TRAIL Promotes the Survival and Proliferation of Primary Human Vascular Endothelial Cells by Activating the Akt and ERK Pathways
Background— TRAIL protein is expressed in the medial smooth cell layer of aorta and pulmonary artery, whereas endothelial cells express all TRAIL receptors (TRAIL-Rs).
Methods and Results— The role of TRAIL/TRAIL-Rs in vascular biology was investigated in primary human umbilical vein endothelial cells (HUVECs) and aortic endothelial cells, which showed comparable surface expression of death (TRAIL-R1 and -R2) and decoy (TRAIL-R3 and -R4) TRAIL-Rs. TRAIL activated the protein kinase Akt in HUVECs, as assessed by Western blot for phospho-Akt. Moreover, experiments performed with a pharmacological inhibitor of the phosphatidylinositol 3-kinase/Akt pathway (LY294002) or a dominant-negative Akt (K179M) demonstrated that TRAIL significantly protected HUVECs from apoptosis induced by trophic withdrawal via Akt and that inhibition of Akt sensitized HUVECs to TRAIL-induced caspase-dependent apoptosis. TRAIL also stimulated the ERK1/2 but not the p38 or the JNK pathways and induced a significant increase in endothelial cell proliferation in an ERK-dependent manner. Conversely, TRAIL did not activate NF-κB or affect the surface expression of the inflammatory markers E-selectin, intercellular adhesion molecule-1, and vascular cell adhesion molecule-1.
Conclusions— The ability of TRAIL to promote the survival/proliferation of endothelial cells without inducing NF-κB activation and inflammatory markers suggests that the TRAIL/TRAIL-R system plays an important role in endothelial cell physiology.
Received November 8, 2002; revision received December 31, 2002; accepted January 21, 2003.
TRAIL is a member of the tumor necrosis factor (TNF) family of cytokines, which play important roles in regulating cell death and inflammation.1 TRAIL, which exists either as a type II membrane or as a soluble protein,2 interacts with 4 high-affinity membrane receptors, belonging to the apoptosis-inducing TNF-receptor (TNF-R) family. TRAIL-R1 and TRAIL-R2 transduce apoptotic signals on binding of TRAIL, whereas TRAIL-R3 and TRAIL-R4 lack intracellular death domain and apoptosis-inducing capability and have been proposed to function as decoy receptors, protecting normal cells, including endothelial cells, from apoptosis.3
Despite its potential as an anticancer therapeutic agent both in vitro and in vivo,4,5 the wide expression of TRAIL and TRAIL-Rs in many normal tissues1 suggests that the physiological role of TRAIL is more complex than merely activating caspase-dependent apoptosis of cancer cells. Although very little is known about possible nonapoptotic functions induced by TRAIL, it has been shown that endothelial cells express the mRNA for all TRAIL receptors3,6 and that TRAIL protein is expressed by the medial smooth cell layer of aorta and pulmonary artery.7 It is also noteworthy that cleavage of membrane TRAIL from the cell surface requires the action of cysteine proteases,2 which are abundantly present in the vessel wall.8
The aim of this study was to investigate the biological activity of TRAIL on vascular endothelial cells, which are among the principal physiological targets of the proinflammatory actions of TNF-α, the prototype of the TNF family of cytokines.9 As a model system of vascular endothelial cells, we used human aortic endothelial cells and human umbilical vein endothelial cells (HUVECs), which were exposed to either TRAIL or TNF-α and characterized in terms of apoptotic/survival and proliferation pathways and expression of inflammatory markers.
Recombinant histidine 6–tagged TRAIL was produced as described previously.10 Recombinant TNF-α and vascular endothelial growth factor (VEGF) were purchased from Peprotec. Polymyxin B, LY294002, PD98059, SB203580, SP600125, and partenolide were from Calbiochem. The optimal concentrations of TRAIL and TNF-α (100 ng/mL each) and of pharmacological inhibitors were determined in preliminary dose-response experiments.
For Western blot analyses, the following antibodies (Abs) were used: rabbit Abs to native (Transduction Laboratories) and mouse monoclonal antibodies (mAbs) to Ser473-phosphorylated forms of Akt; rabbit Abs anti-ERK1/2, anti–phospho-ERK1/2, anti-p38, anti–phospho-p38, anti-JNK, and anti–phospho-JNK (all from New England BioLabs). The enhanced chemiluminescence reagent detection system was from DuPont-NEN.
Flow cytometric analyses were performed by FACScan (Becton Dickinson); the following antibodies were used: mAbs anti–human TRAIL-R1, TRAIL-R2, TRAIL-R3, and TRAIL-R4 (all from Alexis Biochemicals); PE-conjugated anti-mouse secondary Abs (Immunotech); the FITC-conjugated mAbs anti–E-selectin, anti–intercellular adhesion molecule (ICAM)-1 (Bender Medical System), and anti–vascular cell adhesion molecule (VCAM)-1 (Cymbus Biotechnology). Nonspecific fluorescence was assessed by use of normal mouse IgG followed by a second layer (as above) or by incubation with irrelevant isotype-matched conjugated antibodies.
Cell Cultures, Transfection, and Apoptosis Analysis
Primary HUVECs, obtained as described previously,6 were grown on 0.2% gelatin-coated tissue culture plates in M199 endothelial growth medium (BioWhittaker) supplemented with 20% FBS, 10 μg/mL heparin, and 50 μg/mL ECGF (Sigma Chemical). Human aortic endothelial cells were purchased from BioWhittaker and cultured in basal endothelial growth medium supplemented with 2% FBS, 12 μg/mL bovine brain extract, 1 μg/mL hydrocortisone, and 10 ng/mL human ECGF (all from BioWhittaker). In all experiments, cells were used between the third and fifth passages in vitro.
In some experiments, subconfluent HUVECs were transfected with pEGFP-C1 empty vector (Clontech) or EGFP-tagged constructs expressing either a membrane-targeted constitutively active form of Akt (Myr-Akt) or a dominant-negative form carrying a point mutation that ablates the kinase activity (Akt-K179M; both generous gifts of Dr S. Marmiroli, University of Modena) by use of ProFection DEAE-Dextran transfection assay (Promega). Twenty-four hours after transfection, cultures were washed to remove transfection medium and those cells that died as a consequence of toxic effects of DEAE-dextran solution. The percentage of transfected cells, analyzed at this time point by fluorescence microscopy, ranged from 28% to 46% in 4 different experiments. HUVECs were then starved (M199 medium+0.5% FCS) for 18 hours in the absence or presence of TRAIL before the degree of apoptosis was evaluated.
For apoptosis analysis, at different times after treatments, substrate-attached HUVECs were harvested by trypsin treatment and pooled with floating cells to analyze the degree of apoptosis in the entire cell population. Cells were then double-stained with propidium iodide and FITC-conjugated annexin-V (Alexis Biochemicals) according to the manufacturer’s instructions and analyzed by flow cytometry as detailed previously.10
Western Blot Analysis
For Western blot analysis, HUVECs were plated in 10-cm dishes, grown at subconfluence, and subjected to partial FCS reduction (to 0.5%) and complete growth factor withdrawal for 18 hours before the addition of TRAIL or TNF-α. Cells were harvested in lysis buffer containing 1% Triton X-100, Pefablock (1 mmol/L), aprotinin (10 μg/mL), pepstatin (1 μg/mL), leupeptin (10 μg/mL), NaF (10 mmol/L), and Na3VO4 (1 mmol/L). Protein determination was performed by Bradford assay (Bio-Rad). Equal amounts of protein (50 μg) for each sample were migrated in acrylamide gels and blotted onto nitrocellulose filters. Blotted filters were probed with primary antibodies for the phosphorylated Akt, ERK1/2, p38, or JNK. After incubation with peroxidase-conjugated anti-rabbit or anti-mouse IgG, specific reactions were revealed with the enhanced chemiluminescence Western blotting detection system. Membranes were stripped by incubation in Re-Blot 1X antibody stripping solution (Chemicon International) and reprobed for the respective total protein kinase content or β-actin to verify loading evenness.
Densitometric values were expressed in arbitrary units and estimated by ImageQuant software (Molecular Dynamics). Multiple film exposures were used to verify the linearity of the samples analyzed and avoid saturation of the film.
Assays for Caspase Activity and NF-κB DNA Binding
Caspase-8 and caspase-3 activity was measured with caspase colorimetric assay kits (Alexis Biochemicals) as described by the manufacturer. NF-κB induction was measured with the Trans-AM NF-κB p65 kit (Active Motif), which measures the level of the active form of NF-κB contained in cell extracts able to specifically bind to an oligonucleotide containing the NF-κB consensus site (5′-GGGACTTTCC-3′) attached to a 96-well plate. Caspase activity and NF-κB DNA binding were determined as absorbance values measured with a microplate reader (Multiskan Ascent; Dasit).
HUVECs were plated onto 96-well plates at a density of 5×103 cells/well. The next day, the medium was changed to endothelial cell basal medium containing 0.5% FBS and 0.1% BSA (starvation medium). The cells were then pretreated with inhibitors for 1 hour and incubated with cytokines for 30 hours longer. [3H]Thymidine (1 μCi) was added to each well during the last 6 hours of incubation. [3H]Thymidine-labeled DNA was then measured by liquid scintillation counting.
The results were evaluated by ANOVA with subsequent comparisons by Student’s t test for paired or nonpaired data, as appropriate. Statistical significance was defined as a value of P<0.05. Values are reported as mean±SD.
TRAIL Independently Activates the Akt and ERK1/2 Pathways in Primary Vascular Endothelial Cells
Both HUVECs and aortic endothelial cells (data not shown) showed a similar surface expression of all transmembrane TRAIL-Rs (TRAIL-R1, -R2, -R3, and -R4) at flow cytometry (Figure 1). To investigate the intracellular pathways engaged by the interaction between TRAIL and TRAIL-Rs in endothelial cells, cultured HUVECs were made quiescent by growth factor withdrawal and reduction of serum (to 0.5%) for 18 hours before treatments. The ability of TRAIL to modulate the Akt and ERK1/2 pathways was investigated by Western blot analyses using antibodies specific for the residues that are phosphorylated in each kinase on activation.11–15 For comparison, cells were also treated with TNF-α.11,12 Both cytokines stimulated Akt phosphorylation, which peaked at 15 to 30 minutes of treatment with either TRAIL or TNF-α, as assessed by evaluation of total and Ser473-phosphorylated forms of Akt (Figure 2).
To date, ≥3 subgroups of mitogen-activated protein (MAP) kinase (MAPK) family members have been involved in a wide range of cellular responses.13 The first subgroup includes 2 isoforms of the extracellular signal–regulated kinases, ERK1 and ERK2, which are strongly activated by mitogens. The other 2 subgroups, namely SAPK1/JNK1 (stress-activated protein kinase-1/c-Jun NH2-terminal kinase) and SAPK2/p38, are weakly activated by mitogens but are highly stimulated by inflammatory cytokines and environmental stress inducers.13 Exposure to both TRAIL and TNF-α induced ERK1/2 phosphorylation from 5 minutes of treatment onward (Figure 3A). Conversely, although TRAIL did not activate p38 and JNK1 at any time point examined, the relative levels of phospho-p38 MAPK and phospho-JNK1 reached a maximum value after 15 to 30 minutes of TNF-α exposure (Figure 3, B and C). The TRAIL-induced activation of Akt and ERK1/2 pathways was not caused by potential contaminating endotoxin, because polymyxin B failed to block TRAIL activity (data not shown). Moreover, the effect of specific cell-permeable inhibitors, such as LY294002 (40 μmol/L), inhibitor of phosphatidylinositol (PI) 3-kinase/Akt pathway, and PD98059 (50 μmol/L), a commonly used inhibitor of the ERK pathway, was also assessed on the activation of Akt and ERK1/2 induced by TRAIL. Each inhibitor selectively and specifically blocked the ability of TRAIL to activate Akt or ERK1/2 (Figure 3D), supporting the view that TRAIL activates these pathways independently.
TRAIL-Induced Akt Activation Mediates Antiapoptotic Function in Primary Vascular Endothelial Cells
At 6 hours of trophic withdrawal, whereas in untreated, LY294002- and cytokine-treated HUVEC cultures, apoptosis was still <15%, in cells supplemented with TRAIL+LY294002, the percentage of apoptosis was significantly increased (>40%, P<0.01) (Figure 4A). Apoptosis was increased (P<0.01) in cultures treated with TNF-α+LY294002 also but to a lesser extent (P<0.01) than in those treated with TRAIL+LY294002 (Figure 4A). Consistently, the combination of TRAIL+LY294002 induced a significant (P<0.05) increase of caspase-8 and caspase-3 activity (Figure 4A) compared with all other culture conditions.
On prolonged trophic withdrawal, untreated HUVECs showed a progressive loss of viability coupled with a concomitant increase of apoptosis. Indeed, at 18 hours of trophic withdrawal, apoptosis was >40% in untreated cells, whereas it was significantly (P<0.01) lower in cultures treated with either TNF-α or TRAIL, TNF-α showing a greater (P<0.01) prosurvival activity than TRAIL (Figure 4B). At this time point, LY294002 either alone or in combination with cytokines induced massive apoptosis (>70%) and caspase activation (Figure 4B), completely overriding the protective effects of TRAIL and partially that of TNF-α. In contrast, PD98059 had no effect on apoptosis and caspase activation in HUVECs in both the absence and presence of cytokines (data not shown).
The role of Akt in TRAIL-mediated protection of HUVECs from apoptosis was confirmed by transfection experiments performed with Akt constructs. Whereas constitutively active Myr-Akt mimicked the protective effect of TRAIL, expression of the dominant-negative Akt-K179M was associated with increased cell apoptosis and completely abrogated the ability of TRAIL to protect HUVECs from trophic withdrawal–induced apoptosis (Figure 4C).
Overall, these data suggest that Akt plays an important role in mediating survival signaling in HUVECs in response to TRAIL and that endothelial cells can be sensitized to TRAIL-mediated apoptosis by blocking of this pathway.
TRAIL Induces an ERK-Dependent Increase of Endothelial Cell Proliferation
Next, by analyzing the thymidine incorporation in serum-starved cultures (Figure 5), we explored whether the ability of TRAIL to activate Akt and ERK affected endothelial cell proliferation. Of note, TRAIL significantly increased (≈2-fold, P<0.01) HUVEC thymidine incorporation at a similar degree of VEGF, used as positive control.16,17 On the contrary, TNF-α significantly (P<0.01) reduced thymidine incorporation in HUVECs with respect to untreated controls. In these experiments, LY294002 could not be used because of its high toxicity. PD98059 was not toxic and completely blocked cell proliferation in all culture conditions (Figure 5). SB203580 (10 μmol/L) and SP600125 (10 μmol/L), pharmacological inhibitors of p38/MAPK and JNK1, respectively, did not affect the TRAIL stimulatory activity and did not relieve the inhibitory activity of TNF-α (data not shown), suggesting that pathways other than p38 and JNK1 were involved in TNF-α–mediated cell growth inhibition.
TRAIL Does Not Activate NF-κB or Affect the Expression of Adhesion Molecules
We next demonstrated that TRAIL was unable to activate the NF-κB pathway in endothelial cells at any time point examined (Figure 6A). Conversely, as expected,9,18 TNF-α significantly (P<0.01) stimulated the DNA binding activity of NF-κB in HUVECs, and this activation was completely abrogated by partenolide (10 μmol/L), a pharmacological inhibitor of the NF-κB pathway (Figure 6A). In addition, both SB203580 and partenolide partially (P<0.05), and to similar degrees, counteracted the TNF-α–mediated protection from apoptosis. Although partenolide+SB203580 did not show any additive effect with respect to each molecule used alone, partenolide+LY294002 completely abrogated the prosurvival activity of TNF-α, indicating that unlike TRAIL, TNF-α also activated a p38/NF-κB prosurvival pathway in HUVECs (Figure 6B).
One of the major proinflammatory responses in endothelial cells initiated by TNF-α is the NF-κB–dependent upregulation of E-selectin, ICAM-1, or VCAM-1 expression.9 E-selectin, like VCAM-1, is typically not detected in unactivated endothelial cells but is rapidly synthesized in response to proinflammatory stimuli, thus making it a marker of the activated endothelial phenotype. ICAM-1 is present on endothelial cells at all times, but its expression is increased after treatment of cells with inflammatory cytokines.9 Whereas E-selectin supports the initial rolling of leukocytes on activated endothelium, ICAM-1 has an important role in the migration of leukocytes to sites of inflammation, enabling the firm adhesion and diapedesis of leukocytes. Although expression of ICAM-1, E-selectin, and VCAM-1 was not affected by exposure to TRAIL, TNF-α significantly (P<0.01) increased the expression of these antigens (Figure 7A). Moreover, the ability of TNF-α to upregulate inflammatory markers was significantly counteracted by both SB203580 and partenolide but not by SP600125, implying that a p38/NF-κB pathway also played a key role in mediating the upregulation of inflammatory markers induced by TNF-α (Figure 7B).
Many studies have shown that TRAIL is a potent apoptosis inducer in malignant cells, whereas its role in normal cell physiology is much less well understood.1 We have demonstrated here that both aortic endothelial cells and HUVECs exhibit a similar pattern of surface TRAIL-R expression. Moreover, in vascular endothelial cells, TRAIL stimulates the phosphorylation of the serine/threonine kinase Akt in a manner dependent on PI3K activation.14 The ability of TRAIL to activate the antiapoptotic PI3K/Akt pathway in endothelial cells is a completely new and unexpected finding, also in consideration of the large number of studies underlining the proapoptotic activity of TRAIL, at least in malignant cells.1,4,5 It should be underlined that the PI3K/Akt pathway is of central importance in endothelial cell biology, conferring survival to endothelial cells in response to angiogenic cytokine stimulation, fluid shear stress, and matrix attachment signals. This pathway is essential also for endothelial cell differentiation, migration, and NO production, key features of the angiogenic response.19
It is also noteworthy that inhibitors of the PI3K/Akt pathway were able to sensitize endothelial cells to TRAIL-mediated cytotoxicity. A possible explanation of these findings is that the PI3K/Akt pathway impairs the activation of the apical caspases by inhibiting the recruitment of procaspase-8 to the death-inducing signaling complex.20 We consistently observed a robust caspase activation and induction of apoptosis after 6 hours of starvation in the presence of TRAIL+LY294002. Thus, under certain conditions, TRAIL may also be involved in blood vessel regression. Moreover, the differential degree of protection from trophic withdrawal–induced apoptosis between TRAIL and TNF-α, observed both in the absence and in the presence of LY294002, is probably because of the PI3K-independent p38-dependent activation of NF-κB by TNF-α. In fact, although the PI3K/Akt and the ERK pathways have been reported to activate NF-κB under some circumstances,18 TRAIL did not stimulate NF-κB in primary HUVECs, and thus, it did not affect the surface expression of the inflammatory markers.
We have also demonstrated that, among the members of the MAPK family, TRAIL is very effective in activating ERK. Consistent with the previous demonstration that the ERK1/2 pathway is a central element in transducing mitogenic signals in endothelial cells,16 we have also demonstrated that TRAIL increased HUVEC proliferation and that such effect was abrogated by PD98059. Remarkably, TRAIL acts as a mitogen for primary endothelial cells at a degree comparable to VEGF, a major regulator of both physiological and pathological angiogenesis.19 Collectively, these experiments demonstrate that Akt and ERK are critical intermediates in TRAIL-induced protection from apoptosis and proliferation in endothelial cells, respectively.
The ability of TRAIL to promote endothelial cell survival/proliferation is not without precedent, because TWEAK, another TNF family member, has recently been shown to promote endothelial cell survival and angiogenesis.21,22 Given the importance of the TNF family in many immune responses and the contribution of vascular endothelium to immune processes such as inflammation, it is not surprising that TNF ligands can affect vascular tissue. However, data presented here show that TRAIL has a direct effect on endothelial cell survival and proliferation without inducing inflammatory markers.
The magnitude of the TRAIL-induced proliferative response is similar to that of more thoroughly characterized angiogenic factors, such as VEGF. An interesting parallelism between VEGF and TRAIL is that both are produced by the vascular smooth muscle cells of the human vascular wall and most likely modulate endothelial cell functions in a paracrine pathway.7,23 Thus, TRAIL may also be implicated in the homeostatic control of endothelial biology and possibly in angiogenesis.
This study was supported by the Fondi d’Incentivazone per la Ricerca di Base (FIRB) and Associazione Italiana per la Ricerca sul Cancro (AIRC) grants.
Sheridan JP, Marsters SA, Pitti PM, et al. Control of TRAIL-induced apoptosis by a family of signaling and decoy receptors. Science. 1997; 277: 818–822.
Gochuico BR, Zhang J, Ma BY, et al. TRAIL expression in vascular smooth muscle. Am J Physiol. 2000; 278: L1045–L1050.
Collins T, Read MA, Neish AS, et al. Transcriptional regulation of endothelial cell adhesion molecules: NF-kappa B and cytokine-inducible enhancers. FASEB J. 1995; 9: 899–909.
Secchiero P, Gonelli A, Celeghini C, et al. Activation of the nitric oxide synthase pathway represents a key component of tumor necrosis factor-related apoptosis-inducing ligand-mediated cytotoxicity on hematologic malignancies. Blood. 2001; 98: 2220–2228.
Madge LA, Pober JS. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFκB in human endothelial cells. J Biol Chem. 2000; 275: 15458–15465.
Lee SE, Chung WJ, Kwak HB, et al. Tumor necrosis factor-α supports the survival of osteoclasts through the activation of Akt and ERK. J Biol Chem. 2001; 276: 49343–49349.
Suhara T, Mano T, Oliveira BE, et al. Phosphatidylinositol 3-kinase/Akt signaling controls endothelial cell sensitivity to Fas-mediated apoptosis via regulation of FLICE-inhibitory protein (FLIP). Circ Res. 2001; 89: 13–19.
Jones RG, Elford AR, Parsons MJ, et al. CD28-dependent activation of protein kinase B/Akt blocks Fas-mediated apoptosis by preventing death-inducing signaling complex assembly. J Exp Med. 2002; 196: 335–348.
Lynch CN, Wang YC, Lund JK, et al. TWEAK induces angiogenesis and proliferation of endothelial cells. J Biol Chem. 1999; 274: 8455–8459.
Jakubowski A, Browning B, Lukashev M, et al. Dual role for TWEAK in angiogenic regulation. J Cell Sci. 2002; 115: 267–274.
Brogi E, Wu T, Namiki A, et al. Indirect angiogenic cytokines upregulate VEGF and bFGF gene expression in vascular smooth muscle cells, whereas hypoxia upregulates VEGF expression only. Circulation. 1994; 90: 649–652.