PP2A Promotes Endothelial Survival via Stabilization of Translational Inhibitor 4E-BP1 Following Exposure to TNFα

Abstract

Objective

TNFα may change from a stimulator of reversible activation of endothelial cells (ECs) to a killer when combined with cycloheximide (CHX). The means by which endothelial cells are destined to either the survival or the apoptotic pathways are not fully understood. We investigated the role of p38 MAPK and protein phosphatase 2A (PP2A) activation and their regulation of 4E-BP1 stability in ECs to determine whether this pathway contributes to apoptosis induced by TNFα and CHX.

Methods and Results

Apoptosis was induced in human umbilical vein ECs (HUVECs) by treating them with a combination of TNFα and cycloheximide (CHX) [TNFα/CHX]. Activation of p38 MAPK was increased in HUVECs undergoing apoptosis, which was associated with degradation of eIF4E regulator, 4E-BP1, in a p38 MAPK-dependent manner. CHX attenuated a TNFα-stimulated increase in the expression and activity of PP2A. Silencing PP2A expression with siRNA transfection mimicked CHX-sensitization, increasing HUVEC apoptosis with TNFα stimulation, suggesting a protective role for PP2A in the apoptotic process.

Conclusions

Our data suggest i) TNFα stimulates PP2A and that HUVECs elude apoptosis by PP2A-dependent de-phosphorylation of p38 MAPK and ii) CHX-induced inhibition of PP2A leads to maintenance of p38 activity and degradation of 4E-BP1, resulting in enhanced TNFα-induced apoptosis.

Tumor necrosis factor-α (TNFα), a pro-inflammatory cytokine which mediates apoptosis in endothelial cells, is implicated in the pathogenesis of atherosclerosis.¹ Although it has been well-established that endothelial cell (EC) apoptosis is an important process underlying the pathogenesis of atherosclerotic plaque,² the molecular mechanisms responsible for apoptosis of ECs in the setting of TNFα exposure remains elusive, since TNFα simultaneously stimulates pathways for apoptotic response (e.g., caspase activation) and pathways for survival (e.g., activation of the transcription factor NFκB).^{3, 4} Ultimate fate of the cell is determined by the balance between pro- and anti-apoptotic stimuli and is a consequence of crosstalk between the major TNFα-induced signaling pathways.^{5, 6}

A clue to solving the identity of protective factors that are activated with TNFα in ECs has come from the cell culture model; ECs in vitro will not undergo apoptosis in the presence of TNFα unless they are sensitized by exposure to an inhibitor of protein synthesis such as cycloheximide (CHX),⁷ a well-known ribotoxin.⁸ Surprisingly, although CHX sensitizes ECs to apoptosis with TNFα exposure, CHX on its own does not induce EC apoptosis, and has been shown in several in vivo studies to protect the endothelium from the formation atherosclerosis.^9–11 This is therefore an interesting paradox; that while TNFα on its own does not cause apoptosis of ECs, and CHX alone has beneficial effects on atherosclerosis, a combination of the two leads to endothelial cell apoptosis. Various hypotheses have been put forward to explain this phenomenon.

In view of the fact that inhibition of protein synthesis by CHX exposure alone does not induce apoptosis in ECs, we investigated a novel mechanism as to whether TNFα/CHX treatment can modulate 4E-BP1, the inhibitor of eIF4E, and thereby lead to apoptosis of the EC. In addition to its well-known role in initiation of cap-dependent translation, eIF4E has been more recently identified as a master regulator of cell survival.¹² Several investigators have identified 4E-BP1 as an important regulator in maintaining cellular viability during conditions of cellular stress, such as hypoxia.¹³ Energy homeostasis is maintained by 4E-BP1 by its sequestration of eIF4E and resulting reduction of translation initiation. Thus, 4E-BP1 is a well-conserved metabolic brake. In addition, cleavage of the full sized 4E-BP1 polypeptide by caspase generates a peptide fragment that sequesters and inhibits eIF4E even more potently than the full-length 4E-BP1.¹⁴ Peptide products sequester eIF4E by binding to its conserved binding site, involving tryptophan 73.^{12, 15} Studies suggest that increasing intracellular levels of peptides containing a conserved eIF4E-binding motif found within 4E-BP1, with the ability to bind eIF4E, leads to rapid dose-dependent apoptosis that is not linked to inhibition of cap-dependent translation.^{12, 16}

In agreement with other studies, we demonstrate that the p38 pathway plays a role in regulating the TNFα/CHX-induced EC apoptosis. In addition, our data suggest a novel mechanism by which CHX decreases EC resistance to apoptosis with TNFα stimulation. This work presents evidence that CHX, through inhibition of PP2A activity, leads to uninhibited up-regulation of p38 in ECs, which in turn increases the degradation of 4E-BP1. Ultimately we have provided new insights into the mechanisms modulating the vascular endothelial apoptotic response with TNFα/CHX treatment.

Methods

Reagents

Recombinant human TNFα was purchased from R&D Systems (Minneapolis, MN). Cycloheximide (CHX), rapamycin (Rapa), and okadaic acid (OA) were purchased from Sigma (St.Louis, MO). SB203580, was from Calbiochem (San Diego, CA). Cell culture media were purchased from Lonza (Walkersville, MD). Antibodies used for this study were anti-p38, anti-phospho p38 (New England Biolabs, Beverly, MA), anti-4EBP1 and anti-phospho 4EBP1 antibodies (Cell Signaling Technology, Beverly, MA) and anti-Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were purchased from Chemicon (Temecula, CA).

Cell culture

Primary cultures of human umbilical venous ECs (HUVECs) were obtained as described,¹⁷ and used at passage 3 or 4. HUVECs were cultured in endothelial cell basal medium 2 (EBM-2) (Lonza, Walkersville, MD). The medium was supplemented with EGM-2 bullet kit. HUVECs were cultured until 80% confluent and then serum starved overnight. Cells were incubated in 100-mm culture dishes then typsinized. Floating cells were collected, pelleted, and pooled with trypsinized cells, and viability was assessed with trypan blue exclusion.

Fluorescence-activated cell sorter (FACS) analysis

Cells were cultured as previously and after treatment, aliquots of cells (1×10⁶) were resuspended in 500 µl binding buffer and stained with 5 µl fluorescein isothiocyanate (FITC)-labeled annexin-V according to the manufacturer's instructions. Propidium iodide (PI, 5 µl) was added to the samples after staining with annexin-V, and then put in the dark for 30 minutes. Flow cytometry (BD, FACSCalibur, USA) was performed immediately after staining.

Western analysis

Cells were lysed on ice in cell lysis buffer. Protein concentration was measured using Bio-Rad protein assay reagent. Lysates (30 µg) were boiled in Laemmli sample buffer containing β-mercaptoethanol and resolved electrophoretically on 10% SDS-polyacrylamide gel. The gels were electrotransferred to a polyvinylidine difluoride membrane (Bio-Rad) using a tank blot procedure. Membranes were incubated with primary antibodies and respective horseradish peroxidase–linked to secondary antibody. Immunoreactive bands were visualized by enhanced chemiluminescence detection system (Amersham).

Immunoprecipitation

For immunoprecipitation analysis, the Dynabead Protein A kit was utilized (Invitrogen). Briefly, cell lysates were incubated with anti-eIF4E or anti-eIF4G antibody which was bound to the magnetic beads in 200ul PBS/Tween 20 and kit steps were followed. Target antigen was eluted and prepared for western blotting.

Determination of caspase activity

Caspase-3 activity was analyzed by a fluorescence spectrophotometric assay, using the flourogenic peptide DEVD-AMC (BDPharmingen, San Diego, CA) as a substrate. After appropriate cell treatment, adherent cells were harvested.

RNA Interference to Silence Expression of PP2A

Small interfering RNA (siRNA) oligonucleotides, as well as control oligonucleotides, were purchased from Santa Cruz Biotechnology (cat # sc-43509). Briefly, each well of a 6-well plate containing a subconfluent HUVEC culture was transfected with 100nM siRNA using Lipofectin (Invitrogen) according to manufacturer’s protocol. The cells were then incubated at 37°C overnight. After 8 hours, the cells in each well were replaced with fresh growth medium. At 72 h after transfection, the cells transfected with control and experimental siRNA were used for experiments and were harvested separately for extraction of total protein and used for western blot analysis.

Statistics

All data are expressed as means ± SEM, and Student’s t-test was used for statistical analysis of the differences. For statistical analysis of more than two groups, data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test. Analysis was accomplished with PRISM (Graphpad, La Jolla, CA, USA). The experiments were repeated three to five times, and data from representative experiments are shown. (*p < 0.05, **p < 0.01, ***p < 0.001).

Results

Co-treatment with TNFα/CHX induces apoptosis in HUVECs

HUVECs are resistant to the cytotoxic effects of either TNFα or CHX alone at low doses; however TNFα and CHX work synergistically to induce apoptosis when cells are incubated with TNFα and CHX together (Figure 1).⁷ No significant difference was observed on cell viability when HUVECs were treated with TNFα or CHX alone. However, when the cells were treated with a combination of TNFα and CHX, cell viability decreased significantly by 8 hours (Figure 1A). TNFα alone increased the activity of caspase-3 after 4 hours of treatment, but a decrease of caspase-3 activity was seen subsequently at 8 hours, with no significant changes in cell viability compared to untreated cells throughout the incubation period (Figure 1B). CHX had no significant effect if given alone, but simultaneous treatment with TNFα increased caspase-3 activation with a peak at 8 hours. Figure 1C shows Annexin-V and PI staining assays with flow cytometry which quantified the effect of TNFα and CHX on apoptosis induction after 8 hours of treatment. The percent annexin-V positive increased significantly in TNFα/CHX compared to other groups (10.5 +/− 3.0%).

Figure 1.

HUVECs were treated with TNFα (10ng/ml) with or without CHX (10µg/ml) for 8 h. (A) Following treatment, cells were subjected to cell count by trypan blue exclusion method. Results are expressed as mean of three different experiments ± SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001 compared with control. (B) HUVECs were treated with TNFα with or without CHX for the times indicated. Caspase-3 activity was measured as described in “Methods.” Results are expressed as mean of three different experiments ± SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001 compared with control, ***, **p < 0.01 compared with control. (C) HUVECs were treated with TNFα (10ng/ml) with or without CHX (10µg/ml) for 8 h and then subjected to fluorescence-activated cell sorting (FACS) to analyze the percentage of cells in staining positive for annexin-V and propidium iodide. (D) Cells were exposed to TNFα (10ng/ml) with or without CHX (10µg/ml) for 4 hours, and processed for immuno-precipitation as described in “Methods.” Results are representative of three different experiments. (E) Cells were exposed to TNFα (10ng/ml) with or without CHX (10µg/ml) for 4 hours, and processed for western blotting. Results are representative of three different experiments.

Increased binding of eIF4E to 4E-BP1 in TNFα/CHX-induced apoptosis

During cellular conditions favoring apoptosis, 4E-BP1 is hypo-phosphorylated (active form), allowing its binding to eIF4E, thus impeding binding to eIF4G and formation of an active initiation complex. We first examined the levels of initiation complexes in HUVECs treated with TNFα/CHX through immunoprecipitation with anti-eIF4E antibody and anti-eIF4G antibodies. Figure 1D shows that at 4 hours of TNFα/CHX treatment, 4E-BP1-eIF4E complexes were increased (top panel) with a simultaneous reduction in eIF4E-eIF4G complexes (bottom panel), suggesting that formation of cap-dependent translation initiation machinery, the eIF4E-eIF4G protein complex, is reduced when cells were treated with a combination of TNFα and CHX.

4EBP1 is phosphorylated at multiple sites including Thr46, Ser64 and Thr69. p38 MAPK has been shown to phosphorylate several sites on 4E-BP1 including Ser-64. We next analyzed the phosphorylation state of 4EBP1 by western blot at these sites to determine whether 4E-BP1 hypo-phosphorylation correlates with increased binding to eIF4E. Figure 1 E shows that with four hours of treatment, CHX had a stimulatory effect on the level of phosphorylation of 4E-BP1, particularly on Ser64 and Thr69 (Figure 1E, compare lanes 1 and 3). Figure 2A shows a time course revealing that although treatment of HUVECs with TNFα/CHX led to an initial increase of Thr37/46 phosphorylation, there was an overall reduction of 4E-BP1 phosphorylation at these sites by 8 hours (top panel). Remarkably, there was a concurrent and significant reduction of total (full-length) 4E-BP1 at 8 hours following TNFα/CHX treatment (bottom panel). Interestingly, this reduction in total level of 4E-BP1 proteins seen with TNFα/CHX treatment was not observed with TNFα or with CHX alone (Figure 2B). Moreover, TNFα/CHX-mediated 4E-BP1 degradation is not due to inhibition of the classic mTOR pathway since total 4E-BP1 protein levels are stable in the presence of rapamycin, a direct inhibitor of mTOR (Figure 2B).

Figure 2.

(A) Western blotting is shown for HUVECs treated with TNFα (10ng/ml) and CHX (10µg/ml) over a time course from 0–8 hours. (B) Western blotting is shown for HUVECs treated with TNFα (10ng/ml) with and without CHX (10µg/ml) for 8 hr, in the absence or presence of 30 min rapamycin (100nM) pre-treatment.

p38 MAPK Inhibitor SB203580 Decreases TNFα/CHX-induced Apoptosis

The p38 pathway has been shown previously i) to regulate the 4E-BP1 and eIF4E complexes ¹⁸ and ii) to be required for TNFα-mediated apoptosis in ECs.¹⁹ We therefore examined whether p38 signaling is also involved in TNFα/CHX-induced apoptosis in HUVECs. Treatment of HUVECs with TNFα induced phosphorylation of p38 within 15 minutes (Figure 3A). Figure 3A also shows that p38 phosphorylation is not significantly increased with CHX at 15 minutes, but in cells co-treated with TNFα and CHX, there is a significant increase in p38 phosphorylation when compared to cells treated with TNFα alone. To further investigate the involvement of p38, we tested the effects of the p38 inhibitor SB203580 on activation of caspase-3. As compared with TNFα/CHX co-treatment, SB203580 attenuated p38 activation significantly (Figure 3A). In addition, SB203580 treatment led to protection from cell death (Figure 3B, left panel) and a 45% decrease of caspase-3 release at 8 hours (Figure 3B, right panel), suggesting that p38 plays a pro-apoptotic role in TNFα/CHX-induced apoptosis in HUVECs. The increased level of TNFα-induced p38 phosphorylation seen with the addition of CHX (Figure 3A) indicated that CHX facilitates the maintenance of p38 activation.²⁰ Furthermore, pre-treatment with p38 inhibitor SB203580 abrogates 4E-BP1 degradation mediated by TNFα/CHX treatment (Figure 4), suggesting a role for p38/4E-BP1 pathway in TNFα/CHX mediated apoptosis of HUVECs. Rapamycin pre-treatment also partially prevented the TNFα/CHX-induced decrease in 4E-BP1 protein levels. These data taken together suggest that increased phosphorylation of 4E-BP1 may contribute to its disappearance.²¹

Figure 3.

(A) HUVECs were preincubated with or without SB203580 (20µM) for 30 min and further treated for 15 min with TNFα (10ng/ml) with or without CHX (10µg/ml). Total cell lysates were prepared and processed for western blotting with p-p38 and total p38 antibodies. Results are representative of three different experiments. Bottom panel, arbitrary densitometric units; mean of three different experiments; bars, SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001. (B) Left panel, HUVECs were preincubated with or without SB203580 (20µM) for 30 min and further treated with TNFα and CHX for 8 hr and. After treatment, cells were subjected to cell count by trypan blue exclusion method. Results are expressed as mean of three different experiments ± SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001. Right panel, HUVECs were preincubated with or without SB203580 (20µM) for 30 min and further treated with TNFα (10ng/ml) and CHX (10µg/ml) for 8 hr. Caspase-3 activity was measured as described in “Materials and Methods.” Results are expressed as mean of three different experiments ± SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001.

Figure 4.

p38 MAPK inhibitor SB203580 abrogates 4E-BP1 degradation. HUVECs were treated with TNFα (10ng/ml) with or without CHX (10µg/ml) for 8 hr, following rapamycin (100nM) pre-treatment, or SB203580 (20µM) pre-treatment for 30 min as indicated. Total cell lysates were prepared and processed for western blotting with total 4E-BP1 and GAPDH antibodies. Results are representative of three different experiments. Bottom panel, arbitrary densitometric units; mean of three different experiments; bars. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001.

Decreased expression of PP2A with CHX treatment

Our results indicate that CHX may be sensitizing ECs to TNFα-induced apoptosis by enhancing the activation of p38 (Figure 3A). To determine the mechanism behind the enhanced activation of p38, we examined the known negative regulators of p38 phosphorylation. Although MAPK phosphatases (MKP1 and MKP2), primarily localized in the nucleus, play a key role in p38 inactivation,^{22, 23} we chose to study type 2A phosphatases (PP2A) since it catalyzes the dephosphorylation of both p38¹⁸ and 4E-BP1.²¹ The expression of PP2AC subunit was increased with TNFα stimulation, persisting to 8 hours (Figure 5A, left panel). With TNFα/CHX co-treatment, there was an initial increase in total PP2A protein level by 30 minutes but the total protein level of PP2A was reduced significantly by 4–8 hours (Figure 5A, right panel). To confirm that decreased protein levels of PP2A following TNFα/CHX treatment correlate with reduced PP2A activity, we measured PP2A activity by a phosphatase assay using immunoprecipitated PP2A. TNFα stimulation causes a significant increase in PP2A activity (200% compared to untreated cells; Figure 5B). However, CHX treatment and co-treatment with TNFα/CHX inhibited PP2A activity to 50% of the level seen in untreated cells. To determine whether CHX is enhancing TNFα-promoted 4E-BP1 degradation through inhibition of PP2A de-phosphorylation, we examined the total 4E-BP1 protein levels in the presence of okadaic acid (OA), an inhibitor of PP2A. We found a further decrease in the level of 4E-BP1 protein with OA pre-treatment (Figure 5C). OA pre-treatment also mimicked CHX treatment in sensitizing HUVECs to TNFα-induced apoptosis. That is, OA pre-treatment led to increased HUVEC cell death and a 30% increase of caspase-3 release at 8 hours (data not shown).

Figure 5.

(A) HUVECs were exposed to TNFα (10ng/ml)) from 0 to 8 hours as indicated then prepared for western blotting and membranes probed with PP2A and GAPDH antibodies. Right panel, HUVECs were exposed to TNFα (10ng/ml) and CHX (10µg/ml) from 0 to 8 hours as indicated then prepared for western blotting and membranes probed with PP2A and GAPDH antibodies. Results are representative of three different experiments. Bottom panel, arbitrary densitometric units; mean of three different experiments; bars. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001, **, p < 0.01 compared to control. (B) HUVECs were treated with TNFα (10ng/ml) with or without CHX (10µg/ml) for times indicated. Cells were then harvested and PP2A activity measured as described in “Methods.” Results are expressed as mean of three different experiments ± SE. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001, **, p < 0.01 compared to control. (C) HUVECs were preincubated with or without okadaic acid (100 nM) for 30 min and further treated for 8 min with TNFα (10ng/ml) with or without CHX (10µg/ml). Total cell lysates were prepared and processed for western blotting with total 4E-BP1 and GAPDH. Results are representative of three different experiments. Data were analysed by ANOVA, and pairwise comparisons were performed by Bonferroni/Dunn post test ***, p < 0.001, **, p < 0.01 compared to control. (D) HUVECs, transfected with either control or PP2A antisense oligonucleotides, were exposed to TNFα (10ng/ml) for 2 hours (left panel and middle panel) or 8 hours (right panel). Western blotting with PP2A, phospho p-38, anti-cleaved caspase-3, GAPDH, and total 4E-BP1 are shown.

To demonstrate that inhibition of PP2A sensitizes HUVECs to TNFα-induced apoptosis, we transiently transfected HUVECs with PP2A silencing RNA (siRNA) and treated with TNFα alone. Figure 5D (middle panel) shows increased phosphorylation of p38 (top row) and increased levels of cleaved caspase-3 (second row) when transfected cells are treated with TNFα for 2 hours, suggesting that HUVECs avoid TNFα-induced apoptosis through maintenance of negative feedback to activated pathways (p38) through de-phosphorylation. Figure 5D (right panel) demonstrates that silencing of PP2A in the setting of TNFα leads to degradation of 4E-BP1 (8 hours of TNFα stimulation). This supports a model in which CHX sensitizes HUVECs to apoptosis through inhibition of PP2A which activates uninhibited p38 activity and increased degradation of 4E-BP1.

Discussion

In agreement with other investigators,^{3, 4, 6, 7} we demonstrated that TNFα does not induce endothelial cell apoptosis, unless these cells are made susceptible to apoptosis by the addition of cycloheximide (CHX). Our results presented a novel mechanism by which CHX sensitizes HUVECs to TNFα-mediated apoptosis. It is well known that p38 MAPK activation plays a critical role in TNFα-induced EC apoptosis.⁵ TNFα alone only transiently activates p38 MAPK, since its activity is quickly terminated following dephosphorylation by PP2A.²⁰ As shown by other investigators, CHX maintains p38 phosphorylation, through inhibition of phosphatases PP2A²⁰ and MKP-1/2. ²⁴

Our model illustrated that CHX sensitizes ECs to TNFα-induced apoptosis by first, inhibiting PP2A activity, allowing p38MAPK to remain in its active phosphorylated state. Second, treatment of HUVECs with TNFα/CHX decreased the apparent half-life of 4E-BP1, a key component in the general translational machinery. This effect was also due to PP2A inhibition. Whereas 4E-BP1 is normally a stable protein that is long-lasting in the presence of CHX alone,²¹ the combination of TNFα and CHX uniquely led to a rapid degradation of 4E-BP1 in a p38 MAPK-dependent manner (Figure 4). Indeed, PP2A inhibition with both OA and siRNA in our system mimicked the effects seen with CHX treatment and increased the susceptibility of HUVECs to apoptosis. Thus, our data suggest that PP2A inhibition by CHX during TNFα stimulation i) increased the activity of p38-MAPK and ii) induced 4E-BP1 degradation.²¹ Our results are in agreement with previous studies which identified PP2A activity as an important mechanism underlying cell survival, likely through its regulation of 4E-BP1.²⁵

As reviewed by Boudreau,²⁶ the precise role of PP2A in the maintenance of cell survival and apoptosis in response to PP2A inhibition are not well understood. PP2A plays a role in regulating apoptosis in a stimulation dependent and cell-type dependent manner. In many instances PP2A serves as a positive regulator of apoptosis because it activates pro-apoptotic Bcl-2 family members. However, our data support a protective role for protein phosphatase 2A (PP2A). In our model, PP2A behaved as a counter-regulator of p38 MAPK signaling as demonstrated by Cornell et al in epithelial cells.²⁷ In this model, TNFα stimulation also activated both the p38 MAPK pathway and PP2A which terminated ongoing p38 MAPK activation. We hypothesized that HUVECs resist apoptosis with TNFα (alone) stimulation due to the ability of intact PP2A to quickly de-phosphorylate TNFα-activated p38 MAPK. The addition of CHX (TNFα/CHX) disables PP2A’s ability to turn off p38 MAPK. Indeed, PP2A silencing with siRNA in our system mimicked the effect of CHX and increased the susceptibility of HUVECs to apoptosis. Our results agree with a report by Lee et al. showing a direct protein interaction between PP2A and p38 MAPK in endothelial cells and increased p38 MAPK activity with PP2A inhibition.²⁸ In addition, the effect of PP2A may differ depending not only on the cell type but also temporally. Recent in vivo studies in rat hearts reveal that PP2A inhibition immediately prior to the onset of sustained ischemia abolishes protection during reperfusion (this was associated with activation of p38 MAPK), while PP2A inhibition during reperfusion may have a cardioprotective effect (when no upregulation of p38MAPK was observed).²⁹

Protein synthesis inhibition noted in cells fated to undergo apoptosis is associated with a decrease in the proportion of ribosomes in polysomes³⁰ suggesting that there is an inhibition of the initiation phase, characterized by recruitment of the 40 S ribosomal subunit to the mRNA. The initiation phase occurs through ribosomal recognition of the mRNA 5’ cap protein complex, created by interaction of three subunits: eukaryotic initiation factor 4A (eIF4A), eIF4E, and eIF4G. The eIF4E polypeptide, which is involved in binding the mRNA cap to the ribosome, is the rate-limiting component of the initiation of protein translation.³¹ In addition to its role in translation initiation, recent studies have implicated eIF4E as a key anti-apoptotic protein, likely due to its function in the exporting nuclear growth-related mRNAs from the cell nucleus to the cytoplasm.¹² Indeed, there is evidence that eIF4E could represent a ‘checkpoint’ protein by which cells sense the integrity of the translation machinery and perturbing eIF4E function through binding rapidly and directly triggers the apoptotic machinery.¹⁶

eIF4E activity is controlled through reversible interaction with 4E-binding protein-1 (4E-BP1). 4E-BP1 competes with eIF4G for binding to eIF4E.³² In quiescent cells, hypo-phosphorylated 4E-BP1 binds to and inhibits eIF4E, but upon exposure to a variety of extracellular stimuli, 4E-BP1 is phosphorylated, resulting in bond disruption, release of eIF4E, and binding of eIF4E to eIF4G in order to initiate mRNA cap-dependent translation.³³ In MCF-7 and HeLa cells, translation inhibition preceding apoptosis is characterized by de-phosphorylation of 4E-BP1.³⁴

Multiple stimuli affect the phosphorylation of 4E-BP1. The protein kinase mammalian target of rapamycin (mTOR) plays a key role in phosphorylation of 4E-BP1, but other kinases, such as the phosphatidylinositol 3-kinase (PI3K)/Akt,³⁵ and p38 mitogen-activated protein kinase (p38 MAPK)¹⁸ have been implicated. De-phosphorylation of 4E-BP1 is catalyzed by protein phosphatase 2A (PP2A).³⁶

It is well known that 4E-BP1 can bind and inhibit eIF4E when 4E-BP1 is in its hypo-phosphorylated state. However recent studies have revealed a different mechanism by which the cellular apoptotic machinery can use 4E-BP1 to inhibit eIF4E during cellular stress; through degradation of the full-size 4E-BP1 protein. For instance, early in apoptosis, caspase activation leads to cleavage of full sized 4E-BP1 polypeptide. This cleavage generates a fragment that sequesters and inhibits eIF4E even more potently than the full-length 4E-BP1.¹⁴

Increased degradation of 4E-BP1 can also occur through 4E-BP1 hyper-phosphorylation and subsequent ubiquitination leading to proteosomal degradation.²¹ p38 MAPK, which plays a key role in a variety of cellular responses to stresses such as viral infection, osmotic shock and UV irradiation,^37–39 may be increasing turnover of 4E-BP1 protein by hyper-phosphorylation. This effect has also been described in HSV-1 infection of primary human epithelial cells in which p38 MAPK is activated. HSV-1 infection results in hyper-phosphorylation of 4E-BP1 and a reduction of 4E-BP1 steady-state protein levels due to an increase in proteosomal degradation.³⁹ In the quiescent state of the cell, 4E-BP1 binding to eIF4E might not only prevent eIF4E from associating with eIF4G, but may also maintain 4E-BP1 protein levels. Unbound 4E-BP1 is free to become phosphorylated and thus susceptible to degradation. Rapid degradation of phosphorylated 4E-BP1 by viral invasion ensures that PP2A is unable to dephosphorylate 4E-BP1, restoring its ability to bind eIF4E.³⁹ Similarly, the addition of CHX in our model may lead to 4E-BP1 degradation through hyper-phosphorylation. Thus, our model may serve as a potential illustration of a common mechanism of viral inflammation leading to EC dysfunction and apoptosis.

In conclusion, we have shown a novel mechanism by which endothelial cells (ECs) both a) resist apoptosis with exposure to TNFα and b) how CHX induces vulnerability to apoptosis in ECs exposed to TNFα. We postulate that CHX in our model may elucidate the role of putative factors found in atherosclerotic plaque, such as lipid peroxidation products, which contribute to apoptosis by modulating TNFα activation in the endothelium.⁴⁰ Oxidative stress and resulting lipid peroxidation products, such as 4-hydroxynonenal (4HNE), have been shown to play a powerful role in the modulation of cell signalling and inhibition of protein synthesis respectively, thereby facilitating the apoptotic actions of TNFα on endothelial cells in vivo.⁴¹ Co-treatment of HUVECs with TNFα and CHX led to rapid degradation of eIF4E regulator, 4E-BP1, in a p38 MAPK-dependent manner. As shown by others, 4E-BP1 is an important regulator of cell stress and is important for maintaining cell viability.¹³ Degradation of 4E-BP1 is likely to promote apoptosis through several mechanisms. First, loss of the full-length 4E-BP1 may impair energy homeostasis through dysregulation of translation initiation.¹³ In addition, degradation of 4E-BP1 gives rise to a truncated 4E-BP1 peptide that binds to the strongly anti-apoptotic protein eIF4E, thereby leading to potent inhibition of eIF4E.¹⁴ Modulation of eIF4E activity may therefore be a mechanism by which CHX potentiates TNFα-mediated EC apoptosis.