Crosstalk between hydrogen sulfide and nitric oxide in endothelial cells

Hydrogen sulfide (H2S) and nitric oxide (NO) are major gasotransmitters produced in endothelial cells (ECs), contributing to the regulation of vascular contractility and structural integrity. Their interaction at different levels would have a profound impact on angiogenesis. Here, we showed that H2S and NO stimulated the formation of new microvessels. Incubation of human umbilical vein endothelial cells (HUVECs-926) with NaHS (a H2S donor) stimulated the phosphorylation of endothelial NO synthase (eNOS) and enhanced NO production. H2S had little effect on eNOS protein expression in ECs. L-cysteine, a precursor of H2S, stimulated NO production whereas blockage of the activity of H2S-generating enzyme, cystathionine gamma-lyase (CSE), inhibited this action. CSE knockdown inhibited, but CSE overexpression increased, NO production as well as EC proliferation. LY294002 (Akt/PI3-K inhibitor) or SB203580 (p38 MAPK inhibitor) abolished the effects of H2S on eNOS phosphorylation, NO production, cell proliferation and tube formation. Blockade of NO production by eNOS-specific siRNA or nitro-L-arginine methyl ester (L-NAME) reversed, but eNOS overexpression potentiated, the proliferative effect of H2S on ECs. Our results suggest that H2S stimulates the phosphorylation of eNOS through a p38 MAPK and Akt-dependent pathway, thus increasing NO production in ECs and vascular tissues and contributing to H2S-induced angiogenesis.


Introduction
Hydrogen sulfide and NO are known gasotransmitters that contribute to many physiological functions [1]. These gaseous messengers can be produced endogenously to respond to diverse physiological and pathophysiological stimuli [1]. In ECs, H 2 S can be generated from Lcysteine by the enzymatic action of CSE (EC 4.4.1.22) [2]. Hydrogen sulfide-induced relaxation of vascular tissue was partially reduced by the removal of the vascular endothelium or in the presence of L-NAME (an inhibitor of NO synthase) [3]. Nitric oxide can be generated in ECs from L-arginine by eNOS (EC 1.14.13.39) [4]. Being a homodi-meric protein, the activation of eNOS is dependent on intracellular calcium (Ca 2+ ) level and other cofactors like nicotinamide adenine dinucleotide phosphate (NADPH), tetrahydrobiopterin (BH 4 ), flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN) [5]. The activity of eNOS is affected by many post-translational modification mechanisms, such as phosphorylation on multiple amino acids like Ser-1179/1177 (bovine/human) and Thr-495 residues [6,7], whereas eNOS protein can be self-inhibited by high concentrations of NO through S-nitrosylation [8]. Due to its reducing capability, H 2 S may reduce NO to form a thiol-sensitive molecule S-nitrothiols (RSNO) [9]. Conversely, H 2 S has been found to reduce RSNO to release NO from GSNO (S-Nitrosoglutathione) [10]. Moreover, H 2 S and NO interact on each other's catalysing enzymes; NO donor increases the expression and activity of CSE in cultured aortic smooth muscle cells (SMCs) [3]. In rat vascular SMCs, H 2 S had no direct effect on NO production, but it augmented interleukin-induced NO production, and this effect was related to increased iNOS expression (inducible NOS) [11]. NaHS (a H 2 S donor) treatment reduced eNOS activity and expression but not of nNOS (neuronal NOS) and iNOS in isolated rat aortas and HUVECs [12]. NaHS inhibited eNOS-catalysed conversion of [ 3 H]arginine to [ 3 H]-citrulline [13]. NaHS also inhibited iNOS expression and NO production in macrophage cells (RAW264.7) [14]. Na 2 S selectively augmented NO production in chronically ischaemic tissues, by influencing iNOS and nNOS expression and stimulating nitrite reduction to NO via xanthine oxidase (XO) under hypoxic condition [15].
The angiogenic crosstalk between H 2 S and NO in ECs has been unclear. Our present study showed that the pro-angiogenic effect of H 2 S appears to be regulated by both a NO-dependent and an independent mechanism, whereas NO effect on angiogenesis is partially dependent on H 2 S. We demonstrated that H 2 S stimulated NO release by increasing eNOS phosphorylation via a p38 MAPK and Akt-dependent mechanism, which contributes to the stimulatory effect of H 2 S on EC proliferation and angiogenesis.

Measurement of NO production
Total nitrate/nitrite concentrations were measured by conversion of nitrate to nitrite after incubating supernatants with nitrate reductase (10 U/ml) and NADPH (5 mM) for 1 hr at 37°C. The total nitrite was measured with a Griess assay kit (Promega, Madison, WI, USA) using a reference sodium nitrate standard curve [18]. The results obtained with the Griess assay have also been validated by the diaminofluorescein fluorophore system (DAF-FM), which can be deacetylated by intracellular esterases and further reacts with NO to form a fluorescent benzotriazole (DAF fluorescence; Invitrogen, Burlington, ON, Canada). Endothelial cells were incubated with 5 lM DAF-FM for 30 min. at 37°C. The cells were washed to remove excess dye, replaced with fresh medium and observed under a fluorescent microscope as previously described [19]. To detect the production of NO in aortic tissues, isolated aortas were incubated with DAF-FM (5 lM) at 37°C in Kreb's buffer and then rapidly removed and frozen at À20°C. Aortic tissue samples were embed-ded in optimal cutting temperature (OCT) compound until frozen, and sectioned using Leica CM1850 UV microtome-cryostat (Leica Biosystems, Concord, ON, Canada). The tissue blocks were cut into 10-lm thick sections and observed under a fluorescent microscope [20].

Gene knockdown and overexpression
Endothelial cells were seeded in 6-well plates and cultured until they reached 70-80% confluence. The cells were then transfected with specific siRNA to knockdown CSE or eNOS gene (50 nM). Negative siRNA was used as transfection control (50 nM), using Lipofectamine TM RNAi-MAX transfection reagent according to the manufacturer's instruction (Invitrogen). Overexpression experiments were carried out with plasmid DNA containing CSE cDNA (pIRES2-EGFP, 4.0 lg) or eNOS cDNA (pcDNA 3.1 eNOS-GFP, 4.0 lg). Mock empty vector was used as transfection control (Addgene, Cambridge, MA, USA) [21][22][23] using Lipofectamine TM 2000. Forty-eight hours after transfection, the cells or media were collected and evaluated by Western blot or Griess assay analysis.

Western blot analysis
Cultured cells were collected and incubated in a lysis buffer containing 0.5 M EDTA, 1 M Tris-Cl (pH 7.4), 0.3 M sucrose and a protease inhibitors mixture (Sigma-Aldrich). The cell extracts were sonicated three times (5-10 sec./each) on ice using a cell sonicator (Sonic Dismemrator Model 100, Fisher Scientific, Ottawa, ON, Canada) [23]. Cellular extracts were separated by centrifugation at 14,000 9 g for 15 min. at 4°C. Supernatants were collected, and the same amounts of proteins were separated on 10% SDS-polyacrylamide gels and blotted onto nitrocellulose membranes (Pall Corporation, Pensacola, FL, USA). All primary antibody incubations were performed at 4°C overnight. The antibody dilution for phospho-eNOS (Ser1177), eNOS, phospho-ERK, ERK, phospho-p38 MAPK, p38 MAPK, phospho-Akt (S473) and Akt was at 1:1000 (Cell Signaling Technologies). Anti-CSE antibody was used at 1:5000 (Proteintech Group, Chicago, IL, USA), and anti-b-actin antibody was at 1:10000 (Sigma-Aldrich). The membranes were stripped using a buffer containing 100 mM b-mercaptoethanol, 2% SDS and 62.5 mM Tris-HCl (pH 6.8) at 50°C for 30 min. Membranes were visualized using enhanced chemiluminescence Western blotting system (GE Healthcare, Piscataway, NJ, USA). Densitometric quantification was performed using Alpha Digi Doctor Software (Richardson, TX, USA). The protein bands were quantified and normalized against either b-actin or total form levels of the target protein, and expressed as a percentage relative to the controls (equals 100%). The phosphorylation level is defined as the ratio between the phosphorylated target proteins and their total forms and expressed in the summarized bar graphs as the percentage of the untreated controls.

Capillary-like tube formation assay
The Matrigel matrix gel was thawed overnight at 4°C on ice and then added to pre-chilled culture dishes and allowed to polymerize at 37°C for 1 hr. Endothelial cells (2 9 10 4 cells) were incubated with different agents in 500 ll DMEM and then seeded onto the surface of Matrigel (BD Biosciences). After 12 hrs, the formation of capillary-like structure was imaged by light microscope. The total lengths of tube-like structures per field were measured using image analysis software (NIH Image software-Image J).

Cell proliferation assay
Cells were counted using automated cell counter TC10 Tm from BioRad (Mississauga, ON, Canada) and seeded into 96-well plates (1 9 10 4 cells/well). After 24 hrs of initial seeding, cells were incubated with DMEM serum-free medium for overnight [23]. The proliferation rates were evaluated by 5-bromo-2′-deoxyuridine (BrdU) incorporation assay according to the manufacturer's instructions (EMD Biosciences, San Diego, CA, USA).

Microvessel formation assay
Cystathionine gamma-lyase knockout (KO) mice were generated as described previously [2]. Eight-week-old male CSE-KO and wild-type (WT) mice were sacrificed, and aorta were rapidly cleaned off adipose tissues and blood. Aorta were cut into rings (length,~3 mm) and implanted in a fibrin gel obtained by adding 400 ll of a fibrinogen solution (3 mg/ml) and thrombin (1.5 U/ml); (Sigma-Aldrich). The fibrin gels were given 30 minutes to solidify before different treatments were applied. As a control, the effect of medium alone was assayed, and quantitative evaluation of new microvessels was carried out after 72 hrs [24]. All animal experiments were conducted according to the Care and Use of Laboratory Animals Guide (NIH Publication No. 85-23, revised 1996) and approved by Lakehead University Animal Care Committee, Canada.

H 2 S-induced NO production in ECs
Stimulation of ECs with NaHS for 30 min. increased NO production over a concentration range from 10 to 100 lM (Fig. 1A). The effect of NaHS on NO production was blocked when cells were pre-treated with NOS inhibitor Nx-L-NAME (Fig. 1B). NaHS-induced increase in NO production was further confirmed in primarily cultured mouse ECs ( Figure S1). Nitro-L-arginine methyl ester treatment also significantly reduced NO production. We next determined the effect of L-cysteine (H 2 S precursor) on NO production. L-cysteine pre-treatment stimulated NO production in ECs (Fig. 1B). However, blocking of CSE activity by PPG reversed L-cysteine effect (Fig. 1B). Cystathionine gammalyase knockdown using CSE-specific siRNA significantly reduced CSE protein level and attenuated NO production in comparison with the cells transfected with negative siRNA. Moreover, CSE overexpression significantly elevated CSE expression level and resulted in an increase in NO level ( Fig. 1C and D). Nitric oxide data was further confirmed by DAF-FM fluorescence dye showing that NaHS treatment stimulated NO release in aortic tissues and ECs (Fig. 1E and F). NaHS (50 and 100 lM) treatment markedly increased the phosphorylation of eNOS in ECs (Fig. 2A). The stimulatory effect of NaHS on eNOS phosphorylation was time dependent, and the increase in phosphorylated eNOS appeared at 10 min., peaked at 30 min., and gradually declined to baseline over the period of 1-hr NaHS exposure (Fig. 2B). NaHS treatment up to 36 hrs had no significant effect on eNOS expression level (Fig. 2C).

The role of p38 MAPK/Akt in H 2 S-induced eNOS phosphorylation and NO production
Diverse kinases such as Akt, p38-MAPK kinase and ERK are important for NO production and signalling activation [25][26][27]. To elucidate the signalling pathways involved in H 2 S-induced eNOS phosphorylation and the NO production, we examined the roles of Akt, ERK and p38 MAPK in H 2 S-stimulated NO production. Treatment with NaHS at 100 lM enhanced the phosphorylation of p38 MAPK, Akt and ERK to different levels (Fig. 3). SB202190 (a p38 MAPK inhibitor) and LY294002 (a PI3K/Akt inhibitor), but not U0126 (an inhibitor of ERK), significantly reduced H 2 S-induced phosphorylation of eNOS (Fig. 4).
We further found that the stimulatory effect of NaHS on NO production was decreased by the same treatments (SB202190 or LY294002), and neither SB202190 nor LY294002 alone had any detectable effect on NO production (Fig. 5A). In addition, p38 MAPK inhibition by SB202190 attenuated the NaHS-induced phosphorylation of Akt (Fig. 5B), indicating that p38 MAPK might regulate the upstream signalling cascade that leads to Akt activation. These results suggest that p38 MAPK and Akt are required for NO activation by H 2 S.

The role of NO in H 2 S-induced EC proliferation and angiogenesis
NaHS significantly induced EC proliferation (Fig. 6A). To show the effect of endogenously produced H 2 S, CSE knockdown with a siRNA approach attenuated cell proliferation. The knockdown of CSE significantly attenuated the proliferation of EC by about 25% compared with the control group (Fig. 6B). We also found that CSE knockdown significantly decreased, but NaHS induced a similar and comparable increase, in the proliferation of primarily cultured mouse ECs ( Figure S2). The CSE overexpression stimulated EC proliferation  (Fig. 6B). Next, we study the effect of NO on proliferation. The overexpression of eNOS stimulated cell proliferation, which was strengthened by NaHS treatment (Fig. 6C and D).
We then determined whether H 2 S and NO can interact to regulate angiogenesis. Hydrogen sulfide-induced EC proliferation was attenuated by eNOS knockdown (Fig. 7A and B), whereas treatment with NO precursor L-arginine (1 mM) or NaHS (100 lM) alone significantly increased EC proliferation (Fig. 7B). Furthermore, NaHS (100 lM) treatment significantly increased the capillary-like tube formation of EC compared with the untreated cells (Fig. 7C). NaHSinduced increase in tube formation was significantly attenuated by co-treatment with L-NAME (200 lM), whereas L-NAME treatment alone had no significant effect on tube formation (Fig. 7C). The aortic tissues from CSE-KO mice showed a markedly decreased formation of new microvessels compared with WT mice. After treating the embedded aortic rings with NaHS, the sporting of vascular neogenesis was significantly increased in both CSE-KO and WT mice with markedly higher levels in CSE-KO mice (Fig. 7D). Similar to the effect of NaHS, L-arginine (a NO precursor) stimulated vascular neogenesis in both CSE-KO and WT mice (Fig. 7D). Furthermore, the pro-angiogenic effects of H 2 S on aortic rings from both CSE-KO and WT mice were inhibited by L-NAME treatment (Fig. 7D). Nitro-L-arginine methyl ester treatment inhibited new vessel formation from WT aortic rings, but not that from CSE-KO aortic rings (Fig. 7D). Treatment of EC with NaHS (100 lM) increased the capillary-like tube formation, and co-treatment with a p38 or Akt inhibitor (SB202190 or LY294002) significantly reduced the H 2 S effect (Fig. 7E). However, treatment of EC with LY294002 or SB202190 alone had no significant effect on tube formation (Fig. 7E). LY294002 or SB202190 blocked the proliferation induced by H 2 S, and neither LY294002 nor SB202190 alone had any detectable effect (Fig. 7F), demonstrating that p38 MAPK and Akt are responsible for H 2 S-induced EC proliferation and angiogenesis.

Discussion
Gasotransmitters play important roles in angiogenesis [28][29][30]. Angiogenesis is important for the development of the cardiovascular system and sustaining blood supplies, wound healing and fetus development [31][32][33][34][35]. In our present study, we found that H 2 S can interact with NO to induce angiogenesis of both cloned EC line and freshly isolated primary mouse ECs. The mechanisms for H 2 S action are mainly ascribed to the stimulation of the p38 MAPK/Akt and eNOS phosphorylation, which was followed by increased NO production. Phosphorylation activates eNOS [30]. In our study, the phosphorylation of p38 MAPK precedes the phosphorylation of Akt in the H 2 S signalling cascade, which was confirmed when inhibition of p38 MAPK abolished H 2 S-induced phosphorylation of Akt. We also found that H 2 S activated ERK phosphorylation with a time course similar to that for p38 MAPK activation. However, the inhibition of ERK did not affect H 2 S-stimulated NO production. By altering the phosphorylation of eNOS, H 2 S regulated NO production in ECs. Our observation is consistent with another recent finding by Predmore et al. [36] who demonstrated that Na 2 S (150 lM) treatment stimulated NO production in bovine arterial ECs. While these authors illustrated the H 2 Sdependent Akt mechanism that stimulates NO production, the involvement of other kinases, like p38 MAPK, or the synergistic partnership between H 2 S and NO in angiogenesis were not addressed. Conversely, it has been reported that a high concentration of NaHS (300-3000 lM) significantly inhibited the activity of recombinant bovine eNOS [13]. It is worthy noted here that NaHS at this high concentration range unlikely bears physiological relevance.
We explored the possible interaction between H 2 S and NO in angiogenesis regulation. Ex vivo aortic explants isolated from CSE-KO mice showed a remarkable decrease in vascular neogenesis when compared to WT mice. L-arginine treatment stimulated angiogenesis in the WT mice and to a lesser extent, in the CSE-KO mice. On the other hand, L-NAME treatment reduced new vessel formation in WT mice, and this inhibitory effect was not significant in CSE-KO mice, suggesting that the angiogenic effect of NO might be mediated through H 2 S biosynthesis. Cystathionine gamma-lyase overexpression stimulated EC proliferation, whereas CSE knockdown reversed this effect. Interestingly, we found that the pro-angiogenic effect of H 2 S was partially attenuated in the presence of eNOS inhibitor L-NAME, or after eNOS knockdown using siRNA. Taken together, our results suggest that both gasotransmitters are required for optimal angiogenic activity, yet angiogenesis still proceeds in the presence of either H 2 S or NO alone albeit to a reduced degree. Previous studies had reported that H 2 S and NO can mediate angiogenesis without much knowledge about the H 2 S-NO interaction on angiogenesis [24,37]. Recently, one study reported that a mutually dependent relationship between H 2 S and NO is important for physiological control of different vascular function [38]. Our study used different angiogenesis model (CSE-KO mice versus rat) and experimental conditions, and we found that H 2 S and NO, alone or combined, can cause angiogenesis. Hydrogen sulfide-stimulated angiogenesis was partially but not completely inhibited by NO blockage, whereas in CSE-KO mice, NO treatment stimulated angiogenesis but to a reduced level. The exact molecular mechanism underlying H 2 S-mediated NO pro-angiogenic response is not clear.
In summary, our studies demonstrate that H 2 S promotes NO production in ECs via the activation of a cascade of phosphorylation events, starting from p38 MAPK, Akt to eNOS. Hydrogen sulfide promotes EC tube formation, proliferation and angiogenesis by NOdependent and independent mechanisms as outlined in Figure 8. Thus, H 2 S may be a key regulator for angiogenic signalling pathways, whether they required NO or not. The elucidation of the H 2 S-NO relationship in the vascular biology would improve our understanding of the pathogenic mechanisms for cardiovascular disease in general and angiogenic-related diseases in particular.