Interaction Between Neuronal NOS Signaling and Temperature Influences SR Ca²⁺ Leak:Role of Nitroso-Redox Balance

Abstract

Rationale

While nitric oxide (NO) signaling modulates cardiac function and excitation-contraction coupling, opposing results due to inconsistent experimental conditions, particularly with respect to temperature, confound the ability to elucidate NO signaling pathways. Here we show that temperature significantly modulates NO effects.

Objective

Test the hypothesis that temperature profoundly impacts nitroso-redox equilibrium, thereby affecting sarcomeric reticulum (SR) Ca²⁺ leak.

Methods and Results

We measured SR Ca²⁺ leak in cardiomyocytes from wild-type (WT), NO/redox imbalance (NOS1^−/−), and hyper S-nitrosylation (GSNOR^−/−) mice. In WT cardiomyocytes, SR Ca²⁺ leak increased as temperature decreased from 37°C to 23°C, whereas, in NOS1^{−/ −}cells, the leak suddenly increased when the temperature surpassed 30°C. GSNOR^{−/ −} cardiomyocytes exhibited low leak throughout the temperature range. Exogenously added NO had a biphasic effect on NOS1^−/− cardiomyocytes; reducing leak at 37°C but increasing it at sub-physiologic temperatures. Oxypurinol and Tempol diminished the leak in NOS1^{−/ −} cardiomyocytes. Cooling from 37° to 23°C increased ROS generation in WT but decreased it in NOS1^−/− cardiomyocytes. Oxypurinol further reduced ROS generation. At 23°C in WT cells, leak was decreased by tetrahydrobiopterin, an essential NOS cofactor. Cooling significantly increased SR Ca²⁺ content in NOS1^−/− cells but had no effect in WT or GSNOR^−/−.

Conclusions

Ca²⁺ leak and temperature are normally inversely proportional, whereas NOS1 deficiency reverses this effect, increasing leak and elevating ROS production as temperature increases. Reduced denitrosylation (GSNOR deficiency) eliminates the temperature dependence of leak. Thus, temperature regulates the balance between NO and ROS which in turn has a major impact on SR Ca²⁺.

INTRODUCTION

Nitric oxide (NO) exerts diverse regulation of cell signaling¹ through a broad range of post-translational modifications, largely through S-nitrosylation of specific cysteine thiol moieties^2-4. Despite the increased understanding of NO signaling and the identification of various NO Synthase (NOS) isoforms in the cardiac myocyte, a consensus has not emerged on the mechanism underlying NO regulation of excitation-contraction (EC) coupling, and studies have reported diametrically opposite results. For example, NOS1 deficient (NOS1^−/−) mice, showed opposite behavior of cardiomyocytes in Ca²⁺ handling and contractility^5-7 or sarcoplasmic reticulum (SR) Ca²⁺ leak^{8, 9} in comparison to wild type (WT) cells. Similarly, studies showed opposite effects on beta-adrenergic contractile responses^{1, 7}. We hypothesized that other factor(s) can fundamentally change the direction of an NO-based physiologic response. Here we identify temperature as a key determinant of cardiomyocyte response to NO.

Temperature has a broad influence on Ca²⁺ signaling in cardiomyocytes^{10, 11} and there is a close relationship between body temperature and NO production in the pathogenesis of focal cerebral ischemia^{12, 13}. Temperature modulates NO production “in vitro”, directly affecting NOS activity¹⁴. Thus, changes in temperature, by affecting NO production and therefore the nitroso-redox (NO/redox) balance^15-17 in the heart, may be closely related to ryanodine receptor (RyR2)-mediated SR Ca²⁺ leak. Post-translational modifications, such as S-nitrosylation, of RyR2 are influenced by cellular NO/redox state and can both positively and negatively affect SR Ca²⁺ leak^18-20. Signaling defects in Ca²⁺ handling, such as Ca²⁺ leak, contribute to impaired contractility in the failing heart, thus depleting SR Ca²⁺ storage as a consequence of the leak and leading to impaired contractile function of the heart.

The present study focused on the effect of temperature on SR Ca²⁺ leak in isolated cardiomyocytes from WT mice or mouse models of aberrant S-nitrosylation, NOS1^{−/ −}and S-nitrosoglutathione reductase deficient (GSNOR^−/− ), and shows that reactive oxygen/nitrogen species signaling are affected by temperature.

METHODS

Animal models

We studied age matched C57BL/6J mice (wild-type, WT; n=26) and mice with a homozygous deletion of NOS1 (B6;129S4-Nos1^<tm1Plh>J; n=20; Jackson Laboratories, Bar Harbor, ME); or S-nitrosoglutathione reductase (GSNOR^−/− ) (n= 10). All protocols and experimental procedures were approved by the Animal Care and Use Committee of the University of Miami and followed the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-234, revised 2011).

Myocyte isolation

Cardiac myocytes were isolated and prepared from hearts as previously described⁶. Briefly, hearts were harvested and perfused retrogradely in a modified Langendorff system (constant flow 2 mL/min) with an isolation solution (see Online Supplemental Material), bubbled with 5% CO₂ -95% O₂ for at least 15 minutes. Once cleaned, the hearts were perfused with collagenase type 2 (Worthington Biochemical Corporation, Lakewood, NJ) ~315 U/mL and protease type XIV (Sigma-Aldrich Co.) 5.2 U/mL for 10 minutes. After digestion, myocytes were released by gentle mechanical disruption. The extracellular Ca²⁺ was restored by sequential additions of CaCl₂ in a Ca²⁺− free Tyrode solution (see Online Supplemental Material). Cardiomyocytes were resuspended in a 1.8 CaCl₂ Tyrode solution at room temperature and then loaded with Fura-2. Myocytes were stimulated at 0.5 or 1 Hz.

Intracellular Ca²⁺ measurement

Intracellular Ca²⁺ was measured using the Ca²⁺-sensitive dye Fura-2 (Molecular Probes, Eugene, OR, USA) and a dual-excitation spectrofluorometer (IonOptix LLC, Milton, MA, USA), excited with a xenon lamp at wavelengths of 340 and 380 nm. The emission fluorescence was reflected through a barrier filter (510 ± 15 nm) to a photomultiplier tube. The “in vivo” calibration was performed superfusing a free Ca²⁺ and then a Ca²⁺ saturating (5 mmol/L) solutions both containing 10 µmol/L ionomycin (Sigma-Aldrich, St. Louis, MO) until reaching a minimal (Rmin) or a maximal (Rmax) ratio values, respectively. [Ca²⁺]_i was calculated as described previously⁹.

Measurement of SR Ca²⁺ leak and SR Ca²⁺ content

SR Ca²⁺ leakage was assessed with 1 mmol/L tetracaine (Sigma-Aldrich, St. Louis, MO, USA) as described by Shannon et al.²¹. The observed decrease in the Fura-2 ratio in presence of tetracaine compared to the non-tetracaine treated condition was considered the Ca²⁺ leak for a particular myocyte (Online Figure I). After assessing Ca²⁺ leak, tetracaine was washed out by superfusing fresh 0Na⁺/0Ca²⁺ Tyrode solution and SR Ca²⁺ content was assessed as described by Bassani et al.²² by a caffeine challenge.

SR Ca²⁺ contents were calculated considering that SR represents 3.5% and cytosol 65% of the myocyte volume as previously described⁹. SR leak-SR load pairs were grouped by similar SR Ca²⁺ load and expressed as a leak-load relationship fitted by an exponential growth function using the Graph Pad Prism software (version 4.02). Measurements were mostly carried out at 23°C, 25°C, 30°C, 34°C or 37°C.

Treatments

Cardiomyocytes loaded with Fura-2 were pre-incubated for 20 minutes with the following compounds (unless otherwise is specified): Tempol (100 μmol/L, Calbiochem, Calbiochem/EMD Biosciences, San Diego, CA, USA); Oxypurinol (100 μmol/L, Sigma-Aldrich); (±)-S-Nitroso-N-acetylpenicillamine (SNAP) (1 or 50 μmol/L, Santa Cruz, Santa Cruz, CA, USA); N⁵-(1-Imino-3-butenyl)-L-Ornithine (LVNIO) (100 μmol/L, Enzo Life Sciences, Plymouth Meeting, PA, USA); Hydrogen peroxide (H₂O₂) (100 μmol/L, Sigma-Aldrich); (6R)-5,6,7,8-Tetrahydrobiopterin dihydrochloride (BH₄) (300 μmol/L, Sigma-Aldrich). Experiments in control condition (no treatment) were run in parallel to each type of pharmacological intervention for each batch of cardiomyocytes.

Detection of Reactive Oxygen Species (ROS)

ROS were measured by using the sensitive probe 2’,7’-dichlorodihydrofluoresceine (H₂DCF −DA, 10 μM; Molecular Probes) in two different ways. First, fresh isolated mouse cardiomyocytes were placed in the chamber of an IonOptix spectrofluorometer and the background fluorescence (F₀) was acquired and then, cardiomyocytes were incubated during 30 min at 23°C or 37°C with H₂DCF − DA and washed. The initial fluorescence (F_i) and a second measure after 5 minutes (F) were acquired. Myocytes were stimulated at 1 Hz and ROS expressed as:

$$ROS=(\mathrm{F}-{\mathrm{F}}_0)-({\mathrm{F}}_i-{\mathrm{F}}_0)$$

Alternatively, control or 100 μmol/L oxypurinol treated cardiomyocytes were loaded with H₂DCF-DA for 15 minutes on polylysine-coated microscope slides at 23°C or 37°C. After 10 minutes washing with Tyrode solution, cardiomyocytes were fixed with 2% p-formaldehyde in cold phosphate buffered saline (PBS) and then washed once with PBS. Cardiomyocytes treated with 1 mmol/L H₂O₂ at either 23°C or 37°C, from each mouse model, were used as positive control for ROS and were used as maximal fluorescence signal for normalization of each group (F_max). Fluorescence (F) was captured with an excitation wavelength of 488 nm and 525 nm emission. Images were quantified by ImageJ (NIH) software and results were expressed as:

$${\mathrm{F}}_{DCF}=(\mathrm{F}-{\mathrm{F}}_0)∕({\mathrm{F}}_{max}-\mathrm{F}),$$

where F₀ is background.

Nitric oxide measurement

Isolated WT mouse cardiomyocytes were loaded with the NO-sensitive dye 4,5-diaminofluorescein diacetate (5 μmol/L DAF-2 DA; Calbiochem/EMD Biosciences, San Diego, CA, USA) for 20 minutes at room temperature. Cells were set in the perfusion chamber of an IonOptix system and fluorescence (excited at 488 nm and emission collected at 510 ± 15 nm) was recorded during stabilization at 23°C. Then, fluorescence was also acquired at 25°C, 30°C, 34°C and 37°C for each cell. DAF-2 fluorescence intensity (F) was expressed as F/F₀, where F₀ is the fluorescence intensity at 23°C after the stabilization time.

NOS isoform expression

Cardiomyocytes from 4 WT hearts were exposed to different temperature and then collected in RNA later lysis buffer (Qiagen, Valencia, CA). Total RNA was extracted from cells using Pure-Link Micro-to-Midi Total RNA Purification System (Qiagen) and reverse-transcribed using High capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA). All samples were treated with TurboTM DNase (Ambion, Austin, TX). Quantitative real-time PCR was performed in triplicate using a 20 μl reaction mixture containing 10 ng cDNA, TaqMan Universal PCR Master Mix (Roche, Branchburg, NJ) and primer/probe sets for nitric oxide synthase 1 (Nos1, Mm00435171_m1), Nos2 (Mm00440502_m1), and Nos 3 (Mm00435197_g1 ) (TaqMan Gene Expression Assay, Applied Bio systems, Foster City, CA). As an internal control glucoronidase beta (GUSB, Mm01197698_m1) was determined in each reaction. Reactions conditions were performed according to manufacturer instructions: 1 cycle of 50°C for 2 minutes, 1 cycle of 90°C for 10 minutes and 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Software from iQ5 multicolor real-time PCR detection system (Bio-Rad, Hercules, CA) was used for PCR analyses. mRNA relative expression was calculated by 2ΔCt method. qPCR data reflects four different experiments each done at 23°C, 30°C, and 37°C.

Statistical analysis

Data are expressed as mean ± SEM. For Leak-Load relationship, an exponential growth fit, which compares independent fits with a global shared fit, was applied. Two or more groups of data were considered to fit different curves if p < 0.05. For comparisons of two groups, Student's t test was used. For comparison of three or more groups, one or two way-ANOVA was performed. Two-way ANOVA was used when a second variable was involved. Newman-Keuls or Bonferroni's post-hoc tests were used as appropriate by the GraphPad Prism version 4.02 (GraphPad Prism Software Corporation San Diego, CA USA). A p < 0.05 was considered significant.

RESULTS

Leak-load relationship is critically affected by temperature

We measured SR Ca²⁺ leak (Online Figure I) in cardiomyocytes from WT and NOS1^−/− (which have depressed S-nitrosylation and oxidative stress) mice over a broad temperature range from 23°C to 37°C. The SR Ca²⁺ leak response to changes in temperature was different in these two models (Figure 1A). The pattern of SR Ca²⁺ leak-temperature, at matched SR Ca²⁺ load, shows that leak slightly increases (p=0.038) when the temperature is reduced from physiologic temperature through 34°C, 30°C, 25°C and 23°C in WT cardiomyocytes, with the leak at 23°C, being higher than at 37°C (Figure 1B, p=0.014 and 2A). In marked contrast, the temperature-dependent change in leak in NOS1^{−/ −}cardiomyocytes was lower compared to WT at low temperature (Figure 1A and B), in agreement with Wang et al.⁸. However, NOS1^{−/ −}cardiomyocytes exhibited a significant increase in leak when the temperature surpassed 30°C (Figure 1A) as we previously showed⁹. This behavior is summarized in Figure 1B which compares SR Ca²⁺ leak in NOS1^{−/ −}at 37°C with 23°C (Figure 1B, p=0.0054 and 2A).

Figure 1. SR Ca²⁺ leak is temperature sensitive.

(A) Dependence of SR Ca²⁺ leak with the bath temperature in WT and NOS1^−/− cardiomyocytes averaged at matched SR Ca²⁺ load ≈ 75 μmol/L. (B) SR Ca²⁺ load-leak relationship at 23°C or 37°C in WT and NOS1^−/− cardiomyocytes. ** p < 0.01 vs. WT at 37°C; *** p < 0.001 vs. WT at 34°C (Two-way ANOVA). † p < 0.05 NOS1^{−/ −}at 37°C vs. 23°C; †† p < 0.01 WT at 37°C vs. 23°C (Exponential growth fitting).

Reactive oxygen species are associated with the SR Ca²⁺ leak

ROS signaling plays a role in the regulation of RyR channel gating by promoting redox post-translational modifications^{3, 23, 24}. In particular, oxidative stress induces RyR2 to leak Ca²⁺ from the SR¹⁹. We found that SR Ca²⁺ leak at either 23°C or 37°C (Figure 2A) correlated with the rate of ROS production (Figure 2B, as measured with dichlorofluoresceine in an IonOptix system) in both WT and NOS1^{−/ −}cardiomyocytes. ROS was increased at 23°C compared with 37°C in WT, while in NOS1^−/− the leak was lower at 23°C compared to physiologic temperature (Figure 2B). In order to verify that ROS mediates SR Ca²⁺ leak-temperature dependence, we first tested the effect of the superoxide scavenger Tempol on Ca²⁺ leak at 23°C, 30°C and 37°C. Tempol eliminated the Ca²⁺ leak in both WT (p<0.0001) and NOS1^{−/ −} (p<0.0001) cardiomyocytes within the range of temperatures (Figure 2C, D and E), confirming the importance of ROS in the gating of RyR2 channels in both models. We further investigated the involvement of xanthine oxidoreductase (XOR), a superoxide-generating enzyme. As shown above, ROS is primarily elevated in NOS1^{−/ −}at 37°C and WT at 23°C. Oxypurinol reduced ROS in XOR-treated cells (Figure 3A and B) and Ca²⁺ leak in NOS1^{−/ −}cardiomyocytes (Figure 3C). The SR Ca²⁺ leak–load relationship were virtually identical in NOS1^{−/ −}under control conditions or oxypurinol between 23° and 30°C (K_+Oxy= 0.0089±0.0045 vs. K_(-)Oxy= 0.0048±0.0024, p= 0.475; Figure 3D), while at 34° and 37°C oxypurinol− treated NOS1^{−/ −}cardiomyocytes exhibited a leak-load curve that was less steep than NOS1^−/− control (K_+Oxy= 0.0030±0.0025 vs. K_(-)Oxy= 0.0117±0.0015, p= 0.0481 at 37°C; Figure 3E).

Figure 2. Leak is associated to reactive oxygen species in cardiomyocytes.

(A) SR Ca²⁺ leak in WT and NOS1^{−/ −}at 23°C or 37°C [at matched SR Ca²⁺ load ≈ 75 μmol/L] (Student's t-test). (B) ROS in WT and NOS1^{−/ −}at 23°C or 37°C (Student's t-test), as measured by epifluorescence of H₂DCF 5 minutes after the initial reading. (C) Dependence of SR Ca²⁺ leak with temperature in WT or NOS1^−/− cardiomyocytes in the absence or the presence of 1mmol/L Tempol [at matched SR Ca²⁺ load ≈ 75 μmol/L] (* p < 0.05 and ** p < 0.01 ( −-treated WT; . †††† (-) Tempol vs. +Tempol p < 0.0001 (-)Tempol vs. +Tempol-treated NOS1^−/− cardiomyocytes; Two-way ANOVA). (D) SR Ca²⁺ load-leak relationship in WT or NOS1^−/− cardiomyocytes, non-treated and Tempol-treated at 23°C (** p < 0.01 WT (-)Tempol vs. WT +Tempol and †† p < 0.01 NOS1^−/− (-)Tempol vs. NOS1^−/− +Tempol). (E) At 37°C (*** p <0.0001 WT (-)Tempol vs. WT +Tempol and ††† p <0.001 NOS1^−/− (-)Tempol vs. NOS1^−/− +Tempol; exponential growth fitting.

Figure 3. Xanthine Oxidase-derived superoxide mediates the increase of leak in cardiomyocytes.

(A) Representative images of H₂DCF fluorescence in response to ROS in WT or NOS1^−/− cardiomyocytes in the presence or the absence of 100 μmol/L oxypurinol at 23°C or 37°C (Bar, 10 μm). (B) Averaged ROS production in WT or NOS1^−/− cardiomyocytes in the presence or the absence of oxypurinol 100 μmol/L at 23°C or 37°C. Values were normalized by the maximal signal for each strain in the presence of H₂O₂ (catalase-treated cardiomyocytes were used as negative control). * p <0.05, ** p < 0.01 and *** p < 0.001 (-)Oxy vs. +Oxy-treated cardiomyocytes, Student's t-test. (C) Dependence of SR Ca²⁺ leak with temperature in NOS1^−/− cardiomyocytes in the absence or the presence of 100 μmol/L oxypurinol [at matched SR Ca²⁺ load ≈ 75 μmol/L] (** p < 0.01 (-)Oxy vs. +Oxy-treated NOS1^−/−, Two-way ANOVA). (D) SR Ca²⁺ load-leak relationship in non-treated and oxypurinol-treated NOS1^−/− cardiomyocytes at 23°C and (E) at 37°C († p< 0.05 (-)Oxy vs. +Oxy NOS1^−/−; exponential growth fitting).

Involvement of NO on the temperature-mediated changes in SR Ca²⁺ leak

In order to test whether defective NO production was the only cause for the biphasic response of SR Ca²⁺ leak to changes in temperature, we supplemented the bathing buffer with the NO donor SNAP. At temperatures over 30°C, 1 μmol/L SNAP decreased the leak in NOS1^−/− and were equivalent to the levels observed in WT cardiomyocytes. As predicted by our hypothesis, at the lower temperatures (23°C and 25°C), 1 μmol/L SNAP increased the leak in NOS1^{−/ −}to the WT values (Figure 4A). Surprisingly, 50 μmol/L SNAP did not significantly affect the leak in NOS1^{−/ −}cardiomyocytes (Online Figure II). Thus, only 1 μmol/L SNAP was able to restore the normal SR Ca²⁺ leak-temperature pattern in NOS1^−/− cardiomyocytes, suggesting that regulation of SR Ca²⁺ in diastole is extremely sensitive to the level of NO, at least under this particular condition of exogenous supply of NO by SNAP. Next, we used 100 μmol/L L-VNIO to specifically inhibit NOS1 in WT cardiomyocytes. This approach allowed us to test if acute inhibition of this isoform was sufficient to mimic the leak-temperature pattern observed in the genetically modified model. Consistently, L-VNIO reduced the leak at 23°C and increased it at 37°C as expected (Figure 4B).

Figure 4. Lack of NOS1-derived NO underlies the atypical SR Ca²⁺ leak-temperature pattern in NOS1^−/− cardiomyocytes.

(A) Dependence of SR Ca²⁺ leak on temperature in NOS1^−/− cardiomyocytes in the absence or the presence of 1 μmol/L SNAP [at matched SR Ca²⁺ load ≈75 μmol/L] (* p < 0.05(-)SNAP vs. +SNAP-treated NOS1^−/−, Two-way ANOVA). (B) Dependence of SR Ca²⁺ leak with temperature in WT cardiomyocytes in the absence or the presence of 100 μmol/L L-VNIO [at matched SR Ca²⁺ load ≈ 75 μmol/L] (* p < 0.05 (-)L-VNIO vs. +L-VNIO-treate WT, Two-way ANOVA).

Reduced SR Ca²⁺ leak is associated with low ROS levels in a GSNO reductase deficiency model

As shown above, aberrant S-nitrosylation and the redox state of cardiomyocytes affects SR Ca²⁺ leak. Deficiency of GSNO reductase also induced an aberrant NO/redox state with hyper S-nitrosylation of a broad spectrum of proteins²⁵. S-nitrosylation is believed to exert a protective role against an oxidative environment, thus the levels of ROS in GSNOR^{−/ −}cardiomyocytes were lower than those in WT at 23°C but not at 37°C (Figure 5A and B). Consistent with the results shown above, where the extent of the leak is proportional to the amount of ROS, the GSNOR^{−/ −}model exhibited lower leak compared to WT cardiomyocytes at sub-physiologic temperatures (Figure 5C, p=0.001; at matched SR Ca²⁺ load, insert). Importantly, the leak at 37°C in GSNOR^{−/ −}was not different, regardless of the lower ROS level, compared to WT, agreeing with our previous findings²⁶. Surprisingly, the treatment of GSNOR^{−/ −} cardiomyocytes with 100 μmol/L hydrogen peroxide (H₂O₂), an oxidant agent that induces cardiomyocyte oxidative stress, did not change SR Ca²⁺ leak in this model (Figure 5C), suggesting that the supposed protective role of hyper S-nitrosylation might be in part due to enhanced cellular antioxidant mechanisms.

Figure 5. ROS and temperature-dependent SR Ca²⁺ leak are attenuated in GSNOR^−/− cardiomyocytes.

(A) Representative images of H₂DCF fluorescence in response to ROS in WT and GSNOR^−/− cardiomyocytes at 23°C or 37°C (Bar, 10 μm). (B) Averaged ROS production in WT and GSNOR^−/− cardiomyocytes at 23°C or 37°C. Values were normalized by the maximal signal for each strain in the presence of H₂O₂ (* p < 0.05 vs. WT, Student's t-test). (C) Dependence of SR Ca²⁺ leak with temperature in WT control cardiomyocytes or GSNOR^−/− in the absence or the presence of 100 μmol/L H₂O₂, [at matched SR Ca²⁺ load ≈ 75 μmol/L] (* p <0.05 WT vs. GSNOR^−/−, Two-way ANOVA).

SR Ca²⁺ stores are primarily affected by elevated leak under NO/redox imbalance

The SR Ca²⁺ content of cardiomyocytes paced at 1 Hz was estimated by caffeine challenge. WT and GSNOR^{−/ −}cardiomyocytes exhibited a relatively stable Ca²⁺ content throughout the studied range of temperatures. Surprisingly, despite the different behavior of leak in GSNOR^{−/ −}compared with WT cardiomyocytes when the temperature drops, the average Ca²⁺ loads were not different throughout the broad range of studied temperatures (Figure 6A). In contrast, the SR Ca²⁺ content in NOS1^{−/ −}was higher at 23°C–25°C than at 34°C − 37°C. At the low range of temperature, SR Ca²⁺ content was elevated in NOS1^{−/ −} cardiomyocytes compared to WT or GSNOR^−/− . However, at a physiologic range of temperatures (34°C-37°C), the Ca²⁺ load was significantly reduced in NOS1^−/− cells (Figure 6A). Thus, there was a significant correlation between the SR Ca²⁺ leak and the resulting average SR Ca²⁺ content in NOS1^{−/ −}(Figure 6B, R²=0.9014) but not in WT (R²=0.0189) or GSNOR^{−/ −}(R²=0.3553) cardiomyocytes. These results suggest that Ca²⁺ stores in the NO/redox imbalance model are highly dependent on the leak of Ca²⁺ from the SR.

Figure 6. SR Ca²⁺ content is governed by leak in NOS1^−/− cardiomyocytes.

(A) SR Ca²⁺ load in WT compared with NOS1^−/− or GSNOR^−/− cardiomyocytes at different temperatures between 23°C and 37°C (* p <0.05 and ** p <0.01; Student's t-test); (B) Linear regression between SR Ca²⁺ leak versus average SR Ca²⁺ load in WT (black) or NOS1^−/− (red) and GSNOR^−/− (cyan) cardiomyocytes.

Coupling of NO synthase and NO production

NO production was assessed in cardiomyocytes from WT mice in order to verify that intracellular NOS− derived NO generation is dependent on temperature. NO levels increased as the temperature increased from 23°C to 37°C (Figure 7A) as described previously for activity of NOS1, NOS2 and NOS3 “in vitro”¹⁴. NO production at the different temperatures was inversely correlated with the Ca²⁺ leak (at matched SR Ca²⁺load) in WT cardiomyocytes (R²=0.9812; Figure 7B). This association suggests that temperature-dependent NOS activity plays an important role in SR Ca ²⁺ leak regulation. We measured SR Ca²⁺ leak in WT cardiomyocytes at 23°C in the presence of tetrahydrobiopterin (BH₄), a NOS cofactor which also contributes to maintain the functional quaternary structure (dimer) of the enzyme. Interestingly, there was a time-dependent reduction in leak (p=0.066 with 2 minutes pre-incubation and p=0.011 with 20 minutes pre-incubation; Figure 7C), suggesting that the elevated leak at room temperature is mediated by NOS1 uncoupling. Additionally, we examined the relationship between temperature and expression of NOS1, NOS2 and NOS3 in WT cardiomyocytes exposed to 23°C, 30°C and 37°C for 30 minutes. No differences were seen in the amount of mRNA of each isoform in response to temperature change (Figure 7D), further suggesting that dependence of NO production on temperature is regulated by the activity rather than expression of NOS isoforms.

Figure 7. NO synthase uncoupling increases leak at low temperature in WT cardiomyocytes.

(A) NO generation in WT cardiomyocytes measured by detection of DAF-2 fluorescence at 23°C, 25°C, 30°C, 34°C or 37°C. Fluorescence was expressed as relative fluorescence units (F/F₀) by normalization of the DAF-2 signal at each temperature (F) to the fluorescence at 23°C (F₀) (N=4 mice; * p < 0.05 vs. 23°C, one-way ANOVA). (B) Lineal correlation between SR Ca²⁺ leak and NO production in WT cardiomyocytes at all studied temperatures. (C) SR Ca²⁺ leak [at matched SR Ca²⁺ load ≈ 75 μmol/L] in WT cardiomyocytes in the absence (BL) or the presence of 100 μmol/L tetrahydrobiopterin (BH₄) for 2 or 20 minutes (* p < 0.05, Student's t-test). (D) mRNA content of NOS1, NOS2 and NOS3 in WT cardiomyocytes exposed to 23°C, 30°C and 37°C for 30 minutes. (N=4; Student's t-test).

DISCUSSION

The present findings reconcile a central, 15 year old controversy in NO cardiobiology; temperature profoundly affects SR Ca²⁺ leak. These results explain why experiments on isolated myocytes when performed at room temperature yield different findings to those performed at physiologic temperatures^{8, 9}.

We previously showed that leak was increased in the absence of NOS1 at 37°C whereas Wang et al. showed a reduced leak at room temperature compared to WT control cardiomyocytes⁸. Therefore, we have highlighted that the thermal dependence of the cardiomyocyte metabolism tightly affects the redox state of the cell, which modifies RyR2 activity and thereby SR Ca²⁺ leak. At low temperatures, RyR2 gating is favored^{10, 11}, and may be associated with reduced NOS activity¹⁴. It has been proposed that XOR and NOS1 co-localize in the SR in close proximity to RyR2. Thus, NO1-mediated NO production exerts a powerful inhibitory effect on myocardial XOR and tightly controls RyR2 channel activity. Therefore, a decrease or absence of NOS1 activity leads to a rise in superoxide anion²⁷ and dysregulation of RyR2-mediated Ca²⁺ release. We consistently found increased ROS generation in WT cardiomyocytes when cooled from 37°C to 23°C. This effect was primarily XOR-mediated since inhibition of this enzyme with oxypurinol prevented the rise of ROS at 23°C. We speculate that the limited availability of NOS1-derived NO in this microenvironment is still sufficient to account for the differences in the leak between WT and NOS1^{−/ −}at 23°C in the presence of ROS. The interaction of NO with superoxide yields peroxynitrite, a very aggressive molecule that is more RyR2-damaging than ROS^{15, 28}.

Consistent with observations in NOS1^{−/ −}cardiomyocytes that the complete absence of NOS1-derived NO is associated with greatly enhanced XOR activity^{18, 27}, the reduction in ROS by oxypurinol was associated with reduced SR Ca²⁺ leak. In addition, the complete elimination of ROS by Tempol, abolished the leak-temperature dependence either in WT or NOS1^−/− cardiomyocytes, confirming the key role of ROS in this process. Other sources of ROS might play a signaling role, affecting the Ca²⁺ leak control. For instance, mitochondrial-derived ROS may be significantly affected by temperature due to changes in the electron flow through the respiratory chain²⁹. Indeed, hypothermic preconditioning activates mitochondrial ROS release and ERK activation in ventricular myocytes³⁰. Moreover, additional evidence that ROS are associated with leak is the fact that, at sub-physiologic temperatures, where the leak is reduced in GSNOR^{−/ −}compared to WT cardiomyocytes, there is also less ROS. In this regard, RyR2 is a highly redox-sensitive Ca²⁺ channel and its activity is increased by oxidizing agents and decreased by reducing agents. However its regulation is even more complex. Thus, the intracellular NO/redox equilibrium is a critical factor determining the responsiveness to Ca²⁺ of the RyR2 channel. Post-translational modifications of RyR2, such as S-glutathionylation³¹ or S-nitrosylation^{18, 19}, are considered protective against irreversible oxidation of redox-sensitive cysteine residues on this channel. As shown previously²⁶, the leak in GSNOR^{−/ −}cardiomyocytes was not different from WT at 37°C, consistent with there being no difference in ROS, and suggesting that enhanced S-nitrosylation caused by deficient denitrosylation activity, would not affect leak at physiologic temperature. In contrast to WT, decreasing temperature did not increase the leak in GSNOR^{−/ −}cells, which could indicate a redox-protective effect of S-nitrosylation at sub-physiologic temperatures. Further evidence for this hypothesis is the fact that the exposure of GSNOR^{−/ −}cardiomyocytes to an oxidative stimulus was not sufficient to break down this redox-protective shield mediated by hyper S-nitrosylation.

We mentioned above that a lack of NOS1-derived NO production may be the underlying cause of the atypical SR Ca²⁺ leak-temperature pattern in cardiomyocytes from NOS1^{−/ −}mice. We have shown that restoring the NO supply in this model rescues the pattern of high leak at room temperature and lower leak at the physiologic range of temperatures as observed in WT cardiomyocytes. However, there is a complexity to NO signaling which is evidenced in this work by the ability of 1 μmol/L but not 50 μmol/L of SNAP to elicit such an effect. We previously showed a triphasic response of SR Ca²⁺ leak, to increasing doses of nitroglycerin in NOS1^−/− cardiomyocytes⁹, thus the effects of exogenously added NO are highly dependent on the concentration and the nature of the NO donor. Accordingly, the idea that reduced NOS1 activity promotes a local microenvironment for this particular behavior of SR Ca²⁺ leak, explains the specific inhibition of NOS1 in WT cardiomyocytes yielding results consistent with those in the NOS1 knockout model. As expected, the high leak at 23°C in WT was abolished and the low leak at 37°C was increased by L-VNIO. However the effect was less robust at the physiologic temperature since this increase was only half the level of the leak observed in NOS1^{−/ −}at this temperature. We speculate that the intracellular metabolism of the drug might be different at 37°C and therefore, the effective concentration of L-VNIO might not be sufficient to efficiently inhibit NOS1. Alternatively, the time of exposure to the inhibitor may not be adequate to generate the characteristic NO/redox imbalance as seen by the chronic deletion of NOS1. The risk of using a higher dose of L-VNIO is the loss of specificity. Thus, we confirmed that defective NOS1-derived NO production leads to a NO/redox imbalance that differentially affects SR Ca²⁺ leak at physiologic or sub-physiologic temperatures in isolated cardiomyocytes.

Reduced activity of NOS1 at low temperature may be the cause of increased XOR-mediated ROS in WT cardiomyocytes. Moreover, this redox dysregulation would uncouple NOS1, further contributing to ROS production (see the proposed working model in Figure 8). Thus, re-coupling of NOS1, as mediated by BH₄ treatment, restores the redox balance and leads to low levels of SR Ca²⁺ leak. It was recently shown that BH₄ supplementation restores NO production and S-nitrosylation of proteins, in cardiomyocytes from a model of dystrophic cardiomyopathy exhibiting oxidative stress²⁰.

Figure 8. Proposed working model for the effects of the temperature on diastolic RyR2-mediated Ca²⁺ leak.

A) Under physiologic temperature conditions, NOS1 activity is normal and its interaction with XOR maintains RyR2 in a stabilized state by a regulated proportion of S-nitrosylated cysteines. B) At sub-physiologic temperature, NOS1 activity is decreased, which allows increasing superoxide (O₂⁻) production by XOR, leading to BH₄ oxidation, NOS1 uncoupling with further ROS production, which promotes the generation of peroxynitrite (ONOO⁻) and consequent oxidation of thiol moeties on RyR2. Under these conditions, RyR2 becomes more susceptible to Ca²⁺ leak.

It has been demonstrated previously that NO-mediated S-nitrosylation of RyR2, enhances its activity^{3, 32}. However it is important to differentiate activation of the channel from a leaky channel. RyR2 activity is regulated by many factors including Ca²⁺, phosphorylation, S-nitrosylation, oxidation, etc. Despite activation, the channel might be more or less stabilized, exhibiting different levels of diastolic Ca²⁺ leak. Recent studies have associated low S-nitrosylation levels with increased leak measured at 37°C in different models^18-20 suggesting that NO-mediated S-nitrosylation reduces leak. Here, we propose that temperature is a variable that, by affecting the NO/redox balance, may change the response of RyR2 to exogenous NO.

Ca²⁺ stores in NOS1^{−/ −}cardiomyocytes are highly dependent on the leak of Ca²⁺ from the SR, since increased leak at higher temperature is associated with depleted SR and vice versa. In addition to the increased leak at 37°C, a slow Ca²⁺ decay (Online Figure III) also contributes to the depletion of SR Ca²⁺ stores at this temperature in NOS1^−/− . However, Ca²⁺ decay is not different from WT at 23°C (Online Figure III), thus the decreased Ca²⁺ leak would be enough to explain the augmented Ca²⁺ content at low temperature. This dependence allows us to speculate that SR Ca²⁺ stores in this model of NO/redox imbalance are poorly regulated. There was no obvious association between SR Ca²⁺ content and leak within the range of temperatures studied in WT myocytes, suggesting that despite reduced Ca²⁺ decay with decreasing temperature (Online Figure III), it may be enough to compensate for the increased leak in normal mice. Similarly, SR Ca²⁺ content was not significantly affected by leak in GSNOR^{−/ −} cardiomyocytes. We would expect an enhanced SR filling in GSNOR^−/− since they exhibit not only lower leak at sub-physiologic temperatures but also faster Ca²⁺ reuptake than WT (Online Figure III). Paradoxically, we did not observe such a response. We do not have a logical explanation for these results, but just speculations of a very tight set point for the SR filling as a consequence of the hyper S-nitrosylation phenotype. Further investigations on this model of aberrant S-nitrosylation should be addressed to elucidate this aspect.

Potential implications

A number of studies have investigated the effect of hypothermic preconditioning^{30, 33, 34} or hypothermia^35-38 on ischemia-reperfusion and myocardial infarction. These studies demonstrate that therapeutic hypothermia profoundly prevents the deleterious effects induced by these experimental models of myocardial injury. Therapeutic hypothermia is broadly tested in clinical trials and protocols are still being optimized with the main goal of reducing neurologic impairment after cardiac arrest^39-41. However, the understanding of its benefits on cardioprotection is limited. Reducing temperature in a normal healthy heart increases the Ca²⁺ leak which would be detrimental to the myocardium. Although this effect of cooling does not appear to be protective for healthy cardiomyocytes, the perspective changes when the scenario switch to a highly aggressive condition such as ischemia-reperfusion, where a burst of ROS is released and the NO-redox equilibrium is disrupted. Indeed, the latter situation is better represented by the NO/redox imbalanced described for NOS1^{−/ −}hearts at physiologic temperature. Upon reperfusion, the ischemic myocardium undergoes oxidative stress^{42, 43} which may also activate Ca²⁺/calmodulin-dependent protein kinase II^{44, 45}, and leading to Ca²⁺ handling abnormalities including SR Ca²⁺ leak⁴⁶, which is characteristic in cardiomyocytes with NOS1 deletion. Therefore, inhibition of NOS1 during therapeutic hypothermia could avoid peroxynitrite formation, and might yield better results in terms of reducing spontaneous SR Ca²⁺ release into the cytosol, thereby improving cardioprotection and the outcomes of this powerful intervention. Alternatively, according our data, inhibition of GSNOR during hypothermia may also add substantial improvements to this protocol since Ca²⁺ leak is kept low during cooling in the GSNOR deficient model.

Conclusion

Here we show that Ca²⁺ leak increases when the temperature drops, coinciding with an XOR-mediated increase in ROS production and NOS1 uncoupling. In contrast, the leak rises at temperatures >30°C in the absence of NOS1 activity and is also associated with elevated ROS. Deficient denitrosylation has a protective role by maintaining low leak levels independent of the temperature. These results suggest that Ca²⁺ leak from the SR is regulated by a crucial interaction between temperature and NO/redox balance.

Novelty and Significance.

What Is Known?

• Nitric oxide (NO) signaling modulates cardiac function, including calcium cycling in cardiac myocytes.

• Abnormalities in calcium cycling, such as increased calcium leak from intracellular stores, severely affect cardiac performance.

• Changes in temperature affect the metabolism of NO and consequently NO signaling in the heart.

What New Information Does This Article Contribute?

• Calcium leak is dependent on temperature and is strongly associated with NO and reactive oxygen species (ROS).

• The response to temperature of cardiomyocytes lacking neuronal NO synthase activity (NOS1^−/− mouse model) is opposite that of wild type cardiomyocytes whereas deficient denitrosylation (GSNOR^−/− mouse model) protects against the cooling-induced rise in ROS and calcium leak.

Using pharmacologic interventions on isolated cardiac myocytes derived from NOS1^−/−, GSNOR^−/− and wild type mouse models, we demonstrate that calcium leak exhibits changes that are regulated by an interaction between temperature and the NO/redox balance. This work highlights that temperature is a crucial factor in NO cardiobiology, and reconciles controversial findings in this field. Moreover, these results support the proposed cross-talk between xanthine oxidase and neuronal NO synthase in the regulation of the calcium channel gating in the heart. Decrease in temperature toward sub-physiologic conditions disrupts this cross-talk favoring a rise of ROS and therefore increasing calcium leak. Hence, either inhibition of neuronal NO synthase or preventing denitrosylation during therapeutic hypothermia in patients undergoing reperfusion post-ischemia or myocardial infarction, might be beneficial in preventing a temperature-dependent increase in calcium leak, thereby favoring cardioprotection and improving clinical outcomes.

Nonstandard Abbreviations and Acronyms

Ca²⁺

calcium

GSNOR^−/−

S-nitrosoglutathione reductase deficient mice

NO

nitric oxide

NO/redox

nitroso-redox

NOS1^−/−

neuronal nitric oxide synthase deficient mice

ROS

reactive oxygen species

RyR2

cardiac ryanodine receptor

SR

sarcoplasmic reticulum

XOR

xanthine oxidoreductase