Research ArticlePhysiology

H2S production by reactive oxygen species in the carotid body triggers hypertension in a rodent model of sleep apnea

See allHide authors and affiliations

Sci. Signal.  16 Aug 2016:
Vol. 9, Issue 441, pp. ra80
DOI: 10.1126/scisignal.aaf3204

Sleeping with lower blood pressure

Individuals with sleep apnea periodically stop breathing or breathe more shallowly while sleeping. The resulting intermittent decreases in blood oxygen concentrations or hypoxia activates an organ called the carotid body, which sends out signals to increase breathing, but these signals also increase blood pressure and can lead to hypertension. Yuan et al. found that exposure of rodents to bouts of intermittent hypoxia similar to that seen in sleep apnea triggered the production of reactive oxygen species in the carotid body. These reactive oxygen species increased the generation of the gasotransmitter hydrogen sulfide, which stimulates carotid body activity. Inhibiting the enzyme that generates hydrogen sulfide prevented the development of high blood pressure in this rodent model of sleep apnea. Thus, inhibitors of the enzyme that produces hydrogen sulfide could be used to prevent the hypertension associated with sleep apnea.

Abstract

Sleep apnea is a prevalent respiratory disease in which episodic cessation of breathing causes intermittent hypoxia. Patients with sleep apnea and rodents exposed to intermittent hypoxia exhibit hypertension. The carotid body senses changes in blood O2 concentrations, and an enhanced carotid body chemosensory reflex contributes to hypertension in sleep apnea patients. A rodent model of intermittent hypoxia that mimics blood O2 saturation profiles of patients with sleep apnea has shown that increased generation of reactive oxygen species (ROS) in the carotid body enhances the chemosensory reflex and triggers hypertension. CO generated by heme oxygenase-2 (HO-2) induces a signaling pathway that inhibits hydrogen sulfide (H2S) production by cystathionine γ-lyase (CSE), leading to suppression of carotid body activity. We found that ROS inhibited CO generation by HO-2 in the carotid body and liver through a mechanism that required Cys265 in the heme regulatory motif of heterologously expressed HO-2. We showed that ROS induced by intermittent hypoxia inhibited CO production and increased H2S concentrations in the carotid body, which stimulated its neural activity. In rodents, blockade of H2S synthesis by CSE, by either pharmacologic or genetic approaches, inhibited carotid body activation and hypertension induced by intermittent hypoxia. Thus, our results indicate that oxidant-induced inactivation of HO-2, which leads to increased CSE-dependent H2S production in the carotid body, is a critical trigger of hypertension in rodents exposed to intermittent hypoxia.

INTRODUCTION

Sleep apnea is a widespread respiratory disease characterized by brief (15 to 30 s), repeated episodes of cessation of breathing during sleep, which occur because of defective generation of the respiratory rhythm (central sleep apnea) or because of obstruction of the upper airway (obstructive sleep apnea) (13). Repetitive apneas lead to periodic decreases in blood O2 concentrations, resulting in intermittent hypoxia. Sleep apnea patients and rodents exposed to intermittent hypoxia, which simulates the altered blood O2 saturation profiles caused by sleep apnea, exhibit increased sympathetic nerve activity (SNA) and hypertension (4, 5). Studies in humans and rodents suggest that the carotid body, the primary sensory organ for detecting hypoxia in arterial blood, mediates a reflex increase in SNA that results in hypertension (6, 7). Intermittent hypoxia sensitizes the carotid body response to hypoxia (8, 9) and induces sensory long-term facilitation (LTF), which is characterized by a long-lasting increase in sensory nerve activity (10). Sleep apnea patients exhibit marked increases in blood pressure during apneic episodes because of intermittent hypoxia–evoked carotid body hypersensitivity to hypoxia, whereas sensory LTF mediates reflex activation of the sympathetic nervous system and persistent daytime hypertension (6). Ablation of the carotid body blocks sympathetic activation and hypertension in intermittent hypoxia–subjected rats (11, 12).

Intermittent hypoxia increases the generation of reactive oxygen species (ROS) in the carotid body due to increased expression of genes encoding prooxidant enzymes that are regulated by hypoxia-inducible factor 1 (HIF-1) and decreased expression of antioxidant enzyme–encoding genes that are regulated by HIF-2 (1315). Treatment of rats with a ROS scavenger during exposure to intermittent hypoxia normalizes carotid body activity and prevents the development of hypertension (10, 16, 17). Thus, ROS generation in response to intermittent hypoxia triggers carotid body activation and the ensuing cardiovascular pathology. However, the mechanisms by which ROS activates the carotid body by intermittent hypoxia have not been fully delineated. Because intermittent hypoxia does not alter carotid body morphology (10), we hypothesized that the augmented carotid body activity is due to effects of ROS on hypoxia signal transduction. Various theories have been proposed to explain hypoxia sensing by the carotid body (18). Emerging evidence suggests that the gaseous messenger hydrogen sulfide (H2S) mediates carotid body activation by hypoxia (1923). Glomus cells, the primary O2-sensing cells in the carotid body, express at least two major H2S-synthesizing enzymes: cystathionine β-synthase (CBS) and cystathionine γ-lyase (CSE) (19, 24). CBS is itself induced by chronic hypoxia through transcriptional activation by HIFs (25). However, Cse−/− mice display greatly diminished H2S generation and a severely impaired carotid body and physiological responses to acute hypoxia (19), suggesting that CSE-derived H2S is more vital for regulating carotid body activity. We hypothesized that ROS increase the CSE-dependent synthesis of H2S to trigger carotid body activation and hypertension in rodents exposed to intermittent hypoxia.

RESULTS

Intermittent hypoxia increases CSE-dependent H2S production through ROS

H2S concentrations were measured in the carotid bodies of adult rats exposed to alternating cycles of 15 s of hypoxia (5% O2) followed by 5 min of room air (21% O2), nine episodes per hour, and 8 hours per day for 10 days. This intermittent hypoxia paradigm decreases arterial blood–O2 saturation from ~97% to ∼80% with each episode of hypoxia (12), which is comparable to the O2 desaturation observed in sleep apnea patients (2). H2S concentrations were increased in carotid bodies from rats exposed to intermittent hypoxia as compared with controls (Fig. 1A). To assess the contribution of CSE, an H2S-synthesizing enzyme in the carotid body (19), to the observed increase in H2S, we treated rats with l-propargylglycine (l-PAG), a selective inhibitor of CSE (26). H2S concentrations were not increased by intermittent hypoxia in carotid bodies from rats treated with l-PAG (Fig. 1A). In addition, intermittent hypoxia increased H2S concentrations in the carotid bodies of wild-type mice, a response that was absent in CSE knockout mice (Fig. 1B). Thus, pharmacologic and genetic approaches in rats and mice, respectively, demonstrate that intermittent hypoxia increases CSE-dependent H2S production.

Fig. 1 Induction of CSE-dependent H2S production by intermittent hypoxia.

(A) H2S concentrations in the carotid bodies of rats exposed to room air (C), room air with l-PAG (C + PAG), intermittent hypoxia (IH), or IH with l-PAG (IH + PAG) (n = 3 experiments for each treatment; a total of 12 carotid bodies from six rats for each treatment). (B) H2S concentrations in wild-type (WT) and CSE-null (Cse−/−) mice exposed to room air (C) or IH (n = 3 experiments for each treatment per genotype; a total of 24 carotid bodies from 12 mice for each treatment per genotype). (C and D) MDA (C) and H2S amounts (D) in the carotid bodies of rats exposed to room air (C), room air with MnTMPyP [a ROS scavenger (RS)] (C + RS), IH, or IH with RS (IH + RS) (n = 3 experiments for each treatment; a total of 12 carotid bodies from six rats for each treatment). (E) RT-qPCR assay of Cse mRNA normalized to 18S ribosomal RNA (rRNA) in carotid bodies of rats exposed to either room air (C) or IH, normalized to C (n = 3 experiments for each treatment; a total of six carotid bodies from three rats for each treatment). (F) Representative immunoblot of CSE protein abundance in liver tissue of rats exposed to room air (C) or IH (n = 3 experiments for each treatment). (G) Representative immunoblot of CSE expression in HEK-293 cells stably transfected with empty vector or vector encoding CSE (n = 3 experiments). α-Tubulin abundance was assayed in (F) and (G) as a loading control. (H and I) MDA (H) and H2S amounts (I) in CSE-expressing HEK-293 cells exposed to room air (C), room air with RS (C + RS), or IH (n = 4 experiments for each treatment). Data are means ± SEM. **P < 0.01; ***P < 0.001; ns, not significant. See fig. S1 for data on concentrations of H2S and MDA and CSE protein abundance in rat liver.

Malondialdehyde (MDA) amounts, which reflect oxidized lipids (27), were measured in the carotid bodies as an index of ROS generation. MDA amounts were increased in carotid bodies from rats exposed to intermittent hypoxia as compared with controls, an effect that was abrogated by treatment with manganese (III) tetrakis(1-methyl-4-pyridyl)porphyrin pentachloride (MnTMPyP), a membrane-permeable ROS scavenger (28) during the 10 days of intermittent hypoxia exposure (Fig. 1C). MnTMPyP treatment also blocked the intermittent hypoxia–induced increase in H2S concentrations in the rat carotid body (Fig. 1D). However, MnTMPyP treatment had no effect either on MDA amounts or on H2S concentrations of carotid bodies from control rats (Fig. 1, C and D). These findings indicate that ROS generated during intermittent hypoxia stimulates H2S production in the carotid body.

Reverse transcription and quantitative real-time polymerase chain reaction (RT-qPCR) assays revealed no significant change in Cse mRNA abundance in carotid bodies from rats exposed to intermittent hypoxia (Fig. 1E). Because the nature of the carotid body precludes immunoblotting for protein abundance (the wet weight of the rat carotid body is ~50 μg), we determined the effects of intermittent hypoxia on CSE protein abundance by assessing rat liver tissue. We believe liver tissue to be a reasonable substitute, because livers of rats exposed to intermittent hypoxia also exhibited increases in the amounts of H2S and MDA, which were blocked by MnTMPyP treatment (fig. S1, A and B). In the liver, CSE protein abundance was similar between intermittent hypoxia–exposed and control rats (Fig. 1F and fig. S1C). Together, these findings suggest that changes in CSE abundance do not account for the ROS-dependent increase in H2S in rats exposed to intermittent hypoxia.

We next investigated whether ROS generated by intermittent hypoxia directly affected CSE enzymatic activity in human embryonic kidney (HEK)–293 cells heterologously expressing CSE (Fig. 1G). Cells were exposed to 60 cycles of intermittent hypoxia, with each cycle consisting of 30 s of hypoxia (1.5% O2) followed by 5 min of room air (20% O2), as described previously (13). This paradigm of in vitro intermittent hypoxia induces cellular responses similar to those evoked by in vivo intermittent hypoxia (10, 13). Although intermittent hypoxia increased ROS, as evidenced by increased MDA amounts, H2S concentrations remained unchanged (Fig. 1, H and I). H2S concentrations were unaltered by treating control CSE-expressing cells with the ROS scavenger MnTMPyP (Fig. 1I). These observations rule out a direct effect of intermittent hypoxia–evoked ROS on CSE catalytic activity.

Intermittent hypoxia–induced ROS inhibits CO production to increase H2S

In the carotid body, heme oxygenase-2 (HO-2) synthesizes the signaling molecule carbon monoxide (CO) (29). CO inhibits H2S signaling by stimulating protein kinase G (PKG) to phosphorylate and inactivate CSE (21). We hypothesized that intermittent hypoxia–induced ROS may inhibit CO synthesis, thereby decreasing the inhibitory actions of PKG on CSE and augmenting H2S production from CSE. To test this possibility, we measured CO concentrations in carotid bodies and liver tissue from rats exposed to intermittent hypoxia, and we analyzed PKG activity and CSE phosphorylation only in liver tissue because of the limited availability of carotid body tissue. CO concentrations were decreased in the carotid bodies and liver tissue of rats exposed to intermittent hypoxia (Fig. 2, A and B), and these effects were associated with reduced PKG activity and CSE phosphorylation (Fig. 2, C and D). All of these responses to intermittent hypoxia were abolished by administration of the ROS scavenger MnTMPyP (Fig. 2, A to D). However, MnTMPyP had no effect on basal CO concentration, PKG activity, and CSE phosphorylation of carotid bodies and liver tissue from control rats (Fig. 2, A to D). Neither HO-2 mRNA nor HO-2 protein abundances were altered in the carotid bodies and liver tissue, respectively, of rats exposed to intermittent hypoxia (fig. S2, A and B). These observations indicate that ROS generated during intermittent hypoxia inhibits HO-2–CO production without affecting HO-2 abundance, ultimately leading to an increase in H2S production by CSE.

Fig. 2 Intermittent hypoxia increases H2S concentrations through ROS-dependent inhibition of CO production.

(A) CO concentrations in the carotid bodies (CB) of rats exposed to room air (C), room air with MnTMPyP [a ROS scavenger (RS)] (C + RS), intermittent hypoxia (IH), or IH with RS (IH + RS) (n = 3 experiments for each treatment; a total of 12 carotid bodies from six rats for each treatment). (B to D) CO concentrations (B), PKG activity (C), and serine phosphorylation (P) of CSE (D) in liver samples from rats exposed to room air (C), room air with RS (C + RS), IH, or IH with RS (IH + RS) (n = 4 experiments for each treatment). (E to I) Concentrations of MDA (E), CO (F), and H2S (G); activity of PKG (H); and serine phosphorylation of CSE (I) in HO-2– and CSE-expressing HEK-293 cells exposed to room air (C), room air with RS (C + RS), IH, or IH with RS (IH + RS) (n = 5 experiments for each treatment). In (D) and (I), representative immunoblots (upper panel) and densitometry analysis (lower panel) are shown. I.P., immunoprecipitation. Data are means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. See fig. S2 for data on HO-2 mRNA in the rat carotid body and HO-2 protein abundance in the rat liver.

Intermittent hypoxia reduces CO production through direct inhibition of HO-2 enzyme activity by ROS

We next investigated whether intermittent hypoxia reduces CO production by direct inactivation of HO-2 through ROS. Because the limited mass of carotid body precludes rigorous biochemical analysis, we tested whether intermittent hypoxia inhibited CO production in HEK-293 cells heterologously expressing both HO-2 and CSE (21) (fig. S3A) similarly to that in the carotid body. Intermittent hypoxia increased ROS as monitored by MDA amounts (Fig. 2E), decreased CO concentrations (Fig. 2F), enhanced H2S concentrations (Fig. 2G), and diminished both PKG activity (Fig. 2H) and CSE phosphorylation (Fig. 2I). The ROS scavenger MnTMPyP had no effects in control cells, but blocked the increase in MDA amounts and abolished intermittent hypoxia–induced changes in the amounts of CO and H2S, PKG activity, and CSE phosphorylation (Fig. 2, E to I) without affecting the protein abundance of either HO-2 or CSE (fig. S3A). These findings suggest that the effects of intermittent hypoxia on carotid bodies can be at least partially recapitulated in HEK-293 cells.

To assess the mechanism by which ROS generated in response to intermittent hypoxia inhibit CO production, we analyzed the HO-2 enzyme activity by varying the concentration of one of its substrates, hemin. In HEK-293 cells coexpressing HO-2 and CSE, the apparent binding affinity (Km) of HO-2 for hemin was not altered by exposure to intermittent hypoxia (fig. S3B). However, intermittent hypoxia reduced maximal CO synthesis (Vmax), an effect prevented by the ROS scavenger MnTMPyP (Fig. 3, A and B). The C-terminal region of HO-2 contains three heme regulatory motifs that encompass cysteine residues (Cys127, Cys265, and Cys282), respectively (Fig. 3C, upper panel) (30). Oxidation and reduction of cysteine thiols are a major mechanism for redox regulation of proteins and enzymes (30). We reasoned that one (or more) of these cysteines is required for ROS-dependent inhibition of CO production by HO-2. Substituting alanine for Cys265 (C265A), but not Cys127 (C127A) or Cys282 (C282A), abolished the decrease in Vmax and CO generation in response to intermittent hypoxia (Fig. 3, C and D, and fig. S3C). Cells coexpressing CSE and HO-2 (C265A) still exhibited an increase in ROS during intermittent hypoxia, but no longer responded with changes in PKG activity, CSE phosphorylation, or H2S concentrations (Fig. 3, E to H). Unlike HO-2, the inducible CO-synthesizing enzyme HO-1 does not have cysteine-containing heme regulatory motifs (30). Cells expressing HO-1 exhibited increased ROS production during intermittent hypoxia without changes in CO concentration (fig. S4, A and B).

Fig. 3 Intermittent hypoxia alters the kinetic properties of HO-2.

(A and B) CO production as a function of hemin concentration (A) and maximal reaction rate of CO production (Vmax) (B) in HO-2– and CSE-expressing HEK-293 cells exposed to room air (C), room air with ROS scavenger (RS) (C + RS), intermittent hypoxia (IH), or IH with RS (IH + RS) (n = 4 experiments for each treatment). (C and D) Effects of mutation of cysteine residues in HO-2 on Vmax (C) and CO generation (D) in HEK-293 cells exposed to room air (C) or IH. Upper panel in (C) shows schematic location of cysteine residues in heme regulatory motifs (HRMs) of HO-2. MSR, membrane-spanning region. n = 5 experiments for each treatment. (E to H) MDA amounts (E), PKG activity (F), serine phosphorylation (P) of CSE (G), and H2S concentrations (H) in CSE- and HO-2 (C265A)–expressing HEK-293 cells exposed to room air (C), room air with RS (C + RS), IH, or IH with RS (IH + RS) (n = 4 experiments for each treatment). In (G), representative immunoblots (upper panel) and densitometry analysis (lower panel) are shown. (I) CO production in WT HO-2–expressing HEK-293 cells exposed to room air (C), room air with PEG-catalase (C + CAT), IH, and IH with CAT (IH + CAT) (n = 4 experiments for each treatment). (J) Effect of hydrogen peroxide (H2O2) on CO production in HEK-293 cells expressing either WT HO-2 or C265A mutant HO-2 (n = 5 experiments in each group). Data are means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. See fig. S3 for data on CSE and HO-2 protein abundance and hemin binding affinity (Km) for HO-2. See fig. S4 for data on concentrations of MDA and CO in HEK-293 cells expressing HO-1.

We next sought to identify the particular species of ROS that mediated the effects of intermittent hypoxia on HO-2. H2O2, derived from superoxide anion (O2), has been implicated in mediating, in part, the effects of intermittent hypoxia on the carotid body (31). Polyethylene glycol (PEG)–catalase, a scavenger of H2O2, had no effect on CO production in control cells expressing HO-2 and CSE, but rescued inhibition of CO production by intermittent hypoxia (Fig. 3I). Moreover, exogenous H2O2, similar to intermittent hypoxia, inhibited CO production in cells expressing wild-type HO-2 but not in cells expressing the intermittent hypoxia–insensitive C265A mutant HO-2 (Fig. 3J).

H2S mediates carotid body activation by intermittent hypoxia

We next investigated whether increased H2S production contributes to carotid body activation during intermittent hypoxia. Carotid bodies were harvested and sensory nerve activity was recorded ex vivo to exclude the confounding influence of cardiovascular changes caused by intermittent hypoxia in intact animals. Action potentials of the same height, duration, and shape (that is, single unit) were selected for analysis (fig. S5, A to H). To simulate apneic episodes, we challenged carotid bodies ex vivo with 10 episodes of alternating cycles of 30 s of hypoxia (5% O2) and 5 min of room air (20% O2). Carotid bodies isolated from rats exposed to intermittent hypoxia exhibited higher baseline sensory nerve activity, augmented responses to each episode of hypoxia, and a progressive increase in baseline activity after the termination of repetitive hypoxia (that is, sensory LTF). All of these effects were lost in rats treated with l-PAG during the last 2 days of a 10-day exposure to intermittent hypoxia (Fig. 4, A to D). Carotid bodies from wild-type mice, but not those from Cse−/− mice, showed similar responses to intermittent hypoxia (Fig. 4, E to H). Similar to wild-type mice exposed to intermittent hypoxia, carotid bodies of HO-2−/− mice exhibited higher baseline sensory nerve activity, augmented responses to each episode of hypoxia, and sensory LTF, and the 10-day exposure to intermittent hypoxia had no further effect on these responses (fig. S6, A to D).

Fig. 4 Increased H2S production mediates activation of the carotid body by intermittent hypoxia.

(A) Examples of ex vivo carotid body sensory nerve responses during ten 30-s episodes of acute intermittent hypoxia (AIH; at arrows) and after AIH for 60 min [sensory LTF (sLTF)] from rats exposed toroom air (C), room air with l-PAG (C + l-PAG), IH, or IH with l-PAG (IH + l-PAG). (B to D) Average data of baseline activity before AIH (B), hypoxia (Hx) response (mean of 10 episodes of AIH) (C), and sLTF (averaged over 60 min after AIH) (D) (n = 8 experiments for each treatment; 16 single units from eight carotid bodies for each treatment). (E) Examples of ex vivo carotid body sensory nerve responses during five 30-s episodes of AIH (at arrows) and after AIH for 60 min (sLTF) from WT mice exposed to room air (WT-C) or IH (WT-IH) and Cse−/− mice exposed to room air (Cse−/−-C) or IH (Cse−/−-IH). (F to H) Average data of baseline activity before AIH (F), hypoxia (Hx) response (mean of five episodes of AIH) (G), and sLTF (averaged over 60 min after AIH) (H) (n = 8 experiments for each treatment per genotype; 16 single units from eight carotid bodies for each treatment per genotype. Data are means ± SEM. *P < 0.05; ***P < 0.001; ns, not significant; imp/s, impulses per second. See fig. S5 for analysis of carotid body single-unit sensory nerve activity.

Inhibiting H2S synthesis prevents intermittent hypoxia–induced sympathetic activation and hypertension

The chemosensory reflex arising from the carotid body plays a critical role in mediating sympathetic activation and hypertension in rodents exposed to intermittent hypoxia (6). We tested whether blocking H2S synthesis prevents intermittent hypoxia–induced sympathetic activation and hypertension. Blood pressure and splanchnic SNA were recorded in anesthetized rats. Rats exposed to intermittent hypoxia showed increased basal blood pressure and SNA (Fig. 5, A to C). To simulate apnea, we challenged rats with brief episodes of hypoxia (12% inspired O2 for 30 s). Brief episodes of hypoxia markedly increased blood pressure and SNA, responses that were nearly absent in control rats (Fig. 5, A, D, and E). l-PAG treatment during the last 2 days of a 10-day exposure to intermittent hypoxia normalized basal blood pressure and SNA, and also prevented the increases in blood pressure and SNA caused by brief hypoxic challenges, without affecting baseline blood pressure and SNA in control rats (Fig. 5, A to E).

Fig. 5 H2S mediates chemosensory reflex–dependent sympathetic activation and hypertension in rodents exposed to intermittent hypoxia.

(A) Examples of arterial blood pressure (BP; upper panels) and splanchnic SNA (middle and lower panels) responses to acute hypoxia (12% O2 for 30 s, indicated by black bars) of rats exposed to room air (C), room air with l-PAG (C + l-PAG), IH, or IH with l-PAG (IH + l-PAG). Arrows denote baseline BP (upper panel) and SNA (lower panel). Lower panels show SNA on an expanded time scale of 200 ms as compared to 2 min in middle panels to demonstrate action potentials of the splanchnic nerve under conditions of normoxia (Nx) and hypoxia (Hx). (B to E) Average data of baseline BP (B), baseline SNA (C), Hx-induced change in BP (D), and SNA (E) of rats exposed to room air (C), room air with l-PAG (C + l-PAG), IH, or IH with l-PAG (IH + l-PAG) (n = 8 rats for each treatment). (F and G) Baseline BP (F) and plasma epinephrine (E) and norepinephrine (NE) concentrations (G) of WT and CSE-null (Cse−/−) mice exposed to either room air (C) or IH (n = 7 to 9 mice for each treatment per genotype). Data are means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

To further investigate the role of CSE-derived H2S in intermittent hypoxia–induced hypertension, we compared the effects of intermittent hypoxia between 4-week-old wild-type and CSE-null mice. Blood pressure was monitored by the tail-cuff method before and after intermittent hypoxia, and plasma catecholamine (epinephrine and norepinephrine) concentrations were measured as an index of sympathetic activation. Wild-type mice exhibited increased blood pressure and plasma catecholamines in response to intermittent hypoxia exposure, responses that were absent in CSE-null mice (Fig. 5, F and G).

DISCUSSION

Sleep apnea, a clinically prevalent disorder, elicits intermittent hypoxia, which often leads to hypertension. Considerable evidence suggests that a hyperactive carotid body and the ensuing chemosensory reflex can drive this hypertension through increased SNA. Our results establish a mechanism that contributes to carotid body activation and the resulting cardiovascular pathology in a rodent model of sleep apnea. Specifically, we find that ROS generated during intermittent hypoxia in the carotid body depresses CO signaling, ultimately increasing stimulatory H2S signaling (Fig. 6).

Fig. 6 Intermittent hypoxia activates H2S signaling in the carotid body.

Intermittent hypoxia (IH) induces ROS-dependent inhibition of CO production by HO-2, which leads to decreased PKG activity and increased H2S production by CSE that triggers the carotid body chemosensory reflex, leading to sympathetic activation and hypertension.

Challenging rodents with intermittent hypoxia to simulate the blood O2 saturation profiles encountered during sleep apnea increased CSE-dependent H2S production in the carotid body. Scavenging ROS was sufficient to abrogate the increase in H2S concentrations in the carotid body in response to intermittent hypoxia, demonstrating the epistatic importance of ROS. We observed that H2S signaling was increased during intermittent hypoxia due to a decrease in inhibitory CO signaling and not due to changes in CSE abundance. Intermittent hypoxia reduced the rate of CO production through the direct inhibition of HO-2 activity by ROS. Attenuated CO production, in turn, decreased the inhibitory phosphorylation of CSE by PKG, resulting in increased H2S synthesis.

Analysis of the molecular mechanism underlying HO-2 inactivation by intermittent hypoxia revealed that ROS reduced the maximal rate of CO production (Vmax) from the catalysis of hemin degradation without altering the binding of hemin to HO-2. Furthermore, we identified Cys265 in the heme regulatory motif as a critical redox-sensing residue. A role for Cys265 was further supported by the finding that ROS produced during intermittent hypoxia had no inhibitory effect on CO production in cells expressing HO-1, which lacks cysteine containing heme regulatory motif (30). Our results further suggest that H2O2 is one of the species of ROS that mediate HO-2 inactivation during intermittent hypoxia. Because of technical limitations, these analyses were performed in HEK-293 cells heterologously expressing HO-2 and CSE. However, similar signaling is likely to occur in the carotid body because the effects of intermittent hypoxia on ROS, CO, and H2S concentrations were similar between HEK-293 cells and the carotid body.

Acute hypoxia also activates H2S generation through decreased CO production through HO-2 and reduced serine phosphorylation of CSE by PKG (21). However, the reduced CO production under conditions of acute hypoxia is primarily due to the low affinity of HO-2 for O2, rather than an effect on Vmax, which enables the enzyme to transduce changes in O2 availability into changes in CO production (21). Moreover, the effects of acute hypoxia are transient and reversed after a few minutes of reoxygenation (21). In contrast, the effects of intermittent hypoxia on CO production, serine phosphorylation of CSE, and H2S production persist for several hours after terminating the intermittent hypoxia exposure. Our results suggest that the long-lasting effects of intermittent hypoxia on CO-regulated H2S production are due to increased ROS generation during intermittent hypoxia, which is further accentuated by reciprocal changes in the expression of anti- and prooxidant enzymes (14, 31).

A major finding of the present study is that blockade of H2S production is sufficient to prevent carotid body activation and hypertension in intermittent hypoxia–exposed rodents. Although exogenous and endothelial H2S triggers local blood vessel relaxation (3234), we find that H2S exerts excitatory effects on carotid body activity during intermittent hypoxia. Consistent with previous reports (810), carotid bodies from rats exposed to intermittent hypoxia displayed exaggerated hypoxic sensitivity and a persistent increase in baseline activity after repetitive hypoxia (sensory LTF). The enhanced carotid body activity in rats exposed to intermittent hypoxia was reflected in the heightened chemosensory reflex, as evidenced by increased baseline sympathetic activity and blood pressure, as well as marked hypertension and sympathetic activation during simulated apneas. Treating rats with a CSE inhibitor during the last 2 days of intermittent hypoxia exposure restored normal carotid body function, SNA, and blood pressure, and blocked hypertensive responses to simulated apneas. It is important to note that CSE is also present in the brainstem neuronal populations important for regulating blood pressure (35). Whereas our results illustrate that Cse−/− carotid bodies exhibit a tissue-autonomous reduction in activity, the full blood pressure responses of Cse−/− mice may be due to altered signaling from the brainstem regions and the carotid body.

Antioxidant treatment also normalizes carotid body function (10) and prevents hypertension (16), demonstrating that ROS is epistatic to H2S in driving hypertension. However, antioxidants must be administered during the entire 10 days of intermittent hypoxia exposure, and they are ineffective in normalizing carotid body function when given during the last day of intermittent hypoxia (10). Previous studies have implicated endothelin-1 (ET-1) as the cause of the augmented hypoxic sensitivity (3638) and serotonin [5-hydroxytryptamine (5-HT)] in inducing sensory LTF (31) in the carotid bodies of rodents exposed to intermittent hypoxia. Because ET-1 and 5-HT also increase ROS (39, 40), the effects of these transmitters on the carotid body could be mediated through ROS-dependent activation of H2S production. Such a mechanism might explain why blocking H2S production by administration of a CSE inhibitor was sufficient to prevent augmented hypoxic sensitivity and sensory LTF as well as the ensuing chemosensory reflex–mediated hypertension. We have discovered that HO-2−/− mice up-regulate another gaseous messenger, nitric oxide (NO), and its synthesizing enzyme, neuronal NO synthase (nNOS), to maintain oxygen sensing (21). However, carotid bodies of HO-2−/− mice exhibited a similar phenotype as wild-type mice to 10 days of intermittent hypoxia exposure, including increased baseline sensory nerve activity, augmented response to acute hypoxia, and sensory LTF. Given that 10 days of intermittent hypoxia had no further effect on the carotid body sensory activity in HO-2−/− mice, NO signaling from nNOS is unlikely to play a major role in mediating sensory LTF and hypertension, a possibility that requires further investigation.

Population-based studies have shown a strong correlation between the severity of sleep apnea and hypertension (1, 41). At least one-third of patients with obstructive sleep apnea exhibit a heightened chemosensory reflex that adversely contributes to the severity of their hypertension (42). Although continuous positive airway pressure (CPAP) is presently the principal treatment for obstructive sleep apnea, a substantial number of patients do not respond to CPAP (43). In some patients, CPAP induces central apnea and periodic breathing, which places patients at increased risk of hypertension (44). Idiopathic central sleep apnea is also characterized by an exaggerated chemosensory reflex and greatly increased risk of hypertension (45), and treatment options for central apnea are even more limited. In both central and obstructive sleep apnea, the acute spikes in blood pressure associated with apneic episodes may predispose patients to hemorrhagic stroke, whereas chronic hypertension increases the risk of heart failure. Thus, controlling hypertension in sleep apnea patients is a major clinical problem. Our results with intermittent hypoxia, which model both obstructive and central sleep apnea, suggest that inhibiting CSE to reduce H2S signaling in the carotid body may be a novel approach to treat hypertension in patients with sleep apnea.

MATERIALS AND METHODS

Preparation of animals

Experiments were approved by the Institutional Animal Care and Use Committee of the University of Chicago and were performed on age-matched male rats or age-matched male C57/BL6 wild-type and Cse−/− mice (32) unless otherwise noted. Acute experiments were performed on rats and mice anesthetized with urethane (1.2 g/kg, intraperitoneally). At the end of the experiment, rats and mice were euthanized by intracardiac injection (0.1 ml) of euthanasia solution (Beuthanasia-D Special, Schering-Plough). Where indicated, rats were treated by intraperitoneal injection of MnTMPyP (Enzo Life Sciences; 5 mg/kg per day) or l-PAG (Chem-Impex International; 30 mg/kg per day).

Exposure to intermittent hypoxia

The protocols for intermittent hypoxia exposure were described previously (10, 12). Briefly, conscious rats or mice were exposed to alternating cycles of hypoxia (5% O2 for 15 s) and room air (21% O2 for 5 min), between 0900 and 1700 for 10 days. Control experiments were performed on rats and mice exposed to alternating cycles of room air instead of hypoxia. Sixteen hours after terminating the 10th day of intermittent hypoxia exposure, carotid bodies and livers were collected under terminal urethane anesthesia.

Measurements of BP and SNA

Blood pressure was monitored by the tail-cuff method in conscious mice using a noninvasive blood pressure system (IITC Inc.) as described (12). Acute experiments were performed on rats anesthetized with urethane. After tracheal intubation, the femoral artery and vein were cannulated. Blood pressure was monitored through a femoral artery catheter. The rectal temperature was maintained at 37.5 ± 1°C by means of a heating pad. The rats were mechanically ventilated (Harvard Apparatus). The left splanchnic nerve was isolated retroperitoneally, transected, and desheathed. The central cut ends of the nerve were placed on bipolar platinum-iridium electrodes for recording electrical activity as described previously (12). For quantification of SNA, hexamethonium (25 mg/kg) was intravenously administered at the end of the experiment to block the electrical activity for calibration.

Ex vivo carotid body recording

Sensory nerve activity was recorded from ex vivo carotid bodies harvested from anesthetized rats and mice as previously described (10, 27). Briefly, the carotid sinus nerve was treated with collagenase, and several nerve bundles were isolated. Action potentials from one of the nerve bundles were recorded using a suction electrode (~20-μm-diameter tip). In general, two to three action potentials of varying size and amplitude were seen in a given nerve bundle. For analysis of sensory nerve activity, action potentials of the same height, duration, and shape (that is, single unit) were selected using Spike Histogram software (LabChart 7 Pro; fig. S5). To obtain stable baseline sensory nerve activity, we first superfused carotid bodies for 1 hour with room air–equilibrated medium. Baseline sensory nerve activity under normoxia was recorded for 5 min, followed by acute intermittent hypoxia (30 s of hypoxia followed by 5 min of room air for each cycle and a total of 10 cycles in rats and 5 cycles in mice) as described previously (31). Sensory nerve activity was recorded for 60 min after the last cycle of acute intermittent hypoxia. Baseline sensory nerve activity and hypoxic response (cumulative response during the entire cycles of acute intermittent hypoxia) were expressed as impulses per second. Sensory nerve activity during the entire 60 min after the termination of acute intermittent hypoxia was expressed as percentage of baseline activity.

Plasmids, cell culture, and transfection

Complementary DNA (cDNA) encoding rat HO-2 was ligated into pcDNA3.1 (Thermo Fisher, #V790-20) directly 3′ to a nucleotide sequence encoding an N-terminal Myc tag. cDNA encoding mouse CSE was ligated into pCMV-Myc-N (Clontech, #635689). Missense mutations of HO-2 were generated using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies, #200521) according to the manufacturer’s instructions. Mutations were verified by nucleotide sequencing at the University of Chicago core sequencing facility. HEK-293 cells were cultured in Dulbecco’s modified Eagle’s medium (Life Technologies, #11995-065), 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (100 U/ml) in 95% air and 5% CO2 (20% O2) at 37°C. Cells were seeded on 60-mm plates and transfected with 0.5 μg of plasmid DNA encoding either wild-type or mutant HO-2 and/or 2 μg of plasmid DNA encoding CSE using Lipofectamine (Invitrogen, #18324-012) according to the manufacturer’s protocol. HEK-293 cell cultures were exposed to intermittent hypoxia as described previously (13). Where indicated, HEK-293 cells were pretreated with MnTMPyP (50 μM) for 30 min before intermittent hypoxia exposure.

Immunoblot assays for HO-2 and CSE

Immunoblot assays for HO-2 and CSE were performed as previously described (21). The following primary antibodies (and dilution) were used: HO-2 (Abcam, #ab90492; 1:1000), CSE (Novus Biologicals, #NBP1-52849; 1:1000), and α-tubulin (Sigma-Aldrich, #T6199; 1:1000).

Immunoprecipitation assay for CSE phosphorylation

To assess the relative abundance of phosphorylated CSE, we treated either HEK-293 cell or liver extracts with anti-CSE antibody (2 μg; Novus Biologicals, #NBP1-52849) at 4°C for 18 hours. The resulting CSE-antibody complex was pulled down with 50-μl suspension of Protein G Magnetic Beads (Millipore, #LSKMAGG10) by incubating at room temperature for 20 min. The immunoprecipitates were washed with phosphate-buffered saline containing 0.1% Tween 20 and extracted with Laemmli buffer for Western blot analysis. The immunoblots were probed with either anti-phosphoserine antibody (Millipore, #AB1603; 1:1000) or anti-CSE antibody (Novus Biologicals, #NBP1-52849; 1:1000).

Measurement of H2S

H2S generation in the carotid bodies, liver tissues, and HEK-293 cells was determined by in vitro methylene blue assay, as described previously (19), with the following modifications. The absorbance of H2S trapping solution was monitored at 670 nm using Shimadzu UV-VIS Spectrophotometer (UV-2401PC). H2S concentrations were calculated using a molar extinction coefficient of 71,089 M−1 cm−1 at 670 nm and expressed as nanomoles per hour per milligram of protein.

Measurement of CO

CO production in the carotid bodies, liver tissues, and HEK-293 cells was determined using in vitro crystal violet assay, as described previously (20, 46), except that CO concentrations were calculated using a molar extinction coefficient of 87,000 M−1 cm−1 at 590 nm and expressed as nanomoles per hour per milligram of protein.

Measurement of MDA amounts

Carotid bodies, liver tissue, or HEK-293 cells were homogenized in 10volumes of 20 mM phosphate buffer (pH 7.4) at 4°C, and the resulting homogenate was centrifuged at 500g for 10 min at 4°C. MDA levels were analyzed in the supernatant as previously described (27) and were reported in nanomoles of MDA formed per milligram of protein.

Measurement of PKG activity

Cyclic guanosine 3′,5′-monophosphate (cGMP) concentrations were determined by a chemiluminescent enzyme-linked immunosorbent assay (Cell Biolabs, #STA-506) according to the manufacturer’s instructions. Before the experiment, cells were treated with 1 mM 3-isobutyl-1-methylxanthine to inhibit the degradation of cGMP by phosphodiesterases. cGMP concentrations were normalized to protein concentration as determined by bicinchoninic acid assay and were reported in milliunits per milligram of protein.

Measurement of plasma catecholamine concentrations

Blood samples (∼300 μl) were collected from anesthetized mice by cardiac puncture and placed in heparinized (30 U per milliliter of blood) ice-cold microcentrifuge tubes. Plasma was separated by centrifugation and stored at −80°C. Plasma norepinephrine and epinephrine were determined by high-pressure liquid chromatography combined with electrochemical detection using 3,4-dihydroxy-benzyl-amine hydrobromide (Sigma-Aldrich) as an internal standard (27) and expressed as nanograms of catecholamines per milliliter of plasma.

Quantitation of HO-2 and Cse mRNA

HO-2 and Cse mRNA expression in carotid bodies was analyzed by RT-qPCR with SYBR GreenER (Invitrogen, #11764-100) as previously described (20). Briefly, RNA was extracted from rat carotid bodies (two carotid bodies from a single rat) using TRIzol (Thermo Fisher) and was reverse-transcribed using SuperScript III Reverse Transcriptase (Thermo Fisher). Relative mRNA quantification, expressed as fold change, was calculated using the formula 2−ΔΔCT, where ΔCT is the difference between the threshold cycles of the given target cDNA and 18S rRNA, and Δ(ΔCT) is the difference between the ΔCT values under normoxia and intermittent hypoxia. PCR specificity was confirmed by omitting the template and by performing a standard melting curve analysis. The nucleotide sequences of primers used for qPCR are as follows: rat CSE, 5′-AGCGATCACACCACAGACCAA-3′ and 5′-ATCAGCACCCAGAGCCAAAGG-3′; rat HO-2, 5′-ACTGGGAGGAGCAGGTGAAG-3′ and 5′-GGTAGAACTGGGTCCCTTCCC-3′; 18S rRNA, 5′-GTAACCCGTTGAACCCCATT-3′ and 5′-CCATCCAATCGGTAGTAGCG-3′.

Determination of kinetic properties of HO-2

HEK-293 cells coexpressing HO-2 and CSE were homogenized at 4°C with three volumes of 0.05 M tris-HCl buffer (pH 7.4) with 10 mM EDTA and 20% (v/v) glycerol and centrifuged at 100,000g for 70 min at 4°C. CO production was determined as a function of hemin concentration (from 0.01 to 20 μM). The apparent hemin Km and maximum velocity (Vmax) were analyzed as previously described (20).

Drugs

The following drugs were used: MnTMPyP (Enzo Life Sciences, #ALX-430-070-M050), l-PAG (Chem-Impex International, #02726), H2O2 (Sigma-Aldrich, #323381), and PEG-catalase (Sigma-Aldrich, #C4963). All drugs were dissolved in physiological saline and prepared fresh during the experiment. The following concentrations of drugs were used: MnTMPyP, 50 μM; PEG-catalase, 10 U/ml; and H2O2, 100 μM. In experiments involving HEK-293 cells, cells were treated with the desired concentration of drugs 30 min before the experiment. Rats and mice were treated with MnTMPyP (5 mg/kg, intraperitoneally) every day before exposure to intermittent hypoxia. In the experiments with l-PAG, rats were treated with l-PAG (30 mg/kg, intraperitoneally) daily during the last 2 days of exposure to intermittent hypoxia.

Data analysis and statistics

All data are reported as means ± SEM derived from three independent biological experiments, unless otherwise stated in the figure legends. Statistical analysis was performed with either one-way or two-way analysis of variance (ANOVA) with repeated measures followed by post hoc Tukey’s test. For the analysis of normalized data, Mann-Whitney test was used. All P values <0.05 were considered significant.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/9/441/ra80/DC1

Fig. S1. Effect of intermittent hypoxia on H2S, MDA, and CSE protein abundance in rat liver.

Fig. S2. Effect of intermittent hypoxia on HO-2 mRNA and protein abundance.

Fig. S3. Analysis of CSE and HO-2 protein abundance in HEK-293 cells.

Fig. S4. Effect of intermittent hypoxia on CO concentration in HO-1–expressing HEK-293 cells.

Fig. S5. Analysis of carotid body sensory nerve activity.

Fig. S6. Carotid body response to intermittent hypoxia in HO-2−/− mice.

REFERENCES AND NOTES

Acknowledgments: We are grateful to D. Nicolae (University of Chicago) for help with the statistical analysis and R. Wang (Laurentian University, Canada) for providing Cse−/− mice. Funding: This work was supported by NIH grants P01-HL-90554 and UH2-HL-123610. Author contributions: N.R.P. conceived the study; N.R.P. and G.K.K. designed research; G.Y., Y.-J.P., S.A.K., J.N., A.S., and G.K.K. performed experiments; G.Y., Y.-J.P., S.A.K., A.S., J.N., and G.K.K. analyzed the data; C.V. and S.H.S. contributed reagents; N.R.P., G.K.K., G.L.S., and S.H.S. wrote the paper. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article