Hydrogen sulfide replacement therapy protects the
vascular endothelium in hyperglycemia by preserving
Kunihiro Suzukia, Gabor Olaha, Katalin Modisa, Ciro Colettaa, Gabriella Kulpb, Domokos Geröa, Petra Szoleczkya,
Tuanjie Changc, Zongmin Zhoud, Lingyun Wuc, Rui Wange, Andreas Papapetropoulosa,f, and Csaba Szaboa,1
Departments ofaAnesthesiology andbOphthalmology, University of Texas Medical Branch, Galveston, TX 77555;cDepartment of Pharmacology, University
of Saskatchewan, Saskatoon, SK, Canada S7N 5A2;dG.P. Livanos and M. Simou Laboratories, Department of Critical Care and Pulmonary Services,
Evangelismos Hospital, University of Athens, School of Medicine, Athens 10675, Greece;eDepartment of Biology, Lakehead University, Thunder Bay,
ON, Canada P7B 5E1; andfLaboratory of Molecular Pharmacology, Department of Pharmacy, University of Patras, 265 04 Patras, Greece
Edited by Solomon H. Snyder, Johns Hopkins University School of Medicine, Baltimore, MD, and approved July 7, 2011 (received for review April 5, 2011)
The goal of the present studies was to investigate the role of
changes in hydrogen sulfide (H2S) homeostasis in the pathogenesis
of hyperglycemic endothelial dysfunction. Exposure of bEnd3 mi-
crovascular endothelial cells to elevated extracellular glucose
(in vitro “hyperglycemia”) induced the mitochondrial formation of
reactive oxygen species (ROS), which resulted in an increased con-
sumption of endogenous and exogenous H2S. Replacement of H2S
or overexpression of the H2S-producing enzyme cystathionine-
γ-lyase (CSE) attenuated the hyperglycemia-induced enhancement
of ROS formation, attenuated nuclear DNA injury, reduced the ac-
tivation of the nuclear enzyme poly(ADP-ribose) polymerase, and
improved cellular viability. In vitro hyperglycemia resulted in
a switch from oxidative phosphorylation to glycolysis, an effect
that was partially corrected by H2S supplementation. Exposure of
isolated vascular rings to high glucose in vitro induced an impair-
by CSE overexpression or H2S supplementation. siRNA silencing of
Vascular rings from CSE−/−mice exhibited an accelerated impair-
ment of endothelium-dependent relaxations in response to in vitro
hyperglycemia, compared with wild-type controls. Streptozotocin-
of H2S; replacement of H2S protected from the development of en-
H2S protects against the development of hyperglycemia-induced
endothelial dysfunction. We hypothesize that, in hyperglycemic
endothelial cells, mitochondrial ROS production and increased H2S
tects against these alterations, resulting in reduced ROS formation,
improved endothelial metabolic state, and maintenance of normal
diffusible mediator with multiple roles in the cardiovascular
system in health and disease (1–4). Cystathionine-β-synthase
(CBS), cystathionine-γ-lyase (CSE), and 3-mercaptopyruvate
sulfurtransferase are key enzymes involved in its production of
H2S (3–6). CSE is primarily responsible for most of the H2S pro-
duction in the vasculature (2, 3). The roles of H2S in the cardio-
vascular system include vasodilatation (7, 8) and stimulation of
angiogenesis (9, 10).
cells to produce vasorelaxant mediators such as nitric oxide) plays
a key role in the pathogenesis of various diabetic complications
(11, 12). Overproduction of mitochondrial reactive oxygen species
(ROS) is a principal contributor to the pathogenesis of hypergly-
cemic endothelial dysfunction (11–14). Because H2S production
and degradation is a dynamic process in biological systems, and
ROS can enhance the degradation of H2S (15–17), here we tested
whether hyperglycemia results in a H2S-deficient state and exam-
ined whether H2S replacement affects the development of hyper-
glycemic endothelial dysfunction in vitro and in vivo.
ydrogen sulfide (H2S) is an endogenously produced labile
In Vitro Hyperglycemia Is Associated with Increased H2S Degradation
Caused by Mitochondrial ROS Overproduction. Exposure of endo-
thelial cells to elevated glucose for 7 d resulted in a significant
Incubation of the cells with the ROS scavenger Tempol increased
reduced H2S level in hyperglycemia may be attributable to in-
creased consumption of H2S by ROS. Administration of H2S to
culture medium (without cells) resulted in a decline of ambient
H2S concentrations because of the reaction of H2S with oxygen
and culture media constituents. The consumption of H2S was in-
creased in the presence of normoglycemic cells and was further
accelerated in the presence of hyperglycemic cells (Fig. 1B). Ad-
dition of the mitochondrial uncoupling agent carbonyl cyanide 3-
a slower consumption rate of H2S (Fig. 1B), consistent with the
hypothesis that mitochondrially derived ROS production con-
The increased mitochondrial ROS production in hyperglycemic
cells was demonstrated by the redox-sensitive dye MitoSOX red;
thenoyltrifluoroacetone (TTFA) (Fig. 1C), confirming previous
studies (11–14) showing that mitochondria represent a major
source of ROS.
H2S Replacement Exerts Cytoprotective Effects in Hyperglycemic
Endothelial Cells in Vitro. Addition of H2S (100–300 μΜ) for the
last 4 d of the 7 d of the hyperglycemic period provided a con-
centration-dependent protection against cellular ROS production
(Fig. 2 A and B) and attenuated mitochondrial membrane de-
polarization as measured by the fluoroprobe 5,5′,6,6′-tetrachloro-
1,1′3,3′-tetraethylbenzamidazol-carboncyanine (JC-1) (Fig. 2C).
Exposure of the cells to intermittent high-/low-glucose conditions
is known to exacerbate hyperglycemic endothelial dysfunction
(18–20). Accordingly, intermittent high/low glucose induced a
more pronounced increase in ROS production compared with
a steady elevation of glucose, and H2S continued to attenuate
this response (Fig. 2A). H2S administration also reduced DNA
injury (Fig. S1A) and the activation of the nuclear enzyme poly
(ADP-ribose) polymerase (PARP) (Fig. S1B), which are known
downstream consequences of mitochondrial ROS formation in
Author contributions: K.S., L.W., R.W., A.P., and C.S. designed research; K.S., G.O., K.M.,
C.C., G.K., D.G., P.S., T.C., Z.Z., and C.S. performed research; Z.Z. contributed new reagents/
analytic tools; K.S., K.M., C.C., L.W., R.W., and C.S. analyzed data; and R.W., A.P., and C.S.
wrote the paper.
Conflict of interest statement: C.S. is a stockholder in Ikaria Inc., a for-profit organization
involved in the development of H2S-based therapeutics.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
www.pnas.org/cgi/doi/10.1073/pnas.1105121108PNAS Early Edition
| 1 of 6
hyperglycemic endothelial cells (11–14). Overexpression of CSE
in endothelial cells elevated extracellular H2S levels in normo-
glycemic cells (by 32 ± 5%), but only to a smaller degree in hy-
perglycemic cells (by 10 ± 3%), consistently with the increased
consumption of H2S during hyperglycemia, as demonstrated
above. Overexpression of CSE attenuated the hyperglycemia-in-
duced increase in ROS production (Fig. 3A). Both pharmaco-
logical replacement of H2S and overexpression of CSE protected
against the hyperglycemia-induced decline in cellular viability.
For instance, hyperglycemia decreased cell viability by 18 ± 2%
(P < 0.05), whereas in the endothelial cells overexpressing CSE,
cell viability increased by 6 ± 3%, compared with normoglycemic
controls (n = 4).
Mechanisms of the Cytoprotective Effect of H2S in Hyperglycemic
Endothelial Cells. Analysis of the cellular metabolic status of the
cells showed that high glucose induces a shift away from the mi-
tochondrial oxidativephosphorylation toward glycolysis:thebasal
respiratory capacity and the respiratory reserve capacity response
to carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP)
were reduced in hyperglycemic cells, compared with the normo-
glycemic cells (Fig. 4). Treatment of the cells with H2S resulted
cellular glucose because of enhanced consumption by mitochondrially de-
rived reactive species. (A) Extracellular concentration of H2S in endothelial
cells cultured in low (5.5 mM, LG) or high (40 mM, HG) glucose was de-
termined by the amperometric H2S sensor method at 7 d. H2S levels in high
glucose were significantly suppressed (*P < 0.05 compared with low glucose),
an effect that was attenuated by the antioxidant Tempol (100 μM) (#P <
0.05; n = 5). (B) The rate of consumption of exogenous H2S was measured
by the amperometric H2S sensor. Cells were incubated for 7 d in low and
high glucose, respectively, at which point a single concentration of H2S (300
mM) was added to the medium. There was a significant (*P < 0.05) increase in
the rate of decline in H2S concentrations in cells incubated with high glucose,
an effect that was attenuated (#P < 0.05) by treatment of the hyperglycemic
cells with the uncoupling agent CCCP (0.5 μM) (n = 5). (C) MitoSOX red oxi-
dation was increased in cells exposed to high glucose (*P < 0.05), and this
effect was significantly inhibited by H2S (300 μM) (#P < 0.05) or by the
uncoupling agents TTFA (10 μM) and CCCP (0.5 μM) (#P < 0.05; n = 5).
H2S levels are decreased in endothelial cells placed in high extra-
and protects against mitochondrial depolarization in endothelial cells placed
in high extracellular glucose. Mitochondrial ROS production was measured
in low (5.5 mM, LG) or high (40 mM, HG) glucose conditions at 7 d by using
the MitoSOX red method. A shows the responses to H2S (100–300 μM) in no
glucose, high glucose, and alternating high/low extracellular glucose con-
ditions (H/L). H2S afforded a concentration-dependent and significant sup-
pression of MitoSOX oxidation (#P < 0.05). The Western blot Inset in A shows
that the expression of the H2S-generating enzyme CSE was not suppressed
by high glucose. B shows representative flow cytometric and fluorescent
microscopic images for the four respective groups (low and high glucose
with and without 300 μM H2S). C shows the oxidation of JC-1, a dye used to
detect mitochondrial membrane depolarization. In cells placed in high glu-
cose, there was an increase in mitochondrial depolarization (*P < 0.05),
which was attenuated by H2S (P < 0.05) (n = 5).
Replacement of H2S normalizes mitochondrial oxidant production
2 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105121108Suzuki et al.
in an improvement of mitochondrial respiration, whereas the
hyperglycemia-induced increase in the glycolytic activity of the
the cellular ATP levels in hyperglycemic cells, likely because of
the compensatory effect of increased glycolysis (ATP levels de-
creased to 88 ± 1% in hyperglycemic endothelial cells vs. nor-
moglycemic endothelial cells; n = 3, P < 0.05). H2S prevented the
hyperglycemia-induced suppression of cellular ATP levels (104 ±
2% of control normoglycemic cells, n = 3). Thus, H2S treatment
produced a partial reversal of the hyperglycemia-induced meta-
bolic switch and normalized the energetic status of the cells. As
opposed to the effects of H2S administration for 4 d (see above),
a short period of H2S administration (the last 60 min of the hy-
perglycemic period) failed to affect the mitochondrial ROS
overproduction (Fig. S2). Thus, although high concentrations of
H2S are known to inhibit mitochondrial cytochrome c oxidase,
resulting in a suppression of mitochondrial oxidative phosphory-
lation (21–23), the present effects of H2S in decreasing mito-
chondrial ROS production in hyperglycemic endothelial cells are
not mediated by an acute suppression of mitochondrial function.
The expression of CSE (a principal H2S-producing enzyme in
vascular tissues) was not affected by exposure of the endothelial
cells to elevated glucose for 7 d (Fig. 2A). Furthermore, the
effects of H2S in hyperglycemic endothelial cells did not depend
on the activation of the ATP-sensitive potassium (KATP) channel
because the KATPchannel blocker glibenclamide failed to re-
verse the protective effect of H2S (Fig. S3).
H2S Replacement Therapy Improves Endothelial Function in Hyper-
glycemic Endothelial Cells and in Diabetic Rats. Overexpression of
CSE in rat aortic rings produced a significant protection against
the development of endothelial dysfunction induced by elevated
extracellular glucose (Fig. 5). Pharmacological supplementation of
H2S yielded a similar protective effect (Fig. S4). Streptozotocin-
diabetic rats exhibited a decrease in their blood H2S concen-
trations (Fig. 6A) without any change in the tissue expression of
CSE or CBS (Fig. S5). H2S replacement therapy did not affect
circulating glucose levels in the diabetic animals (Fig. 6B) but
protected against the development of diabetes-induced endothe-
lial dysfunction ex vivo (Fig. 6C).
In endothelial cells where endogenous H2S production was sup-
pronounced degree of ROS production than in the corresponding
control cells (Fig. 3B). Moreover, in thoracic aortic rings from
CSE−/−mice, incubation in extracellular glucose for 24 h caused
a more pronounced impairment of endothelium-dependent relax-
ations than in corresponding rings from wild-type mice (Fig. 7).
H2S against hyperglycemic endothelial dysfunction.
The formation of ROS from endothelial cells is a key factor in
the pathogenesis of diabetic complications (11, 12). In addition,
increased ROS formation and endothelial dysfunction has been
linked to various forms of critical illness, to postoperative
conditions, as well as to impaired glucose tolerance conditions
and postprandial hyperglycemia (11, 12, 24–31). Mitochondrial
electron transport is recognized as a key source of ROS in
hyperglycemic endothelial cells (11, 12). ROS, on their own
and by combining with endothelial nitric oxide to form the
reactive oxidant peroxynitrite, can induce DNA damage and
activation of suicidal pathways governed by the nuclear enzyme
The biosynthesis of H2S, as well as the biological degradation
(consumption) of H2S, is a dynamic process (3, 5, 15). The
current results point to the existence of a crucial interplay be-
tween endothelial H2S formation and ROS production in main-
taining mitochondrial function: elevated glucose perturbs this
balance. Our data demonstrate that the consumption of H2S is
accelerated in endothelial cells placed in elevated glucose, an
effect that depends on mitochondrial ROS formation (because it
can be attenuated by mitochondrial uncoupling). It is conceiv-
able that this accelerated H2S consumption is responsible for
the lower baseline levels of H2S detected in the medium of
cells placed in elevated extracellular glucose and for the de-
creased H2S levels measured in the circulation of streptozotocin-
diabetic rats. On the other hand, down-regulation of CSE does
not occur in hyperglycemia and diabetes under our experimental
conditions and, therefore, is not responsible for the reduced
H2S, as a reducing agent and an antioxidant molecule, has
been previously shown to protect various cell types from oxida-
tive injury (33–36). Based on the current results, we hypothesize
that H2S provides a reducing/antioxidant intracellular environ-
ment that contributes to the maintenance of normal mitochon-
drial function. This balance is perturbed when mitochondrial
ROS production is stimulated by high concentrations of glucose.
We hypothesize that the ROS from hyperglycemic mitochondria
directly reacts with and consumes the intracellular H2S, which
then creates additional mitochondrial dysfunction, possibly by
oxidative modification to mitochondrial proteins. Such a positive
feed-forward cycle may then culminate in a dysfunctional mito-
chondrial state where molecular oxygen is used to produce ROS
(as opposed to ATP) and where mitochondrial efficacy is di-
minished. These events will lead to a loss of mitochondrial
membrane potential and, finally, a spillage of ROS to the cyto-
solic and nuclear compartments.
in endothelial cells placed in high extracellular glucose, whereas silencing of
CSE exacerbates this response. Mitochondrial ROS production was measured
in low (LG) or high (HG) glucose at 7 d by using the MitoSOX red method. In
response to elevated extracellular glucose, an increase in MitoSOX red oxi-
dation was observed (*P < 0.05), which was attenuated when cells were
overexpressing CSE (#P < 0.05; A) and enhanced when CSE was silenced with
siRNA (#P < 0.05; B) (n = 4). The Western blot Insets confirm efficient CSE
overexpression and knockdown, respectively.
Overexpression of CSE attenuates mitochondrial oxidant production
Suzuki et al. PNAS Early Edition
| 3 of 6
Previous studies have demonstrated that antioxidant depletion
is a hallmark of hyperglycemia in endothelial cells (37–39). It has
also been demonstrated previously that endothelial ROS over-
production leads to oxidative and nitrosative protein mod-
ifications, DNA injury, and activation of secondary deleterious
cellular cycles of injury, such as the one governed by the acti-
vation of PARP (11–14, 18–20, 32). The beneficial effects of
antioxidants on the endothelial function in hyperglycemia may
be, at least in part, related to the preservation of the endothelial
The current bioenergetic findings, in agreement with a recent
analysis of bioenergetic alterations in rat retinal endothelial
cells placed in high extracellular glucose (40), demonstrate that
bEnd3 endothelial cells placed in high extracellular glucose ex-
hibit a reduced oxygen consumption rate (i.e., reduced mito-
chondrial oxidative phosphorylation), an effect that is partially
counterbalanced by an up-regulation of glycolysis. Our results
also demonstrate that H2S replacement therapy protects against
this pathophysiological switch between oxidative phosphoryla-
tion and glycolysis. We conclude that restoration of oxidative
phosphorylation, coupled with an improvement of cellular ATP
levels, mitochondrial depolarization, and mitochondrial ROS
production are the key intracellular events through which H2S
replacement is able to restore normal cellular function in hy-
perglycemic endothelial cells and prevent the development of
The results of the current study demonstrate that replacement
of the H2S, either by supplementation into the culture medium
or by overexpression of the H2S-producing enzyme CSE, is able
to protect from the deleterious consequences of hyperglycemia.
On the other hand, siRNA silencing of CSE, or the deletion of
the CSE gene, results in conditions where hyperglycemia induces
an exacerbated endothelial response (more ROS production and
more severe loss of endothelium-dependent relaxant function),
consistent with the hypothesis that endogenous H2S plays a pro-
tective role against the deleterious consequences of hyperglyce-
mia in endothelial cells.
Our in vivo/ex vivo observations, showing that supplementa-
tion of H2S to streptozotocin-diabetic rats improves the endo-
thelium-dependent relaxant function of vascular rings, are
consistent with our in vitro findings in endothelial cells. Also,
previous studies have demonstrated that pharmacological inter-
ventions (e.g., ROS scavenging, peroxynitrite neutralization,
PARP inhibition) that protect hyperglycemic endothelial cells
are also able to improve endothelium-dependent relaxations in
diabetic rodents (11, 12, 29–32), and, therefore, prevention of
the activation of these downstream pathways is likely to con-
tribute to the effects of H2S in the current experimental system.
The translational value of the current findings is enhanced by the
observation that circulating H2S levels are lowered not only in
streptozotocin-diabetic and in NOD mice (a model of type 1
diabetes) but also in patients with diabetes (41–43). We conclude
that hyperglycemia produces a H2S-deficient state in endothelial
cells: H2S replacement therapy in hyperglycemic conditions may
be of therapeutic potential.
Cell Culture. The bEnd3 microvascular endothelial cell line was purchased
from the American Type Culture Collection, cultured at 37 °C at 5% CO2, in
a humidified chamber, with 5.5. mM glucose containing DMEM with 10%
FBS, 2 mM glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, and 1%
placed in low (LG) or high (HG) glucose for 7 d in the absence or presence of H2S treatment (300 μM) was performed by using the Seahorse Biosciences XF24
Analyzer system. In A, a time course for measurement of oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) is shown under basal
conditions, followed by the sequential addition of oligomycin (1 μg/mL), FCCP (0.3 μM), and antimycin A (2 μg/mL). B shows oxygen consumption rate and
mitochondrial reserve capacity, and C shows extracellular acidification rate and glycolytic rate values, representing oxidative phosphorylation and glycolysis,
0.05; n = 15 wells (mean ± SEM) from n = 3 experiments performed on different experimental days].
H2S reduces the degree of the bioenergetic derangements in endothelial cells placed in high extracellular glucose. Bioenergetic analysis of the cells
dysfunction in thoracic aortic rings placed in elevated extracellular glucose.
Rat aortic rings were incubated in low (LG) or high (HG) glucose for 48 h.
High glucose suppressed endothelium-dependent relaxant responses (*P <
0.05), an effect that was attenuated in rings overexpressing CSE (#P < 0.05;
n = 4). (Inset) Representative Western blots for CSE in rings exposed to ad-
enovirus expressing GFP or CSE, respectively.
CSE overexpression protects against the development of endothelial
4 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105121108Suzuki et al.
nonessential amino acids. Cells were plated in six-well plates at 1 × 105cells
per well in a final volume of 2 mL with 5.5 mM DMEM. On the following day,
cells were placed in DMEM containing 5.5 mM (low-glucose control) or 40
mM glucose (high glucose). The culture medium was changed every day.
After 3 d, 100–300 μM NaHS was added to the wells and reapplied every 8 h
until the completion of the experiments. Treatment with 10 μM TTFA, 0.5
μM CCCP, or 10 μM glibenclamide was applied for 3 d. Cells were incubated
at 37 °C at 5% CO2, in a humidified chamber for 7 d, followed by analysis. In
another experimental series, the effect of 100–300 μM NaHS was tested on
ROS production induced by alternating (12-h cycle) high-/low-glucose con-
ditions. To overexpress CSE in endothelial cells, cells were transfected with
an adenoviral plasmid as previously described (44). For CSE silencing, Silencer
Select siRNA for CSE and nonsense control siRNA were obtained from
Ambion and were transfected with Lipofectamine 2000.
Measurement of H2S Levels. Amperometric H2S sensors (WPI) were used for
the real-time measurement of dissolved H2S concentration in the medium
(15). The amperometric H2S sensor was calibrated before each experiment
with freshly prepared (anoxic) NaHS stock solution (0–300 μM).
Measurement of Mitochondrial ROS Production. MitoSOX red (Invitrogen),
a mitochondrion-specific hydroethidine-derivative fluorescent dye, was used
to assess mitochondrial O2−production in situ (45).
Western Blotting. Whole-cell lysate were made by using RIPA buffer with
EDTA with a protein protease inhibitor mixture. Equal amounts of protein
lysate were separated with 8–12% SDS/PAGE gels, transferred to a 0.45-μm
nitrocellulose. The membrane was blocked with 5% low-fat milk in PBS or
Tris-buffered saline containing 0.05% Tween-20 and incubated with the
primary antibody overnight at 4 °C. Primary antibodies for CSE, CBS, and
PARP were from Santa Cruz Biotechnology, and for actin, from Sigma.
Measurement of Membrane Potential by the Fluorescent Dye JC-1. We assessed
mitochondrial membrane potential by using JC-1 as described (39).
Quantitation of DNA Strand Breaks. DNA strand breaks were detected with
a single-cell gel electrophoresis assay (14). DNA strand breaks were quanti-
tated by examining the fixed and stained cells under a fluorescence micro-
scope. The mean length of the DNA tail was determined by measuring 20
cells for each condition.
Cell Viability Assay. The mitochondrial-dependent reduction of 3-(4,5-dime-
thylthiazol-2-yl)-2,5-diphenyltetrazolium bromide was used to measure mito-
chondrial respiration, as an indicator of cell viability (46).
Bioenergetic Analysis. The XF24 Analyzer (Seahorse Biosciences) was used to
measure bioenergetic function in intact bEnd3 cells. The XF24 creates
a transient 7-μL chamber in specialized microplates that allows for oxygen
(A) Streptozotocin-diabetic vehicle-treated rats (STZ/V) exhibit reduced blood
H2S levels (*P < 0.05), an effect that is normalized by supplementation of H2S
using the H2S-releasing minipumps (STZ/S) (#P < 0.05). (B) The hyperglycemic
response is unaffected by H2S-releasing minipumps: *P < 0.05 shows signifi-
cant and comparable degree of hyperglycemia in STZ rats treated with ve-
hicle or H2S-releasing pumps. (C) The thoracic aortas of streptozotocin-
diabetic rats (STZ/V) exhibit reduced endothelium-dependent relaxant func-
tion in response to acetylcholine (1 nM to 30 μM) (*P < 0.05); supplementa-
tion of H2S using the H2S-releasing minipumps (STZ/S) attenuated the degree
of this endothelial dysfunction (#P < 0.05; n = 4–6).
Improvement of endothelial function by H2S in diabetic rats ex vivo.
function in thoracic aortic rings placed in elevated extracellular glucose.
ACh-induced relaxations in aortic rings from wild-type (CSE+/+; A) or CSE−/−
(B) mice placed in low (LG) or high (HG) glucose for 24 h. *P < 0.05 shows
significant inhibition of relaxations in CSE−/−rings. C shows a comparison of
the percentage decrease of the relaxation by high glucose between CSE+/+
and CSE−/−mice; *P < 0.05 shows a higher degree of impairment of the
relaxations in the CSE−/−rings than in wild-type rings (n = 6–12).
CSE deficiency exacerbates the development of endothelial dys-
Suzuki et al. PNAS Early Edition
| 5 of 6
consumption rate and extracellular acidification rate or proton production
rate to be monitored in real time (40, 47). To measure indices of mito-
chondrial function, oligomycin, FCCP, and antimycin A were injected se-
quentially at the final concentrations of 1 μg/mL, 0.3 μM, and 2 μg/mL,
respectively. Using these agents, we determined the basal level of oxygen
consumption, the amount of oxygen consumption linked to ATP production,
the level of non-ATP–linked oxygen consumption (proton leak), the maximal
respiration capacity, and the nonmitochondrial oxygen consumption. Cel-
lular ATP content was measured by a luminescent assay (48).
Vascular Studies of in Vitro Hyperglycemia. Thoracic aortic rings from Sprague-
Dawley rats were incubated for 48 h under normoglycemic or hyperglycemic
conditions in DMEM as described above, in thepresence or absence of200 μM
NaHS, applied every 8 h, followed by the determination of endothelium-
dependent relaxations (29). Adenoviral CSE overexpression in vascular rings
was performed as described for endothelial cells above, followed by in-
cubation of the rings for 48 h under normoglycemic or hyperglycemic con-
ditions in DMEM. Vascular studies from thoracic aortae wild-type and CSE−/−
mice were performed as described (8); rings were incubated either in nor-
moglycemic or hyperglycemic conditions in DMEM for 24 h, followed by the
determination of acetylcholine-induced relaxations.
Vascular Studies in Diabetic Rats. Diabetes in male Sprague-Dawley rats was
induced with a single streptozotocin injection of 60 mg/kg of body wt i.p.
prepared in citrate buffer (pH 4.5). On day 14, animals were implanted with
osmotic pumps (Alzet) filled with NaHS (releasing a dose of 16 μg/kg per min)
or vehicle. Rats were divided into groups as follows: control group (CTL/V,
n = 11, nondiabetic rats treated with vehicle), control with H2S (CTL/S, n = 12;
nondiabetic rats treated with H2S), streptozotocin-induced diabetes group
(STZ/V, n = 7; diabetic rats treated with vehicle), and streptozotocin-induced
diabetes group (STZ/S, n = 9; diabetic rats treated with H2S). Minipumps were
replaced at 2 wk. H2S or vehicle treatment lasted for 28 d. Blood glucose and
blood H2S levels were measured with an Accu-Chek Advantage (Roche) and
the amperometric H2S sensors.
Statistical Analysis. Data are expressed as means ± SEM. Statistical analysis
was performed by ANOVA.
ACKNOWLEDGMENTS. This study was supported by grants from the
National Institutes of Health (R01GM060915), the Shriners Burns Hospitals
(#8661), and the Juvenile Diabetes Foundation (17-2010-542) (all to C.S.),
from the Thorax Foundation (to Z.Z.), and from the Canadian Institutes of
Health Research (to L.W. and R.W.).
1. Fiorucci S, Distrutti E, Cirino G, Wallace JL (2006) The emerging roles of hydrogen
sulfide in the gastrointestinal tract and liver. Gastroenterology 131:259–271.
2. Li L, Moore PK (2008) Putative biological roles of hydrogen sulfide in health and
disease: A breath of not so fresh air? Trends Pharmacol Sci 29:84–90.
3. Szabó C (2007) Hydrogen sulphide and its therapeutic potential. Nat Rev Drug Discov
4. Wang R (2002) Two’s company, three’s a crowd: Can H2S be the third endogenous
gaseous transmitter? FASEB J 16:1792–1798.
5. Szabo C (2010) Gaseotransmitters: New frontiers for translational science. Sci Transl
6. Shibuya N, Mikami Y, Kimura Y, Nagahara N, Kimura H (2009) Vascular endothelium
expresses 3-mercaptopyruvate sulfurtransferase and produces hydrogen sulfide.
J Biochem 146:623–626.
7. Zhao W, Zhang J, Lu Y, Wang R (2001) The vasorelaxant effect of H2S as a novel
endogenous gaseous KATPchannel opener. EMBO J 20:6008–6016.
8. Yang G, et al. (2008) H2S as a physiologic vasorelaxant: Hypertension in mice with
deletion of cystathionine γ-lyase. Science 322:587–590.
9. Papapetropoulos A, et al. (2009) Hydrogen sulfide is an endogenous stimulator of
angiogenesis. Proc Natl Acad Sci USA 106:21972–21977.
10. Szabó C, Papapetropoulos A (December 30, 2010) Hydrogen sulfide and angiogenesis:
Mechanisms and applications. Br J Pharmacol, 10.1111/j.1476-5381.2010.01191.x.
11. Giacco F, Brownlee M (2010) Oxidative stress and diabetic complications. Circ Res 107:
12. Szabo C (2009) Role of nitrosative stress in the pathogenesis of diabetic vascular
dysfunction. Br J Pharmacol 156:713–727.
13. Nishikawa T, et al. (2000) Normalizing mitochondrial superoxide production blocks
three pathways of hyperglycaemic damage. Nature 404:787–790.
14. Du X, et al. (2003) Inhibition of GAPDH activity by poly(ADP-ribose) polymerase ac-
tivates three major pathways of hyperglycemic damage in endothelial cells. J Clin
15. Doeller JE, et al. (2005) Polarographic measurement of hydrogen sulfide production
and consumption by mammalian tissues. Anal Biochem 341:40–51.
16. Geng B, et al. (2004) Endogenous hydrogen sulfide regulation of myocardial injury
induced by isoproterenol. Biochem Biophys Res Commun 318:756–763.
17. Carballal S, et al. (2011) Reactivity of hydrogen sulfide with peroxynitrite and other
oxidants of biological interest. Free Radic Biol Med 50:196–205.
18. Piconi L, et al. (2004) Intermittent high glucose enhances ICAM-1, VCAM-1, E-selectin
and interleukin-6 expression in human umbilical endothelial cells in culture: The role
of poly(ADP-ribose) polymerase. J Thromb Haemost 2:1453–1459.
19. Piconi L, et al. (2006) Constant and intermittent high glucose enhances endothelial
cell apoptosis through mitochondrial superoxide overproduction. Diabetes Metab Res
20. Horváth EM, et al. (2009) Rapid ‘glycaemic swings’ induce nitrosative stress, activate
poly(ADP-ribose) polymerase and impair endothelial function in a rat model of di-
abetes mellitus. Diabetologia 52:952–961.
21. Nicholls P, Kim JK (1982) Sulphide as an inhibitor and electron donor for the cyto-
chrome c oxidase system. Can J Biochem 60:613–623.
22. Nicholson RA, et al. (1998) Inhibition of respiratory and bioenergetic mechanisms by
hydrogen sulfide in mammalian brain. J Toxicol Environ Health A 54:491–507.
anion and cyanide anion in primary rat hepatocyte cultures. Toxicology 188:149–159.
24. Piconi L, Quagliaro L, Ceriello A (2003) Oxidative stress in diabetes. Clin Chem Lab
25. HorváthEM, BenkoR,Gero D, Kiss L, Szabó C (2008) Treatment withinsulin inhibits poly
(ADP-ribose)polymerase activation in a rat model of endotoxemia. Life Sci 82:205–209.
26. Lehr HA, et al. (2006) Consensus meeting on “Relevance of parenteral vitamin C in acute
27. Wallace JP, Johnson B, Padilla J, Mather K (2010) Postprandial lipaemia, oxidative
stress and endothelial function: A review. Int J Clin Pract 64:389–403.
28. Standl E, Schnell O, Ceriello A (2011) Postprandial hyperglycemia and glycemic vari-
ability: Should we care? Diabetes Care 34(Suppl 2):S120–S127.
29. Szabó C, et al. (2002) Part I: Pathogenetic role of peroxynitrite in the development of
diabetes and diabetic vascular complications: Studies with FP15, a novel potent per-
oxynitrite decomposition catalyst. Mol Med 8:571–580.
30. Pacher P, Szabó C (2006) Role of peroxynitrite in the pathogenesis of cardiovascular
complications of diabetes. Curr Opin Pharmacol 6:136–141.
31. Sivitz WI, Yorek MA (2010) Mitochondrial dysfunction in diabetes: From molecular
mechanisms to functional significance and therapeutic opportunities. Antioxid Redox
32. Garcia Soriano F, et al. (2001) Diabetic endothelial dysfunction: The role of poly(ADP-
ribose) polymerase activation. Nat Med 7:108–113.
33. Yin WL, He JQ, Hu B, Jiang ZS, Tang XQ (2009) Hydrogen sulfide inhibits MPP+-
induced apoptosis in PC12 cells. Life Sci 85:269–275.
34. Tyagi N, et al. (2009) H2S protects against methionine-induced oxidative stress in brain
endothelial cells. Antioxid Redox Signal 11:25–33.
35. Tang XQ, et al. (2010) Hydrogen sulfide antagonizes homocysteine-induced neuro-
toxicity in PC12 cells. Neurosci Res 68:241–249.
apoptosis via preservation of mitochondrial function. Mol Pharmacol 75:27–34.
37. Price KD, Price CS, Reynolds RD (2001) Hyperglycemia-induced ascorbic acid deficiency
promotes endothelial dysfunction and the development of atherosclerosis. Athero-
38. Weidig P, McMaster D, Bayraktutan U (2004) High glucose mediates pro-oxidant and an-
tioxidant enzyme activities in coronary endothelial cells.Diabetes Obes Metab 6:432–441.
39. Ungvari Z, et al. (2009) Resveratrol attenuates mitochondrial oxidative stress in cor-
onary arterial endothelial cells. Am J Physiol Heart Circ Physiol 297:H1876–H1881.
40. Trudeau K, Molina AJ, Guo W, Roy S (2010) High glucose disrupts mitochondrial
morphology in retinal endothelial cells: Implications for diabetic retinopathy. Am J
41. Brancaleone V, et al. (2008) Biosynthesis of H2S is impaired in non-obese diabetic
(NOD) mice. Br J Pharmacol 155:673–680.
42. Jain SK, et al. (2010) Low levels of hydrogen sulfide in the blood of diabetes patients
and streptozotocin-treated rats causes vascular inflammation? Antioxid Redox Signal
43. Whiteman M, et al. (2010) Adiposity is a major determinant of plasma levels of the
novel vasodilator hydrogen sulphide. Diabetologia 53:1722–1726.
44. Bucci M, et al. (2010) Hydrogen sulfide is an endogenous inhibitor of phosphodies-
terase activity. Arterioscler Thromb Vasc Biol 30:1998–2004.
45. Mukhopadhyay P, Rajesh M, Yoshihiro K, Haskó G, Pacher P (2007) Simple quantita-
tive detection of mitochondrial superoxide production in live cells. Biochem Biophys
Res Commun 358:203–208.
46. Virág L, Salzman AL, Szabó C (1998) Poly(ADP-ribose) synthetase activation mediates
mitochondrial injury during oxidant-induced cell death. J Immunol 161:3753–3759.
47. Ferrick DA, Neilson A, Beeson C (2008) Advances in measuring cellular bioenergetics
using extracellular flux. Drug Discov Today 13:268–274.
48. Módis K, et al. (2009) Cytoprotective effects of adenosine and inosine in an in vitro
model of acute tubular necrosis. Br J Pharmacol 158:1565–1578.
6 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105121108Suzuki et al.
Suzuki et al. 10.1073/pnas.1105121108
(ROS) production. (A) DNA strand breakage was measured in low (5.5 mM, LG) or high (40 mM, HG) glucose conditions at 7 d by using the Comet assay. High
glucose induced an increase in DNA strand breakage compared with low glucose (*P < 0.05), and H2S (300 μM) afforded a significant suppression of this
response (#P < 0.05). (Inset) Representative images are shown for the four respective groups (low/high glucose with and without 300 μM H2S). (B) Activation of
the nuclear enzyme poly(ADP-ribose) polymerase (PARP) was measured by detection of the poly(ADP ribose) polymers by using Western blotting. High glucose
induced an increase in PARP activation (*P < 0.05), and H2S (300 μM) afforded a suppression of this response (#P < 0.05). (Inset) Representative Western blot is
shown for the four respective groups (low and high glucose with and without 300 μM H2S) (n = 5).
Replacement of hydrogen sulfide (H2S) attenuates cellular responses that lay downstream from hyperglycemic mitochondrial reactive oxygen species
Suzuki et al. www.pnas.org/cgi/content/short/1105121108 1 of 3
high extracellular glucose. Mitochondrial ROS production was measured in low (5.5 mM, LG) or high (40 mM, HG) glucose conditions at 7 d by using the
MitoSOX red method, and H2S (100–300 μM) was administered for 1 h at the end of the experiment. High glucose increased MitoSOX red oxidation (*P < 0.05),
but, when applied according to this protocol, H2S failed to affect this response (n = 4).
Acute administration of H2S at the end of the hyperglycemic period does not affect mitochondrial oxidant production in endothelial cells placed in
extracellular glucose. Mitochondrial ROS production was measured in low (5.5 mM, LG) or high (40 mM, HG) glucose conditions at 7 d by using the MitoSOX
red method in the presence or absence of H2S (300 μM), with and without glibenclamide (10 μM, GLB) pretreatment. Glibenclamide slightly attenuated
hyperglycemia-induced MitoSOX red oxidation but failed to influence the protective effect of H2S on this response. *P < 0.05 indicates significant increases
in MitoSOX red oxidation in high glucose compared with low glucose, and#P < 0.05 indicates significant suppression of this response by H2S (n = 4).
The ATP-sensitive potassium (KATP) channel inhibitor glibenclamide does not prevent the protective effect of H2S in endothelial cells placed in high
were incubated in low (5.5 mM, LG) or high (40 mM, HG) glucose for 72 h. H2S (200 μM) was administered every 8 h. High glucose induced a suppression of
endothelium-dependent relaxant responses (*P < 0.05), an effect that was prevented by H2S (#P < 0.05; n = 4).
H2S protects against the development of diabetic endothelial dysfunction in rat aortic rings placed in elevated extracellular glucose. Rat aortic rings
Suzuki et al. www.pnas.org/cgi/content/short/11051211082 of 3
Fig. S5. Download full-text
tissues. Representative Western blots for CSE and CBS, loading control (actin), and densitometric analysis are shown in healthy control rats and in rats after
6 wk of streptozotocin-induced diabetes. Only CSE was detected in the thoracic aorta; CSE is more abundant than CBS in the heart, whereas only CBS was
detected in the brain (n = 6).
Streptozotocin-induced diabetes for 6 wk does not affect the expression of cystathionine-γ-lyase (CSE) and cystathionine-β-synthase (CBS) in various
Suzuki et al. www.pnas.org/cgi/content/short/11051211083 of 3