Activation of Store-Operated I_CRAC by Hydrogen Peroxide

SUMMARY

Reactive oxygen species such as hydrogen peroxide (H₂O₂) play a role in both innate immunity as well as cellular injury. H₂O₂ induces changes in intracellular calcium ([Ca²⁺]_i) in many cell types and this seems to be at least partially mediated by transient receptor potential melastatin 2 (TRPM2) in cells that express this channel. Here we show that low concentrations of H₂O₂ induce the activation of the Ca²⁺-release activated Ca²⁺ current I_CRAC. This effect is not mediated by direct CRAC channel activation, since H₂O₂ does not activate heterologously expressed CRAC channels independently of stromal interaction molecule (STIM). Instead, I_CRAC activation is partially mediated by store depletion through activation of inositol 1,4,5 trisphosphate receptors (IP₃R), since pharmacological inhibition of IP₃ receptors by heparin or molecular knock-out of all IP₃ receptors in DT40 B cells strongly reduce H₂O₂-induced I_CRAC. The remainder of H₂O₂-induced I_CRAC activation is likely mediated by IP₃R-independent store-depletion. Our data suggest that H₂O₂ can activate Ca²⁺ entry through TRPM2 as well as store-operated CRAC channels, thereby adding a new facet to ROS-induced Ca²⁺ signaling.

INTRODUCTION

Reactive oxygen species (ROS) are a group of molecules and ions with the potential of causing cellular damage due to their highly reactive characteristics. Hydrogen peroxide (H₂O₂), an oxidizing agent and member of ROS, is often used in experimental models of oxidative stress, although accumulating evidence indicates that H₂O₂ may also function as an important signaling molecule in diverse cellular processes such as cell development, proliferation, signal transduction and protein regulation [1]. Thus, ROS and intracellular Ca²⁺ have shown interdependent relations in numerous processes [2], thereby linking ROS to one of the most diverse second messengers in the cell. H₂O₂ is constantly produced in the cell as a byproduct of aerobic metabolism in the mitochondria. To prevent toxic ROS overload, several enzymatic mechanisms will catalyze the conversion of H₂O₂ to water and oxygen, including peroxisomal catalase and cytosolic peroxiredoxins and glutathione peroxidase. For a long time, H₂O₂ was believed to freely cross membranes, but more recent evidence suggests that the membrane permeability is regulated by the composition of the membrane as well as diffusion facilitation through aquaporins [3].

The broad effects of H₂O₂ on enzymes, growth factors, transcription factors and ion channels are believed to mainly occur through redox modification of reactive thiol groups in cysteine residues [4]. Although various ion channels are modulated by ROS [4, 5], the only known ion channel that can be gated through actions of H₂O₂ is the Ca²⁺ permeable non-selective cation channel TRPM2 (transient-receptor potential melastatin-2) [6–8]. The channel’s primary agonist is adenosine diphosphoribose (ADPR) [9, 10]. Intracellular Ca²⁺, H₂O₂ or related nucleotides such as cyclic adenosine diphosphoribose (cADPR) and nicotinic acid adenine dinucleotide phosphate (NAADP) have limited if any direct gating activities [7, 11], although they can synergize with ADPR to activate TRPM2 currents at lower ADPR concentrations [7, 12]. For H₂O₂-mediated activation of TRPM2 both ADPR-dependent and -independent mechanisms have been suggested [8, 13]. Additional channel targets of H₂O₂ include the ryanodine receptor (RyR), which functions as a Ca²⁺-release channel in muscle cells and neurons [14] and an outwardly rectifying cation current, I_OGD, in murine cortical neurons, presumably involving TRPM7 channels [15]. Additionally, H₂O₂ at millimolar concentration can induce a sustained nonselective cation current, I_LiNC, independent of ion channels [16].

The major route of Ca²⁺ entry in nonexcitable cells is via store-operated channels (SOC). The best characterized SOC is the Ca²⁺ release-activated Ca²⁺ channel (CRAC), a highly selective low-conductance Ca²⁺ channel, which is activated by depletion of internal Ca²⁺ stores [17]. Molecularly, this involves stromal interaction molecule 1 (STIM1), which senses ER Ca²⁺ levels [18, 19] and upon tore depletion activates the plasma membrane channel CRACM1 (also called Orai1) [20–22]. Store-operated Ca²⁺ entry (SOCE) is a widely expressed mechanism in many cells that respond to ROS, however, little is known about effects of ROS on the store-operated CRAC current (I_CRAC). We set out to assess possible effects of ROS on I_CRAC using calcium imaging and whole-cell electrophysiology. We report that extracellular application or intracellular perfusion of H₂O₂ at micromolar concentrations activates I_CRAC in several native cell lines independent of the presence or absence of TRPM2-like currents.

METHODS

Cell culture

Cells were incubated at 37°C with 5% CO₂ in the appropriate cell media. Tetracycline-inducible HEK293 TRPM2-expressing cells were cultured in DMEM with 10% fetal bovine serum (FBS) supplemented with blasticidin (5 µg/ml, Invitrogen) and zeocin (0.4 mg/ml, Invitrogen) [9], RBL-2H3 cells in DMEM with 10% FBS, Jurkat T-lymphocytes in RPMI 1640 with 10% FBS, HEK293 CRACM1-overexpressing cells in DMEM with 10% FBS [23], and DT40 B-lymphocytes in RPMI 1640 with 10 % FBS supplemented with 5% chicken serum and 2 mM L-Glutamine. For induction of TRPM2 expression, HEK293 cells were resuspended in medium containing 1 µg/ml tetracycline (Invitrogen) 4–8 hrs before experiments.

Solutions and chemicals

For fluorescence and patch-clamp measurements cells were kept in standard extracellular saline solution (in mM): 140 NaCl, 2.8 KCl, 2 MgCl₂, 11 glucose, 10 HEPES-NaOH. CaCl₂ concentration was 1 mM in calcium imaging experiments and 10 mM in patch-clamp experiments unless otherwise stated. Neutral pH was between 7.2 and 7.3, adjusted with NaOH and osmolarity was 300 mOsm. For the TRPM2 inhibition panel the divalent and trivalent cations (Ba²⁺, Cd²⁺, Co²⁺, Cu²⁺, Ni²⁺, Sr²⁺, Zn²⁺, La³⁺) were added as chloride salts into the standard extracellular solution at 1 mM. In some experiments 1 µM LaCl₃ was included in the extracellular solution. H₂O₂ (30% stock solution) was added in various concentrations to extra- and intracellular solutions.

Standard pipette filling solutions contained (in mM): 140 Cs-glutamate, 8 NaCl, 1 MgCl₂, 10 HEPES-CsOH. In patch-clamp recordings with DT40 cells MgCl₂ concentration was increased to 3 mM to inhibit endogenous TRPM7 currents. Ca²⁺ concentration was buffered to 150 nM, 200 nM or 500 nM in whole-cell experiments, calculated with WebMaxC (http://www.stanford.edu/~cpatton/webmaxc/webmaxcS.htm), with 10 mM Cs-BAPTA and 4 mM, 4.55 mM or 6.9 mM CaCl₂, respectively, or left unbuffered. ADPR was added to the pipette solution in TRPM2 experiments. In some experiments heparin was added to the pipette solution. IP₃ and thapsigargin (dissolved in DMSO) was used in some experiments to activate I_CRAC. All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), except H₂O₂ (Fisher Scientific, Hampton, NH, USA).

Electrophysiology

Patch-clamp experiments were performed in whole-cell configuration at 21–25°C. Patch pipettes were pulled from glass capillaries (Kimble Products, Fisher Scientific, USA) on a DMZ-Universal Puller (DAGAN, Minneapolis, MN) with pipette resistances in the range of 2 to 3 MΩ. Data were acquired with Patchmaster software controlling an EPC-10 amplifier (HEKA, Lambrecht, Germany). Voltage ramps of 50 ms spanning the voltage range of −100 mV to 100 mV were applied from a holding potential of 0 mV at a rate of 0.5 Hz, unless otherwise stated. Experiments were recorded typically over periods of 100–600 s. Voltages were corrected for liquid junction potentials of 10 mV. Capacitive currents and series resistances were determined and corrected before each voltage ramp. The low-resolution temporal development of currents for a given potential was extracted from the leak-corrected individual ramp currents by measuring current amplitudes at voltages of −80 mV and +80 mV, unless otherwise stated.

Single channel measurements

TRPM2 single channel recordings in Jurkat T cells were performed in the whole-cell configuration. A ramp from −100 to +10 mV over 11 s was applied continuously. Currents were filtered at 50 Hz. Linear fits were performed from −100 to 0 mV in Igor Pro (Wavemetrics, Oregon, USA) to determine the single channel conductance b according to a line function a + bx.

Fluorescence measurements

For measurement of cytosolic Ca²⁺ concentration cells were loaded with 5 µM Fura-2 AM (acetoxymethylester, Molecular Probes) for 30 min in 37°C, washed and kept in external solution. Experiments were performed with an Olympus BX2 fluorescence microscope equipped with a dual excitation fluorometric imaging system (TILL-Photonics). Data acquisition and computation was controlled by TILLvisION software. Dye-loaded cells were excited by wavelengths of 340 and 380 nm for 20 ms each, produced by a monochromator (Polychrome IV). The fluorescence emission of several single cell bodies was simultaneously recorded with a video camera (TILL-Photonics Imago) with an optical 440 nm long-pass filter. The signals were sampled at 0.5 Hz and computed into relative ratio units of the fluorescence intensity at the different wavelengths (340/380 nm).

Subcloning and overexpression

Full length human STIM1 was subcloned as described earlier [24]. For electrophysiological analysis, STIM1 proteins were over-expressed in HEK293 cells stably expressing CRACM1 [25] using lipofectamine 2000 (Invitrogen) and the GFP expressing cells were selected by fluorescence. Experiments were performed 24 – 48 hours post transfection.

Data Analysis

Data was analyzed with FitMaster (HEKA, Lambrecht, Germany) and Igor Pro (WaveMetrics, Oregon, USA). Where applicable, statistical errors of averaged data are given as means ± SEM with n determinations. Single ramps were plotted as current-voltage relationships. Currents were normalized to cell size in pF, unless otherwise stated. Half-maximal activation time was calculated with a minimum-maximum dose-response function, f(t) = Y_min+(Y_max-Y_min)×(1/(1+(K_d/t)ⁿ), where Y_min is current basal level, Y_max is the current plateau phase, K_d is half-maximal activation time and n is an integer. Statistical calculations were performed using Igor Pro. Significance levels were determined by student’s t-test. P < 0.05 was considered to be significant.

RESULTS

TRPM2 currents are not significantly affected by divalent or trivalent ions

H₂O₂ is a well-known activator of TRPM2 currents [6–8]. We pursued two strategies to differentiate H₂O₂-induced Ca²⁺ influx through TRPM2 from other Ca²⁺ entry pathways, including I_CRAC. The first strategy involved the use of cell lines that differentially express I_CRAC and TRPM2 and the second was based on differential pharmacological inhibition. Known blockers of TRPM2, such as 2-amino-ethoxydiphenyl borate (2-APB), flufenamic acid or clotrimazole [26, 27] also interfere with store-operated calcium entry [28–30]. Lanthanum (La³⁺), however, potently blocks I_CRAC [31], but reportedly does not inhibit TPRM2 activity [32]. We tested various divalent cations and La³⁺ for inhibitory actions on TRPM2 currents in whole-cell patch-clamp experiments in HEK293 cells stably expressing TRPM2 under a tetracycline-inducible promoter. Cells were kept in a standard NaCl-based extracellular solution with 1 mM CaCl₂ and perfused with a standard Cs-glutamate-based intracellular solution supplemented with 500 µM ADPR (see methods). Within seconds of whole-cell break-in, large currents developed (Fig. 1A) with the typical linear current-voltage (I/V) relationship of TRPM2 (Fig. 1B), reaching several nA at −80 mV. In order to measure the inhibition of TRPM2, currents were allowed to reach the plateau phase before application of 1 mM of various divalent cations or La³⁺ at 60 s. When applying Ba²⁺, Cu²⁺, Ni²⁺, Cd²⁺, Sr²⁺, or Co²⁺ current reduction was less than 5% and Zn²⁺ and La³⁺ reduced currents by just 5–7% (Fig. 1C). In each case, the current reduction was reversible as seen by the increase in current following removal of the cations at 120 s. This demonstrates that TRPM2 is not markedly affected by divalent ions or La³⁺, in marked contrast to many other ion channels, including I_CRAC [31].

Figure 1. Inhibition of TRPM2 current by various divalent and trivalent cations.

A) Average TRPM2 currents in tetracycline-induced HEK293 cells (1 µg/ml, 4–8 hrs) perfused with Cs-glutamate based pipette solution containing 500 µM ADPR with unbuffered internal Ca²⁺. Currents were normalized to the data point prior to cation application. The bar indicates application of 1 mM of each cation (n=5 for each condition), displayed is the Ba²⁺ and La³⁺ trace. Currents were measured with a voltage ramp from −100 to +100 mV over 50 ms at 0.5 Hz intervals from a holding potential of 0 mV. Inward currents were extracted at −80 mV, averaged and plotted versus time. B) Representative current-voltage (I/V) relationship of ADPR-activated TRPM2 current in HEK293 cells at t = 60 s and upon application of 1 mM La³⁺ at t = 120 s. C) Degree of inhibition of TRPM2 by 1 mM of various di- and trivalent ions. Values represent mean (± SEM) over the 60 s period of application.

H₂O₂ causes Ca²⁺ release and Ca²⁺ influx independent of TRPM2

Using Fura-2 calcium imaging, we assessed Ca²⁺-permeable ion channels as candidates for H₂O₂-mediated Ca²⁺ signals in various cell types that express I_CRAC with or without TRPM2. Jurkat T lymphocytes express both I_CRAC and TRPM2 natively [11, 33], whereas RBL-2H3 (Fig. S1A; [34]) and wild-type (WT) HEK293 express only I_CRAC [9, 21]. In the latter cell type, we additionally overexpressed TRPM2 heterologously. Cells were bathed in standard external NaCl-based bath solution with 1 mM Ca²⁺ and then exposed for 3 min to 100 µM H₂O₂ in Ca²⁺-free bath solution to reveal internal Ca²⁺ release, followed by readmission of 1 mM Ca²⁺ to reveal possible Ca²⁺ entry. In RBL-2H3 cells, H₂O₂ induced both Ca²⁺ release and a relatively small Ca²⁺ influx component when returning to 1 mM Ca²⁺ (Fig. 2A). Since RBL-2H3 cells do not produce ADPR-induced TRPM2 currents (Fig. S1A) and H₂O₂ caused large Ca²⁺ release, we assessed Ca²⁺ entry through I_CRAC by performing identical experiments in the presence of 1 µM La³⁺, a highly potent blocker of I_CRAC [31]. This revealed that a portion of the Ca²⁺ signal observed after readmission of Ca²⁺ was mediated by La³⁺-sensitive Ca²⁺ influx (Fig. 2A).

Figure 2. H₂O₂ induces Ca²⁺ response in RBL-2H3 cells, HEK293 cells and Jurkat T cells.

A) Averaged changes in [Ca²⁺]_i measured as ratios of Fura-2 fluorescence excited at 340 and 380 nm in Fura-2 AM loaded RBL-2H3 cells in response to H₂O₂. Cells were kept in a saline solution with 1 mM CaCl₂ in the absence (blue trace, n=34) or presence of 1 µM LaCl₃ (red trace, n=35). Black bar indicates application of 100 µM H₂O₂ in nominally Ca²⁺ free saline solution. In controls H₂O₂ was omitted from the Ca²⁺ free saline solution (black trace, n=57) and CaCl₂ was omitted from the readmission solution (green trace, n=42). Cells were loaded with 5 µM Fura-2 AM at 37° C for 30 min. Traces are representative of three independent experiments performed on different days. B) Experimental protocol as in A) but for Fura-2 AM loaded HEK293 WT cells in the absence (blue trace, n=24) or presence of 1 µM LaCl₃ (red trace, n=6). Black bar indicates application of 20 µM H₂O₂ in Ca²⁺ free saline solution. In controls H₂O₂ was omitted from the Ca²⁺ free saline solution (black trace, n=28) and CaCl₂ was omitted from the readmission solution (green trace, n=52). Traces are representative of three independent experiments performed on different days. C) Experimental protocol as in A) but for Fura-2 AM loaded tetracycline-induced (1 µg/ml, 21–22 hrs) HEK293 cells expressing TRPM2 in the absence (blue trace, n=16) or presence of 1 µM LaCl₃ (red trace, n=13). Black bar indicates application of 20 µM H₂O₂ in Ca²⁺ free saline solution. In controls H₂O₂ was omitted from the Ca²⁺ free saline solution (black trace, n=9) and CaCl₂ was omitted from the readmission solution (green trace, n=28). D) Experimental protocol as in A) but for Fura-2 AM loaded Jurkat T cells in the absence (blue trace, n=74) or presence of 1 µM LaCl₃ (red trace, n=126). Black bar indicates application of 20 µM H₂O₂ in nominally Ca²⁺ free saline solution. In controls H₂O₂ was omitted from the Ca²⁺ free saline solution (black trace, n=49) and CaCl₂ was omitted from the readmission solution (green trace, n=15). Traces are representative of three independent experiments performed on different days.

This was confirmed in HEK293 WT cells, where H₂O₂-induced Ca²⁺ release was followed by a Ca²⁺ signal upon Ca²⁺ readmission that could also be inhibited by 1 µM LaCl₃ (Fig. 2B). It should be noted that in HEK293 WT cells 20 µM H₂O₂ was sufficient to evoke reliable Ca²⁺ responses, whereas in RBL-2H3 cells the H₂O₂ concentration had to be increased to 100 µM to reliably induce Ca²⁺ entry upon Ca²⁺ readmission. We next investigated Ca²⁺ release and Ca²⁺ influx in tetracycline-induced HEK293-TRPM2 cells. Applying 20 µM H₂O₂ resulted in Ca²⁺ release and substantial Ca²⁺ influx both in the absence and presence of 1 µM LaCl₃ (Fig. 2C). This confirms that H₂O₂ can activate TRPM2 and that La³⁺ fails to block TRPM2.

We also investigated Jurkat T cells, which express TRPM2 and CRAC channels natively. Here, application of 20 µM H₂O₂ revealed Ca²⁺ release as well as Ca²⁺ influx when returning to 1 mM Ca²⁺ (Fig. 2D). To assess whether the Ca²⁺ entry upon readmission of Ca²⁺ involved TRPM2 activity, we performed the same experiment in the presence of 1 µM LaCl₃, which has negligible inhibitory effects on TRPM2 (as seen in Fig. 1 and Fig. 2C). Again H₂O₂ induced a Ca²⁺ release, but Ca²⁺ entry was completely blocked (Fig. 2D), demonstrating that in Jurkat T cells the H₂O₂-mediated Ca²⁺ influx caused by 20 µM H₂O₂ was not related to TRPM2.

For all cell types investigated, the H₂O₂-induced Ca²⁺ signals trended back to the baseline in control experiments without Ca²⁺ readmission, but did not return entirely. The incomplete return to baseline may be due to an additional effect of H₂O₂ on the plasma membrane Ca²⁺ ATPase (PMCA). It has been previously reported that La³⁺ can cause an apparent prolongation of the release response due to a block of PMCA [35]. However, our data would imply that H₂O₂ is less potent than 1 µM La³⁺ in blocking PMCA. Our control data also provides evidence that Ca²⁺ entry through I_CRAC is more significant than the difference between the H₂O₂ and H₂O₂ + La³⁺ traces suggests. Together, these findings indicate that extracellular application of H₂O₂ at low micromolar concentrations can activate a Ca²⁺ influx pathway in Jurkat T cells, HEK293 WT cells and RBL-2H3 cells that is not related to TRPM2 activity.

H₂O₂ induces I_CRAC in RBL-2H3 and Jurkat T cells

To determine whether the H₂O₂-activated Ca²⁺ influx observed in Ca²⁺ imaging experiments was mediated by I_CRAC, we performed whole-cell patch-clamp experiments. We used standard experimental conditions optimized for measuring I_CRAC, i.e. NaCl-based extracellular solution with 10 mM CaCl₂ and Cs-glutamate-based intracellular solution buffered to 150 nM free Ca²⁺ with Cs-BAPTA (see methods). In whole-cell patch-clamp experiments with RBL-2H3 cells, 10 mM CsCl was included in the bath solution to inhibit the inward-rectifier potassium current. We first tested for I_CRAC activation in RBL-2H3 cells by extracellular application of 40 µM H₂O₂ from a wide-tipped puffer pipette 1 min after break-in. As shown in Fig. 3A, H₂O₂ application activated an inward current in 12 out of 14 cells, reaching a plateau of −2.5 ± 0.3 pA/pF at 600 s with a half-maximal activation time of 281 ± 1.2 s. No current developed in two cells. The current exhibited inward rectification that is typical for I_CRAC [17] as illustrated in a representative I/V relationship (Fig. 3B). When including 1 µM LaCl₃ in the bath and application solution no inward current developed (Fig. 3A). In control experiments without H₂O₂ application, no significant inward currents developed (Fig. 3A).

Figure 3. H₂O₂ activates I_CRAC in RBL-2H3 cells and Jurkat T cells.

A) Average current development in RBL-2H3 cells induced by extracellular application of H₂O₂ in the absence (filled circles, n=12) or presence of 1 µM LaCl₃ in the bath solution (open circles, n=5). In control experiments H₂O₂ was omitted (open triangles, n=9). Experiments were performed using Cs-glutamate based pipette solution with [Ca²⁺]_i clamped to 150 nM free Ca²⁺ with 10 mM Cs-BAPTA. The bar indicates application of 40 µM H₂O₂. Currents were measured with a voltage ramp from −100 to +100 mV over 50 ms at 0.5 Hz intervals from a holding potential of 0 mV. Inward currents were extracted at −80 mV and plotted versus time. Error bars indicate SEM. B) Representative I/V relationship of H₂O₂-activated CRAC current in RBL-2H3 cells at t = 600 s in the absence (black trace) or presence of 1 µM LaCl₃ (grey trace). C) Average current development in Jurkat T cells by extracellular application of H₂O₂ in the absence (filled circles, n=5) or presence of 1 µM LaCl₃ in the bath solution (open circles, n=9). In control experiments H₂O₂ was omitted (open triangles, n=5). [Ca²⁺]_i was clamped to 200 nM free Ca²⁺ with 10 mM Cs-BAPTA and [Mg²⁺]_i was 3 mM. The bar indicates application of 80 µM H₂O₂. Currents were analyzed as in A). D) Representative I/V relationship of H₂O₂-activated CRAC current in Jurkat T cells at t = 600 s in the absence (black trace) or presence of 1 µM LaCl₃ (grey trace). E) Average current density of I_CRAC extracted at −80 mV in RBL-2H3 cells perfused with 20 µM IP₃ (n=5). The bar indicates application of 40 µM H₂O₂. F) Average current density of I_CRAC extracted at −80 mV in Jurkat T cells perfused with 20 µM IP₃ (n=5). The bar indicates application of 40 µM H₂O₂.

Similar experiments were performed using Jurkat T cells. To effectively suppress spontaneous activation of I_CRAC in these cells, [Ca²⁺]_i was buffered to 200 nM and [Mg²⁺]_i was increased to 3 mM to suppress activation of endogenous TRPM7 currents. Application of 40 µM H₂O₂ did not activate inward currents consistently, however, 80 µM H₂O₂ activated an inward current in 5 of 5 cells. Currents reached a plateau of ~−1.2 pA/pF at 600 s with a half-maximal activation time of 115 ± 3.3 s (Fig. 3C) and exhibited the typical inwardly rectifying I/V relationship for I_CRAC (Fig. 3D). The H₂O₂-induced current could be completely blocked by 1 µM LaCl₃ in the bath solution (Fig. 3C). No significant inward current developed in control experiments when cells were not exposed to H₂O₂ (Fig. 3C). In summary, these data provide evidence for the ability of H₂O₂ to activate I_CRAC in Jurkat T cells and in RBL-2H3 cells. It should be noted that none of the cells showed any sign of seal breakdown or increased leak current due to application of H₂O₂ at the concentrations used and neither did we observe development of any significant non-specific currents.

ROS are known to have a potential damaging effect on molecules such as DNA, proteins and lipids. We therefore investigated if H₂O₂, in addition to its activating qualities, exerted a negative effect on I_CRAC. We activated I_CRAC by perfusing cells with 20 µM IP₃ and internal Ca²⁺ buffered to 150 nM and then applied H₂O₂ after I_CRAC had fully developed. In RBL-2H3 cells, we did not see any inhibiting effect of 40 µM H₂O₂ on I_CRAC (Fig. 3E) and in Jurkat T cells, CRAC currents were slightly reduced by about ~15% during application of 40 µM H₂O₂ (Fig. 3F). The relative lack of effect of H₂O₂ after I_CRAC activation by IP₃ is further evidence that the current activated by H₂O₂ is exclusively I_CRAC and not caused by other non-specific or previously unknown currents.

H₂O₂-induced I_CRAC activation is partially mediated by IP₃ receptors

Next, we sought to determine the underlying mechanism of I_CRAC activation by H₂O₂. First, we addressed whether H₂O₂ activates I_CRAC due to store-independent and direct interaction with the pore-forming unit CRACM1. This was investigated in HEK293 cells stably overexpressing CRACM1 with only endogenous STIM molecules present [23]. These cells produce small IP₃-induced I_CRAC that is limited in size by the endogenous STIM molecules, as both CRACM and STIM proteins are needed to form CRAC channels that are regulated by store-depletion [24]. Whole-cell patch-clamp experiments were performed under conditions similar to those used in RBL-2H3 (Fig. 3). Application of 40 µM H₂O₂ produced a large inward current of −26.7 ± 10.6 pA/pF at 300 s in STIM1-transfected CRACM1-overexpressing HEK-293 cells and no significant current response in non-transfected cells (Fig. 4A). Representative I/V relationships for these cells are shown in Fig. 4B. This indicates that H₂O₂ does not activate CRAC channels directly, but rather through store depletion and STIM1-dependent gating.

Figure 4. Involvement of IP₃ receptors in H₂O₂-activated I_CRAC.

A) Average current density extracted at −80 mV in HEK293 cells stably overexpressing CRACM1 with (filled circles, n=4) and without (open circles, n=5) transfection of STIM1. The bar indicates application of 40 µM H₂O₂. Experiments were performed using Cs-glutamate based pipette solution with [Ca²⁺]_i clamped to 150 nM free Ca²⁺ with 10 mM Cs-BAPTA. Currents were measured with a voltage ramp from −100 to +100 mV over 50 ms at 0.5 Hz intervals from a holding potential of 0 mV. Error bars indicate SEM. B) Representative I/V relationship in response to application of 40 µM H₂O₂ at 300 s in HEK293 cells overexpressing CRACM1 with (black trace) and without transfection of STIM1 (grey trace).C) Average current density extracted at −80 mV in RBL-2H3 cells activated by extracellular application of 40 µM H₂O₂ with perfusion of 100 µg/ml heparin (filled circles, n=5). The bar indicates application of 40 µM H₂O₂. For comparison the H₂O₂-activated I_CRAC trace (Fig. 3A) is included in the figure (dashed line). D) Representative I/V relationship of H₂O₂-activated CRAC current in RBL-2H3 cells at ~600 s when perfused with 100 µg/ml heparin (mean of 4 ramps).

H₂O₂ can induce Ca²⁺ release in a variety of cells and one proposed mechanism is through activation of IP₃ receptors (IP₃R) [36, 37], leading to activation of I_CRAC [17] (see Fig. 3A). To determine if the IP₃ pathway is involved in the activation of I_CRAC by H₂O₂, we perfused RBL-2H3 cells with 100 µg/ml heparin, an antagonist of IP₃ receptors [38]. When applying 40 µM H₂O₂, heparin slowed the activation kinetics of I_CRAC and reduced its amplitude as illustrated in Fig. 4C. In the presence of heparin, I_CRAC developed with a half-maximal activation time of 313 ± 1.8 s compared to 281 ± 1.2 s in its absence, and reaching an amplitude of −1.6 ± 0.2 pA/pF at 600 s compared to −2.5 ± 0.3 pA/pF in controls without heparin. A representative I/V relationship of CRAC currents in the presence of heparin is shown in Fig. 4D. These results further indicate that IP₃R activity may be partly responsible for mediating the H₂O₂-activated I_CRAC, although at present we cannot distinguish between direct IP₃R effects and/or possible effects on IP₃ mobilization.

H₂O₂-induced activation of I_CRAC independent of IP₃R

To further investigate the role of IP₃ receptors in H₂O₂-activated I_CRAC, we conducted patch-clamp experiments in a wild-type chicken DT40 B-lymphocyte cell line (DT40 WT) expressing all three IP₃ receptor isoforms (type I, II and III) as well as a genetically modified DT40 cell line in which all three IP₃ receptor isoforms are knocked out (DT40 KO) [39]. As DT40 cells develop very small I_CRAC, we optimized experimental conditions by increasing the external CaCl₂ concentration to 20 mM and expanding the voltage ramp to −150 mV, extracting current amplitudes at −130 mV [40]. Furthermore internal MgCl₂ concentration was increased to 3 mM to inhibit endogenous TRPM7 currents. In DT40 cells, TRPM2 does not contribute to the H₂O₂ effects, since perfusing these cells with 1 mM ADPR failed to induce any currents (Fig. S1B).

We tested the two cell lines for IP₃-activated I_CRAC by perfusing cells with 20 µM IP₃. In WT cells, I_CRAC developed immediately in 4 out of 4 cells reaching a plateau of −1.2 ± 0.2 pA/pF at ~75 s, whereas KO cells did not show any current development (Fig. 5A). Neither WT nor KO cells developed I_CRAC in the absence of IP₃ with Ca²⁺ buffered to 150 nM (Fig. 5B). However, the DT40 KO cells have previously been shown to activate endogenous I_CRAC when stores are depleted independent of IP₃ receptors [40]. To confirm these previous observations we applied 10 µM thapsigargin (Tg) at 60 s resulting in I_CRAC development (Fig. 5C). Representative I/V relationships for the IP₃ and Tg experiments are displayed in Fig. 5D. Next, we assessed whether H₂O₂ could activate I_CRAC in DT40 cells by applying 40 and 100 µM H₂O₂ from a wide-tipped puffer pipette at 60 s, but as this resulted in no clear I_CRAC (data not shown), we included 40 µM H₂O₂ in the pipette solution. In WT cells this resulted in a slowly developing I_CRAC-like current in 4 out of 4 cells reaching an amplitude of ~−1.2 pA/pF at 300 s (Fig. 5E). In DT40 KO cells an inward current with smaller amplitude of ~−0.5 pA/pF at 300 s could be observed (Fig. 5E). This is significantly smaller than the amplitude in DT40 WT cells (P < 0.05). The averaged current of cells responding with I_CRAC (5 out of 10 cells) yielded an inward current of ~−0.8 pA/pF at 300 s (data not shown). Representative I/V relationships for the H₂O₂ experiments are shown in Fig. 5F. These results suggest that activation of I_CRAC by H₂O₂ occurs mainly through a pathway involving IP₃ receptors, but that an IP₃ receptor-independent pathway can also lead to I_CRAC.

Figure 5. Activation of I_CRAC in wild-type and complete IP₃ receptor-knockout DT40 B cells.

A) Average current density in DT40 wild-type (WT) (filled circles, n=4) and complete (type I, II, III) knock-out (KO) cells (open circles, n=4) when perfused with 20 µM IP₃. Extracellular CaCl₂ concentration was 20 mM and intracellular MgCl₂ concentration was 3 mM. Currents were measured with a voltage ramp from −150 to +100 mV over 50 ms at 0.5 Hz intervals from a holding potential of 0 mV. Inward currents were extracted at −130 mV, averaged and plotted versus time. B) Average current density extracted at −130 mV in DT40 WT (filled circles, n=7) and KO cells (open circles, n=8) with internal Ca²⁺ buffered to 150 nM. C) Average current density extracted at −130 mV in DT40 KO cells in response to application of 10 µM thapsigargin (Tg) as indicated by the bar (n=6). D) Representative I/V relationship of IP₃-perfused DT40 WT cells at t = 150 s and Tg-stimulated DT40 KO cells (red trace, mean of 6 ramps) at t = 240 s. E) Average current density extracted at −130 mV in DT40 WT (filled circles, n=4) and KO (open circles, n=10) when internally perfused with 40 µM H₂O₂. F) Representative I/V relationship of H₂O₂-perfused DT40 WT (black trace, mean of 4 ramps) and KO cells (red trace, mean of 10 ramps) at t = 300 s.

H₂O₂-induced activation of TPRM2 in Jurkat T lymphocytes

The activation of TRPM2 by H₂O₂ is well established [6–8], but with the activation of I_CRAC our results add a new aspect to the experimental use of H₂O₂. Jurkat T cells express both TRPM2 and I_CRAC endogenously [11, 33]. However, the H₂O₂-induced Ca²⁺ influx observed after re-introduction of external Ca²⁺ could completely be blocked by 1 µM La³⁺ (Fig. 2B) and therefore is unlikely to represent Ca²⁺ influx through TRPM2 channels. Two questions therefore remain, why was TRPM2 mediated Ca²⁺ influx not apparent in Ca²⁺ imaging experiments, and if it is a matter of concentration, at what concentration does H₂O₂ activate TRPM2 in Jurkat T cells? We set out to measure H₂O₂-activated TRPM2 currents in patch-clamp experiments. Conditions used to investigate the activation of I_CRAC with intracellular Ca²⁺ clamped to 150 nM did not cause any TRPM2 activation in addition to I_CRAC, even when applying 500 µM H₂O₂ (data not shown). We therefore left intracellular Ca²⁺ unbuffered and assessed H₂O₂-induced single-channel activity that might reflect biophysical properties of TRPM2. This is possible in the whole-cell configuration because of the large single-channel conductance and characteristically long open times of TRPM2 channels [9, 11, 41]. Cells were kept in standard sodium solution with 1 mM CaCl₂ and 1 µM La³⁺. Immediately after break-in, a recording with an 11 s long ramp protocol from −100 mV to 10 mV was started (Fig. 6) and 100 µM H₂O₂ was applied from the extracellular side. This produced only very few single channel openings in 3 out of 3 cells after 60 s. In two of the cells we observed activation of only 2 and 3 single channels and in another cell about 8 channels were active. Fig. 6 shows four ramp measurements from a representative recording with the activity of two single channels displaying a linear current with a reversal potential of 0 mV. We fitted the current traces of the first open channels in every cell with linear regression fits between −100-0 mV, yielding a mean single channel conductance of 65 ± 5.7 pS. This is very similar to the single channel conductances reported in Jurkat T cells when activating TRPM2 with ADPR, 67 pS, and cADPR, 69 pS [11]. These data, taken together with the measurements in intact cells (Fig. 2), indicate that H₂O₂ is a poor activator of TRPM2 in Jurkat T cells and the contribution of H₂O₂-activated TRPM2 channels to Ca²⁺ signals can therefore be regarded as negligible in this cell system.

Figure 6. Activation of TRPM2 channels by H₂O₂ in Jurkat T cells.

Single channel activity of a representative cell measured with long voltage ramps in the whole-cell configuration. Two single channels activated when exposed to 100 µM H₂O₂ from 65 s (dotted lines indicate channel open levels). First single channel activity was observed at 195 s. The data revealed a single channel conductance of 65 ± 5.7 pS (n = 3) at negative potentials. 1 µM LaCl₃ was included in the bath and application solution.

DISCUSSION

The current study provides evidence for H₂O₂ as a novel activator of the Ca²⁺-selective CRAC current and thereby adds new knowledge to the involvement of H₂O₂ in regulating cellular calcium levels. Using calcium imaging and whole-cell patch-clamp experiments we demonstrate that endogenous I_CRAC can be activated by extracellular application and intracellular perfusion of micromolar concentrations of H₂O₂ in Jurkat T cells, HEK293 cells, RBL-2H3 cells and DT40 B cells. This can be blocked by 1 µM LaCl₃. Heparin reduced the H₂O₂-activated I_CRAC in RBL-2H3 cells suggesting that IP₃ receptors play an important role in the Ca²⁺ release that leads to activation of I_CRAC. Furthermore electrophysiology experiments in DT40 WT and KO cells support the involvement of both IP₃-dependent and -independent mechanisms. Lastly, we confirm the activation of TRPM2 ion channels by H₂O₂ in Jurkat T cells, demonstrating that H₂O₂ within the same cell line is capable of activating both endogenous TRPM2 and CRAC channels.

Application of H₂O₂ has been reported to have different effects on ion channels, manifesting themselves as activation, potentiation or inhibition, but we are not aware of a report that would implicate activation of CRAC channels. There may be several explanations as to why other groups have not observed activation of I_CRAC by H₂O₂ in their experiments: First, the cell type in the given experiment may not produce significant I_CRAC. Second, endogenous I_CRAC is an exceedingly small current and requires optimized experimental conditions for electrophysiological detection, i.e. an enhanced Ca²⁺ gradient across the plasma membrane and buffering of cytosolic Ca²⁺ to prevent Ca²⁺-induced inactivation. Third, due to its small amplitude, I_CRAC may not be detected in the presence of larger H₂O₂-activated currents.

In whole-cell patch-clamp experiments we investigated possible activation pathways of H₂O₂-induced I_CRAC. The lack of current response in HEK293 cells stably overexpressing CRACM1 speaks against a direct activation of CRAC channels by H₂O₂. Instead, it appears that Ca²⁺ release from intracellular stores may underlie this effect, since depletion of calcium stores is the main activation mechanism of I_CRAC and H₂O₂ has been reported to induce Ca²⁺ release via activation of IP₃ receptors in human platelets [36] and human endothelial cells [37]. This would also explain the observed reduction in H₂O₂-mediated I_CRAC amplitudes and delayed current activation kinetics when including heparin in the pipette solution. The involvement of IP₃R in H₂O₂-induced I_CRAC was further investigated in DT40 cells in which I_CRAC-like current development was observed in both WT and IP₃R KO cells when including H₂O₂ in the pipette solution. Based on these findings it is tempting to propose that H₂O₂ can also activate I_CRAC through IP₃-independent pathways

A previous report suggested that H₂O₂ activates IP₃ receptors via oxidation of thiol groups, as a reduction in Ca²⁺ release was observed in endothelial cells when the reducing agent, dithiothreitol (DTT), was added to the bath solution [36]. The same study additionally ruled out that the H₂O₂-induced Ca²⁺ release occurred via an increased IP₃ production, although H₂O₂ may induce IP₃ production through activation of tyrosine kinases [42]. The latter effect may be further amplified through H₂O₂-mediated inhibition of protein tyrosine phosphatases [43]. Our results with heparin, which partially inhibits H₂O₂-induced activation I_CRAC (see Fig. 4), suggest that at least part of the H₂O₂ effect is mediated by IP₃ production, although we cannot rule out that heparin could also suppress a more direct activation of IP₃R.

In addition to the proposed involvement of IP₃R, H₂O₂ has also been reported to inhibit SERCA activity [36, 44–46] and this could well be a contributing factor to Ca²⁺ store depletion leading to I_CRAC. The underlying mechanisms are not clear, but it has been suggested that this also occurs via oxidation of thiol groups, as one study observed a protective effect of DTT on SERCA [36]. Another report, however, found DTT to have equivocal effects on H₂O₂-induced inhibition of SERCA [45]. In the latter, a Fenton reaction (Fe²⁺ + H₂O₂ → OH^• + OH⁻ + Fe³⁺) was used to generate hydroxyl radicals and thereby inducing oxidative stress in sarcoplasmic reticulum vesicles from rabbit skeletal muscle cells. Inhibition of SERCA could not be prevented by DTT when using Fe²⁺ in the Fenton reaction, but when replacing Fe²⁺ with Cu²⁺, inhibition was completely abolished. Inhibition of SERCA by hydroxyl radicals has also been suggested to be due to a direct attack on the ATP-binding site [46]. In conclusion, both IP₃ receptors and SERCA may be responsible for H₂O₂-mediated store depletion and subsequent activation of I_CRAC.

Another possible contributor to store depletion by H₂O₂ could be increased membrane permeability through other pathways. The membrane of sarcoplasmic reticulum in pig coronary artery smooth muscle cells had a higher passive Ca²⁺ permeability when exposed to H₂O₂ than the plasma membrane [44], which would be in accordance with activation of I_CRAC by H₂O₂ while still maintaining cell integrity. Additional Ca²⁺-release channels could be involved, e.g. RyR. This release channel can be activated by H₂O₂ [14] and induce I_CRAC in DT40 B cells [47], but RBL-2H3 cells do not express RyR [48] and it is therefore not involved in activation of I_CRAC in these cells. This does not exclude RyR from contributing to H₂O₂-activated I_CRAC in other RyR expressing cell lines, e.g. DT40 B cells. It should be noted that the H₂O₂ concentration required for activation was in the millimolar range [14], which speaks against a significant contribution by RyR to the CRAC currents demonstrated in this study. It has recently been demonstrated that TRPM2 channels also function as lysosomal Ca²⁺-release channels in β-cells [49]. Our data do not exclude that TRPM2-mediated Ca²⁺ release can contribute to the activation of I_CRAC, but at least in Jurkat T cells, H₂O₂ is a rather weak activator of TRPM2 and not likely to contribute either to Ca²⁺ release or Ca²⁺ entry via TRPM2. Another possible source for Ca²⁺ release could be mitochondria, since several studies have reported mitochondrial Ca²⁺ release by H₂O₂ [36, 37, 50]. However, mitochondria are not considered as activators of I_CRAC and their role is to function as a Ca²⁺ buffer, thereby facilitating more extensive store depletion as well as reducing Ca²⁺-dependent slow inactivation of I_CRAC [51]. Finally, it has been demonstrated that H₂O₂ releases Ca²⁺ from a thapsigargin-insensitive non-mitochondrial Ca²⁺ store in endothelial cells [37], raising the possibility that H₂O₂ could target an as yet unidentified CRAC store.

Lanthanides exert a blocking effect on several Ca²⁺-conducting channels. Here we have used 1 µM LaCl₃, a potent blocker of I_CRAC with an estimated K_D of 58 nM [52], to confirm the H₂O₂-induced inward current as I_CRAC. It has been reported that exposing isolated sarcoplasmic reticulum vesicles to 15 µM La³⁺ also blocks IP₃ receptors [53], which could potentially lead to an indirect block of I_CRAC. This, however, does not appear to be the case in the present study, as we did not observe any major differences in the Ca²⁺-release response of any of the cell lines when comparing H₂O₂ traces in the absence or presence of 1 µM LaCl₃. In addition, IP₃ receptors in our cells are unlikely to have been exposed directly to significant concentrations of La³⁺, since La³⁺ does not easily cross membranes [54].

In this study we demonstrate the activation of I_CRAC both by external application of H₂O₂ as well as by internal perfusion. Since H₂O₂ appears to activate I_CRAC through store depletion, it likely interacts with components in intracellular stores and therefore requires H₂O₂ to cross the plasma membrane. This may explain differences in activation times and H₂O₂ concentrations required for activation of I_CRAC across different cell types, as they may differ in both membrane composition and expression of H₂O₂-transporting aquaporins. Additionally, differences in efficiency of cellular H₂O₂-eliminating mechanisms may influence the effective H₂O₂ concentration obtained for the activation of I_CRAC in a given cell. Interestingly, we found that activation of I_CRAC required a much lower H₂O₂ concentration than TRPM2 under optimized experimental conditions in Jurkat T cells where both mechanisms are present. This confirms that H₂O₂ may in principle serve as an activator of TRPM2 currents but in Jurkat T cells it does so with much lower potency and efficacy than activating I_CRAC.

This study is the first to describe the activation of I_CRAC by H₂O₂, which adds new perspectives to the cross talk between calcium homeostasis and ROS as well as to the use of H₂O₂ in experimental settings. This also means that the results of previous studies using H₂O₂ may have had Ca²⁺ contribution from I_CRAC, which could potentially affect both internal Ca²⁺ concentration as well as inward current amplitude in patch-clamp experiments. If I_CRAC activation is unwanted in experiments involving H₂O₂, addition of low concentrations of LaCl₃ may be a useful tool until selective I_CRAC inhibitors become available.

Given the broad spectrum of examples demonstrating cooperativity between H₂O₂ and Ca²⁺ it is highly probable that the findings of this study may be relevant in a number of processes, e.g. inflammation. Neutrophils produce substantial amount of H₂O₂ during the respiratory burst via NADPH oxidase and are highly dependent on intracellular Ca²⁺ as a trigger of processes such as adhesion, differentiation, chemotaxis, phagocytosis, oxidase activation and apoptosis [55]. Our findings provide a possible coupling between the H₂O₂ production and the cellular Ca²⁺ requirement. A similar scenario in which H₂O₂-activated I_CRAC could be relevant is the respiratory burst in mast cells in which ROS production, including H₂O₂, is accompanied by an increase in [Ca²⁺]_i via store-operated Ca²⁺ entry [56]. Furthermore it has been estimated that H₂O₂ during inflammation reaches concentrations of 10–100 µM in the microenvironment surrounding macrophages [57]. This falls within the concentrations used in this study for activation of I_CRAC. Finally, a recent study demonstrates that wound healing in zebra fish results in increased H₂O₂ concentration released by epithelial cells, extending as a concentration gradient 100–200 µm from the site of injury, presumably reaching concentrations of 0.5–50 µM gradually diminishing over 1–2 hrs [58]. The study further demonstrates that the released H₂O₂ functions as a chemotactic signal recruiting leukocytes to the damaged area. Although this contrasts the general concept that ROS produced during inflammation mainly originates from the respiratory burst of phagocytes, it provides an interesting scenario in which H₂O₂ can potentially function as a central mediator of Ca²⁺-dependent processes in different immune cells.