Involvement of gap junctions between smooth muscle cells in sustained hypoxic pulmonary vasoconstriction development: a potential role for 15-HETE and 20-HETE

Abstract

In response to hypoxia, the pulmonary artery normally constricts to maintain optimal ventilation-perfusion matching in the lung, but chronic hypoxia leads to the development of pulmonary hypertension. The mechanisms of sustained hypoxic pulmonary vasoconstriction (HPV) remain unclear. The aim of this study was to determine the role of gap junctions (GJs) between smooth muscle cells (SMCs) in the sustained HPV development and involvement of arachidonic acid (AA) metabolites in GJ-mediated signaling. Vascular tone was measured in bovine intrapulmonary arteries (BIPAs) using isometric force measurement technique. Expression of contractile proteins was determined by Western blot. AA metabolites in the bath fluid were analyzed by mass spectrometry. Prolonged hypoxia elicited endothelium-independent sustained HPV in BIPAs. Inhibition of GJs by 18β-glycyrrhetinic acid (18β-GA) and heptanol, nonspecific blockers, and Gap-27, a specific blocker, decreased HPV in deendothelized BIPAs. The sustained HPV was not dependent on Ca²⁺ entry but decreased by removal of Ca²⁺ and by Rho-kinase inhibition with Y-27632. Furthermore, inhibition of GJs decreased smooth muscle myosin heavy chain (SM-MHC) expression and myosin light chain phosphorylation in BIPAs. Interestingly, inhibition of 15- and 20-hydroxyeicosatetraenoic acid (HETE) synthesis decreased HPV in deendothelized BIPAs. 15-HETE- and 20-HETE-stimulated constriction of BIPAs was inhibited by 18β-GA and Gap-27. Application of 15-HETE and 20-HETE to BIPAs increased SM-MHC expression, which was also suppressed by 18β-GA and by inhibitors of lipoxygenase and cytochrome P450 monooxygenases. More interestingly, 15,20-dihydroxyeicosatetraenoic acid and 20-OH-prostaglandin E₂, novel derivatives of 20-HETE, were detected in tissue bath fluid and synthesis of these derivatives was almost completely abolished by 18β-GA. Taken together, our novel findings show that GJs between SMCs are involved in the sustained HPV in BIPAs, and 15-HETE and 20-HETE, through GJs, appear to mediate SM-MHC expression and contribute to the sustained HPV development.

hypoxic pulmonary vasoconstriction (HPV) is the acute local process in the pulmonary artery tree that optimizes pulmonary ventilation-to-perfusion matching in response to alveolar hypoxia diverting blood away from poorly ventilated regions of the lung (46). It is a unique adaptive mechanism of the pulmonary circulation, since under the same conditions, systemic arteries demonstrate a vasodilatation that improves blood supply and substrate delivery to hypoxic tissue (46). In pathological conditions associated with global hypoxia, however, such as in respiratory disease or at altitude, HPV causes a detrimental increase in pulmonary vascular resistance and pulmonary artery pressure leading to increased afterload on the right heart (46). However, despite many years of extensive research, the precise mechanisms underlying HPV remain unresolved (46).

In intrapulmonary arteries of mammals, the contractile response to hypoxia typically is biphasic, characterized by a rapid transient vasoconstriction that is followed by a slowly developing sustained contraction (4, 16, 18, 19, 32, 34). It has been shown that in a majority of animal species the sustained phase of HPV in intrapulmonary arteries has endothelium dependency (18, 19, 32, 34). In contrast, the sustained hypoxic vasoconstriction in bovine intrapulmonary arteries (BIPAs) is endothelium independent (16, 24).

Coordination of cellular responses in vascular tissues is mediated by multiple signaling mechanisms, including intercellular communications. There is growing evidence that direct communication within and between endothelial cells (ECs) and smooth muscle cells (SMCs) through gap junctions (GJs) is an important modulator of tone at all levels of the vascular tree and essential in the control and coordination of normal vascular function (12). GJs are composed of intercellular channel clusters that directly connect the cytoplasm of adjacent cells, allowing the direct passage of electrical current and small signaling molecules up to 1.2 kDa between adjacent cells (12). GJs consist of connexins (Cxs), a family of proteins that form channels linking the cytoplasm of adjacent cells (12). In blood vessels, including pulmonary artery, in both ECs and SMCs, GJs are made up of three Cx molecules, namely Cx37, Cx40, and Cx43, designated according to molecular mass in kiloDaltons (5, 6, 31). In the arterial wall, the endothelial cell layer is coupled to the inner layer of the media by myoendothelial gap junctions (MEGJs) that form at the end of thin cytoplasmic projections from the ECs (9). The electrical and biochemical properties of gap junction channels are regulated by phosphorylation and dephosphorylation of Cxs (12). It has been shown that GJs in pulmonary vascular wall are important for its remodeling and reactivity changes under chronic hypoxia (5, 22) and MEGJs are involved in the sustained HPV development in rodent models (30, 49). However, such a contribution of GJs to the sustained HPV phase in BIPAs is not yet reported.

Eicosanoids are vasoactive compounds produced by the metabolism of arachidonic acid (AA) (52). While products of the cyclooxygenase (COX) pathway can modulate vascular function, there is no clear evidence supporting their role as a mediator of HPV (46). 20-Hydroxyeicosatetraeonic acid (20-HETE) synthesis is inhibited by hypoxia and inhibition of 20-HETE synthesis in the rabbit lungs promotes HPV (46). On the other hand, hypoxia activates 15-lipoxygenase (15-LOX) and increases 15-hydroxyeicosatetraenoic acid (15-HETE). Increased 15-HETE elicits constriction of neonatal rabbit pulmonary arteries (53). COX-derived eicosanoids bind to G_q-coupled receptors and, thereby, increase inositol-1,4,5-trisphosphate (IP₃) and intracellular calcium concentration (2). However, to the best of our knowledge, no receptors have yet been identified for the 15-HETE. HETEs are highly lipophilic molecules and small enough to diffuse through the GJs and/or activate GJ signaling between SMCs. These observations led to our hypothesis that GJ-mediated communications between SMCs triggered by hypoxia-induced 15-HETE and 20-HETE could be involved in the sustained HPV development in BIPAs.

METHODS

Vascular tone measurements and hypoxic protocol.

Third-order (0.3- to 0.5-mm inner diameter) and second-order (1.0- to 2.0-mm inner diameter) pulmonary artery branches were harvested from bovine right and left lungs purchased from a local slaughterhouse. After the heart and lungs were quicly excised from the animal, they were placed in normal Tyrode solution (in mM: 135 NaCl, 5.4 KCl, 1.8 CaCl₂, 1.0 MgCl₂, 5 HEPES, and 11 glucose; pH was adjusted to 7.40 with NaOH) and transported to the laboratory on ice. Isolated pulmonary arterial rings, with and without intact endothelium, were dissected from bovine lungs, cleaned of connective tissue and studied for changes in isometric force. BIPA rings (2 mm in length) were then mounted to force transducers in individually thermostated (37°C) 10-ml baths (Metro Scientific) at an optimal passive tension of 5 g in Krebs bicarbonate buffer solution (pH 7.4) containing the following (in mM): 118 NaCl, 4.7 KCl, 1.5 CaCl₂, 25 NaHCO₃, 1.1 MgSO₄, 1.2 KH₂PO₄, and 5.6 glucose, and gassed with 21% O₂-5% CO₂-5% N₂ (normoxia, Po₂ ∼140 Torr). After 40 min of incubation, arterial preparation viability and equilibration were assessed from the response to 120 mM KCl (120K). A brief depolarization of BIPAs with 120K also increases the reproducibility of subsequent vascular responses. In experiments with endothelium-denuded arterial rings, the endothelium was disrupted by gently rubbing the luminal surface of the BIPA rings. To elicit a full contractile response to hypoxia, the arteries were precontracted with 30 mM KCl (30K) to induce a constriction of ∼60% of that to 120K before and during the hypoxic challenge. Hypoxia was induced by gassing of 30K-containing bicarbonate buffer solution with 95% N₂-5% CO₂ (Po₂ = 20–40 Torr) for 11–14 h. All drugs and chemical compounds were from Sigma-Aldrich (Milwaukee, WI) and Cayman Chemical (Ann Arbor, MI).

Western blot analysis.

Proteins were extracted from frozen tissue in lysis buffer, after which Western blot analysis using the specific antibodies indicated in the individual figures was performed as described previously (16). We measured expression of smooth muscle-myosin heavy chain (SM-MHC), myocardin, phosphorylated myosin light chain (P-MLC), and total MLC. Normalization was done by either using α-tubulin or β-actin. Using ImageJ software, we did densitometry analysis.

Mass spectrometry.

Third-ordered BIPAs were incubated in Krebs buffer (3 ml) after giving them a passive tension in the presence or absence of 20-HETE. 20-HETE and its metabolites in the Krebs buffer were extracted using the solid phase extraction method as described (50). The extracted metabolites were reconstituted in methanol (20 μl) and subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses using a Q-trap 3200 linear ion trap quadrupole LC-MS/MS spectrometer equipped with a Turbo V ion source operated in negative electrospray mode (Applied Biosystems, Foster City, CA). The mobile phase used was A: acetonitrile/methanol/water/acetic acid (30/10/60/0.01) and B: acetonitrile/methanol/acetic acid (90/10/0.01). The high-performance liquid chromatography gradient started with 100% A to 96% A in 2.5 min at a flow of 0.3 ml/min and then changed to 20% A from 2.5 to 7 min at a flow of 0.4 ml/min.

Fluorescent 20-HETE analogs in bovine pulmonary artery endothelial cells.

Bovine pulmonary artery endothelial cells passages 2–5 were primarily isolated and cultured as previously described by us (28). They were plated on glass coverslides. When subconfluent, they were rinsed with serum-free media before incubation for 5 min at 37° with the following test compounds: 1) serum-free media, 2) EtOH {vehicle for labeled 20-HETE agonist; 20-hydroxy-N-((4-(4-((7-nitrobenzo[c][1,2,5]oxadiazol-4-yl)amino)butoxy)phenyl)sulfonyl)eicosa-5(Z),14(Z)-dienamide [20-HEDE-NBD]}, 3) 10 or 50 μM short-chain fatty acid labeled with NBD (hexanoyl-NBD), 4) 10 or 50 μM HEDE-NBD, or 5) coverslide without cells with NBD labeled 20-HETE analog. After being rinsed with PBS twice more, cells were imaged using fluorescence microscopy with excitation wavelengths of 490 and emission wavelengths of 510. Exposures were for 1.5 or 3 s.

Statistical analysis.

For registration of vascular ring contractile activity and its following analysis, LabChart 8 (ADInstruments) software was used. Tension is presented as a percentage of the maximum constriction obtained to the exposure to 120K during the equilibration procedure. Results are expressed as means ± SE of these measurements. Student's t-test for paired or unpaired data were made, as appropriate, to evaluate statistical significance. Differences were assumed to reach statistical significance if the confidence range was no <95% or P < 0.05.

RESULTS

The sustained HPV in isolated BIPAs is endothelium-independent.

Prolonged hypoxia (Po₂ = 20–40 Torr) during 13 h elicited a constrictor response in third-order (Fig. 1, A and B) and second-order (Fig. 1C) BIPAs precontracted with 30K. Typically, HPV in BIPAs consisted of a transient dilatation followed by a sustained slowly developing contraction (the sustained HPV). Occasionally, the rapid transient HPV peaked within ∼30 min of the onset of hypoxia, which was observed only in third-order BIPAs. A typical record of HPV in third-order BIPAs is shown in Fig. 1A. To evaluate the involvement of the endothelium in HPV development, we tested the effect of prolonged hypoxia on the tone of endothelium-denuded arterial rings. Mechanical removal of the endothelium in third- and second-order BIPAs had no effect on hypoxia-induced BIPA contractions (Fig. 1, B and C).

Fig. 1.

Effect of prolonged hypoxia on tone of endothelium-denuded and intact bovine intrapulmonary arteries (BIPAs). A: recording of endothelium-intact BIPAs precontracted with 30 mM KCl (30K) and exposed to hypoxia (40 Torr) or normoxia (140 Torr). B: summarized data showing the effect of prolonged hypoxia on tone of third-order BIPAs precontracted with 30K (E+, n = 13, and E−, n = 13 from 8 animals). C: effect of prolonged hypoxia on tone of second-order BIPAs precontracted with 30K (E+, n = 4 from 4 animals, and E−, n = 7 from 7 animals). Each symbol represents the mean ± SD. E+, endothelium intact; E−, endothelium removed.

The sustained HPV development in BIPAs requires intracellular Ca²⁺ release and Ca²⁺ sensitization but not extracellular Ca²⁺ entry activation.

The aim of the next set of experiments was to determine Ca²⁺ sources that are involved in the mechanisms of the sustained HPV development in BIPAs. Selective voltage-dependent L-type Ca²⁺ channel inhibitors verapamil (10 μM) and nitrendipine (100 nM) had no effect on the sustained HPV development in endothelium-denuded third-order BIPAs (Fig. 2, A and B). Nonspecific Ca²⁺ entry inhibition with CdCl₂ (1 mM) and LaCl₃ (1 mM) also had no effect on the amplitude of the sustained HPV (Fig. 2, C and D). All Ca²⁺ entry inhibitors suppressed vasoconstriction evoked by 30K but had no effect on the sustained HPV development.

Fig. 2.

Effect of Ca²⁺ entry inhibition, Ca²⁺-free Krebs buffer solution, and Rho-associated coiled-coil kinase (ROCK) inhibition on the sustained hypoxic pulmonary vasoconstriction (HPV) development in endothelium-denuded (E−) third-order BIPAs precontracted with 30K. A: effect of verapamil (10 μM, n = 6 from 4 animals) compared with control (n = 6 from 4 animals). B: effect of nitrendipine (100 nM). C: effect of CdCl₂. D: effect of LaCl₂. In B, C, and D, n = 7 from 7 animals for all treatments and control. E: sustained HPV development in containing BAPTA (200 μM) and EGTA (100 μM) Ca²⁺-free Krebs buffer solution (n = 4 from 4 animals) compared with control (n = 4 from 4 animals). F: effect of ROCK inhibition with Y-27632 (1 μM, n = 4 from 4 animals) compared with control (n = 25 from 9 animals). *P < 0.05.

To evaluate the dependency of the sustained HPV development on Ca²⁺ we further studied the effect of prolonged hypoxia on BIPAs in Ca²⁺-free Krebs buffer containing 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA; 200 μM) and ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid tetrasodium salt (EGTA; 100 μM) to scavenge extra- and intracellular Ca²⁺. The data in Fig. 2E demonstrate that the amplitude of the sustained HPV in endothelium-denuded third-order BIPAs was significantly decreased in the absence of Ca²⁺.

Rho-associated coiled-coil kinase (ROCK) is the main kinase responsible for increasing Ca²⁺ sensitivity of contractile proteins (Ca²⁺ sensitization) in vascular SMCs (48). Y-27632 (1 μM), a selective inhibitor of ROCK, significantly suppressed the amplitude of the sustained HPV in endothelium-denuded third-order BIPAs (Fig. 2F).

GJs are involved in the sustained HPV development in endothelium-denuded isolated BIPAs.

It has been shown previously that MEGJs are involved in the sustained HPV development in rodent intrapulmonary artery (30, 49). Since the sustained HPV in BIPAs is endothelium-independent, we have postulated that signaling via GJs between SMCs can be involved in the sustained HPV mechanisms in BIPAs. To determine this, we tested the effect of GJ inhibitors on HPV development in endothelium-denuded isolated BIPAs. Exposure to a nonselective inhibitor of GJs, 18β-glycyrrhetinic acid (18β-GA; 30 μM), significantly suppressed the amplitude of HPV in both endothelium-denuded third- and second-order BIPAs (Fig. 3, A and B). Another nonselective GJ inhibitor, heptanol (3.5 mM), also significantly reduced HPV amplitude in endothelium-denuded third-order BIPAs (Fig. 3C). Similarly, a highly selective GJ inhibitor, connexin mimetic peptide Gap-27 (100 μM), applied 1 h before the hypoxia exposure, significantly decreased HPV in third-order BIPAs (Fig. 3D). It is noteworthy that 18β-GA (30 μM) did decrease 30K-induced BIPAs contraction in normoxia (control: 49 ± 11% vs 18β-GA: 46 ± 18%).

Fig. 3.

Effect of gap junction (GJ) inhibition on the sustained HPV development in endothelium-denuded BIPAs. A: effect of 18β-glycyrrhetinic acid (18β-GA; 30 μM, n = 7 from 5 animals) on HPV in deendothelized (E−) second-order BIPAs precontracted with 30K (control, n = 8 from 8 animals). B: effect of 18β-GA (23 from 9 animals) on HPV in E− third-order BIPAs (control, n = 25 from 9 animals). C: effect of heptanol (3.5 mM, n = 10 from 4 animals) on HPV in E− third-order BIPAs (control, n = 7 from 4 animals). D: effect of Gap-27 (100 μM, n = 5 from 5 animals) on HPV in E− third-order BIPAs (control, n = 11 from 8 animals). *P < 0.05.

GJ inhibition decreased SM-MHC expression.

MEGJs mediate endothelial cell-derived serotonin-signals required for maintaining transforming growth factor β (TGFβ)-dependent SMCs differentiation (20, 21). Therefore, we also measured SM-MHC and myocardin expression levels in the endothelium-denuded third-order BIPAs. While 18β-GA (30 μM) did not change myocardin (Fig. 4A), it decreased SM-MHC (Fig. 4B) expression in BIPAs exposed to hypoxia. Gap-27 (100 μM) also decreased SM-MHC expression (Fig. 4C) and myosin light chain (MLC) phosphorylation (Fig. 4D).

Fig. 4.

Effect of GJ inhibition on expression of smooth muscle-myosin heavy chain (SM-MHC), myocardin, and phosphorylation of myosin light chain (MLC) in the endothelium-denuded third-order BIPAs. A and B: effect of 18β-glycyrrhetinic acid (18β-GA, 30 μM) on myocardin (n = 10 from 5 animals) and SM-MHC expression (n = 10 from 5 animals). C and D: effect of Gap-27 (100 μM) on SM-MHC expression (n = 10 from 5 animals) and on MLC phosphorylation (n = 10 from 5 animals). NS, not significant (P > 0.05).

AA metabolites contribute to the sustained HPV development in endothelium-denuded isolated BIPAs.

AA metabolites contribute to HPV development (46). Therefore, in the next set of experiments, we used inhibitors of different AA metabolic pathways to determine the contribution of AA products to the sustained HPV development in precontracted endothelium-denuded third-order BIPAs. The sustained HPV development was significantly decreased by 1) 15-lipoxygenase inhibitor 1 (15-LOXI1, 25 μM; Fig. 5A), which prevents 15-HETE formation from AA (3, 13); 2) DDMS (1 μM; Fig. 5B), which inhibits 20-HETE synthesis from AA by the cytochrome P450-CYP4A2 (33); and 3), indomethacin (1 μM; Fig. 5C), COX inhibitor that inhibits synthesis of eicosanoids from AA (25).

Fig. 5.

Effect of arachidonic acid metabolism inhibitors on the sustained HPV development in endothelium-denuded (E−) third-order BIPAs precontracted with 30K compared with control (n = 25 from 9 animals) and effect of 18β-GA (30 μM, 23 from 9 animals). A: effect of 15-lipoxygenase inhibitor 1 (15-LOXI1; 25 μM, n = 6 from 6 animals). B: effect of DDMS (1 μM, n = 6 from 6 animals). C: effect of indometacin (1 μM, n = 6 from 6 animals). *P < 0.05.

15-HETE and 20-HETE inhibitors decreased SM-MHC expression in endothelium-denuded BIPAs.

To determine whether eicosanoids play a role in maintaining the expression of SM-MHC in BIPAs exposed to hypoxia, we first tested the effect of 15-LOXI1 and DDMS on SM-MHC expression levels for 12 h. Both 15-LOXI1 and DDMS significantly reduced the expression of SM-MHC in BIPAs (Fig. 6). Indomethacin did not affect SM-MHC expression (Fig. 6).

Fig. 6.

Changes in SM-MHC expression in precontracted with 30K endothelium-denuded third-order BIPAs. A: representative Western blots showing the effect of 15-LOXI1 (25 μM), indomethacin (Indo; 1 μM), and DDMS (1 μM) on SM-MHC expression. B: summary of 15-LOXI, Indo, and DDMS on SM-MHC expression is shown (n = 10 from 5 animals).

GJ inhibition suppressed vasoconstrictor effect of 15-HETE and 20-HETE in endothelium-denuded BIPAs.

Because AA metabolites produced in vascular tissue contribute to the sustained HPV development in BIPAs, and their molecular size is small enough (<1.2 kDa) to pass through GJs, we speculated that AA metabolites transmitted between SMCs contribute to the sustained HPV. Alternatively, AA metabolites are lipophilic and hence they can embed in lipid layers and activate GJ signaling. We tested whether vasoconstrictor responses of BIPAs to some AA metabolites can be affected by GJ inhibition in SMCs. Four different AA metabolites with closely related structure, 5-, 12-, 15-, and 20-HETE, all evoked transient contraction of endothelium-denuded third-order BIPAs (Fig. 7, A–D), but contractions were most prominent in response to 15- and 20-HETE (Figs. 7, A and B). As represented in Fig. 7A, the contractile response of BIPAs to 15-HETE (1 nM) was significantly reduced by 18β-GA (30 μM) and Gap-27 (100 μM). Similarly, the contraction evoked by 20-HETE (1 μM) was also significantly inhibited by 18β-GA (Fig. 7B). In contrast, BIPA contraction evoked by 5-HETE and 12-HETE was unaffected by 18β-GA (Fig. 7, C and D). Additionally, we also tested the effect of 18β-GA on angiotensin-II induced contraction of BIPAs and, as expected, found that angiotensin-II induced contraction of BIPAs was not reduced by GJ inhibition (Fig. 7E).

Fig. 7.

Effect of GJ inhibition on the vasoconstrictor effects of arachidonic acid metabolites in endothelium-denuded (E−) third-order BIPAs. A: effect of 18β-GA (30 μM, n = 6 from 6 animals) and GAP-27 (100 μM, n = 3 from 3 animals) on constriction evoked by 15-hydroxyeicosatetraenoic acid (15-HETE; 1 nM, control, n = 6 from 6 animals). B: effect of 18β-GA (30 μM, n = 6 from 6 animals) on constriction evoked by 20-HETE (1 μM, control, n = 5 from 5 animals). C: effect of 18β-GA (30 μM, n = 6 from 6 animals) on constriction evoked by 5-HETE (1 μM, control, n = 6 from 6 animals). D: effect of 18β-GA (30 μM, n = 4 from 4 animals) on constriction evoked by 12-HETE (1 μM, control, n = 4 from 4 animals). B: effect of 18β-GA (30 μM, n = 5 from 5 animals) on constriction evoked by angiotensin II (Ang-2; 1 μM, control, n = 5 from 5 animals). *P < 0.05.

Fluorophore-tagged 20-HETE analog is localized in the cytoplasm.

Bright-field images confirmed cells in the area of imaging (Fig. 8A). We observed cytoplasmic fluorescence sparing the nuclear areas on coverslips of cells labeled with 20-HEDE-NBD, a stable 20-HETE analog covalently linked to NBD (Fig. 8, Bi and Bii). Comparable cytoplasmic fluorescence was not observed using other (control) treatments including labeling with a short-chain ester of the linker-NBD fluorophore moiety, hexanoyl-NBD (Fig. 8, Ci and Cii). Data are shown for cells incubated with 50 uM 20-HEDE-NBD; a similar distribution with less intense fluorescence was observed with cells incubated with 10 uM compound.

Fig. 8.

A: bright-field image confirmed cultured cells in the field despite no visible fluorescence. Bi: structure of 20-HETE analog. Bii: cultured endothelial cells on coverslips were incubated with 20-HEDE-NBD and then imaged with fluorescence miscropscopy. Shown are endothelial cells with cytoplasmic fluorescence with a perinuclear sparing. Ci: cells incubated with hexanoyl-NBD, a short chain fatty acid with the same linker and fluorophore, label showed no fluorescence. Not shown are control studies with cells incubated with labeled 20-HETE carrier (ethanol) or coverslips without cells incubated with NBD labeled 20-HETE analog.

20-HETE is converted to 15,20-DiHETE and 15-LOXI1 inhibits 20-HETE-mediated contractions in the endothelium-denuded BIPAs.

Since the dose of 20-HETE required to contract BIPAs was 1,000-fold higher than that of 15-HETE, this gave us a cue that 20-HETE was indirectly acting on the BIPA. 20-HETE is a substrate for COX (44), is converted to 20-OH-PGE₂ (29), and can be a potential substrate for lipoxygenases (26). To test a possibility that 20-HETE is metabolized in BIPAs, we performed mass spectrometry analysis to determine whether 20-HETE-derived 15-LOX-dependent 15,20-DiHETE and COX-dependent 20-OH-PGE₂ were generated by BIPAs in the presence and absence of 20-HETE. Characteristics of multiresidue methods (MRMs) used for 20-HETE and 20-OH-PGE₂ quantitation were 319.2/245.2 and 367/287.3, respectively, based on the optimization results from standard chemicals (Cayman Chemical). For identification of the proposed 15,20-DiHETE, two MRMs, 335.2/219.2 and 335.2/287.2, were set up based on its chemical structure and the characteristic fragmentation pattern for 15-HETE and 20-HETE, respectively. 20-OH-PGE₂ with a MRM of 367/287.3 was increased over 22-fold with the addition of 20-HETE (1 μM) to the incubation (0.37 ± 0.19 to 8.32 ± 2.10 ng, n = 4; Fig. 9, A and B). Additionally, both of the characteristic MRMs proposed for 15,20-DiHETE were detected in incubations with 20-HETE with a peak retention time at 4.5 min, while the MRMs from control incubations showed only background noise (Fig. 9, C and D). The amount of 15,20-DiHETE produced could not be accurately calculated without a standard, but the signal intensity measured as count per second for 335.2/219.2 was about half of the MRM for 20-OH-PGE₂, suggesting that the yield for 15,20-DiHETE was in the same level as for the production of 20-OH-PGE₂. Interestingly, production of 20-OH-PGE₂ was completely blocked by 18β-GA (Fig. 9E). Along these lines, the 15,20-DiHETE levels, relative to internal standard 8-iso-PGF2α [15,20-DiHETE-to-8-iso-PGF2α ratio], were also significantly reduced 0.47 (20-HETE) and 0.16 (20-HETE + 18β-GA) by the GJ inhibitor 18β-GA (Fig. 9F). If this observation is correct, then a 15-LOX inhibitor should block the biological actions of 20-HETE. Therefore, we tested whether 15-LOXI1 blocks the 20-HETE-mediated contraction of BIPAs. 15-LOXI1 (25 μM) blocked the 20-HETE (1 μM)-mediated contraction of BIPAs (20-HETE: 22 ± 5% and 20-HETE+15-LOXI: 5 ± 1%; P < 0.05; n = 5–7).

Fig. 9.

Mass spectrometry detection of 15,20-DiHETE and 20-OH-PGE₂ and the effect of 15-LOXI1 on 20-HETE mediated contraction of the endothelium-denuded (E−) third-order BIPAs. A: metabolites of 20-HETE are shown. B: tracing showing production of 20-OH-PGE₂ in the presence of 20-HETE. C and D: 15,20-DiHETE could be detected in the tissue bath with endothelium-denuded third-order BIPAs when 20-HETE (1 μM) was applied to the BIPAs. 15,20-DiHETE was not detectable in the absence of 20-HETE. E and F: 20-OH-PGE₂ and 15,20-DiHETE-to-8-iso-PGF_2α ratio decreased by the GJ inhibitor 18β-GA. ND, not detected.

18β-GA inhibited 15-HETE-induced expression of SM-MHC in endothelium-denuded BIPAs.

Additionally, we applied 15-HETE (1 nM, Fig. 10A) and 20-HETE (1 μM; Fig. 10B) to endothelium-denuded third-order BIPAs exposed to normoxia for 12 h and then determined SM-MHC levels. Both eicosanoids significantly increased SM-MHC expression in BIPAs. The GJ inhibitor 18β-GA (30 μM) decreased (P < 0.05) 15-HETE-, but not 20-HETE-, induced SM-MHC expression (Fig. 10, A and B). Since activation of the Ca²⁺-sensitive ROCK and nuclear factor of activated T-cells (NFAT) pathways increase SM-MHC expression in vascular smooth muscle cells (45, 47), we examined whether ROCK and NFAT pathways mediated 15-HETE-elicited SM-MHC expression. Interestingly, Y27632 (1 μM; Fig. 10C), a ROCK inhibitor (27), did not prevent 15-HETE-induced expression of SM-MHC. However, l-methionyl-l-alanylglycyl-l-prolyl- l-histidyl-l-prolyl-l-valyl-l-isoleucyl- l-valyl-l-isoleucyl-l-threonylglycyl-l- prolyl-l-histidyl-l-α-l-glutamic acid (NFATI; 100 nM; Fig. 10D), a selective inhibitor of calcineurin-mediated dephosphorylation of NFAT (43), prevented 15-HETE-induced expression of SM-MHC.

Fig. 10.

Changes in SM-MHC expression in precontracted with 30K endothelium-denuded third-order BIPAs. A and B: effect of 18β-GA (30 μM) on 15-HETE (1 nM) and 20-HETE (1 μM) induced SM-MHC expression (n = 10 from 5 animals). C and D: effect of Y27632 (1 μM) and NFATI (100 nM) on 15-HETE (1 nM) induced SM-MHC expression (n = 10 from 5 animals).

DISCUSSION

The salient findings of the present study are 1) GJ signaling between SMCs is mediated at least partly by AA metabolites, viz., 15-HETE, 20-HETE; and 2) novel derivatives of 20-HETE, contribute to the (i) sustained HPV development and the (ii) expression of SM-MHC in BIPAs.

HPV in isolated vessels of different animal species is usually characterized by an initial transient vasoconstriction phase followed by a slowly developing sustained phase, which is attributed to the development of the hypoxic pulmonary hypertension and right heart failure (46). As it has been reported previously (16, 24) and confirmed in the present study, in contrast to the majority of animal species (18, 19, 32, 34), the sustained phase of HPV in BIPAs is not dependent on signals from the endothelium. Since the removal of the endothelium in third and in second-order BIPAs had no effect on the sustained HPV development, we propose that endothelium-independence of HPV can be the specific feature of whole BIPA tree.

Studies have shown that the sustained HPV in rats is dependent on Ca²⁺ release from intracellular ryanodine-sensitive stores (18, 40) and ROCK-mediated Ca²⁺ sensitization (30, 40, 41, 48) but not extracellular Ca²⁺ entry (1, 18, 42), in SMCs. Consistently, we found that the sustained HPV in BIPAs is ROCK-mediated Ca²⁺ sensitization dependent and extracellular Ca²⁺ entry independent.

MEGJs play an important role in the sustained HPV development in rodents by facilitating myofilament Ca²⁺ sensitization in intrapulmonary artery SMCs (30, 49). Interestingly, this involves transfer of yet unidentified small signaling molecules from endothelial cells to SMCs, rather than electrical coupling (30). In contrast to rodents, GJ signaling between SMCs contributed to the sustained HPV development in BIPAs. Commonly used nonspecific, 18β-GA and heptanol, as well as highly selective, Gap-27, GJ inhibitors significantly reduced the sustained HPV amplitude, SM-MHC expression, and phosphorylation of MLC in BIPAs. A previous report suggested that GJ signals between rat pulmonary artery endothelial cells and SMCs increase expression of SM-MHC and maintain the contractile phenotype of SMCs (20). Since our data are obtained in endothelium-denuded vessels, they suggest that GJ signaling between SMCs, but not MEGJs, contributed to increased SM-MHC expression in BIPAs exposed to prolonged hypoxia.

Gap-27 is a synthetic connexin-mimetic peptide (SRPTEKTIFII), which is an analog of the extracellular loop 2 of Cx37 and Cx43, and selectively inhibits GJs that contain these Cxs (5). In pulmonary arterial wall, both in the endothelium and SMCs, Cx37, Cx40, and Cx43 are expressed (5, 6, 31), but Cx43 is the predominant Cxs isoform in SMCs layer of major arteries (17) and its expression levels have been shown to increase in rat intrapulmonary artery in chronic hypoxia (5). Since Gap-27 decreased the sustained HPV in endothelium-denuded BIPAs, we hypothesize that Cx37/Cx43 GJ-mediated signaling in SMCs may contribute to intracellular Ca²⁺ release, ROCK-mediated myofilament Ca²⁺ sensitization, and expression of SM-MHC in BIPAs.

We discovered that small lipid molecules derived from AA, 15- and 20-HETEs, triggered GJ signaling between SMCs. Inhibition of 15-LOX, 20-HETE synthase, and COX significantly decreased sustained HPV. The GJ inhibitor did not further block/reduce HPV in the presence of LOX and 20-HETE synthase inhibitors. Since both 18β-GA and Gap-27 blocked the 15- and 20-HETE-mediated contraction of endothelium-denuded BIPAs, these findings suggest that 15- and 20-HETE potentially evoked signal transduction through GJs between SMCs. Previous studies have reported that 15-HETE evokes contraction of rabbit pulmonary arteries (53) and participates in the pathogenesis of hypoxic pulmonary hypertension (8, 36). Altogether our data show that in the sustained HPV development GJ-mediated action of 15-HETE contributed to increase intracellular Ca²⁺, triggered ROCK-mediated myofilament Ca²⁺ sensitization, and potentiated MLC phosphorylation in BIPAs. To support the last statement, it has been shown that AA-derived 15-LOX-dependent products induce ROCK-mediated phosphorylation of protein phosphatase-1 inhibitor CPI-17 and MLC phosphatase inhibition leading to myofilament Ca²⁺ sensitization in vascular SMCs (51). Thus our results provide evidence that 15-HETE-dependent GJ signaling between SMCs in BIPAs may contribute to sustained HPV development.

Inhibition of 15-LOX and 20-HETE synthase, but not COX, decreased SM-MHC expression in endothelium-denuded BIPAs. 18β-GA decreased SM-MHC expression induced by 15-HETE, but not by 20-HETE. This indicates that 15-HETE-elicited signals through GJs between SMCs increased SM-MHC expression. A previous study suggested that Cx43 in MEGJs conducted signals that stimulated TGFβ/SMAD-mediated SMCs differentiation (20, 21). In our hands, GJ inhibitors did not affect phosphorylation of SMAD2 and SMAD3 (data not shown), and the ROCK inhibitor did not reduce 15-HETE induced expression of SM-MHC. Instead, inhibition of NFAT blocked 15-HETE-elicited SM-MHC expression. The SM-MHC gene has a NFAT binding site and NFAT is known to activate SM-MHC gene transcription (37). Therefore, our novel findings suggest that NFAT, but not SMAD and ROCK, signaling potentially participated in 15-HETE-elicited GJ communication between SMCs that increased SM-MHC expression.

15-HETE induces inflammation and proliferation of pulmonary artery SMCs (35, 36) and, in contrast, 20-HETE relaxes pulmonary arteries (7). However, our knowledge about their role in regulating SMCs phenotype or SMC contractile protein expression is limited. A recent study suggests that 20-HETE-derived COX-2-dependent 20-OH-PGE enhances mature inflamed adipocyte hypertrophy in mesenchymal stem cell undergoing adipogenic differentiation (29). Furthermore, other studies have suggested that PGE₂ released by MLO-Y4 cells functions in an autocrine manner to regulate GJ function and Cx43 expression (15) and PGE₂ receptor-4 signaling modulates expression of a number of unique pathways involved in SMCs proliferation and migration (23). Therefore, it is plausible that derivatives of 20-HETE in BIPAs likely mediate HPV and SM-MHC expression. Since 1,000-fold higher dose of 20-HETE than of 15-HETE was required to contract BIPAs and to increase SM-MHC expression, we speculated that 20-HETE-derived COX-dependent 20-OH-PGE₂ or 15-LOX-dependent 15,20-DiHETE mediates the physiological actions of 20-HETE in BIPAs. Mass spectrometry data supported this notion and confirmed that 20-OH-PGE₂ increased by 22-fold when 20-HETE was incubated with BIPAs. Similarly, 15,20-DiHETE was detected in the tissue bath fluid when BIPAs were incubated with 20-HETE. It is noteworthy that this metabolite was undetectable in the absence of 20-HETE. Intriguingly, no cell surface receptors have yet been identified for 15-HETE or 20-HETE. Furthermore, COX and 15-LOX are intracellular enzymes. COX-1/2 is localized to sarco(endo)plasmic reticulum in most cell types (39). 15-LOX is mostly cytoplasmic and is also found on the nuclear envelope in corneal cells (10, 14) and on the inner plasma membrane in blood cells (11). Based on the localization of COX and LOX, it reasonable to speculate that extracellular 20-HETE has to potentially reach these enzymes in the intracellular compartments to generate 20-OH-PGE₂ and 15,20-DiHETE derivatives. Therefore, we propose 20-HETE perhaps passes through the GJs. Consistently, 18β-GA blocked generation of 20-OH-PGE₂ and 15,20-DiHETE. Since neither 15-HETE nor 20-HETE elicited contraction of endothelium-intact BIPAs (data not shown), we speculate that 20-HETE potentially entered the SMCs through MEGJs exposed by removal of endothelium. This idea is further reinforced by our results indicating the fluorescent-tagged 20-HETE analog 20-HEDE-NBD is localized in the cytoplasm of cultured cells and GJ inhibition blocked 20-HETE evoked contraction of BIPAs. Furthermore, 15-LOXI1 blockade of the 20-HETE-mediated contractions also supports the LC-MS/MS results and indicates that 15-LOX mediates the conversion of 20-HETE to 15,20-DiHETE in BIPAs. The 20-HETE-induced increase of SM-MHC was not blocked by 18β-GA. This suggests that intercellular communication was not involved in 20-HETE-stimulated SM-MHC expression, but instead 20-HETE or a 20-HETE-derivative, 15,20-DiHETE or 20-OH-PGE₂, elicited SM-MHC expression in an autocrine/paracrine manner. 20-OH-PGE₂ and 15,20-DiHETE are polar molecules and can easily conduct out or diffuse into extracellular space between SMCs/bath fluid and then exert their physiological action through activation of GJ signaling or cell surface receptors. In this regards, it is noteworthy that a 20-HETE synthase inhibitor blocked HPV by 80–90% in the presence of GJ blockade. Our results are consistent with the previous findings that suggest PGE₂ receptor-2 is a critical component of the signaling cascade between mechanical strain and GJ-mediated communication between osteocytes (15) and by which maturation and increased contractility of the vessel are coupled to the potent smooth muscle dilatory actions of PGE₂ (23). We report here detection of 15,20-DiHETE in BIPAs and describe a potential association of 20-OH-PGE₂ and 15,20-DiHETE with GJ signaling between SMCs that are involved in the development of sustained HPV and in SMC contractile protein expression.

It has been shown previously that the SMC population in pulmonary artery wall is not uniform/homogenous (38) and we suppose that SMCs located closer (peripheral) to the intima and adventitia are more sensitive to hypoxia. Because of this, activation of 15-LOX can be attributed predominantly to peripheral SMCs and then 15-HETE/15,20-DiHETE produced by 15-LOX can be transmitted via GJs to the SMCs in deeper layers away from the blood flow leading to their contraction and the sustained HPV development. A scheme illustrating potential signaling pathways involved in GJ-mediated intercellular signaling between SMCs in sustained HPV of BIPAs according to our findings is represented in Fig. 11.

Fig. 11.

A schematic demonstrating potential signaling pathways involved in GJ-mediated intercellular signaling between SMCs in sustained hypoxic vasoconstriction of BIPAs. Hypoxia induces activation of 15-LOX in peripheral SMCs of BIPAs, which leads to 15-LOX-mediated formation of 15-HETE and transformation of 20-HETE to 15,20-HETE. These metabolites of AA then pass through GJs to deeply situated SMCs or through GJs stimulate SMCs in deeper medial layer where activate expression of SM-MHC, MLC phosphorylation, and possibly ROCK-mediated myofilament Ca²⁺ sensitization and intracellular Ca²⁺ ([Ca²⁺]_i) contributing to the sustained HPV development in BIPAs.

In summary, our novel findings demonstrate that 1) GJs between SMCs play a critical role in mediating signals that elicit sustained HPV, 2) 15-HETE- and 20-HETE-mediate GJ signaling in the SMCs eliciting sustained HPV and increasing the expression of SM-MHC, and 3) 20-HETE-derived LOX-dependent 15,20-DiHETE and the COX-dependent 20-OH-PGE₂ might play a potential role in mediating sustained HPV and expression of SM-MHC in BIPAs.

GRANTS

This study was supported by The Fulbright Program and the US Department of State's Bureau of Education and Cultural Affairs (to I. V. Kizub and S. A. Gupte; 2013–2014); National Heart, Lung, and Blood Institute Grants HL-115124 (to M. S. Wolin), HL-034300 (to M. L. Schwartzman), HL-116530 (to E. R. Jacobs), and HL-111392 (to J. R. Falck); Veterans Administration Merit Review (BX001681 to E. R. Jacobs); and the Robert A. Welch Foundation (I-0011 to J. R. Falck).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

I.V.K., M.S.W., J.R.F., E.R.J., M.L.S., and S.A.G. conception and design of research; I.V.K., A.L., V.D., S.R.J., H.J., J.R.F., S.R.K., R.E., and E.R.J. performed experiments; I.V.K., A.L., V.D., S.R.J., H.J., and E.R.J. analyzed data; I.V.K., S.R.J., M.S.W., E.R.J., M.L.S., and S.A.G. interpreted results of experiments; I.V.K., A.L., V.D., E.R.J., and S.A.G. prepared figures; I.V.K. and S.A.G. drafted manuscript; I.V.K., A.L., S.R.J., M.S.W., J.R.F., E.R.J., M.L.S., and S.A.G. edited and revised manuscript; I.V.K., A.L., V.D., S.R.J., H.J., M.S.W., J.R.F., E.R.J., M.L.S., and S.A.G. approved final version of manuscript.