Abstract
Sodium hydrosulfide and dimethylsulfide duplicate the effects of hydrogen sulfide in causing coma in Sprague-Dawley rats and are additive for lethality. Nitrite, pyruvate and dithiothreitol had no significant effect on coma or lethality but bicarbonate with and without glucose reduced duration of coma. This finding suggests an antidotal treatment.
INTRODUCTION
Hydrogen sulfide (H2S) is a highly toxic chemical comparable to cyanide in lethality (Burnett et al. 1977; Evans 1967; Sollman 1957). It is a toxic hazard in more than seventy varieties of workplace, including the oil and gas industry, agriculture, and sewage treatment facilities (WHO 1981). Hydrosulfide (HS–) is the active agent in H2S as well as in its inorganic analogues, NaHS and Na2S. The three main toxic manifestations of acute exposure to H2S are death, apnea and unconsciousness. The latter is called “knockdown” in the gas and oil industry and is here referred to as “coma”. Of the three, coma is the most common in humans but least studied in animal models (Warenycia et al. 1990). In contrast, organic analogues of H2S, such as methyl mercaptan (MeSH), dimethyl sulfide (Me2S) and dimethyl disulfide (Me2S2) can reproduce the toxic effects of H2S in rats, including coma (Ljunggren and Norberg 1943; Tansy et al. 1981; Zieve et al. 1974).
Me2S occurs naturally as a decomposition product of plants and animals (Sandmeyer 1981). It is the methyl analogue of H2S, and has been shown to produce similar effects with a markedly diminished potency of about 50-fold for white female rats and 40-fold for Sprague Dawley rats (Ljunggren and Norberg 1943; Warenycia et al. 1990; Zieve et al. 1974). Me2S is also a product of metabolism, and is found in the breath and urine of patients in hepatic coma (Chen et al. 1970).
The pathophysiological explanation that death from H2S is a consequence of the inhibition of cytochrome oxidase in the brain stem became conventional wisdom with time, despite the absence of supporting evidence (Guidotti 1994, 2006, 2007; Reiffenstein et al. 1992; WHO 1981). Current treatment options for acute H2S toxicity are based on three assumptions: 1) that its mechanism of action bears similarity to that of cyanide, 2) that apnea is a consequence of inhibition of the respiratory centre in the brain stem, and 3) that both H2S and cyanide inhibit cytochrome oxidase in the brain stem. (Sollman 1957; Evans 1967; Beauchamp et al. 1984; Burnett et al. 1977; Cittadini et al. 1972; Guidotti 2006, 2007; Haggard 1925; Reiffenstein et al. 1992; Smith 1986; WHO 1981). However, direct evidence in support of these assumptions is lacking.
Evidence against this postulate has accumulated gradually. For example, the concentration of sulfide in gills was shown to exceed that in brain of fish exposed to H2S (Torrans and Clemens 1982); secondly, no correlation was found between enzyme activity and inhibition of respiratory activity of mitochondria exposed to cyanide; and finally, the inhibition of respiratory rhythm of the isolated neo-natal rat brain stem by HS–or by cyanide was not achieved with doses that inhibited respiration (Greer et al. 1995; Greer and Carter 1995). More recently, in vivo studies with rats have shown that delivery of HS– to the lung results in a much greater apneic response and at much lower concentrations than does delivery to the brain; also, the blockade of neural signals from lung to brain prevents apnea due to HS–(Almeida and Guidotti 1997Almeida and Guidotti 1999). Therefore, although apnea is executed by the respiratory center, it appears to be signaled by the lung.
In recent years, sodium nitrite at dosage levels of 75 mg/kg, (Smith et al. 1976) pyruvate at dosage levels of 1 g/kg (Dulaney and Hume 1988) and dithiothreitol at dosage levels of 80 mg/kg (Warenycia et al. 1990) have all been studied as antidotes to sulfide toxicity in rats, with some reports of use in humans (Gunn and Wong 2001; Hall and Rumack 1997; Nam et al. 2004).
Two alternative antidotal therapies are independent of the cyanide-analogy hypothesis. These are the use of hyperbaric oxygen and the administration of sodium bicarbonate. Although hypoxemia and a reduced oxygen-carrying capacity in hemoglobin have not been demonstrated in human victims, oxygen and hyperbaric oxygen are still used for the treatment of human victims, and oxygen for apnea is a reasonable treatment even without an underlying theory of mechanism (Guidotti 2007; Gunn and Wong 2001; Nam et al. 2004). An alternate approach to antidotal therapy is the use of bicarbonate, which was attempted in a controlled study to prevent hyperpnea, apnea and death on anesthetized Sprague Dawley rats. Bicarbonate was successful against lethality, but coma was not studied (Almeida and Guidotti 1999).
In this study, we have examined the acute toxicity of Me2S on Sprague Dawley rats with respect to incidence and duration of coma, and its interaction with NaHS. The effects of three antidotal interventions, nitrite, pyruvate and dithiothreitol, on the duration and incidence of coma and on lethality were also examined. Finally, the effects of bicarbonate and glucose as pre-treatments and rescue treatments in acute Me2S toxicity, and as pretreatment in NaHS lethality were examined as a possible treatment option.
MATERIALS AND METHODS
Chemicals
Anhydrous Me2S and NaHS•nH2O were obtained from Aldrich Chemical Company (USA); NaNO2 (nitrite), NaCl, NaHCO3 (bicarbonate) and glucose were obtained from Fisher Scientific Company (Canada); CH3COCO2Na (pyruvate) and dithiothreitol (DTT) were obtained from Sigma Chemical Company (Canada). The Me2S was used as supplied. An aqueous stock solution of NaHS ca 1 M was quickly prepared on receipt of the reagent, and 0.5 ml aliquots sealed under nitrogen in glass ampoules to prevent the rapid decomposition of NaHS. The concentration was determined by iodometric assay and remained constant for up to 2 years with the ampoules stored at room temperature. A fresh ampoule was used for each experiment.
Animals
Sprague Dawley male rats, 300–350 g, were obtained from the Health Sciences Lab Animal Services of the University of Alberta. The animals are bred and nurtured on the premises. They are housed at room temperature, fed PMI diet #5001, a Commercial preparation, and receive tap water to drink. The use of animals was approved by the Health Sciences Animal Welfare Committee of the University of Alberta, which strictly adheres to the requirements of the Canadian Council on Animal Care. Approval of protocols by the Committee is made after strict review, with particular attention paid to the number of animals, their treatment and the method of euthanasia.
Methods
All agents were administered by intraperitoneal injections. Nitrate 75 mg/kg, DTT 50 mg/kg in saline, and pyruvate 1g/kg in aqueous solution, were injected as a pre-treatment or within 30 seconds of loss of consciousness. Animals that did not become comatose were excluded from analysis. Bicarbonate 1.4% in aqueous solution, with and without glucose 5%, was injected as a pre-treatment or within 30 seconds of coma induction. Animals were observed constantly for the first 30 minutes after injection and every 15 minutes thereafter for at least 2 hours.
Coma was defined operationally as a loss of the righting reflex with subsequent recovery to consciousness, as previously demonstrated for mercaptans (Zieve et al. 1974). Animals that died before recovery were not included in analyses for the duration of coma.
Dose-mortality curves were prepared by the cumulative method of Reed and Meunch (1938) which uses fewer animals. In this method, which is a requirement of our institutional animal research committee, the assumption is made that animals responding to a low dose would also respond to a higher dose (and conversely for non-responding animals), so that responding animals in the lower dosage groups are therefore added to the tally of animals responding to higher doses. Since individual animals may therefore appear in several dose groups, inferential statistical methods, which assume the independence of data in each group, cannot be used.
Characterization of Toxicity of Me2S and NaHS
Dimethyl sulfide (Me2S) was tested in the dose range 375 mg/kg to 1500 mg/kg. Animals were observed for induction and duration of coma, and for lethality within 24 hours of injection with Me2S.
Coma induction by NaHS in rats has not been reported. This possibility was examined by using it in combination with Me2S. Low doses of NaHS, 7.5 and 10.0 mg/kg which represent 0.5 and 0.67 times the LD50 respectively (Warenycia et al. 1990), were administered ip, and followed 3 min later by low doses of Me2S. Pre-treatment with NaHS, 7.5 mg/kg was followed by Me2S, 562, 640 or 724 mg/kg, and pre-treatment with NaHS, 10 mg/kg was followed by Me2S, 436, 501 or 562 mg/kg.
Pathological changes were examined in a limited study. Three groups of three animals each were used. Two groups received Me2S, at 938 mg/kg, a submaximal coma dose, and 537 mg/kg, the LD50 determined by regression analysis. The third received a lethal dose of NaHS, 30 mg/kg (Warenycia et al. 1990). All animals were examined macroscopically, and some tissues microscopically, for possible changes that could lead to death. Animals were counted as experiencing a lethal effect if they died within 24 hours of exposure.
Effect of Purported Antidotal Treatments
Duration of coma and lethality were used as the toxic responses, and the three agents nitrite, dithiothreitol and pyruvate were tested for their antidotal efficacy. They were given ip within 30 seconds of onset of coma or as a pretreatment 10 minutes before the dose of Me2S. A sub-lethal dose of Me2S, 938 mg/kg, was selected so as to permit observation of either decreases or increases in duration of coma.
Bicarbonate as an Intervention
Bicarbonate was shown to be highly effective as a pre-treatment in preventing lethality from NaHS in rats under general anesthesia. (Almeida and Guidotti 1999) A more extensive examination of the effects of this agent on coma induced by Me2S 938 mg/kg was undertaken. Bicarbonate, 2 ml of 1.4% aqueous solution was given either at the onset of coma (protocol A), or as a pre-treatment 2 min before the Me2S (protocol B), and in both cases, with or without 5% glucose. Normal saline was given as a control. The antidotal effect of bicarbonate with and without glucose was also tested against NaHS.
RESULTS
Effects on Consciousness
Table 1 shows the incidence of coma from Me2S. The dose range for incidence (Table 1) is relatively small from about 700 mg/kg to a maximal effect at about 1200 mg/kg. The ED50 calculated by regression analysis from a semi-log plot was about 800 mg/kg. At the highest dose, death always preceded recovery from coma. Coma was not induced with doses below about 560 mg/kg. At the second highest dose one animal died before regaining consciousness. The dose dependent curve for coma incidence is normal with a linear mid-range.
Table 2 shows the duration of coma from a range of doses of Me2S alone and in combination with NaHS, at 7.5 or 10.0 mg/kg. Coma duration is short, suggesting a low efficacy for Me2S, with a maximal duration of about 20 minutes, and a limited drug range from about 750 mg/kg to about 1100 mg/kg.
Data from use of a combination of Me2S and NaHS are also shown in Table 2. Previous studies have shown that when given alone, NaHS at 7.5 and 10.0 mg/kg did not produce coma (Warenycia et al. 1990), and this was also observed here. However, when followed by a dose of Me2S, 562 mg/kg or greater, coma was produced. Thus, no coma was induced by either of the agents if administered alone; in combination however, coma lasting for 1.6 min and 3.4 min was produced with NaHS at 7.5 and 10.0 mg/kg respectively, presumably by synergistic action. The coma produced by Me2S 750 mg/kg is increased by more than 80% if NaHS is administered with it. The NaHS, presumably by synergistic action, increases the potency of Me2S with respect to induction of coma.
Lethality of Me2S
Survival after coma lasted for two or more hours depending on the dose of Me2S that was administered. At doses below 300 mg/kg no deaths were recorded. At doses slightly above this, some animals survived for as long as five days. Using a 24-hour post treatment duration, a dose curve for the lethality of Me2S was calculated according to the method of Reed and Meunch (1938) and is presented in Figure 1. The LD50 was 537 mg/kg and the LD100 about 1000 mg/kg with a linear range from about 400 mg/kg to 700 mg/kg. It is noted that the apparent non-lethal dose of about 300 mg/kg applies only to an end point set at 24 h after treatment. It is also noted that the LD50 of 537 mg/kg is smaller than the ED50 for coma at 813 mg/kg (see Figure 1).
Effects of Antidotal Interventions on Lethality and Duration of Coma
The data in Table 3 describes the efficacy of the antidotes. In all cases, regardless of the treatment protocol applied, or the antidote used, no biologically significant difference in the duration of the coma induced by Me2S at 938 mg/kg was observed as compared with control conditions. Furthermore, lethality was also unaffected.
In the study with nitrite, 3 of the pre-treated animals tested did not lose consciousness (i.e. duration = 0), and 3 died without regaining consciousness and so were not included. This did not occur with the other pre-treatment tests.
Bicarbonate as an Intervention in Coma Induced by Me2S
Table 4 summarizes the data on pre-treatment and rescue with bicarbonate after challenge with Me2S at 938 mg/kg. The control experiment showed a mean duration of coma equal to 10.4 min; no animals survived the treatment by 24 hours. Bicarbonate given to animals as a pre-treatment, or at the onset of coma, reduced the duration of coma by about 17% but the lethality of Me2S remained at 100%. Bicarbonate, when given together with glucose, produced a very different result. If given at onset (protocol A), the duration of coma was reduced by about 27% and the mortality fell by 50%. When given as a pre-treatment however (protocol B), the duration of coma was reduced by 80% and lethality was abolished.
This protocol of administration of the two antidotal agents as a pre-treatment was also tested in conscious rats against a lethal dose of NaHS, 30 mg/kg (Almeida and Guidotti 1999) and the results are shown in Table 5. As expected, no animals experienced coma. Bicarbonate alone caused a reduction in lethality of 20% and in combination with glucose of 75%. These results are comparable to those described previously for anesthetized animals (Almeida and Guidotti 1999).
Other Observations
When dimethyl sulfide (Me2S) was tested in the dose range 375 mg/kg to 1500 mg/kg, there was no immediate effect at the lowest dose. With higher doses, ataxia, either alone or accompanied by subsequent loss of the righting reflex was observed, along with an increase in heart rate and rate of respiration. In most cases there was also a period of salivation that increased in duration with the dose of sulfide. On recovery from coma, the animals sustained tremor-like movements of the head and spasms of the neck muscles that lasted for up to 10 minutes. No full-body seizures were observed. Following the period of muscle spasms, the animals were either asleep or alert but quiescent. Animals in this state of recovery survived for varying periods depending on the treatment that was administered. If death occurred, it did so during a period of apparent sleep, or during a brief episode of gasping that lasted less than 30 seconds.
As noted above, three groups of animals were used for a limited study of pathology outcomes. Two groups received Me2S, at 938 and 537 mg/kg, and the third received a lethal dose of NaHS, 30 mg/kg. At the higher dose of Me2S, 938 mg/kg, all animals became unconsciousness within 2.5 minutes. Of these one died in 2 min while the other two regained consciousness after 11 and 8 min respectively, and died within 2.5h. At the lower dose of Me2S, 537 mg/kg, none became unconscious and two died within 24 hours. All animals given a lethal dose of NaHS died after entering coma within two minutes.
Whether examined macroscopically or microscopically, examination revealed no obvious differences between the three groups of animals. The abdominal cavity contained some bloodstained fluid, and the lungs were usually fully inflated with irregular mottling and multiple focal petechial haemorrhages present throughout. The liver and meninges were mildly congested. All other organs were normal within gross limits. The animals were all injected in the lower right abdominal quadrant and with one exception, no injury was detected that could be associated with the injection. One animal had bloody liquid in the mid-jejunum mixed with ingesta. The blood was undigested and appeared to be the result of acute bleeding into the intestine. This may have been an artefact of injection, although no entrance wound was found.
Microscopic studies were carried out on the heart, lung, liver, adrenal glands, various levels of the brain and spinal cord and decalcified sections of middle and inner ear. Focal areas of interstitial cellular accumulation within the myocardium, with intense congestion of the myocardial capillaries and irregular areas of intracellular edema, were noted. Focal areas of hepatocellular degeneration were also noted, with some cells that seemed to be ischemic. No other microscopic anomalies were found.
DISCUSSION
Toxic effects of organic analogues of H2S, namely Me2S, Me2S2 and MeSH have been characterized previously and shown to reproduce important effects of hydrogen sulfide toxicity and to be more easily managed under laboratory conditions (Ljunggren and Norberg 1943; Tansy et al. 1981; Zieve et al. 1974). In particular, it has been shown from studies with female white rats that these agents have a lower lethal potency than H2S and that they do cause coma (Ljunggren and Norberg 1943). From these cited studies and from the data presented here, it appears that Me2S is a suitable agent for study of knockdown in animals.
The coma so induced is dependent on dosage and is readily quantified, both by incidence and duration. The lethal potency of Me2S, LD50at 537 mg/kg, is greater than the potency of the agent for coma, ED50 = 817 mg/kg, so lethality at the lower doses is not accompanied by coma. Thus, it appears that the mechanisms of the two effects are distinct.
The additive or synergistic effect shown by Me2S and NaHS for coma, suggests that a similar mechanism is operating for this outcome for these sulfide compounds. Similarity of the two compounds was also supported by the absence of macroscopic or microscopic changes post mortem examination after lethal doses.
Nitrite, pyruvate, and dithiothreitol, currently proposed and used as treatments, had no effect on the duration of coma and or the lethal potency of Me2S. This and the equivocal data from clinical use, has not prevented their use (Gunn and Wong 2001; Nam et al. 2004).
Bicarbonate, which previously had been shown to provide absolute protection against lethality from NaHS for anesthetised rats (Almeida and Guidotti 1999), had only a marginal effect on coma and no effect on lethality from Me2S. By contrast, in this study of rats treated with Me2S, bicarbonate and glucose prevented lethality and reduced the duration of coma by 81%. The treatment also reduced the lethality of NaHS in conscious rats by 75%.
The data presented here emphasize the importance of selecting the end-point that is used for the determination of lethality. The LD50 of 537 mg/kg for Me2S was obtained with an end-point set at 24 hours. If death during the period of coma had been set as the end-point, the dose curve would almost be a vertical line with LD1 and LD99 equal to about 1500 mg/kg. It would resemble the steep lethality-dose curve seen in many species for H2S or NaHS. (Guidotti, 1996) Two animals from the lowest dose group died two days after the 24-hour limit, showing that an endpoint of lethality after 72 hours would have resulted in a very different dose-response curve for lethality. At some lower dose levels of Me2S, animals will recover from coma and appear to be in good health yet succumb to lethal effects days later. Thus, it may be postulated that the greater the lethal potency of a sulfide analogue the less likely it will be to demonstrate a stage of coma prior to death.
If coma is primarily a consequence of damage to the RAS or the reticulo-cerebral function, by a toxic mechanism or consequence of metabolic derangement such as anoxia or hypoglycaemia, then the site of action for coma is unlikely to be a single receptor. Me2S-induced coma is not unlike that of a general anaesthetic with weak potency, as shown by its ED50 of 813 mg/kg, a dose range for coma of about 500 mg and duration of about 20 minutes. These characteristics suggest a non-specific mechanism and argue against a specific toxin-receptor interaction as a mechanism of action for coma. Brain damage from H2S at high concentrations cannot be attributed to defined lesions in the brain (Tvedt et al. 1991). Perhaps a broad distribution of the sulphide moiety as toxin occurs in the brain which causes some non-specific membrane perturbation. H2S is known to be a neuromodulator (Abe and Kimura 1996; Dello Russo et al. 2000; Trevisani et al. 2005) and is also known to induce changes in tissue levels of some amino acids in the brain. (Reiffenstein 1989).
The presumed efficacy of nitrite as an antidote relies primarily on a study by Smith et al. (1976), who showed that rats treated with sodium sulfide 50 mg/kg, had a mortality rate of 78% which was reduced to 20% in animals post-treated with nitrite. Although an improvement in survival was claimed, the authors still noted that “antidoted animals exhibited severe signs of sulfide poisoning”. Our data show that nitrite reduced neither the incidence nor the duration of coma in rats treated with Me2S. It also failed to prevent death. A similar lack of effect would be expected for H2S and Na2S. It is likely that most if not all of the animals in the study by Smith et al. (1976), would have died if a later endpoint had been used.
Other recent studies have also challenged the use of nitrite as an antidote. The mechanism of action of nitrite is claimed to be through the formation of methemoglobin, with subsequent sequestration of the sulfide, rendering it unavailable to inhibit cytochrome oxidase. In an in vitro study it was shown that a methemoglobin-sulfide complex was indeed formed but it was short-lived. In the presence of air or oxygen, the formation of methemoglobin by nitrite was slow and only 5% of the available haemoglobin was converted after 20 minutes (Beck et al. 1981). The authors concluded that nitrite was not an appropriate antidote to sulfide.
This study does not support the use of pyruvate as an antidote. Pyruvate was also proposed as an antidote to hydrogen sulfide toxicity, on the assumption of similarity of the toxicities of cyanide and sulfide (Cittadini et al. 1972). Pyruvate binds directly to cyanide by a nucleophilic addition and was reported to have caused a two-fold increase in the LD50 (Dulaney and Hume 1988), presumably as a result of diminished inhibition in the brainstem. However, using the isolated brainstem and spinal cord from neonatal rats, Greer and Carter (1995), showed that a dose of cyanide that was lethal to rats in vivo failed to abolish the respiratory rhythm in that preparation. Pettersen and Cohen (1993) found a similar result because cytochrome oxidase inhibition did not correlate with diminished respiratory function in fish exposed to H2S.
The notion that dithiothreitol could serve as an antidote to hydrogen sulfide toxicity derives from the assumption that per-sulfide moieties protect sulfhydryl groups at the active site of enzyme or receptor proteins from inhibitory anions like HS–(Warenycia et al. 1990). In our study, dithiothreitol did not reduce lethality or duration of coma from Me2S, irrespective of the protocol used.
Bicarbonate afforded full protection when administered either as a pre-treatment or by simultaneous iv infusion to anaesthetised rats against a lethal dose of NaHS (Almeida and Guidotti 1999), but a mechanism of action was not identified. In conscious but un-anesthetized animals, bicarbonate had a modest effect on coma induced by Me2S and no effect on lethality, which was unexpected given the 40-fold lower toxicity of Me2S. In combination with glucose, however, it had a pronounced effect. This study does not provide an explanation for the phenomenon, but it does suggest a role for glucose in the mechanisms for coma and lethality caused by H2S. The human brain, in its basal state has an average requirement for glucose of about 6 mg per minute per 100 g of brain tissue (Reivich et al. 1979). About 85% of this glucose is combusted to produce energy, with H2O and CO2formed as waste products. Thus, carbonic anhydrase and enzymes of glycolysis and glucogenesis may be involved. A sudden and profound but reversible drop in glucose availability to the central nervous system, might explain the observed effects and the apparently complete recovery in both the animal model and in humans who experience “knockdown”.
Bicarbonate and glucose would be very benign, low-risk treatments but require venous access, which is difficult under field conditions.
CONCLUSION
NaHS, an inorganic analog for H2S is a safer substitute for the gas and is routinely used for laboratory study. In this study NaHS and the organic analog Me2S, have been used to study the efficacy of purported antidotes that are currently used for H2S toxicity; these agents do not appear to work in laboratory models. In contrast, bicarbonate and glucose is protective against coma and lethality, and is likely to be protective against apnea (Almeida and Guidotti 1999). Life-threatening toxicity due to H2S often occurs in remote locations (such as isolated natural gas wells), on large agricultural enterprises, and in confined spaces (such as ships’ holds) where discovery and rescue may be delayed and medical treatment may not be readily available. A mixture of bicarbonate and glucose could be applied as antidote in such cases.
Footnotes
Figure and Tables
Acknowledgements
The authors thank the Tripartite Fund for Occupational Health, the University of Alberta Faculty of Pharmacy, Health Sciences Laboratory Animal Services, and the Department of Public Health Sciences of the University of Alberta Faculty of Medicine for support of this work.