OUP user menu

Ascorbic acid and tetrahydrobiopterin potentiate the EDHF phenomenon by generating hydrogen peroxide

Ambroise Garry, David H. Edwards, Ian F. Fallis, Robert L. Jenkins, Tudor M. Griffith
DOI: http://dx.doi.org/10.1093/cvr/cvp235 218-226 First published online: 10 July 2009


Aims Our objective was to investigate whether pro-oxidant properties of ascorbic acid (AA) and tetrahydrobiopterin (BH4) modulate endothelium-dependent, electrotonically mediated arterial relaxation.

Methods and results In studies with rabbit iliac artery (RIA) rings, NO-independent, endothelium-derived hyperpolarizing factor (EDHF)-type relaxations evoked by the sarcoplasmic endoplasmic reticulum Ca2+-ATPase inhibitor cyclopiazonic acid and the G protein-coupled agonist acetylcholine (ACh) were enhanced by AA (1 mM) and BH4 (200 µM), which generated buffer concentrations of H2O2 in the range of 40–80 µM. Exogenous H2O2 potentiated cyclopiazonic acid (CPA)- and ACh-evoked relaxations with a threshold of 10–30 µM, and potentiation by AA and BH4 was abolished by catalase, which destroyed H2O2 generated by oxidation of these agents in the organ chamber. Adventitial application of H2O2 also enhanced EDHF-type dilator responses evoked by CPA and ACh in RIA segments perfused intraluminally with H2O2-free buffer, albeit with reduced efficacy. In RIA rings, both control relaxations and their potentiation by H2O2 were overcome by blockade of gap junctions by connexin-mimetic peptides (YDKSFPISHVR and SRPTEK) targeted to the first and second extracellular loops of the dominant vascular connexins expressed in the RIA. Superoxide dismutase attenuated the potentiation of EDHF-type relaxations by BH4, but not AA, consistent with findings demonstrating a differential role for superoxide anions in the generation of H2O2 by the two agents.

Conclusion Pro-oxidant effects of AA and BH4 can enhance the EDHF phenomenon by generating H2O2, which has previously been shown to amplify electrotonic hyperpolarization-mediated relaxation by facilitating Ca2+ release from endothelial stores.

  • Connexin-mimetic peptides
  • Gap junction
  • Superoxide anion

1. Introduction

In many arteries, NO-independent relaxations can be inhibited by synthetic connexin-mimetic peptides that interrupt intercellular communication via myoendothelial and homocellular smooth muscle gap junctions, suggesting that such responses are electrotonic in nature, rather than mediated by a freely diffusible endothelium-derived hyperpolarizing factor (EDHF).13 We have previously provided evidence that H2O2 can potentiate such ‘EDHF-type’ relaxations in rings of rabbit iliac artery (RIA) when these are evoked by cyclopiazonic acid (CPA), an agent that promotes store-operated endothelial Ca2+ entry by inhibiting the sarcoplasmic endoplasmic reticulum (ER) Ca2+-ATPase pump of the ER Ca2+ store.4 This novel action of H2O2 may reflect enhanced Ca2+ store depletion secondary to sensitization of the InsP3 receptor, with the resulting increase in Ca2+ mobilization promoting the opening of the hyperpolarizing endothelial KCa channels that are widely recognized to underpin the EDHF phenomenon.46 We have also shown that the inhibitory effects of connexin-mimetic peptides against EDHF-type relaxations are attenuated by ascorbic acid (AA) and R-5,6,7,8-tetrahydrobiopterin (BH4),7,8 whose ability to improve endothelial function in patients with hypertension, hypercholesterolaemia, diabetes, and heart failure has widely been attributed to their ability to prevent uncoupling of the constitutive eNOS and thereby reduce production of the superoxide anion (·O2) by the oxygenase component of the enzyme.913 It is well known, however, that AA and BH4 can generate H2O2 following oxidation by molecular oxygen in aqueous solution.1315 To examine whether this pro-oxidant activity also modulates endothelium-dependent relaxation, we have correlated the ability of AA and BH4 to generate H2O2 with potentiation of EDHF-type relaxations evoked by CPA and acetylcholine (ACh) in the RIA. In this vessel, H2O2 cannot be regarded as an EDHF because H2O2-evoked changes in smooth muscle membrane potential are much smaller than those associated with endothelium-dependent smooth muscle hyperpolarization, and H2O2-evoked relaxations of endothelium-denuded preparations are unaffected by blockade of K+ channels.4,16 Most experiments were conducted with ring preparations in which the endothelium was directly exposed to H2O2, but H2O2 was also applied adventitially in perfused arterial segments to mimic the in vivo situation where systemic administration of pharmacological doses of AA have been shown to generate high concentrations of H2O2 in interstitial fluid, but circulating H2O2 is efficiently destroyed by red cell glutathione peroxidase and catalase.17,18

2. Methods

Experiments were performed with iliac arteries from male NZW rabbits (2–2.5 kg) sacrificed by injection of sodium pentobarbitone (120 mg/kg i.v.). Protocols conformed to UK Home Office regulations and the Guide for the Care and Use of Laboratory Animals issued by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). Tissues were transferred to oxygenated (95% O2, 5% CO2) Holman's buffer containing (in mM): NaCl 120, KCl 5, NaH2PO4 1.3, NaHCO3 25, CaCl2 2.5, glucose 11, and sucrose 10. Myograph experiments were conducted with oxygenated Holman's solution (95% O2, 5% CO2) at 37°C and pH 7.4. To evaluate EDHF-type responses, in all experiments, the cyclooxygenase inhibitor indomethacin (10 µM) and NG-nitro-l-arginine methyl ester (300 µM) were added to the buffer 40 min before tone was induced by phenylephrine (PE, 1 µM). All pharmacological agents were obtained from Sigma, UK.

2.1 Tension myography

Rings (2–3 mm wide) were mounted in a myograph (model 610M; Danish MyoTechnology, Aarhus) placed under a resting tension of 1 mN and then allowed to equilibrate for ∼30 min with readjustments of tension to compensate for stress relaxation. Cumulative concentration–response curves to CPA and ACh were constructed under control conditions and after 30 min incubation with AA (1 mM) or BH4 (200 µM) in the presence and absence of catalase (Cat No. C9322 derived from bovine liver) or superoxide dismutase (SOD) (Cat No. S7571 derived from bovine erythrocytes). Some rings were incubated with H2O2 (10, 30, or 100 µM) for 30 min before constriction. The role of gap junctions was investigated by pre-incubating for 20 min with 43Gap26 (VCYDKSFPISHVR; 100 µM), the truncated peptide YDKSFPISHVR, and the short peptides SRPTEK or its unnatural enantiomeric d-isomer srptek (each at 300 µM). Stock solutions were prepared in buffer with the exception of CPA (DMSO), indomethacin (5% ethanol), 43Gap 26 and YDKSFPISHVR (dH2O) and SRPTEK/srptek (10% acetic acid). Peptides were confirmed to be of greater than 85% purity by HPLC.

2.2 Perfusion myography

Iliac artery segments (4–5 mm in length) were cannulated onto two glass micropipettes in a pressure myograph (Living Systems Instrumentation, USA). Flow and pressure were fixed at 0.5 mL/min and 75 mmHg, respectively, giving a basal external arterial diameter 1500 µm. PE (1 µM), CPA and ACh were administered intraluminally and in some experiments 100 µM H2O2 was added to the extraluminal Holman's solution in the myograph chamber 30 min before constriction. At the end of each experiment, arteries were perfused with sodium nitroprusside (SNP, 100 µM) to induce full dilatation. Diameter changes were recorded on PowerLab 400 using Chart v4.1.2 software (AD Instruments, UK).

2.3 Hydrogen peroxide assay

Briefly, 300 µL samples of buffer were collected at the beginning and end of relaxation protocols in experiments with ACh or CPA (corresponding to AA/BH4 incubation times of 30 and 60 min). These were added to Amplex Red (10 µM) and horseradish peroxidase (0.6 U/mL) in a 96-well plate and incubated in the dark at room temperature for 15 min. Absorbance was read at 560 nm using a Fluostar optima spectrophotometer (BMG Labtech) and H2O2 concentrations derived from a standard curve. Experiments were also performed in the absence of arterial rings with buffer maintained at 37°C and either oxygenated with 95% O2/5% CO2 or exposed to air. In perfusion experiments, buffer was sampled from the myograph chamber and the effluent from the artery under study.

2.4 Mass spectra

Peptide mass spectra were recorded using either a Waters 1525 µ HPLC or UPLC Aquity autosamplers equipped with LCT Premier XE or Q-Tof micro mass sensitive detectors, respectively. Spectra were obtained in electrospray positive ion mode (ESI+) using a carrier solvent of 50:50 mixture of acetonitrile:deionized water. Formic acid (0.1% of eluent) was used as the proton source.

2.5 Statistics

In tension experiments, the maximal percentage reversal of PE-induced constriction (Rmax) by CPA or ACh and concentrations giving 50% reversal of this constrictor response (IC50 for CPA) or 50% of maximal relaxation (EC50 for ACh) were determined for each experiment. The use of IC50 rather than EC50 values was necessary to allow for a small initial CPA-induced constriction observed in ring experiments.4,8 In perfusion studies, dilatations evoked by CPA and ACh were expressed as a function of the response to SNP, to calculate EC50 and Dmax. All parameters were calculated as mean ± SEM and compared by the Student's t-test or ANOVA followed by a Bonferroni post-test. P < 0.05 was considered significant; n denotes the number of animals studied or assays performed for each data point.

3. Results

3.1 Effects of AA and BH4 on CPA- and ACh-evoked relaxation

Pre-incubation of RIA rings with 1 mM AA or 200 µM BH4 caused leftward shifts in the concentration–relaxation curves for CPA and ACh without affecting Rmax and increased buffer [H2O2] to ∼40 and ∼60 µM after 60 min, respectively (Figure 1AD and Table 1). Increases in myograph [H2O2] and the potentiating effects of AA and BH4 on relaxation were abolished by 1000 U/mL catalase, which did not itself modulate control responses to either CPA or ACh (Figure 1AD and Table 1).

Figure 1

Effects of AA and BH4 on EDHF-type relaxations to CPA and ACh in RIA rings. (A and B) 1 mM AA and (C and D) 200 µM BH4 each elevated bath [H2O2] (bar graphs) and resulted in a catalase-sensitive potentiation of relaxation. [H2O2] was measured at the conclusion of each experiment, i.e. after 60 min incubation with AA or BH4. ***P < 0.001 for specific concentrations of CPA or ACh compared with control.

View this table:
Table 1

Effects of pharmacological interventions on EDHF-type relaxations evoked by CPA and ACh by in the presence and absence of H2O2 or catalase

CPAnpIC50Rmax %AChnpEC50Rmax %
Control54.61 ± 0.0192.8 ± 2.1Control56.55 ± 0.0975.8 ± 2.7
Catalase54.66 ± 0.0793.8 ± 2.4Catalase56.64 ± 0.0877.2 ± 3.3
AA55.12 ± 0.05***92.3 ± 2.3AA57.00 ± 0.06*81.9 ± 1.2
AA+catalase54.59 ± 0.1191.5 ± 3.0AA+catalase56.48 ± 0.1075.3 ± 2.2
Control74.70 ± 0.1079.6 ± 5.4Control66.43 ± 0.0876.6 ± 3.8
Catalase74.74 ± 0.0485.8 ± 2.5Catalase66.39 ± 0.1073.1 ± 3.6
BH475.39 ± 0.15***85.1 ± 2.1BH466.91 ± 0.04**79.7 ± 3.1
BH4+catalase74.76 ± 0.0583.9 ± 2.5BH4+catalase66.42 ± 0.0871.2 ± 3.9
Control74.64 ± 0.0484.3 ± 3.1Control86.50 ± 0.0674.9 ± 4.4
H2O2 (10 µM)74.80 ± 0.0490.3 ± 1.6H2O2 (10 µM)76.55 ± 0.0872.7 ± 5.2
H2O2 (30 µM)74.96 ± 0.1091.6 ± 2.5H2O2 (30 µM)66.87 ± 0.2076.8 ± 5.5
H2O2 (100 µM)75.36 ± 0.17***86.0 ± 2.5H2O2 (100 µM)97.25 ± 0.15***74.7 ± 4.1
Control44.80 ± 0.0392.6 ± 1.4Control76.46 ± 0.0969.0 ± 3.0
SOD (1200 U)44.71 ± 0.0484.1 ± 2.0SRPETK726.5 ± 7.3***
AA45.53 ± 0.19**87.5 ± 3.6srptek56.29 ± 0.2468.9 ± 5.6
AA+SOD45.53 ± 0.15**90.2 ± 2.0
Control44.67 ± 0.1693.4 ± 3.2Control76.63 ± 0.1077.9 ± 5.4
SOD (1200 U)44.66 ± 0.0680.0 ± 7.9SRPTEK734.2 ± 8.9**
BH446.01 ± 0.20***88.1 ± 7.1SRPTEK+H2O2539.0 ± 11.1*
BH4+SOD45.39 ± 0.21*85.5 ± 9.1H2O2 (100 µM)57.16 ± 0.1**74.6 ± 7.1
Control64.92 ± 0.1178.0 ± 7.6
43Gap26 (100 µM)6-30.6 ± 6.9**
43Gap26+H2O245.03 ± 0.1773.5 ± 10.4
H2O2 (100 µM)46.31 ± 0.15***72.8 ± 7.9
Control64.73 ± 0.0583.1 ± 3.8
YDKSFPISHVR (300 µM)4-35.2 ± 14.1***
YDKSFPISHVR+H2O244.59 ± 0.0180.2 ± 3.3
H2O2 (100 µM)65.30 ± 0.11***84.3 ± 4.2
  • Potency (negative log IC50 or EC50) and maximal relaxation (Rmax) expressed as a percentage of the constrictor response to phenylephrine are given as mean ± SEM.

  • *P < 0.05 compared with control.

  • **P < 0.01 compared with control.

  • ***P < 0.001 compared with control.

3.2 Effects of H2O2 on CPA- and ACh-evoked relaxation/dilatation

Pre-incubation of endothelium-intact rings with increasing [H2O2] progressively lowered IC50/EC50 values for CPA and ACh with a threshold for potentiation between 10 and 30 µM H2O2, but exerted no effect on Rmax even at 100 µM H2O2 (Figure 2A; Table 1). Assay of buffer on completion of the relaxation protocols confirmed that intrinsic antioxidant mechanisms did not reduce applied myograph [H2O2] (Figure 2A). In perfused RIA segments, concentrations of CPA and ACh causing 50% of maximal dilatation were similar to the IC50/EC50 values observed in rings, and adventitial application of 100 µM H2O2 potentiated these responses to an extent intermediate between 30 and 100 µM H2O2 in rings, again without affecting Dmax (Figure 2B; Tables 1 and 2). H2O2 was undetectable in buffer exiting from the lumen of segments exposed to 100 µM H2O2 on their adventitial surface (Figure 2B).

Figure 2

Concentration-dependent potentiation of EDHF-type relaxations/dilatations to CPA and ACh in the presence of exogenous H2O2. (A) In tension myograph experiments, buffer [H2O2] was unchanged at the conclusion of the experiments (bar graph, illustrated for ACh). (B) In perfusion experiments, intraluminal and extraluminal [H2O2] at the conclusion of each experiment were similarly unchanged (bar graphs). ** and *** denote P < 0.01 and 0.001 for specific concentrations of CPA or ACh compared with control.

View this table:
Table 2

Effects of adventitially applied H2O2 on EDHF-type dilatations evoked by CPA and ACh

Control54.79 ± 0.0488.0 ± 2.3
H2O2 (100 µM)55.09 ± 0.05*91.4 ± 2.9
Control56.30 ± 0.0878.2 ± 4.7
H2O2 (100 µM)56.75 ± 0.09*82.6 ± 2.0
  • Potency (negative log EC50) and maximal dilatation (Dmax) expressed as a percentage of the dilator response to SNP are given as mean ± SEM.

  • *P < 0.01 compared with control.

3.3 Role of superoxide anions

SOD (1200 U/mL) did not affect the potentiation of CPA-evoked relaxation caused by 1 mM AA, but significantly attenuated the potentiation observed with 200 µM BH4 (Figures 3A and 4A; Table 1). Corresponding assays of myograph [H2O2] demonstrated that generation of H2O2 from AA was unaffected by SOD at 30 or 60 min (i.e. the start and completion of the relaxation protocols), whereas H2O2 accumulation from BH4 was reduced by ∼50% after 30 min, but unaffected at 60 min (Figures 3A and 4A).

Figure 3

Effects of SOD (1200 U/mL) on concentration–response curves for CPA and generation of H2O2 from 1 mM AA. (A) SOD did not affect the potentiating effects of AA on relaxation or elevations in myograph [H2O2] at 30 and 60 min (insets). (B) Time course of the generation of H2O2 in oxygenated buffer. (C) DTPA reduced the formation of H2O2 in well-oxygenated buffer and no generation of H2O2 was evident in comparative experiments performed in deionized water (dH2O). (D) SOD did not affect the generation of H2O2 either in well-oxygenated buffer or air-exposed buffer (open columns). ** and *** denote P < 0.01 and 0.001 compared with the appropriate control.

Figure 4

Effects of SOD (1200 U/mL) on concentration–response curves for CPA and generation of H2O2 from 200 µM BH4. (A) SOD attenuated the potentiating effects of BH4 on relaxation and the associated increase in myograph [H2O2] after 30 but not 60 min (insets). (B) Time course of H2O2 generation in oxygenated buffer. (C) DTPA did not affect the formation of H2O2 in oxygenated buffer. (D) Inhibitory effects of SOD were much less evident in well-oxygenated buffer than air-exposed buffer (open columns). *, ** and *** denote P < 0.05, 0.01, and 0.001 compared with control.

In experiments conducted in well-oxygenated buffer (95% O2/5% CO2) in the absence of arterial rings, generation of H2O2 from 1 mM AA rose to a plateau at 60–90 min (Figure 3B). The transition metal chelator DTPA caused a concentration-dependent reduction in H2O2 accumulation at 30 min, and no H2O2 was detectable in experiments employing oxygenated deionized water rather than buffer (Figure 3C). Generation of H2O2 from 1 mM AA was insensitive to the presence of SOD and was two- to three-fold greater in well-oxygenated buffer compared with buffer exposed to air (Figure 3D).

In analogous experiments with 200 µM BH4, generation of H2O2 was rapid (∼60 µM after 5 min) and rose to a peak at ∼90 µM at 30 min before subsequently declining at 60 and 90 min (Figure 4B), but was unaffected by DTPA, in contrast to AA (Figure 4C). The high concentrations of DTPA required to inhibit the accumulation of H2O2 in experiments with AA are likely to reflect competition between Ca2+ and other metal ions present in the buffer (see Supplementary material online).

Accumulation of H2O2 after 30 min was similar at ∼80 µM whether the buffer was well-oxygenated or exposed to air, but in buffer exposed to air accumulation of H2O2 was almost abolished by SOD at concentrations ≥300 U/mL after 30 min, whereas in well-oxygenated buffer H2O2 production was depressed only at SOD concentrations ≥900 U/mL, with 1200 U/mL SOD causing an overall reduction of ∼50% (Figure 4D).

3.4 Role of gap junctions

Pre-incubation of RIA rings with 100 µM 43Gap26 attenuated relaxations to CPA, reducing Rmax to ∼30%, as previously reported.3,7,8 Potentiation of CPA-evoked relaxation by 100 µM H2O2 was lost in the presence of this peptide, which normalized relaxation (Figure 5A). Electrospray mass spectrometry demonstrated that samples of 43Gap26 treated with H2O2 (both at 100 µM) at pH ∼7 that were allowed to stand for 60 min either in buffer or deionized water displayed substantial oxidation to cystine-bridged dimers as evidenced by the presence of the triply protonated species [(VCYDKSFPISHVR)2 - 2H]+3H+ (calculated m/e for C140H212N38O38S2+ 3H+ {M+3H+} 1033.52; found 1033.46), the triple sodium adduct [(VCYDKSFPISHVR)2 - 2H]+3Na+ (calculated m/e for C140H212N38O38S2+3Na+ {M+3Na+} 1055.50; found 1055.80), and mixed proton/sodium adducts at intermediate mass values.

Figure 5

Effects of connexin-mimetic peptides on control EDHF-type relaxations and their potentiation by exogenous H2O2. (A) Attenuation of CPA-evoked relaxations by 43Gap 26 and its truncated form YDKSFPISHVR were both reversed by 100 µM H2O2. (B) SRPTEK almost abolished relaxations to ACh, both in the presence and absence of 100 µM H2O2, whereas srptek was inactive. (C) 43Gap 26 but not SRPTEK reduced applied [H2O2] in oxygenated buffer. ** and *** denote P < 0.01 and 0.001 compared with control.

Pre-incubation with a cysteine-free truncated form of 43Gap26 (YDKSFPISHVR at 300 µM) closely mimicked the effects of 43Gap26 at 100 µM (Figure 5A). Pre-incubation with the short peptide SRPTEK at 300 µM markedly attenuated relaxations to ACh, reducing Rmax to less than 30%, whereas the corresponding d-isomer srptek was inactive; SRPTEK similarly abolished the potentiated ACh-evoked relaxations observed in the presence of 100 µM H2O2 (Figure 5B). Mass spectrometry confirmed the lack of dimerization of the YDKSFPISHVR or SRPTEK oligopeptides (data not shown).

Direct chemical interaction between 43Gap26 and H2O2 was confirmed by the demonstration that there was ∼50% consumption of 100 µM H2O2 30 min after addition to air-exposed Holman's buffer containing 100 µM 43Gap26, whereas consumption of H2O2 was not evident in buffer containing 100 µM SRPTEK (Figure 5C).

4. Discussion

The major finding of the present study is that AA and BH4 potentiate EDHF-type relaxations of rabbit arteries evoked by CPA and ACh through a mechanism that is sensitive to catalase. Control and potentiated responses were both inhibited by connexin-mimetic peptides. It follows that H2O2 generated by the oxidation of AA and BH4 can amplify NO-independent arterial relaxations mediated by the spread of endothelial hyperpolarization via gap junctions.

Incubation of RIA rings with 1 mM AA or 200 µM BH4 enhanced EDHF-type responses with the threshold for relaxation decreasing from 10–30 µM to ∼1 µM with CPA and from ∼0.1 µM to ∼0.01 µM with ACh. Assay of myograph [H2O2] after 60 min incubation revealed conversion rates of ∼4% and ∼30% for AA and BH4, respectively, consistent with evidence that BH4 is more readily oxidized by molecular oxygen than AA in physiological buffer.13 The potentiating effects of AA and BH4, but not control responses, were inhibited by catalase and could be mimicked by exogenous H2O2, which itself enhanced EDHF-type relaxations at a threshold concentration of 10–30 µM. The observation that H2O2 amplifies relaxant responses to ACh, which mobilizes Ca2+ from the ER store via the formation of InsP3, is consistent with evidence that H2O2 enhances Ca2+ release by sensitizing the InsP3 receptor and extends previous findings with CPA which elevates endothelial [Ca2+]i by blocking ER Ca2+ uptake.46 EDHF-type relaxations evoked by CPA and ACh can also be potentiated by the sulphydryl reagent thimerosal, which amplifies Ca2+ release from the ER by oxidizing critical thiol groups present in the InsP3 receptor, thus raising the possibility of a molecular target common to H2O2 and thimerosal.4,1921 It should be appreciated that the concentrations of buffer H2O2 generated from AA and BH4 in the present study are likely to correspond to intracellular levels within the suggested physiological range (1–10 µM), since glutathione peroxidase, catalase, and other mechanisms are thought to limit cytosolic [H2O2] to 1–15% of that applied extracellularly.22

In the rat, microdialysis techniques have shown that systemic administration of AA at pharmacological doses sufficient to achieve circulating AA levels of 1–10 mM causes extravascular accumulation of H2O2 and the ascorbyl radical (a marker of AA oxidation) at concentrations that correlate directly with plasma [AA], with interstitial fluid [H2O2] rising to 20–150 µM but H2O2 remaining undetectable in the intravascular compartment.17,18 To mimic this in vivo situation, 100 µM H2O2 was selectively applied to the adventitia of perfused RIA segments and found to potentiate NO-independent CPA- and ACh-evoked dilatations, although less effectively than in ring preparations where the endothelium was directly exposed to the same concentration of H2O2. This reduction in potency is likely to reflect a concentration gradient of H2O2 across the wall of the segments because their intraluminal perfusate did not contain H2O2 and the RIA possesses ∼10 layers of smooth muscle cells and is therefore relatively thick-walled.3 It thus seems likely that putative H2O2-dependent pro-oxidant effects of AA and BH4 on endothelial function in vivo would be most pronounced in the microcirculation because (i) the adventitial–endothelial concentration gradient of H2O2 would be smaller than in conduit arteries and (ii) gap junction-dependent mechanisms can dominate over NO-mediated vasodilatation in resistance arteries,23 consistent with evidence that the number of myoendothelial connections per endothelial cell is highest in small arteries.24

4.1 Mechanisms involved in the generation of H2O2

To gain insights into the pathways whereby AA and BH4 generate H2O2 in physiological buffer, experiments were conducted to evaluate the contribution of the superoxide anion (·O2), which has been implicated in the autoxidation of BH4, but whose role in the oxidation of AA is controversial.14,15,25,26 In myograph studies with rings, SOD attenuated the potentiating effects of BH4 on CPA-evoked relaxations and associated formation of H2O2, whereas the generation of H2O2 from AA and its potentiating effects on relaxation were unaffected. Spin trap analysis has shown that BH4 reacts with molecular oxygen at physiological pH to generate ·O2, which then drives a chain reaction involving the reduction of molecular oxygen via an intermediary BH4 radical that accelerates the rate of BH4 oxidation by ∼5-fold.15 Loss of this chain reaction is likely to explain the ability of SOD to attenuate H2O2 formation and relaxation in the present study, since SOD causes a marked reduction the rate of BH4 autoxidation.15 In contrast, aqueous solutions of AA are stable at pH ∼7 unless trace concentrations of Fe3+ or Cu2+ ions (normally present in commercially available salts) are present as contaminants that catalyze its oxidation.25,27 We were thus unable to detect formation of H2O2 from AA in deionized water and found that the generation of H2O2 from AA in buffer was substantially reduced in the presence of DTPA, a polydentate scavenger of Fe3+ and Cu2+ ions,25,27 whereas formation of H2O2 from BH4 was unaffected. Conversely, 1 µM Fe3+ or Cu2+ catalyzed the generation of H2O2 from 1 mM AA, with Cu2+ being by far the more active cation (see Supplementary material online). Although metal-catalyzed oxidation of AA has been suggested to involve the interaction of ·O2 with ascorbate ions and the ascorbyl radical, the effects of SOD on the rate of AA oxidation were modest (less than 2-fold decrease),14 and it has also been proposed that metal-catalyzed oxidation (e.g. of 6-hydroxydopamine and 1,2,4-benzenetriol) can proceed via a 2-electron mechanism in which H2O2 is generated directly from molecular oxygen (rather than ·O2) and is therefore insensitive to SOD.28,29 The impact of oxygenation on H2O2 accumulation was examined in experiments conducted with buffer exposed to air, because isolated arterial preparations are conventionally maintained in buffer gassed with O2 at levels well above the physiological range. Such studies demonstrated a ∼50% reduction in H2O2 generated from AA after 30 min compared with oxygenated buffer, whereas H2O2 formation from BH4 was unaltered. Major differences in the role of intermediary ·O2 in the oxidation of AA and BH4 were nevertheless highlighted by observations that the generation of H2O2 from BH4 in buffer exposed to air was suppressed by SOD, whereas its formation from AA under the same conditions was unaffected.

4.2 Role of gap junctions

Studies with connexin-mimetic peptides have confirmed the underlying electrotonic nature of NO- and prostanoid-independent responses to CPA and ACh in rabbit arteries, and shown that such peptides do not attenuate direct NO-mediated relaxation or endothelial release of endogenous NO in sandwich bioassay experiments.3,7,8,30 As previously reported, CPA-evoked relaxations of the RIA were attenuated by the peptide VCYDKSFPISHVR (43Gap26; 100 µM), which possesses homology with the first extracellular loop of Cx43, the dominant connexin expressed in the media of the RIA and interrupts the spread of CPA-evoked endothelial hyperpolarization through the vessel wall.3,7,8 In the present study, the inhibitory effects of 100 µM 43Gap 26 and the potentiating effects of 100 µM H2O2 on CPA-evoked relaxation cancelled when both agents were present simultaneously, with relaxation then being maintained at control levels. This may, at least in part, reflect oxidation of the thiol group of the cysteine residue of 43Gap26 by H2O2, because dimerization via the formation of an intermolecular cystine bridge was confirmed by mass spectrometry and there was 50% consumption of applied H2O2 in the presence of equimolar peptide concentrations, consistent with the reaction 2R-SH+H2O2→R-S-S-R+2 H2O. However, we also found that a truncated peptide YDKSFPISHVR mimicked the effects of 43Gap26 when applied at 300 µM, suggesting specific gap junction blockade because oxidation/dimerization of this peptide was prevented by deletion of the N-terminal valine and cysteine residues. Furthermore, the peptide SRPTEK, which corresponds to a highly conserved sequence in the second extracellular loops of the principal endothelial connexins expressed in the RIA (Cx37 and Cx40), as well as Cx43,3 caused near-complete inhibition of ACh-evoked EDHF-type relaxations, even when these were potentiated by H2O2. This short peptide may therefore be considered a general, redox-insensitive inhibitor of gap junction signalling because (i) SRPTEK did not dimerize in the presence of H2O2, (ii) assay revealed no consumption of buffer H2O2, and (iii) the d-isoform (srptek) was biologically inactive, suggesting that loss of relaxation driven by SRPTEK involves a specific molecular recognition event at the cell surface.

We have previously shown that 100 µM BH4 opposes the ability of equimolar concentrations of 43Gap26 or the related peptide 37,40Gap26 (VCYDQAFPISHIR, which targets Cx37 and Cx40) to inhibit CPA-evoked EDHF-type relaxations and smooth muscle hyperpolarization in the RIA.7 At this concentration, BH4 generates myograph H2O2 concentrations of ∼40 µM (data not shown) and could therefore contribute to the normalization of EDHF-type responses both via the potentiation of ER Ca2+ mobilization and peptide oxidation. However, at concentrations normally found in plasma (50–100 µM), which are lower than those employed in the present experiments, AA can also oppose the inhibition of CPA-induced relaxation and smooth muscle hyperpolarization by 43Gap26 and 37,40Gap26.8 It should also be noted that in non-vascular cells H2O2 has been variously shown to enhance or inhibit intercellular coupling via gap junctions constructed from Cx43, probably by alternating the phosphorylation/oxidation status of residues present in the intracellular cytoplasmic tail of this connexin subtype.3135 Indeed, the hyperphosphorylation of Cx43 that follows administration of H2O2 or phorbol esters (which generate ·O2) can be prevented by a spectrum of antioxidants, including AA, with preservation of channel function.36,37 Further studies are therefore necessary to evaluate the effects of competing pro- and antioxidant mechanisms on gap junctional communication in the endothelial and smooth muscle layers of the vessel wall, and how their contributions might vary under different experimental conditions because AA and BH4 are both capable of reducing H2O2 to H2O as well as generating H2O2.14,15,38,39

4.3 Conclusions

Clinical studies have suggested that AA and BH4 can both improve endothelial dysfunction in human conduit arteries by increasing the bioavailability of NO, provided that pharmacological doses are administered systemically.912 The present findings raise the possibility that high concentrations of AA and BH4 might also reverse endothelial dysfunction by amplifying the EDHF-type responses that have been postulated to compensate for loss of NO-dependent dilatation. It should thus be noted that the concentration of AA employed in the present in vitro studies (1 mM) is lower than the venous concentrations (1.5–3.2 mM) associated with restoration of endothelial responsiveness to ACh/methacholine in forearm plethysmographic studies in patients with hypertension or peripheral arterial disease following intra-brachial arterial administration of AA.10,11 As noted above, such concentrations of circulating AA can elevate interstitial fluid [H2O2] to levels that potentiated NO-independent dilatation when applied to the adventitia of RIA segments. Corresponding measurements of interstitial [H2O2] are not available for BH4, although the infusion rates found necessary to prevent endothelial dysfunction following ischaemia–perfusion injury in the human forearm lead to total circulating biopterin concentrations of ∼100 µM and appear to involve mechanisms distinct from the role of BH4 as a co-factor for eNOS.40 Further studies are therefore needed to assess the effects of extracellularly generated H2O2 on endothelial function in vivo. Because H2O2 production from ·O2 is normally attenuated by NO through an interaction that results in the formation of peroxynitrite, it also remains to be determined if the reduction in NO bioavailability that characterizes many vascular disease states upregulates the EDHF phenomenon directly.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Conflict of interest: none declared.


The study was supported by the British Heart Foundation (Grant PG/05/133/19892) and the Cardiff Institute of Tissue Engineering (CITER). Funding to pay the Open Access publication charges for this article was provided by the British Heart Foundation.

The online version of this article has been published under an open access model. Users are entitled to use, reproduce, disseminate, or display the open access version of this article for non-commercial purposes provided that the original authorship is properly and fully attributed; the Journal, Learned Society and Oxford University Press are attributed as the original place of publication with correct citation details given; if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated. For commercial re-use, please contact journals.permissions@oxfordjournals.org.


View Abstract