H₂S Increases Survival during Sepsis: Protective Effect of CHOP Inhibition

Sepsis is a systemic inflammatory response to suspected or identified infection. Severe sepsis and septic shock remain the tenth most common cause of death in intensive care units, with >210,000 deaths annually in the United States (1, 2). This high mortality demonstrates the lack of effective therapies. Therefore, a deeper understanding of the pathophysiology of this complex syndrome is necessary to develop new and more effective treatments.

The response to infection depends on the initial host–microbial interaction. Activation of the innate immune system directs eradication of the infection while limiting the inflammatory response. At the molecular level, several signal-transduction pathways are induced, as well as transcription factors, such as NF-κB, which is a master regulator of several inflammatory mediators (3).

However, a new role is emerging for endoplasmic reticulum (ER) stress signaling and the unfolded protein response (UPR) in the regulation of inflammation and innate immunity (4, 5). Specific ER stress mediators were shown to increase the inflammatory response of dendritic cells and macrophages (6–8). Additionally, signaling from the UPR intersects with major inflammatory pathways, such as JNK and NF-κB (9). The ER exerts a central role in synthesis, maturation, and trafficking of proteins. Under stressful cellular conditions, accompanied by protein misfolding, alterations in ER Ca²⁺, or oxidative stress, the ER elicits a complex adaptive response by activation of the UPR to re-establish cellular homeostasis. Ultimately, if the ER stress is not alleviated and cellular function is compromised, apoptosis is initiated.

C/EBP homologous protein 10 (CHOP) is a transcription factor and mediator of the UPR that plays a major role in inducing apoptosis (4, 9, 10). However, evidence indicates a new role for CHOP in the inflammatory response (11–13). CHOP expression is triggered in lungs of mice injected with LPS and after hemorrhagic trauma (14, 15). This transcription factor is crucial for induction of caspase-11 (16) that is involved in maturation of pro–IL-1β through caspase-1 (17). Accordingly, genetic loss of Chop results in decreased LPS-induced secretion of IL-1β (14). Thus, it is plausible that CHOP plays a major role in the pathogenesis of sepsis, although direct evidence is lacking.

H₂S is a biologically active gas serving as a messenger molecule (18). H₂S is endogenously generated in many cell types, mainly from cysteine, by two enzymes: cystathionine-β-synthase and cystathionine γ-lyase (CSE). It has a pivotal role in modulating cardiovascular function, under physiologic and pathologic conditions. Mice lacking the Cse gene develop hypertension (19), whereas cardiac-specific overexpression of CSE protects against myocardial ischemia-reperfusion injury (20). The role exerted by H₂S in inflammation remains controversial because it was reported to have both proinflammatory and anti-inflammatory effects (18). In animal models of sepsis, both the pharmacological inhibition of H₂S synthesis, as well as its administration, were shown to improve survival of septic mice (21, 22).

The present investigation examined the therapeutic benefit of administration of H₂S during polymicrobial sepsis induced by cecal ligation and puncture (CLP) in mice and the mechanistic basis for this effect. Exogenous H₂S inhibited CHOP expression, and genetic deletion of Chop identified this transcription factor as a key molecule in the pathogenesis of sepsis.

Male wild-type (WT) (6–8 wk old) C57BL6/J mice and CHOP-KO mice on C57BL6/J background were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were anesthetized with isoflurane via a vaporizer at 1.5–2.5 minimum alveolar concentrations. Under sterile conditions, a 1.5-cm incision was made in the lower abdominal region, and the cecum was exposed. The distal portion of the cecum was completely ligated 1 cm from the end with a 3-0 silk suture, punctured once with an 18-gauge needle, and then replaced in the peritoneal cavity. Subsequently, the peritoneal wall and skin were closed with double sutures. A 1-ml s.c. injection of sterile saline (0.9%) was administered to mice after surgery. Sham mice underwent abdominal incision and cecal exposure without ligation and puncture. After the procedure, mice had access to water and food ad libitum. Survival was monitored for 8 d, while for cytokines, CFU, and organ harvesting, animals were sacrificed after 5 or 18 h.

Mice were injected i.p. with 25 mg/kg LPS (from Escherichia coli 055:B5), and survival was monitored for 8 d. All procedures were carried out in accordance with the Johns Hopkins University Animal Care and Use Committee–approved protocols that conformed to the Guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health.

H₂S was generated by dissolving NaHS into saline just before use (21). One hundred microliters of the solution was injected s.c. in mice 1 h before CLP (100 μmol/kg). In the H₂S posttreatment survival experiment, mice received H₂S s.c. 2 h after the induction of sepsis, and H₂S was administered again after 24 h for the following 2 d.

Peritoneal lavage was performed with 5 ml sterile PBS. Serial dilutions of peritoneal fluid and blood were plated on brain and heart infusion agar plates (Becton Dickinson, Franklin Lakes, NJ) at 37°C. Bacterial colonies were counted after 24 h.

Animals were sacrificed at the time specified in the figures, and blood was drawn by cardiac puncture. Plasma was diluted and assessed for cytokine levels by ELISA (eBioscience, San Diego, CA).

Leukocyte accumulation in the lung was measured as previously described, with some modification (23). Briefly, lung tissue was homogenized in a solution containing 0.5% hexa-decyl-trimethyl-ammonium bromide and centrifuged for 30 min at 20,000 × g at 4°C. An aliquot of the supernatant was incubated with a solution of tetra-methyl-benzidine (1.6 mM) and 0.1 mM H₂O₂. The rate of change in absorbance was measured at 405 nm. Myeloperoxidase (MPO) activity was defined as the quantity of enzyme degrading 1 μmol H₂O₂/min at 25°C (expressed as U/g weight of tissue).

Aspartate aminotransferase levels were measured as an indicator of liver damage from mouse serum using the Vet Ace Clinical Chemistry analyzer system.

Cells were isolated from mice as previously described (24) and plated on cover slips; nonadherent cells were washed off after 2 h. The following day, cells were stimulated with LPS (10 μg/ml), alone or in combination with H₂S (200 μM) for 18 h. RAW 264.7 mouse macrophages were cultured in DMEM medium with penicillin, streptomycin, and 10% FBS. All cells were maintained at 37°C in the presence of 5% CO₂ in a humidified incubator.

For measurement of mRNA levels, total RNA was extracted from cells using TRIzol Reagent (Life Technologies). RNA was reverse transcribed into cDNA using oligonucleotide (dT)₂₀ primers and the SuperScript III First-Strand Synthesis System (Life Technologies Corporation). The resulting cDNA was amplified using SYBR Green Real-Time PCR and detected on an ABI 7500 system (Life Technologies) with the following amplification conditions: 95°C for 10 min and then 40 cycles of 95°C for 15 s and 60°C for 60 s. Gene expression was calculated relative to the housekeeping gene β-actin using the ΔΔCt algorithm (primers used: Chop forward: 5′-GAGCCAGAATAACAGCCGGA-3′, Chop reverse: 5′-CTGTCAGCCAAGCTAGGGAC-3′).

Harvested organs were flash frozen in liquid nitrogen and stored at −80°C. Tissue was homogenized in RIPA buffer with protease and phosphatase inhibitors, and protein content was determined. Immunoblotting was performed as previously described (25). Briefly, proteins were separated on a gradient SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and blocked with 5% milk buffer. Membranes were incubated with primary Ab against CHOP, Nrf2, ATF4, or AFT6 (Santa Cruz Biotechnology, Dallas, TX), activated caspase 3 (Cell Signaling Technology, Danvers, MA), or tubulin (Abcam, Cambridge, MA) overnight at 4°C. After washing, membranes were incubated with specific secondary HRP-conjugated Abs for 1 h and incubated with HRP substrate, and signal was detected with a Kodak Image Station 4000B.

Fixed specimens were embedded in paraffin and stained with H&E for light microscopic evaluation. Activation of caspase-3 was detected by immunoperoxidase staining, using an Ab that only recognizes the activated form. CHOP expression was identified using a rabbit anti-mouse Ab against CHOP (Santa Cruz Biotechnology). Sections were counterstained with hematoxylin. Slides were viewed with an inverted microscope (Axio Observer.A1; Zeiss), and images were acquired with a Olympus DP72 color camera using CellSense software.

Peritoneal macrophages were fixed in cold methanol, and immunostaining was carried out as previously described (26). A rabbit anti-mouse Ab (against CHOP) was used as primary Ab at 4°C overnight. After washing, fluorescent-labeled secondary Ab was added for 1 h at room temperature. Cells were washed again and mounted, and fluorescence was observed with a confocal microscope (Zeiss LSM 510 Meta).

Data are expressed as mean ± SD, and statistical significance was determined by an unpaired t test with Welch correction. A p value < 0.05 was considered significant. Survival was analyzed using the Kaplan–Meier method. Statistical analyses were performed using the log-rank test.

Previous in vivo studies assessing the impact of exogenous H₂S during sepsis reported conflicting results, revealing both a protective and a deleterious effect of H₂S (21, 22). We induced sepsis by performing CLP, a widely used animal model of polymicrobial sepsis that has the advantage of closely resembling sepsis in humans (27). One group of mice was treated s.c. with NaHS (100 μmol/kg) to generate H₂S. One hour later, severe sepsis was induced by single puncture with an 18-gauge needle in the H₂S-treated and untreated groups. We carried out a mortality study, and monitored survival for 8 d. Remarkably, we found that >80% (17 of 21) of the septic mice treated with H₂S survived compared with <25% (5 of 22) of untreated septic mice (p < 0.0001) (Fig. 1A). Thus, adopting the same dosage and delivery of H₂S used in the study by Spiller et al. (21), we confirmed a significant increase in the survival of septic mice treated with H₂S.

To assess the potential therapeutic use of H₂S, we evaluated the efficacy of H₂S treatment after induction of sepsis. For this purpose, sepsis was induced by CLP, and H₂S was administered s.c. 2 h later. We found that H₂S was equally effective in reducing mortality when given after the induction of sepsis (Fig. 1B).

Given the relevance of cytokines in regulating the inflammatory response (28, 29), we next evaluated the levels of the proinflammatory TNF-α and the anti-inflammatory IL-10 in the plasma of septic mice by ELISA. Blood levels of TNF-α were increased 5 h after CLP, as expected. H₂S treatment attenuated the increase by ∼50% (Fig. 2A). TNF-α levels continued to rise 18 h after CLP to 10 times the 5-h level, and H₂S blunted this increase. Also, plasma levels of IL-10 were elevated 5 h after CLP, with a further modest increase at 18 h. H₂S attenuated the increase in IL-10 levels, as well. Then, we determined the concentration of these mediators in the peritoneum, the site of infection. A peritoneal lavage was carried out, and TNF-α and IL-10 levels were determined at 5 and 18 h post-CLP. Similar patterns of change in the levels of these cytokines were found in the peritoneal cavity (Fig. 2B) as in the plasma. Thus, H₂S treatment significantly lowered both TNF-α and IL-10 levels in the blood and at the infected site.

Then, we assessed MPO activity as a measure of neutrophil infiltration in the lung 18 h after CLP induction (Fig. 2C, left panel). Septic mice showed a significant increase in MPO activity in the lung compared with sham animals. Treatment with H₂S lowered this activity to basal levels. In addition, we assessed liver function by determining aspartate aminotransferase (AST) levels (Fig. 2C, right panel). We found that CLP increased the level of AST compared with sham; however, H₂S treatment did not prevent this increase.

Next, we evaluated morphologic changes in the lung, liver, and spleen using H&E staining at 18 h after CLP. The histological sections of lung from septic mice showed signs of injury, including thickening of the alveolar septa and interstitial edema (Supplemental Fig. 1, upper panels). In the liver sections, swelling of hepatocytes and infiltrating neutrophils were observed (Supplemental Fig. 1, middle panels). Spleen from septic mice displayed numerous macrophages containing phagocytized karyorrhectic debris, likely from apoptotic cells in the white pulp (Supplemental Fig. 1, bottom panels). To verify cell death by apoptosis, we carried out immunostaining of spleen sections using an Ab to detect activated caspase-3 (Fig. 3A). We also examined caspase-3 activation in tissue extracts by immunoblotting (Fig. 3B). Caspase-3 activation was induced in the white pulp of spleen from septic mice. H₂S decreased its cleavage and prevented most of the above-described histological changes in organs due to CLP. Caspase-3 activation was also examined in lung and liver of septic mice by immunostaining of tissue sections and by Western blotting analysis. However, no sign of caspase-3 activation in lung or liver of septic mice was detected ≥18 h after CLP (data not shown).

Bacterial clearance is an important component in the resolution of infection. Therefore, we next evaluated the number of live bacteria in the blood and in the peritoneal cavity and how H₂S treatment would modulate bacterial load (Fig. 4). H₂S did not significantly alter the number of live bacteria 5 h after CLP, either in blood or peritoneum. At 18 h, the CLP group displayed an increase in live bacteria, whereas H₂S treatment resulted in an almost complete elimination of the bacterial burden in both blood and peritoneal cavity. Thus, H₂S treatment was associated with a significantly enhanced rate of bacterial clearance.

Recent findings highlighted the links between the inflammatory response and the ER stress-signaling pathways (4, 30). CHOP is emerging as a key molecule involved in the inflammatory response (11–13). Additionally, it was reported that H₂S can reduce ER stress in cardiac tissue (31) and in a neuroblastoma cell line (32). However, no study has tested the impact of H₂S on CHOP modulation during inflammation or sepsis. To elucidate the mechanism of H₂S protection, we first examined whether CHOP is expressed in vivo after CLP and whether H₂S modulates its expression. CLP-induced sepsis led to a significantly increased expression of CHOP in spleen, lung, and liver of septic mice 18 h after CLP compared with sham-operated mice (Fig. 5A). H₂S treatment in vivo was associated with decreased expression of CHOP in all of these organs.

Then, to investigate which cell type in the organs expressed CHOP, we performed immunostaining of tissue sections (Fig. 5B). The lungs of septic mice display CHOP expression in a disseminated pattern, probably both in inflammatory cells and lung parenchymal cells. Indeed, a previous study (14) showed that alveolar type I and type II cells express CHOP after LPS injection. In the liver, CHOP staining did not appear in hepatocytes or endothelial cells. Cells expressing CHOP are most likely tissue macrophages, such as Kupffer cells and infiltrating inflammatory cells. In the spleen, CHOP staining was evident in the sinusoidal areas outside the white pulp where macrophages mostly reside. Higher magnification shows that the CHOP⁺ cells were large round cells with abundant cytoplasm, consistent with the morphology of macrophages. Thus, under septic conditions, CHOP was expressed mainly by inflammatory and reticulo-endothelial cells in these organs.

We also investigated whether other members of the ER stress response were affected by H₂S treatment in vivo. Activation of the canonical UPR is mediated by three specific ER membrane proteins, including PERK and ATF6. Activation of the PERK pathway is implicated in attenuation of global translation initiation while increasing translation of specific transcription factors, such as ATF4, which is a main inducer of CHOP transcription. ATF6 signaling has a major role in chaperone induction and may also induce CHOP expression (9). Induction of these pathways is complex and modulates numerous proteins intersecting with a variety of inflammatory and stress signaling. In the spleen, we found that H₂S decreased CLP-induced ATF6 activation, as well as ATF4 (Fig. 6). Thus, other members of the UPR were directly or indirectly modulated by H₂S in vivo.

We used CHOP-KO mice to determine whether CHOP is an important mediator of the inflammatory response. A previous study (14) showed that CHOP plays a role in the pathogenesis of lung inflammation after LPS injection; however, its role in the overall outcome of mice treated with LPS has not been established. Thus, we first sought to test whether CHOP-KO mice would be more resistant to LPS-induced mortality (Fig. 7). To address this, we injected WT mice and CHOP-KO mice with LPS i.p. (25 mg/kg) and observed them for 8 d. Survival of CHOP-KO mice (87.5%) was significantly higher than that of WT mice (12.5%, p = 0.001).

Next, we investigated whether CHOP-KO mice are more resistant to CLP-induced sepsis. We found that septic CHOP-KO mice showed significantly improved survival compared with septic WT mice (p = 0.01, Fig. 8). Furthermore, we tested whether there would be an additive effect produced by H₂S treatment in combination with Chop genetic deletion. Treatment of septic CHOP-KO mice with H₂S did not show significant additional benefit compared with septic CHOP-KO mice (p = 0.9). These results clearly demonstrate that CHOP is a major negative factor in LPS- and CLP-induced mortality and provide evidence that H₂S-increased survival is mediated through CHOP inhibition.

Next, we examined the general histologic profile of liver, spleen, and lung at 18 h post-CLP in WT and CHOP KO mice (Supplemental Fig. 2). Septic CHOP KO mice displayed a lesser degree of tissue damage compared with septic WT mice, corroborating that inhibiting CHOP expression increases survival. We confirmed apoptosis by immunostaining of sections of spleen using an Ab to detect activated caspase-3 (Fig. 9). Although septic WT mice displayed activation of caspase-3, CHOP-KO mice were almost completely negative. Then, we explored whether CHOP is involved in the release of TNF-α and IL-10 in septic mice. The levels of these cytokines were significantly reduced in septic CHOP-KO mice compared with septic WT mice, both in plasma and in the peritoneal cavity (Fig. 10A, 10B). Lastly, we determined bacterial burden in CHOP-KO mice. After CLP, bacterial clearance was markedly improved in septic CHOP-KO mice compared with WT animals, in the blood as well as the peritoneum (Fig. 10C).

Thus, genetic deletion of Chop increases survival after CLP by reducing the systemic inflammatory response, decreasing apoptosis, and improving bacterial clearance.

We used macrophages to explore the mechanism by which H₂S reduces CHOP expression, because these cells play a central role in the innate immune response. In addition, our data suggest that this cell type shows increased CHOP expression during sepsis (Fig. 5B).

First, we assessed CHOP expression and the effects of H₂S in the macrophage cell line RAW264.7 and in peritoneal mouse macrophages. RAW cells were stimulated with LPS and with constituents of Gram-positive bacterial peptidoglycan (PGN). Incubation of cells for 18 h with these inducers led to CHOP expression. H₂S treatment blunted this effect (Fig. 11A), along with a reduction in Chop mRNA transcript (Fig. 11B). Immunostaining of peritoneal macrophages demonstrated that LPS treatment resulted in increased CHOP expression, in agreement with a previous study (16). H₂S treatment reduced CHOP expression in primary macrophages (Fig. 11C). Next, we investigated whether H₂S inhibits the ER stress induced by tunicamycin (TN) treatment, a widely used inducer of the ER stress response. H₂S was able to reduce CHOP expression by TN (Fig. 11D). Thus, H₂S decreased CHOP expression in macrophages stimulated with bacterial products and with a direct ER stress inducer.

Subsequently, we investigated the mechanism by which H₂S decreases CHOP expression. A recent study (33) showed that the protective effect of H₂S in the heart is mediated by increased Nrf2 levels. Nrf2 is a transcription factor implicated in improving cell survival following ER stress, and it was shown to inhibit CHOP expression (34, 35). We first examined whether H₂S induces Nrf2 in macrophages. We incubated RAW cells with H₂S for up to 24 h and assessed Nrf2 expression. Nrf2 was induced after 1 h of H₂S treatment, increased with time from 1 to 8 h, and started to decrease at 24 h (Fig. 11E).

To establish whether H₂S decreases CHOP expression through Nrf2 activation, we used silencing RNA directed against Nrf2 (siNrf2). RAW cells were transfected with a scrambled silent control (siCon) oligonucleotide or siNrf2 for 24 h and then incubated with medium (Bas) or H₂S for 18 h. Nrf2 was effectively downregulated by the siNrf2 compared with siCon (Fig. 11F). Then, we stimulated the cells with LPS in the presence or absence of H₂S. In the siCon group, LPS-induced CHOP expression was decreased by H₂S treatment. In the siNrf2 group, the basal level of CHOP was increased, and treatment with H₂S did not decrease LPS-induced CHOP expression (Fig. 11G). Our data indicate that Nrf2 modulates CHOP expression under basal conditions, and it is a major regulator of the H₂S-mediated decrease in CHOP expression.

In this study, we report that administration of H₂S markedly improves survival of experimental polymicrobial sepsis when given prior to the onset and, more importantly, even 2 h after CLP. The improved outcome is associated with decreased cytokine production, reduced apoptosis, and increased bacterial clearance. CLP induced CHOP expression, but this effect was attenuated by H₂S treatment. We found that CHOP is a major regulator of the inflammatory response, because CHOP-KO mice are significantly resistant to LPS- and CLP-induced mortality. We also demonstrated that the protection afforded by genetic deletion of Chop is not affected by H₂S treatment, suggesting that CHOP may be a key mediator of H₂S-induced protection. Finally, we demonstrated that H₂S inhibition of CHOP expression is mediated by Nrf2 activation in macrophages.

Two previous investigations on the effect of exogenous H₂S on experimental sepsis led to conflicting results. Zhang et al. (22) showed that H₂S injection into septic mice resulted in increased mortality, whereas Spiller et al. (21) reported a beneficial effect. The divergent outcome of these studies may be explained by the different route of H₂S administration and the dosage used. Although a beneficial effect was observed when H₂S was administered s.c. at a dose of 100 μmol/kg (∼5.6 mg/kg), an i.v. bolus (10 mg/kg) resulted in increased mortality. As the study from Spiller et al. (21) showed, an i.v. bolus of H₂S provokes a significant decrease in mean arterial pressure, whereas s.c. administration avoids this effect. The latter is likely due to its slow release into the circulation. Therefore, during sepsis, an already compromised vascular tone could be dysregulated even further by a direct i.v. bolus of H₂S, which induces vasorelaxation in different vascular beds (18). Thus, although other factors may also have a role, our study confirms the primary importance of the route of administration in dictating whether H₂S is beneficial or deleterious in the setting of septic shock.

Our study shows that H₂S treatment resulted in decreased production of the proinflammatory TNF-α, as well as decreased leukocyte infiltration in the lungs and liver, in line with other studies (21, 36–38). Because IL-10 is known to have anti-inflammatory properties (39), we were surprised to find that H₂S treatment decreased the release of this cytokine. Previously, H₂S was reported to increase IL-10 levels in vitro and in a model of acute lung injury (37, 40), although H₂S decreased IL-10 levels during hepatic ischemic-reperfusion injury (41). It is plausible that the impact of H₂S on IL-10 levels varies, depending on the cell type or animal model used. However, and more importantly, it was shown that neutrophils from septic patients have reduced bactericidal activity when treated with IL-10, leading to impaired host defense with overgrowth or persistence of pathogens (39, 42–44). Moreover, neutralization of IL-10 in septic mice improves survival (45) and enhances phagocytosis by macrophages (46). Thus, the attenuated release of this cytokine could be important, because the overproduction of this cytokine, together with other factors, could be responsible for the immune-suppressive state that is characteristic of sepsis (47). Therefore, H₂S treatment may help to maintain an effective, but not overactive, host immune response leading to increased survival.

Another important aspect of our study is that treatment with H₂S was associated with a virtually complete inhibition of apoptosis induced by CLP. H₂S is known to decrease apoptosis in different cell types, such as neurons (48), cardiomyocytes (20), and cancer cells (49). However, this effect in a model of experimental sepsis has not been reported. Recent studies highlighted that apoptosis during sepsis plays an important role. Indeed, septic spleens from patient autopsies and from animal studies show widespread lymphocyte apoptosis during sepsis (50–55). Conversely, inhibition of lymphocyte apoptosis in experimental models leads to increased survival (53, 54, 56–58). Among the mechanisms of the antiapoptotic effect of H₂S, the preservation of mitochondrial function through the opening of the K_ATP channels was shown to play an important role (18). Nevertheless, during sepsis, both the death receptor pathway and mitochondrial-mediated apoptosis have been described, as well as cross-talk between these two pathways (51, 54, 55, 57, 59). Taken together, our results suggest that the beneficial effect of H₂S is likely due to modulation of different pathways that are often interconnected, targeting apoptosis, as well as inflammatory mediators, and resulting in increased survival.

Septic mice exhibited increased expression of CHOP, and treatment with H₂S was associated with its downregulation. Although CHOP is a major inducer of apoptosis in response to ER stress (4, 9, 10), recent evidence identifies a new role for CHOP as a mediator of the inflammatory response (14–17). Our data demonstrate that this transcription factor plays a pivotal role in the pathogenesis of sepsis. Indeed, CHOP-KO mice displayed virtually no apoptosis, augmented bacterial clearance, and increased survival after CLP or LPS injection. All of the beneficial features observed in mice with genetic deletion of Chop were also seen in septic WT mice treated with H₂S. Furthermore, when we administered H₂S to septic CHOP-KO mice there was no additional improvement in survival compared with septic untreated CHOP-KO mice. This evidence supports the concept that suppression of CHOP may be a major factor in the beneficial effect afforded by H₂S. It is important to note that we observed no colocalization of caspase-3 activation with CHOP expression in the spleen of septic mice. Although caspase-3 activation was mainly in the white pulp of the spleen, CHOP expression was present in the outer marginal zones of this organ. Additionally, we also identified augmented CHOP expression in lung and liver; however, no detectable activation of caspase-3 was observed. Altogether, these data suggest that CHOP may not play a direct role in the apoptotic events that we observed in the spleen of septic mice, indicating that its primary role may be related to the inflammatory response per se. As shown by Endo and colleagues (16), CHOP has a critical role in the induction of caspase-11, which is needed for the production of IL-1β. In this regard, our data add new insights supporting the concept that CHOP contributes to the inflammatory response, in agreement with other studies (11–14, 16, 17). Importantly, to our knowledge, we show for the first time that genetic deletion of Chop leads to increased survival after CLP or LPS injection.

It is unknown whether H₂S directly affects CHOP expression. One possible target of H₂S was demonstrated in a study by Lefer and colleagues (33), who found that, in myocardial ischemic injury, H₂S treatment provides protection in the heart through Nrf2 signaling, a master regulator of cytoprotective genes. Furthermore, mouse embryonic fibroblasts deficient in Nrf2 display augmented CHOP expression after glucose deprivation (60). Thus, we hypothesized that H₂S could inhibit LPS-induced CHOP by increasing Nrf2 signaling. Our results show that H₂S increases Nrf2 levels in macrophages and that silencing of Nrf2 abolishes the inhibitory effect of H₂S on LPS-induced CHOP upregulation. A molecular mechanism by which H₂S modulates Nrf2 was recently described. Under basal conditions, Nrf2 is maintained in the cytoplasm by binding to its negative regulator Keap1. H₂S S-sulfhydrates Keap1 at specific cysteine residues, resulting in dissociation of Keap1 from Nrf2 and enhanced Nrf2 nuclear translocation and transcriptional activity (61).

How does CHOP regulate the inflammatory response? CHOP is an atypical transcription factor because it can inhibit or enhance transcription of a target gene, depending on the interaction with other transcription factors (62, 63). A very recent study (11) demonstrated that CHOP can modulate NF-κB activation by decreasing IκB degradation and p65 translocation. This mechanism leads to a decrease in NF-κB target genes involved in apoptosis and inflammation. Thus, modulation of CHOP by H₂S can interfere with different pathways involved in the immune response during sepsis.

In summary, we report that H₂S treatment increased survival in a model of experimental sepsis by reducing the inflammatory response and the immunosuppressive state induced by apoptosis of immune cells, leading to a more effective and balanced innate and acquired immune response. Our data also highlight a major role for CHOP, which may act as an amplifier of the inflammatory response in the pathogenesis of sepsis, and the ability of H₂S treatment to counter CHOP signaling. Finally, H₂S inhibits CHOP expression, at least in part, through a mechanism involving Nrf2 activation. Sepsis is a complex and dynamic syndrome. These results increase our knowledge of the pathogenetic mechanisms of sepsis and simultaneously shed light on new targets and suggest innovative strategies for the treatment of this syndrome.

We thank Dr. F. Spiller (Federal University of Florianopolis, Santa Catarina, Brazil) and Dr. F. Cunha (School of Medicine of Ribeirao Preto, University of Sao Paulo, Sao Paulo, Brazil) for their helpful advice on CLP procedure and H₂S treatment.

The authors have no financial conflicts of interest.