Hydrogen-Rich Medium Ameliorates Lipopolysaccharides-Induced Mitochondrial Fission and Dysfunction in Human Umbilical Vein Endothelial Cells (Huvecs) via Up-Regulating HO-1 Expression

Sepsis is dened as life-threatening organ dysfunction caused by a dysregulated host response to infection. It has been showed that the change of mitochondrial dynamics has been proved to be one of the main causes of death in patients with severe sepsis. And hydrogen has been proved to exert its protective effects against sepsis via heme oxygenase-1 (HO-1). This study was designed to demonstrate that whether the benet effects of hydrogen can maintain the dynamic process of mitochondrial fusion/ssion to mitigate human umbilical vein endothelial cells (HUVECs) injury exposed to endotoxin through HO-1.


Introduction
Sepsis is a common complication after burns, trauma and major surgery. It is the main cause of death in acute and critical patients [1]. Sepsis has caused enormous global burden in terms of incidence rate and mortality. Lipopolysaccharide / endotoxin (LPS), which exists in the outer membrane of Gram-negative bacteria, binds to receptors on the surface of endothelial cells, resulting in microcirculation disturbance, septic shock and multiple organ failure [2,3]. Endothelial cell injury has been considered as one of the important pathological features of sepsis [4]. So far, the potential mechanism of LPS induced endothelial cell injury has not made a breakthrough, which is a medical problem to be solved.
Mitochondria is a highly mobile organelle, which frequently occurs division / fusion, biosynthesis and autophagy. The dynamic process of mitochondrial division and fusion is called mitochondrial dynamics [5][6][7][8]. Its balance is a necessary condition for maintaining the normal morphology and function of mitochondria. Mitochondrial dysfunction has been proved to be one of the main causes of death in patients with severe sepsis, and the key cause of mitochondrial dysfunction is the change of mitochondrial dynamics. The disorder of mitochondrial dynamic balance plays an important role in the pathogenesis of sepsis. The prognosis of sepsis is closely related to the early activation of mitochondrial biosynthesis [9].
Heme oxygenase (HO) is recognized as a core component of the mammalian stress response and antioxidant stress defense [10][11][12][13]. As an inducible stress response protein, HO-1 is highly concentrated in tissues closely related to heme protein catabolism, and can degrade heme into carbon monoxide (CO), free iron and biliverdin [14]. The protein is also induced in response to various stimuli, such as reactive oxygen species (ROS) production [15]. In the process of oxidative stress, induction of HO-1 exerts powerful antioxidant, anti-in ammatory and anti-apoptotic properties [16]. Previous studies have shown that the endogenous defense effect of HO-1 pathway on sepsis may be related to mitochondrial dynamics [17]. Mitochondrial ssion is regulated by dynein related protein 1 (Drp1), which is a highly conserved dynamin-related GTPases [18]. However, the overall regulatory mechanism remains unclear.
Molecular hydrogen (H 2 ), the smallest gas molecule with molecular weight, is colorless, odorless, insoluble in water. It has a strong ability to penetrate and enter cells and an extremely high biological safety pro le, which can react with ROS [19]. In recent years, it has been demonstrated that molecular hydrogen, with selective antioxidant, anti-in ammatory and anti-apoptotic properties, has therapeutic effects on a variety of diseases [20][21][22][23][24]. Hydrogen has been reported to ameliorate sepsis induced damage of important organs [25], but the speci c mechanism by which it exerts protective effects remains to be revealed.
Herein, we tested the hypothesis that HO-1 played a critical role in maintaining the dynamic process of mitochondrial fusion/ ssion to mitigate human umbilical vein endothelial cells (HUVECs) injury exposed to endotoxin.

Cell apoptosis
HUVECs were cultured on 10 mm glass dishes. After stimulation with or without LPS and hydrogen-rich medium. The cell death was tested by Dead Cell Apoptosis Kit with FITC annexin V and PI (Invitrogen, USA) and was performed by following the manufacturer's instruction. The cells were concentrated, and washed twice with cold PBS. Basically, collected individual cells were stained with FITC annexin V and PI, then analyzed by confocal microscope (Nikon). Cytomembrane with bright green uorescence was apoptotic cell.

Mitochondria imaging
HUVECs were cultured on 10 mm glass dishes. After stimulation with or without LPS, hydrogen-rich medium and Znpp, cells were incubated with 500 nM MitoTracker Orange (Molecular Probes-Invitrogen, USA) for 5 min, then washed with PBS, and were examined under confocal microscope (Nikon). HUVECs in 3 independent experiments were used for quanti cation of mitochondrial aspect ratio (length/width) using ImageJ-3D Object counter plug-in. Each experiment was done at least four times and each time 16-25 cells per condition were quanti ed. Aspect ratio was de ned as length/width.

Mitochondrial membrane potential
HUVECs were cultured on 10 mm glass dishes. After stimulation with LPS, cells were incubated with 500 nM JC-1 for 20 min (Biyuntian, Nanjng, China), washed with PBS, and were examined under confocal microscope (Nikon). HUVECs in 3 independent experiments were used for quanti cation. Each experiment was done at least four timess.

Drp1 and mitochondria colocalization
HUVEC cells cultured on 10 mm glass dishes were stimulated with LPS for 12h. During the last 5 min of treatment 500 nM MTO was added. Cells were washed with PBS and xed with PBS containing 4% paraformaldehyde and incubated for 10 min in ice-cold. 0.3% Triton X-100 (Sigma-Aldrich, USA) were used for 10 min for permeabilization. Nonspeci c sites were blocked with 5% Goat Serum in PBS for 1 h and then the cells were incubated with anti-Drp1 (1:50) antibody (Cell Signalling, USA). Secondary antibody was anti-rabbit (Molecular Probes-Invitrogen, USA). For the colocalization analysis only one focal plane was analyzed with a Zeiss LSM-5 Pascal 5 Axiovert 200 microscope. Images obtained were deconvolved and background was subtracted using the ImageJ software. Colocalization between the Drp1 and mitochondria was quanti ed using the Manders' algorithm, as detailed elsewhere [28,29].

Statistical analysis
Data are shown as mean ± SD (standard deviation). Data were analyzed by one-way ANOVA and differences among groups was detected using a Dunnett's test. Statistical signi cance was de ned as P < 0.05.

Results
Hydrogen-rich medium restrained LPS-induced cell death in HUVECs Exposure to 10 μg/ml LPS for 12h induced a signi cant increase in Annexin V-FITC-positive cells compared with control cells (P < 0.05), indicating that LPS induced cell apoptosis in HUVECs. However, hydrogen-rich medium attenuated LPS-induced cell apoptosis in HUVECs (P < 0.05) (Fig. 1).

Hydrogen-rich medium attenuated LPS-induced mitochondrial dysfunction in HUVECs
To assess the changes in mitochondrial oxidative phosphorylation, we measured oxygen consumption rates (OCR) at baseline and after maximal uncoupling. Basal OCR was observed no signi cantly change after LPS or hydrogen-rich medium treatment (Fig. 3a-c). Decline in OCR was due to a decreased mitochondrial oxidative capacity, because FCCP-stimulated maximal OCR was signi cantly decreased with LPS treatment, which was improved by hydrogen-rich medium (Fig. 3d-f). Furthermore, LPS treatment decreased mitochondrial memberain potential (Fig. 2) and ATP content ( Fig.   3g-i) while hydrogen-rich medium ameliorated these damages. These last data suggest that hydrogenrich medium attenuates LPS-induced mitochondrial dysfunction in HUVECs.

Hydrogen-rich medium inhibits LPS induced mitochondrial ssion
Treatment of HUVEC with LPS stimulated the appearance of short fragment mitochondria while hydrogen-rich medium enlongated the mean length of mitochondria and decreased its fragmentation after LPS stimulation (Fig. 4). Taken together, these results suggest that LPS stimulates mitochondrial ssion in HUVECs while hydrogen-rich medium inhibits these changes.
Hydrogen-rich medium depressed Drp1 expression level and activity Treatment with LPS increased mitochondrial ssion protein Drp1 expression and hydrogen-rich medium restrained LPS-induced Drp1 over expression (Fig. 5a-d). Mdivi-1 is a selective inhibitor of Drp1. We used Mdivi-1 to understand the role of Drp1 in LPS-treated HUVECs. LPS induced Drp1 over expression were decreased by pretreatment with 25 μM Mdivi-1 and hydrogen-rich medium played a similar role to Mdivi-1 in this experiment (Fig. 5e-f). Previous studies have established the migration of Drp1 from the cytosol to the mitochondrial surface as an initial step in mitochondrial ssion [30]. Therefore, we next evaluated whether mitochondrial ssion triggered by LPS was associated with changes in the subcellular distribution of Drp1. Considered that LPS stimulated Drp1 express level was highest at 12 h (Fig. 5b), we choosed 12 h after LPS stimulation to observe the co-localization of Drp1 and mitochondria. Immuno uorescence studies showed that the Drp1 was recruted to mitochondria after 12 h of treatment with LPS. Both hydrogen-rich medium and Mdivi-1 restrained the recruitment of Drp1. Hydrogen-rich medium and Mdivi-1 used together showed no signi cance compared with used hydrogen-rich medium or Mdivi-1 alone (Fig. 5e-g). Taken together, these data suggest that LPS induced Drp1 over expression and recruitment to mitochondria triggered mitochondrial ssion. Hydrogen-rich medium played a similar role to Mdivi-1, both of them could depressed Drp1 expression and recruitment.
Hydrogen-rich medium attenuates LPS-induced Drp1 activation, mitochondrial dysfunction and mitochondrial ssion by up regulating HO-1 expression HO-1, namely, heat shock protein [31], is reported to impede oxidative cellular injury [32]. To further study the role of HO-1 in LPS-stimulated HUVEC, we detected the expression of HO-1 by western blot. The result showed that HO-1 expression was higher in LPS group compared with sham group and in LPS+H 2 group, HO-1 increased more percent compared with LPS group (Fig. 6a-b). HO-1 expression didn't be in uenced by Mdivi-1 (Fig. 6c-d). Pretreatment of 10 μM Znpp decreased LPS induced up-regulation of HO-1 and reversed the effects of hydrogen-rich medium on HO-1 content (Fig. 6e-f). Znpp also reversed the inhibition effect of hydrogen-rich medium on total Drp1 level (Fig. 6g-h). Based on these data, we got a conclusion that hydrogen-rich medium down-regulate Drp1 by increasing HO-1 expression. To con rm if HO-1 have protective effects on mitochondrial activities, Znpp also had been used to in mitochondrial activity experiment and mitochondrial morphology experiment. The results were consistent with previous, Znpp restricted the protective effects of hydrogen-rich medium on mitochondrial membrane potential ( Fig. 7a-b), ATP content (Fig. 7c), maximal respiration rate (Fig. 7d) and aggravated mitochondrial ssion (Fig. 7e-f). Znpp had no in uence on basal respiration rate.

Discussion
Sepsis is a systemic in ammatory response syndrome induced by infection, which is the leading cause of deaths in intensive care patients, and there is no effective treatment. In sepsis, as the target cells and the effect cells of the in ammatory response, vascular endothelial cells were activated continuousely, and were severe damaged, eventually leading to organs damage and failure [33]. Therefore, vascular endothelial cells as a targets for sepsis are of great signi cance. Based on this, we used LPS to induce an in ammatory model in HUVECs and successfully prepared the in ammatory state of the cells to perform subsequent relevant experiments from the cellular level.
Hydrogen is a small moleculor gas, has antioxidative, anti-in ammatory, and antiapoptotic effects. A large amount of research show that H 2 have bene cial effects to multi-organ injury induced by severe sepsis [20][21][22][23][24]. Consistent with our experiment, after LPS stimulation, the V-FITC positive cells increased signi cantly (P < 0.05), the maximum OCR, mitochondrial membrane potential and ATP content decreased signi cantly (P < 0.05), the mitochondrial fragments increased signi cantly (P < 0.05), and the expressions of Drp1 also increased signi cantly (P < 0.05), while the hydrogen-rich medium can reverse the above changes of various indexes caused by LPS treatment (P < 0.05) at 3 h, 12 h and 24 h. The trend of HO-1 at 12 h was same to that of Drp1.
Mitochondria are dynamic organelles, which undergo continuous ssion and fusion [34]. Fission events are regulated by Drp1 and fusion events are regulated by the large dynamin-related GTPases known as mitofusin 1 and 2 (Mfn1 and 2) as well as optic atrophy1 (OPA1) [35,36]. Alterations in mitochondrial dynamics signi cantly impact mitochondrial numbers and shape, oxygen consumption rate and ATP production [37]. A large amount of research nd that Imbalances of mitochondrial ssion and fusion can induces mitochondrial dysfunction (impaired respiration, mitochondrial memberian potential and ATP production) in cardio vascular disease, diabetes, neuropaphy and other disease [28, 29, 32, 38, 39]. As can also be seen from our experiments, LPS treatment of HUVECs caused changes in intracellular mitochondrial dynamics, resulting in a signi cant decrease in maximal OCR, while also affecting mitochondrial membrane potential and ATP content, demonstrating that LPS induces mitochondrial dysfunction (Fig. 2, 3). Another indicator of altered mitochondrial dynamics, the shape of mitochondria, was also changed after LPS treatment of HUVECs, which stimulated mitochondrial ssion with increased mitochondrial fragmentation (Fig. 4), again demonstrating that LPS causes mitochondrial dysfunction.
Drp1 is one of the most important proteins involved in the regulation of mitochondrial ssion. Drp1 translocates from the cytosol to the mitochondria and interacts with mitochondrial outer membrane proteins (Fis-1, Caf-4, MDV-1 in yeast and Mff in metazoans) [30,40]. After generating a ring around the organelle, the GTP hydrolysis causes their constriction triggering the mitochondrial ssion [40]. Our data support the idea that LPS-induced mitochondrial ssion depends on Drp1 activation. This view was further demonstrated using Mdivi-1. Mdivi-1 is a selective inhibitor of Drp1. LPS induced Drp1 overexpression was decreased by pretreatment of 25 µM Mdivi-1. Previous studies have determined that the migration of Drp1 from cytoplasm to mitochondrial surface is the rst step of mitochondrial division [38]. Therefore, the relationship between LPS induced mitochondrial division and the change of Drp1 subcellular distribution was evaluated through the co-localization of Drp1 and mitochondria. It was found that Drp1 was recruited to mitochondria after LPS treatment, which was inhibited by hydrogen-rich medium and Mdivi-1, However, the simultaneous treatment of hydrogen and Mdivi-1 did not produce superposition effect (Fig. 5). It is proved that LPS induced Drp1 over-expression and mitochondrial recruitment triggered mitochondrial division, which could be inhibited by hydrogen.
Previous studies have shown that the endogenous defense effect of HO-1 pathway on sepsis may be related to mitochondrial dynamics [17]. HO-1 is reported to impede oxidative cellular injury [32].
Consistent with the method of verifying Drp1, Znpp, an inhibitor of HO-1, was used for veri cation. LPS induced increased HO-1 expression, while in LPS + H 2 group, HO-1 expression increased more than that in LPS group (Fig. 6a-b). Pretreatment of 10 µM Znpp decreased LPS induced HO-1 up-regulation and reversed the effect of hydrogen-rich medium on HO-1 content (Fig. 6e-f). ZnPP also reversed the inhibitory effect of hydrogen-rich medium on total Drp1 level ( Fig. 6g-h). It was proved that hydrogen could downregulate Drp1 by increasing HO-1 expression. At the same time, we explored whether HO-1 was protective or harmful to mitochondria. ZnPP limited the protective effect of hydrogen-rich medium on mitochondrial membrane potential, ATP content, maximum respiratory rate and increased mitochondrial division (Fig. 7), which inversely proved the protective effect of HO-1 on mitochondria and the protective effects of H 2 is relevant for HO-1 gene regulation.

Conclusions
In summary, our results showed that the protective effect of H 2 on LPS-treated HUVECs could be involved in the mitochondrial dynamics dependent mechanism mediated by HO-1 and Drp1. Ethics approval and consent to participate Not applicable.

Consent for publication
All listed authors consent to the submission, and all data are used with the consent of the person generating the data.

Availability of data and materials
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.    Effects of LPS and H2 on oxygen consumption rate (OCR). HUVECs were stimulated with or without 10 ug/ml LPS and 0.6 mM hydrogen-rich medium. After 3 h, 12 h and 24 h of stimulation, OCR were tested by seahorse xf24 extracellular ux analyzer. Analysis result of basal respiration rate (a-c), maximal respiration rate (d-f) and ATP content (g-i). Maximal respiration rate and ATP content decreased in LPS group compared with sham group. Hydrogen-rich medium alleviated LPS induced mitochondria respiration dysfunction. Basal respiration rate didn't be in uenced by LPS or hydrogen-rich medium. *P<0.05; **P<0.01 vs. group Sham; #P < 0.05; ##P < 0.01vs. group LPS. Effects of LPS and H2 on mitochondrial morphology. HUVECs were stimulated with or without 10 ug/ml LPS and 0.6 mM hydrogen-rich medium. a After 3 h, 12 h and 24 h of stimulation, cells were incubated with 500 nM mitotracker orange for 5 min and then photographed by confocal scope. b-d Analysis result of aspect ratio (length/width). Exposure to LPS induced mitochondrial fragmentation in HUVECs. *P<0.05; **P<0.01 vs. group Sham; #P < 0.05; ##P < 0.01vs. group LPS.

Figure 5
Exposure of HUVECs to LPS altered the expression of the mitochondrial ssion protein Drp1. a-d LPS stimulated Drp1 expression while hydrogen-rich medium inhibited it. HUVECs were pretreated for 1 h with 25 μM mdivi-1. e-f LPS and hydrogen-rich medium incubated 12 h. Western blot showed hydrogen-rich medium had a similar effect to Mdivi-1, both of them could inhibit Drp1 expression . g Imuno ourescence also been used to con rm the activity of Drp1. Exposure to LPS activated Drp1 and promoted Drp1 recruitment to mitochondria. Hydrogen-rich medium and Mdivi-1 inhibit Drp1 activity. *P<0.05; **P<0.01 vs. group Sham; #P < 0.05; ##P < 0.01vs. group LPS.