Hydrogen peroxide depolarizes mitochondria and inhibits IP₃-evoked Ca²⁺ release in the endothelium of intact arteries

Graphical abstract

Highlights

• •H₂O₂ is produced by several cell processes including mitochondria and may act as an intracellular messenger and cell-cell signalling molecule.
• •Spontaneous local Ca²⁺ signals and IP₃-evoked Ca²⁺ increases were inhibited by H₂O₂.
• •H₂O₂ suppression of IP₃-evoked Ca²⁺ signalling may be mediated by mitochondria via a decrease in the mitochondrial membrane potential.
• •H₂O₂-induced mitochondrial depolarization and inhibition of IP₃-evoked Ca²⁺ release, may protect mitochondria from Ca²⁺ overload during IP₃-linked Ca²⁺ signals.

Abstract

Hydrogen peroxide (H₂O₂) is a mitochondrial-derived reactive oxygen species (ROS) that regulates vascular signalling transduction, vasocontraction and vasodilation. Although the physiological role of ROS in endothelial cells is acknowledged, the mechanisms underlying H₂O₂ regulation of signalling in native, fully-differentiated endothelial cells is unresolved. In the present study, the effects of H₂O₂ on Ca²⁺ signalling were investigated in the endothelium of intact rat mesenteric arteries. Spontaneous local Ca²⁺ signals and acetylcholine evoked Ca²⁺ increases were inhibited by H₂O₂. H₂O₂ inhibition of acetylcholine-evoked Ca²⁺ signals was reversed by catalase. H₂O₂ exerts its inhibition on the IP₃ receptor as Ca²⁺ release evoked by photolysis of caged IP₃ was supressed by H₂O₂. H₂O₂ suppression of IP₃-evoked Ca²⁺ signalling may be mediated by mitochondria. H₂O₂ depolarized mitochondria membrane potential. Acetylcholine-evoked Ca²⁺ release was inhibited by depolarisation of the mitochondrial membrane potential by the uncoupler carbonyl cyanide 3-chlorophenylhydrazone (CCCP) or complex 1 inhibitor, rotenone. We propose that the suppression of IP₃-evoked Ca²⁺ release by H₂O₂ arises from the decrease in mitochondrial membrane potential. These results suggest that mitochondria may protect themselves against Ca²⁺ overload during IP₃-linked Ca²⁺ signals by a H₂O₂ mediated negative feedback depolarization of the organelle and inhibition of IP₃-evoked Ca²⁺ release.

1. Introduction

The endothelium is the single layer of cells that lines the entire cardiovascular system and it is exposed constantly to a wide range of mechanical and chemical stimuli. The endothelium responds to these stimuli by releasing Ca²⁺-dependent vasoactive factors that include nitric oxide, prostacyclin, endothelium-derived contracting factors, von Willebrand factor, tissue plasminogen activator and endothelial derived hyperpolarising factor (1, 2). These vasoactive factors allow the endothelium to regulate almost all cardiovascular activities including vascular tone, immune responses, angiogenesis and vascular remodelling [1].

There is accumulating evidence that reactive oxygen species (ROS) also regulates endothelial function. ROS modulates endothelial cell growth, proliferation, endothelium-dependent relaxation, cytoskeletal reorganization, inflammatory responses and endothelium-regulated vascular remodelling. Among various ROS, hydrogen peroxide (H₂O₂) fulfils the prerequisites for serving as an intracellular messenger and acting as a cell-cell signalling molecule. H₂O₂ is a small and non-polar molecule produced by several cell processes that include mitochondria and NADPH oxidase [17,29,76]. During mitochondrial ATP production, the electron transport chain leaks electrons from complexes I and III resulting in the formation of superoxide anion radical (${\mathrm{O}}_{2^{-}}$) [11]. ${\mathrm{O}}_{2^{-}}$ generates H₂O₂ spontaneously, or by the activity of superoxide dismutases. Although ${\mathrm{O}}_{2^{-}}$ is not membrane permeable, H₂O₂ can diffuse across biological membranes or may cross membrane boundaries via channels like aquaporins [8] to regulate physiological and pathological cellular processes [3,62,75,85].

Many of the systems that produce H₂O₂, such as mitochondria, are modulated by the cytoplasmic Ca²⁺ concentration [6,24,36]. Ca²⁺ released from IP₃Rs may result in Ca²⁺signals that propagate into the mitochondrial matrix [35,59]. Changes in mitochondrial Ca²⁺ may lead to enhanced ATP synthesis [74] and, as a result, increased ROS production [68]. Conversely, increased H₂O₂ generated by mitochondrial activity may modulate Ca²⁺ signalling to exert control on endothelial function. For example, in various cultured cell lines, H₂O₂ evokes Ca²⁺ release from the internal Ca²⁺ store [27,37,77].These observations raise the possibility that there may be feedback regulation of mitochondrial ATP production by changes in the cell activity mediated via the cytoplasmic Ca²⁺ concentration. To explore this possibility we measured the effects of H₂O₂ on Ca²⁺ signalling in the endothelium in large numbers of endothelial cells in intact blood vessels. We show that H₂O₂ depolarises mitochondria and suppresses IP₃ evoked Ca²⁺ signalling.

2. Methods

2.1. Animals

All animal husbandry and euthanasia were carried out in accordance with the prior approval of the University of Strathclyde Animal Welfare and Ethical Review Body and under relevant UK Home Office Regulations, [Schedule 1 of the Animals (Scientific Procedures) Act 1986, UK]. Strathclyde BPU is a conventional unit which undertakes FELASA quarterly heath monitoring. Male Sprague-Dawley rats (10–12 weeks old), from an in-house colony, were used in the study. Animals were housed 3 per cage (RC2F cages, North Kent Plastics Company, UK), provided with enrichment (aspen wood chew sticks and hanging huts), nesting material (Sizzle nest, LBS Technology, UK), and fresh water and chow (RM1, Special Diet Services, UK) were available ad libitum. Room temperature was 19–23 °C (set point 21 °C), humidity was 45–65 %, and a 12 h light cycle was used. Rats were euthanatized by intraperitoneal injection of pentobarbital sodium (200 mg/kg, Pentaject, Merial Animal Health Ltd, UK).

2.2. Endothelial Ca²⁺ imaging

First order mesenteric arteries were isolated, placed into a physiological saline solution (PSS), cleaned of adherent fat and then used immediately. Each artery was then cut open and pinned flat on a sylgard block, with endothelial cells facing upward (en face preparation). The endothelium was then loaded with acetoxymethyl ester form of the Ca²⁺ indicator, Cal-520 (5 μM) and 0.02% pluronic F-127 in DMSO, for 30 min at 37 °C [55,[79], [80], [81]]. Following incubation, arteries were gently washed before the Sylgard block was inverted and placed in a custom-made bath chamber. The bottom of the chamber was a 0-thickness glass coverslip and two (0.2 μm diameter) steel pins were set between the coverslip and the block to prevent endothelial cells from contacting the coverslip, and to allow solutions to flow across the endothelium. Ca²⁺ images were acquired at 10 Hz on an inverted fluorescence microscope (TE300, Nikon, Japan) using a 40×, 1.4 NA oil immersion lens and a back-illuminated electron-multiplying charge-coupled device (EMCCD) camera (1024 × 1024 13 μm pixels; iXon 888; Andor, UK). Fluorescence excitation (488 nM wavelength) illumination was provided by a monochromator (Horiba, UK).

2.3. Localized flash photolysis

In some experiments, the endothelial Ca²⁺ response to local photolysis of caged IP₃ was examined. In these experiments, the endothelium was loaded membrane permeant, caged IP₃ (5 μM) for 30 min at 37 °C. A xenon flash lamp (Rapp Optoelecktronic, Germany) was used to uncage IP₃ [13,47,80]. The output light was filtered using a UG-5 filter to select ultraviolet light. The light was focused and merged into the excitation light path via a fibre optic bundle and long pass dichroic mirror attached to the lens part of the microscope’s epi-illumination attachment [13,53,58]. The area of the photolysis site (∼80 μm diameter) resulted from the fiber optic diameter and the objective lens magnification (40x).

2.4. Imaging endothelial mitochondria

To assess mitochondrial membrane potential, arteries were pinned out in a Sylgard coated chamber designed for use on an upright microscope. Mitotracker Green FM (100 nM) was added to the PSS and the endothelium was incubated for 20 min followed by 20 min washing. Tetramethylrhodamine ethyl ester (TMRE) (60 nM) was added to the PSS and the endothelium was incubated 10 min [20,21,80]. TMRE (60 nm) was subsequently present in all perfusion solutions. Minimal photobleaching of TMRE was observed over the 5 min recording periods used. TMRE and Mitotracker Green images (10 Hz) were acquired on an upright microscope (Eclipse FN1; Nikon, Japan) equipped with a 60× water immersion objective (1.0 numerical aperture) and an EMCCD camera (iXon 888; Andor, UK).

2.5. Experimental protocols

The effect of H₂O₂ on basal endothelial Ca²⁺ activity was studied using a non-cumulative concentration response in the same preparation. In these experiments, a Ca²⁺-free PSS was used and H₂O₂ added to the Ca²⁺-free perfusate. To prevent depletion of internal Ca²⁺ stores, arteries were incubated in Ca²⁺-containing PSS between each exposure to H₂O₂.

The effect of H₂O₂ (with or without catalase, 1000 U ml⁻¹) on evoked (acetylcholine, ACh; 100 nM) endothelial Ca²⁺ activity was studied in paired experiments. In these experiments, a control response (5-minute recording) to ACh (flowed rate: 1.5 ml min⁻¹) was obtained before the tissue was washed for 5 min, and allowed to equilibrate for 10 min. ACh was then applied a second time, together with H₂O₂ (with or without catalase), and the responses compared. ACh and H₂O₂ (with or without catalase) were applied via separate syringe pumps each at 0.75 ml min⁻¹. The effects of various pharmacological interventions on ACh-evoked Ca²⁺ signalling were also studied in paired experiments in which control responses were first obtained and then the endothelium was incubated with each antagonist for 20 min. Following the incubation period, ACh was applied a second time and responses compared to control. Each pharmacological agent was present throughout the second exposure to ACh.

Experiments utilising caged-IP₃ also used a paired experimental design. An initial response to photolysis was recorded, and the tissue was then rested for 10 min. The endothelium was then incubated with H₂O₂ for 20 min before a second response, using the same photolysis location, was obtained.

2.6. Data analysis

Automated analysis of endothelial Ca²⁺ imaging recordings was carried out using custom-written Python routines [44,80,81]. In brief, average intensity projections were used to generate regions-of-interest (ROI) around each cell. The Ca²⁺ response of each endothelial cell was then extracted by averaging the fluorescence intensity within each ROI, for each cell and image in the dataset. Each ROI/cell was assigned an identification number so that the response of each cell could be compared within experimental series. Fluorescence signals are expressed as ratios (F/F₀) of fluorescence counts (F) relative to baseline (control) values before stimulation (F₀). The baseline (F₀) was identified automatically as the 100 frame (10 s) period exhibiting the lowest noise prior to the introduction of any agonist. The total number of oscillations, and the amplitude of each oscillation were then extracted for each cell using a zero-crossing peak-detection algorithm [82] for signals exceeding 3 times the standard deviation of baseline noise.

2.7. Statistics

All data are presented as mean ± SEM of n biological replicates. Data were analysed using repeated measures one-way ANOVA with Geisser-Greenhouse correction and Dunnett’s multiple comparisons test, or paired t-test as appropriate. A p value less than 0.05 was considered statistically significant. All statistical analysis was performed using GraphPad Prism version 6.0 (GraphPad Software, USA) was used to run the statistical analysis.

2.8. Reagents and chemicals

The PSS consisted of (in mM):145 NaCl, 2.0 MOPS, 4.7 KCl, 1.3 NaH₂PO₄, 5.0 Glucose, 1.17, MgCl, 2.0 CaCl, 0.02 EDTA (pH adjusted to 7.4 with NaOH). In experiments using Ca²⁺ free PSS, CaCl₂ was replaced with MgCl₂ on an equimolar basis and EGTA (1 mM) was included. Caged-IP₃ (caged-IP₃ 4,5-dimethoxy-2-nitrobenzyl) was obtained from Sichem (Germany). Cal-520 was obtained from Abcam (UK). Pluronic F-127 was obtained from Invitrogen (UK). Mitotracker Green FM was obtained from Invitrogen (UK). All other drugs and chemicals were obtained from Sigma (UK). Stock solutions of ACh, catalase-polyethylene glycol and H₂O₂ were prepared by dissolving each chemical in double-distilled, dionized water. 2-aminoethoxydiphenyl borate (2-APB), caged-IP₃, Cal-520, carbonyl cyanide m-chlorophynyl hydrazine (CCCP), oligomycin, TMRE and Mitotracker Green FM were dissolved in DMSO.

2.9. Data availability

All data underpinning this study is available from the authors upon reasonable request.

3. Results

To determine if H₂O₂ alters spontaneous Ca²⁺ release from the internal store in intact mesenteric arteries, H₂O₂ (100 nM, 1 μM, 10 μM and 100 μM) was applied in a Ca²⁺ free PSS (Fig. 1). Between H₂O₂ applications arteries were washed in PSS (containing Ca²⁺) to allow the internal Ca²⁺ stores to refill. As the concentration of H₂O₂ increased, spontaneous Ca²⁺ release events decreased (Fig. 1). As spontaneous Ca²⁺ release arises from IP₃-receptor activity [44,80,81], these results suggest that H₂O₂ may suppress Ca²⁺ release from the internal Ca²⁺ store.

Fig. 1.

Effect of H₂O₂ on spontaneous local Ca²⁺ signals arising from the internal Ca²⁺ stores. (A) Pseudo-colour images of spontaneous Ca²⁺ signals activity (red high, blue low amplitude) over a 5 min period in control (Ca²⁺-free PSS) and with H₂O₂ (100 nM to 100 μM). Between each recording, H₂O₂ was washed out and Ca²⁺ was restored to the bathing medium (10 min) to permit the store to refill. Scale bar: 20 μM. (B) Ca²⁺ signals measured in ∼200 cells shown in A. (C) Density plot of mean peak value of Ca²⁺ signalling in Ca²⁺ free MOPS and increasing concentrations of H₂O₂. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values. (D) Summary of mean peak value of Ca²⁺ signalling in all cells. (E) Summary of number of Ca²⁺ signalling peaks in all cells n = 5, *p<0.05 vs Ca²⁺-free MOPS.

To further examine the effect of H₂O₂ on Ca²⁺ release from the store, the effects of the free radical were examined on ACh-evoked Ca²⁺ release. ACh (100 nM) evoked substantial Ca²⁺ signals that were heterogeneous across the endothelium and the amplitude and frequency of Ca²⁺ oscillations varied across cells. (Fig. 2A) (see also [2,38,44,49,52,55,81]). After washing out ACh, the endothelium was allowed to rest for 10 min and then challenged again with ACh (100 nM) and H₂O₂ (100 μM) applied simultaneously. H₂O₂ suppressed several aspects of ACh-evoked Ca²⁺ signalling. There was a reduction in the percentage of cells responding to ACh, a decrease in the amplitude, and a reduction frequency of oscillations in the presence of H₂O₂ when compared to controls (ACh alone; Fig. 2B-F). The effect of H₂O₂ on endothelial cells was also heterogeneous, and the free radical affected the Ca²⁺ response of some cells more than others’ (see Fig. 2Aiv). In the absence of H₂O₂, ACh (100 nM; 10 min. apart) evoked reproducible Ca²⁺ signals (Figure S1).

Fig. 2.

ACh-evoked Ca²⁺ release was suppressed by H₂O₂. (Ai) Pseudo-colour images (green) of Ca²⁺ signalling evoked by ACh (100 nM), and ACh (100 nM) in the presence of H₂O₂. (100 μM). Scale bar: 20 μM. (Aii) Overlaid Ca²⁺ signalling traces from ∼200 cells (shown in A) with the average shown as the black line in response to ACh and ACh+H₂O₂. Individual Ca²⁺ traces are coloured according to the magnitiude of the control (ACh) response. (Aiii) Rastergram plot of Ca²⁺ signals. Each red dot represents a Ca²⁺ peak in each cell (shown on the left axis) ACh (100 nM) left-side and ACh (100 nM) + H₂O₂ (100 μM) right-side. (Aiv) The effect of H₂O₂ varied on Ca²⁺ signals across cells. The panel shows traces of Ca²⁺ signalling from three typical cells from (Ai). The blue lines are the Ca²⁺ signals evoked by ACh (100 nM) and the red lines ACh (100 nM) + H₂O₂ (100 μM). Cell 3 was largely unaffected by H₂O₂. (B) Density plot of mean peak value of Ca²⁺ signalling from cells treated with ACh (100 nM) and ACh (100 nM) + H₂O₂ (100 μM). Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values (C) Summary of mean peak value of Ca²⁺ signalling in all cells. (D) Density plot of the frequency of Ca²⁺ signals. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values (E) Summary of frequency of Ca²⁺ oscillations in all cells. (F) Summary of percentage of ACh-responsive cells. For all summary data (D–F) n = 6; *p < 0.05.

To determine if the internal Ca²⁺ store content was altered by H₂O₂, ionomycin (2 μM, in Ca²⁺ free PSS) was applied in the absence or presence of H₂O₂ (100 μM). Ionomycin-evoked Ca²⁺ signals were not significantly altered by H₂O₂ (Fig. 3A). Two measurements were used in this analysis; the amplitude of ionomycin-induced Ca²⁺ release and the area under the curve (Fig. 3B, C). Each measure was unchanged suggesting that H₂O₂ did not deplete the internal Ca²⁺ store.

Fig. 3.

Internal Ca²⁺ store content was unchanged in H₂O₂. (A) The Ca²⁺ store content was assessed using ionomycin (2 μM) applied in a Ca²⁺-free PSS. The left panel shows the averaged ionomycin-induced Ca²⁺ transient in the absence of H₂O₂ (100 μM) while the right panel is in the presence of H₂O₂. The black lines is the mean of five independent preparations blue and red lines show the standard error of the mean (SEM). (B–C) Summary data of mean peak response (B) and area under the curve (C) of ionomycin-induced Ca²⁺ increase (n = 5).

To confirm that the suppression of ACh-evoked Ca²⁺ signalling in endothelial cells arose from H₂O₂, catalase-peg (1000 U/ml) was used to breakdown H₂O₂. Catalase by itself did not alter the Ca²⁺ signal evoked by ACh (100 nM) when compared to controls (Fig. 4A-F). Furthermore, in the presence of catalase, H₂O₂ (100 μM) did not alter the amplitude (Fig. 4B, D), or the frequency (Fig. 4C,F) of the ACh-evoked Ca²⁺ signals, nor did it alter the percentage of cells activated by ACh (Fig. 4E). These data suggest that H₂O₂ supresses ACh-evoked Ca²⁺ signals in native endothelial cells.

Fig. 4.

Catalase eliminated of the effect of H₂O₂ on ACh-evoked Ca²⁺ signalling. (Ai) Pseudo colour (green) images of Ca²⁺ signalling evoked by ACh (100 nM), ACh (100 nM) + Catalase (1000 U/ml) and ACh (100 nM) + Catalase (1000 U/ml) + H₂O₂ (100 μM). Scale bar: 20 μM. (Aii) Overlaid Ca²⁺ signalling traces from each endothelial cell with treatment of ACh (100 nM), ACh (100 nM) + Catalase (1000 U/ml) and ACh (100 nM) + Catalase (1000 U/ml) + H₂O₂ (100 μM). Individual Ca²⁺ traces are coloured according to the magnitiude of the control (ACh) response. (B) Density plot of mean peak value of Ca²⁺ signalling. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values (C) Density plot of number of the frequency Ca²⁺ signalling. (D) Summary data showing mean peak value of Ca²⁺ signalling in all cells. (E) Summary data showing the percentage of active cells. (F) Summary data showing the frequency of Ca²⁺ signals in all cells (n = 6 for all summary data).

In native endothelial cells, ACh-evoked Ca²⁺ release requires activation of IP₃Rs [1]. In support, ACh-evoked Ca²⁺ release was rapidly blocked by 2-APB (Fig. 5A-F). 2-APB significantly attenuated the amplitude (97% reduction; Fig. 5A, B, C) and frequency (99% reduction; Fig. 5E) of ACh-evoked Ca²⁺ signals, and the percentage of active cells activated by ACh (97% reduction; Fig. 5D).

Fig. 5.

2-APB inhibits ACh-evoked Ca²⁺ signalling. (Ai) Pseudo colour (green) images of Ca²⁺ signals evoked by ACh (100 nM) and ACh (100 nM) with 2-APB (100 μM). Scale bar: 20 μM. (Aii) Overlaid Ca²⁺ signalling traces of ACh (100 nM; left) and ACh (100 nM) with 2-APB (100 μM; right). Individual Ca²⁺ traces are coloured according to the magnitiude of the control (ACh) response. (B) Density plot of mean peak value of Ca²⁺ signalling. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values. (C) Summary data of the mean peak value of Ca²⁺ signalling (E) percentage of active cells and (D) number of peaks of Ca²⁺ signalling. For all summary data (D–E), n = 5, * p<0.05.

To determine which part of the IP₃ pathway Ca²⁺ release was modified by H₂O₂, we performed experiments using the membrane-permeant, photoactivateable form of IP₃ (caged-IP₃5 μM). IP₃, released via the photolysis of caged-IP₃, directly activates IP₃Rs [13,51] and evoked a Ca²⁺ response (Fig. 6A-E). H₂O₂ (100 μM) significantly attenuated the Ca²⁺ response to photolysis of caged-IP₃ (22% reduction; Fig. 6C-E). Again, there was heterogeneity in the sensitivity to H₂O₂ and some cells were less affected than others (Fig. 6C). These results demonstrate that H₂O₂ reduces ACh-evoked Ca²⁺ release by altering either the activity of IP₃ receptors or the interaction between IP₃ and IP₃Rs.

Fig. 6.

H₂O₂ suppresses IP₃ receptor activity. (A) Representative image showing endothelial cells of an intact artery. The red circle demarks the area that was selected for photolysis of caged IP₃. Scale bar: 20 μM. (B) Heat map of Ca²⁺ signalling stimulated by photolysis of caged IP₃ in the preselected area. (C) Uncaging IP₃-evoked Ca²⁺ signalling traces of two representative single cells (cell 1 and cell 2 from B) before and after 20 min incubation of H₂O₂ (100 μM). As with ACh there was heterogeneity in the responses of cells to H₂O₂. The peak value of the Ca²⁺ signal was either unaltered (cell 2) or suppressed (cell 1) by H₂O₂. (D) Summary data showing the Ca²⁺ signal peak after uncaging. Control: blue; H₂O₂ (100 μM): red (n = 5). (E) Averaged peak value of Ca²⁺ signals from five different experiments. Control: blue; H₂O₂ treated: red, (n = 5; *p < 0.05).

Since H₂O₂ is reported to increase the activity of IP₃Rs [11], the question arises as to how H₂O₂ is able to decrease IP₃-evoked Ca²⁺ release. Mitochondria exert profound control of IP₃-evoked Ca²⁺ release [58,80] and H₂O₂ has been shown to alter mitochondrial function [56]. These observations raise the possibility that H₂O₂ may exert effects on IP₃R indirectly. To determine if mitochondria mediate the effects of H₂O₂, we investigated the effect of uncoupling mitochondria on Ca²⁺ release from the internal Ca²⁺ store. To do this, the uncoupler, CCCP, and the complex I inhibitor, rotenone, were used in separate experiments. Each drug was used in combination with the ATP synthase inhibitor oligomycin, to prevent reversal of the ATP synthase. CCCP (5 μM) and oligomycin (6 μM) inhibited ACh-evoked Ca²⁺ signalling (Fig. 7A, B & Figure S2); the inhibition remained even after CCCP and oligomycin wash out (Fig. 7A, B). CCCP and oligomycin significantly reduced the amplitude and frequency of ACh-evoked Ca²⁺ signals, and the percentage of cells activated by ACh (Fig. 7C-F). Similarly, rotenone (2 μM) and oligomycin (6 μM) also inhibited ACh-evoked Ca²⁺ signaling (Fig. 8A-F). These results demonstrate that mitochondria regulate IP₃-mediated Ca²⁺ release and that mitochondrial membrane potential depolarization inhibits IP₃-evoked Ca²⁺ release.

Fig. 7.

The uncoupler, CCCP, and ATP synthase blocker, oligomycin, inhibited ACh-evoked Ca²⁺ signals. (Ai) Pseudo colour (green) images of Ca²⁺ activity evoked by ACh (100 nM; left), ACh (100 nM) + CCCP (5 μM) + oligomycin (6 μM; middle), and after washout of the CCCP and oligomycin (right). Scale bar: 20 μM. (Aii) Overlaid single cell Ca²⁺ traces from the endothelial cells shown in Aii. Individual Ca²⁺ traces are coloured according to the magnitiude of the control (ACh) response. (B) Density plot of mean peak value of the Ca²⁺ signal. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values (C) Density plot of the frequency of Ca²⁺ signals. (D) Summary of the mean peak value of Ca²⁺ signals in all cells, (E) Percentage of active cells (F) and the frequency of signals in all cells. For all summary data (D–F), n = 3, *p<0.05.

Fig. 8.

Inhibition of complex I blocked ACh-evoked Ca²⁺ signalling. (Ai) Pseudo colour (green) images of Ca²⁺ signals evoked by ACh (100 nM), ACh (100 nM) + Rotenone (2 μM) + oligomycin (6 μM) and after washout of rotenone and oligomycin. Scale bar: 20 μM. (Aii) Overlaid Ca²⁺ signalling traces, from the cells shown in Ai, evoked by ACh (100 nM), ACh (100 nM) + Rotenone (2 μM) + oligomycin (6 μM) and after washout of rotenone and oligomycin. (B) Density plot of mean peak value of Ca²⁺ signalling. Individual data points have been coloured (from blue, low to red, high) according to the density (i.e. occurrence) of particular values (C) Density plot of number of peaks of Ca²⁺ signalling. (D) Summary of the mean peak value of Ca²⁺ signals in all cells, (E) Percentage of active cells (F) and the frequency of signals in all cells. For all summary data (D–F), n = 3, *p<0.05.

To explore the role H₂O₂ plays in mitochondria-regulated Ca²⁺ signaling, mitochondrial membrane potential was assessed using TMRE. TMRE is a lipophilic cation that is rapidly sequestered by the negatively-charged (∼−180 mV) mitochondrial membrane potential [21]. TMRE was imaged for 30 min to ensure the stability of the indicator. After 30 min, H₂O₂ (100 μM) was introduced and TMRE imaged for a further 30 min. H₂O₂ caused a significant decrease in TMRE fluorescence intensity (Fig. 9A, B). These findings suggest that H₂O₂ may suppress IP₃-evoked Ca²⁺ release by depolarizing mitochondria. In a control experiments, to confirm mitochondrial localization, TMRE (60 nM) was loaded together with mitotracker green (100 nM). The two mitochondrial indicators largely overlapped in their localization (Figure S3). As expected from a mitochondrial localization of the dyes, mitochondrial membrane potential depolarization with CCCP (5 μM; applied with oligomycin (6 μM)) dispersed punctuate TMRE staining and reduced mitotracker green labelling (Figure S3).

Fig. 9.

Depolarisation of mitochondria by H₂O₂. (A) Pseudo colour images (red) of the mitochondrial membrane potential (reported by the membrane potential sensitive dye TMRE). In control (0 min ∼ 30 min) and after H₂O₂ (100 μM; 30 min–60 min). H₂O₂ decrease the mitochondrial membrane potential as revealed by the decrease in mitochondrial TMRE fluorescence intensity. Scale bar: 20 μM. (B) Quantification of the normalized intensity of TMRE (n = 6). *p < 0.05.

4. Discussion

Interaction between the internal Ca²⁺ store and mitochondrial are critical in regulating cell signalling and cell performance. Several diffusible mediators communicate between the two organelles to control cell and tissue function. Of these, Ca²⁺ and H₂O₂ are of particular significance. Mitochondria are a major source, and the internal Ca²⁺ store a target, for H₂O₂. H₂O₂ modulates Ca²⁺ transport mechanisms on the internal Ca²⁺ store [5,12,24,60] and in turn, Ca²⁺ release from the store modulates mitochondrial function by regulating the enzymes of the Krebs cycle and oxidative phosphorylation [41]. Ca²⁺-induced changes in metabolic rate result in altered oxygen consumption, respiratory chain electron leakage and H₂O₂ levels [14]. Here, we have demonstrated H₂O₂ depolarises mitochondria and inhibits spontaneous and agonist-evoked IP₃-induced Ca²⁺ signals. We suggest that suppression of IP₃-evoked Ca²⁺ release arises from a H₂O₂-induced the decrease in mitochondrial membrane potential (Fig. 10).

Fig. 10.

Proposed mechanism for mitochondria regulation Ca²⁺ signalling in endothelial cells. The mitochondrial membrane potential is depolarized by H₂O₂. The depolarized membrane potential limits the driving force for Ca²⁺ uptake by mitochondria. As result, the concentration of Ca²⁺ in microdomain area between mitochondria and endoplasmic reticulum is increased and the ion suppresses Ca²⁺ release through IP₃ receptor from the endoplasmic reticulum. Limited Ca²⁺ endothelial and artery function. On the other hand, reduced uptake of Ca²⁺ into mitochondria will lead to less production of ATP, therefore, reduced the metabolic production of H₂O₂. The negative feedback loop may help endothelial cells maintain normal metabolism and cell survival.

Mitochondria are potent modulators of IP₃-evoked Ca²⁺ release. Ca²⁺ uptake by mitochondria may promote Ca²⁺ release from IP₃Rs [18,19,23,43,54,65,72,73,78], limit IP₃-evoked Ca²⁺ signals [4,34] or slow IP₃-evoked Ca²⁺ wave progression [10,15,32,61,69,84]. At least two mechanisms have been proposed to account for mitochondrial control of IP₃R activity. First, at sites of close contact between the internal Ca²⁺ store and mitochondria [26,48], channels on the internal Ca²⁺ store and mitochondrial channels (e.g. the uniporter and voltage-dependent anion-selective channel) may cluster, and Ca²⁺ uptake into mitochondria occurs at these sites [25,33,63,64]. Ca²⁺ uptake depends critically on the mitochondrial membrane potential. As the membrane potential decreases, so does mitochondrial Ca²⁺ uptake. Mitochondrial Ca²⁺ uptake limits a negative feedback process that operates at IP₃ receptors to maintain Ca²⁺ release [19,54]. In smooth muscle, mitochondrial Ca²⁺ uptake is fast enough to regulate local spontaneous Ca²⁺ signals arising from IP₃Rs (Ca²⁺ puffs) [57] and regulates store-operated Ca²⁺ entry [59], demonstrating tight functional coupling between IP₃Rs and mitochondria. However, in other studies, close coupling between the internal store and mitochondria was not required for mitochondrial control of Ca²⁺ release to occur [80]. In the second mechanism proposed to account for mitochondrial control of the internal store, mitochondrial ATP production modulates Ca²⁺ release. When ATP production was restricted Ca²⁺ release was inhibited. This mechanism permits mitochondria to control Ca²⁺ release while being positioned far from the internal Ca²⁺ store [80]. Several studies show that ATP maintains IP₃-mediated Ca²⁺ release. ATP potentiates IP₃-induced Ca²⁺ release in permeabilized cells and from native endoplasmic reticulum vesicles, and it enhances activation of IP₃-gated channels and purified, reconstituted IP₃Rs [28,39,50,66] by increasing the open time of the channel s[7]. Thus factors provided by mitochondria may diffuse to IP₃R to maintain IP₃R activity.

The present results highlight another control mechanism which may operate between mitochondria and the internal Ca²⁺ store i.e. diffusion of H₂O₂. Our results suggest that the control of IP₃ release by H₂O₂ is indirect and mediated by depolarization of mitochondria so combines aspects from both proposals (diffusible substance and mitochondrial depolarization). H₂O₂-mediated depolarization of the mitochondrial membrane potential will decrease the driving force for Ca²⁺ uptake by mitochondria and so limit negative feedback control of IP₃-evoked Ca²⁺ release. Several proposals exist for the mechanisms by which H₂O₂ may depolarize mitochondria [30,31] and include inhibition of alpha-ketoglutarate dehydrogenase [22,56] or succinate dehydrogenase [56], or activation of the permeability transition pore [83]. Interestingly, the effect of H₂O₂ was not homogeneous across all cells and some were affected more than others. The reason for the heterogeneity is not clear, but perhaps differences in basal levels of H₂O₂, or metabolic state of the cells may contribute. Alternatively, antioxidant enzymes whose activities are directed at reducing hydrogen peroxide, such as catalase, glutathione peroxidase, and thioredoxin peroxidase, may vary across cells.

While the present results suggest that H₂O₂ inhibition of IP₃R is indirect and mediated via mitochondria, ROS may also directly alter IP₃ receptor activity. However, H₂O₂ is often reported to promote rather than inhibit IP₃R activity. When H₂O₂ transients were prevented, Ca²⁺ oscillations were inhibited in some cells, implying that H₂O₂ makes subsequent Ca²⁺ release more likely to occur [11]. In support of these observations, in various cultured endothelial cell lines (HAECs, LECs, HUVECs and CVECs), H₂O₂ induces Ca²⁺ release from the internal store [27,37,77]. Various exogenously added oxidants stimulate rather than inhibit IP₃R-mediated Ca²⁺release. Thimerosal [12,16,42], t-butylhydroperoxide [9], and diamide [45,46] each increase IP₃-evoked Ca²⁺ release. In the case of thimerosal, the proposed mechanism involves an increased sensitivity of the receptor to [IP₃] that results in Ca²⁺oscillations occurring at basal [IP₃] in unstimulated cells [5,40]. Although sensitization to IP₃may be a general mechanism responsible for the action of other oxidants, it has also been suggested that they (oxidants) may alter the sensitivity of IP₃R to Ca²⁺ [9,45]. Our results suggest that the indirect effect of H₂O₂ on endothelial mitochondria may dominate control of IP₃R. Since H₂O₂ may increase IP₃R activity, these findings buttress the proposal that inhibition of Ca²⁺ release by H₂O₂ is indirect.

H₂O₂ regulates several key modulators of cell activity including cell proliferation, migration, and differentiation, and because H₂O₂ is membrane permeable, the ROS may exert widespread control across many endothelial cells to regulate cardiovascular function. H₂O₂ mediates at least part of its effects through changes in Ca²⁺ signalling. Ca²⁺ influx in native and in cultured endothelial cell lines is evoked by H₂O₂. In the native endothelium of superior epigastric arteries, H₂O₂ evoked Ca²⁺ increases result from Ca²⁺ influx via TRPV4 channels [67]. In mouse and human lung microvascular endothelial cell lines (MLMVEC, HPAE, H5V and HLMVEC) H₂O₂ evokes Ca²⁺ influx via TRPV4 or TRPM2 [70,71]. Our results reveal an additional complexity in the effects of H₂O₂. H₂O₂ may inhibit IP₃-evoked Ca²⁺ release in native endothelial cells. The inhibition of IP₃ evoked Ca²⁺ by H₂O₂ may be an indirect and mediated via depolarization of the mitochondrial membrane potential. This process may serve as a negative feedback modulation of mitochondrial function. Ca²⁺ increases associated with cell activity may initially stimulate mitochondrial ATP production. However, an increase in electron transport activity will result in increased ROS production, with a decrease in mitochondrial membrane potential and inhibition of IP₃-evoked Ca²⁺ release as a result. H⁠₂O⁠₂ mediated mitochondrial depolarization may be a mechanism by which the ogranelles inhibit IP⁠₃-evoked Ca²⁺ signalling to protect themselves against Ca^⁠2+ overload .

Author contributions

XZ, MDL,CW & JGM developed the concept. XZ performed the experiments. JGM & XZ drafted the manuscript. JGM, ZX, CW & MDL edited and revised the manuscript. CW & JGM sourced funding. All authors approved the final version of the manuscript.

Declaration of Competing Interest

None.

Appendix A. Supplementary data

The following is Supplementary data to this article: