Hydrogen sulfide is gaining acceptance as an endogenously produced modulator of tissue function. The present paradigm of H2S (diprotonated, gaseous form of hydrogen sulfide) as a tissue messenger consists of H2S being released from the desulfhydration of l-cysteine at a rate sufficient to maintain whole tissue hydrogen sulfide concentrations of 30 μM to >100 μM, and these tissue concentrations serve a messenger function. Utilizing physiological concentrations of l-cysteine and aerobic conditions, we found that catabolism of hydrogen sulfide by mouse liver and brain homogenates exceeded the rate of enzymatic release of this compound such that measureable hydrogen sulfide release was less with tissue-containing vs. tissue-free buffers. Analyses of the gas space over rapidly homogenized mouse brain and liver indicated that in situ tissue hydrogen sulfide concentrations were only about 15 nM. Human alveolar air measurements indicated negligible free H2S concentrations in blood. We conclude rapid tissue catabolism of hydrogen sulfide maintains whole tissue brain and liver concentrations of free hydrogen sulfide that are three orders of magnitude less than conventionally accepted values and only 1/5,000 of the hydrogen sulfide concentration (100 μM) required to alter cellular function in vitro. For hydrogen sulfide to serve as an endogenously produced messenger, tissue production and catabolism must result in intracellular microenvironments with a sufficiently high hydrogen sulfide concentration to activate a local signaling mechanism, while whole tissue concentrations remain very low.
- cell signaler
- nitric oxide
- cytochrome c oxidase
until recently, the physiological significance of hydrogen sulfide was that of a malodorous toxin with a median lethal dose for rodents comparable to that of cyanide (26). However, in 1989, Warenycia et al. (37) reported that brain tissue of rats normally contained relatively high concentrations (mean: 54 μM) of free hydrogen sulfide, and subsequent studies showed similar high concentrations in a variety of mammalian tissues. In addition, it has been shown that desulfhydration of l-cysteine by cystathionine β-synthase (EC 4.2.22) and cystathionine γ-lyase (EC 220.127.116.11) provides an endogenous source of hydrogen sulfide (30). Finally, in vitro studies have found that exposure to hydrogen sulfide altered multiple tissue functions (1, 8, 16, 17, 18, 27, 32–34, 38–41). These observations have elevated the status of hydrogen sulfide from that of a malodorous toxin to an endogenously produced gaseous mediator of cellular activity in various tissues. The burgeoning scientific literature supporting the concept that hydrogen sulfide is a physiological messenger has been the subject of several review articles (10, 35).
A major piece of evidence supporting the concept that hydrogen sulfide serves as an endogenously produced signaling molecule is the multiple in vitro studies demonstrating that cellular activity is altered when hydrogen sulfide is added to the bathing solution. (In aqueous solutions at pH 7.4, about one-fourth of free hydrogen sulfide exists as H2S and three-fourths as HS−; in this paper the sum of these two forms will be referred to as free hydrogen sulfide, with the term H2S reserved for the diprotonated, gaseous form.) Some of the brain functions influenced by exposure to hydrogen sulfide are reduction of hypothalamic release of corticotrophin (8), facilitation of long-term potentials in the hippocampus (1), and inhibition of synaptic transmission in microglial cells of the brain stem (17). In addition, it has been shown that hydrogen sulfide causes increased intestinal secretion (27), increased intracellular calcium concentration (18), increased substance P in pancreatic cells (32), inhibition of cell proliferation (38), opening KATP channels in arterial tissue (33, 39), and relaxation of arterial (16, 40) and ileal (34) smooth muscle. Of considerable importance is that virtually all of these experiments required a bathing solution of hydrogen sulfide concentration of at least 100 μM to induce an appreciable tissue response.
Normal whole tissue is said to contain appreciable concentrations of hydrogen sulfide: e.g., brain: 47 μM to 166 μM (12, 14, 21, 25, 37); liver: 26 μM (23), 144 μM (21); kidney: 40 μM (23), 200 μM (21); heart: 130 μM (23); and lung: 30 μM (9). Recently reported plasma concentrations of H2S include values of 38 μM (42), 34 μM (5), 46 μM (41), and 274 μM (14). Thus, the conventional model of hydrogen sulfide production/distribution consists of the major organs producing this compound via the enzymatic degradation of l-cysteine at a rate sufficient to maintain whole tissue hydrogen sulfide concentrations of 30 μM to >100 μM. Presumably, this production raises the plasma level to roughly 40 μM, which would result in perfusion of the entire body with sizeable hydrogen sulfide concentrations.
The purpose of the present report is to describe the results of studies suggesting that under physiological conditions, hydrogen sulfide resulting from tissue desulfhydration of l-cysteine is rapidly catabolized such that whole tissue and blood concentrations of free hydrogen sulfide are orders of magnitude less than conventionally accepted values.
MATERIALS AND METHODS
Animal and human studies were approved by the appropriate Review Committees of the Minneapolis Veterans Administration Medical Center.
Release of H2S by tissue homogenates at varying l-cysteine concentrations.
Measurements were made using homogenized tissue incubated in a media similar to that used initially by Stipanuk and Beck (30) and, subsequently, by multiple other investigators (1, 16, 23, 40, 42 ). Immediately after death with CO2, liver or brain tissue was removed from male C57BL/6 mice (weight, 20–30 g; Harlan-Teklad, Madison, WI) and homogenized in ice-cold potassium phosphate buffer (pH 6.8) to yield a tissue concentration of 100 mg tissue per ml of buffer. Aliquots of the tissue homogenate (0.3 ml containing 30 mg of tissue) were added to prewarmed (37°C) 50 ml polypropylene syringes, along with 0.1 ml of 20 mM pyridoxal 5′ phosphate and 0.5 ml of potassium phosphate buffer (pH 7.4) and 0.1 ml of an aqueous solution containing 100 mM, 10 mM, or 1 mM l-cysteine (yielding final concentrations of l-cysteine of 10, 1, and 0.1 mM). Control syringes contained no tissue, but all other contents were identical to that described for the active preparation. For aerobic experiments, 50 ml of air was added to the syringes, which were sealed with stopcocks and incubated at 37°C for 30 min, when an aliquot of the gas space was removed for H2S analysis. For anaerobic experiments, the syringes were first flushed with 50 ml of N2, and then 50 ml of N2 was added to the syringe, which was sealed and handled in identical fashion to that described for the aerobic studies.
Concentration of hydrogen sulfide in brain and liver tissue of mice.
Liver or cerebral tissue was obtained from mice either immediately after death via CO2 exposure or while the animal was under pentobarbital anesthesia. The liver was exposed via a midline laparotomy, and the brain was exposed via removal of the posterior portion of the cranium. Immediately upon exposure of the tissue, a roughly 0.1 g portion of liver or whole brain was removed and rapidly transferred to a 5-ml polypropylene syringe and the barrel of the syringe was reinserted, such that about 5 ml of air remained in the syringe. This tissue-containing syringe was attached via a three-way stopcock to a second 5-ml polypropylene syringe containing 0.5 ml of ice-cold 150 mM phosphate buffer (pH 5.7). The gas in the syringe (∼5 ml), buffer, and tissue were then very rapidly injected back and forth ∼25 times between the two syringes. This process finely minced the tissue as it repeatedly passed through the small opening in the stopcock and thoroughly mixed the gas space with the tissue homogenate. A gas sample for H2S analysis was then aspirated into a 1-ml syringe via a needle inserted through an injection septum attached to the third arm of the stopcock, and a 0.3 ml aliquot of the gas space was analyzed. The time elapsing between removal of the tissue segment from the body and aspiration of the gas sample for analysis averaged about 30 s. Five minutes following the initial analysis, after vigorous mixing, a second gas sample was removed for H2S analysis. The syringes, stopcock, septum, and buffer were weighed prior to and after the addition of the tissue, and tissue weight was determined by difference. The volume of gas in the syringe was noted. To test the accuracy of the calculation of the hydrogen sulfide content of tissue from measurement of H2S concentration in the gas space, recovery studies were performed in which 0.1 ml of a pH 7.4 solution containing 2.7 μM sodium sulfide was added to the homogenization device containing 0.5 ml buffer and 5 ml of gas. The solutions and gas were mixed using the same procedure used to homogenize tissue, and the gas space was then sampled and analyzed for H2S concentration. The quantities of hydrogen sulfide in the gas and liquid phases were calculated and summed and the total compared with the quantity added to determine %recovery, which averaged 94 ± 6%.
Concentration of H2S in alveolar air.
The relatively high lipid and water solubility of H2S allows this gas to rapidly permeate the alveolar membrane, and a near-perfect equilibrium between blood and alveolar air would be expected. Ten healthy volunteers held their breath for a 15-s period with their mouths wide open to minimize the accumulation of bacterially produced H2S in the oral cavity. The subjects then sealed their lips around the arm of a stopcock attached to a 20-ml polypropylene syringe and rapidly exhaled around the stopcock. Near the end of exhalation a 20-ml sample of end alveolar air was aspirated into the syringe. The end expiratory collection was assayed for CO2 and H2S. The H2S concentration of a concurrently collected atmospheric sample also was determined.
Sodium sulfide was diluted in 150 mM, pH 7.4, phosphate buffer to yield hydrogen sulfide concentrations of 50, 5, and 1 μM. Two hundred milliliters of each of these solutions and a solution containing only buffer were placed in four coded 250-ml polypropylene bottles sealed with a gas-tight lid. Three blinded observers removed the lids, sniffed the air above the top of the bottle, and recorded their impressions of the odor.
The concentration of CO2 in alveolar air samples was determined using infrared spectroscopy (Capstar-100; CWE, Ardmore, PA). The concentration of H2S in the various samples was determined using a gas chromatograph equipped with a chemiluminescence sulfur detector (model 355; Sievers, Boulder, CO). A 0.3-ml gas sample was injected onto a 2.4-m long, 3.1-mm diameter Teflon column packed with Chromosil 330, maintained at 80°C. The carrier gas was nitrogen with a flow rate of 25 ml/min. The concentration of H2S was determined via comparison of peak area of the unknown to that of known standards. The sensitivity of the machine is such that a 1-ppb concentration yields a peak roughly three times that of background noise.
The aqueous solubility of H2S, per se, is independent of the pH of the solution (4). Aqueous solubilities of H2S at various temperatures, obtained from the paper of Carroll and Mather (4), are 4.3, 2.2, and 1.6 ml of H2S/ml water/760 Torr H2S at temperatures of 4°C, 20°C, and 37°C, respectively. The concentration of HS− in solution for a given concentration of dissolved H2S was calculated from the pH of the solution and the pKa for H2S ↔ HS− of 6.9 (7). At physiological pH, the ratio of H2S to HS− is roughly 1:3, while at pH 5.7 the ratio is about 25:1. The total hydrogen sulfide (H2S + HS−) in a solution was calculated from the sum of H2S dissolved in solution plus the HS− that would exist for that concentration of H2S, based on the pH of the solution. In calculating hydrogen sulfide release from homogenates, we summed the H2S in the gas space and the H2S and HS− calculated to exist in the liquid phase.
Measurements of tissue release of hydrogen sulfide at varying concentrations of l-cysteine.
The rates of release of hydrogen sulfide during aerobic incubation of tissue-containing or tissue-free media containing variable quantities of l-cysteine are shown in Table 1. Results are expressed as: 1) hydrogen sulfide released during incubation of the 1 ml of tissue-containing and tissue-free buffer; 2) hydrogen sulfide released per gram of tissue during incubation of the tissue-containing buffer, and 3) the rate of release of hydrogen sulfide attributable to tissue activity, calculated from the difference between hydrogen sulfide release by the tissue-containing and tissue-free buffers.
Aerobic incubation of liver tissue and 10 mM l-cysteine released readily detectable hydrogen sulfide (260 ± 6.0 pmol·min−1·ml−1 of incubate or 8.7 nmol·min−1·g of tissue−1). However, appreciable hydrogen sulfide release (190 ± 31 pmol·min−1·ml−1) was observed for the 10 mM l-cysteine-containing, tissue-free buffer control, and the difference between the active and control incubates indicated that liver tissue activity released hydrogen sulfide at a mean rate of 2.3 ± 0.1 nmol·min−1·g of tissue−1. Hydrogen sulfide release was undetectable (<0.1 pmol·min−1·ml−1) when liver homogenates were incubated with 1.0 mM and 0.1 mM l-cysteine. Since there was appreciable hydrogen sulfide release by the tissue-free media at these l-cysteine concentrations, the difference between tissue-containing and tissue-free media was negative, i.e., tissue inhibited hydrogen sulfide release.
Aerobic incubation of brain tissue resulted in low rates of hydrogen sulfide release at all l-cysteine concentrations; however, the release of hydrogen sulfide was always greater with the tissue-free control, yielding negative values for hydrogen sulfide production attributable to tissue activity at each l-cysteine concentration, i.e., brain tissue inhibited hydrogen sulfide release.
Incubation of liver tissue and 10 mM l-cysteine under anaerobic conditions demonstrated copious hydrogen sulfide release relative to that of the tissue-free buffer, yielding hydrogen sulfide production by liver tissue of 106 nmol·min−1·g of tissue−1 (see Table 2). However, during incubations with 1 mM and 0.1 mM cysteine, hydrogen sulfide release declined by > 1,000-fold to values below that of the tissue-free buffer (i.e., once again negative values for tissue activity were obtained). At all l-cysteine concentrations, anaerobic incubation of brain tissue released quantities of hydrogen sulfide that were greater than that of the tissue-free controls, yielding low, but measureable rates of hydrogen sulfide release attributable to tissue activity (1.2 to 0.017 nmol·min−1·g of tissue−1).
Hydrogen sulfide concentration generated in buffer.
The hydrogen sulfide concentration generated in the buffer during aerobic incubation of liver with varying concentrations of l-cysteine (calculated from the H2S concentration of the gas space) averaged 725 ± 16 nM for 10 mM l-cysteine and was undetectable < 0.1 nM for l-cysteine concentrations of 1 mM and 0.1 mM. For aerobic incubation of brain tissue, the hydrogen sulfide concentrations averaged 111, 15, and 5.3 nM, respectively, for the 10, 1.0, and 0.1 mM l-cysteine concentrations. For anaerobic incubations, the hydrogen sulfide concentrations of the liver-containing buffer averaged 8,950, 27, and 5.3 nM, respectively, for 10, 1.0, and 0.1 mM l-cysteine concentrations with comparable values for brain-containing buffer of 248, 28, and 5.3 nM, respectively.
Concentration of hydrogen sulfide in brain and liver tissue of mice.
Studies were carried out with liver and brain tissue obtained from mice immediately after death via CO2 inhalation or while under anesthesia with pentobarbital. The calculated hydrogen sulfide concentrations of tissue removed under these two sets of conditions showed no significant differences, and the results of all measurements obtained for each type of tissue were combined and are shown in Table 3. The calculated concentrations of free hydrogen sulfide in brain and liver tissue following the 30-s homogenization process averaged 14 ± 3 nM and 17 ± 3 nM, respectively. Repeat measurements obtained 5 min following the initial measurements showed a decline in H2S concentration in the gas space with a concomitant decline in calculated tissue hydrogen sulfide concentration of brain and liver tissue of about 15% and 47%, respectively.
Typical gas chromatographic tracings of the H2S concentrations of the gas space above a brain homogenate (0.078 g of brain) are shown in Fig. 1. The tracings represent the H2S peaks obtained: at the initial measurement (A); 5 min after the initial measurement (B), and following instillation of 0.1 ml of a 50 μM sodium sulfide solution (the putative concentration in normal brain) into the homogenate-containing syringe (C). As shown in the figure, in contrast to the small H2S peaks derived from the tissue, the sulfide-containing solution yielded an H2S peak that was far off scale and not quantifiable due to saturation of the detector at this concentration of H2S.
Concentration of H2S in atmosphere and alveolar air.
All expired air samples had a CO2 concentration of > 5%, indicating that they represented alveolar air. The concentration of H2S in the atmosphere at the time of collection of the samples averaged 55 pmol/l of air, which is equivalent to a concentration of about 1.2 ppb. The concentrations of H2S in the alveolar air samples of the 10 healthy subjects were only marginally greater than that of the concurrently assessed atmospheric concentrations with values ranging from 0 to 67 pmol/l (mean: 15.3 ± 8.5 pmol/l or 0.33 ± 0.16 ppb) above that of the atmosphere.
Each of the three blinded judges rated the odor of the 50 μM hydrogen sulfide solution to be very obnoxious, the 5 μM solution to be unpleasant, and the 1 μM solution to be detectable, but weak. The control sample containing no hydrogen sulfide was always identified as having no odor, relative to that of the other samples.
The principal reaction yielding H2S in tissue is thought to be desulfhydration of l-cysteine catalyzed by cystathionine β-synthase and cystathionine γ-lyase, enzymes found in varying concentrations in a wide variety of tissue. Virtually every study demonstrating such enzymatic production of H2S release has employed tissue homogenates incubated anaerobically with 10 mM l-cysteine (1, 16, 23, 40, 42). However, the physiological situation is aerobic with intracellular cysteine concentrations maintained at < 0.1 mM (19, 31). The commonly employed 10 mM l-cysteine concentration is toxic to cells (20, 24, 36). Finally, l-cysteine spontaneously decomposes to hydrogen sulfide in aqueous solutions (see Tables 1 and 2), and accurate assessment of enzymatically mediated hydrogen sulfide production requires appropriate correction for this spontaneous decomposition.
We measured the rate of H2S release into the gas space of a syringe during incubation of buffer containing mouse liver or brain homogenates and l-cysteine concentrations ranging from 10 mM (the conventionally employed concentration) to 0.1 mM (the physiological concentration) under aerobic and anaerobic conditions. The release of H2S observed during incubation of tissue-free buffer was employed to correct for spontaneous (nontissue mediated) decomposition of l-cysteine to hydrogen sulfide. As shown in Table 1, the rate of release of hydrogen sulfide during aerobic incubation of liver homogenates with 10 mM l-cysteine averaged 260 ± 6.0 pmol·min−1·ml−1 of incubation media or 8.7 ± 0.20 nmol·min−1·g liver tissue−1. When corrected for the appreciable hydrogen sulfide release of the nontissue-containing control, release of free hydrogen sulfide attributable to tissue activity averaged 2.3 ± 0.10 nmol·min−1·g tissue−1. The unexpected observation was that hydrogen sulfide release by the liver-containing buffer was undetectable (< 0.003 nmol·min−1·g tissue−1) at l-cysteine concentrations of 1 mM or 0.1 mM, while the tissue-free media released readily detectable H2S at these l-cysteine concentrations. Thus, at l-cysteine concentrations of 1.0 mM and 0.1 mM, the presence of liver tissue actually inhibited measurable hydrogen sulfide release.
During aerobic incubation with 10 mM l-cysteine, the release of free hydrogen sulfide by buffer containing brain tissue was appreciably less than that observed with liver (see Table 1). This release rate declined sharply when l-cysteine concentrations were reduced to 1.0 and 0.1 mM, but release but was still detectable, in contrast to the undetectable release observed with liver. However, spontaneous (nontissue mediated) hydrogen sulfide release exceeded that observed for the brain-containing incubates at all l-cysteine concentrations. Thus, brain tissue inhibited measureable hydrogen sulfide release at all l-cysteine concentrations employed.
Anaerobic incubation of liver and brain tissue (see Table 2) yielded higher tissue hydrogen sulfide release rates at all l-cysteine concentrations compared with that observed with aerobic incubation. However, values for liver incubated with 1.0 or 0.1 mM l-cysteine were very low, less than those observed during incubation of tissue-free buffer. Anaerobic incubation of brain tissue showed low rates of hydrogen sulfide release that were greater than the spontaneous production at all l-cysteine concentrations, i.e., low rates of tissue-mediated release of hydrogen sulfide from l-cysteine was demonstrated at all l-cysteine concentrations for brain tissue.
These results raise the question of how liver and brain tissue could inhibit, rather than enhance, observed hydrogen sulfide release relative to that of tissue-free buffer when incubations were carried out under physiological conditions (aerobically with l-cysteine concentrations of 1.0 or 0.1 mM). The answer appears to be provided by the observations that l-cysteine spontaneously decomposes in the buffer, releasing hydrogen sulfide (see Tables 1 and 2) and that hydrogen sulfide is rapidly catabolized by a variety of tissues (11). Thus, the tissue-containing homogenates are simultaneously producing and catabolizing hydrogen sulfide, and the observed accumulation of H2S in the gas space reflects the net of absolute production minus consumption. With 10 mM l-cysteine, the absolute production of hydrogen sulfide (enzymatic plus spontaneous production) in the liver-containing homogenate exceeded the rate of consumption, and net production of hydrogen sulfide was observed. However, when the l-cysteine concentration was reduced to ≤ 1 mM, hydrogen sulfide release by tissue was negligible. Tissue enzymes presumably were releasing hydrogen sulfide at these lower cysteine concentrations, but the rate of release was sufficiently slow that consumption of hydrogen sulfide by liver tissue consumed all available hydrogen sulfide. Since spontaneous decomposition of l-cysteine to hydrogen sulfide in the tissue-free buffer was not subject to tissue catabolism, hydrogen sulfide release from tissue-free buffer exceeded that of the tissue-containing buffer. Thus, liver tissue actually inhibited net hydrogen sulfide release from these lower concentrations of l-cysteine. While detectable hydrogen sulfide release was observed during aerobic incubation of brain tissue with each concentration of l-cysteine, these values were less than those observed with tissue-free buffer, i.e., brain tissue also inhibited the net production of hydrogen sulfide.
Surprisingly, the finding that liver and brain homogenates may actually inhibit, rather than enhance, hydrogen sulfide release from incubates containing l-cysteine appears to be a novel observation. This cannot be explained by a difference in incubation technique since we used a media identical/similar to that employed by previous investigators, although our H2S measurements were performed on the gas space over the homogenate, whereas most previous studies trapped H2S in a zinc acetate solution and then measured hydrogen sulfide in the solution. We believe that the failure of previous investigators (1, 16, 23, 40, 42) to note that tissue may inhibit H2S release is probably attributable to their use of anaerobic conditions and 10 mM concentrations of l-cysteine in contrast to our studies that were carried out aerobically using lower, more physiological concentrations (1 mM and 0.1 mM) of l-cysteine. Both of these alterations favor hydrogen sulfide consumption over production (see Tables 1 and 2). Of note is that the one previous report (31) that measured H2S release by tissue at varying l-cysteine concentrations observed markedly diminished H2S release when the l-cysteine concentration was reduced from higher concentrations to 2 mM.
The concentration of hydrogen sulfide generated in our homogenates during incubation with l-cysteine was calculated from the H2S concentration observed in the gas space. Under aerobic conditions these values at 10 mM l-cysteine averaged 109 nM and 725 nM for brain and liver homogenates, respectively. However, at potentially physiological l-cysteine concentrations of 1.0 mM and 0.1 mM, these values were, respectively, only 15 nM, and 5.5 nM for brain homogenates and < 0.1 nM for liver homogenates, concentrations < 1/1,000 of the 30 μM to > 100 μM free hydrogen sulfide that have been reported to be present in whole tissues and blood.
The low free hydrogen sulfide concentrations achievable in tissue homogenates during aerobic incubation with physiologic concentrations of l-cysteine raised a question as to the validity of the relatively high whole tissue free hydrogen sulfide concentrations reported in the literature. Hydrogen sulfide can exist in free or bound forms, a distinction that is important since bound sulfhydryl groups cannot function as a gaseous messenger. Publications touting H2S as a regulator of cellular function have treated reported 30 μM to >100 μM tissue and plasma hydrogen sulfide concentrations as a measure of the free form of this molecule.
Multiple theoretical arguments can be raised against the concept that whole tissue normally contains free hydrogen sulfide concentrations of 30 μM to > 100 μM. The human nose is very sensitive to the characteristic odor of H2S (2) and blinded observers reported that a pH 7.4 solution containing 50 μM hydrogen sulfide reeked of H2S, 5 μM had a distinctly unpleasant odor, and 1.0 μM had a faint but detectable H2S odor. Since tissue and blood do not have an H2S odor, tissue and blood concentrations seemingly must be > 1 μM. The toxicity of hydrogen sulfide results, in part, from its ability to inhibit cytochrome c oxidase. Brain cytochrome c oxidase is 80% inhibited at an H2S concentration of 1 μM (22); thus, the conventionally accepted whole tissue hydrogen sulfide concentrations of > 30 μM tissue would render cytochrome c in a perpetual state of inhibition. Brain death following administration of hydrogen sulfide occurred with brain concentrations of hydrogen sulfide only 20% greater than the reported normal 54 μM concentration (37). Subsequently reported normal brain hydrogen sulfide concentrations have far exceeded the observed lethal concentration. In addition, a recent study showed that mice exposed to an inhaled H2S concentration of 80 ppm went into a state resembling suspended animation, and body temperature fell significantly at an inhaled concentration of only 20 ppm (3). The maximal tissue free hydrogen sulfide concentrations that could result from 80 ppm and 20 ppm of H2S are about 25 μM and 6 μM, respectively, concentrations below those said to normally exist in tissue. The body seemingly would not normally maintain hydrogen sulfide concentrations that are close to or exceed the concentrations that induce suspended animation or death. Finally, given the rapid metabolism of hydrogen sulfide by many tissues, an enormous quantity of l-cysteine would be required to maintain a > 30 μM tissue and blood concentration of free hydrogen sulfide. For example, at a 30 μM hydrogen sulfide concentration, we (11) found that rat liver-containing homogenates metabolized hydrogen sulfide at a rate of 2 nmol·min−1·mg protein−1. Assuming a similar metabolic rate for human liver, this value extrapolates to about 0.9 mol/day for the whole liver, a value well in excess of the daily requirement of l-cysteine plus methionine [about 0.01 mol (1.0 g) per day].
Previous measurements of H2S concentration in tissue have employed various extraction techniques to remove hydrogen sulfide from tissue. In contrast, in the present study, we rapidly homogenized tissue in a gas-tight syringe, and the concentration of hydrogen sulfide initially present in tissue was calculated from the H2S concentration in the gas space over the homogenate. Liver and brain were found to have mean free hydrogen sulfide concentrations of only 17 ± 8 nM and of 14 ± 9 nM, respectively, values that are three orders of magnitude less than conventionally accepted values. The enormous discrepancy between the very low H2S concentration we observed in the gas space over a brain tissue homogenate vs. that expected if brain had the commonly accepted concentration of 50 μM hydrogen sulfide is “gas chromatographically” demonstrated in Fig. 1.
The whole tissue concentration of hydrogen sulfide reflects a balance between ongoing production and removal of this compound in the tissue. Hydrogen sulfide can be removed from tissue via uptake into the blood or oxidation to thiosulfate and sulfate. A potential criticism of our measurement of hydrogen sulfide in tissue, as well as all previous measurements, is that the in vivo balance between tissue production and catabolism of hydrogen sulfide may have been altered by the experimental procedure. Our finding that blood contains negligible free H2S indicates that the lack of blood perfusion of tissue during the analytical process could only increase our observed tissue hydrogen sulfide concentration relative to that in situ. The effect of our homogenization process on hydrogen production vs. catabolism is not known. For that reason, we minimized the time elapsing between removal of tissue from the body and analysis to a period of only 30 s. In addition, tissue homogenization was carried out in buffer at ice water temperature to minimize both enzymatic production and consumption of hydrogen sulfide. Given that gas space measurements obtained 5 min after the initial assay showed only moderate declines in hydrogen sulfide concentration, we conclude that our 30-s measurement provided a reasonably accurate assessment of the in situ, whole tissue concentration of free hydrogen sulfide (about 15 nM) in liver and brain tissue. For comparison, measurements of NO concentration of normal tissue ranges from 1 nM to 50 nM (13), and nM changes in tissue NO concentration are associated with alterations in tissue function (28).
We also reassessed the concept that the hydrogen sulfide content of blood is on the order of 40 μM. The lung was employed as an equilibrator to obtain in vivo measurements of the free H2S concentration in the blood of healthy human subjects. The relatively high lipid and water solubility of H2S allows this gas to rapidly permeate the alveolar membrane, and a near perfect equilibrium between blood and alveolar air would be expected. The concentrations of H2S in 10 alveolar air samples was found to be only marginally greater than that of the atmospheric concentration, with a mean concentration of 15 pmol/l above atmospheric. This value corresponds to a free hydrogen sulfide concentration (H2S plus HS−) in the blood of only about 100 pM, > 1/100,000 of reported values of 40 μM value. This very low blood concentration provides additional evidence against the existence of free hydrogen sulfide concentrations of 30–100 μM in all the major organs since equilibration of perfusing blood with such tissue H2S concentrations would be expected to produce μM concentrations of H2S in the blood.
The very low tissue and plasma hydrogen sulfide concentrations found in our studies prompts the question of why multiple investigators have reported tissue hydrogen sulfide concentrations that are orders of magnitude greater than the values observed in the present study. It appears that the extraction processes previously employed removed both bound (such as acid labile hydrogen sulfide) and free forms of this molecule. Clearly this was the case with several studies (15, 23) frequently cited as showing tissue H2S concentrations of > 30 μm. In addition, the extraction processes might have allowed tissue H2S production to occur while metabolism was inhibited. Seemingly the only explanation for the reports of plasma H2S concentrations of ∼40 μM measured with a hydrogen sulfide-sensitive electrode is that the electrode was not specific for free hydrogen sulfide.
Our finding that tissue free hydrogen sulfide concentrations are on the order of 15 nM (rather than the presently accepted values of 30 μM to >100 μM) have implications for the concept that H2S serves as an endogenously produced gaseous intracellular messenger. Given that virtually all in vitro studies have required 100 μM hydrogen sulfide to alter cell function, whole tissue concentrations of free hydrogen sulfide seemingly are insufficient to serve a messenger function. Thus, it is necessary to postulate that intracellular hydrogen sulfide production and consumption result in tissue microenvironments with free hydrogen sulfide concentrations (presumably 100 μM) sufficient to induce a biological effect, while a much lower concentration is maintained in whole tissue. This situation requires the existence of a mechanism in this microenvironment whereby hydrogen sulfide can act as a local effector as opposed to a mechanism that requires diffusion of hydrogen sulfide throughout the entire cell or between cells.
The low whole tissue and blood free hydrogen sulfide concentrations observed in our studies would greatly facilitate the ability of this H2S to function as a rapidly adapting mediator of cellular activity. If organs normally were maintained at the presently accepted whole tissue concentrations 30 μM to >100 μM hydrogen sulfide and were perfused with blood containing 40 μM hydrogen sulfide, enormous changes in endogenous production would be required to appreciably alter tissue hydrogen sulfide concentration. In contrast, relatively minor changes in endogenous release could rapidly induce sizeable local changes in concentration in the situation where the overall ambient free hydrogen sulfide concentration of tissue is maintained at low (nM) levels and the free hydrogen sulfide concentration in blood perfusing the tissue is negligible.
Perspectives and Significance
The present study demonstrates that whole tissue concentrations of free hydrogen sulfide are orders of magnitude less than the concentrations required to alter tissue function. To serve as an intracellular messenger, yet-to-be-described mechanisms must exist that allow very localized high intracellular concentrations of free hydrogen sulfide to act as an effector, i.e., a system comparable to that of a ligand and its receptor. Elucidation of the effector mechanism should provide important insights into how hydrogen sulfide performs its varied messenger functions and how this function might be altered for therapeutic purposes.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society