Mild Hyperthermia-Induced Thermogenesis in the Endoplasmic Reticulum Defines Stress Response Mechanisms

Abstract

Previous studies reported that a mild, non-protein-denaturing, fever-like temperature increase induced the unfolded protein response (UPR) in mammalian cells. Our dSTORM super-resolution microscopy experiments revealed that the master regulator of the UPR, the IRE1 (inositol-requiring enzyme 1) protein, is clustered as a result of UPR activation in a human osteosarcoma cell line (U2OS) upon mild heat stress. Using ER thermo yellow, a temperature-sensitive fluorescent probe targeted to the endoplasmic reticulum (ER), we detected significant intracellular thermogenesis in mouse embryonic fibroblast (MEF) cells. Temperatures reached at least 8 °C higher than the external environment (40 °C), resulting in exceptionally high ER temperatures similar to those previously described for mitochondria. Mild heat-induced thermogenesis in the ER of MEF cells was likely due to the uncoupling of the Ca²⁺/ATPase (SERCA) pump. The high ER temperatures initiated a pronounced cytosolic heat-shock response in MEF cells, which was significantly lower in U2OS cells in which both the ER thermogenesis and SERCA pump uncoupling were absent. Our results suggest that depending on intrinsic cellular properties, mild hyperthermia-induced intracellular thermogenesis defines the cellular response mechanism and determines the outcome of hyperthermic stress.

1. Introduction

Eukaryotic cells have two evolutionarily highly conserved systems to cope with environmental or pathophysiological stress conditions: the heat-shock response (HSR) and the unfolded protein response (UPR) [1,2,3]. The effects of stress depend on the stress dose, such as intensity and duration. By characterizing its effect on Chinese Hamster Ovary (CHO) cells, we have previously classified heat stress into three distinct categories, namely: severe, involving induction of heat-shock proteins (HSPs) together with major macromolecular damage or even cell death; moderate (negligible protein denaturation with less intense HSP induction), and mild, eliciting “eustress” (without HSP induction) [4]. We also showed that mild heat triggers a distinct, dose-dependent remodeling of the cellular lipidome, followed by the expression of HSPs only at higher temperatures [5]. A significant elevation in the relative concentration of saturated membrane lipid species and specific lysophosphatidylinositol and sphingolipid species suggests rapid membrane microdomain reorganization and an overall time-dependent increase in membrane rigidity in response to fluidizing heat. Our RNAseq experiments revealed that mild heat initiated stress-related signaling cascades in the endoplasmic reticulum (ER) in agreement with previous reports [5,6,7], resulting in dose-dependent lipid rearrangement and an elevated resistance to membrane fluidization by benzyl alcohol [5]. To protect cells against lethal, protein-denaturing high temperatures, the classical HSP response evolved. The presence of distinct layers of stress response elicited by different heat dosages highlights the capability of cells to utilize multiple tools to resist and survive potentially lethal stress conditions.

The ER consists of a dynamic membrane network that is required for the synthesis and modification of proteins and lipids. The accumulation of unfolded proteins in the ER lumen activates an adaptive unfolded protein response, UPR, as a mechanism to restore homeostasis. The UPR is mediated by three ER-localized transmembrane proteins: inositol-requiring 1a (IRE1a), PKR-like endoplasmic reticulum kinase (PERK), and activating transcription factor 6 (ATF6) [3,8,9,10]. Interestingly, recent findings show that lipid perturbation is also a direct activator of the UPR, independent of protein unfolding [11,12,13,14,15,16]. Although the mechanism of the UPR process has been extensively investigated, the relationship between heat stress, especially mild hyperthermia, and the ER homeostatic response remains unclear.

Our previous observation that mild, non-protein-denaturing heat induced a type of UPR in mammalian cells [5] initiated our studies on intracellular heat production. At the subcellular level, the mitochondrion is the organelle best known for thermogenesis [17]. Using a temperature-sensitive fluorescent indicator targeted to mitochondria, it has been shown that mitochondria are physiologically maintained at close to 50 °C [18]. This high organelle temperature is maintained in the face of a variety of metabolic stresses, including substrate starvation or modification or decreased ATP demand [19]. The ER has also gained attention as another organelle responsible for heat production, mediated by a Ca²⁺-ATPase (SERCA) pump [20]. Although these studies were not the first to suggest that mammalian mitochondria and the ER could be warmer than their surroundings, such discoveries fostered attention and discussion, especially considering the thermal physics of the phenomenon [21]. While there is continuing debate on sizzling organelles, we aimed to investigate the intriguing possibility that the high metabolic demand of the cellular stress response could generate thermogenesis in the ER, further elevating its temperature and ultimately leading to a heat-shock response.

2. Materials and Methods

2.1. Cell Lines and Culture Conditions

U2OS (human osteosarcoma) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (Gibco, Waltham, MA, USA), 4.5 g/L glucose, 1% L-glutamine, penicillin (200 units/mL) and streptomycin (200 µg/mL) (Sigma–Aldrich, Burlington, MA, USA) at 37 °C in a humidified atmosphere of 5% CO₂. These growing conditions were used for all types of U2OS cells. Cells were routinely tested for mycoplasma contamination using qPCR (MycoQuant Mycoplasma quantification kit, catalog no. MQ-50 AVIDIN, Szeged, Hungary). The U2OS mEOS cell line was created by Belyy et al. and expressed transmembrane IRE1 protein tagged with a C-terminal mEos4b in U2OS IRE1α_KO cells. The cell line was a kind gift from Dr. Belyy, and its construction is described in [22].

To construct the U2OS cell line expressing fluorescently labeled XBP1 (mNeonGreen), 200 ng linearized pLHCX-XBP1 mNeonGreen plasmid NLS (Addgene plasmid #115971; http://n2t.net/addgene:115971; RRID: Addgene_115971), a gift from David Andrews [23], was transfected into 10⁶ U2OS cells using the Amaxa Nucleofector System with the Cell Line Nucleofector Kit V (catalog no: VCA-1003, Lonza, Cologne, Germany) utilizing program X-001 according to the manufacturer’s protocol. Twenty-four hours after transfection, the cells were plated at 20% confluence in complete DMEM supplemented with 150 µg/mL hygromycin B (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA). During selection, the cells were fed with a fresh selection medium, and resistant colonies arose from single cells and were transferred to 96-well plates (Thermo Fisher Scientific, Waltham, MA, USA). Expanded colonies were tested by inducing ER stress with 200 nM thapsigargin for 6 h (Sigma–Aldrich, Burlington, MA, USA) using confocal microscopy (Zeiss LSM800 controlled by the ZEN 2.3 software, Carl Zeiss Microscopy GmbH, Jena, Germany). Colonies without a fluorescent signal in the absence of ER stress but showing a bright green signal upon induction were considered positive as U2OS XBP1 (UXB3) and used in further experiments.

MEF (mouse embryonic fibroblast) cells were routinely grown in complete Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS (Gibco, Waltham, MA, USA), 4.5 g/L glucose, 1% L-glutamine, penicillin (200 units/mL) and streptomycin (200 µg/mL) (Sigma–Aldrich, Burlington, MA, USA) in a 37 °C incubator (Galaxy 170S-Eppendorf/New Brunswick) containing 5% CO₂.

All types of cells were plated at a density of 250 cells/mm² and cultivated in tissue culture plates (VWR, Radnor, PA, USA) or glass-bottom dishes (Greiner Bio-One, Kremsmünster, Austria) 24 h prior to experiments. Cells were at 80–90% confluence at the time of the experiments, ensuring logarithmic growth during subsequent treatments.

2.2. Fluorescent Labeling for Measuring ER Temperature and Ca²⁺ Levels

2.2.1. Temperature Determination with ER Thermo Yellow (λex/λem 510/591 nm)

U2OS and MEF cells in culture medium were incubated with 250 nM ER thermo yellow (diluted from 1 mM stock solution in DMSO) at 37 °C in a humidified atmosphere of 5% CO₂ for 30 min. After incubation, the cells were washed once with pre-warmed PBS (phosphate-buffered saline). ER thermo yellow indicator was a kind gift from Satoshi Arai (National University of Singapore) [20]. Thapsigargin (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) was used as a positive control, and the cells were pre-treated with 1 µM thapsigargin for 1 h, washed in culture medium, then labeled with ER thermo yellow.

2.2.2. Nuclear Staining with Hoechst 33342 (λex/λem 350/461 nm)

To facilitate cell segmentation on microscopic images, nuclei were labeled with 1 µg/mL Hoechst 33342 (bisbenzimide trihydrochloride, Sigma–Aldrich, Burlington, MA, USA).

2.2.3. Determination of ER Ca²⁺ Levels with Mag-Fluo-4 ER (λex/λem 493/517 nm)

Mag-Fluo-4 (M14206, Thermo Fisher Scientific, Waltham, MA, USA) is a low-affinity fluorescent indicator suitable for evaluating the Ca²⁺ concentration in cellular compartments with elevated Ca²⁺ levels, such as the ER [24]. U2OS and MEF cells were incubated in culture medium containing 1 µM Mag-Fluo-4 (1.5 µL of a 1 mM stock solution in DMSO added to 1500 µL of pre-warmed culture medium) at 37 °C for 30 min in an incubator containing 5% CO₂. After incubation, cells were washed with pre-warmed PBS, detached with trypsin, and resuspended in 500 µL of culture medium.

2.3. Cell Fixation

Cells were washed three times with ice-cold PBS, then fixed with 4% PFA (paraformaldehyde) solution for 7 min at room temperature, and washed again with ice-cold PBS. Fresh, pre-warmed RPMI 1640 (Sigma–Aldrich, Burlington, MA, USA) was added to the cells. RPMI 1640 contains no FBS (Gibco, Waltham, MA, USA) and no phenol-red indicator, so the culture medium does not interfere with fluorescent imaging.

2.4. Treatments

2.4.1. Heat-Shock Experiments

Twenty-four hours prior to heat-shock treatment, cells were seeded in tissue culture plates (VWR, Radnor, PA, USA) or glass-bottom dishes (Greiner Bio-One, Kremsmünster, Austria) at a density of 250 cells/mm² in the appropriate complete culture medium. Dishes were sealed with Parafilm and placed into a thermostatically regulated water bath (WNB 7, Memmert) set to different temperatures (39 °C, 40 °C, 41 °C, 42 °C, 43 °C, 44 °C). Samples were analyzed either immediately after the heat shock or after six hours of recovery time at 37 °C.

2.4.2. Treatment with ER Stressors

The culture medium on cells was replaced with medium containing 5 μg/mL tunicamycin (Sigma–Aldrich, Burlington, MA, USA) or 100 nM thapsigargin (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA), and, after one hour at 37 °C, rinsed and replaced with fresh medium.

2.5. RNA Isolation and Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR)

U2OS and MEF cells were heat-treated at different temperatures (37 °C, 39 °C, 40 °C, 41 °C, 42 °C, 43 °C, 44 °C) or treated with 5 µg/mL tunicamycin or 100 nM thapsigargin for one hour. After one hour of recovery, RNA was isolated utilizing an RNA and protein purification kit (Macherey-Nagel, Düren, Germany) as per the manufacturer’s guidelines. RNA was converted into cDNA using the high-capacity cDNA reverse transcription kit (Thermo Fisher Scientific, Waltham, MA, USA). Each reaction mix consisted of 15 µL mRNA (1 µg), 3 µL primer, 1.5 µL reverse transcriptase, 1.2 µL dNTPs, 3 µL buffer, and 6.3 µL RNase-free water. The reverse transcription process was carried out using a T100 thermocycler (BioRad, Hercules, CA, USA) according to the following protocol: 10 min incubation at 25 °C, reverse transcription for 2 h at 37 °C, and 5 min of inactivation at 85 °C. The resulting cDNA product was diluted 1:20 for use as a template in a 20 µL RT-qPCR reaction containing 9 µL of cDNA, 1 µL of primer mix (250 nM for both forward and reverse primers), and 10 µL of 2x Power SYBR-Green PCR master mix (Thermo Fisher Scientific, Waltham, MA, USA). The reactions were run on a RotorGene 3000 instrument (Qiagen, Hilden, Germany) with the following settings: initial denaturation at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s, and annealing at 60 °C for 60 s. Melting curve analysis spanning 50–95 °C was performed to confirm the specificity of amplification. Primer sequences utilized in RT-qPCR reactions are listed in Table S3. Mouse GAPDH was used for MEF cell samples, while human GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and human RPL27 (ribosomal protein L27) genes were used for the U2OS cell line as an internal normalization control. Relative gene expression levels were determined using the ΔΔCt method.

2.6. Flow Cytometry Analysis

2.6.1. Measuring XBP1 Protein Levels

U2OS cells stably expressing fluorescently tagged XBP1 (mNeonGreen) protein were seeded in 35 mm glass-bottom dishes (Greiner Bio-One, Kremsmünster, Austria) at a density of 250 cells/mm² 48 h prior to experiments. Cells were heat-treated at 40 °C and 42 °C or treated with tunicamycin (5 µg/mL) for 1 h, followed by 6 h of recovery at 37 °C, then trypsinized and washed with culture medium. All cell analyses were performed using a BD Accuri C6 flow cytometer (BD Biosciences, CA, USA) equipped with a standard laser and filter setup. Data were obtained with linear amplification of the forward (FSC) and side scatter (SSC) light signals. Fluorescence signals emitted by the mNeonGreen-tagged XBP1 proteins were measured by the FL1 detector for green fluorescence with logarithmic amplification. The light emission from 10,000 cells per sample was analyzed, and the frequency distribution histograms were created with the web-based program floreada.io (https://floreada.io) in FL1. The fluorescence intensity values of the treated cells were compared to the intensity values of the untreated cells (maintained at 37 °C).

2.6.2. Measuring ER Ca²⁺ Levels

Mag-fluo-4 labeled MEF and U2OS cells were seeded in 35 mm dishes (Greiner Bio-One, Kremsmünster, Austria) at a density of 250 cells/mm² 24 h prior to experiments. Cells were trypsinized and washed with culture media. Fluorescence intensity of Mag-fluo-4-loaded cells was measured using a BD Accuri C6 flow cytometer (BD Biosciences, CA, USA) with an FL1 detector for green fluorescence and logarithmic amplification. The light emission from 10,000 cells per sample was analyzed, and the frequency distribution histograms were created with the web-based program floreada.io (https://floreada.io) in FL1. Cells were heated to 40 °C in 10 min, and the mean fluorescence intensity of the treated cells was compared to the values of the untreated cells (maintained at 37 °C). Cells were also pre-treated with the SERCA pump inhibitor, 1 µM thapsigargin, for one hour.

2.7. Fluorescence Microscopy Assays

2.7.1. ER Temperature Measurements by ER Thermo Yellow Fluorescence

Fluorescence intensity changes in ER thermo yellow-labeled cells were monitored using a Cell-Discoverer 7 fluorescence microscope (Zeiss, Jena, Germany) equipped with the Axiocam 506 imaging device for controlling light, temperature, and CO₂. For the ER thermo yellow indicator, we used λex528/λem553, and for the Hoechst stain, λex348/λem455, with 25% and 6% light source intensity, respectively. One image was acquired at 37 °C and within 10 min, and another image was acquired at 40 °C to reduce the possibility of photobleaching. Approximately 1000 cells were acquired from one image, and they were accurately segmented for further analysis utilizing MATLAB 2019b (MathWorks Inc., Natick, MA, USA), ImageJ 1.54f (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, MD, USA), and CellProfiler 4.2.1 (Broad Institute, Cambridge, MA, USA). Fluorescence intensity values of individual cells were averaged, and the average intensity of the heat-treated sample (F) was normalized to the intensity value of the pre-heated sample (F0), resulting in the F/F0 ratio in which F/F0 = 1 meant no change in ER temperature, F/F0 < 1 meant increased temperature, and F/F0 > 1 meant decreased temperature). Changes in F/F0 of fixed cells give information about the external temperature (37–40 °C) change, and any additional change in live cells provides information about the changes in the ER temperature. Increasing ambient temperature from 37 °C to 40 °C resulted in a 10% decrease in the F/F0 ratio of fixed cells, which was used to calibrate the temperature changes in live cells.

2.7.2. Super-Resolution Microscopy

Super-resolution fluorescence measurements were performed on a custom-built inverted microscope system based on a Nikon Eclipse Ti-E frame. Images were captured under epi-illumination with an oil immersion objective (Nikon CFI Apo 100x, NA = 1.49). Fluorescent proteins were photoconverted by a 405 nm laser (Nichia, 60 mW) and were localized by a readout laser operating at 561 nm (Cobolt Jive, 300 mW). A fluorescent filter set (Semrock, LF405/488/561/635-A-000, Rochester, NY, USA) and an additional emission filter (Semrock, FF01-600/37) were used to separate the excitation and emission wavelengths. Images were captured by an EMCCD digital camera (Andor iXon3 897, 512 × 512 pixels, 16 µm pixel size). Typically, 3000–6000 frames were captured with an exposure time of 100 ms (until no more photoconversion was detectable). Image stacks were evaluated with the rainSTORM 3.1.8. localization software [25]. Only localizations with precisions of <45 nm were used for further analysis.

Density-based spatial clustering of applications with density-based cluster analysis (DBSCAN) was performed on the localization datasets for quantitative evaluation. After thorough tests (Figure S4), the input parameters for this algorithm were set to ε = 45 (the maximum distance between two adjacent points) and N = 10 (the minimum number of points that form a cluster). Cluster parameters like cluster area and number of localizations per cluster were calculated and saved for further analysis.

3. Results

3.1. Distinct Cell Types Show Differential Stress Transcriptome Profiles upon Heat Treatment

The transcription profiles of key HSP and UPR genes were described in MEF and U2OS cells in response to heat treatment for 1 h at different temperatures (37 °C, 39 °C, 40 °C, 41 °C, 42 °C, 43 °C, 44 °C) (Figure S1, Figure 1 and Figure 2, Table S1). RT-qPCR analysis revealed a distinct, cell-type-dependent response for the different levels of heat stress.

In MEF cells, Hsp25 and Hsp70 genes exhibited significant upregulation already at mild (39–40 °C) temperatures, while in U2OS cells the elevated expression of HSP25 could only be observed at higher (41–42 °C) temperatures (Figure 1A). Transcription of the HSP genes was much higher in MEF cells than in U2OS cells, while the expression of the HSR master regulator HSF1 was significantly increased only in U2OS cells (Table S1 and Figure S1).

To investigate the mechanism of the UPR, we focused on its three main signaling branches, represented by the IRE1, PERK, and ATF6 genes. Relative expression of these genes, along with XBP1 (X-box binding protein 1), the downstream target of IRE1, and ATF4 and CHOP, the targets of PERK, were analyzed. In U2OS cells, there was no change in the expression of IRE1, PERK, and ATF6 (Table S1 and Figure S2), but our results revealed a significant increase in the expression of XBP1 and CHOP upon heat treatment at 39–40 °C (Figure 1B,C). Notably, PERK-dependent Atf4 expression was significantly downregulated in MEF cells at mild temperatures (39–42 °C) and normalized under more severe heat conditions (43–44 °C) (Figure 1C).

Upon activation, IRE1 undergoes autophosphorylation and catalyzes RNase-mediated unconventional splicing of the XBP1 mRNA [22]. Our RT-qPCR results in U2OS cells revealed an increase in total XBP1 expression, concomitant with an increased amount of spliced XBP1 (XBP1-s) mRNA (Table S1 and Figure 2B). In parallel, the amount of unspliced XBP1 (XBP1-us) slightly increased in the mild temperature region (39–41 °C) but behaved oppositely in response to more severe heat treatment (43–44 °C) by displaying a much more pronounced decrease (Figure S2). In MEF cells, there was no alteration either in the total or in the unspliced Xbp1 level under mild heat stress (39–42 °C), while the spliced Xbp1 was downregulated. At higher temperatures (43–44 °C), we observed downregulation of the unspliced Xbp1 and upregulation of the spliced form.

Pearson correlation analysis was conducted to examine the relationship between the induced HSR and UPR genes upon exposure of MEF and U2OS cells to heat, and distinct correlation patterns were observed (Table S2 and Figure S3). MEF cells exhibited moderate to very strong positive correlations between the inductions of the examined HSP family members (r = 0.96 for Hsp25 vs. Hsp70). In U2OS cells, the less pronounced HSP gene expression was paralleled by weaker correlations (r = 0.65 for HSP25 vs. HSP70). In contrast, the association was strong between certain UPR genes in U2OS but not in MEF cells: for Xbp1 total vs. Chop, r = 0.84 in U2OS and r = 0.32 in MEF. On the other hand, we found a strong negative correlation between the induction of HSR and UPR genes in MEF cells (r = −0.9 for Hsp90aa1 vs. Atf4), whereas the same association was moderately positive in U2OS cells (r = 0.4) (Table S2 and Figure S3).

3.2. IRE1 Clustering in Response to Mild Heat Treatment of 40 °C

IRE1, one of the key UPR signal activators, is known to assemble into larger oligomers in response to treatment with tunicamycin, which induces ER stress [26]. In view of this finding, we asked whether a similar IRE1 clustering would result from mild, fever-like temperature stress. U2OS IRE1-mEOS cells, in which IRE1 transmembrane proteins were tagged with a fixation-resistant photoactivatable fluorescent mEos4 tag, were treated at 40 °C or 42 °C for one hour and fixed. Imaging by dSTORM super-resolution microscopy revealed that the number of small clusters (<5000 nm²) decreased, while the number of larger clusters increased in response to the heat treatments (Figure 2). The increase in cluster area was highly significant at 40 °C, and may be correlated with enhanced UPR signaling activity.

3.3. IRE1-Dependent XBP1 Expression Is Upregulated upon Mild Heat Treatment

To further investigate the intricate involvement of IRE1 in UPR signaling, IRE1-dependent XBP1 protein levels were measured during heat treatments using a fluorescence reporter construct in U2OS cells [23]. Flow cytometry demonstrated significant increases in IRE1-dependent XBP1-mNeonGreen levels in cells subjected to heat treatments (Figure 3). The mean fluorescence intensity of the reporter has increased by +13.6% and +60.2% in cells treated at 40 °C and 42 °C, respectively. These observations supported the results of RT-qPCR and dSTORM analyses and corroborated the activation of the UPR pathway under fever-like conditions.

3.4. Mild Heat Could Induce Thermogenesis in the ER

According to our previous findings and a literature search, the induction of HSP genes is a response typically associated with higher temperatures. To explain the unexpected increase in the induction of HSP genes in MEF cells elicited by mild heat, we proposed intracellular thermogenesis as the potential source. To test this hypothesis, we used ER thermo yellow, a fluorescent probe selectively targeting the ER, to measure the intracellular temperature. Since the probe’s intensity is not affected by enzymatic activity, metabolic activity, and cytosolic Ca²⁺ concentration, a decrease in the probe’s fluorescence intensity indicates increased temperature in the ER [20].

High-resolution fluorescence microscopy images of the ER thermo yellow-labeled MEF and U2OS cells (heat-stressed while on the microscope stage) showed significant differences in the ER temperature of heat-treated cells (Figure 4A). In live U2OS cells, the 40 °C treatment did not increase the ER temperature more than in fixed cells. In contrast, a significant ER temperature elevation was observed in MEF cells. A calibration curve was used to estimate the extent of temperature increase, calculated from intensity changes in the fixed cells (Figure 4C). This estimation revealed temperatures reaching at least 8 °C higher than the external environment (40 °C) in the ER of live MEF cells.

The elevated ER temperature in MEF cells might result from SERCA pump uncoupling. In this state, the pump continues hydrolyzing ATP but without effectively transporting calcium ions across the ER membrane, with the remaining energy dissipating as heat. To test this possibility, we treated groups of cells with the non-competitive SERCA inhibitor, thapsigargin, as a positive control [27]. Thapsigargin treatment eliminated the increase in ER temperature in live MEF cells (Figure 4A), suggesting that SERCA activity was indeed the source of the mild heat-induced ER thermogenesis.

3.5. ER Ca²⁺ Levels Decrease upon Mild Heat Treatment

To further support the mild heat-induced uncoupling of the SERCA pump in MEF but not in U2OS cells, we measured ER Ca²⁺ levels in both cell lines after 40 °C heat treatment. The low-affinity calcium probe, Mag-fluo-4, registers changes in Ca²⁺ level in a cell by changes in fluorescent intensity as measured by flow cytometry. Upon heat treatment, a significant decrease in ER Ca²⁺ levels was detected in MEF cells, contrasted with the absence of detectable changes in Ca²⁺ levels in U2OS cells (Figure 5). This observation further strengthens our hypothesis that the observed temperature change in MEF cells is linked to SERCA pump uncoupling caused by fever-like heat treatment.

Cells were also pre-treated with thapsigargin, which, as expected, resulted in decreased ER Ca²⁺ levels, disrupting cellular calcium homeostasis. TG-treated cells were then subjected to mild heat stress. Inhibition of the SERCA pump by thapsigargin prevented further Ca²⁺ depletion by heat, thus supporting the role of SERCA in intracellular heat production upon mild heat stress.

4. Discussion

4.1. Interplay between the UPR and HSR upon Heat Stress

Previously, we described how Chinese Hamster Ovary (CHO) cells use distinct layers of stress response to protect against different levels of environmental stress [5]. We also identified the UPR as a first line of defense during mild hyperthermia, resulting in membrane rearrangement and protection against various environmental challenges [4,5]. To further generalize our previous findings, we used real-time quantitative PCR and super-resolution fluorescence microscopy to dissect the interrelationship and the mechanism of the switch between UPR and HSR in mouse embryonic fibroblast (MEF) and human osteosarcoma (U2OS) cell lines. By applying a range of heating challenges from mild to more severe (39–42 °C), we detected the induction of both types of cellular defense, UPR, and HSR, but with distinctive, cell-specific mechanisms. While UPR seems to be sufficient for withstanding mild hyperthermia in U2OS cells, this is not the case for MEF cells, where the HSR is induced at much lower temperatures (Figure 1). Among the three branches of the UPR, the IRE1 (XBP1) and PERK (CHOP, ATF4) pathways were activated by mild heat, while expression of the ATF6 branch seemed to be silent in U2OS cells (Table S1 and Figure S2). Previously, it was shown for AD293 cells that a longer (1–16 h) heat stress at 40 °C could induce the downstream targets of ATF6; however, they were not induced for other cell types tested including two hepatocyte cell lines, a pancreatic beta cell line, and MEF cells [7]. As ATF6 activation is influenced by its oxidation state, glycosylation state, and proteosome-dependent turnover, it seems that stimuli preferentially affecting these aspects of ER protein processing do not occur during mild, fever-like hyperthermia but only with higher temperatures [28].

Pearson analysis has shown a negative correlation between the expression (including splicing) of specific HSR and UPR genes in MEF cells, with particularly strong correlation between the levels of Hsp90aa1 mRNA and UPR mRNAs of Xbp1_s (r = 0.8) and Atf4 (r = −0.9) (Table S2 and Figure S3). Since HSP90 was shown to stabilize both PERK and IRE1 (but not ATF6) by associating with their cytosolic domains this phenomenon may be the result of the dissociation of HSP90 from the two ER sensor kinases stimulated by intensive intracellular thermogenesis [29,30]. It was reported earlier for five different mammalian cell types that exposure of cells to longer periods of hyperthermia at 40 °C resulted in a partial or full ER stress pathway induction as well as an HSR, while at 43 °C, the ER stress pathway was inhibited [7]. In contrast, our results indicate that under extreme conditions, when the stress cannot be overcome, a different pathway of the UPR can be activated for the purpose of initiating apoptosis, cell death, and clearance (Figure 6) [31]. These observations suggest that there are first and second waves in the UPR, depending on the severity of the stress (Figure 6).

Under mild hyperthermia, U2OS cells activate both the UPR and HSR to maintain stability, while severe heat triggers a second UPR wave. In MEF cells, ER thermogenesis occurring at mild temperature initiates a stronger HSR and a consequent downregulated UPR. It was previously reported that HSR decreases UPR activation in Saccharomyces cerevisiae mainly through the facilitation of protein folding and secretion [32]. In both yeast and mammalian cells, the transport of conformationally defective proteins to the vacuole/lysosome for degradation is another mechanism for relieving ER stress [33]. The downregulation of UPR genes in MEF cells could also be explained by the general suppression of gene expression upon HSP gene induction [34].

Previously, we reported the molecular details of cellular “eustress”, wherein CHO cells adapted to mild heat by maintaining membrane homeostasis, activating lipid remodeling, and redistributing chaperone proteins [4]. Recent studies have revealed an alternative ER stress activation known as lipid bilayer stress (LBS), which elicits a UPRLBS independent from the classical UPRPT and is responsible for managing proteotoxicity and restoring protein homeostasis. Thibault et al. proposed that LBS serves as a gateway between acute and chronic ER stress [35]. Our results suggest that there could be two waves of UPR. UPR^LBS is switched on to maintain membrane integrity in U2OS cells as a result of mild heat-induced ER membrane perturbation, while a greater temperature increase would initiate UPR^PT, as was observed in both cell lines above 42 °C (Figure 1B). Most of the data on UPR^LBS were obtained from treatments resulting in increased membrane stiffness or decreased membrane fluidity via lipotoxicity, palmitic acid supplementation, or inositol depletion [16,36]. However, there are some indications that compounds that increase membrane fluidity, like benzyl alcohol, tergitol, and palmitoleic acid, produce an effect similar to heat and can also initiate the UPR^LBS independently of unfolded protein accumulation in yeast [37]. It was suggested that it is crucial to maintain the biophysical properties of the ER membrane, especially its compressibility [38]. It may be that since heat alters membrane thickness [39], lateral diffusion [4], bending and stiffness [40], and membrane compressibility, UPR^LBS is switched on during mild hyperthermia to maintain the biophysical integrity of all cellular membranes, including the ER.

The induction of XBP1 splicing indicates that the IRE1 pathway is activated by mild heat in U2OS cells (Figure S1 and Figure 1B). Our super-resolution fluorescence microscopy data showed that mild ER heat stress induced the clustering of IRE1 similarly to the UPR^PT observed by [22]. It was suggested that the pre-assembled, resting state IRE1 dimers form tetramers (dimers of dimers) upon activation, followed by the appearance of massive clusters if IRE1 expression remains high [22]. The increased IRE1 cluster size in the current experiments demonstrates activation of the IRE1 pathway upon mild hyperthermia, which was also confirmed by the downstream induction of the XBP1 protein.

In vivo experiments and MD simulations showed that the architecture of the IRE1 molecule could induce compression of the lipid bilayer, making it capable of sensing the physical properties of the membrane environment, which determine the lifetimes of the dimeric and oligomeric configurations of the ER sensor [14]. While previous studies suggested that IRE1 dimerization was induced by increased saturation and thickening of the bilayer [14,41], our super-resolution experiments demonstrated that the clusterization of IRE1 can also happen in a thinner, more fluid membrane and that the membrane perturbation induced by mild heat plays a role in the initiation of the UPR^LBS. The induction of UPR genes upon more severe heat treatment indicates a second wave of responses initiating the classical UPR^PT function to either restore proteostasis or induce apoptosis [1].

4.2. Mild Heat-Induced Intracellular Thermogenesis

Our novel observation that mild, non-protein-denaturing heat induced an HSR in a cell-type-specific manner initiated our studies on intracellular thermogenesis. While the role of mitochondria in non-shivering thermogenesis has already been widely accepted [42], the high temperature [19] of the organelle is still under debate [21]. There has been increasing evidence that the sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) pump plays a role in non-shivering thermogenesis in the muscle [43], as uncoupling of the SERCA pump results in non-productive activity and increased ATP hydrolysis and heat production [44,45].

We hypothesized that a mild environmental temperature increase in MEFs could uncouple the SERCA pump, therefore increasing ER temperature to the levels of HSR initiation. In a reconstituted liposome system, it was shown that the optimal functioning of the SERCA pump depended on the conformation of the lipid head groups and the deformation of the bilayer [46,47,48]. Using the temperature-sensitive, ER-specific ER thermo yellow probe and high-content fluorescence imaging, we detected temperatures in excess of 48 °C in the ER of MEF cells (Figure 4) during mild heating, which could induce protein denaturation in the lumen of the ER and initiate UPRPT [49].

Further supporting this hypothesis, the ER Ca2+ levels drop as a result of mild heat treatment (Figure 5), and TG inhibition of the SERCA pump suppresses ER thermogenesis (Figure 4). There are few studies reporting elevated ER temperatures. Using ER thermo yellow, Arai et al. (2014) estimated that the SERCA-uncoupling Ca²⁺ ionophore, ionomycin, caused an ER temperature elevation of 1.7 ± 0.4 °C while using a derivative of this probe, ERthermAC, the induction of thermogenesis by isoproterenol or forskolin raised the temperature from 25 °C to 42 °C in WT-1 brown adipocyte cells [50]. Recently, it was reported that enhanced Ca²⁺ cycling by activation of the SERCA2b–RyR2 pathway stimulated UCP1- and mitochondria-independent thermogenesis in beige adipocytes from mice, humans, and pigs [51]. These results from the literature, together with our current finding of an ER temperature increase of 11 °C, support the hypothesis of a greater role of the ER in thermogenesis.

Although lipid bilayer stress is now established as an ER stress inducer parallel to and independent of proteotoxicity, our understanding of the mechanism of this phenomenon and the signaling pathways involved remains limited, and we are only beginning to understand how the lipidome is influenced by the response of the UPR-ER system to stress. We anticipate that applying the tools of proteomics, metabolomics, and lipidomics will lead to new findings and a more comprehensive understanding. Different environmental, psychological, or pathophysiological stresses induce membrane perturbations, which have to be corrected [52], and this process involves increased metabolic activity to replace or reconfigure lipids and proteins, which could also induce hyperthermia [53].

Taken together, the literature, the data presented here, and the observation that the activity of mitochondrial respiratory chain enzymes was found to be maximal at or slightly above 50 °C [18], it may be necessary to reevaluate the concept of homogeneous intracellular temperature distribution, and how membrane integrity can be maintained at these high temperatures. Our previous lipidomics results suggested the existence of a rapid compensatory mechanism involving specific lipid rearrangements by which cells could reestablish normal dynamic membrane phase behavior [5]. The extent of the changes in the lipidome, however, does not suggest a homeoviscous adaptation to very high temperatures; therefore, these studies should be extended to the level of the organelles and membrane proteins potentially involved in thermogenesis. We should also consider the possibility of significant local membrane fluidity alterations because of heterogeneous spatial and temporal temperature fluctuations. Lipids with special properties that could counterbalance local membrane perturbations should also be examined.

The current study revealed that mild hyperthermic stress, like a fever, can lead to a disturbance of ER homeostasis and, depending on the cell type, generate a significant intracellular temperature rise via thermogenic pathways. The stress-induced high internal temperatures could modify the cellular response, and the intrinsic cellular properties could determine the outcome of hyperthermic stress. The data presented here demonstrate that the increased ER temperature initiates a cytosolic HSR, suggesting the existence of ER-associated HSR sensor(s). A role for Hsf1 and other components of the HSR in ER quality control was proposed before to regulate protein folding in the ER and vesicular transport [54]. The literature and our previous studies have lent support to a model in which thermal stress is transduced into a signal at the membrane level [55,56]. It was also demonstrated that rapid membrane relocalization of HSPs can maintain membrane fluidity and polymorphism [57]; therefore, it is likely that the HSR can also protect ER membranes.

The ER contributes to the production and folding of about a third of cellular proteins and is thus inextricably linked to the maintenance of cellular homeostasis and the fine balance between health and disease. Our findings may have implications in the development of therapeutic strategies to combat diseases in which the UPR was shown to play an important role, like cancer [57], neurodegenerative [58] and ocular diseases [2], metabolic diseases like obesity and diabetes [59] and Duchenne’s muscular dystrophy [60]. Knowledge of cell-to-cell variation in response to hyperthermia could also be relevant for preventing heat stress caused by global warming. The variation in the cellular stress response will affect not only the progression of the disease but also the effectiveness of existing therapeutics (e.g., antibiotics), and also directly affects cell physiology and the ability of the organism to survive extreme heat waves [61].