CNGA2 Channels Mediate Adenosine-Induced Ca2+ Influx in Vascular Endothelial Cells
Objectives— Adenosine is a cAMP-elevating vasodilator that induces both endothelium-dependent and -independent vasorelaxation. An increase in cytosolic Ca2+ ([Ca2+]i) is a crucial early signal in the endothelium-dependent relaxation elicited by adenosine. This study explored the molecular identity of channels that mediate adenosine-induced Ca2+ influx in vascular endothelial cells.
Methods and Results— Adenosine-induced Ca2+ influx was markedly reduced by L-cis-diltiazem and LY-83583, two selective inhibitors for cyclic nucleotide-gated (CNG) channels, in H5V endothelial cells and primary cultured bovine aortic endothelial cells (BAECs). The Ca2+ influx was also inhibited by 2 adenylyl cyclase inhibitors MDL-12330A and SQ-22536, and by 2 A2B receptor inhibitors MRS-1754 and 8-SPT, but not by an A2A receptor inhibitor SCH-58261 or a guanylyl cyclase inhibitor ODQ. Patch clamp experiments recorded an adenosine-induced current that could be inhibited by L-cis-diltiazem and LY-83583. A CNGA2-specific siRNA markedly decreased the Ca2+ influx and the cation current in H5V cells. Furthermore, L-cis-diltiazem inhibited the endothelial Ca2+ influx in mouse aortic strips, and it also reduced 5-N-ethylcarboxamidoadenosine (NECA, an A2 adenosine receptor agonist)-induced vasorelaxation.
Conclusion— CNGA2 channels play a key role in adenosine-induced endothelial Ca2+ influx and vasorelaxation. It is likely that adenosine acts through A2B receptors and adenylyl cyclases to stimulate CNGA2.
Adenosine is an endogenous nucleoside with potent vasodilatory capacities in many vascular beds.1 Adenosine can be released from myocardium, endothelial cells, and skeletal muscles as a result of metabolism. The released adenosine then elicits vasorelaxation either by directly stimulating A2-adenosine receptors in vascular smooth muscle cells, causing subsequent vasorelaxation,1 or by indirectly acting on vascular endothelial cells, triggering endothelium-dependent vasorelaxation.1–3
Adenosine induces the endothelium-dependent vasorelaxation either via a Ca2+-dependent mechanism3–5 or a Ca2+-independent mechanism.2 In the former case, the adenosine-induced [Ca2+]i rise stimulates endothelial cells to release vasorelaxants such as nitric oxide (NO)3,6 and endothelium-derived hyperpolarizing factor (EDHF).7 These vasodilators diffuse to nearby smooth muscle cells, causing vascular smooth muscle relaxation. At least in some arteries, this [Ca2+]i rise is a prerequisite for adenosine-induced vasodilator release and vasorelaxation.3–5 For example, in skeletal muscle arterioles, adenosine elicits an endothelial [Ca2+]i rise,3,5 and chelation of endothelial [Ca2+]i with intraluminal perfusion of BAPTA-AM abolishes the adenosine-induced vasorelaxation, indicating an obligatory role of endothelial [Ca2+]i in this relaxation.3,5 In another study, it was found that adenosine induces NO release from rat aortic endothelium in situ, and this NO release requires Ca2+ influx.4
Virtually nothing is known about the molecular identity of channels that mediate adenosine-induced Ca2+ influx in endothelial cells. Multiple Ca2+-permeable channels are expressed in vascular endothelial cells. These include TRP channels8 and CNG channels.9–12 CNG channels are activated by cAMP and cGMP,13 the levels of which are elevated when endothelial cells are exposed to adenosine.1,14,15 The activity of CNG channels can be inhibited by L-cis-diltiazem and LY-83583.16,17 The L-cis isomer of diltiazem selectively blocks CNG channels, whereas the D-cis-isomer inhibits L-type Ca2+ channels.16 Among 3 functional CNG isoforms (CNGA1-A3), CNGA2 has a much higher sensitivity to cAMP than other CNG isoforms.13 This property makes CNGA2 a more likely downstream target for adenosine. Previously, most CNGA2-related studies were carried out in olfactory sensory neurons, where CNGA2 form heterotetramers with 2 modulatory subunits CNGA4 and CNGB1.13 In the present study, we tested the hypothesis that CNG channels are involved in adenosine-induced Ca2+ influx in vascular endothelial cells. Our results demonstrated that CNG channels, especially CNGA2, mediate the adenosine-induced Ca2+ influx in these cells, and that CNG channels may play an important role in endothelium-dependent vasorelaxation in response to A2 adenosine receptor agonists.
Materials and Methods
Cell Culture and Ca2+ Dye Loading
H5V cells, which were derived from murine embryonic heart endothelium, and were grown in 90% DMEM and 10% FBS. Before experiments, the endothelial cells were loaded with 10 μmol/L Fluo-3/AM in a physiological saline solution (PSS) that contained in mmol/L: 140 NaCl, 5 KCl, 1 CaCl2, 10 glucose, 5 Hepes, pH 7.4.
Vessel Preparation and Ca2+ Dye Loading
The animal study was conducted in conformity with the Guide for animal Care and Use of Laboratory Animals published by the US National Institute of Health. For [Ca2+]i measurements, thoracic aorta from male C57 mice was cut into a small strip, and then mounted onto an experimental chamber with endothelial surface facing the objectives. The endothelial layer was then fluorescently loaded with 10 μmol/L Fluo-4/AM.
CNGA2-Specific siRNA and Transfection
The vector-based siRNA strategy was used. The CNGA2-specifc siRNA sequence was designed using Ambion siRNA Target Finder. A pair of inverted repeat sequences containing the 19-nt siRNA was then synthesized. The sequence for the strand 1 was 5′-TGGCAAAGATGACCACAGGTTCAAGAGACCTGTGGTCATCTTTGC-CATTTTTT-3′, and that for strand 2 was 5′-AATTAAAAA-ATGGCAAAGATGACCACAGGTCTCTTGAACCTGTGGTCAT-CTTTGCCAGGCC-3′. The CNGA2-specific nucleotides are underlined. These 2 strands were annealed and then cloned into a self-constructed siRNA expression vector pcDU6C. H5V cells were stably transfected with the pcDU6C containing either the CNGA2-specific siRNA or a control siRNA with scrambled sequence.
Briefly, whole-cell proteins were extracted. 100 μg proteins were loaded onto each lane and separated on a SDS/PAGE gel. Proteins were transferred to a polyvinylidene fluoride (PVDF) membrane, and blotted with the primary anti-CNGA2 (1:1000), anti-CNGA4 (2 μg/mL), or anti-A2B receptor antibody (1:100). The bound antibodies were detected with horseradish peroxidase-conjugated secondary antibody.
Briefly, the experimental chambers containing either cultured endothelial cells or isolated aortic strips were placed on the stage of an inverted microscope (Olympus IX81), and the [Ca2+]i fluorescence was measured using a FV1000 laser scanning confocal imaging system. Changes of [Ca2+]i were displayed as a ratio of fluorescence relative to the fluorescence before the application of adenosine or 8-BrcAMP (F1/F0).
[Ca2+]i rise was initiated by applying adenosine (100 μmol/L) to the cultured cells bathed in PSS. For aortic strips, the Fluo-4/AM-loaded vessel strips were bathed in a solution that contained in mmol/L: 118 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 KH2PO4, 25.2 NaHCO3, 11.1 glucose, pH 7.4, bubbled with 95% O2-5% CO2. When appropriate, inhibitors were added 10 minutes before the initiation of adenosine-induced Ca2+ influx. A slightly higher concentration of L-cis-diltiazem was used for the mouse aortic strips to overcome the relative poor penetration of the chemical in this tissue (100 μmol/L for tissue versus 50 μmol/L for H5V cells).
Whole-cell current was recorded using an EPC9 patch clamp amplifier (HEKA) in voltage-clamp mode. Patch pipette contained in mmol/L: 145 Na+-glutamate, 5 CsCl, 5 EGTA, 5 Hepes, pH 7.4. Bath solution contained in mmol/L: 145 Na+-glutamate, 5 Hepes, 10 glucose, pH 7.4. Adenosine (100 μmol/L) was applied to the cells held at −80 mV. Instantaneous I-V relationships were obtained by applying an ascending ramp protocol from −100 mV to +100 mV for 500 ms before and after adenosine application. L-cis-diltiazem and LY-83583 were added 10 minutes before adenosine application.
Arterial Tension Measurement
Segments of aorta were dissected from male C57 mice and mounted in a myograph. Isometric tension was measured. Cumulative concentration-response relationships for the relaxant effect of NECA were determined in aortic rings after steady contraction with 11-deoxy prostaglandin (PG) F2α. The concentration of 11-deoxy PGF2α varied from 60 to 300 nmol/L to achieve similar degree of constriction in different arteries. If needed, L-cis-diltiazem (100 μmol/L), L-NAME (100 μmol/L), or charybdotoxin (50 nmol/L) plus apamin (50 nmol/L) was added 30 minutes before NECA application. In some arterial rings, the endothelial layer was mechanically removed.
For detailed experimental procedures, please see the supplemental materials (available online at http://atvb.ahajournals.org).
Role of CNG Channels in Adenosine-Induced Ca2+ Influx in H5V Cells
Application of adenosine (100 μmol/L) elicited a [Ca2+]i rise in H5V cells bathed in PSS (Figure 1A). This [Ca2+]i rise was attributable to Ca2+ influx, because (1) removal of extracellular Ca2+ abolished the [Ca2+]i rise; (2) adenosine also induced Mn2+ influx as will be described later. We then explored the possible involvement of CNG channels in this Ca2+ influx. L-cis-diltiazem, a highly selective inhibitor for CNG channels,16 caused a dose-dependent inhibition on this Ca2+ influx (supplemental Figure I). L-cis-diltiazem at 50 μmol/L markedly reduced the percentage of cells displaying adenosine-induced Ca2+ influx (Figure 1D). This agent also decreased the magnitude of [Ca2+]i rise among the responding cells (Figure 1B and 1E). Another CNG channel inhibitor LY-8358317 (20 μmol/L) had similar effect to that of L-cis-diltiazem (Figure 1C through 1E).
Because both L-cis-diltiazem and LY-83583 inhibit multiple CNG isoforms, they cannot be used to differentiate specific CNG isoform(s) involved. Thus, siRNA strategy was used to study the possible involvement of CNGA2 in H5V cells. CNGA2 was chosen as the possible target based on its high sensitivity to cAMP13 and its expression in vascular endothelial cells.9,11,12 In immunoblot experiments, a CNGA2-specific antibody recognized a protein band with molecular size of ≈80 kDa, which correlates well with that of mouse CNGA2 in Genbank (NM_007724).18 Stable expression of a CNGA2-specific siRNA reduced the CNGA2 protein level by 80±2% (n=5) (Figure 2A and 2C). In contrast, stable expression of a control siRNA had no effect on the CNGA2 protein level (Figure 2A). The effect of CNGA2-siRNA was specific to CNGA2, because it had no effect on the expression of another CNG isoform CNGA4 (≈66 kDa, NM_001033317) (Figure 2B and 2C). Functionally, the CNGA2-specific siRNA not only reduced the percentage of cells displaying adenosine-induced Ca2+ influx (Figure 2F), but also decreased the magnitude of adenosine-induced [Ca2+]i rise among the responding cells (Figure 2E and 2G). In contrast, the control siRNA had no effect on this Ca2+ influx (Figure 2F and 2G). Mn2+ influx studies were performed to verify the results of Ca2+ influx studies. The adenosine-induced Mn2+ influx was also reduced by the CNGA2-specific siRNA but not by the control siRNA (supplemental Figure II). Taken together, these data strongly suggest that CNGA2 is the main channel responsible for adenosine-induced Ca2+ influx in H5V cells.
Role of CNG Channels in Adenosine-Induced Inward Current in H5V Cells
Whole cell patch clamp was used to verify the involvement of CNGA2 in the adenosine-induced responses. Application of adenosine (100 μmol/L) elicited an inward current in cells clamped at −80 mV (Figure 3A). Because the adenosine-induced inward current was transient (Figure 3A), an ascending voltage-ramp protocol was applied from −100 mV to +100 mV for 500 ms to obtain instantaneous I-V relationships before and after the adenosine application (Figure 3A and 3B). The results show that adenosine could stimulate the outward current at positive membrane potential and enhance the inward current at negative membrane potential (Figure 3B and 3G). More importantly, treatment with L-cis-diltiazem (50 μmol/L; Figure 3C, 3D, and 3H) and CNGA2-specific siRNA (Figure 3E, 3F, and 3H) diminished the adenosine-induced current in H5V cells. In contrast, the control siRNA had no effect (Figure 3H). These data agree with the results from the fluorescent Ca2+ influx experiments and support a role of CNGA2 channels in the adenosine-induced current.
Note that, even in the presence of L-cis-diltiazem, there are some background cation current in H5V cells (Figure 3D). It is likely that this background current might be contributed by other cation channels such as TRP channels.8
Involvement of Adenylyl Cyclases and A2B Receptors
Adenosine may act on A2 receptors to stimulate adenylyl cyclases, resulting in an increased cytosolic cAMP level.1,15 Adenosine may also stimulate guanylyl cyclases, thus elevating cytosolic cGMP level.14 Both cAMP and cGMP can activate CNGA2 channels. To test possible involvement of adenylyl cyclases and guanylyl cyclases, we used SQ-22536 and MDL-12330A, both of which inhibit adenylyl cyclases, and ODQ, which inhibits guanylyl cyclases. SQ-22536 caused a dose-dependent inhibition on adenosine-induced Ca2+ influx (supplemental Figure I). SQ-22536 at 300 μmol/L and MDL-12330A at 10 μmol/L drastically reduced the percentage of cells displaying adenosine-induced Ca2+ influx (11±4% [n=9] for SQ-22536-treated, 14±6% [n=7] for MDL-12330A-treated versus 88±3% [n=13] for the control cells), and they also markedly decreased the magnitude of [Ca2+]i rise among the responding cells (F1/F0 of 1.6±0.1 [n=9] for SQ-22536-treated, 1.5±0.2 [n=7] for MDL-12330A-treated versus 3.2±0.3 [n=13] for the control cells). In contrast, ODQ treatment (50 μmol/L) had no effect either on the percentage of cells displaying adenosine-induced Ca2+ influx (87±3% [n=6] versus 88±3% [n=13] for the control cells) or on the magnitude of [Ca2+]i rise among the responding cells (F1/F0 of 2.8±0.4 [n=6] versus 3.2±0.3 [n=13] for the control cells). These data suggest an involvement of adenylyl cyclases but not guanylyl cyclases.
Adenosine may act either on A2A or A2B receptors to stimulate adenylyl cyclases.15 An A2A-specific inhibitor SCH-58261,15 an A2B-specific inhibitor MRS-1754,15 and a relatively A1- and A2B-selective inhibitor 8-SPT19 were used to differentiate the involvement of A2A and A2B. Both MRS-1754 and 8-SPT caused a dose-dependent inhibition on adenosine-induced Ca2+ influx (supplemental Figure I). MRS-1754 at 50 nmol/L and 8-SPT at 1 μmol/L (a concentration that may only inhibit A2B and A1 receptors)19 drastically reduced the percentage of responding cells (41±7% [n=4] for MRS-1754-treated, 3±1% [n=4] for 8-SPT-treated versus 88±3% [n=13] for the control cells), and they also markedly decreased the magnitude of [Ca2+]i rise among the responding cells (F1/F0 of 2.0±0.1 [n=4] for MRS-1754-treated, 1.6±0.1 [n=4] for 8-SPT-treated versus 3.2±0.3 [n=13] for the control cells). In contrast, SCH-58261 had no effect (supplemental Figure I). The expression of A2B receptors in H5V cells was also confirmed by RT-polymerase chain reaction (PCR) and immunoblot experiments (supplemental Figure III). These data suggest an involvement of A2B but not A2A receptors.
Role of CNGA2 in 8-BrcAMP-Induced Ca2+ Influx and Cation Current in H5V Cells
Exogenous cAMP was used to further study the regulatory role of cAMP in CNGA2-mediated Ca2+ influx and cation current. Application of membrane-permeant 8-Br-cAMP (100 μmol/L) induced a [Ca2+]i rise (supplemental Figure IVA and IVB) and a cation current in H5V cells (supplemental Figure IVC through VE). Both the [Ca2+]i rise and the cation current were inhibited by the CNGA2-specific siRNA, L-cis-diltiazem and LY83583 (supplemental Figure IV).
Role of CNG Channels in the Primary Cultured BAECs
H5V is a cell line. There is a concern that endothelial phenotype may change during prolonged cell culture and cell passage conditions. Thus, we used the primary cultured BAECs to verify the above findings. Similar to H5V cells, adenosine induced a [Ca2+]i rise in BAECs (supplemental Figure V), and this [Ca2+]i rise was abolished in the presence of L-cis-diltiazem and LY-83583 (supplemental Figure V). Patch clamp experiments also showed that the adenosine-stimulated current was inhibited by L-cis-diltiazem and LY-83583 in BAECs (supplemental Figure VI).
Role of CNG Channels in Adenosine-Induced Endothelial Ca2+ Influx in Mice Aorta
About one fourth of the endothelial cells in mice aortic strips were found to display spontaneous [Ca2+]i transients, some cells with a single [Ca2+]i transient and others with repetitive [Ca2+]i oscillations. The rest (three fourths) cells were quiescent. Application of adenosine (100 μmol/L) elicited a transient [Ca2+]i rise in 12±2% (n=8) of endothelial cells, which were previously quiescent (Figure 4A through 4C). Some cells responded with a single [Ca2+]i transient whereas others responded with repetitive [Ca2+]i transients. This adenosine-induced endothelial [Ca2+]i rise was almost completely abolished in aortic strips that were pretreated with L-cis-diltiazem (100 μmol/L) for 10 minutes (Figure 4C). We also tested LY-83583 (20 μmol/L). However, LY-83583 itself greatly facilitated the spread of Ca2+ signals to neighboring cells presumably because of its action on gap junctions (n=4), thus this agent was not used further. Because of the difficulty in determining the exact cell boundary of individual endothelial cells in mouse aortic strips with our confocal microscope system, which significantly hampered our ability of accurately estimating the change in [Ca2+]i, we did not attempt to analyze the magnitude of [Ca2+]i rise. Instead, only the effect of L-cis-diltiazem on the percentage of responding cells was compared (Figure 4C).
One may query about the physiological importance of these adenosine-induced [Ca2+]i transients, because relatively low percentage of endothelial cells (12±2%, n=8) in isolated aortic strips responded to adenosine. However, note that the endothelial cells in situ are known to display heterogeneity, with different population of cells responding to different agonists.20 Therefore, it is expected that only certain percentage of cells can respond to adenosine. For comparison, in our experiment, a well-recognized vasoactive agonist acetylcholine (1 μmol/L) elicited [Ca2+]i transients in 21±5% (n=8) of the endothelial cells in aortic strips (Figure 4C), and furthermore this acetylcholine-induced [Ca2+]i rise is insensitive to L-cis-diltiazem (100 μmol/L) (Figure 4C).
Role of CNG Channels in A2 Adenosine Receptor-Mediated Vasorelaxation
We further examined the functional role of CNG channels in adenosine receptor-mediated vasorelaxation. In agreement with the results from other groups,21 NECA, a selective A2 adenosine receptor agonist, induced concentration-dependent vasorelaxation in mouse aortic segments preconstricted with 11-deoxy PG F2α (Figure 4D, supplemental Figure VIIA). The relaxation was mostly endothelium-dependent, because removal of the endothelium greatly reduced NECA-induced relaxation (Figure 4D). Importantly, L-cis-diltiazem (100 μmol/L) markedly inhibited the vasorelaxation to NECA in normal (endothelium-intact) aortic rings (Figure 4D, supplemental Figure VII), but it had no effect on the residual small relaxation to NECA in endothelium-denuded aortic rings (Figure 4D). These data suggest that CNG channels play a key role in the endothelium-dependent relaxation induced by NECA. We also explored the vasodilator(s) involved in NECA-induced relaxation. L-NAME (100 μmol/L) reduced the NECA-induced relaxation in endothelium-intact aortic rings to a level similar to that of endothelium-denuded rings (supplemental Figure VIIC). Furthermore, combined application of charybdotoxin (50 nmol/L) and apamin (50 nmol/L), a treatment considered to be a hallmark of EDHF response,7 had no effect on the NECA-induced relaxation (supplemental Figure VIIC). These results suggest that the NECA-induced vasorelaxation is mediated by NO but not EDHF.
Note that the adenosine-induced [Ca2+]i rise and cation current were transient (Figures 1A and 3⇑A), whereas the NECA-induced vasorelaxation lasted much longer (supplemental Figure VIIA). This is reasonable, because the vasorelaxation involves the Ca2+-NO-protein kinase G signaling pathway and the protein kinase G-mediated protein phosphorylation,22 which has a much long-lasting effect.
Adenosine is a key metabolite involved in metabolic hyperemia in many vascular beds including coronary circulation, cerebral circulation, and skeletal muscle circulation.1 The roles of adenosine in vasculature are most prominent during hypoxia, ischemia, and reactive hyperemia.1 One important action of adenosine is to induce endothelium-dependent vasorelaxation.1–4 It is previously shown that, at least in some arteries, such as skeletal muscle arteries, a rise in [Ca2+]i is required for the endothelium-dependent vasorelaxation in response to adenosine.3–5 In the present study, we explored the possible role of CNG channels in adenosine-induced Ca2+ influx in vascular endothelial cells. Our results show that the adenosine-induced Ca2+ influx was markedly reduced by CNG-specific inhibitors L-cis-diltiazem and LY-83583 in H5V cells and in the primary cultured BAECs. Whole cell patch clamp recorded an adenosine-induced current that was sensitive to L-cis-diltiazem in both cell types. Furthermore, a CNGA2-specific siRNA almost completely abolished the adenosine-induced Ca2+ influx current in H5V cells. Moreover, L-cis-diltiazem inhibited adenosine-induced endothelial [Ca2+]i rise in isolated mouse aortic strips, and it also markedly reduced the endothelium-dependent vasorelaxation to NECA, an A2 adenosine receptor agonist. It was also found that the NECA-induced vasorelaxation was mediated by NO but not by EDHF. These data provide compelling evidence that CNG channels, CNGA2 in particular, play a key role in adenosine-induced endothelial Ca2+ influx and subsequent vasorelaxation.
CNG channels are Ca2+-permeable nonselective cation channels activated by cyclic nucleotides cAMP and cGMP.13 It is previously documented that adenosine may increase cAMP and cGMP by stimulating adenylyl cyclases and guanylyl cyclases in endothelial cells.1,14,15 In the present study, adenosine-induced endothelial Ca2+ influx was drastically reduced by adenylyl cyclase inhibitors MDL-12330A and SQ-22536, but not affected by guanylyl cyclase inhibitor ODQ. Exogenous application of membrane-permeant 8-BrcAMP also induced Ca2+ influx. These data suggest that cAMP but not cGMP is the second messenger that links adenosine to CNGA2 activation. Adenosine may interact with 4 G protein-coupled receptors, A1, A2A, A2B, and A3. Stimulation of A1 and A3 receptors inactivates adenylyl cyclases, resulting in a decreased cAMP level, whereas stimulation of A2A and A2B receptors activates adenylyl cyclases, leading to an increased cAMP level.15 In the present study, the adenosine-induced Ca2+ influx was inhibited by MRS-1754 and 8-SPT but not by SCH-58261, suggesting an involvement of A2B but not A2A receptors. Taken together, it appears that adenosine binds to A2B receptors, causing activation of adenylyl cyclases, which elevate cytosolic cAMP level, leading to an increased activity of CNGA2 channels.
In 2 published functional studies, Zhang et al12 proposed that CNGA2 contributes to the store-operated Ca2+ influx induced by thapsigargin in pulmonary artery endothelial cells, whereas Wu et al11 suggested that CNGA2 mainly allows Na+ entry, resulting in membrane depolarization, which reduces the driving force for Ca2+ entry. In the present study, we showed for the first time that CNGA2 mediates adenosine-induced endothelial Ca2+ entry and vasorelaxation. This finding fits well with a unique property of CNGA2, ie, its higher sensitivity to cAMP than other CNG isoforms.13 Note that vascular endothelial cells in vivo are also the targets of several other physiologically important cAMP-elevating agents including calcitonin gene-related peptide,23 adrenomedullin,23 and adrenaline (through β-receptors).24 In future, it will be interesting to explore whether CNG channels also participate in the vasorelaxation in response to other cAMP-elevating agents.
Previous studies also found the expression of 2 other CNG isoforms, CNGA1 and CNGA4, in vascular endothelial cells.9–12 However, CNGA1 is unlikely to be involved in the adenosine-induced responses, because CNGA1 is not sensitive to cAMP.13 CNGA4 is not a functional subunit and cannot form a functional channel by itself. The role of CNGA4, if any, is to modulate the function of native endothelial CNG channels. There is a possibility that CNGA4 may form heteromeric channels with CNGA2 in endothelial cells as in the case for olfactory neurons.13 Further experiments are needed to clarify whether such an association between CNGA2 and CNGA4 indeed exists in endothelial cells.
In conclusion, we demonstrated that CNG channels, especially CNGA2, play a key role in adenosine-induced endothelial Ca2+ influx and subsequent vasorelaxation. It is likely that adenosine acts through A2B receptors and adenylyl cyclases to stimulate CNGA2 in endothelial cells.
Sources of Funding
This study was supported by Hong Kong Research Grant Council (CUHK4526/06m and Li Ka Shing Institute of Health Sciences.
Original received May 18, 2007; final version accepted February 11, 2008.
Wyatt AW, Steinert JR, Wheeler-Jones CP, Morgan AJ, Sugden D, Pearson JD, Sobrevia L, Mann GE. Early activation of the p42/p44MAPK pathway mediates adenosine-induced nitric oxide production in human endothelial cells: a novel calcium-insensitive mechanism. FASEB J. 2002; 16: 1584–1594.
Duza T, Sarelius IH. Conducted dilations initiated by purines in arterioles are endothelium dependent and require endothelial Ca2+. Am J Physiol Heart Circ Physiol. 2003; 285: H26–H37.
Govers R, Rabelink TJ. Cellular regulation of endothelial nitric oxide synthase. Am J Physiol Renal Physiol. 2001; 280: F193–F206.
Yao X, Garland CJ. Recent developments in vascular endothelial cell transient receptor potential channels. Circ Res. 2005; 97: 853–863.
Yao X, Leung PS, Kwan HY, Wong TP, Fong MW. Rod-type cyclic nucleotide-gated cation channel is expressed in vascular endothelium and vascular smooth muscle cells. Cardiovasc Res. 1999; 41: 282–290.
Zhang J, Xia SL, Block ER, Patel JM. NO upregulation of a cyclic nucleotide-gated channel contributes to calcium elevation in endothelial cells. Am J Physiol Cell Physiol. 2002; 283: C1080–C1089.
Kaupp UB, Seifert R. Cyclic nucleotide-gated ion channels. Physiol Rev. 2002; 82: 768–824.
Fredholm BB, IJzerman AP, Jacobson KA, Klotz KN, Linden J. International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. Pharmacol Rev. 2001; 53: 527–552.
Leinders-Zufall T, Rand MN, Shepherd GM, Greer CA, Zufall F. Calcium entry through cyclic nucleotide-gated channels in individual cilia of olfactory receptor cells: spatiotemporal dynamics. J Neurosci. 1997; 17: 4136–4148.
Bonigk W, Bradley J, Muller F, Sesti F, Boekhoff I, Ronnett GV, Kaupp UB, Frings S. The native rat olfactory cyclic nucleotide-gated channel is composed of three distinct subunits. J Neurosci. 1999; 19: 5332–5347.
Gao Z, Li BS, Day YJ, Linden J. A3 adenosine receptor activation triggers phosphorylation of protein kinase B and protects rat basophilic leukemia 2H3 mast cells from apoptosis. Mol Pharmacol. 2001; 59: 76–82.
Huang TY, Chu TF, Chen HI, Jen CJ. Heterogeneity of [Ca2+]i signaling in intact rat aortic endothelium. FASEB J. 2000; 14: 797–804.
Lincoln TM, Dey N, Sellak H. cGMP-dependent protein kinase signaling mechanisms in smooth muscle: from the regulation of tone to gene expression. J Appl Physiol. 2001; 91: 1421–1430.
Guimaraes S, Moura D. Vascular adrenoceptors: an update. Pharmacol Rev. 2001; 53: 319–356.