H₂S concentrations in the heart after acute H₂S administration: methodological and physiological considerations

Administration of exogenous H₂S was unable to increase the pools of combined H₂S in the heart of rats, even after H₂S intoxication. Any long-lasting effects of H₂S on the heart, previously reported in the literature, do not require a measurable accumulation of H₂S in cardiac tissue.

Abstract

In this study, we have tried to characterize the limits of the approach typically used to determine H₂S concentrations in the heart based on the amount of H₂S evaporating from heart homogenates—spontaneously, after reaction with a strong reducing agent, or in a very acidic solution. Heart homogenates were prepared from male rats in control conditions or after H₂S infusion induced a transient cardiogenic shock (CS) or cardiac asystole (CA). Using a method of determination of gaseous H₂S with a detection limit of 0.2 nmol, we found that the process of homogenization could lead to a total disappearance of free H₂S unless performed in alkaline conditions. Yet, after restoration of neutral pH, free H₂S concentration from samples processed in alkaline and nonalkaline milieus were similar and averaged ∼0.2–0.4 nmol/g in both control and CS homogenate hearts and up to 100 nmol/g in the CA group. No additional H₂S was released from control, CS, or CA hearts by using the reducing agent tris(2-carboxyethyl)phosphine or a strong acidic solution (pH < 2) to “free” H₂S from combined pools. Of note, the reducing agent DTT produced a significant sulfide artifact and was not used. These data suggest that 1) free H₂S found in heart homogenates is not a reflection of H₂S present in a “living” heart and 2) the pool of combined sulfides, released in a strong reducing or acidic milieu, does not increase in the heart in a measurable manner even after toxic exposure to sulfide.

NEW & NOTEWORTHY

Administration of exogenous H₂S was unable to increase the pools of combined H₂S in the heart of rats, even after H₂S intoxication. Any long-lasting effects of H₂S on the heart, previously reported in the literature, do not require a measurable accumulation of H₂S in cardiac tissue.

exogenous hydrogen sulfide (H₂S) or H₂S releasing agents have been administered to various animal models and in vitro preparations in an attempt to determine whether this molecule possesses physiological and pharmacological properties beneficial to human health, such as preventing ischemic cardiac insult (27, 33, 51, 52). This novel field of research must, however, resolve a difficult conundrum (11, 39): as there is a clear overlap between the blood and tissue concentrations of H₂S that have been reported to be associated with measurable physiological/pharmacological effects and those producing toxic effects (15, 24), the fundamental question remains of whether it is possible for H₂S to increase in a measurable manner in tissues without producing some signs of toxicity (11). In addition, it is difficult to reconcile long-term effects that could result from the direct accumulation of H₂S in tissues with the very rapid disappearance of dissolved H₂S from the blood (2, 24) and tissue (58) after exogenous administration. The latter disappears in minutes because of the high rate of H₂S oxidation in the blood (10, 60, 62) and its profound interaction with hemoglobin in the blood (16, 59). As a result and since only the dissolved form of exogenous H₂S (at neutral pH, H₂S/HS⁻) is able to diffuse from the blood into the tissues, we can anticipate that only a very small fraction of this dissolved/free exogenous H₂S/HS⁻ that is administered can eventually diffuse into tissues (16, 24). Furthermore, the pool of H₂S/HS⁻ diffusing into any tissue is rapidly oxidized in the mitochondria (17) or can combine with metallo-compounds (45). It comes therefore as no surprise that many studies have suggested that 1) only a trivial amount of free/dissolved H₂S (in the nM range at best), if any, can be found in baseline conditions in various tissues including the heart (6, 18, 28, 57) and 2) no significant change in the pools of combined H₂S should be expected to be recovered after administration of exogenous H₂S. More recent studies have, however, reported values that were at odds with this view, showing concentrations of free H₂S in the high micromolar range in the heart (22, 25, 42) and much higher levels of combined H₂S after administration of a relatively small quantity of exogenous H₂S (34, 41). The origin of this discrepancy, which is sometimes of several orders of magnitude, is not clear.

Most of the studies designed to determine the concentrations of H₂S in the heart after H₂S administration rely on a similar principle: the determination of gaseous H₂S diffusing from heart homogenates. The premise of this approach is that free H₂S in cardiac tissues, i.e., dissolved H₂S, is present in the form of gaseous H₂S and HS⁻ in a pH-dependent manner. Since the acidic dissociation constant (pK_a) of H₂S is ∼7 (1, 32), gaseous H₂S, which is about a third of the total amount of dissolved H₂S at physiological pH, evaporates in keeping with the difference in partial pressure of H₂S between the milieu where it is supposed to be present and the gaseous phase. HS⁻ converts to gaseous H₂S as the latter evaporates, until all soluble H₂S/HS⁻ disappears from the milieu. For the pools of sulfide that are unable to evaporate spontaneously (20), different methods have been used to free combined HS⁻ from proteins and metalloproteins. For instance, sulfane sulfur, bound sulfur (42, 55, 54), or sulfurated/sulfhydrated proteins (31, 40, 56)—terms coined to describe a pool of H₂S combined with the cysteine residues of proteins—have been suggested to be produced after sulfide administration. Strong reducing agents, such as dithiothreitol (DTT) or tris(2-carboxyethyl)phosphine (TCEP), by cleaving disulfide bonds have been shown to be able to release this pool of sulfide (8, 30, 44, 61), which then could be measured as gaseous H₂S. It should be pointed out that the view that H₂S/HS⁻ can per se create in vivo sulfhydrated proteins has been challenged on various grounds (55); rather, it has been suggested that it is the oxidized form of H₂S, S⁰, that can achieve such sulfhydration (55).

Another pool of combined sulfide has been described as “acid labile,” from which gaseous H₂S can only be released at low pH, typically below 4 (18). This “acid-labile” pool comprises various metallo-sulfide compounds, which reflect the propensity of free H₂S to combine with the multitude of metalloproteins present in and outside cells. H₂S released at extremely low pH should not be confused with the effect of decreasing pH on the transformation of HS⁻, already present in dissolved form, into gaseous H₂S. In addition to these pools, polysulfides can be formed after exposure to exogenous H₂S and have been suggested to contribute to some of the effects of H₂S (21, 32). More importantly, most free H₂S is rapidly oxidized and eventually eliminated as sulfate (7) and thiosulfate (7, 17, 29) and cannot be recovered by evaporation. We have estimated that the rate of disappearance of free H₂S from the blood is so rapid that soluble H₂S concentration in the blood can fall to zero within minutes after sulfide intoxication (16).

The purpose of the present study was to determine the fate of exogenous H₂S in the heart in control conditions and after infusion of levels of H₂S producing an acute cardiac failure during which H₂S must interact with cardiac proteins (19) and metalloproteins (16), using a methodology based on the principle described above, i.e., H₂S evaporation. We sought to characterize the limits of this approach. We have focused our interest on the processes involved in preparing tissue homogenates, which require an extreme alteration of dying cardiac tissues, which are cut, sliced, crushed, and then homogenized, conditions during which a significant amount of free sulfide could evaporate. Our general objective was to determine whether the changes in pools of H₂S in the heart can be used as a tool to evaluate the benefits of H₂S at low levels (26, 52) as well as its toxicity at higher levels of exposure (2, 12, 16). Our specific aims were to describe 1) the effects of the process of homogenization on H₂S evaporation; 2) the use of alkaline solution while processing the cardiac tissue to limit this evaporation; 3) the effects of DTT, the reducing agent typically used to free H₂S from cysteine residues; and 4) the change in various pools of H₂S in the heart in rats exposed to H₂S-induced cardiac toxicity. Finally, we sought to integrate and discuss these data in light of models accounting for the interaction of H₂S with proteins and metallo-compounds in vitro and in vivo.

METHODS

Determination of H₂S

Description of apparatus and equipment.

Samples [10 or 20 ml of sodium hydrosulfide (NaSH) solutions or heart homogenate] were placed in a 50-ml leakproof chamber sealed with an airtight plastic stopper and equipped with two catheters and inlet-outlet ports for gas flow. Samples were introduced into the chamber with these catheters; unless otherwise mentioned, the pH of the solution or the homogenate was maintained between 6 and 7, even when the samples were processed at higher pH (see below). The “fluid phase” was constantly agitated with a magnetic stirrer bar, while the headspace was ventilated by a continuous flow of N₂ to limit the oxidation of H₂S. This flow was set at 500 ml/min to overcome the rate of sampling of the gas analyzer while maintaining the highest possible H₂S partial pressure gradient between the fluid and gas phases. In few instances, a 500-ml Erlenmeyer flask was used for some validation experiments as described below. The actual flow rate of N₂ was determined with a pneumotachograph connected to a pressure transducer (Pneumotach amplifier 1, series 1100, Hans Rudolph, Shawnee, KS). The concentration of H₂S flowing out of the headspace was directly measured by a H₂S gas analyzer [Interscan RM series, Interscan, range 0–1 parts per million (ppm), resolution 0.001 ppm]. The analog output of the H₂S analyzer and the flow rate signal were recorded with an A/D data acquisition system (PowerLab 16/35, ADInstruments) and were analyzed on LabChart7 (ADInstruments). The concentration of H₂S (ppm) in the gas phase multiplied by the gas flow rate through the chamber headspace (l/s) allowed the computation of the rate of H₂S diffusing from the solution. The total amount of H₂S evaporating at any given time t could then be computed by integrating the rate of diffusion of H₂S during this time t. Results are expressed as concentration of H₂S per gram of tissue or per liter of solution as well as total quantity of H₂S produced.

Validation of the measurement.

The linearity and limits of detection of our experimental setup were tested with known concentrations of H₂S in 10 ml or 20 ml of a buffer solution (final concentration of 0.05, 0.1, 2, 20, 100, and 200 μM H₂S) added in the chamber as described in Description of apparatus and equipment. Stock solution of H₂S was freshly prepared from NaSH (Sigma) in PBS.

To determine potential artifacts produced on the H₂S signal related to the nature and pH of the solution, the same analysis was repeated by injecting 2–5 ml of solution containing HCl (final 0.3 M), glycine-NaOH buffer (25 mM), TCEP (5 mM), or DTT (50 mM) into the chamber containing 10 or 20 ml of buffer solution (for TCEP or DTT) without any H₂S present. The total quantity of H₂S is presented as “equivalent H₂S.” All compounds were tested in triplicate.

Effects of pH of solution on H₂S and effects of homogenization.

With a pK_a around 7, soluble H₂S is present in solution in the forms of HS⁻ and gaseous H₂S, with about one-third of gaseous H₂S at physiological pH (1, 32). At pH of 9, most H₂S is under the form HS⁻, minimizing the loss of H₂S in tissue, but with little or no S⁰ (1, 32). Since homogenization was done at a pH of 9 in fresh tissue, we have determined the maximal rate of disappearance of H₂S from a solution containing 100 nmol of H₂S that was constantly agitated at a pH of 6, 7, or 9. In addition, to clarify the effect of homogenization per se on the loss of H₂S during tissue processing, solutions of H₂S in PBS (5 ml) or 50 mM glycine-NaOH buffer (pH 9.3) were exposed to the same protocol as heart homogenates, with a tapered glass tissue homogenizer (Wheaton) for 15 min, and then were introduced into the chamber to measure H₂S remaining after this process. A solution of 100 mM sodium phosphate buffer was used to restore neutral pH. Again, all our measurements were done at neutral pH, even when the processing of fresh tissue was performed in alkaline solution.

In Vivo Animal Experiments and Tissue Sampling

Animal preparation.

Fifteen male Sprague-Dawley rats (Charles River) weighing 499 ± 95 g (∼12–16 wk) were used. All experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (8th ed., National Research Council Institute for Laboratory Animal Research). The study was approved by the Penn State Hershey Medical Center Institutional Animal Care and Use Committee.

Rats were anesthetized by inhalation of 3–5% isoflurane in O₂ for few minutes before intraperitoneal injection of urethane (700 mg/kg) + α-chloralose (60 mg/kg). The rats were tracheostomized and were breathing spontaneously. Expiratory flow was measured with a pneumotachograph (series 1100, Hans Rudolph), and mixed expired O₂, CO₂, and H₂S fractions were determined by specific analyzers as previously described (14, 24). A catheter (PE-50) was inserted into the right femoral artery to measure arterial blood pressure (ABP). A similar catheter was advanced to the left ventricle through the right carotid artery to monitor ventricular pressure via pressure transducers (46). An additional catheter was inserted into the right femoral vein for continuous H₂S or saline infusion. Electrocardiogram was recorded by a bioelectric amplifier (Differential AC Amplifier, model 1700, A-M Systems). All signals were digitized with PowerLab and were visualized online. Data were stored for further analysis with LabChart7.

Six animals were used as control, and nine animals were exposed to H₂S by continuous infusion of NaSH solution. H₂S infusion was started at a rate of 0.4 mg/min (5 μmol/min, 0.8 mg/ml NaSH in saline), and then the rate of infusion was increased by 0.08 mg/min (1 μmol/min) every 2 min. In three rats, the infusion was stopped when mean ABP (MABP) reached 50 mmHg with a decrease in left ventricular contractility (dP/dt_max) (cardiogenic shock, CS). These three rats were euthanized precisely 5 min after the end of sulfide infusion. The animals were transcardially perfused with 60 ml of cold PBS to wash out the blood, and the heart was removed from the rat within a minute and weighed. The heart was immediately immersed in a total volume of 5 ml of 50 mM glycine-NaOH buffer (pH 9.3). In the six remaining animals, the infusion was continued until the animal presented a complete cardiac asystole (CA), typically by pulseless electrical activity (PEA). Animals presenting apnea during the infusion were mechanically ventilated. These six rats were perfused with PBS according to the same protocol as soon the PEA occurred. Three of these six hearts were immediately immersed in glycine-NaOH buffer, while the three other hearts were snap-frozen by liquid nitrogen and stored in a −80°C freezer. In the six control animals, three hearts were immediately used for H₂S measurement, while the other three hearts were snap-frozen.

H₂S measurement in the heart.

The fresh hearts were homogenized at a pH of 9 with a tapered tissue grinder and transferred to the 50-ml chamber. The pH of the tissue homogenate was restored to ∼7 as soon as it was introduced into the chamber with sodium phosphate buffer. As described in Effects of pH of solution on H₂S and effects of homogenization, the total amount of free H₂S spontaneously evaporating was determined at neutral pH in all fresh samples. The total time between the death of the animal and the recording was <30 min.

After 1 h of spontaneous H₂S evaporation, TCEP was added to the chamber at a final concentration of 5 mM (pH was adjusted to neutral by NaOH) while H₂S signal was continuously recorded. An additional injection of 50 mM TCEP was performed 10 min later, and the total TCEP-released signal was measured over 30 min. The pool of acid-labile H₂S was then identified by injecting HCl (final concentration of 0.3 M) into the chamber. A small aliquot (∼50 μl) of the sample was withdrawn from the chamber through the catheter each time H₂S signal returned to baseline for determination of the pH of the solution. With this approach, the pH of the solution reached values in each instance <1.5. Each test was performed in triplicate.

Frozen cardiac tissues were homogenized with a mortar. Ground tissue was mixed with 10 ml of saline and kept in an airtight syringe. The homogenate was then injected into the 50-ml chamber to measure free H₂S.

The “equivalent H₂S” level produced by glycine-NaOH buffer injection or the injections of TCEP or HCl was subtracted from the corresponding H₂S signal. The quantity of H₂S in the heart homogenate was expressed as nanomoles per gram of tissue wet weight. To compare our data to the levels of free or bound H₂S in the heart reported in the literature, we assumed that the weight of proteins was ∼10% of the total wet weight of the heart in rats (28, 48).

H₂S in Vitro Interaction with Proteins and Metalloproteins.

We also tested the limit of detection for H₂S bound to protein (albumin) or metalloprotein [hydroxocobalamin or methemoglobin (MetHb)]. The flask was prefilled with N₂ for 20 min, and both air inlet and outlet ports were clamped to allow 10 min of incubation. H₂S (final concentration of 200 μM) was injected into the 500-ml chamber containing a solution of 10% albumin (1.52 mM; bovine albumin fraction V, DOT Scientific), hydroxocobalamin (1 mM; Sigma), or 10% MetHb (1.55 mM; human-derived Hb, Sigma). Free H₂S was determined for 30 min; then TCEP was injected, followed by an injection of HCl (0.3 M). The same protocol was repeated in albumin with 60 min of incubation instead of 10 min. Each condition was studied in triplicate.

Statistical Analysis

Linear regression and Bland-Altman plots were used to compare the values of H₂S in our solutions and expected concentrations obtained by weighing NaSH. The time constant of the change in the concentrations of H₂S in solution was determined at different pH values as the time required for the concentrations to decrease by 63%. The concentrations of H₂S evaporating from the solutions were compared between PBS and 10% albumin by unpaired t-test. The effects of incubation time (10 min or 1 h) in PBS or albumin solution were also analyzed by unpaired t-test. H₂S recoveries in solution containing albumin, MetHb, hydroxocobalamin, or PBS alone were also compared with one-way ANOVA followed by Bonferroni's post hoc comparisons. P < 0.05 was regarded as significant for any of these comparisons.

All statistical analyses were conducted with GraphPad Prism 5 (GraphPad Software, La Jolla, CA).

RESULTS

Validation of H₂S Detection Method

Detection limit and linearity in H₂S determination.

The linearity and detection limits were established by injecting known quantities of H₂S into the chamber (Fig. 1). The expected concentrations of H₂S were determined by precise weighing of NaSH (see methods) corresponding to a final concentration of 200, 100, 20, and 2 μM (in 20 ml) and 0.1 and 0.05 μM (in 10 ml) of H₂S. The amount of H₂S injected ranged therefore from 0.5 nmol to 4 μmol. The total quantities of free H₂S evaporating/recovered over a 30-min period are given in Fig. 1 for each concentration (in triplicate) as a function of the expected values. Dropping the pH of the solution to 1.2 after 30 min of recording (not shown in Fig. 1), transforming any remaining HS⁻ to gaseous H₂S, allowed the recovery of an additional 5% across all H₂S levels. Of note, the recovery of total H₂S for the lowest levels of H₂S that we used, i.e., 0.5 and 1 nmol (corresponding to 50 and 100 nM) were 0.48 and 0.86 nmol, respectively, as shown in Fig. 1.

Fig. 1.

A: H₂S signal expressed in parts per million (ppm; top) and the corresponding quantity in nanomoles (bottom) in response to injection in the chamber of solutions containing different concentrations of H₂S (200, 100, 20, and 2 μM in 20 ml and 100 and 50 nM in 10 ml of buffer solution). Arrowheads indicate the moment of injection. B: correlation between the concentrations of H₂S solution injected (expected values obtained by weighing NaSH) into the chamber and H₂S recovered from evaporation (measured values). Each data point is the average of results obtained in triplicate. C: Bland-Altman plot of the data shown in B. The approach used to determine H₂S concentrations allows a recovery of most of the H₂S present in solution. Inset: magnification of the plot for the lowest concentrations (0.05 and 0.1 μM H₂S).

Effects of saline, HCl, glycine-NaOH buffer, and TCEP solutions on H₂S signal.

As illustrated in Fig. 2A, injection of 2 ml of saline into 10 or 20 ml of PBS produced a small artifact, corresponding to an equivalent level of H₂S of 0.089 ± 0.047 nmol (or 4.5 ± 2.4 nM in the 20-ml chamber). Injection of HCl (final 0.3 M) into the chamber also produced a very small and transient artifact, which averaged 0.204 ± 0.126 nmol (10.2 ± 6.3 nM), while the injection of 25 mM glycine-NaOH buffer (buffer used in our fresh homogenate) produced an artifact of 0.123 ± 0.061 nmol (6.2 ± 3.1 nM). These artifacts from HCl and glycine-NaOH buffer were similar to saline. Finally, in contrast to DTT, 5 mM TCEP injection produced 0.068 ± 0.041 nmol (3.4 ± 2.1 nM in the chamber), which was similar to the error obtained after injection of saline. TCEP at 50 mM produced the same artifact as the solution containing 5 mM. Data presented below are reported after subtraction of this artifact.

Fig. 2.

A: H₂S signal after injection in the chamber of 2–5 ml of solutions containing saline, HCl (0.3 M final concentration), glycine-NaOH buffer (25 mM), TCEP (5 mM), and DTT (50 mM). H₂S signal is expressed in ppm (top) and the corresponding quantity in nanomoles (bottom). Arrowheads indicate the moment of injection. B: averaged signals (triplicate measurement) expressed in equivalent H₂S produced by saline, HCl, glycine-NaOH, TCEP, and DTT. Note that the injection of saline, HCl, glycine-NaOH buffer, and TCEP produced an artifact ranging from 0.07 to 0.20 nmol of equivalent H₂S (in 10 or 20 ml, see methods). In contrast, significant levels of H₂S appear to be produced by DTT. C: averaged signals produced by saline and DTT which was prepared at the time of the experiment (D0) or 1 day (D1) or 3 days (D3) before the experiment. Freshly prepared DTT solution (D0) released an amount of H₂S that was 150 times higher than the artifact produced by saline. H₂S recovered from the DTT solution was even much larger when 1- to 3-day-old DTT was used. D: H₂S remaining in solution at neutral and alkaline pH while constantly agitated at relevant pH values. At pH of 6 and 7 very similar kinetics of H₂S were found (with time constant of 2.7 and 4.1 min, respectively), and H₂S remaining in tissues reached the same level at 15 min. Using a solution pH of 9 dramatically slowed down the rate of evaporation, with a longer time constant (21.8 min). E: H₂S measured at neutral pH from 100 nmol of H₂S-mixed solution after 15 min of “processing” in glass homogenizer. Almost all the H₂S virtually disappeared from the solution prepared with PBS after 15 min of grinding. In contrast, the loss of H₂S was prevented by using an alkaline solution, allowing the recovery of 65% of H₂S present in solution.

Amount of H₂S produced by DTT.

As developed in methods, DTT has been used in many studies as the primary reducing agent to release H₂S from cysteine residues (18, 42, 61). DTT produced very significant H₂S artifact: injecting 50 mM DTT (1 mmol) into the chamber produced an error corresponding to 201.88 ± 187.13 nmol of H₂S, or a concentration of 10.1 ± 9.4 μM (Fig. 2, A and B). Fresh DTT immediately prepared before the measurement produced an average of 15 nmol H₂S, leading to error in concentration in our setting of 0.75 μM and 10 times more if the solution was >1 day old (Fig. 2C).

Effect of pH on rate of H₂S evaporation.

The maximal rate of H₂S evaporation obtained during vigorous and constant agitation was similar in solutions mixed with 100 nmol H₂S measured at pH of 6 and 7, with a time constant of disappearance from the solution of 2.7 and 4.1 min, respectively (Fig. 2D). At a pH of 9, the maximal rate of H₂S evaporation was dramatically slowed down, with a much more longer time constant (21.8 min).

Mechanical effects of tissue grinder on the loss of dissolved H₂S.

The mechanical effects of grinding on the recovery of H₂S from a known concentration of H₂S in PBS and in alkaline solution (100 nmol, 10 μM in 10 ml) are illustrated in Fig. 2E: all dissolved H₂S has virtually vanished after 15 min of grinding when performed at neutral pH. In contrast, the loss of H₂S was limited by grinding at pH of 9 (with glycine-NaOH buffer), allowing recovery of 65 ± 12% of the expected H₂S amount. Of note, all measurements were done at neutral pH.

Summary of characterization.

Our method of determination of H₂S was able to accurately determine concentrations of free H₂S in solutions across a range of 50 nM to 200 μM. The limit of detection that we considered in this study was 0.2 nmol. The addition of DTT released a large amount of H₂S into the measurement chamber, in contrast to TCEP. Finally, the use of alkaline solution at pH 9 during the process of homogenization prevented in part, but not totally, the loss of H₂S from the sample solution.

Determination of H₂S Concentrations in Heart After Sulfide Intoxication-Induced Cardiac Failure in Rats

Circulatory effects of H₂S.

Nine animals received an intravenous infusion of H₂S at a rate of ∼0.8 mg/min (10 μmol/min), which we found to produce a rapid depression in cardiac function (19, 46, 47). In six of these nine rats the infusion was continued until the animal presented CA, while in three rats the infusion was stopped as soon as MABP reached 50 mmHg (Fig. 3). As presented in methods, six rats received saline infusion and were used as the control group.

Fig. 3.

Examples of the effects of continuous infusion of H₂S, starting at 0.4 mg/min (5 μmol/min) NaSH in saline and then increasing by 0.08 mg/min (1 μmol/min) every 2 min, on respiratory flow (V̇), arterial blood pressure (ABP), and concentration of gaseous H₂S in the blood (CgH₂S) estimated from expired H₂S. H₂S infusion induced decrease in ABP and increase in CgH₂S. A: in the cardiogenic shock (CS) protocol, infusion was interrupted when mean ABP dropped below 50 mmHg. In this example, the rat was mechanically ventilated (MV, gray area) because it presented apnea. ABP recovered rapidly. Expired H₂S disappeared within a few minutes after the cessation of the infusion. Animals were euthanized 5 min after the cessation of sulfide infusion. B: cardiac asystole (CA) conditions. H₂S infusion was maintained, increasing the rate of infusion by 0.08 mg/min (1 μmol/min) every 2 min, until CA. CgH₂S reached a peak of 40 μM before sharply decreasing, likely related to a decrease in pulmonary blood flow.

The effects of H₂S infusion were identical to those we previously reported (46, 47). Baseline MABP was 95 ± 18 mmHg with oxygen consumption (V̇o₂) of 11 ± 2 ml/min. MABP decreased within <5 min, reaching 50 mmHg at 12 ± 4 min into the infusion, while estimated arterial concentrations of gaseous H₂S (CgH₂S) increased to 5.8 ± 1.6 μM. Of note, V̇o₂ was still unchanged at this time (12 ± 4 ml/min). Cardiac contractility decreased during this period, as reflected by a drop in dP/dt_max from 4,394 ± 1,011 to 3,134 ± 636 mmHg/s. In the three rats in which infusion was stopped at this level, MABP and dP/dt_max continued to decrease, reaching their nadir (35 ± 13 mmHg and 1,779 ± 1,270 mmHg/s, respectively) within 2–3 min. Two of three animals presented apnea requiring mechanical ventilation. When the animals were euthanized (5 min after the cessation of H₂S exposure), all the circulatory parameters had already recovered (Fig. 3), while V̇o₂ remained unchanged (12 ± 6 ml/min). MABP and dP/dt_max were back to baseline (94 ± 18 mmHg and 4,593 ± 1,180 mmHg/s, respectively). Expired H₂S was undetectable at the time of euthanasia. The total amount of H₂S that was infused averaged 94 ± 26 μmol (7 ± 2 mg) in these three rats.

In the six remaining rats, H₂S infusion was maintained with all the animals under mechanical ventilation. MABP and dP/dt_max continued to decrease until a PEA or ventricular fibrillation occurred. Cardiac arrest was observed 18 ± 4 min into the infusion. CgH₂S reached a peak value of 37 ± 10 μM. The total amount of H₂S injected under this condition was 178 ± 58 μmol (13 ± 4 mg).

H₂S concentrations in heart.

control animals.

Figure 4A shows original traces of H₂S evaporating from control heart homogenates. Free H₂S was measured in three control animals from freshly harvested hearts (1.1–1.8 g) after homogenization in glycine-NaOH buffer and from frozen hearts in the other three rats. The amount of H₂S that we could recover from fresh control hearts (processed in alkaline solution) averaged 0.50 ± 0.46 nmol (0.15–1.02 nmol), corresponding to a concentration of free H₂S of 0.18 ± 0.24 nmol/g tissue (Fig. 4B). The total amount of H₂S released after TCEP (5 mM) was within our limits of detection, corresponding to 0.001 ± 0.002 nmol/g. Higher concentrations of TCEP (50 mM) had no additional effects. Gaseous H₂S abruptly increased after addition of 0.3 M HCl (286.5 ± 68.5 nmol). The concentration of H₂S produced by an acidic environment (pH < 1.5) was 214.5 ± 11.6 nmol/g.

Fig. 4.

A: example of the changes in H₂S concentrations in ppm (left) and in the quantity of H₂S expressed in nanomoles and nanomoles per gram of tissue (right), released from control heart homogenates. Top: H₂S levels were obtained without any intervention (spontaneous release of free H₂S). Gray line presents signal produced by the vehicle solution. Middle: TCEP (5 mM) injection did not produce any measurable change in H₂S. Bottom: effects of 0.3 M HCl injection on release of H₂S. Note that reducing the pH abruptly to very low levels produced the release of H₂S corresponding to concentrations of H₂S of ∼200 nmol/g. B: average + SD concentrations of H₂S released from hearts that were freshly homogenized or snap-frozen and homogenized. The levels of free H₂S (spontaneously released) were similar in hearts harvested from the animals with H₂S-induced cardiogenic shock (CS) and the control group. Higher levels of H₂S were found in the hearts obtained from the animals with cardiac asystole (CA). In contrast, H₂S levels measured after the injection of 5 mM TCEP and 0.3 M HCl were similar in all groups. Note that H₂S was not detected (ND) in the fresh CA hearts and the frozen control hearts after TCEP in control conditions. The release of H₂S and the effect of TCEP and HCl injection were similar in the frozen samples and the fresh samples in control conditions.

Results were of the same order of magnitude with the three pooled frozen hearts (total 2.9 g) while the hearts were processed at neutral pH. Free H₂S averaged 0.37 nmol/g, while H₂S released after TCEP and HCl injection were 0 and 128.3 nmol/g, respectively.

intoxicated animals.

The level of free H₂S in the three hearts (0.9–1.6 g) harvested after H₂S-induced acute CS and processed at pH 9 averaged 0.25 ± 0.10 nmol/g and was not different from the levels found in the control hearts. In contrast, in the three hearts (1.1–1.6 g) obtained from the animals that died from H₂S-induced complete CA the level of free H₂S was ∼10 times higher, reaching 14.34 ± 12.48 nmol/g. Much higher levels were even found in the pooled homogenate from the hearts (2.9 g) obtained from the three remaining animals that died from H₂S exposure (106 nmol/g, ∼600 times both the control level and the level found during sublethal intoxication).

However, 5 mM TCEP released no H₂S (0.001 ± 0.001 nmol/g, under detection limit) from the hearts that presented acute cardiac failure or CA. HCl (0.3 M) injection induced a large increase in H₂S from samples of both acute cardiac failure and CA, averaging 186.1 ± 13.2 and 229.9 ± 34.6 nmol/g, respectively, values that were similar to the levels found in the frozen hearts (Fig. 4B).

Interaction Between H₂S with Albumin and Metallo-compounds

Ten percent albumin solution.

A 10% albumin solution (1.52 mM) was incubated for 10 min with 2, 20, and 200 μM H₂S and for 1 h with 200 μM H₂S only (Fig. 5 and Fig. 6). After 10 min of incubation the recovery of gaseous H₂S was only 6 ± 5%, 35 ± 4%, and 74 ± 5% of the initial amount of H₂S for 2, 20, and 200 μM, respectively. Addition of 5 mM TCEP and 0.3 M HCl did not produce any relevant recovery of H₂S; as a result, 80 ± 5%, 62 ± 3%, and 25 ± 5% of the initial H₂S concentration was not recovered after incubation with 2, 20, and 200 μM H₂S. After 1 h of incubation, the recovery of a solution of 200 μM H₂S was only 6 ± 2% (Fig. 6). H₂S evaporating after TCEP and HCl represented not more than 2% and 1% of the total initial amount of H₂S, respectively.

Fig. 5.

H₂S recovery from PBS and 10% albumin solution after 10 min of incubation with various concentrations of H₂S (2, 20, 200 μM). A, left: H₂S recovered is expressed in absolute values. Right: H₂S recovered is expressed in % of initial concentrations. The level of H₂S recovered in albumin solution was decreased compared with PBS, and although administration of 5 mM TCEP allow recovery of some H₂S from the protein solution as a function of the initial H₂S concentration, the largest part of H₂S was not recovered by TCEP. *P < 0.05, significant difference between PBS and 10% albumin. B: various pools of H₂S recovered from PBS or albumin solution at various H₂S concentrations. This representation illustrates that the major part of H₂S “disappearing” from H₂S mixed with albumin was not recovered after TCEP or HCl.

Fig. 6.

H₂S recovery in PBS and 10% albumin solution after 1 h of incubation with 200 μM H₂S (data obtained at 10 min are given for comparison and are the same as in Fig. 5). Data presented similarly to Fig. 5. A: in PBS, recovery of free H₂S was 97% and 83% of original concentration after 10 min and 1 h of incubation, respectively. TCEP (5 mM) and HCl (0.3 M) injection did not produce any significant levels of H₂S. In albumin, 94% of initial concentration of free H₂S was not recovered after 1 h of incubation. H₂S produced by TCEP or HCl corresponded to a tiny proportion of H₂S that was not recovered. *P < 0.05, significant difference between 10-min and 1-h incubation. B: various pools of H₂S recovered from PBS or albumin solution after 10-min and 1-h incubation. This representation illustrates that the major part of H₂S “disappearing” from H₂S mixed with albumin was not recovered after TCEP or HCl.

MetHb.

MetHb solution (1.55 mM) was incubated with 200 μM (4,000 nmol) H₂S for 10 min. As illustrated in Fig. 7, only 2 ± 1% of the initial quantity of H₂S was spontaneously recovered. Injection of 5 mM TCEP had virtually no effect and allowed the recovery of no more than 1 ± 2% of initial H₂S. However, in major contrast to the solution of albumin and as shown in Fig. 7, 70 ± 12% of H₂S incubated with MetHb solution was immediately recovered after addition of 0.3 M HCl.

Fig. 7.

H₂S recovery from a solution of hydroxocobalamin (B12; 1 mM) and methemoglobin (MetHb; 1.55 mM) incubated for 10 min. Data presented similarly to Fig. 6. A: 98% of the initial amount of H₂S was not recovered from hydroxocobalamin solution, and neither 5 mM TCEP nor 0.3 M HCl released any additional H₂S. Similarly, most of the free H₂S (98% of initial concentration) was not recovered in MetHb solution. However, ∼70% of H₂S was recovered after addition of HCl. *Significantly different from PBS (P < 0.05); ^$significantly different from albumin (Alb) (P < 0.05); ^#significantly different from hydroxocobalamin (P < 0.05). B: various pools of H₂S recovered from the solution of hydroxocobalamin and MetHb after 10-min incubation. This panel shows that the major part of H₂S “disappearing” from H₂S mixed with the cobalt compound was not recovered after TCEP or HCl, in contrast to the reaction with the ferric iron of MetHb.

Hydroxocobalamin.

As shown in Fig. 7, incubating a 1 mM solution of hydroxocobalamin with 200 μM H₂S solution for 10 min led to the spontaneous recovery of only 2 ± 1% of initial H₂S concentration, akin to the effects of MetHb, and no additional H₂S (0.003 ± 0.003%) was recovered after addition of 5 mM TCEP. However, the addition of 0.3 M HCl did not produce any relevant evaporation of H₂S (0.12 ± 0.07%), leaving 98% of the initial H₂S unrecovered.

DISCUSSION

To determine the concentrations of the various pools of H₂S present in the heart after sulfide intoxication, our approach, which has been used previously by different groups (18, 28, 42), relies on the assumption that molecules of dissolved H₂S present in an homogenate of tissues, in this instance cardiac tissue, are capable of diffusing outside the tissue and then can be measured as gaseous H₂S. Various pools of H₂S have been found to be present in tissues based on this method (6, 18, 28, 42, 44, 57): 1) a pool of dissolved H₂S (gaseous H₂S and HS⁻) spontaneously diffusing from the tissue homogenate and 2) a pool of combined H₂S that must be first transformed into a dissolved form before being measured. Two different interventions are typically applied: A first approach uses a strong reducing agent (typically DTT or TCEP) with the aim of recovering H₂S combined on cysteine residues of proteins. This approach was first used in brain tissues by Warenycia et al. (61). A second procedure consists of acutely decreasing pH to extremely low levels (<2) and, by doing so, releasing H₂S from a pool of sulfide, which comprises sulfide-metals (18). It should be pointed out that this methodology is being used on “dead tissue” homogenate, sometimes prepared in alkaline conditions to avoid initial evaporation of H₂S (6).

What we found is that numerous limitations exist with this approach, affecting our interpretations of the presence of dissolved and combined H₂S and the effects of infusion of exogenous sulfide, as discussed in the following paragraphs.

Baseline Levels of H₂S in Heart: Is Free Sulfide Present in the Heart in Control Conditions?

As shown in Fig. 4, the level of H₂S spontaneously released from the heart in baseline conditions was within the same order of magnitude as in the studies of Levitt et al. (28) and Ubuka et al. (57), in which levels of baseline H₂S were found to be extremely small (0.055 nmol/g tissue) or close to the detection limit. However, as mentioned in the introduction and as illustrated in Fig. 8, some studies have reported baseline levels ranging from 10,000 to 40,000 nmol/g (22, 25, 34, 41, 42). In most of these studies, the concentration of gaseous H₂S was typically measured by gas chromatography with a single sampling from the headspace in a closed vial. The debate on the actual amount of “free” sulfide present in tissues is not a new one (9, 38). In 2008, Furne et al. (6) and Whitfield et al. (62) independently published two papers aimed at challenging the view that gaseous/dissolved H₂S endogenously produced is present in the blood and in tissues at concentrations previously suggested to range from a few to hundreds of micromolar (39). A similar argument has been made by Olson (37), who also confirmed, by direct measurement of the level of gaseous H₂S “diffusing” from tissue homogenates or from the blood, that the concentration of free H₂S was, if anything, lower than a few hundred nanomolar. This contention is also supported by the observation that concentrations in the low micromolar range of gaseous/dissolved H₂S in solution are enough to impede mitochondrial function (3, 64). Furthermore, from a toxicological standpoint, H₂S poisoning provokes life-threatening symptoms at concentrations of gaseous H₂S in the blood around 5 μM (16, 24, 47).

Fig. 8.

Concentrations of H₂S in rodent hearts from the literature. Concentrations are expressed in micromoles per kilogram. Whenever data were expressed per gram of protein, we assumed a 10% ratio between the weight of the heart and the weight of proteins (28, 48). Free H₂S, H₂S produced in reducing environment, and H₂S released in very acidic condition were identified according to the methods used in each paper. H₂S concentrations from references a, b, and c are the concentrations measured in the present study in control, CS, and CA conditions, respectively. ND, undetectable. This figure shows that 2 different ranges of concentrations of free H₂S can be found in the literature (see discussion): a first group of studies including ours identified levels of free H₂S below 0.2 μmol/kg, while a second group of studies reported levels between 6,000 and 40,000 μmol/kg. Similarly very high concentrations of H₂S have been found after addition of a reducing agent, which in most cases was DTT (in contrast to TCEP in our study), corresponding to levels of H₂S even higher than the entire amount of H₂S that was administered (see also Table 1). The acid-labile H₂S concentrations are all within a similar range (high micromolar) in all studies.

In our hands, the process of preparing the homogenate tissue, if performed at neutral pH, got rid of all free H₂S present by evaporation, as expected from previous works from Olson's group (4). As shown in Fig. 2, using NaOH during the heart processing clearly slowed down this evaporation process, in keeping with the expected effects of an alkaline pH on H₂S-to-HS⁻ ratio (32). However, loss by evaporation is still present even at pH 9 during the grinding process.

Finding free H₂S in postmortem conditions may therefore say little about the presence of free H₂S in a living heart, as it is the result of the balance between H₂S released and unable to be oxidized by dying tissues and H₂S evaporating, trapped, and/or oxidized before, during, and after the homogenization procedure. The alkaline condition that was used to limit evaporation could have also contributed to the postmortem H₂S/HS⁻ production from tissue homogenate containing cysteine molecules (36, 62). Of note, we found similar levels of H₂S in our frozen tissues that were not prepared in glycine-NaOH buffer (Fig. 4), in which all dissolved H₂S should have been expected to evaporate.

No additional H₂S was recovered after incubation with TCEP in baseline conditions. This strong reducing agent was preferred over DTT, which produced a large H₂S signal “artifact” (Fig. 2). A similar effect was observed previously (39) and has already been suggested to take place in other tissues (53), making disputable any inference about the tissular concentrations of H₂S found in the micromolar range after DTT.

Decreasing pH below 2 with ∼0.3 M HCl led to the recovery of a relatively large amount of gaseous H₂S, corresponding to a tissue concentrations of sulfide around the 200 μM (or rather μmol/kg) range. These levels of H₂S after treatment with a strong acid are consistent with the results of Levitt et al. (28), Ubuka et al. (57), and Ogasawara et al. (35) (see Fig. 8). This pool of H₂S, by analogy with the effects of HCl on our solution of MetHb, may have originated from H₂S already present/combined on myoglobin or may have originated from a large pool of thiols or proteins denatured at very low pH.

Effects of H₂S Administration

Only after lethal intoxication (CA conditions) was the level of free H₂S elevated (∼100 times higher than in control and CS conditions). Although this still represents a very low level (low μM), these concentrations would be consistent with the levels we found in the blood to be associated with high toxicity for the heart (15, 16, 47). Keeping in mind the limitations of this approach, these data suggest that the ability of the heart and most tissues to oxidize H₂S disappears as soon as the mitochondrial function is abolished (as in CA conditions), allowing the accumulation of exogenous H₂S. Nevertheless, the total amount of infused H₂S leading to an acute CS in a 500-g animal averaged 94,000 nmol delivered within 10 min; 178,000 nmol was required to produce a terminal CA. Free H₂S found in the heart from our intoxicated animals was not more than 0.04 nmol and 20.18 nmol in CS and CA conditions, respectively. In other words, at best 0.00004% and 0.011% of the total amount of H₂S that was infused was found in the heart in CS and CA conditions, respectively, despite the massive H₂S exposure.

Table 1 illustrates the different levels of H₂S found in the heart after H₂S administration in keeping with the dose administered in various studies. In contrast to studies reporting the quantity of H₂S recovered in combined forms that were sometimes higher than the amount administered, our main finding was that the amount of combined H₂S in the hearts harvested a few minutes after recovery from a sulfide exposure that led to a profound depression in cardiac function (CS conditions) was not different from that in control animals. The same amount of H₂S was recovered after TCEP or after HCl (Fig. 4 and Fig. 8) in all conditions, despite the fact that H₂S must have interacted with the heart, at least with L-type calcium channels, during H₂S-induced cardiac failure (19, 50, 65), leading to a profound depression in cardiac contractility. Clearly, data on the cobalt compounds (Fig. 7) suggest that the interactions between metallo-compounds and H₂S cannot be reduced to a simple trapping effect, as H₂S oxidation in air may have taken place during the few minutes the heart was harvested, a phenomenon that could have been catalyzed by the presence of some metallo-compounds. The lack of effects of TCEP also requires some clarification. Following original publication of Warenycia et al. (61), it has been accepted that in a reducing environment H₂S combined with the free cysteine residues of proteins can be released. Although the existence of this mechanism in vivo has been challenged, strong reducing agents appear to be able to release free H₂S from proteins (18, 42, 44). We found that this pool does not represent a significant sink for H₂S as was previously hypothesized (63), at least within a range of concentrations of H₂S producing a cardiac arrest. The fate of H₂S in a simple solution of albumin, which may possess one free cysteine for creation of a disulfide bond with H₂S per molecule of albumin (49), does support the view that most exogenous H₂S reacting in tissues can be recovered (Fig. 5 and Fig. 6). This implies that the major part, if not all H₂S, could have been oxidized. Indeed, while H₂S mixed with PBS for 1 h had a recovery rate that averaged ∼80% (Fig. 6), free H₂S released from an albumin solution barely reached a few percent. In our study, >90% of H₂S was still unrecovered from the albumin solution even after TCEP, indicating that the presence of albumin may have catalyzed H₂S oxidation into sulfide compounds that do not reversibly release H₂S under reducing or acidic conditions.

Table 1.

Quantity of H₂S in the heart after H₂S administration

Pool of H2S	Reference	Species	Protocol of Administration	Amount of Exogenous H2S Administered, nmol	Change in Amount of H2S in Heart, nmol	H2S Concentration in Heart, % dose administered
Free	b	Rats	IV H2S-induced CS	94,000	+0.04	+0.00004
	c	Rats	IV H2S-induced CA	178,000	+20.18	+0.011
	34	Mice	Intra-LV and IV 7 days	224	+1,000	+446
	41	Diabetic mice	IV 1 day	64	+50	+78
			IV 7 days	448	+50	+11
Bound (reducing agent)	b	Rat (TCEP)	IV H2S-induced CS	94,000	0	0
	c	Rat (TCEP)	IV H2S-induced CA	178,000	0	0
	34	Mice (DTT)	Intra-LV and IV 7 days	224	+8,000	+3,571
	41	Diabetic mice (DTT)	IV 1 day	64	+1,000	+1,563
			IV 7 days	448	+3,500	+781

Values are amount of H₂S injected (mice or rats) and change in the amount of H₂S found in the heart from baseline conditions expressed in nanomoles or in % of the amount administered from the literature. References b and c summarize our data obtained during H₂S-induced acute cardiogenic shock (CS) and cardiac asystole (CA), respectively. Note that in many studies the additional amount of H₂S recovered from the heart after H₂S administration was much higher than the amount of H₂S that was administered. IV, intravenous; LV, left ventricle.

Finally, as mentioned in the introduction, it should be kept in mind that the amount of H₂S required to interact with the heart to acutely reduce its contractility is actually extremely small.

Perspectives

This study does not support the view that exogenous H₂S increases the pools of free or combined H₂S in the heart in a significant manner, even at toxic levels. This conclusion does not preclude the formation of such pools, which if affecting key proteins may exert important physiological as well as toxic effects, but we interpret our finding as evidence that this pool represents, if anything, a trivial quantity within the limit of detection of methods using H₂S evaporation. Obviously, considering that only a very small fraction of exogenously administered H₂S would be “available” to the heart, it would require a considerable amount of H₂S (higher than in lethal intoxication) to increase the pool of H₂S in the heart anyway. At these levels, the brain, the most sensitive organ to H₂S toxicity, would long have been intoxicated. The previously reported prevention of ischemia-reperfusion injury by H₂S (41, 43) could well be mediated by mechanisms that are not directly related to the accumulation of H₂S per se in the heart but via specific interaction with key proteins or key mechanisms of transduction. It is also possible that the acute effects of H₂S on mitochondrial respiration (3), via decrease in ATP (23) along with reactive oxygen species production (5), could be effective by triggering for instance a preconditioning effect in the heart rather than by acting through a persistent accumulation of a measurable pool of sulfide.

In terms of H₂S poisoning, H₂S provokes a very rapid depression in cardiac contractility. The therapeutic agents trapping H₂S, cobalt compounds or MetHb (15), are most effective when H₂S is present in a free form. We have shown that the rapidly diffusible nature of H₂S creates a small window for antidotes to be administered with some effectiveness (2, 12, 16). Our present results indicate that free H₂S can be found in the heart but only after lethal exposure conditions in which the animals presented a PEA (12, 46) but not in the animals surviving the intoxication. These observations, along with the still unresolved issue of whether free H₂S found in the heart is actually present in vivo, challenge the use of trapping agents alone in the treatment of sulfide intoxication-induced acute CS (13, 47) if administered minutes after H₂S exposure.

Conclusions

Methods using evaporation of free H₂S from cardiac homogenates do not allow a meaningful determination of dissolved H₂S in the heart. In addition, H₂S released after application of the strong reducing agent TCEP was hardly detectable in postmortem conditions. Acid-labile H₂S was found in the high micromolar range, but its origin remains unclear and this pool of H₂S was not affected after H₂S administration even at high or toxic levels. These data can be understood in light of in vitro interactions between H₂S and proteins or metallo-compounds and the very low levels of H₂S able to diffuse into tissues during and after H₂S administration even in toxic ranges. These results suggest that the major part, if not all, of exogenous H₂S administered in vivo vanishes in forms that cannot be recovered.

GRANTS

This work was supported by the National Institutes of Health Office of the Director and the National Institute of Neurological Disorders and Stroke (Grants 1R21 NS-090017-02 and U01 NS-097162).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

T.S. and P.H. performed experiments; T.S. and P.H. analyzed data; T.S. and P.H. interpreted results of experiments; T.S. and P.H. prepared figures; T.S. and P.H. drafted manuscript; T.S. and P.H. edited and revised manuscript; T.S. and P.H. approved final version of manuscript; P.H. conceived and designed research.