Hemorrhagic shock (HS) is a life-threatening complication in both trauma patients and in the operating room 1 2 . The pathophysiology of HS is complex, especially during the reperfusion phase 3 . During HS, the state of vasoconstriction turns into vasodilatory shock. According to Landry et al. 4 , this phenomenon is related to tissue hypoxia as well as to a proinflammatory immune response 4 . In addition, during the reperfusion phase, cellular injuries induced by ischemia are enhanced, and are associated with excessive production of radical oxygen species (ROS), leading to a further systemic inflammatory response 5 .

Hydrogen sulfide (H₂S), is known as an environmental toxic gas 6 , but has also recently been recognized as a gasotransmitter 7 , similar to nitric oxide (NO) and carbon monoxide (CO). H₂S is endogenously synthesized 8 and may play a crucial role in critical care according to the recent review of Wagner et al. in 2009 9 . Depending on the selected models, H₂S has been reported to exhibit pro- and anti-inflammatory properties and to display opposite effects in various shock conditions 10 11 12 13 . H₂S has also been reported to induce direct inhibition of endothelial nitric oxide synthase (eNOS) 14 . However, this effect was linked to the concentration of H₂S, whereby H₂S caused contraction at low doses and relaxation at high doses in both rat and mouse aorta precontracted by phenylephrine 14 . This dual effect was related, at low dosage, to the inhibition of the conversion of citrulline into arginine by eNOS (contraction) and at high dosage by activation of K⁺ _ATP channels or due to NO quenching 15 . Blackstone et al. 10 11 recently suggested that inhalation of H₂S induced a "suspended animation-like" state which protected animals from lethal hypoxia. Furthermore, Morrison et al. 16 demonstrated that pre-treatment with inhaled or intravenous (i.v.) H₂S prevented death and lethal hypoxia in rats subjected to controlled but unresuscitated hemorrhage.

Conversely, Mok et al. 17 reported the hemodynamic effects of the inhibition of H₂S synthesis, along with a rapid restoration in mean arterial pressure (MAP) and heart rate (HR), in a model of unresuscitated hemorrhage in rats.

As the vascular effects of H₂S are still a matter of debate, and since all of these data originated from unresuscitated hemorrhage, we therefore tested the hypothesis that the H₂S donor sodium hydrosulfide (NaHS), infused before retransfusion in a model of a controlled hemorrhagic rat, may improve hemodynamics and attenuate oxidative and nitrosative stresses, as well as the inflammatory response during reperfusion. Since the role of the cardiovascular system during shock becomes critical, we therefore focused on the inflammatory response as well as on the oxidative and nitrosative stresses in the heart and aorta.

The animal protocol was approved by the regional animal ethics committee (CREEA-Nantes, France). The experiments were performed in compliance with the European legislation on the use of laboratory animals.

In the present study, we report the beneficial effects of NaHS as an H₂S donor, prior to retransfusion, in a rodent model of controlled hemorrhage. The key findings were that a single i.v. NaHS bolus immediately before retransfusion of shed blood (i) limited the I/R induced-decrease in MAP and (ii) was associated with reduced inflammatory and oxidative stress responses.

Although H₂S is usually considered as an endogenous vasodilatator, this effect nevertheless remains a matter of debate. At low concentrations (10 to 100 μM H₂S), Ali et al. 15 found a vasoconstrictor effect of H₂S on rodent aorta, whereas Dombkovski 20 reported that H₂S was responsible for either vasodilatation or vasoconstriction, according to species and organ requirements. Furthermore, data reported in the literature are highly conflicting: indeed, Mok et al. 17 reported an increase in MAP in unresuscitated HS treated with H₂S synthesis blockers (DL-propargylglycine and μ-cyanoalanine) whereas Morrison et al. 16 , using an opposite experimental approach, reported beneficial effects of H₂S on survival in rats submitted to lethal unresuscitated HS. In the present study, compared to the HS-saline group, a single i.v. bolus of NaHS produced a substantial increase in MAP in hemorrhagic rats. All rats were well oxygenated (PaO₂ >100 mm Hg, data not shown), an observation that was not reported in the studies by Mok et al. 17 and Morrison et al. 16 .

The absence of a detrimental effect on stroke volume has already been reported by others 11 21 22 . Herein, heart rate was not altered in either group while carotid blood flow was higher in the HS-NaHS group. Since blood flow was decreased in HS-saline, this would suggest a higher stroke volume in HS-NaHS treated rats, although this conclusion could be challenged since cardiac output was not directly measured in this study. Nevertheless, this result is in agreement with improved ejection fraction in a model of myocardial I/R injury 23 .

In the present study, NaHS treatment limited the metabolic acidosis induced by I/R. Simon et al. 21 also reported similar metabolic effects in pigs. Whether this effect is due to reduced metabolic demand induced by the sulfide donor or to a direct effect on mitochondrial K⁺ _ATP channels remains speculative since metabolic rate was not measured.

It is well documented that cardiovascular dysfunction during I/R is partly linked to the activation of the NF-κB/Rel pathway. This mechanism has been demonstrated in recent investigations 24 , allowing the expression of iNOS and subsequent overproduction of NO in cardiovascular tissues 25 . As reported by others 26 , we show herein that NaHS induced an in vivo down-expression of iNOs, with subsequent decrease in NO overproduction.

The effects of H₂S on inflammation are also a matter of contention 25 27 28 . In the present model, we report a predominant inflammatory modulation effect. Indeed, NaHS was found to limit cardiovascular NF-κB activation as well as decrease I-CAM expression in aorta. These results confirm in vitro experiments which demonstrated that NaHS as well as other H₂S endogenous donors modulate leukocyte-mediated inflammation 25 29 by decreasing leukocyte adhesion and leukocyte infiltration 23 through activation of K⁺ _ATP channels 25 .

In the present study, infusion of a NaHS bolus attenuated oxidative stress induced by I/R, as mirrored by a decreased release of O₂ ^- in tissues. H₂S is known to react with the four different reactive oxygen species 30 31 32 . Since increased ROS formation is implicated in lipid peroxidation and oxidation of thiol groups, H₂S, by decreasing ROS overproduction, may in fact limit tissue damage. Our results show that O₂ ^- production was decreased in both aorta and heart, suggesting a protective effect on cardiovascular tissues. These results are in agreement with the observations of Sivarajah et al. 33 , who recently reported that the cardioprotective effects of NaHS in a model of I/R on isolated cardiomyocytes were related to antioxidative and anti-nitrosative properties.

Nrf2 could contribute to adaptive and cytoprotective responses to various cell damages 31 34 . Different antioxidant cellular pathways are associated with Nrf2 expression such as the heme oxygenase enzymes, HO-1 and HO-2. Indeed, Maines et al. 30 reported increased levels of HO-1 in I/R injuries; moreover, HO-1 was found to improve resistance to oxidative stress 32 and modulate inflammatory response, particularly in hemorrhagic shock 35 . HO-2, meanwhile, is found in almost all tissues and is known as a potential O₂ sensor in addition to playing a role in the maintenance of vascular tone 32 . Conversely to aortic tissues, there were no changes in Nrf2, HO-1 or HO-2 in the heart samples. In the present experimental design, rats were anesthetized and warmed but not overheated for ethical reasons in accordance with our animal care regulatory agency. The metabolic rate was not measured. In the studies of Blackstone et al. 10 11 and Morisson et al. 16 , animals were awake. The difference between the two experimental protocols does not exclude a metabolic effect in our experiments. However, since body temperature remained constant throughout the study period, the putative effect of hypothermia did not significantly contribute to the observed results, which are related to reduced inflammatory and oxidative stress pathways. Consequently, the beneficial effect of NaHS is unlikely the result of a hibernation-like metabolic state of "suspended animation" as reported previously 10 11 16 22 . The present observation, however, confirms other studies in which H₂S donors NaHS and Na₂S protected against ischemia reperfusion injury 23 33 36 37 38 39 40 41 and burn injury 29 independently of core temperature.

The present in vivo experimental study of I/R following resuscitated hemorrhagic shock in rats demonstrates that a single i.v. bolus of NaHS limited the decrease in MAP during early reperfusion and down-regulated NF-κB, iNOS and I-CAM expressions. These anti-inflammatory effects were associated with decreased NO and O₂ ^- production. Such beneficial effects of H₂S donors warrant further experimental studies.

• The results of this in vivo experimental study demonstrate that a single i.v. bolus of hydrogen sulfide (considered as the third gaseous transmitter) donor, NaHS, prevented ischemia reperfusion (I/R)-induced hemodynamic dysfunction in a model of controlled hemorrhage in rats.

• NaHS reduced NO production and I/R-dependent iNOS expression and improved metabolic dysfunction.

• NaHS down-regulated NF-κB, iNOS and I-CAM expressions in this model.

• NaHS reduced I/R-induced oxidative stress.

CBF: carotid blood flow; CO: carbon monoxide; eNOS: endothelial nitric oxide synthase; EPR: electron paramagnetic resonance; FeDETC: N, N D-Ethyldithiocarbamate and Fe³+ citrate complex HO-1: heme-oxygenase-1; HO-2: heme-oxygenase-2; HR: heart rate; HS: hemorrhagic shock; H₂S: hydrogen sulfide; iNOS: inducible NOS; I/R: ischemia-reperfusion; i.v.: intravenous; MAP: mean arterial pressure; NaHS: sodium hydrosulfide; NO: nitric oxide; Nrf2: nuclear respiratory factor 2; O₂ : superoxide anion; PI-κB: phosphorylated I-κB; PMSF: phenylmethylsulfonyl fluoride; ROS: radical oxygen species; SD: standard deviation.

The authors declare that they have no competing interests.

FG participated in the surgical procedure, in in vitro measurements and in the design of the protocol, and drafted the manuscript. MB carried out the Western blotting. MdlB and LF carried out the surgical procedure and in vitro measurements. OD participated in the laboratory investigations. AM, PC and DH helped to design the study. PR helped to design the study and to draft the manuscript. LL participated in in vitro measurements. PA designed the study, and coordinated and drafted the manuscript. FM participated in the design of the study, performed the statistical analysis and helped to draft the manuscript.