Saturated, but Not Unsaturated, Fatty Acids Induce Apoptosis of Human Coronary Artery Endothelial Cells via Nuclear Factor-κB Activation

HCAECs were purchased from Clonetics/BioWhittaker (Verviers, Belgium) and cultured in the commercially available endothelial cell growth medium EGM-2-MV (EBM-2 supplemented with EGM-2-MV SingleQuots; Clonetics/BioWhittaker). This medium contained 5% (vol/vol) FCS. Cells were kept in this medium during all experiments and were not serum starved. Only cells from passages 2 and 3 were used for experiments. Cells were obtained from healthy donors (aged 16–56 years) who had given their informed consent. In most experiments, cells were incubated with NEFA. NEFAs (Sigma-Aldrich, Taufkirchen, Germany) were bound to fatty acid–free BSA, as previously described (10). In brief, NEFAs (200 mmol/l in ethanol except for stearate, which was dissolved to 100 mmol/l) were diluted 1:25 into Krebs-Ringer HEPES buffer containing 20% (wt/vol) BSA. This mixture was gently agitated at 37°C under nitrogen overnight. Control medium containing ethanol and BSA was prepared analogously. These stock solutions were stored in aliquots under nitrogen at −20°C. At the NEFA concentration used, BSA reached a concentration of 2.5 or 5% (wt/vol) in the medium. The SN50 inhibitor peptide and hypoestoxide were obtained from Calbiochem (Schwalbach, Germany), trichodion from Alexis (Grünberg, Germany), and human tumor necrosis factor (TNF)-α from R&D Systems (Wiesbaden, Germany). Triacsin C, cycloserine, fumonisin B1, etomoxir, cycloheximide, actinomycin D, and the neutralizing anti–TNF-α antibody were obtained from Sigma-Aldrich.

Confluent cells were treated as indicated. Detached cells were harvested from the supernatant by centrifugation and added to the adherent cells harvested by trypsinization. Cells were washed with PBS, fixed in 70% ice-cold ethanol, centrifuged, and washed again with PBS. After staining with propidium iodide (50 μg/ml) diluted in PBS containing ribonuclease A (100 μg/ml), cells were subjected to flow cytometric analysis of DNA content using a Becton Dickinson FACScalibur cytometer. Percentages of cells in the different cell cycle phases were calculated by CellQuest software (Becton Dickinson, Heidelberg, Germany).

Nuclear proteins were prepared as described previously (11). Synthetic oligonucleotides containing a high-affinity binding site for nuclear factor κB (NFκB), 5′-GTTAGTTGAGGGGACTTTCCCAGGC-3′, were end-labeled with [α-³²P]dATP (3,000 Ci/mmol) and Klenow enzyme and incubated with up to 10 μg nuclear protein in 20 μl of 22 mmol/l HEPES-KOH, pH 7.9, 70 mmol/l KCl, 2.2 mmol/l dithiothreitol, and 10% glycerol on ice for 20 min. Polydeoxyinosinic-deoxycytidylic acid (0.05 mg/ml) was added as unspecific competitor. The samples were run on a 5% nondenaturing polyacrylamide gel in a buffer containing 25 mmol/l Tris-HCl, pH 8.0, 190 mmol/l glycine, and 1 mmol/l EDTA. Gels were dried and analyzed by autoradiography.

Intracellular and secreted proteins (TNF-α and interleukin [IL]-1β) were quantified with Quantikine enzyme-linked immunosorbent assays from R&D Systems. To measure cytokines in cell lysates, cell monolayers were washed with PBS and scraped off in PBS supplemented with 1 mmol/l EDTA, 1 mmol/l phenylmethylsulfonylfluoride, 10 μg/ml aprotinin, 0.5 μg/ml leupeptin, and 0.7 μg/ml pepstatin. After cell lysis by sonication, lysates were cleared by centrifugation. Measurements were performed in duplicate. For standardization, cellular protein of the cell lysates was determined with the Bradford method. Cytokines in cell culture supernatants were measured after centrifugation.

Data were analyzed by ANOVA with Bonferroni’s post hoc test. A P value <0.05 was considered statistically significant. For these tests, the statistical software package SigmaStat for Windows 1.0 (Jandel, San Rafael, CA) was used.

HCAECs were treated for 24 h with different saturated and unsaturated NEFAs (1 mmol/l each), and apoptosis was measured by flow cytometric cell cycle analysis (quantification of cells with subG1 DNA content). The saturated NEFAs stearate (C18:0) and, to a lesser extent, palmitate (C16:0) significantly induced apoptosis in HCAECs (Fig. 1A). The apoptotic mechanism was recently confirmed by demonstration of caspase-3 activation after palmitate and stearate treatment (9). We were not able to detect saturated NEFA-induced cytochrome c release, which is seen in most, but not all, forms of apoptosis (data not shown). Compared with these saturated NEFAs, the carrier BSA (2.5 and 5%, data not shown) and the monounsaturated NEFA palmitoleate (C16:1 ω7) and oleate (C18:1 ω9) and the polyunsaturated NEFA linoleate (C18:2 ω6) had no proapoptotic effect (Fig. 1A). Moreover, all unsaturated NEFAs tested significantly prevented stearate-induced apoptosis (Fig. 1B). In palmitate-induced apoptosis, unsaturated NEFAs also tended to be protective. Due to the weak proapoptotic effect of palmitate, protection by unsaturated NEFAs, however, did not reach the level of significance (Fig. 1B). Thus, saturated and unsaturated NEFAs display differential effects on HCAEC survival; saturated NEFAs trigger apoptosis, whereas unsaturated NEFAs prevent this kind of lipoapoptosis.

As palmitate induces apoptosis in other cell systems via ceramide biosynthesis (rev. in 12), we investigated the role of ceramide formation in lipoapoptosis of HCAECs. To this end, we treated HCAECs with specific inhibitors of several steps of the ceramide de novo synthesis pathway. Triacsin C was recently shown to efficiently inhibit the long-chain acyl-CoA synthetase activity (palmitoyl-CoA formation) of HCAECs at 5–10 μmol/l (13). In our experimental setting, triacsin C (10 μmol/l) revealed cytotoxic effects in the absence of NEFA (Fig. 2). This might be due to the reduction of basal intracellular acyl-CoA concentrations below a critical level necessary for the maintainance of cellular viability. Apart from this, triacsin C significantly prevented palmitate- and stearate-induced apoptosis (Fig. 2). In contrast, inhibitors of enzymes more downstream in the ceramide biosynthesis pathway, i.e., cycloserine (1 mmol/l) and fumonisin B1 (50 μmol/l), had no protective effect (Fig. 2). In addition, treatment of cells with exogenous C2-ceramide (10 μmol/l) did not induce apoptosis in HCAECs (data not shown). Therefore, NEFA activation via fatty acyl-CoA formation, the first step of NEFA metabolism, but not ceramide de novo synthesis, seems to be required for lipoapoptosis in HCAECs. Treating HCAECs with etomoxir (100 μmol/l), an inhibitor of carnitine palmitoyltransferase I that catalyzes the rate-limiting step of the β-oxidation, had no significant effect on lipoapoptosis (Fig. 2). This suggests that β-oxidation is not involved in the mechanism used by saturated NEFAs to induce lipoapoptosis in these cells.

As treatment of HCAECs with the inhibitor of gene transcription actinomycin D (10 μg/ml) or the inhibitor of protein biosynthesis cycloheximide (20 μg/ml) significantly prevented stearate-induced apoptosis (Fig. 3), de novo synthesis of protein(s) seems to be required for HCAEC lipoapoptosis. Therefore, we investigated potential candidate transcription factors for apoptotic gene expression. One transcription factor known to be activated by NEFA is NFκB (14–16). Figure 4 shows that the proapoptotic saturated NEFA stearate promotes activation of NFκB. In agreement with its minor proapoptotic properties, the saturated NEFA palmitate had only a weak effect on NFκB activity (data not shown). Linoleate, which was chosen as a representative unsaturated NEFA shown above to lack proapoptotic effects in HCAECs, is not able to induce NFκB activation (Fig. 4). Thus, activation of NFκB appears to be restricted to the proapoptotic saturated NEFA. Consistent with its antilipoapoptotic effect, linoleate reduced stearate-induced NFκB activation to the level of BSA (Fig. 5B). Furthermore, treatment of HCAECs with the NFκB inhibitors SN50 (18 μmol/l) and trichodion (50 μmol/l) protected HCAECs from lipoapoptosis (Fig. 5A) and reduced stearate-induced NFκB activation even below the level of BSA (Fig. 5B). Inhibitor of κB kinase (IKK) is the best characterized upstream kinase regulating NFκB activity via phosphorylation of inhibitor of κB. Treatment of HCAECs with hypoestoxide (100 μmol/l), a selective and direct inhibitor of IKK, prevented lipoapoptosis (Fig. 6). Taken together, these data suggest that palmitate and stearate exert their lipoapoptotic effects via activation of IKK and NFκB.

To investigate whether saturated NEFAs activate NFκB via induction of NFκB-activating cytokines, such as TNF-α or IL-1β, we determined TNF-α and IL-1β production after treament of HCAECs with palmitate or stearate (1 mmol/l each). Under these conditions, HCAECs did not produce measurable amounts of intracellular or secreted IL-1β protein (data not shown). TNF-α was detected in minor amounts in the supernatant (2.5% BSA: 12.5 ± 1.8 pg/ml; 5% BSA: 13.5 ± 2.7 pg/ml; n = 3). However, neither palmitate nor stearate was able to significantly increase the amount of secreted TNF-α protein over the respective BSA control (1 mmol/l palmitate: 12.4 ± 2.2 pg/ml; 1 mmol/l stearate: 23.1 ± 10.5 pg/ml; n = 3). In addition, the detected TNF-α concentrations (in the picograms per milliliter range) are considered too low to significantly induce apoptosis in HCAECs, as exogenously applied TNF-α concentrations up to 500 ng/ml were not effective in triggering substantial apoptosis (data not shown). Furthermore, a neutralizing anti–TNF-α antibody was not able to significantly reduce palmitate- or stearate-induced apoptosis in HCAECs (data not shown). Taken together, these data suggest that saturated NEFAs do not activate NFκB via stimulation of TNF-α or IL-1β synthesis/secretion.

Consistent with our recent report (9), the common plasma saturated NEFAs palmitate and stearate, at 1 mmol/l, clearly reveal proapoptotic effects on HCAECs. Taking into account that 1) the total plasma NEFA concentration in the fasting state can reach 1.6 mmol/l, as detected in some healthy participants of the TÜF (Tübingen family study for type 2 diabetes) (A. Fritsche, N. Stefan, unpublished data) and that 2) palmitate and stearate comprise ∼28 and 12%, respectively, of total NEFA, as measured in 54 TÜF participants (17), the concentrations used here are, at least in the case of palmitate, not too far from the physiological situation. As for stearate, it can be estimated that 1 mmol/l is about fivefold higher than its plasma concentration in healthy subjects. Even though respective data are lacking, it is conceivable that this stearate concentration can be reached in pathological situations associated with markedly increased lipolytic rates, such as morbid obesity and type 2 diabetes. The proapoptotic effects of saturated NEFAs on HCAECs is in agreement with earlier studies in HUVECs (5,6). In these cells, however, Artwohl et al. (5) additionally demonstrated that unsaturated NEFAs, at a concentration of 300 μmol/l, act proapoptotically as well. Moreover, these authors point out that the proapoptotic properties of unsaturated NEFAs are a function of the degree of desaturation with linoleate being comparably effective as stearate. In HCAECs, we were not able to detect any proapoptotic effects of unsaturated NEFAs (linoleate included), even at a concentration of 1 mmol/l. Moreover, we provide data demonstrating potent antilipoapoptotic effects of these NEFA species. As to these discrepancies, we speculate that endothelial cells from different sources, i.e., vein versus artery and umbilical versus coronary vessels, diverge in their apoptotic susceptibility toward unsaturated NEFAs. With regard to the importance of endothelial cell apoptosis in atherogenesis, plaque erosion, and acute coronary syndromes (7,8), our findings, which clearly demonstrate antilipoapoptotic properties of palmitoleate, oleate, and linoleate in HCAECs, could provide a mechanistic explanation for the previously observed cardioprotective effects of these non-ω3 unsaturated NEFAs (18–20).

Furthermore, we provide evidence that saturated NEFAs, after being activated to fatty acyl-CoA, induce HCAEC apoptosis via activation of IKK and NFκB. We show that this signaling pathway does neither involve conversion of NEFA to ceramide nor NEFA degradation via β-oxidation. In addition, we tend to exclude nitric oxide as a mediator of saturated NEFA-induced apoptosis since recent attempts to inhibit lipoapoptosis by blocking nitric oxide synthase isoforms using pharmacological inhibitors, i.e., N^G-monomethyl-l-arginine, N^G-nitro-l-arginine methyl ester, and N^G-(1-iminoethyl)-l-lysine, were not successful (data not shown). Further studies are, however, needed to substantiate this finding. It was recently demonstrated that saturated NEFAs are able to activate NFκB in several cell systems, such as murine and human skeletal muscle cells (14,21), bovine eye pericytes (22), bovine aortic endothelial cells (16), and HUVECs (23). Since NFκB activation by various physiological stimuli, e.g., IL-18 (24), high glucose (25), hypoxia (26), and TNF-α (27), is reported to trigger apoptosis in endothelial cells, the mechanism of lipoapoptosis described in this study appears plausible. Consistent with the antiapoptotic effect of unsaturated NEFAs observed in this study, unsaturated NEFAs are reported to inhibit NFκB activation in endothelial cells (28) and macrophages (29) and to prevent saturated NEFA-induced apoptosis in rat insulinoma and human islet cells (30). In addition, our findings point to NEFA-stimulated and NFκB-directed induction of proapoptotic genes. Preliminary data using the Apoptosis Oligo GEArray from SuperArray Bioscience support this suggestion: among 112 spotted genes, 7 were found regulated by stearate in HCAECs. Increased were, for example, the proapoptotic genes Bcl-2–like protein 13, lymphotoxin-β receptor, TWEAK (TNF-related weak inducer of apoptosis) receptor, and TRAF1 (TNF receptor–associated factor 1). Studies further evaluating the role of these genes in HCAEC lipoapoptosis are currently on the way. As to the molecular mechanisms upstream of IKK and NFκB that constitute the divergent behavior of saturated and unsaturated NEFA in HCAECs, we can at the moment only speculate. However, the very recently reported findings on the differential potential of these NEFA classes to provoke an endoplasmic reticulum stress response (31) or to stimulate formation of diacylglycerol (32,33), a well-known activator of classical and novel protein kinase C isoforms, represent good starting points for further investigations. Finally, we show that the effect of saturated NEFAs is not mediated by NFκB-dependent induction of proapoptotic cytokines, such as TNF-α and IL-1β; 1) IL-1β was not expressed by HCAECs, as was previously demonstrated (17,34), and 2) we detected only low amounts of secreted TNF-α protein, which were found to be ineffective in provoking apoptosis and, in addition, were not regulated by saturated NEFAs. Very recently, it was reported that saturated, but not unsaturated, NEFAs induce insulin receptor and insulin receptor substrate 1 expression in human aortic endothelial cells and confer insulin sensitivity to these cells (35). This finding could also imply enhanced survival factor signaling in these cells in response to saturated NEFAs. Whether such a protective mechanism is absent or counterregulated in HCAECs cannot be answered at the moment and awaits further studies.

In conclusion, we show here that saturated and unsaturated NEFAs display differential effects on HCAEC survival; saturated NEFAs, such as palmitate and stearate, are proapoptotic, whereas unsaturated NEFAs, such as palmitoleate, oleate, and linoleate, are antilipoapoptotic. Mechanistically, promotion of HCAEC apoptosis by saturated NEFAs requires fatty acyl-CoA formation and subsequent NFκB activation. As endothelial cell apoptosis is supposed to play an important role in plaque erosion and acute coronary syndromes, our findings could provide a theoretical explanation for the increased incidence of such life-threatening events in medical disorders associated with chronically elevated plasma NEFA concentrations.

This study was supported in part by a grant from the German Research Foundation (KFO 114/1-1) and the European Community’s FP6 EUGENE 2 (LSHM-CT-2004-512013).