Cell culture

Human coronary artery endothelial cells (HCAECs, #CC-2585, Lonza) were cultured in EBM-2 basal medium (#CC-3156, Lonza) supplemented with EGM-2 MV microvascular endothelium cell growth medium (#CC-3202, Lonza). Cells were maintained at 37°C in a 5% CO₂ humidified incubator and used between passages 3 and 6. To simulate febrile conditions, cells were cultured under static conditions at either 37°C or 40°C, depending on the experiment.

Treatments

To investigate the molecular mechanisms underlying the response to hyperthermia, HCAECs were treated with the following compounds: B-I09 (0.2–20 µ M, #6009, Tocris), an IRE1α inhibitor, to block XBP1 splicing and assess the role of the unfolded protein response (UPR); Chloroquine (100 µ M, Sigma, C6628), was used to inhibit autophagy by blocking lysosomal function. Staurosporine (1 µ M, #1285, Bio-Techne), was used as a pro-apoptotic-inducing agent (positive control).

RFP-GFP-LC3B transfection

100,000 HCAECs were grown until 70µl confluency in µ -Slide 8 Well high Glass Bottom (Ibidi), and transfected with 10µl of Premo autophagy tandem sensor RFP-GFP-LC3B mix (Molecular Probes, P36239) diluted in 300 µl of culture medium per well. Transfected cell were grown overnight in normal conditions and then cultured at 37°C or 40°C for 3 or 24h. After nuclei stain with Hoechst (1:1000), and mounting with 200µl ProLong Gold antifade reagent, both RFP and GFP positive puncta were quickly visualized by using an AxioObserver microscope (Zeiss, x63) and images were processed and quantified with Fiji (ImageJ) software for quantification.

RNA isolation and microarray analysis

For transcriptomic analysis, HCAECs were seeded at a density of 5 × 10⁵ cells per well in 6-well plates and cultured in complete medium for 6, 12, or 24 hours at 37°C or 40°C. Total RNA was extracted using the miRNEasy mini kit (#217004, Qiagen) according to the manufacturer’s instructions. RNA quality was assessed using the Agilent 2100 Bioanalyzer, and microarray analysis was performed using the SurePrint G3 Human Gene Expression v3 8x60K Microarray Kit (Agilent). Differential gene expression was determined by comparing the 40°C and 37°C conditions (S1 Table), and significant changes were defined as a fold change of ≥±2 with a p-value < 0.05, using Partek Flow (Illumina). Pathway analysis was performed using (Ingenuity Pathway Analysis, Qiagen).

qRT-PCR

Total RNA was reverse-transcribed using the SuperScript VILO cDNA Synthesis Kit (#11754050, Invitrogen). qRT-PCR was carried out with TaqMan Universal PCR Master Mix (#4369016, Invitrogen) on a Stratagene Mx3000P detection system (Agilent). Primer sets used in our study are listed in the S2 Table. Relative gene expression was calculated using the 2^-ΔΔCt method.

Western Blot

Cells were lysed in ice-cold cell extraction buffer (#FNN0011, Invitrogen) supplemented with protease and phosphatase inhibitors (#P8340 and #P0044 Sigma). Protein concentration was determined using the Pierce BCA Protein Assay Kit (#23225, Thermo Fisher). Proteins (1–10 µg per sample) were separated by SDS-PAGE and transferred to nitrocellulose or PVDF membranes for subsequent immunoblotting. Primary antibodies were used against sXBP1, LC3B, p62, VE-Cadherin and phospho-VE-Cadherin (S3 Table). Detection was performed using HRP-conjugated appropriate secondary antibodies and chemiluminescence substrate (GE Healthcare).

Immunofluorescence

After treatments, cells were fixed and permeabilized with Cytofix/Cytoperm solution (#554714, Becton Dickinson). Immunolabeling was performed using primary antibodies against LC3B, LAMP2, VE-Cadherin, and cleaved caspase-3 (S3 Table). After incubation with secondary fluorescent antibodies (Thermo Fisher), nuclei were stained with Hoechst (1:1000), and slides were mounted with ProLong Gold antifade reagent. Images were captured using an AxioObserver microscope (Zeiss) and processed with Fiji (ImageJ) software for quantification.

Electron microscopy

Cells were fixed 1h in a mix of 1% glutaraldehyde, 2,5% paraformaldehyde in PBS, then washed 3 x 5 minutes in PBS. Samples were then post-fixed during 1h in reduce Osmium tetroxide 1% Potassium Ferrocyanide 1,5%, then washed 3 x 5 minutes in water. Dehydration was performed using graded series of ethanol in water for 10 minutes 30%, 50%, 70%, 80%, 90% (x2), 100% (x3). Resin infiltration was performed by incubating 1h in an Agar low viscosity resin (Agar Scientific Ltd) and ethanol mix (1:1), followed by overnight incubation in pure resin at 4°C. The resin was then changed and the samples further incubated during 3 hours prior to flat inclusion in gelatin capsules and overnight polymerization at 60°C. 70 nm sections were obtained using an EM-UC6 ultramicrotome (Leica), post-stained in 4% aqueous uranyl acetate and lead citrate, and observed at 120 kV with a Tecnai12 transmission electron microscope (ThermoFisher Scientific) equipped with a 4K × 4K Oneview camera (Gatan).

Mass spectrometry analysis

MS grade Acetonitrile (ACN), MS grade H2O and MS grade formic acid (FA) were from ThermoFisher Scientific (Waltham, MA, USA). Sequencing grade trypsin was from Promega (Madison, WI, USA). Dithiothreitol (DTT) and iodoacetamide (IAA) were from Sigma-Aldrich (Saint-Louis, MO, USA). Six times volume of cooled acetone (−20°C) were added to 10 µg of protein sample lysates. Vortexed tubes were incubated overnight at −20°C then centrifuged for 10 min at 11,000 rpm, 4°C. The protein pellets were dissolved in urea 8M - NH4HCO3 25mM buffer. Samples were then reduced with DTT 10 mM and alkylated with IAA 20 mM before a 16-fold dilution in NH4HCO3 25mM buffer. Digestion was performed overnight at 37°C by sequencing grade trypsin (enzyme to sample ratio of 1/20). The peptides were loaded and desalted on Evotips, according to manufacturer’s procedure, prior to LC-MS/MS analysis. Samples were analyzed on an Orbitrap Fusion mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA) coupled with an Evosep one system (Evosep, Odense, Denmark) operating with the 30SPD method developed by the manufacturer. Briefly, the method is based on a 44-min gradient and a total cycle time of 48 min with a C18 analytical column (0.15 x 150 mm, 1.9µm beads, ref EV-1106) equilibrated at room temperature and operated at a flow rate of 500 nl/min. H20/0.1% FA was used as solvent A and ACN/ 0.1% FA as solvent B. The mass spectrometer was operated by data-dependent MS/MS mode. Peptide masses were analyzed in the Orbitrap cell in full ion scan mode, at a resolution of 120,000, a mass range of m/z 350–1550 and an AGC target of 4.105. MS/MS were performed in the top speed 3s mode. Peptides were selected for fragmentation by Higher-energy C-trap Dissociation (HCD) with a Normalized Collisional Energy of 27% and a dynamic exclusion of 60 seconds. Fragment masses were measured in an Ion trap in the rapid mode, with and an AGC target of 1.104. Monocharged peptides and unassigned charge states were excluded from the MS/MS acquisition. The maximum ion accumulation times were set to 100 ms for MS and 35 ms for MS/MS acquisitions respectively. Label Free quantitation was performed using Progenesis QI for proteomics software version 4.2 (Waters, Milford, MA, USA). The software was allowed to automatically align data to a common reference chromatogram to minimize missing values. Then, the default peak-picking settings were used to detect features in the raw MS files and a most suitable reference was chosen by the software for normalization of data following the normalization to all protein method. A between-subject experiment design was chosen to create groups of biological replicates. MS/MS spectra were exported and searched for protein identification using PEAKS STUDIO Xpro software (Bioinformatics Solutions Inc., Waterloo, ON, Canada). The following parameters were used: trypsin as enzyme (specific), carbamidomethylation (C) as fixed and deamidation (NQ)/oxidation (M) as variable modifications. Precursor and fragment mass tolerances were set to respectively 15 ppm and 0.5 Da. Data were searched against the Human SwissProt database (release 2020 06, 20395 entries). The maximum of missed cleavages was set to 2. Spectra were filtered using a 1% FDR at peptide level. Identification results were then imported into Progenesis to convert peptide-based data to protein expression data using the Hi-3 based protein quantification method.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 9.0. Data are presented as mean ± standard error of the mean (SEM), and comparisons between two groups were made using the Mann-Whitney U test (S4 Table). A p-value < 0.05 was considered statistically significant. N represents the number of different experiments performed on different days. For mass spectrometry, Log2 transformed data were finally used for statistical analysis, i.e., evaluation of differentially present proteins between two groups using a Student’s bilateral t-test. A p-value better than 0.05 was used to filter differential candidates.

Prolonged febrile range hyperthermia impairs HCAEC survival adherence and function

HCAECs cultured at 40°C exhibited significant structural and functional alterations compared to cells maintained at 37°C. As early as 6 hours, transmission electron microscopy revealed the formation of large lysosomal structures (Fig 1A). After 12 hours, cells began to lose their junctions, and after 24 hours, there was a marked detachment of cells, along with the formation of apoptotic bodies and of microparticles (Fig 1A). Quantification of VE-Cadherin staining confirmed progressive junctional disruption over time at 40°C (Fig 1B). A significant increase in apoptosis was observed in flow cytometry, with Annexin V-positive cells rising by 30% at 12 hours and by 20% at 24 hours in the 40°C condition (Fig 1C). Staurosporine was used as a positive control in these experiments and the impact of temperature was 55 and 40% of the maximal value obtained with staurosporine at 12 and 24h, respectively. Additionally, cleaved-caspase 3 staining indicated increased apoptotic signaling in cells lacking continuous junctions (Fig 1D). Supernatant analysis confirmed a steady rise in both detached cells and apoptotic bodies or microparticles after 24 hours of culture at 40°C (Fig 1E).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Febrile range hyperthermia impairs EC survival, adherence and function in vitro.

A. Ultrastructural analysis (TEM, down) of HCAECs subjected to hyperthermia for 6, 12 or 24h. B. Cell-cell junction shown by VE-Cadherin staining over time in HCAECs subjected or not to FRH (left) and semi-quantitated (right). C. Apoptosis levels evaluated by flow cytometry using Pi Annexin V staining in HCAECs subjected or not to FRH for 6, 12 and 24h. D. VE-cadherin and cleaved caspase-3 immunostaining in HCAECs subjected to FRH for 12h. Corresponding semi quantitative analysis for cleaved caspase-3 signal (MFI) in HCAECs subjected or not to FRH. Arrowheads show cell junction perturbations. E. Number of detached cells (left) and microvesicles or apoptotic bodies (MVs, ABs, right) in the media of HCAECs subjected to FRH for 3, 6, 12 and 24h, evaluated by flow cytometry. Each observation is made of n = 3 independent experiments. * p < 0.05; **p < 0.01; ***p < 0.001.

https://doi.org/10.1371/journal.pone.0315119.g001

Transcriptomic and proteomic analyses reveals molecular mechanisms underlying the response to febrile-range hyperthermia

To dissect the molecular basis of these phenotypic changes, we conducted transcriptomic and proteomic profiling of HCAECs exposed to 40°C. Microarray analysis revealed that 243, 607, and 532 mRNA probes were downregulated at 6, 12, and 24 hours, respectively, while 378, 267, and 286 probes were upregulated (Fig 2A and B and S1 Table). Principal component analysis (PCA) indicated a clear segregation of gene expression patterns between the 37°C and 40°C conditions, with significant clustering based on temperature (Fig 2C). The most significantly upregulated pathways included the unfolded protein response (UPR) and endoplasmic reticulum (ER) stress, with IRE1α and XBP1 identified as key early upstream regulators (Fig 2D and E, and S1 Fig). The effect of FRH was also investigated at the proteomic level, using an ion trap mass spectrometry analysis (Orbitrap). Proteomic analysis indicated an enrichment of proteins involved in selective autophagy, adherens junctions, and ER-related processes (Fig 2F and G). Differentially expressed proteins (DEPs) identified through an ANOVA analysis are summarized in a volcano plot comparing cells cultured at 37°C and 40°C (Fig 2F and S4 Table, Fold change >1.5). The identified DEPs (10 downregulated, 24 upregulated) allowed a biological process enrichment analysis of most significant modules of the protein-protein interaction (PPI) network (CluegGO, Cytoscape) and showed that early response to fever-range hyperthermia upon proteomic profiling involved selective autophagy (52%), along with adherens junctions (4%), interconversion of nucleotide di-and triphosphate (8%), ADP metabolic process (8%), eukaryotic 48S preinitiation complex (12%), and snRNP assembly (16%) (Fig 2G).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Unbiased transcriptomic and proteomic analysis unravels molecular mechanisms involved in EC response to febrile range hyperthermia.

A. Volcano plots illustrating the differentially expressed genes (DEGs) over time (40°C vs 37°C). The most significant DEGs are highlighted. B. Venn diagram showing the relationship between regulated targets over time (40°C vs 37°C). C. Principal component analysis color-coded for temperature. Dot size represents different time points. Predicted canonical pathways activated at 6h and 24h (D) and most significant upstream regulators identified (E). F. Volcano plot illustrating the differentially expressed proteins (DEPs) at 6h (40°C vs 37°C). The most significant DEPs are highlighted. G GO/pathway terms enrichment for the given DEPs.

https://doi.org/10.1371/journal.pone.0315119.g002

ER stress and IRE1α are activated in response to fever-range hyperthermia.

Electron microscopy showed signs of ER stress, with notable swelling of the endoplasmic reticulum in cells cultured at 40°C (Fig 3A). qRT-PCR analysis revealed that the expression of ER stress-associated genes such as GRP78 and CHOP was transiently upregulated between 2 and 6 hours (Fig 3B). IRE1α activation was assessed by quantifying both spliced (s) and unspliced (u) XBP1 mRNA. sXBP1 levels increased significantly at 1 and 3 hours in the 40°C condition, while uXBP1 levels remained unchanged (Fig 3C). The transient increase in XBP1 protein was confirmed and quantified in Western Blot. Fig 3D and E). Treatment with the IRE1α inhibitor B-I09 effectively blocked XBP1 splicing in a dose-dependent manner (Fig 3F). Among several targets of the UPR, the expression of DNAJB9, a transcript coding a member of the DNAJ (Hsp40) family with various proteostatic functions, was significantly decreased after treatment by B-I09 (S2 FigA). Of note, hyperthermia also activated the PERK branch of UPR, as the phosphorylation of EiF2α was found higher in cells cultured at 40°C (S2 FigB).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. ER stress response and XBP1 splicing in HCAECs subjected to FRH.

A. Ultrastructural analysis of HCAEC by TEM showing endoplasmic reticulum swelling over time. These pictures are representative of 10 pictures/ conditions, in 3 independent experiments. B. Expression kinetic of GRP78 et CHOP mRNAs over 24h during FRH. C. Expression kinetic of spliced (s)XBP1 and unspliced (u)XBP1 mRNA over 6 hours during FRH as compared to 37°C. D. Western Blot representative of XBP1s protein kinetic expression in HCAECs exposed to moderate hyperthermia (40°C vs 37°C). β-actin was used as a loading control. This blot is representative of 3 independent experiments. E. Quantitative analysis of the kinetic of XBP1s protein expression in HCAECs exposed to 37°C or 40°C (N = 3 experiments). F. RT-qPCR showing the spliced XBP1 (sXBP1) mRNA abundance in HCAEC cultured at 37°C or 40°C for 2h with or without B-I09 (20 µ M) (N = 3 experiments). *P < 0.05, **P < 0.01, ***P < 0.001.

https://doi.org/10.1371/journal.pone.0315119.g003

Hyperthermia increases autophagy flux in ECs.

Hyperthermia-induced autophagy was investigated using LC3B immunostaining and Western Blot analysis. HCAECs cultured at 40°C exhibited a modest but significant increase in LC3B expression over time (Fig 4A, n = 3). LC3B increase was further amplified by chloroquine treatment, which blocks lysosomal degradation, indicating enhanced autophagic flux. p62 levels, a marker of autophagic degradation, were also elevated at 3, 6, and 12 hours in the 40°C condition, confirming the increased autophagic activity (Fig 4B). Immunofluorescence staining revealed co-localization of LC3B⁺ autophagosomes with LAMP2⁺ lysosomes, suggesting active autophagosome-lysosome fusion during hyperthermia (Fig 4C, n = 10 per condition). RFP-GFP-LC3B transfection experiments confirmed transient autophagy induction, with a significant increase in autolysosome formation at 3 hours, which normalized by 24 hours (Fig 4D). HCAEC cultured at 40°C harbored abnormal number of amphisome-like structures carrying high number of autophagosomes, lysosomes and possibly endosomes (Fig 4E). In parallel, cells had increased vesicle trafficking along stages of FRH, suggesting frequent endocytosis and a membrane protein turnover (Fig 4F). Immunostaining revealed colocalization of LC3B⁺ autophagosomes with VE-Cadherin when cells were cultured at 40°C, suggesting that internalization vesicles are taken over by autophagy (Fig 4G). Altogether, these data show that temperature dysregulates autophagy, increasing its flux in an extent that could be detrimental for the cell function and survival.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Heat transiently increases autophagy flux in HCAECs.

A. Experimental design and LC3B levels by Western Blot to assess autophagy flux in HCAEC in response or not to FRH, in presence or absence of 100 µ M chloroquine (CQ). Semi-quantitative measurement is shown on the bottom. B. Experimental design and LC3B or p62 levels by Western Blot to assess autophagic flux within a defined time window by sequential CQ addition. Semi-quantitative measurement is shown on the right. C. LAMP2/ LC3B co-immunolabelling in HCAEC subjected to FRH over time (n = 10 per condition. Images representative of 3 different experiments). D. Fluorescence detection of GFP-RFP-LC3B transfected in HCAEC subjected or not to FRH during 3 or 24h, with autolysosomes in red (RFP+) and autophagosomes in yellow (GFP + RFP+). Semi-quantitative measurement is shown on the right. E. Cytosol ultrastructure in HCAEC subjected to FRH for 12h. F. Plasma membrane ultrastructure in HCAECs subjected to FRH over time. Arrows indicate endosomal formation. G. VE-Cadherin/ LC3B co-immunolabelling in HCAEC subjected or not to FRH for 12h. Arrowheads point at colocalization of autophagosomes with VE-Cadherin. * p < 0.05, ** p < 0.01, *** p < 0.001, 40°C vs 37°C in CQ-treated cells. # p < 0.05, ### p < 0.001, CQ-treated cells vs untreated cells in group 40°C. Student t-test.

https://doi.org/10.1371/journal.pone.0315119.g004

IRE1α inhibition alleviates hyperthermia-induced autophagy and preserves junction integrity.

To evaluate the role of IRE1α in hyperthermia-induced dysfunction, we treated HCAECs with the IRE1α inhibitor B-I09 (0.2–20 µ M). Inhibition of IRE1α effectively prevented cell detachment and reduced microparticle and apoptotic body production after 24 hours at 40°C (Fig 5A and B). Interestingly, B-I09 did not reduce heat-mediated apoptosis, suggesting that cell detachment is not directly mediated by apoptotic mechanisms (Fig 5C). Additionally, B-I09 treatment normalized autophagy flux, as evidenced by reduced LC3B levels and decreased autophagosome formation (Fig 5D and E). Western Blot analysis confirmed that B-I09 treatment restored VE-Cadherin levels, and immunofluorescence showed improved junction integrity compared to untreated cells during hyperthermia (Fig 5F–G). Collectively, these data demonstrate that febrile-range hyperthermia impairs endothelial cell function, leading to cell detachment, junctional disruption, and increased autophagy. Inhibition of IRE1α by B-I09 mitigates these effects by preventing cell detachment, reducing the autophagic burden, and restoring the integrity of cellular junctions, suggesting a critical role of the IRE1α-autophagy axis in maintaining endothelial cell stability under thermal stress (Fig 5H).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. IRE1 inhibition by B-I09 alleviates HCAEC detachment and junction alteration in response to FRH.

Effect of B-I09 on cell detachment (A) and microparticle production (B) and apoptosis (C) by HCAECs in response to FRH. Stau: staurosporine. *p < 0.05, ***p < 0.001 (40°C vs 37°C), # p < 0.05 (B-I09 vs mock). Effect of B-I09 in HCAECs subjected to FRH for 6h on LC3B levels (Western Blot, D) or autophagosome/ autolysosome presentation (LC3B/LAMP2 immunofluorescence, E). F. Levels of total VE-Cadherin assessed by Western Blot in presence of chloroquine (CQ) or B-I09 in HCAECs in response to FRH for 6h. G. Cell junction quality assessed by VE-Cadherin immunolabeling in HCAEC subjected or not to FRH. Corresponding semi quantitative analysis showing intact junction percentage on the right. H. Summary diagram of the study. *p < 0.05, ***p < 0.001 between the indicated groups, by student unpaired t-test.

We acknowledge Rémi Le Borgne and the ImagoSeine core facility of the Institut Jacques Monod, member of the France BioImaging infrastructure (ANR-10-INBS-04) and GIS-IBiSA, for their assistance with cell processing and the transmission electron microscopy observations. We thank Guillaume Chevreux and the Proteomics Core Facility (Institut Jacques Monod UMR 7592) for their assistance with proteomic analysis.

For the purpose of Open Access, a CC-BY public copyright license has been applied by the authors to the present document distributed under a Creative Commons Attribution | 4.0 International license (https://creativecommons. org/licenses/by/4.0/).

https://doi.org/10.1371/journal.pone.0315119.g005

Limitations of the study

This study primarily relies on in vitro models of hyperthermia in cultured endothelial cells, which, while valuable, may not fully recapitulate the complexity of in vivo microvascular dynamics. Future research should include in vivo models to validate these findings, particularly in the context of infection-induced fever and its effects on systemic vascular health. In vitro, some experiments using flow conditions could be used to validate these finding, since shear stress plays an important role in EC biology and survival. Additionally, while B-I09 proved effective in our experiments, the broader physiological consequences of IRE1α inhibition remain to be explored, as this pathway plays a role in various cellular processes beyond autophagy and ER stress management.

Perspectives

Future research should explore the verification of our findings in other endothelial cell lines or primary cells, as these could provide broader biological validation and translational relevance to our work. Additionally, the precise role of XBP1s versus IRE1α itself warrants further investigation through genetic approaches such as XBP1 silencing or knockout experiments. These studies could clarify the specific contributions of IRE1α-mediated UPR signaling versus other pathways in driving endothelial dysfunction under FRH. Given the multifaceted role of the UPR in endothelial cells, further research should explore the involvement of other UPR branches, such as PERK and ATF6, in the cellular response to hyperthermia. It will be important to elucidate how these pathways interact with autophagy and whether they contribute to endothelial dysfunction in the same manner as IRE1α. Moreover, future studies should aim to identify the long-term consequences of modulating UPR pathways, particularly in chronic febrile conditions or in diseases where persistent low-grade inflammation is coupled with endothelial stress.

Moreover, exploring mitochondrial function as part of the endothelial response to FRH presents an intriguing avenue. Assessing mitochondrial dynamics, respiration, and potential crosstalk with the ER could offer further insights into the cellular mechanisms underpinning thermal stress responses.

Another promising avenue of research involves testing the effects of IRE1α inhibition in vivo, both in animal models of febrile infection coupled with microvascular dysfunction and to extend it to more complex models with vascular inflammation.