected. Right panel, Shows the results of two EL-4 stable cell lines expressing shERK5. B, 5 × 10⁶ control or shERK5 cells were stimulated with PMA (100 ng/ml) plus ionomycin (1 μg/ml) for 30 min. After subcellular fractionation, the presence of the different proteins in the nuclear and cytosolic fractions was analyzed by immunoblotting. C, 2 × 10⁵ wt or shERK5 were treated with Fas ligand (1 ng/ml). Twenty-four hours later, cells were stained with annexin V-PE and analyzed by flow cytometry. Mean percentages ± SEM of specific apoptosis from three independent experiments are presented. Student’s t test, ∗, p < 0.01; compare with control cells. Specific apoptosis was calculated by subtracting the percentage of apoptosis under resting conditions from the percentage of apoptosis after treatment. These values were 11.4 and 14.8% for control or shERK5 cells, respectively. The experiments were performed three times with similar results.

The immune system clears tumor cells by inducing apoptosis through different mechanisms. One involves activation of death receptors on the surface of tumor cells. In contrast, some tumor cells escape immune system surveillance by becoming resistant to apoptosis. Because shERK5 rendered leukemic T cells more sensitive to death receptor-induced apoptosis, we injected EL-4-shERK5, as well as control cells, in C57B6 mice and investigated tumor formation. Wild-type (wt) EL-4 cells induced tumors in all mice (Fig. 9⇓). The tumors were first observed 1 wk after cell injection and progressed rapidly, as shown by the plots in Fig. 9⇓. In contrast, EL-4-shERK5 cells failed to develop tumors (Fig. 9⇓) in four independent experiments (n = 20). The second EL-4-shERK5 cell line (EL-4-shERK5 B) also failed to induce tumors (data not shown). In contrast, EL-4 cells stable expressing small hairpin RNA for luciferase induced tumors like the wt cells. shERK5-injected mice sacrificed at day 17 contained no tumor cells. Moreover, mice infected with EL-4-shERK5 were healthy 2 mo after injection, showing that the absence of tumors was not related to slower proliferation, but more likely reflected clearance of tumor cells. Furthermore, these results indicate that ERK2 up-regulation in shERK5-expressing cells could not compensate for the down-regulation of ERK5 in vivo.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Down-regulation of ERK5 blocks tumor development in mice. A, Five C57B6 mice were injected with 2.5 × 10⁵ control or shERK5 cells. Every 2 days, tumor size was measured. The wt mice with large tumors were euthanized 17 days after inoculation. Two months after cell injection, shERK5-injected mice were still alive and healthy. This experiment was performed four times with identical results (n = 20). B, Tumors of euthanized mice injected with different EL-4 cell lines: wt, expressing small hairpin RNA for luciferase or expressing shERK5 were weight and the average ± SD is depicted in the graphic.