Hydrogen Sulfide Reduces Ischemia and Reperfusion Injury in Neuronal Cells in a Dose- and Time-Dependent Manner

Background: The ischemia-reperfusion injury (IRI) of neuronal tissue, such as the brain and retina, leads to possible cell death and loss of function. Current treatment options are limited, but preliminary observations suggest a protective effect of hydrogen sulfide (H2S). However, the dosage, timing, and mechanism of inhaled H2S treatment after IRI requires further exploration. Methods: We investigated possible neuroprotective effects of inhaled H2S by inducing retinal ischemia–reperfusion injury in rats for the duration of 1 h (120 mmHg), followed by the administration of hydrogen sulfide (H2S) for 1 h at different time points (0, 1.5, and 3 h after the initiation of reperfusion) and at different H2S concentrations (120, 80, and 40 ppm). We quantified the H2S effect by conducting retinal ganglion cell counts in fluorogold-labeled animals 7 days after IRI. The retinal tissue was harvested after 24 h for molecular analysis, including qPCR and Western blotting. Apoptotic and inflammatory mediators, transcription factors, and markers for oxidative stress were investigated. Histological analyses of the retina and the detection of inflammatory cytokines in serum assays were also performed. Results: The effects of inhaled H2S were most evident at a concentration of 80 ppm administered 1.5 h after IRI. H2S treatment increased the expression of anti-apoptotic Bcl-2, decreased pro-apoptotic Bax expression, reduced the release of the inflammatory cytokines IL-1β and TNF-α, attenuated NF-κB p65, and enhanced Akt phosphorylation. H2S also downregulated NOX4 and cystathionine β-synthase. Histological analyses illustrated a reduction in TNF-α in retinal ganglion cells and lower serum levels of TNF-α in H2S-treated animals after IRI. Conclusion: After neuronal IRI, H2S mediates neuroprotection in a time- and dose-dependent manner. The H2S treatment modulated transcription factor NF-κB activation and reduced retinal inflammation.


Introduction
Diseases associated with ischemia-reperfusion injury (IRI) of the central nervous system lead to neuronal cell death. Well-known examples of this disease group are traumatic brain injury and stroke, but diseases of the eye, such as retinal vascular occlusion, diabetic retinopathy, and glaucoma, are also associated with IRI. The retinal ganglion cells are very vulnerable to ischemia and have little potential for regeneration following injury. As in other neuronal cells, the IRI of retinal ganglion cells activates several apoptotic and inflammatory signaling pathways that ultimately trigger cell death [1].

H 2 S Inhalation Protects Retinal Ganglion Cells from IRI-Induced Apoptosis at Specific Concentrations and Times
IRI induced a loss of approximately 50% of the retinal ganglion cells (IRI: 1540 ± 78 vs. untreated: 2592 ± 54; Figure 1, Col. 1 and 2). Inhaled H 2 S was able to attenuate this effect at concentrations of 120 and 80 ppm at specific time points (Figure 1, Col 4,6,7). At a concentration of 120 ppm, only the inhalation administered at 1.5 h after IRI had a protective effect, as indicated by a significantly reduced cell death (IRI + 120 ppm H 2 S at 1.5 h: 2039 ± 99, p < 0.001). The immediate inhalation (time point of 0 h), or inhalation after a time delay of 3 h, had no protective benefit (IRI + 120 ppm H 2 S at 0 h: 1540 ± 78; IRI + 120 ppm H 2 S at 3 h: 1686 ± 105). The administration of 80 ppm H 2 S had a strong protective effect at two time points (0 and 1.5 h), with the best results occurring after inhalation with a 1.5 h delay (IRI + 80 ppm H 2 S at 0 h: 2039 ± 99 and IRI + 80 ppm H 2 S at 1.5 h: 2234 ± 181; both p < 0.001). However, no protection against cell death was observed after a time delay of 3 h following IRI. The administration of 40 ppm H 2 S also failed to prevent retinal ganglion cell death at any time point after IRI (IRI + 40 ppm H 2 S at 0 h: 1702 ± 140; IRI + 40 ppm H 2 S at 1.5 h: 1671 ± 92; IRI + 40 ppm H 2 S at 3 h: 1569 ± 103). Representative images of vital retinal ganglion cells and the induction of cell death by IRI, as well as the protective effects of hydrogen sulfide, are shown in Figure 2.

H2S Inhalation Decreased Inflammatory Cytokine Expression
H2S treatment decreased the pro-inflammatory cytokine response, as indicated by the decreased expression of the inflammatory cytokines IL-1β and TNF-α, as determined by qPCR ( Figure 5). The extent of the downregulation depended on the specific concentration and timing of the H2S treatment. Significant changes were noted at 80 ppm at 1.5 h and 0 h, and at 120 ppm at 1.5 h (IL-1ß-IRI: 2.12 ± 0.24 vs. IRI + 80 ppm H2S at 1.5 h: 0.64 ± 0.12,

H2S Inhalation Inhibits TNF-α Expression in the Ganglion Cell Layer
The cytokine TNF-α is a key regulator of the inflammatory response among others. TNF-α was detected mainly in the inner nuclear layer and the area of the ganglion cell layer in retinal cross-sections of IRI eyes ( Figure 6). H2S treatment (80 ppm, time interval 1.5 h) resulted in weaker staining, indicating decreased TNF-α expression.

H 2 S Inhalation Inhibits TNF-α Expression in the Ganglion Cell Layer
The cytokine TNF-α is a key regulator of the inflammatory response among others. TNF-α was detected mainly in the inner nuclear layer and the area of the ganglion cell layer in retinal cross-sections of IRI eyes ( Figure 6). H 2 S treatment (80 ppm, time interval 1.5 h) resulted in weaker staining, indicating decreased TNF-α expression.
(a) (b) Figure 6. TNF-α expression after unilateral IRI and H 2 S inhalation. Double staining was performed to clarify in which areas of the retina the expression of TNF-α occurs. Cross-sections of the retinae 7 days after unilateral IRI showed that IRI led to an increased expression of TNF-α mainly in the retinal ganglion cell layer, as well as in the inner nuclear layer. After the application of H 2 S, the expression was attenuated. Abbreviations: NFL = nerve fiber layer; GCL = ganglion cell layer; INL = inner nuclear layer; ONL = outer nuclear layer.

H 2 S Inhalation Reduces NF-κB Phosphorylation
We also examined the activation of the transcription factor NF-κB, which plays a critical role in inflammatory and apoptotic processes (Figure 9a). We used 80 ppm H 2 S, as this showed to have the best protective effect on retinal ganglion cells. Inhalation immediately and 3 h after retinal IRI only slightly reduced NF-κB phosphorylation compared to IRI without H 2 S treatment (IRI: 1.00 ± 0.09 vs. IRI + 80 ppm H 2 S at 0 h: 0.90 ± 0.10 and IRI + 80 ppm H 2 S at 3 h: 0.93 ± 0.11, n.s.). However, inhalation 1.5 h after IRI reduced NF-κB phosphorylation significantly (IRI: 1.00 ± 0.09 vs. IRI + 80 ppm H 2 S at 1.5 h: 0.77 ± 0.09, p < 0.05).

H 2 S Inhalation Induced Akt Phosphorylation
Western blot analysis showed that immediate treatment with 80 ppm H 2 S enhanced Akt phosphorylation (Figure 9b; IRI: 0.82 ± 0.24 vs. IRI + 80 ppm H 2 S at 0 h: 1.10 ± 0.11, p < 0.05), and this effect was even more pronounced when H 2 S was applied with a delay of 1.5 h (IRI + 80 ppm H 2 S at 1.5 h: 1.26 ± 0.23, p < 0.001). By contrast, extending the treatment delay time to 3 h caused only a mild and not statistically significant increase (IRI + 80 ppm H 2 S at 3 h: 0.95 ± 0.16, n.s.).

Discussion
This study investigated the neuroprotective effect of inhaled H2S. The following are the key findings: (1) The highest protective effect of H2S was observed with treatment at 80 ppm applied 1.5 h after IRI; (2) H2S treatment decreased the expression of the proinflammatory cytokines IL-1β and TNF-α; (3) H2S treatment reduced the expression of the pro-apoptotic Bax protein, increased the expression of the anti-apoptotic BCL-2, and attenuated NOX4 expression; (4) H2S treatment also diminished CBS expression; (5) H2S treatment reduced NF-κB phosphorylation and increased Akt phosphorylation.
The results of this study demonstrate that inhaled H2S has protective effects after the IRI of the retina, but the extent of neuroprotection depends on the dosage and the time of application. The application of 80 ppm H2S 1.5 h after IRI showed the best protection against retinal ganglion cell death. An immediate treatment with 80 ppm H2S significantly reduced cell death too, although to a lesser extent than the 80 ppm treatment at 1.5 h. A concentration of 120 ppm H2S provided a favorable effect only after a time delay of 1.5 h, whereas treatment at 40 ppm had no effect on ganglion cell survival after retinal IRI at any time point.
To date, studies on the neuroprotective effects of inhaled H2S have been performed predominantly at concentrations of 40 or 80 ppm. For example, Wei et al. administered

Discussion
This study investigated the neuroprotective effect of inhaled H 2 S. The following are the key findings: (1) The highest protective effect of H 2 S was observed with treatment at 80 ppm applied 1.5 h after IRI; (2) H 2 S treatment decreased the expression of the proinflammatory cytokines IL-1β and TNF-α; (3) H 2 S treatment reduced the expression of the pro-apoptotic Bax protein, increased the expression of the anti-apoptotic BCL-2, and attenuated NOX4 expression; (4) H 2 S treatment also diminished CBS expression; (5) H 2 S treatment reduced NF-κB phosphorylation and increased Akt phosphorylation.
The results of this study demonstrate that inhaled H 2 S has protective effects after the IRI of the retina, but the extent of neuroprotection depends on the dosage and the time of application. The application of 80 ppm H 2 S 1.5 h after IRI showed the best protection against retinal ganglion cell death. An immediate treatment with 80 ppm H 2 S significantly reduced cell death too, although to a lesser extent than the 80 ppm treatment at 1.5 h. A concentration of 120 ppm H 2 S provided a favorable effect only after a time delay of 1.5 h, whereas treatment at 40 ppm had no effect on ganglion cell survival after retinal IRI at any time point.
To date, studies on the neuroprotective effects of inhaled H 2 S have been performed predominantly at concentrations of 40 or 80 ppm. For example, Wei et al. administered inhaled H 2 S to rabbits at a concentration of 80 ppm for 1 h immediately after induced cardiac arrest and resuscitation and after the return of spontaneous circulation (ROSC). They found that H 2 S treatment reduced the adverse histopathological changes in the brain, improved neurological function, and improved survival in H 2 S-treated animals compared with a sham group [15]. Another study tested two H 2 S concentrations (40 or 80 ppm) in rats immediately after induced cardiac arrest and subsequent ROSC. Both concentrations showed a neuroprotective effect in terms of reduced neuronal cell loss, improved neurological function, and better overall survival [16]. A series of experiments in rats by Wei et al. showed that middle cerebral artery occlusion for 2 h followed immediately by a 3 h treatment with H 2 S at either concentration at the onset of reperfusion resulted in improved neurologic function and reduced infarct size at 24 h compared to the sham group. Both concentrations showed a similar protective effect. Similarly, Geng and colleagues studied the neuroprotective effect of 80 ppm H 2 S applied with a time delay of 1 h after the return of spontaneous circulation (ROSC) after induced cardiac arrest and cardiopulmonary resuscitation. They also showed sufficient protection and reduced permeability of the blood-brain barrier, decreased cerebral edema, and improved neurological outcome and survival [12]. However, the neuroprotective effect of H 2 S at a concentration of 120 ppm or at other time intervals following IRI has not been studied previously.
In our study, the neuroprotective effect of H 2 S was investigated for the first time in a dose-and time-dependent manner. The best effect was given with 80 ppm, either immediately after IRI, or after a time delay of 1.5 h from the onset of reperfusion. A concentration of 40 ppm seemed too low to protect against retinal ganglion cell death after IRI. Similarly, a time lag of 3 h seemed too long to protect against an early-onset injury mechanism at all concentrations. The reason why no neuroprotective effect was observed for 120 ppm H 2 S administered immediately after the termination of ischemia is unknown. One may speculate that a toxic component of H 2 S might counteract the protective effect, possibly at the mitochondrial level, but this was not investigated in detail in the present study. No further loss of vital retinal ganglion cells due to H 2 S was observed in our animals.
Our investigation of the underlying neuroprotective mechanism demonstrated that H 2 S treatment increased Akt phosphorylation and reduced the IRI-induced increase in NF-κB phosphorylation, with each response showing a dose-and time-dependent effect of H 2 S. Akt and the transcription factor NF-κB are significantly involved in anti-apoptotic and inflammatory processes in neuronal IRI [17]. Hypoxia resulting from IRI leads to the activation of Akt, which triggers effects such as pro-survival functions and the inhibition of apoptotic signaling pathways in mitochondria [18,19]. Intracellular signaling triggers an apoptotic stimulus that causes the oligomerization of the Bcl-2 antagonist and Bax, thereby inactivating Bcl-2. Activated Bax then alters the mitochondrial membrane potential, stimulates the release of cytochrome c, and activates caspases [20,21].
Another key regulator in neuronal ischemia is nuclear factor kappa B (NF-κB), which, in its phosphorylated active form, regulates the expression of pro-inflammatory and proapoptotic genes [22][23][24]. We observed a reduced expression of the pro-inflammatory cytokines IL-1ß and TNF-α, and found a decreased expression of the pro-apoptotic protein Bax and an increased expression of the anti-apoptotic Bcl-2. The respective changes depended on the H 2 S concentrations and time intervals used, with the strongest effect at 80 ppm and a delay of 1.5 h.
The underlying neuroprotective mechanism of inhaled H 2 S after acute neuronal damage has been previously analyzed in a few studies. Our results are in accordance with the findings reported by Wei et al. [16], who showed that inhaled H 2 S mediates cerebroprotective effects after induced cardiac arrest. They determined that H 2 S caused a reduction in NF-κB phosphorylation and decreased the release of the pro-inflammatory cytokines IL-6 and TNF-α. Similarly, Zhang et al. demonstrated the anti-apoptotic effects of exogenously supplied H 2 S in their study on traumatic brain injuries in mice. They found that H 2 S applied after injury reduced caspase-3 cleavage and increased Bcl-2 expression in injured brain tissue [25]. A neuroprotective effect of exogenous H 2 S was also demonstrated in a mouse model of induced Parkinson's disease by the promotion of neurogenesis via Akt [26].
In the present study, we were able to show that inhaled H 2 S inhibits the NADPH oxidase 4 (NOX4) enzyme. The complex process by which IRI leads to the death of neuronal ganglion cells involves oxidative stress and pathogenesis caused by the formation of oxygen radicals [27,28]. A protective, anti-oxidative effect of H 2 S has previously been demonstrated in a variety of organs, including neuronal tissues such as the retina [29][30][31][32]. The production of reactive oxygen species by neutrophil granulocytes occurs via NOX4 activity. For example, Wang and colleagues demonstrated that exogenously supplied H 2 S significantly reduced cerebral NOX4 expression in a mouse model of middle cerebral artery occlusion [33].
Endogenic hydrogen sulfide is produced in many tissues and seems to play an important role in physiological and pathophysiological processes [34,35]. In the retina, H 2 S is predominantly produced by CBS and cystathionine γ-lyase (CSE), which are localized in the ganglion cell layer [36]. The function of endogenous CBS in the retina is unclear at present. However, some evidence supports a role for endogenously produced H 2 S in ischemic brain injury and that elevated H 2 S levels contribute to poor outcomes [37]. We demonstrated that the inhalation of 120 ppm H 2 S strongly reduced the expression of CBS after IRI. The lack of a protective effect strengthens the hypothesis that a certain toxic component, such as the inhibition of the respiratory chain, counteracts the H 2 S effects-at least in a minor way.

Animals
We used adult male and female Sprague-Dawley rats (1:1male-to-female ratio, body weights of 280-350 g, Charles River, Sulzfeld, Germany) for these experiments. The animals were fed a standard diet ad libitum and kept on a light/dark cycle of 12 h of each. All interventions performed were in accordance with the ARRIVE guidelines for the use of animals in research and were approved a priori by the Committee for Animal Care of the University of Freiburg (approval number: 35-9185.81/G-16/46). All types of operations and manipulations were performed as previously described [38][39][40]. The number of animals for RGC quantification and molecular analysis was n = 8 per group. For the analysis of the mRNA and protein expression and immunohistology, the retinal tissue was collected 24 h after H 2 S inhalation.

Retrograde Labeling of RGC
Sprague-Dawley rats were anesthetized with isoflurane and placed in a stereotactic apparatus (Stoelting, Kiel, Germany). The skin over the skull was incised and retracted. For drilling three holes on each side of the Sagittal suture, the Lambda and Bregma sutures served as landmarks. Then, a total amount of 7.8 µL of fluorogold (FG) (Fluorochrome, Denver, CO, USA)-dissolved in DMSO/PBS-was injected through the drilling holes into both superior colliculi. To ensure the adequate retrograde transport of FG into the perikarya of the RGC, further experimental procedures were conducted seven days later.

Retinal Ischemia/Reperfusion Injury and H 2 S Treatment
After randomization, rats received intraperitoneal sedation. Next, the anterior chamber of the left eye was cannulated with a 30-gauge needle connected to a reservoir containing 0.9% NaCl. The intraocular pressure was raised to 120 mm Hg due to the hydrostatic pressure of the water column. The immediate onset of ocular ischemia was confirmed by the microscopical detection of interrupted retinal blood flow. After 60 min, reperfusion was initiated by the removal of the needle tip. Failure to immediately restore retinal perfusion at the end of ischemia resulted in the exclusion of the animal, as did lens injury since it prevents RGC death and promotes axonal regeneration [41]. To evaluate the neuroprotective effect of inhaled H 2 S, the animals were randomized to receive either room air or H 2 S at concentrations of 40, 80, or 120 ppm in 21% oxygen and the appropriate amount of nitrogen (Air Liquide, Kornwestheim, Germany) for 60 min. Therefore, animals were placed in an airtight chamber (the dimensions of the chamber were 20 × 35 × 27 cm; the flow rate was 8 L/min initially for 10 min, then reduced to 2 L/min). Treatment with hydrogen sulfide was initiated either immediately after reperfusion or with a delay of 1.5 or 3 h.

RGC Quantification
For RGC quantification, the animals were sacrificed by CO 2 inhalation 7 days after ischemia. Immediately removed retinal tissue was placed in ice-cold Hank's balanced salt solution and processed further for whole-mount preparation. The retinas were carefully placed on a nitrocellulose membrane with the ganglion cell layer (GCL) on top. Subsequently, remnants of the vitreous body were removed and the retinae were fixed in 4% paraformaldehyde for 1 h and embedded in mounting medium (Vectashield; Axxora, Loerrach, Germany). The densities of FG-positive RGC were determined in a blinded fashion using a fluorescence microscope (AxioImager; Carl Zeiss, Jena, Germany) and the appropriate bandpass emission filter (FG: excitation/emission, 331/418 nm), as described previously [38,39,42]. Briefly, we photographed 3 standard rectangular areas (0.200 mm × 0.200 mm = 0.04 mm 2 ) at 1, 2, and 3 mm from the optic disc in the central regions of each retinal quadrant. Thus, we evaluated an area of 0.48 mm 2 per retina. To calculate the average RGC density in cells/mm 2 , we multiplied the number of cells analyzed/0.04 mm 2 by 25. Secondary fluorogold-stained-activated microglial cells (AMC) after RGC phagocytosis were identified according to morphologic criteria and excluded from the calculation. All data are presented as the mean RGC densities (cells/mm 2 ) ± SD.

Real-Time Polymerase Chain Reaction
From retinal tissue harvested 24 h after ischemia, the total RNA from one-quarter of the retina was extracted using a column purification-based kit (RNeasy Micro Kit, Qiagen, Hilden, Germany) according to the manufacturer's instructions. Reverse transcription was performed with 50 ng of total RNA using random primers (High Capacity cDNA Reverse Transcription Kit, Thermo Fisher Scientific, Waltham, MA, USA). Real-time polymerase chain reactions (RT-PCR) were done with a TaqMan ® probe-based detection kit (TaqMan ® PCR Universal Master Mix, Thermo Fisher Scientific, Waltham, MA, USA). The following primers were used: Bax #Rn01480161_g1, BCL-2 #Rn99999125_m1, IL-1ß #Rn00580432_m1, TNF-α #Rn01525859_g1, NOX4 #Rn00585380_m1, CBS #Rn00650948_m1 (all from Thermo Fisher Scientific, Waltham, MA, USA). The PCR assays were then performed on an RT-PCR System (StepOnePlus ® , Thermo Fisher Scientific, Waltham, MA, USA) with the following cycling conditions: 95 • C for 10 min, 40 cycles of 95 • C for 10 s, and 60 • C for 1 min. Reaction specificity was confirmed by running the appropriate negative controls. The cycle threshold (CT) values for each gene of interest were normalized to the corresponding CT values for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (∆CT). The relative gene expression in IR injured retinal tissue, either with the inhalation of hydrogen sulfide or room air, was calculated in relation to the corresponding gene expression in the non-injured retinal tissue of each individual animal (∆∆CT).

Immunohistochemical Staining
For this analysis, eyes were enucleated 24 h after ischemia, immediately placed in 4% paraformaldehyde for 1 h at 4 • C. After washing in Dulbecco's phosphate-buffered saline (D-PBS) before and after post-fixation in 20% sucrose for 4 h at room temperature, the eyes were embedded in Tissue-Tek (Sakura-Finetek, Torrance, CA, USA) and frozen in liquid nitrogen. Frozen sections (10 µm) were cut through the middle third of the eye and collected on gelatinized slides. Immunohistochemistry was performed according to standardized protocols [38]. A polyclonal Rabbit anti-TNF-alpha antibody was used (#ab6671, dilution 1:500, Abcam, Cambridge, UK) and detected with a corresponding fluorophore-conjugated secondary antibody (Alexa Fluor 647, mouse anti-rabbit, dilution 1:1000, Jackson ImmunoResearch Europe, Newmarket, UK). Retinal cell nuclei were stained by adding 4 ,6-diamino-2-phenylindole dihydrochloride hydrate (DAPI; Sigma) to Mowiol mounting medium. The slides were examined under a fluorescence microscope (Axiophot, Carl Zeiss, Jena, Germany).

ELISA
Serum samples were analyzed using a commercial kit for TNF-α (#RTA00, R&D Systems, Minneapolis, MN, USA) according to the manufacturer s instructions. OD was analyzed at 450 nm on a microplate reader (Tecan ® , Männedorf, Switzerland).

Western Blot Analysis
After 24 h of ischemia, retinal tissue was collected for protein expression analysis. The total protein from three-quarters of the retina was extracted and processed for Western blot, as previously described [42]. The membranes were blocked with 5% skim milk in Tween20/PBS and incubated in the recommended dilution of protein-specific antibody (p-NF-κB#3033, p-Akt#9271, Cell Signaling Technology, Danvers, MA, USA) overnight at 4 • C. After incubation with a horseradish peroxidase-conjugated anti-rabbit secondary antibody (GE Healthcare, Freiburg, Germany), the proteins were visualized using enhanced chemiluminescence Western blotting detection reagent (Western Lightning plus enhanced chemiluminescence, #NEL104001EA, PerkinElmer, Waltham, MA, USA) following the manufacturer s instructions. For normalization, the blots were re-probed with total NF-κB (#8242, Cell Signaling) and total Akt (#9272, Cell Signaling). Relative changes in protein expression in IRI retinas after either hydrogen sulfide or room air inhalation were calculated relative to corresponding nonischemic retinas. Recordings and densitometric analysis were performed using Fusion Fx ® (Vilber, Collegién, France).

Statistical Analysis
A computerized statistics program (SigmaPlot Version 11.0, Systat Software Inc., San Jose, CA, USA) was used to evaluate the data. After checking the normal distribution of the data, the results were presented as mean values (SD). Between-group comparisons were performed with one-way ANOVA using the post hoc Holm-Sidak test and Kruskal-Wallis test for data with missing normal distribution. Statistical significance was assumed at p < 0.05.

Conclusions
We demonstrated that inhaled hydrogen sulfide protects retinal ganglion cells against ischemia-reperfusion injury. The effect depends on the time of application as well as on the administered dose. H 2 S mediates part of its effect via anti-oxidative, anti-inflammatory, and anti-apoptotic properties, while its molecular mechanism may be strengthened by an NF-κB and Akt dependent pathway.