Temporal Cross Talk Between Endoplasmic Reticulum and Mitochondria Regulates Oxidative Stress and Mediates Microparticle-Induced Endothelial Dysfunction

T-lymphocyte MPs induce endothelial dysfunction through ER stress–response activation

In response to treatment by MPs from apoptotic lymphocytes, the UPR pathway activation was assessed using the ER stress inhibitor tauroursodeoxycholic acid (TUDCA), a chemical chaperone known to enhance adaptive response of ER. MPs activated the three canonical pathways associated with ER stress response: PERK, IRE1α, and ATF6. Indeed, MPs resulted in an increased phosphorylation of PERK and eukaryotic translation initiation factor 2 α (eIF2α), as well as mRNA (not shown) and protein CHOP expressions (Fig. 1A–C). In addition, MP treatment enhanced X-box binding protein 1 (XBP1) splicing downstream to IRE1α (Fig. 1D) and induced the nuclear translocation of ATF6 (Fig. 1E). Importantly, all these MP-evoked effects were prevented in the presence of the ER stress inhibitor TUDCA (Fig. 1A–E).

OPEN IN VIEWER

FIG. 1.

T-lymphocyte MPs induce ER stress response. Effects of MPs on the three canonical arms of ER stress response. TUDCA was used to inhibit ER stress response. (A–C) Anti-phospho-PERK, anti-phospho-eIF2α, anti-CHOP, and anti-α actin antibody immunoblots of human aortic endothelial cells. (D) Semiquantitative polymerase chain reaction of XBP-1. (E) Immunofluorescent staining showing the nuclear translocation of ATF6. (F) NO^• production measured by EPR. (G) Endothelial-dependent vasorelaxation induced by Ach on mouse aortic rings. *p < 0.05, **p < 0.01. For statistical analyses, one-way analysis of variance and Mann–Whitney U test or analysis of variance for repeated measures and subsequent Bonferroni post hoc test for Ach-induced vasorelaxation were used. n = 3–4. Ach, acetylcholine; ATF6, activating transcription factor 6; eIF2α, eukaryotic translation initiation factor 2 α; EPR, electronic paramagnetic resonance; ER, endoplasmic reticulum; MP, microparticle; NO^•, nitric oxide; PERK, pancreatic endoplasmic reticulum kinase; TUDCA, tauroursodeoxycholate; Tun, tunicamycin; XBP1, x-box binding protein 1. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

It should be noted that the ER stress inducer tunicamycin was able to activate the same three classical ER stress pathways, which were partially prevented by TUDCA (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/ars). Altogether, these data indicate that MPs from apoptotic lymphocytes lead to the activation of all three arms of ER stress response.

To verify whether ER stress–response activation accounts for the endothelial dysfunction induced by T-lymphocyte MPs (25), in vitro NO^• production was evaluated in the absence or presence of TUDCA. As previously described, T-lymphocyte MPs significantly decreased NO^• production in human aortic endothelial cells (HAoECs) (Fig. 1F). Interestingly, the inhibition of ER stress with TUDCA abolished the ability of MPs to reduce NO^• production.

To establish the pathophysiological relevance of MPs, endothelium-dependent relaxation in response to acetylcholine was evaluated following MP incubation with mouse aortic rings in the absence or presence of TUDCA. MPs significantly reduced the maximal acetylcholine-evoked relaxation, which was improved by TUDCA (Fig. 1G). Finally, similarly to MPs, tunicamycin was able to reduce both in vitro NO^• release by endothelial cells and acetylcholine-induced relaxation of mouse aortic rings. These effects were partially prevented in the presence of TUDCA (Supplementary Fig. S1).

Collectively, these results indicate that the activation of ER stress response is implicated in endothelial dysfunction induced by T-lymphocyte MPs.

Involvement of membrane receptors in T-lymphocyte MP-induced ER stress in endothelial cells

We have previously shown that T lymphocyte MPs, through their Fas ligand (FasL), induce vascular smooth muscle cell inflammation by directly interacting with Fas of the recipient cells (31). Also, Yang et al. (33) have reported the involvement of low-density lipoprotein receptor (LDL-R) in the effects induced by T-lymphocyte MPs on human retinal endothelial cells. To determine the mechanism(s) by which MPs would induce ER stress in HAoECs, the implication of both Fas/FasL interaction and LDL-R was analyzed on MP-induced eIF2α phosphorylation. Interestingly, we observed that neutralizing either FasL on MPs or LDL-R on endothelial cells significantly reduced the MP-induced phospho-eIF2α (Fig. 2A). Altogether, these data indicate that both Fas/FasL interaction and LDL-R pathway are involved in MP-mediated ER stress induction.

OPEN IN VIEWER

FIG. 2.

ER stress response induced by T-lymphocyte MPs is mediated by neutral SMase activation via Fas/FasL interaction and LDL-R. (A–C) Effects of MPs on ER stress response (phospho-eIF2α) and SMase expression are inhibited by anti-FasL or LDL-R antibodies. (D–H) Pharmacological inhibition of SMase with GW4869 or silencing with siRNA abolished MP-induced ER stress response and endothelial dysfunction. *p < 0.05. For statistical analyses, one-way analysis of variance and Mann–Whitney U test were used. n = 3–4. FasL, Fas ligand; LDL-R, low-density lipoprotein receptor; SMase, sphingomyelinase.

Neutral sphingomyelinase is the link between MP-activated receptor and ER stress–response induction

Since sphingomyelinase (SMase) cascade has been implicated in the regulation of ER stress response (15), we assessed the involvement of SMase on MP-induced effects. MP treatment induced an increase of neutral SMase expression, which was prevented after neutralizing FasL on MPs (Fig. 2B) or when LDL-R on endothelial cells was blocked (Fig. 2C). Treatment of endothelial cells by MPs for 24 h, but not for 4 h, increased phospho-eIF2α. Interestingly, silencing neutral SMase with a specific siRNA abolished MP-induced increase of eIF2α phosphorylation after 24 h of MP treatment (Fig. 2D, E).

Moreover, using either the specific neutral SMase inhibitor GW4869 or siRNA against neutral SMase completely prevented the MP-induced reduction of eNOS activity as reflected by the phospho-Ser eNOS/phospho-Thr eNOS ratio (Fig. 2F, G). Furthermore, MP-induced NO^• reduction was abolished when neutral SMase was inhibited with GW4869 or silenced by siRNA (Fig. 2H). Altogether, these results suggest that the neutral SMase pathway controls the effects of MPs via Fas/FasL and LDL-R interaction upstream of ER stress, eNOS activation, and NO^• production.

Cytosolic ROS production is upstream of T-lymphocyte MP-induced ER stress in endothelial cells

The possible involvement of oxidative stress in T-lymphocyte MP-induced ER stress activation was then assessed by incubating HAoECs with various inhibitors, the sources of ROS before MP stimulation. Indeed, intracellular enzymatic systems with the capacity to generate ROS were inhibited by allopurinol, apocynin, rotenone, or Nω-nitro-l-arginine (l-NA), which are inhibitors of xanthine oxidase, NADPH oxidase, respiratory chain complex I, and NOS, respectively. At 24 h of treatment, none of the ROS inhibitors modified the phosphorylation of eIF2α in the absence of MPs (Fig. 3A). By contrast, all the inhibitors significantly decreased eIF2α phosphorylation induced by MP treatment. These results suggest that MP-induced ROS production from different sources is required for the activation of ER stress response (Fig. 3B).

OPEN IN VIEWER

FIG. 3.

Cytosolic production of ROS by MPs is upstream of ER stress response. (A, B) Effects of 24-h treatment with inhibitors of ROS sources on ER stress response (phospho-eIF2α) in the absence (A) or presence of MPs (B). (C–I) Cytosolic ROS production measured by EPR (C–E) and DHE labeling (F–I). (J) Effects of inhibitors of ROS sources on ROS production after 4 h of MP treatment. (K, L) Effects of silencing SMase on MP-induced p47phox expression. *p < 0.05, **p < 0.01, ***p < 0.001. For statistical analyses, one-way analysis of variance and Mann–Whitney U test were used. n = 3–7. DHE, dihydroethidium; ROS, reactive oxygen species. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

To establish the implication of cytoplasmic and mitochondrial ROS in the effects of MPs, time course (i.e., 2, 4, and 24 h) changes of ROS levels were measured. Analysis of ROS levels, using electronic paramagnetic resonance (EPR) and dihydroethidium (DHE) labeling, showed that T-lymphocyte MPs had no effect on ROS production at 2 h, increased at 4 h, and returned to basal levels at 24 h (Fig. 3C–I). This effect of T-lymphocyte MPs on ROS production was not affected by TUDCA. It is important to note that tunicamycin treatment induced similar patterns of ROS production as those obtained with MPs (Supplementary Fig. S2).

Interestingly, except for the mitochondrial ROS scavenger (mito-TEMPO), all of the ROS inhibitors tested (allopurinol, apocynin, rotenone, and l-NA), reduced the early increase of cytosolic ROS production (Fig. 3J). To further analyze the enzyme involved in this ROS production, we found that lymphocyte MPs increased the expression of p47phox, a regulatory subunit of NAPDH oxidase at 4 h, but not at 24 h of treatment (Fig. 3K). This increase was partially prevented after silencing of neutral SMase (Fig. 3K).

Altogether, lymphocyte MPs activate neutral SMase and induce an early increase of cytosolic ROS production, upstream of ER stress activation via at least the NADPH oxidase activation.

Cross talk between ER and mitochondria is involved in the effects induced by lymphocyte MPs in endothelial cells

To monitor mitochondrial ROS production, the MitoSOX probe was used. Quantification of confocal images showed that mitochondrial ROS fluorescence was reduced at 2 and 4 h, but was greatly enhanced after 24 h of lymphocyte MP treatment (Fig. 4A–D). Interestingly, the ER stress inhibitor TUDCA did not modify the reduction of MitoSOX fluorescence at 2 and 4 h but completely abolished the increase at 24 h. These results suggest that lymphocyte MPs induce a late increase of mitochondrial ROS by a mechanism downstream of ER stress activation. To confirm this, the mitochondrial ROS scavenger mito-TEMPO was used and did not show any effect on the increase of phospho-eIF2α induced by lymphocyte MPs (Fig. 4E).

OPEN IN VIEWER

FIG. 4.

Mitochondrial production of ROS by MPs is downstream of ER stress response. (A–D) Mitochondrial ROS production measured by MitoSOX labeling. (E) Scavenging mitochondrial ROS with MitoTEMPO (Mito-T) has no effect on ER stress response. (F–I) Effects of MPs on mitochondrial respiration and complex IV activity. (J, K) Effects of MPs on mitochondria-associated membranes. (L) Mito-T improves NO^• production induced by MPs. *p < 0.05, **p < 0.01. For statistical analyses, one-way analysis of variance and Mann–Whitney U test were used. n = 4–6. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

The effects of MPs on mitochondrial O₂ consumption were then analyzed to decipher the role of mitochondria on the MP-induced endothelial dysfunction. Basal O₂ consumption was not modified either by MPs or MPs + TUDCA (Fig. 4F). However, MP treatment significantly reduced both oligomycin and carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) mitochondrial respirations, which were abolished when ER stress was inhibited with TUDCA (Fig. 4G, H). These findings suggest that MPs are able to modify mitochondrial respiration by a mechanism implicating ER stress. Exposure to MPs resulted in an increase of complex IV activity (Fig. 4I) but had no effect on the activity of the other complexes of the electron transport chain (not shown).

To confirm the interaction between ER and mitochondria through the formation of tight physical junctions, MAM integrity was analyzed following MP treatment. As shown in Figure 4J and K, MPs induced MAM disruption characterized by an overexpression of mitofusin 2 (Mfn2) and a decreased expression of voltage-dependent anion channel 1 (VDAC1). Moreover, these effects were prevented in the presence of the ER stress inhibitor TUDCA, suggesting that MP-induced ER stress leads to MAM disruption on endothelial cells. Finally, to verify whether the MP effects on mitochondria lead to endothelial dysfunction, NO^• production was measured in the presence of mito-TEMPO. Scavenging mitochondrial ROS with mito-TEMPO abolished the decrease in NO^• release induced by MPs (Fig. 4L). This result suggests that alteration of mitochondrial function with respect to respiration and ROS production, as a consequence of ER stress, leads to endothelial dysfunction.

MPs from metabolic syndrome patients induce endothelial dysfunction via Fas/FasL pathway and neutral SMase activation leading to ER stress activation in vitro and ex vivo

We have previously described that MPs from metabolic syndrome patients, but not those of nonmetabolic syndrome subjects, induced endothelial dysfunction (2). In this study, we tested whether molecular pathways involved in the effects of MPs from metabolic syndrome required also the cross talk between ER and mitochondria and oxidative stress similarly to lymphocyte MPs. Therefore, MPs from metabolic syndrome patients or nonmetabolic syndrome subjects were isolated, for whom detailed clinical information is described in Supplementary Table S1. As expected, metabolic syndrome patients showed greater visceral obesity as given by waist circumference, enhanced triglyceridemia, and increased blood pressure.

The total number of circulating MPs was significantly increased in patients with metabolic syndrome compared to nonmetabolic syndrome subjects (Supplementary Table S2). Phenotypical characterization of cellular origin of MPs showed an increase of ∼6-fold (annexin V⁺), ∼3.3-fold (CD41⁺), and ∼9.6-fold (CD146⁺) of the circulating level of procoagulant-, platelet-, and endothelial-derived MPs in metabolic syndrome patients compared to nonmetabolic syndrome subjects, respectively. Also, levels of erythrocyte- and leukocyte-derived MPs (CD235a⁺, CD45⁺) were elevated (∼9- and ∼2.7-fold, respectively) in these patients.

As shown in Figure 5A and B, MPs from metabolic syndrome patients, but not from nonmetabolic syndrome individuals, induced an increase in phosphorylation of eIF2α and nuclear translocation of ATF6, which were prevented in the presence of TUDCA. Similarly and as expected, MPs from metabolic patients, in comparison to nonmetabolic syndrome subjects, caused a reduction in NO^• release from endothelial cells, which was abolished when ER stress was inhibited by TUDCA (Fig. 5C). Moreover, incubation of mouse aortic rings with MPs from metabolic syndrome patients, but not from nonmetabolic syndrome subjects, was associated to a reduced endothelium-dependent relaxation in response to acetylcholine (Fig. 5D). This effect was prevented when vessels were incubated in the presence of TUDCA.

OPEN IN VIEWER

FIG. 5.

MPs from MetS (MetS MPs) induce ER stress response via Fas/FasL interaction and neutral SMase pathway. Effects of MPs from nMetS (nMetS MPs) and MetS MPs on the canonical arms of ER stress response. TUDCA was used to inhibit ER stress response. (A) Anti-phospho-eIF2α antibody immunoblots of human aortic endothelial cells. (B) Immunofluorescent staining showing the nuclear translocation of ATF6. (C) NO^• production measured by EPR. (D) Endothelial-dependent vasorelaxation induced by Ach on mouse aortic rings. (E, F) Effects of MPs on ER stress response (phospho-eIF2α) and SMase expression are inhibited by anti-FasL antibodies. (G, H) Silencing SMase with siRNA abolished MP-induced ER stress response. *p < 0.05, **p < 0.01. For statistical analyses, one-way analysis of variance and Mann–Whitney U test or analysis of variance for repeated measures and subsequent Bonferroni post hoc test for Ach-induced vasorelaxation were used. n = 3–4. MetS, metabolic syndrome patients; nMetS, nonmetabolic syndrome subjects. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Taken together, these results indicate that ER stress plays a key role in the induction of endothelial dysfunction induced by MPs from metabolic syndrome patients.

To establish the receptor involved in the effects of MPs from metabolic syndrome patients on endothelial cells, we neutralized either FasL on MPs or LDL-R on HAoECs. Interestingly, neutralization of FasL but not LDL-R on HAoECs, prevented the phosphorylation of eIF2α by metabolic syndrome MPs (Fig. 5E) and abolished their ability to enhance the expression of SMase (Fig. 5F). Finally, it is important to note that silencing neutral SMase with siRNA significantly abolished the increase of eIF2α phosphorylation by metabolic syndrome MPs both at early (4 h) and late stage (24 h) of treatment (Fig. 5G, H), in contrast to T-lymphocyte MPs that were able to increase eIF2α at 24 h only (Fig. 2E).

Cytosolic and mitochondrial ROS productions are downstream of metabolic syndrome MP-induced ER stress in endothelial cells

As illustrated in Figure 6A–H, MPs from metabolic syndrome induced oxidative stress by increasing both cytosolic and mitochondrial ROS. DHE staining showed that MPs from metabolic syndrome patients, in comparison to nonmetabolic syndrome subjects, enhanced cytosolic ROS at 2 and 4 h, but not at 24 h (Fig. 6A–D). Also, MPs from metabolic syndrome patients but not those from nonmetabolic syndrome subjects increased MitoSOX fluorescence at 24 h, but not at 2 or 4 h of treatment of HAoECs (Fig. 6E–H). Cytosolic ROS, but not mitochondrial ROS production, was abolished in the presence of ER stress inhibitor TUDCA. Interestingly, MPs induced p47phox expression at 4 h, but not at 24 h, which was abolished after SMase silencing (Fig. 7A, B). These results suggest that MPs from metabolic syndrome patients activate neutral SMase and induce a spatiotemporal increase of cytosolic ROS at early stage via the NADPH oxidase that concur to ER stress and mitochondrial ROS at late stages of stimulation. All of these effects participate in metabolic syndrome MP-induced endothelial dysfunction.

OPEN IN VIEWER

FIG. 6.

Cytosolic and mitochondrial production of ROS by MetS MPs is downstream of ER stress response. (A–D) Cytosolic ROS production measured by DHE staining. (E–H) Mitochondrial ROS production measured by MitoSOX labeling. For statistical analyses, one-way analysis of variance and Mann–Whitney U test were used. n = 4. *p < 0.05. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

OPEN IN VIEWER

FIG. 7.

Neutral SMase controls p47phox expression. siRNA against neutral SMase prevents p47phox expression induced by MetS MPs at 4 h (A), but not at 24 h (B). *p < 0.05. For statistical analyses, one-way analysis of variance and Mann–Whitney U test were used. n = 4.

MP production from T cells

The human lymphoid CEM T-cell line was used for MP production. Cells were seeded at 10⁶ cells/ml and cultured in serum-free X-VIVO 15 medium (Lonza). MPs were produced as previously described (20, 25). Briefly, cells were treated with actinomycin D (1 μg/ml; Sigma-Aldrich) for 24 h. The supernatant was obtained by centrifugation at 750 g for 15 min and then at 1500 g for 5 min to remove cells and large debris, respectively. MPs from the supernatant were washed after three centrifugation steps (45 min, 14,000 g) and recovered in 400 μl of sterile NaCl 0.9%. Last, a washing medium was used as vehicle. The determination of the amount of MPs was conducted by measuring MP-associated proteins using the DC Protein Assay (Bio-Rad). MPs were used at 10 μg/ml corresponding to 1330 ± 312 × 10³ MPs. This concentration of lymphocyte MPs is found in plasma from patients undergoing carotid endarterectomy (0–2 × 10⁶ MPs/ml plasma) (16).

Metabolic syndrome patients

This study was approved by the Ethics Committee of the University Hospital of Angers (France; NCT: 00997165). A total of 30 patients were included with metabolic syndrome from the METABOL cohort at the Department of Endocrinology and Nutrition of the University Hospital of Angers. Patients were eligible for inclusion, according to the National Cholesterol Education Program-Adult Treatment Panel III, when they had at least three of the following five criteria: (i) waist circumference >102 or 88 cm for men and women, respectively; (ii) systolic and diastolic pressures ≥130/85 mmHg; (iii) fasting glycemia ≥1.1 g/l; (iv) triglycerides ≥1.5 g/l; and (v) high-density cholesterol lipoprotein <0.4 g/l in men or <0.5 g/l in women (Supplementary Table S1). Patients with a history of cardiovascular diseases, preexistent chronic inflammatory disease, and cancer were excluded. Normal controls consisted of 20 subjects who met less than two of the metabolic syndrome criteria (70% without any component of metabolic syndrome).

MP collection from patients

Peripheral blood (20 ml) from nonmetabolic syndrome or metabolic syndrome subjects was collected in ethylenediaminetetraacetic acid tubes (Vacutainers; Becton Dickinson) from a peripheral vein using a 21-gauge needle to minimize platelet activation and was processed within 2 h. After a 20-min centrifugation at 270 g, platelet-rich plasma was separated from whole blood, which was then centrifuged for 20 min at 1500 g to obtain platelet-free plasma (PFP). Two hundred microliters of PFP was frozen and stored at −80°C until subsequent use for MP characterization by flow cytometry. Remaining PFP was subjected to three series of centrifugations each at 21,000 g for 45 min to pellet MPs for in vitro studies, and the supernatant was replaced by 200 μl of sterile 0.9% NaCl and stored at 4°C until subsequent use. MPs from nonmetabolic syndrome or metabolic syndrome subjects were used at the circulating concentration detected for each individual, as previously described (1, 2).

Characterization of MP phenotype

Membrane MP subpopulations were discriminated in PFP according to the expression of membrane-specific antigens (1, 2). Irrelevant human immunoglobulin G (IgG) was used as an isotype-matched negative control for each sample. For numeration studies, 10 μl of PFP was incubated with 5 μl of specific antibody (Beckman Coulter). Annexin V binding was used to numerate phosphatidylserine-expressing circulating MPs (2 μl of annexin V/5 μl PFP). After 30 min at room temperature, samples were diluted in 300 μl of sterile 0.9% NaCl or annexin-V labeling buffer, respectively. Then, an equal volume of sample and Flow-Count beads was added and samples were analyzed in a flow cytometer 500 MPL system (Beckman Coulter).

Cell culture

HAoECs were maintained in culture in an endothelial cell growth medium MV2 (PromoCell) supplemented with 1% streptomycin/penicillin (Lonza). Cells were treated for 24 h with ER stress inducer tunicamycin (0.8 μg/ml; Sigma-Aldrich), 10 μg/ml of T-lymphocyte MPs, or those from nonmetabolic syndrome or metabolic syndrome subjects, in the absence or presence of the ER stress inhibitor, TUDCA (100 μmol/l; Sigma-Aldrich). In another set of experiments, 30 min before MP treatment, cells were pretreated with the NADPH oxidase inhibitor, apocynin (Apo, 100 μmol/l; Calbiochem), the inhibitor of xanthine oxidase, allopurinol (Allo, 50 μmol/l; Sigma-Aldrich), the inhibitor of mitochondrial complex I NADH dehydrogenase, rotenone (Rot, 5 μmol/l; Sigma-Aldrich), the NO^• synthase inhibitor, l-NA (100 μmol/l; Sigma-Aldrich), the mitochondria-targeted scavenger mito-TEMPO (25 nmol/l; Santa Cruz Biotechnology), or the inhibitor of neutral SMase, GW4869 (10 μmol/l; Sigma-Aldrich). All products were used at concentrations at which no cytotoxicity was observed, as deduced from trypan blue exclusion assay.

Western blot

After treatment, cells were homogenized and lysed. Proteins (40 μg) were separated on 4–12% NuPAGE gels (ThermoFisher Scientific). Blots were probed with antibodies against PERK, phospho-PERK, eIF2α, phospho-eIF2α, CHOP, phospho-Thr495 eNOS (Cell Signaling), Mfn2, VDAC1α (Abcam), p47-phox, and phospho-Ser1177 eNOS (BD Biosciences). A polyclonal mouse anti-human β-actin antibody (Sigma-Aldrich) was used for standardization of protein gel loading.

In another set of experiments, cells were pretreated with the LDL-R antibody (5 μg/ml; R&D Systems) or its control IgG antibody for 30 min before MP treatment. Also, FasL carried by MPs was preincubated with human anti-FasL (5 μg; BD Biosciences) for 30 min at 4°C to allow neutralization of MP FasL, after being washed two times with NaCl 0.9% to remove unbound anti-FasL antibody.

Nuclear translocation of ATF6 by confocal microscopy

Cells were seeded on glass slides (Ibidi) for 24 h and then treated with either tunicamycin or MPs for 24 h in the absence or presence of TUDCA. Cells were then washed and fixed with 4% paraformaldehyde. Then, cells were treated with 5% bovine serum albumin, permeabilized by 0.1% Triton, incubated with anti-ATF6 (Abcam) for 1 h, followed by a 1-h incubation with the Alexa fluor-546 secondary antibody (ThermoFisher Scientific) at room temperature. Finally, 4′,6-diamidino-2-phenylindole (DAPI, 300 mmol/l; Santa Cruz Biotechnology) was added for 5 min. Cells were washed and fluorescence was measured with a confocal microscopy (CLMS 700; Zeiss, ZEN software). All images were acquired using 40× objective.

Quantitative and semiquantitative polymerase chain reaction

Cells were treated for 6 h as described above, and then, frozen cell pellets were used to investigate the expression of mRNA for CHOP by real time (RT)-polymerase chain reaction (RT-PCR; human forward primer GAACGGCTCAAGCAGGAAAT, human reverse primer TTCACCATTCGGTCAATCAGAG). Total RNA was isolated using the miRNeasy Qiagen MicroKit. cDNA was generated using the PrimeScript TM RT Reagent Kit (Takara/Clontech) from 500 ng of RNA with random hexameric primers. RT-PCR analyses were performed using a CFX96™ Real-Time PCR Detection System and SYBR Green detection (Bio-Rad). Semiquantitative PCR was made for the expression of XBP-1 (human forward primer AAACAGAGTAGCAGCTCAGACTGC, human reverse primer TCCTTCTGGGTAGACCTCTGGGAG) by using PTC-200 (MJ Research). β-Tubulin was used as the housekeeping reference gene.

siRNA transfection

Transient transfection of HAoECs was done according to the manufacturer's protocol (Santa Cruz Biotechnology). Briefly, HAoECs were treated with either control siRNA or neutral SMase siRNA for 6 h in a serum- and antibiotic-free medium. Cells were then washed and normal fresh medium was added for 24 h before treatment with MPs for 24 h.

NO^• measurement

After treatment for 24 h, as described above, cells were loaded with 4, 5-diaminofluorescein diacetate (10 μmol/l; Santa Cruz Biotechnology) for 30 min at 37°C, and fluorescence was measured.

Endothelial function

Aortic rings were obtained from male Swiss mice bred at the animal facility at the University of Angers. After 24 h of MP incubation, aortic rings were mounted on a wire myograph filled with physiological salt solution, as previously described (20). Endothelium-dependent vasodilatation was studied by cumulative application of acetylcholine (1 nmol/l–10 μmol/l; Sigma-Aldrich) in aortas with functional endothelium precontracted with U46619 (Sigma-Aldrich).

Anion superoxide determination by EPR studies

The detection of anion superoxide production was performed by the EPR technique using 1-hydroxy-3-methylcarbonyl-2, 2, 5, 5 tetramethyl pyrrolidine (CMH; Noxygen) as spin trap. Briefly, after 24 h of treatment, the medium was replaced with 500 μl of deferoxamine-chelated Krebs–Hepes solution containing CMH (500 μmol/l), deferoxamine (25 μmol/l; Sigma-Aldrich), and sodium diethyldithiocarbamate (5 μmol/l; Sigma-Aldrich), and incubated for 45 min at 37°C. ROS measurement was performed on a tabletop x-band spectrometer Miniscope (MS200; Magnettech). Recordings were made at 77°K, using a Dewar flask. Instrument settings were 10 mW of microwave power, 1 mT of amplitude modulation, 100 kHz of modulation frequency, 150 s of sweep time, and three scans. The quantitative measurement of the O₂ ^•− signal amplitude was reported to the relative units for protein concentration (amplitude/μg/μl).

ROS measurement by DHE or MitoSOX staining

Endothelial cells were grown on glass slides (Ibidi), and after different treatments, they were washed and then incubated with either the oxidative fluorescent dye DHE (5 μmol/l; Sigma-Aldrich) for 30 min at 37°C or MitoSOX red (5 μmol/l; ThermoFisher Scientific) for 10 min at 37°C. Then, cells were fixed by 4% paraformaldehyde for 20 min at room temperature. DAPI was added for 5 min. Finally, cells were visualized with confocal microscopy (CLMS 700; Zeiss, ZEN software). All images were acquired using a 63× objective.

Measurement of whole-cell respiration

Respiration rates were measured in treated HAoECs. The cells were collected by trypsinization, washed once in culture medium, and centrifuged (500 g, 5 min). The pellet was resuspended in a respiratory medium, corresponding to normal culture medium without fetal bovine serum and antibiotics, both reagents known to interfere with mitochondrial respiration. Mitochondrial oxygen consumption was measured at 37°C using a high-resolution Oxygraph-2K respirometer (OROBOROS). Oxygraphy experiments were performed on cells cultured in T75 flasks, corresponding to an average of 5 × 10⁶ cells per flask. The basal respiration rate of cells was determined by measuring the linear rate of oxygen consumption. Oligomycin (1 μg/ml) was then added to determine the nonphosphorylating respiration rate. The uncoupling respiration rate was also recorded by stepwise addition of FCCP (0.2–1 μmol/l) up to the optimal concentration representing the maximal capacity of the respiratory chain without the regulation of ATP synthase. Finally, the nonmitochondrial oxygen consumption rate was determined by adding antimycin A (1 μg/ml). At the end of the experiment, a standard volume of respiratory medium is collected and centrifuged for 2 min at 16,000 g. The cell pellet was then resuspended in NaCl (0.9%) and protein concentration was determined spectrophotometrically. Respiration rates were expressed in nmoles of O₂ consumed per minute and per mg protein.

Mitochondrial enzyme activity measurements

The activity of the different mitochondrial respiratory chain complexes was measured on cell homogenates, at 37°C, using an SAFAS spectrophotometer (SAFAS).

Statistics

Data are expressed as mean ± standard error of the mean. Statistical analyses were performed by one-way analysis of variance (ANOVA), then the Mann–Whitney U test or ANOVA for repeated measures, and subsequently, the Bonferroni post hoc test. p < 0.05 was considered to be statistically significant.