• Open Access

Neuroprotective effects of hydrogen sulfide on Parkinson’s disease rat models


Jin-Song Bian, MD, PhD, Associate Professor, Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117456. Tel.: +65 6516 5502; fax: +65 6873 7690; e-mail: phcbjs@nus.edu.sg


Parkinson’s disease (PD) is a neurodegenerative disorder characterized by a progressive loss of dopaminergic neurons in the substantia nigra (SN). The present study was designed to examine the therapeutic effect of hydrogen sulfide (H2S, a novel biological gas) on PD. The endogenous H2S level was markedly reduced in the SN in a 6-hydroxydopamine (6-OHDA)-induced PD rat model. Systemic administration of NaHS (an H2S donor) dramatically reversed the progression of movement dysfunction, loss of tyrosine-hydroxylase positive neurons in the SN and the elevated malondialdehyde level in injured striatum in the 6-OHDA-induced PD model. H2S specifically inhibited 6-OHDA evoked NADPH oxidase activation and oxygen consumption. Similarly, administration of NaHS also prevented the development of PD induced by rotenone. NaHS treatment inhibited microglial activation in the SN and accumulation of pro-inflammatory factors (e.g. TNF-α and nitric oxide) in the striatum via NF-κB pathway. Moreover, significantly less neurotoxicity was found in neurons treated with the conditioned medium from microglia incubated with both NaHS and rotenone compared to that with rotenone only, suggesting that the therapeutic effect of NaHS was, at least partially, secondary to its suppression of microglial activation. In summary, we demonstrate for the first time that H2S may serve as a neuroprotectant to treat and prevent neurotoxin-induced neurodegeneration via multiple mechanisms including anti-oxidative stress, anti-inflammation and metabolic inhibition and therefore has potential therapeutic value for treatment of PD.


Hydrogen sulfide (H2S) is an active molecule and has been recognized as a new biological gas alongside nitric oxide and carbon monoxide. It is produced endogenously by two pyridoxal-dependent enzymes: namely, cystathionine β-synthase (CBS) and cystathionine γ-lyase (CSE). CBS has long been considered to be the predominant enzyme to produce H2S in the central nervous system. Recently, it was found that 3-mercaptopyruvate sulfurtransferase in combination with cysteine aminotransferase also produces H2S from cysteine in neurons (Shibuya et al., 2009). H2S plays important roles in regulating brain function. Physiologically, H2S enhances the induction of LTP by activating NMDA receptors in neurons (Abe & Kimura, 1996), elicits Ca2+ waves in astrocytes (Nagai et al., 2004) and increases Ca2+ levels in microglia (Lee et al., 2006). The protective effects of H2S on neurons against oxidative stress were first reported by Kimura & Kimura (2004). We subsequently found that the anti-oxidant effect of H2S may be mediated by the enhanced glutamate uptake function (Lu et al., 2008). In addition, our group also found that H2S produces anti-inflammatory effects in microglia (Hu et al., 2007) and anti-apoptotic effects in neurons (Hu et al., 2009). All these findings imply that H2S may be of therapeutic value in treating neurodegenerative diseases.

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by a progressive loss of dopaminergic neurons in the substantia nigra (SN) and depletion of the neurotransmitter dopamine in the striatum (Moore et al., 2005). It leads to impairment of movement that includes slow movement, resting tremor, rigidity and disturbance of gait (Parkinson, 2002). Although the mechanism for dopaminergic neuronal degeneration in PD is not completely understood, oxidative stress and neuroinflammation have been implicated as important contributors to the pathogenesis of PD.

Oxidative damage is considered as a primary pathogenic mechanism of nigral dopaminergic cell death in PD (Yoo et al., 2003). 6-Hydroxydopamine (6-OHDA) was the first dopaminergic neurotoxin discovered and has been used experimentally in models of PD for over 30 years (Uversky, 2004; Bove et al., 2005). The neuronal damage induced by 6-OHDA is mainly due to the massive oxidative stress caused by the toxin. Similar to dopamine, 6-OHDA has high affinity for the dopamine transporter, which carries the toxin into the dopaminergic neurons. Once in the neuron, 6-OHDA selectively kills dopaminergic neurons by generating ROS such as superoxide radicals (Tolwani et al., 1999). For this reason, unilateral injection of 6-OHDA is a good model for studying oxidative stress in the pathogenesis of PD.

Rotenone, an inhibitor of mitochondrial complex I, has been widely used as an herbicide and fish destroyer in reservoirs. Given that complex I deficiency persisting over the lifespan of a human being causally contributes to PD, rotenone treatment mimics the pathological process that develops in humans and thus represents a unique model with construct validity (Greenamyre et al., 2003; Sherer et al., 2003b,c). More importantly, the rotenone model recapitulates most of the mechanisms that thought to be important in PD pathogenesis, including loss of dopaminergic neuron in SN and neuroinflammation in nigrostriatal dopaminergic pathway (Sherer et al., 2003a,c). Thus, the rotenone-treated rat is another suitable model for investigating new possible therapeutic targets (e.g. neuroinflammation) for PD.

In this study, we, therefore, investigated the therapeutic and prophylactic effects of H2S on dopaminergic neurodegeneration by examining its anti-oxidant effect in the 6-OHDA-induced PD rat model and its anti-inflammatory effect in rotenone-induced PD model.


Role of endogenous H2S in 6-OHDA-lesioned rodent striatum

Unilateral 6-OHDA lesion leads to obvious neurotoxicity in the rodent model. As shown in Fig. 1A, the time-course of the apomorphine-induced rotations showed that the animal behavior was aggravated gradually from week 4 to 7 after unilateral injection of 6-OHDA into the left striatum. No rotation (0 turns per 30 min) was observed throughout all 7 weeks in the sham group. We further examined the endogenous H2S level in the striatum at the end of 7 weeks of observation. As shown in Fig. 1B, H2S was significantly decreased by ∼25% in the 6-OHDA-lesioned unilateral striatum when compared with that in the sham-operated rats. These data suggest that endogenous H2S production was decreased during the development of PD.

Figure 1.

 Behavioral test and endogenous hydrogen sulfide (H2S) level in the development of Parkinson’s disease in 6-hydroxydopamine (6-OHDA)-lesioned rats. (A) Time course for the contralateral rotations developed in the 6-OHDA-lesioned rats induced by apomorphine. Mean ± SEM. n = 8 in sham group, n = 12 in 6-OHDA group. #P < 0.05, ##P < 0.01 vs. the value at week 4 in the same group. (B) H2S level was decreased in the injured striatum of 6-OHDA-lesioned rat. n = 8–12. Mean ± SEM. #P < 0.05 vs. the corresponding value in the sham group.

Effect of NaHS treatment on behavioral symptoms in 6-OHDA-lesioned rats

To evaluate the therapeutic effect of H2S, rats in which the PD model had been successfully induced (≥ 7 turns per min contralateral rotation induced by apomorphine; PD rats) at week 4 after 6-OHDA lesion were treated with vehicle (saline) or NaHS (1.68 or 5.6 mg kg−1) for a further 3 weeks. As shown in Fig. 2, significant protective effects were observed at the end of 2 weeks of treatment with the high dose of NaHS (5.6 mg kg−1) and at the end of 3 weeks of treatment with both the low (1.68 mg kg−1) and high (5.6 mg kg−1) doses of NaHS. These data suggest that systematic administration of NaHS may alleviate the progress of apomorphine-induced rotational behavior in the unilateral 6-OHDA-lesioned rodent model of PD.

Figure 2.

 Treatment with NaHS ameliorated the contralateral rotational behavior in the development of Parkinson’s disease in unilateral 6-hydroxydopamine (6-OHDA)-lesioned rats. NaHS (1.68 or 5.6 mg kg−1 day−1, i.p) was given daily from the 4th to 6th week after 6-OHDA lesion. Mean ± SEM. n = 11–12, #P < 0.05, ##P < 0.01 vs. the corresponding value in week 4. *P < 0.05, **P < 0.01 vs. the values in Vehicle in the same group respectively.

Effect of NaHS treatment on dopaminergic neuronal degeneration in both SN and striatum

To examine the effect of H2S on neuronal degeneration, immunostaining and Western blotting were used to explore the extent of tyrosine-hydroxylase positive (TH+) neuron loss in both NaHS treated and untreated PD rats. As shown in Fig. 3A,B, unilateral 6-OHDA lesion induced a dramatic loss of dopaminergic neurons in both SN pars compacta (SNc, Fig. 3A) and striatum (Fig. 3B) of the injured hemisphere. The impairment was observed in neither region in the intact hemisphere. NaHS at both 1.68 and 5.6 mg kg−1 doses markedly attenuated the loss of TH+ neurons induced by 6-OHDA. These findings were further confirmed by Western blotting (Fig. 3C,D). These results indicated that H2S may have potential protective effects on 6-OHDA-induced dopaminergic neuron degeneration.

Figure 3.

 Effect of NaHS on 6-hydroxydopamine (6-OHDA)-induced tyrosine-hydroxylase positive (TH+) neuronal degeneration. (A, B) Immunohistochemistry showing NaHS (1.68 and 5.6 mg kg−1 day−1, i.p) alleviated TH+ neuron loss in both SN pars compacta (SNc) (A) and striatum (B) of 6-OHDA-lesioned Parkinson’s disease rats. Photos were taken at ×50 (A) and ×100 magnification (B). (C, D) Western blots showing H2S attenuated the down-regulation of TH expression induced by 6-OHDA in SNc (C) and striatum (D). n = 4 for both experiments, L: left (injured) side; R: right (intact side).

NaHS treatment suppresses excessive lipid oxidation in striatum of 6-OHDA-induced PD rats

We further investigated the anti-oxidative effect of H2S on 6-OHDA-induced dopaminergic neuron degeneration by examining the level of malondialdehyde (MDA), a marker for lipid peroxidation in striatum. As shown in Fig. 4, 6-OHDA stimulated massive MDA production in lesioned striatum compared with that in the sham group. This was also significantly higher than that in the intact side in the same PD rats. Treatment with NaHS (5.6 mg kg−1 day−1) abolished the elevated MDA formation, suggesting that the protective effects of H2S on 6-OHDA-induced neuron degeneration may be mediated by its anti-oxidative effect.

Figure 4.

 Effect of NaHS on malondialdehyde formation in striatum of 6-hydroxydopamine (6-OHDA)-induced Parkinson’s disease rat. Mean ± SEM, n = 5–7, #P < 0.05 vs. the corresponding value in the sham group, *P < 0.05 vs. injured side of striatum in the 6-OHDA-lesioned rat.

Neuroprotective effects involve ERK/NADPH oxidase pathway

gp91phox (also known as NOX2) is a primary membrane subunit of NADPH oxidase, which can be activated to generate large amounts of superoxide and leads to the downstream production of other ROS. In this study, we found that treatment with NaHS for 3 weeks suppressed the upregulation of gp91phox expression in both injured and intact striatum of 6-OHDA-induced PD rats (Fig. 5A). Treatment with NaHS (100 μm) or PD98059 (20 μm) for 6 h significantly inhibited the elevation of gp91phox levels induced by 6-OHDA in SH-SY5Y neuroblastoma cells, a dopaminergic neuronal cell line (Fig. 5B), suggesting a role of ERK in mediating the upregulation of gp91phox expression.

Figure 5.

 Effects of NaHS on 6-hydroxydopamine (6-OHDA)-induced activation of NADPH oxidase and ERK. (A) Treatment with NaHS (1.68 and 5.6 mg kg−1 day−1, i.p) for 3 weeks attenuated the upregulated gp91phox expression in 6-OHDA-lesioned striatum, n = 5. (B) Treatment with NaHS (100 μm) or/and PD98059 (20 μm) for 30 min attenuated 6-OHDA-induced upregulation of gp91phox in SH-SY5Y cells. (C) Treatment with NaHS (1.68 and 5.6 mg kg−1 day−1, i.p) for 3 weeks attenuated p47phox translocation in 6-OHDA-lesioned striatum, n = 5. (D) Treatment with NaHS (100 μm) or/and PD98059 (20 μm) for 2 h attenuated 6-OHDA-induced translocation of p47phox in SH-SY5Y cells. (E) Treatment with NaHS (1.68 and 5.6 mg kg−1 day−1, i.p) for 3 weeks attenuated ERK1/2 activation in 6-OHDA-lesioned striatum. (F, G) Treatment with NaHS (100 μm) for 30 min attenuated 6-OHDA-induced ERK1/2 activation in SH-SY5Y cells. *P < 0.05, ***P < 0.001 vs. 6-OHDA group. L: Left (injured) side; R: right (intact side). Mean ± SEM, n = 4–5. ##P < 0.01, ###P < 0.001 vs. control without 6-OHDA, *P < 0.05, **P < 0.01 and ***P < 0.001 vs. 6-OHDA group.

NADPH oxidase activation requires translocation of cytosolic subunits from cytosol to membrane. As shown in Fig. 5C,D, 6-OHDA stimulated translocation of p47phox in both 6-OHDA-lesioned striatum (Fig. 5C) and 6-OHDA-treated SH-SY5Y cells (Fig. 5D). This effect was significantly attenuated by NaHS treatment and blockade of ERK1/2 with PD98059, suggesting that H2S may target the cytosolic subunits of NADPH oxidase via ERK1/2 pathway.

To confirm the involvement of ERK1/2, we observed ERK1/2 activation in the presence and absence of 6-OHDA or NaHS. As shown in Fig. 5E, unilateral 6-OHDA lesion evoked activation of ERK1/2 in both injured and intact sides of PD rat brains. Treatment with NaHS at both 1.68 and 5.6 mg kg−1 for 3 weeks dose dependently attenuated this effect. The similar results were also found in SH-SY5Y cells. 6-OHDA (50 μm, 30 min)-induced ERK1/2 phosphorylation was significantly attenuated by either NaHS (100 μm) or PD98059 (20 μm), an ERK1/2 inhibitor (Fig. 5F,G). These data confirm that the therapeutic effect of H2S may also involve suppression of ERK1/2.

NaHS reverses 6-OHDA-induced oxygen consumption and ATP depletion

We also observed the involvement of metabolic inhibition in the neuroprotective effects of H2S. As shown in Fig. 6A, NaHS alone markedly decreased oxygen consumption in mitochondria isolated from striatum. Pretreatment with NaHS for 10 min significantly suppressed 6-OHDA-elevated oxygen consumption rate. However, NaHS alone had no significant effect on ATP synthesis, but markedly attenuated ATP depletion induced by 6-OHDA (Fig. 6B). These data suggest that metabolic inhibition may also be another mechanism for the neuroprotective effects of NaHS in the 6-OHDA-induced PD model.

Figure 6.

 Effect of NaHS on oxygen consumption (A) and ATP synthesis (B) in 6-hydroxydopamine (6-OHDA) treated tissues. (A) NaHS, which alone decreased oxygen consumption, reversed 6-OHDA-stimulated oxygen consumption in the striatum. (B) NaHS significantly attenuated the decreased ATP level in 6-OHDA treated SH-SY5Y cells. Mean ± SEM, n = 6–8, *P < 0.05, **P < 0.01 vs. the corresponding value in the control group, #P < 0.05 vs. the corresponding value in the 6-OHDA group.

NaHS alleviates rotenone-induced rat Parkinsonian symptoms and dopaminergic neuron degeneration in SN

To substantiate the therapeutic effect of H2S on PD, we established another PD model with rotenone. As shown in Fig. 7, NaHS (1.68 mg kg−1 day−1) was given 3 days before and throughout the rotenone administration for 4 weeks. The neurotoxic effect was examined with the behavioral test. Both total traveling distance (Fig. 7A) and rearing frequency (Fig. 7B) in rotenone-treated rats were significantly reduced compared to those in the sham group. These effects were remarkably attenuated by NaHS treatment.

Figure 7.

 Effects of NaHS on rotenone-induced rat behavioral disorders (A, B) and loss of TH-positive dopaminergic neuronal loss in the SN (C, D). NaHS treatment attenuated rotenone-induced impairment in the total travel distance during 5 min in an open arena (A, locomotor activity test) and rearing frequencies (B). Both immunohistochemistry (C) and Western blots (D) showed that NaHS treatment protected rotenone-induced loss of TH-positive dopaminergic neurons in the rat SNpc. Sham, vehicle-injected group; Rot, rotenone-injected group; NaHS, NaHS (1.68 mg kg−1 day−1) + rotenone-treated group (n = 6–15). Data are presented as means ± SEM. #P < 0.05; ##P < 0.01 vs. sham group; *P < 0.05, ***P < 0.001 vs. rotenone group.

Immunohistochemical study showed that rotenone also induced a dramatic loss of TH positive cells in SN (Fig. 7C-b vs. the sham group in Fig. 7C-a), confirming the successful establishment of the PD model. This effect was attenuated by NaHS treatment (Fig. 7C-c). The protective effects of NaHS were further confirmed by our Western blotting data which showed that NaHS prevented rotenone-induced down-regulation of TH protein in the SN (Fig. 7D).

NaHS inhibits rotenone-induced microglial activation and accumulation of inflammatory mediators in the rotenone-induced PD model

The persistent activation of microglia accompanied by the sustained secretion of inflammatory mediators is thought to be involved in the neuronal injury in PD. We therefore examined the microglia activation in these rats. As shown in Fig. 8A, treatment with NaHS suppressed the upregulated expression of ED-1, a marker for microglia activation, in the rat SN. The beneficial effect of NaHS was further supported by the inhibitory effect of NaHS on rotenone-induced release of inflammatory mediators. As shown in Fig. 8B,C, rotenone significantly elevated the levels of NOx (NO3 and NO2) and TNF-α in the striatum. NaHS treatment inhibited the accumulation of these mediators. Taken together, our results suggest that NaHS may produce therapeutic effect on PD via inhibition of rotenone-induced microglial activation.

Figure 8.

 Effect of NaHS on rotenone-induced microglial activation in the nigrostriatal pathway in rats. (A) Immunostaining of ED 1 showed that NaHS treatment (1.68 mg kg−1 day−1) reduced the number of ED 1-positive cells in the SNpc in rotenone-treated rats. (B, C) NaHS administration suppressed the accumulation of TNF-α and NO in the striatum in rotenone-induce Parkinson’s disease model rat. N = 10–11, Means ± SEM. ###P < 0.001vs. sham group; *P < 0.05, **P < 0.01 vs. rotenone group.

NaHS attenuates rotenone-induced activation of NF-κB pathway

NF-κB plays a critical role in the inflammatory pathogenesis. The activation of NF-κB was determined by the phosphorylation and degradation of IκB. In this study, we found that rotenone (10 nm) induced phosphorylation (middle panel) and degradation (upper panel) of IκBα in BV-2 microglial cells in a time-dependent manner (Fig. 9A). The maximal response was reached at 15 min after rotenone addition. Both effects were significantly suppressed by NaHS pretreatment (Fig. 9B).

Figure 9.

 Inhibitory effects of NaHS on rotenone-induced NF-κB activation in microglial cells. (A) Time-course for rotenone-induced phosphorylation and degradation of IκBα. (B) Pretreatment with NaHS (100 μm) suppressed the rotenone-induced phosphorylation and degradation of IκBα. n = 5. (C, D) Western blots (C, n = 4) and confocal microscopy (D) showed that pretreatment with NaHS (100 μm) for 10 min reversed rotenone (10 nm, 4 h) induced translocation of p65/RelA from cytosolic to nuclear fraction. Histone H3 and β-actin were used here as loading control for nuclear and cytosolic fractions respectively. Representative confocal microscopy photos from two independent experiments. Means ± SEM. ##P < 0.01, ###P < 0.001 vs. control group; **P < 0.01 vs. rotenone group.

As NF-κB activation requires nuclear translocation of p65/RelA subunit of NF-κB, we further examined the effect of NaHS on the cytosolic and nuclear pools of p65/RelA protein in BV-2 cells. It was found that rotenone treatment markedly increased the nuclear p65/RelA protein level and decreased the cytosolic p65/RelA protein pool. NaHS treatment abolished the nuclear translocation of p65/RelA (Fig. 9C). This was confirmed by confocal microscopy. As shown in Fig. 9D, p65/RelA (green) was mainly localized in the cytoplasm in untreated cells. Rotenone induced translocation of p65/RelA (merged color: light blue) from cytoplasm to nuclear which was stained by Hoechst 33342 (dark blue). Pretreatment with NaHS 100 μm, which alone had no effect, significantly attenuated rotenone-induced accumulation of p65 in the nuclear in BV-2 cells.

NaHS attenuates the microglia-mediated neurotoxicity

This series of experiments was designed to examine whether the protective effects of NaHS on dopaminergic neurons are mediated by suppression of microglia activation or not. As shown in Fig. 10, exposure to the conditioned media from rotenone (10 nm, 36 h)-treated microglia significantly decreased the cell viability of SH-SY5Y by 35%. However, the conditioned media from the microglia treated with both NaHS and rotenone only reduced the cell viability of SH-SY5Y by 16%, which was significantly lower than that caused by the conditioned media from rotenone alone-stimulated microglia. Neither rotenone (10 nm) alone for 36 h nor supernatants from unstimulated microglial cells induced obvious neuronal cell death. These data suggest that the neuroprotective effect of NaHS was secondary to the suppression of toxic mediators release during microglial activation.

Figure 10.

 Effects of NaHS on microglia-mediated neurotoxicity. Cell viabilities of SH-SY5Y cells were measured 36 h after treatment with conditioned media from BV-2 microglial cells exposed to rotenone (10 nm) with or without NaHS (100 μm) pretreatment. N = 6–8, means ± SEM. ###P < 0.001 vs. control group; *P < 0.05 vs. rotenone group.


Parkinson’s disease is the second most prevalent age-related neurodegenerative disease. Although the etiology of PD has been intensively pursued for several decades, genetic and epigenetic factors leading to initiation and progression of the disease remain elusive (Miller et al., 2009). Among all the pathological factors, oxidative stress and neuroinflammation remain to be the leading theory for explaining the development of PD. Oxidative stress has been demonstrated to be an early sign that often precedes and triggers neuronal death in PD (Shibata & Kobayashi, 2008). Thus, inhibition of oxidative injury and subsequent dopaminergic neuronal loss may offer prospective clinical therapeutic benefit for PD and other oxidative stress-related neurodegenerative disorders.

To study the role of endogenous H2S in the progress of PD, we measured H2S levels in the striatum at the end of week 4 after 6-OHDA lesion. It was found that the endogenous H2S level in the injured striatum was significantly lower than that in sham group. The lowered endogenous H2S level was also found in SN in rotenone-induced PD model (data not shown). However, it is worth noting that 6-OHDA also slightly lowered H2S levels in the intact side as well. Thus, the possibility that the decreased level of H2S only correlates with toxicity of 6-OHDA but may not be associated with the mechanism of its toxicity cannot be excluded. Although our current study may not be able to tell whether the lowered H2S in these models is a causative mechanism or just a correlative finding in the development of PD, these interesting findings impel us to continue to study the therapeutic value of exogenous application of H2S.

We found in the present study that systemic treatment with NaHS significantly ameliorated neurotoxin-induced rotational behavior in the 6-OHDA-induced rats and movement dysfunction in the rotenone induced PD models. Our immunohistochemical data further reveal that NaHS may protect brain from TH+ neuronal loss in both SNc and striatum. These findings indicate that H2S may be a potential therapy for PD.

By generation of neurotoxic oxidative stress, NADPH oxidase (phox) has long been associated with neurodegenerative diseases. NADPH oxidase consists of the membrane subunits [gp91phox (also known as NOX-2) and p22phox], and the regulatory cytosolic subunits (p67phox, p47phox, p40phox and the small GTPase Rac1/2). Functional activity of this enzyme requires the assembly of cytosolic and membrane subunits. Upon activation, the cytosolic subunits (p40, p47 and p67) translocate to and associate with the membrane bound heterodimer p22-gp91. The fully assembled enzyme catalyzes the electron transfer from NADPH to molecular oxygen to form superoxide (McDonald et al., 1997) and induces downstream pathological progress. This process in the context of PD may either occur in microglia or in neurons. In clinical research, postmortem analysis of the brains in PD patients reveals an up-regulation of the catalytic subunit of gp91phox. The finding confirms the relationship between activation of NADPH oxidase and PD. In the present study, we found that NaHS significantly attenuated the excessive translocation of p47phox, upregulation of gp91phox and consequent ROS production caused by 6-OHDA, suggesting that H2S may protect neurons against 6-OHDA-induced injury via suppression of NADPH oxidase.

Apart from oxidative stress, activation of neuroinflammatory cells, especially microglia, with subsequent production of pro-inflammatory cytokines and molecules also aggravates PD genesis (Gao et al., 2002; Block & Hong, 2007; Wilms et al., 2007). Clusters of activated microglial cells located around neurons have been observed in the SN in different Parkinsonian syndromes. The pro-inflammatory mediators such as NO and TNF-α have been shown to synergistically contribute to neuronal damage and death in vitro and in vivo (Gao et al., 2003; Loane et al., 2009). Moreover, TNF-α level has been shown to be increased in postmortem brain and CSF from patients with PD (Mogi et al., 2000; Nagatsu et al., 2000; McGuire et al., 2001; Vila et al., 2001). Therefore, microglial activation may not just be insignificant epiphenomena of neuronal death but most likely contributes to secondary damage and further progression of neurodegeneration. For this reason, the inhibition of microglial activation and subsequent production of pro-inflammatory mediators may offer prospective therapeutic value for the treatment of neuroinflammation-related neurodegenerative disorders including PD. In this study, we found that H2S suppressed microglia activation in SN and reduced the subsequent release of endogenous inflammatory mediators (NO and TNF-α) in striatum in rotenone-treated PD model. NF-κB plays a critical role in the inflammatory pathogenesis. The anti-inflammatory effect of NaHS was, therefore, further confirmed by examining its inhibitory effect on NF-κB activity. We found in the present study that NaHS suppressed the phosphorylation and degradation of IκB and the consequent translocation of p65/RelA subunit of NF-κB. These data confirm that H2S exerted anti-inflammatory effects via the inhibition of NF-κB pathway.

Our above experiments indicate that the neuroprotective effects of NaHS may be also secondary to the suppression of microglia activation. This was confirmed by our microglia-mediated neurotoxicity experiment. The conditioned media from microglia treated with rotenone decreased cell viability in SH-SY5Y neuronal cells, but this effect was significantly alleviated in the neuronal cells treated with the conditioned medium from NaHS plus rotenone treatment group. More importantly, rotenone at the same concentration (10 nm) failed to decrease cell viability in SH-SY5Y cells. These data clearly suggest that the protective effect of NaHS on neurons in the current study was, at least in part, from the suppression of rotenone-induced microglia activation and the subsequent released proinflammatory mediators.

As NaHS was reported to produce metabolic inhibition (Nicholson et al., 1998; Elrod et al., 2007; Kiss et al., 2008), we also investigated the effect of H2S on oxygen consumption and ATP synthesis. We found that H2S markedly decreased oxygen consumption but had no significant effect on ATP production. The underlying mechanism for this effect is still not clear but may be related to mitochondrial transitions. Like Na+/K+ ATPase inhibitors, H2S may also be able to induce the active-to-resting transition, which causes higher efficiency in NADH production (Balaban et al., 1980). This may induce more ATP production but with low oxygen consumption (Hahn et al., 2000; Guzy et al., 2003). Our finding that H2S inhibits oxygen consumption is consistent with the previous reports (Nicholson et al., 1998; Elrod et al., 2007). However, the effect of H2S on ATP production in our experimental situation is different from the observation of Kiss et al. (2008). Kiss et al. reported that H2S acutely (∼3 min) decreases ATP production in vascular tissue. The discrepancy between our and Kiss’s findings may probably result from the different tissues used and different time points observed in the two studies.

In line with previous reports (Thakar & Hassan, 1988), we also found that 6-OHDA inhibited ATP production but increased oxygen consumption in our experiments. The mechanisms for this effect may be related to 6-OHDA-induced mitochondrial oxidative phosphorylation uncoupling (Thakar & Hassan, 1988; Noelker et al., 2005). When H2S was given before 6-OHDA, we found that H2S reversed the effect of 6-OHDA on oxygen consumption and ATP production. Taken together, our data suggest that H2S-induced metabolic inhibition may also contribute to its therapeutic effects on PD.

In conclusion, the present study demonstrates for the first time that systemic administration of H2S ameliorates behavioral symptoms and dopaminergic neuronal degeneration in both 6-OHDA and rotenone induced PD models. H2S may serve as a neuroprotectant to treat neurotoxin-induced neurodegeneration via inhibition of NADPH oxidase, anti-microglia activation and suppression of oxygen consumption therefore has potential clinical therapeutic value for treatment of PD.

Experimental procedures

The experimental protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of National University of Singapore.

6-OHDA induced PD model and the behavioral test

Male Sprague–Dawley (SD) rats (180−220 g) were anesthetized with ketamine (75 mg kg−1, i.p.) and xylazine (10 mg kg−1, i.p.). After anesthesia, animals were placed in a stereotaxic apparatus (Stoelting Instruments, Wood Dale, IL, USA). Unilateral injection of 6-OHDA (8 μg 6-OHDA hydrochloride in 4 μL of 0.02% ascorbic acid saline solution) was performed in the left striatum (coordinates from bregma: AP, +3.0 mm; ML, +1.0 mm; DV, −4.5 mm) with a Hamilton syringe (0.46 mm in diameter) at a rate of 0.5 μL min−1, the needle was left in the place for 3 min after injection. Sham-lesioned rats were infused with 4 μL saline containing 0.02% ascorbic acid into the left striatum and served as controls. The rats after surgery were kept in cages with constant temperature (25 °C) and humidity for 4 weeks. They were exposed to a 12:12-h light–dark cycle and had unrestricted access to tap water and food.

At the end of the third week after the lesion, the animals’ tendency to rotate in response to apomorphine (0.5 mg kg−1, s.c.) was tested. Contralateral rotations induced by apomorphine were measured with a video camera two times at weekly intervals. Only in those animals showing at least 7 turns per min in both tests was the model considered to be successfully induced. The successful model rats were divided into three groups: a 6-OHDA model vehicle treatment group, a NaHS low dose (1.68 mg kg−1) treatment group and a NaHS high dose (5.6 mg kg−1) treatment group. Additionally, sham-operated rats received vehicle treatment. All rats received vehicle or NaHS treatment for 3 weeks. At the end of treatment, apomorphine (0.5 mg kg−1, s.c.) was administered to test for changes in the rotational behavior.

The doses of NaHS were selected based on the findings from previous publications (Warenycia et al., 1989; Han et al., 2005; Qu et al., 2006). In our preliminary experiments, we found that the plasma H2S level in the NaHS high dose (5.6 mg kg−1 day−1) group was increased by 20% when compared with that in the vehicle rats.

Rotenone-induced PD model and the behavioral test

Male SD rats aged 7 weeks (180–220 g) were chosen for experiments. Dimethylsulfoxide (DMSO)/polyethylene glycol (1:1) was used as vehicle for rotenone. NaHS, an H2S donor, was dissolved in sterile saline. Rats were randomly divided into three groups. Control rats (sham group) received vehicle only. Rats of model group were subcutaneously injected with rotenone (2.5 mg kg−1) daily for 4 weeks and the daily dose of rotenone was administered t.i.d. at 8 am, 2 pm and 8 pm respectively. The injection volume was 0.1 mL per 100 g body weight. The NaHS-treatment group rats were intraperitoneally injected with NaHS (1.68 mg kg−1 day−1) for three consecutive days prior to rotenone administration. From the fourth day on, NaHS was pretreated 1 h before the rotenone injection on a daily basis for 4 weeks.

A battery of behavioral tests sensitive to varying degrees of dopamine loss in the striatum and SN were used in this study to assess motor function and evaluate the effects of NaHS on rotenone-induced Parkinsonian symptoms. All rats were tested at two time points, before and after 4 weeks treatment. As the rats of rotenone group underwent an obvious mortality (50%) after 4 weeks treatment, only rats that went through all 4 weeks procedure were included in the following statistical analyses.

Locomotor activity was detected by measuring animal travel distance in an open-field arena. The movement activity of each rat in the arena was automatically monitored and detected with a computer-based video tracking system during a 5-min test session. The arena was cleaned after each session. Traveling distance information was processed online with EthoVision software (Noldus, Wageningen, The Netherlands).

To measure rearing activity, animals were placed in a clear Plexiglas cylinder (17.6 cm inside diameter, 34 cm height). All animals were tested and videotaped. While in the cylinder, animals typically rear and engage in exploratory behavior by placing their forelimbs along the wall of the cylinder. Because rotenone affects motor behavior bilaterally, activity was measured by counting the number of rears made by each animal in a 5-min period without recording specific limb use. Videotapes were rated and analyzed by an experimenter blind to experimental condition.

Measurement of hydrogen sulfide

Hydrogen sulfide concentration in brain tissue sample was measured essentially as described in previous publications (Li et al., 2005). Briefly, brain tissues were lyzed and homogenized in chilled KHPO4 buffer (1:10). Aliquots (200 μL) of homogenized supernatant were mixed with potassium phosphate buffer (pH 7.4, 350 μL), zinc acetate (1% w/v, 250 μL), followed by incubation with N,N-dimethyl-p-phenylenediamine sulphate (20 mmol L−1, 133 μL) in 7.2 mol L−1 HCl and FeCl3 (30 mmol L−1, 133 μL) in 1.2 m HCl. Reactions were terminated by trichloroacetic acid (10% w/v, 250 μL) after 15 min color development. The resulting solutions (300 μL) were transferred to 96-well plate and the absorbance of the mixture (670 nm) was measured. H2S was calculated against a calibration curve of NaHS (0.01–100 μmol L−1).

Immunohistochemistry staining

At the end of experiments, animals were perfused with 4% paraformaldehyde (PFA). Brain samples were collected and postfixed in 4% PFA at 4 °C overnight. They were transferred to 15% sucrose in phosphate-buffered saline (PBS) overnight and then to 30% sucrose overnight till the brain sunk to the bottom of the tube. The brain tissues were then sectioned on a cryostat at 30 μm section and mounted on the poly-l-lysine coated slides. Immunostaining was performed by a commercial kit [EnVision+ Dual Link System-HRP (DAB+); Dako, Carpinteria, CA, USA]. Briefly, the sections were permeabilized with PBS + 0.3%Triton X-100 for 10 min and blocked with 10% goat serum in PBS for 30 min and then incubated with primary antibody (anti-TH, 1:500; Sigma, St. Louis, MO, USA; anti-ED 1, 1:100; Serotec, Raleigh, NC, USA) for 2 h and appropriate anti-rabbit secondary antibody for 1 h. Immunostaining was visualized by using substrate-chromogen solution for 10 min. Sections were then counterstained with hematoxylin and ammonia.

Cell cultures

BV-2 and SH-SY5Y cells were incubated under humidified 5% CO2 and 95% O2 at 37 °C in Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA) medium containing 10% fetal bovine serum and 1% streptomycin and penicillin (Invitrogen, Carlsbad, CA, USA).

Cell fractionation and detection of protein translocation

A cell fractionation technique was adopted from previous studies (Mackay & Mochly-Rosen, 2001; Weber et al., 2005). After treatment, SH-SY5Y cells were lyzed with 90 μL ice-cold lysis buffer (Cell signaling, Danvers, MA, USA) and shaken on ice for 1 h. The cell lysate was centrifuged at 1000 × g at 4 °C for 10 min for rough partition between cytosolic and membrane fractions. The supernatant was recentrifuged at 16 000 × g at 4 °C for 15 min to get rid of contaminating pellet materials and collected as cytosolic fraction. The initial pellets were resuspended in 45 μL cell lysis buffer containing 1% Triton X-100. After shaking on ice for another 60 min, samples were then centrifuged at 16 000 × g at 4 °C for 15 min. The second supernatant was collected as plasma membrane fraction. Epitopes were exposed by boiling the protein samples at 95 °C water for 5 min. Each fraction was analyzed for protein content by the Bradford method. Equal amounts of protein were loaded and electrophoresed with 10% SDS-polyacrylamide gel for Western blot assays as described below. The membrane and cytosolic fractions were probed with antibody against p47phox (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA).

TNF-α and nitric oxide assay

A sandwich enzyme-linked immunosorbent assay was used for detecting TNF-α (R&D system, Minneapolis, MN, USA) in rat brain tissue extracts with detection limit at 5 pg mL−1. The production of nitric oxide was determined by measuring the accumulated level of its metabolite, nitrate and nitrite, with the Griess reagent (Cayman, Ann Arbor, MI, USA), with detection limit of 0.5 μm.

Western blot assays

Cells were washed twice with PBS after treatment and solubilized in RIPA lysis buffer (Cell Signaling). Tissue samples were homogenized in tissue lysis buffer (1:10, w/v; Sigma). Protein concentrations were determined by the Bradford method. Protein samples (30 μg) were separated by 10% SDS/PAGE and transferred onto a nitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, USA). After blocking in 10% milk with TBS-T buffer (10 mm Tris–HCl, 120 mm NaCl, 0.1% Tween-20, pH 7.4) for 1 h at room temperature, the membrane was incubated with various primary antibodies (1:1000) at 4 °C overnight. Membranes were then washed three times in TBST buffer, followed by incubation with 1:10 000 dilutions of HRP-conjugated anti-rabbit/goat IgG at room temperature for 1 h, and washed three times in TBST. Visualization was carried out using an ECL® (advanced chemiluminescence) kit (GE Healthcare, Bucks, UK). The density of the bands on the Western blots was quantified by densitometric analysis of the scanned blots using ImageQuant software. The relative phosphorylation was normalized to total protein.

NF-κB activation assay

Cell fractionation was performed with a commercial compartment protein extraction kit from Chemicon according to the manufacturer’s instructions. Protein samples were then subjected to Western blot analysis as described above for p65 nuclear translocation study. This was confirmed by confocal microscopy combined with immunocytochemistry to visualize p65 nuclear translocation. Briefly, BV-2 cells were fixed with ice-cold methanol and were permeabilized with 0.25% Triton X-100/PBST. After blocking with 1% BSA in PBST for 60 min, the coverslip was incubated for 2 h at room temperature with the p65/RelA antibody diluted in 1% BSA (1:50). The coverslip was then washed three times with PBS and further exposed to the FITC-conjugated anti-rabbit IgG at 1:100 at room temperature for 60 min. After that, the coverslip was stained with 1 μg mL−1 Hoechst 33342 for 1 min and the fluorescent confocal image was acquired using a Leica TCS SP5 confocal system (Leica Microsystems GmbH, Wetzlar, Germany).

Microglia-mediated neurotoxicity assay

BV-2 microglial cells were pretreated with or without NaHS (100 μm) for 10 min followed by the addition of rotenone (10 nm). After incubation for 24 h, cell-free culture supernatant was collected as conditioned medium and then transferred into SH-SY5Y cell dishes. SH-SY5Y cells were further incubated with the conditioned medium for another 36 h before cell viability was assayed with MTT method. The absorbance was obtained using a Tecan M200 plate reader at a wavelength of 570 nm with a reference wavelength of 630 nm.

Intracellular ATP assay

Intracellular ATP content assay was performed according to the kit instruction (BioThema, Handen, Sweden). SH-SY5Y cells were treated with vehicle or 6-OHDA for 30 min. In NaHS and NaHS + 6-OHDA groups, NaHS was given 10 min before addition of vehicle or 6-OHDA respectively. After adding 20 μL ATP Eliminating Reagent to each well of 96-well plate, cells were transferred into each well and incubated for 10 min with occasional shaking at room temperature. Twenty microliter extractant B/S and 160 μL ATP reagent HS were then added into samples followed by mixing. The reaction is: ATP + D-luciferin + O2→AMP + PPi + oxyluciferin + CO2 + light. Light emission of samples was measured in a luminometer at wavelength of 485 nm. The intensity of the light is proportional to the amount of intracellular ATP.

Mitochondria preparation and oxygen consumption analysis

The method was described in detail in a recent publication (Will et al., 2006). Briefly, striatum was rapidly excised and homogenized in ice-cold isolation buffer I (210 mm mannitol, 70 mm sucrose, 5 mm HEPES, 1 mm EGTA, 0.5% w/w BSA, pH 7.4). After centrifugation, the mitochondrial pellet was washed by re-suspending it in 20 mL of isolation buffer I once and re-suspended in 0.7 mL of isolation buffer II (210 mm mannitol, 70 mm sucrose, 10 mm MgCl2, 5 mm K2HPO4, 10 mm MOPS, 1 mm EGTA, pH 7.4). Same amount of mitochondria was added into each well of a black 96-well plate containing of 100 μL of respiration buffer (250 mm sucrose, 15 mm KCl, 1 mm EGTA, 5 mm MgCl2, 30 mm K2HPO4, pH 7.4.) and MitoXpress probe (Luxcel, Ireland). Fifty microliter of substrate solution was given to form a final concentration of 12.5/12.5 mm glutamate/malate containing 1.65 mm ADP. One hundred microliter of prewarmed heavy mineral oil was quickly added into each well using a syringe dispenser. The fluorescent signal of each well was read with a Safire microplate reader (TECAN, excitation at 380 nm and emission at 650 nm) for 30 min at 1 min intervals. The fluorescence-time profiles of different treatment were linearized using the following coordinate scale [Y: I(t0)/(I(t)-I(t0)), where I(t0) and I(t) represent fluorescence intensity signals at the start and at time t of monitoring respectively. X:1/t (min−1)]. Linear regression was applied to each transformed profile and the slope represents the oxygen consumption rate.

Assessment of lipid peroxidation

To detect the level of lipid peroxidation in brain homogenates, MDA, a marker of oxidative stress, was measured using a commercially available thiobarbituric acid reactive substances assay kit (Cayman Chemical). The assay was performed according to the manufacturer’s instructions. In brief, brain tissues were lyzed with chilled RIPA buffer and sonicated for 15 s at 40 V over ice. After centrifugation at 1600 g for 10 min at 4 °C, the supernatant was collected for further analysis. The MDA-TBA adduct formed by the reaction of MDA in samples and TBA supplied in the assay kit under high temperature (100 °C) and acidic conditions. Reaction product is measured colorimetrically at 540 nm with a spectrophotometer (Tecan M200). The content of MDA in samples was calculated using a MDA standard curve and expressed as micromolar of MDA produced per gram of protein.

Statistical analysis

All data were presented as mean ± SEM. Statistical significance was assessed with one-way analysis of variance followed by a post hoc (Turkey) test for multiple group comparison. Differences with P-values of less than 0.05 were considered statistically significant.


The authors gratefully thank Tan Choon Ping for the technical assistance. This work was supported by research grants from Singapore National Medical Research Council (1183/2008) and the National Natural Science Foundation of China (No. 30873055).