Connexin 43 Plays a Role in Pulmonary Vascular Reactivity in Mice

Pulmonary arterial hypertension (PAH) is a chronic condition characterized by vascular remodeling and increased vaso-reactivity. PAH is more common in females than in males (~3:1). Connexin (Cx)43 has been shown to be involved in cellular communication within the pulmonary vasculature. Therefore, we investigated the role of Cx43 in pulmonary vascular reactivity using Cx43 heterozygous (Cx43+/−) mice and 37,43Gap27, which is a pharmacological inhibitor of Cx37 and Cx43. Contraction and relaxation responses were studied in intra-lobar pulmonary arteries (IPAs) derived from normoxic mice and hypoxic mice using wire myography. IPAs from male Cx43+/− mice displayed a small but significant increase in the contractile response to endothelin-1 (but not 5-hydroxytryptamine) under both normoxic and hypoxic conditions. There was no difference in the contractile response to endothelin-1 (ET-1) or 5-hydroxytryptamine (5-HT) in IPAs derived from female Cx43+/−mice compared to wildtype mice. Relaxation responses to methacholine (MCh) were attenuated in IPAs from male and female Cx43+/− mice or by pre-incubation of IPAs with 37,43Gap27. Nω-Nitro-L-arginine methyl ester (l-NAME) fully inhibited MCh-induced relaxation. In conclusion, Cx43 is involved in nitric oxide (NO)-induced pulmonary vascular relaxation and plays a gender-specific and agonist-specific role in pulmonary vascular contractility. Therefore, reduced Cx43 signaling may contribute to pulmonary vascular dysfunction.


Introduction
Pulmonary arterial hypertension (PAH) is a progressive disease in which the mean resting pulmonary artery pressure rises above 25 mmHg with a mean resting capillary wedge pressure lower than 15 mmHg [1]. This increase in pressure is associated with both constriction and remodeling of the distal pulmonary vasculature and eventually leads to right-sided heart failure. Prognosis is poor and survival has been reported to only 68% after three years on therapy [2]. PAH is far more common in females than in males (~3:1) [3]. Dysregulation of cell-to-cell communication particularly between pulmonary artery endothelial cells (PAECs) and pulmonary artery smooth muscle cells (PASMCs) is thought to play an important role in both constriction and remodeling of the pulmonary vasculature in PAH [4]. For example, PAECs from patients with PAH have increased gene and protein expression of tryptophan hydroxylase 1 (Tph1), which is the rate-limiting enzyme in the synthesis of 5-hydroxytryptamine (5-HT) [5]. 5-HT causes contraction of pulmonary arteries and proliferation of PASMCs [6]. In addition, PAECs from PAH patients produce decreased amounts of nitric oxide, which is a potent vasodilator that suppresses proliferation of PASMCs [7].

Gene Expression of Connexins in Pulmonary Arteries from Cx43 heterozygous Mice
First, quantitative real time PCR (qPCR) was performed to confirm reduced gene expression of Cx43 in pulmonary arteries of male and female Cx43 +/− mice. Cx43 expression was higher in females than in males in both WT and Cx43 +/− mice ( Figure 1A). Afterward, pulmonary arterial gene expression levels of Cx43 with Cx37, Cx40, and Cx45 in wildtype mice were compared. Cx43 was the predominant vascular connexin in female mice. In male mice, there was a trend towards Cx43 being expressed at greater levels than Cx37 and Cx40, but this was not significant. Cx45 was expressed at lower levels in both male and female mice ( Figure 1B). than in males in both WT and Cx43 +/− mice ( Figure 1A). Afterward, pulmonary arterial gene expression levels of Cx43 with Cx37, Cx40, and Cx45 in wildtype mice were compared. Cx43 was the predominant vascular connexin in female mice. In male mice, there was a trend towards Cx43 being expressed at greater levels than Cx37 and Cx40, but this was not significant. Cx45 was expressed at lower levels in both male and female mice ( Figure 1B).

Pulmonary Arterial Contractile Responses
In the intra-lobar pulmonary arteries (IPAs) of male mice, endothelin-1 (ET-1) was more potent (had a lower median effective concentration or EC 50 value) in Cx43 +/− mice compared to WT mice ( Figure 3A; Table 1). Maximal response to ET-1 (E max ) was, however, unchanged between WT and Cx43 +/− mice. There was no global shift in the concentration response curve ( Figure 3A; Table 1). IPAs from male Cx43 +/− mice showed similar contractile responses to 5-hydroxytryptamine (5-HT) as those from WT mice ( Figure 3B; Table 1). Contractile responses to both ET-1 and 5-HT were similar in IPAs from female Cx43 +/− mice compared to female WT mice (Figure 3C,D; Table 1).

Figure 3.
Pulmonary vascular contractility to ET-1 and 5-HT in intralobar pulmonary arteries (IPAs) from male and female wildtype (WT) and Cx43 heterozygous (Cx43 +/− ) mice. ET-1 was more potent in IPAs from male Cx43 +/− mice than WT mice (A). There was no difference in contractile response to 5-HT in IPAs from male WT and Cx43 +/− mice (B). There was no difference in ET-1 (C) or 5-HT (D) induced contractile response in IPAs from female WT or Cx43 +/− mice. Data are shown as mean ± S.E.M. Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC50) and maximal contractile responses (Emax) between two different groups were analyzed by using the Student's t-test. * EC50 is significantly (p < 0.05) reduced in male Cx43 +/− mice, n = 5-7 per group.  Pulmonary vascular contractility to ET-1 and 5-HT in intralobar pulmonary arteries (IPAs) from male and female wildtype (WT) and Cx43 heterozygous (Cx43 +/− ) mice. ET-1 was more potent in IPAs from male Cx43 +/− mice than WT mice (A). There was no difference in contractile response to 5-HT in IPAs from male WT and Cx43 +/− mice (B). There was no difference in ET-1 (C) or 5-HT (D) induced contractile response in IPAs from female WT or Cx43 +/− mice. Data are shown as mean ± S.E.M. Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC 50 ) and maximal contractile responses (E max ) between two different groups were analyzed by using the Student's t-test. * EC 50 is significantly (p < 0.05) reduced in male Cx43 +/− mice, n = 5-7 per group.

Pulmonary Arterial Relaxation Responses
The relaxation response produced by methacholine (MCh) was significantly reduced in IPAs of both male and female Cx43 +/− mice when compared to WT mice ( Figure 4A,B; Table 2). Pharmacological inhibition of Cx43 with 37,43 Gap27 also significantly attenuated the relaxation responses produced by MCh in IPAs of both male and female mice ( Figure 4C,D; Table 2).
We then confirmed that MCh-induced relaxation responses were dependent upon nitric oxide (NO) since the L-NAME completely inhibited MCh-induced relaxation ( Figure 4E; Table 2). Then we assessed the role of Cx43 in isoprenaline-induced relaxation. Isoprenaline is classically thought to induce relaxation via the cAMP pathway. In these experiments, the relaxation induced by isoprenaline was partially inhibited by 37,43 Gap27 ( Figure 4F; Table 2). Furthermore, we showed nitric oxide plays a role in isoprenaline-induced relaxation since the L-NAME partially attenuated isoprenaline-induced relaxation ( Figure 4G; Table 2).

Pulmonary Arterial Relaxation Responses
The relaxation response produced by methacholine (MCh) was significantly reduced in IPAs of both male and female Cx43 +/− mice when compared to WT mice ( Figure 4A,B; Table 2). Pharmacological inhibition of Cx43 with 37,43 Gap27 also significantly attenuated the relaxation responses produced by MCh in IPAs of both male and female mice ( Figure 4C,D; Table 2).
We then confirmed that MCh-induced relaxation responses were dependent upon nitric oxide (NO) since the L-NAME completely inhibited MCh-induced relaxation ( Figure 4E; Table 2). Then we assessed the role of Cx43 in isoprenaline-induced relaxation. Isoprenaline is classically thought to induce relaxation via the cAMP pathway. In these experiments, the relaxation induced by isoprenaline was partially inhibited by 37,43 Gap27 ( Figure 4F; Table 2). Furthermore, we showed nitric oxide plays a role in isoprenaline-induced relaxation since the L-NAME partially attenuated isoprenaline-induced relaxation ( Figure 4G; Table 2).   37,43 Gap27 also reduced the relaxation response in male (C) and female (D) mice. MCh-induced relaxation was ablated in the presence of L-NAME (E). 37,43 Gap27 partially inhibited isoprenaline-induced relaxation in IPAs from male mice (F) as did L-NAME (G). Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC50) and maximal relaxation responses (Rmax) between two different groups were analyzed by using the Student's t-test. Data are shown as mean ± S.E.M. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, n = 5-6 per group. Statistical symbols shown on the right hand side of graphs indicate global shifts in the concentration response curves. Statistical symbols underneath curves indicate changes in maximal relaxation (Rmax) values. Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC50) and maximal relaxation responses (Rmax) between two different groups were analyzed by using the Student's t-test. Data are shown as mean ± S.E.M. * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

Gene Expression of Bone Morphogenetic Protein Receptor Type II, Tryptophan Hydroxylase 1, and Endothelial Nitric Oxide Synthase in Pulmonary Arteries from Cx43 Heterozygous Mice
Since Cx43 +/− mice displayed dysregulated pulmonary vascular reactivity, gene expression BMPRII (encoded by BMPR2), eNOS (encoded by NOS3), and Tph-1, mediatorsimportant f  37,43 Gap27 also reduced the relaxation response in male (C) and female (D) mice. MCh-induced relaxation was ablated in the presence of L-NAME (E). 37,43 Gap27 partially inhibited isoprenaline-induced relaxation in IPAs from male mice (F) as did L-NAME (G). Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC 50 ) and maximal relaxation responses (R max ) between two different groups were analyzed by using the Student's t-test. Data are shown as mean ± S.E.M. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, n = 5-6 per group. Statistical symbols shown on the right hand side of graphs indicate global shifts in the concentration response curves. Statistical symbols underneath curves indicate changes in maximal relaxation (R max ) values. Global differences in concentration response curves were compared by two-way ANOVA. Changes in logarithm of median effective concentration (Log EC 50 ) and maximal relaxation responses (R max ) between two different groups were analyzed by using the Student's t-test. Data are shown as mean ± S.E.M. * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

Gene Expression of Bone Morphogenetic Protein Receptor Type II, Tryptophan Hydroxylase 1, and Endothelial Nitric Oxide Synthase in Pulmonary Arteries from Cx43 Heterozygous Mice
Since Cx43 +/− mice displayed dysregulated pulmonary vascular reactivity, gene expression of BMPRII (encoded by BMPR2), eNOS (encoded by NOS3), and Tph-1, mediatorsimportant for regulating pulmonary vascular function, were assessed. Expression of BMPR2, NOS3 , and Tph-1 were not significantly altered in either male or female Cx43 +/− mice compared to WT littermates ( Figure  The expression of NOS3 and BMPR2 were significantly lower in female WT mice when compared to male WT mice ( Figure 5A,B). regulating pulmonary vascular function, were assessed. Expression of BMPR2, NOS3 , and Tph-1 were not significantly altered in either male or female Cx43 +/− mice compared to WT littermates ( Figure 5A-C). The expression of NOS3 and BMPR2 were significantly lower in female WT mice when compared to male WT mice ( Figure 5A,B). . BMPR2 and NOS3 were significantly downregulated in female WT mice compared to male WTs. Data are presented as mean ± S.E.M. and were analysed by using two-way ANOVA. * p < 0.05, n = 6 per group with each sample run in triplicate.

Effects of Hypoxia on Pulmonary Vascular Contractility in Male Cx43 Heterozugous Mice
The effects of chronic hypoxia on pulmonary vascular reactivity in Cx43 +/− mice were investigated. The hypoxic experiments were carried out in male mice since it has been previously shown that male mice are more susceptible to hypoxic-induced PH than female mice [25]. Both WT and Cx43 +/− mice developed the right ventricular hypertrophy after two weeks of chronic hypoxic exposure, which verifies the mouse hypoxic model ( Figure 6A). IPAs derived from chronic hypoxic Cx43 +/− mice showed an increased sensitivity to ET-1, which was assessed by a reduced EC50 value and a global leftward shift in the contractile response. In addition, the maximal contractile effect produced by ET-1 was significantly greater in IPAs from hypoxic Cx43 +/− mice than from hypoxic WT mice ( Figure 6B; Table 3). There was no difference in a contractile response to 5-HT in IPAs from hypoxic Cx43 +/− mice ( Figure 6C; Table 3). (C) in pulmonary arteries of male and female WT and Cx43 +/− mice. No differences were observed in expression of BMPR2, NOS3, or Tph1 between WT and Cx43 +/− mice (either male or female). BMPR2 and NOS3 were significantly downregulated in female WT mice compared to male WTs. Data are presented as mean ± S.E.M. and were analysed by using two-way ANOVA. * p < 0.05, n = 6 per group with each sample run in triplicate.

Effects of Hypoxia on Pulmonary Vascular Contractility in Male Cx43 Heterozugous Mice
The effects of chronic hypoxia on pulmonary vascular reactivity in Cx43 +/− mice were investigated. The hypoxic experiments were carried out in male mice since it has been previously shown that male mice are more susceptible to hypoxic-induced PH than female mice [25]. Both WT and Cx43 +/− mice developed the right ventricular hypertrophy after two weeks of chronic hypoxic exposure, which verifies the mouse hypoxic model ( Figure 6A). IPAs derived from chronic hypoxic Cx43 +/− mice showed an increased sensitivity to ET-1, which was assessed by a reduced EC 50 value and a global leftward shift in the contractile response. In addition, the maximal contractile effect produced by ET-1 was significantly greater in IPAs from hypoxic Cx43 +/− mice than from hypoxic WT mice ( Figure 6B; Table 3). There was no difference in a contractile response to 5-HT in IPAs from hypoxic Cx43 +/− mice ( Figure 6C; Table 3). In panels B and C, global differences in CRCs were compared by two-way ANOVA. Changes in the logarithm of median effective concentration (Log EC50) and maximal contractile responses (Emax) between two different groups were analyzed by using the Student's t-test * p < 0.05, ** p < 0.01, *** p < 0.001, n = 5-7 per group. The statistical symbol shown on the right hand side of graph B indicates a global shift in the CRC. The symbol underneath the curve indicates changes in the median effective concentration (EC50) while the symbol above the curve indicates changes in the maximal response (Emax). Hypoxic Cx43 +/− −6.5 ± 0.03 130.3 ± 4.5 7 Log EC50 indicates the logarithm of median effective concentration, Emax maximal contractile effect. Global differences in concentration response curves were compared by using two-way ANOVA. Changes in logarithm of median effective concentration (Log EC50) and maximal contractile responses (Emax) between two different groups were analyzed by using the Student's t-test. NS: not significant * p < 0.05, ** p < 0.01, *** p < 0.001. Data are shown as mean ± S.E.M.  Log EC 50 indicates the logarithm of median effective concentration, E max maximal contractile effect. Global differences in concentration response curves were compared by using two-way ANOVA. Changes in logarithm of median effective concentration (Log EC 50 ) and maximal contractile responses (E max ) between two different groups were analyzed by using the Student's t-test. NS: not significant * p < 0.05, ** p < 0.01, *** p < 0.001. Data are shown as mean ± S.E.M.

Effects of Hypoxia on Expression of Cx43 in Mouse Lung and Pulmonary Artery
Afterward, the effects of chronic hypoxia on Cx43 gene and protein expression were assessed. Using qPCR, it was shown that hypoxia significantly down-regulated Cx43 gene expression in both WT and Cx43 +/− male mice ( Figure 7A). We then used immunofluorescence to visualize the effects of hypoxia on Cx43 protein expression. Lung expression of Cx43 (green fluorescence) was reduced in hypoxic WT mice when compared to normoxic WT mice ( Figure 7B). Furthermore, Cx43 immunoreactivity was further reduced in lungs of hypoxic Cx43 +/− mice when compared to normoxic Cx43 +/− mice ( Figure 7B).

Effects of Hypoxia on Expression of Cx43 in Mouse Lung and Pulmonary Artery
Afterward, the effects of chronic hypoxia on Cx43 gene and protein expression were assessed. Using qPCR, it was shown that hypoxia significantly down-regulated Cx43 gene expression in both WT and Cx43 +/− male mice ( Figure 7A). We then used immunofluorescence to visualize the effects of hypoxia on Cx43 protein expression. Lung expression of Cx43 (green fluorescence) was reduced in hypoxic WT mice when compared to normoxic WT mice ( Figure 7B). Furthermore, Cx43 immunoreactivity was further reduced in lungs of hypoxic Cx43 +/− mice when compared to normoxic Cx43 +/− mice ( Figure 7B).

Effects of Hypoxia on Gene Expression in Pulmonary Arteries Derived from WT and Cx43 +/− Mice
As chronic hypoxia mediated increased vascular reactivity to ET-1 and decreased Cx43 expression, the effects of chronic hypoxia on the expression of other vascular connexins and Panx1 were assessed (Figure 8 A-D). Hypoxia mediated a downregulation of Cx40 gene expression in WT mice ( Figure 8B) while gene expression of Cx45 was up-regulated by hypoxia in Cx43 +/− mice ( Figure 8C). Hypoxia had no effect on the expression of Cx37 or Panx1 in WT or Cx43 +/− mice ( Figure 8A,D).
Gene expression of BMPR2, Tph1, and NOS3 in response to hypoxia in Cx43 +/− mice was also assessed. BMPR2 was downregulated by hypoxia in WT mice and Tph1 was up-regulated by hypoxia in Cx43 +/− mice. There were no differences in BMPR2, NOS3, or Tph1 expression between hypoxic WT and hypoxic Cx43 +/− mice ( Figure 8E-G).

Effects of Hypoxia on Gene Expression in Pulmonary Arteries Derived from WT and Cx43 +/− Mice
As chronic hypoxia mediated increased vascular reactivity to ET-1 and decreased Cx43 expression, the effects of chronic hypoxia on the expression of other vascular connexins and Panx1 were assessed (Figure 8 A-D). Hypoxia mediated a downregulation of Cx40 gene expression in WT mice ( Figure 8B) while gene expression of Cx45 was up-regulated by hypoxia in Cx43 +/− mice ( Figure  8C). Hypoxia had no effect on the expression of Cx37 or Panx1 in WT or Cx43 +/− mice ( Figure 8A,D).
Gene expression of BMPR2, Tph1, and NOS3 in response to hypoxia in Cx43 +/− mice was also assessed. BMPR2 was downregulated by hypoxia in WT mice and Tph1 was up-regulated by hypoxia in Cx43 +/− mice. There were no differences in BMPR2, NOS3, or Tph1 expression between hypoxic WT and hypoxic Cx43 +/− mice ( Figure 8E-G). (G) in pulmonary arteries of wildtype (WT) and Cx43 heterozygous (Cx43 +/− ) mice under normoxic and chronic hypoxic conditions. All data are presented as mean ± S.E.M. and were analyzed by two-way ANOVA. * p < 0.05, ** p < 0.01, n = 6 per group with each sample run in triplicate.

Discussion
To our knowledge, this is the first study to show that mice genetically deficient in Cx43 develop pulmonary vascular dysfunction. It is also the first study to show that Cx43 plays a role in NO-induced pulmonary vascular relaxation. These findings add to the growing body of evidence that suggests that Cx43 is involved in regulation of the pulmonary vasculature. Since PAH is associated with altered pulmonary vascular reactivity, these data suggest Cx43 is worthy of further investigation as a novel therapeutic target for PAH.
Relaxation responses to MCh were reduced in IPAs from both male and female Cx43 +/− mice. Subsequently, we found that pharmacological inhibition of Cx43 using 37,43 Gap27 also attenuated (G) in pulmonary arteries of wildtype (WT) and Cx43 heterozygous (Cx43 +/− ) mice under normoxic and chronic hypoxic conditions. All data are presented as mean ± S.E.M. and were analyzed by two-way ANOVA. * p < 0.05, ** p < 0.01, n = 6 per group with each sample run in triplicate.

Discussion
To our knowledge, this is the first study to show that mice genetically deficient in Cx43 develop pulmonary vascular dysfunction. It is also the first study to show that Cx43 plays a role in NO-induced pulmonary vascular relaxation. These findings add to the growing body of evidence that suggests that Cx43 is involved in regulation of the pulmonary vasculature. Since PAH is associated with altered pulmonary vascular reactivity, these data suggest Cx43 is worthy of further investigation as a novel therapeutic target for PAH.
Relaxation responses to MCh were reduced in IPAs from both male and female Cx43 +/− mice. Subsequently, we found that pharmacological inhibition of Cx43 using 37,43 Gap27 also attenuated MCh-induced relaxation. 37,43 Gap 27 inhibits both Cx37 and Cx43 and both Cx37 and Cx43 are down-regulated in male Cx43 +/− mice. However, the vascular connexins of only Cx43 are downregulated in female Cx43 +/− mice. Therefore, reduction of MCh-induced relaxation in female Cx43 +/− mice suggests an important role for Cx43 in this effect. Our studies confirmed that MCh-induced relaxation was NO-dependent since the MCh mediated relaxation responses were completely abolished in the presence of L-NAME. It is widely considered that NO can diffuse through the endothelial and smooth muscle cell membranes to activate guanylate cyclase within the smooth muscle cell and mediate vasodilation. It has been shown, however, that diffusion of NO across the vascular cell membrane requires specific connexin channels [28]. NO opened and permeated hemichannels expressed in HeLa cells transfected and selected to express Cx43, Cx40, or Cx37. In addition, the blockade of connexin channels abolished myoendothelial NO transfer and NO-dependent vasodilation induced by acetylcholine in rat mesenteric arteries [28]. There is also mounting evidence that NO can interact with gap junctions in a complex and inter-dependent fashion [29]. Cx43 is constitutively S-nitrosylated at cysteine 271 by NO at the myoendothelial gap junctions. This nitrosylation keeps the myoendothelial gap junction open and denitrosylation closes the gap junction channel [30]. Additionally, NO has been shown to enhance gap junction coupling and increase trafficking of Cx40 to the membrane in endothelial cells via the protein kinase A activation [31]. Cx40 has been shown to co-localize with eNOS in the mouse aorta and is involved in conducting vasodilation [32]. Cx40 knock out (Cx40 −/− ) mice showed reduced basal and acetycholine induced NO release and reduced eNOS expression in aortas [32]. However, NO can act as a negative regulator of Cx37, which reduces Cx37-mediated dye transfer and electrical coupling in human umbilical vascular endothelial cells (HUVEC) and mouse microvascular endothelial cells [33,34]. Therefore, it is possible that Cx43 and NO interact in such a fashion that when Cx43 is genetically downregulated or pharmacologically inhibited, it leads to a reduction in NO-mediated vasorelaxation. In the present study, pulmonary arterial gene expression of eNOS was unchanged in male and female Cx43 +/− mice when compared to WT controls. It would, however, be of interest to investigate eNOS protein levels and also MCh-induced NO release in pulmonary arteries from WT and Cx43 +/− mice. It would also be of interest to investigate potential interactions between eNOS and Cx43 using co-immunoprecipitation. 37,43 Gap27 also mediated a small but significant inhibition of isoprenaline-induced relaxation.
Isoprenaline is classically thought to act through the cAMP pathway. In the present study, isoprenaline-induced relaxation was, however, partially inhibited by L-NAME, which suggests that NO plays a role in isoprenaline-induced relaxation. A previous study has also shown β 2 adrenoceptor induced relaxation in mouse pulmonary arteries are attenuated by the L-NAME, endothelial denudation, or deletion of eNOS [35]. β 2 adrenoceptors have been detected in the mouse pulmonary endothelial layer by immunostaining [35]. The inhibitions of isoprenaline relaxation mediated by 37,43 Gap27 and L-NAME were of a similar magnitude. It is, therefore, possible that inhibition of connexin mediated communication via 37,43 Gap27 inhibits the NO component of isoprenaline relaxation. In future studies, it would be of interest to analyze whether smooth muscle responses to NO are affected by downregulation of Cx43. This could be achieved using an NO donor such as sodium nitroprusside. Enhanced ET-1 mediated contraction was observed in IPAs derived from both normoxic and hypoxic male Cx43 +/− mice compared to their WT counterparts. Normoxic female Cx43 +/− mice did not show an increased contractile response to ET-1. However, male Cx43 +/− mice showed downregulation of Cx37, Cx40, Cx45, and Panx1. These effects were not observed in female Cx43 +/− mice. Downregulation of Cx37, Cx40, Cx45, or Panx1 may, therefore, play a role in the increased contractile response to ET-1 observed in the male Cx43 +/− mice. The reduced Cx43 expression observed in the pulmonary arteries of male mice compared to female mice observed in this study may also contribute to the increased effects of ET-1 in male Cx43 +/− mice. Gender differences in the endothelin system may also contribute to contractile differences between male and female Cx43 +/− mice [36]. Studies have shown that plasma endothelin levels were higher in men when compared to women [37,38]. At the receptor level, the ratio of ET A and ET B receptors in the endothelium of saphenous vein was greater in male subjects (3:1) than in female subjects (1:1) [39]. Haemodynamically, a rat model showed pressor responses induced by intravenous administration of ET-1 were greater in male rats than in female rats [40]. In vitro cell culture studies have also confirmed that 17-β oestradiol (E2) inhibited ET-1 gene expression through the extracellular signal regulated kinase (ERK) pathway [41].
5-HT mediated contractile responses in the IPAs were not affected by the partial loss of Cx43 in either male or female mice under either normoxic or hypoxic conditions. Therefore, the effects of reduced Cx43 gene expression varied according to the agonist used. These results agree with previously published data showing that the effects of pharmacological inhibition by 37,43 Gap27 on contractile responses in IPAs were dependent on the agonist used [18]. In line with the current results, Billaud et al. showed that 37,43 Gap27 had no effect on 5-HT induced contraction in IPAs derived from hypoxic rats. However, they did report that 37,43 Gap27 could inhibit 5-HT induced contraction in IPAs from normoxic rats. In contrast with the results reported here showing that downregulation of Cx43 enhanced ET-1 mediated contraction in IPAs form both normoxic and hypoxic Cx43 +/− mice, Billaud et al. found that 37,43 Gap27 had no effect on ET-1 induced contraction in IPAs derived from normoxic or hypoxic rats [18]. 5-HT and ET-1 have different mechanisms of contraction. Since connexins are modulated by PKA, PKC, and calcium [42], the role of connexin mediated communication in the contractile response differs according to the agonist used. A possible reason for the disparity between the Billaud study and our own could be due to the ratiometric changes in the balance of Cx43:Cx40:Cx37 expression between rats and mice. A number of studies have reported tissue and species specific variation in connexin expression profiles and 37,43 Gap27 is defined as being specific to the SRPTEKTIFII sequence-conserved between Cx43 and Cx37 but different in Cx40 [43][44][45][46]. In addition, Cx43 +/− mice will have compensatory changes in other genes, which may affect the contractile response while 37,43 Gap27 inhibits Cx37 as well as Cx43. This also potentially affects the contractile response.
In the gene expression studies, Cx43 expression was higher in female mice than in male mice. This is in keeping with a report which showed that Cx43 expression was higher in rat cardiomyocytes derived from females than males [47]. Furthermore, treatment with oestradiol increased Cx43 expression in human myometrium [48]. In addition to this, there is direct evidence that estrogen can regulate Cx43 gene expression since the promotor region of Cx43 gene contains an estrogen response element [49].
Chronic hypoxic rodents are a commonly used model for PAH. In the current study, the chronic hypoxia suppressed Cx43 gene and protein expression are located in the pulmonary arteries of both WT and Cx43 +/− mice. Hypoxia can regulate Cx43 expression post-translationally by phosphorylation [50]. For example, one study showed exposure to hypoxia is associated with an increase in phosphorylation of Cx43-serine 368 (Ser368) in human microvascular endothelial cells and this phosphorylation was associated with downregulation Cx43 protein expression [51]. As hypoxia downregulates Cx43 expression, it would be of interest in future studies to assess the effects of hypoxia on MCh-induced relaxation in Cx43 +/− mice. The current study showed Cx40 expression was also downregulated in hypoxic WT mice. Previous studies have shown that Cx40 expression was reduced during PAH in rats and treatment with sildenafil increased Cx40 expression via BMP signaling [52,53]. We found Cx45 expression to be upregulated in the Cx43 +/− mice under chronic hypoxic conditions. The function of Cx45 in the vasculature remains unknown even though it has long been known to be expressed in vascular smooth muscle cells [54].
It is interesting to note that Cx43 +/− mice did not develop increased right ventricular hypertrophy (RVH) in response to hypoxia compared to their wildtype counterparts. It has previously been shown by ourselves and others that changes in pulmonary vascular pressures and pulmonary vascular remodeling do not always lead to the development of RVH in mice. For example, increased pulmonary pressures and remodeling have been observed in the absence of RVH in mice that over-express the serotonin transporter, mice that over-express Mts 1, and mice that are dosed with dexfenfluramine [55][56][57]. In addition, Cx43 is highly expressed in cardiac myocytes and is thought to play an important role in hypertrophy of these cells. Expression and localization of Cx43 in cardiac myocytes has been shown to be dynamically regulated in various animal models of cardiac hypertrophy [58]. Therefore, it is possible that cardiac myocytes from Cx43 +/− mice are functionally abnormal and, therefore, have an atypical hypertrophic response to hypoxia.
Female Cx43 +/− mice did not exhibit the compensatory reduction in gene expression of Cx37, Cx40, Cx45, or Panx1 that was observed in male Cx43 +/− mice. Multiple lines of evidence show that Cx43, Cx40, and Cx37 are interdependent on each other and compensatory changes occur upon connexin deletion. For instance, in Cx40 knock out (Cx40 −/− ) mice both total and smooth muscle Cx43 protein expression was reduced in the mouse aortas [59]. In addition, in Cx40 −/− mice, the pericellular component of Cx43 staining was lost and there was increased redistribution of Cx43 in the perinuclear region [59]. This suggests deletion of Cx40 not only leads to Cx43 downregulation but also affects its trafficking. Conversely, another group found that Cx40 −/− mice showed upregulation of Cx37 and Cx43 in the aortic endothelium [60]. In Cx40 −/− neonatal mice, Cx37 protein expression was downregulated in the endothelium and there was an increased Cx37 and Cx43 in the medial layer [61]. Genetic deletion of Cx37 in mice showed reduction in endothelial Cx40 [61]. In the present study, we have assessed changes in gene expression in the whole pulmonary artery. It would be of interest to assess cell type specific changes in future studies.
Gene expression of eNOS, BMPRII, or Tph1 were not affected by the loss of Cx43 under normoxic or hypoxic conditions. Among the normoxic WT mice, eNOS and BMPRII expression was reduced in the females compared to male WT mice. The literature on eNOS expression in PAH patients is contradictory. eNOS was initially reported to be decreased in lungs of PAH patients [62], but evidence later found eNOS expression was unchanged or even increased in PAH patients [63,64]. Furthermore, NO levels have been shown to be reduced in PAH patients [65] and it has been reported that the activity of eNOS rather than its expression was altered in a murine model of PAH [63]. Our findings show that BMPRII expression is reduced in pulmonary arteries from female mice, which is in line with proof that activation of the estrogen response element in the promoter region of BMPRII can downregulate BMPRII expression [66].
In conclusion, this study has shown that Cx43 plays a role in NO-dependent vasodilation in the pulmonary vasculature. Cx43 is also involved in pulmonary vascular contractility. However, effects on contractility are gender-dependent and agonist-dependent. Hypoxia has been shown to decrease Cx43 expression in mouse pulmonary arteries, which is an effect that may contribute to the increased vasoreactivity observed under hypoxic conditions.

Ethical Statement
All experimental procedures were carried out in accordance with the United Kingdom Animal Procedures Act (1986) and with the "Guide for the Care and Use of Laboratory Animals" published by the US National Institutes of Health (NIH publication no.85, eighth edition). Ethical approval was granted by the Glasgow Caledonian University Animal Welfare and Ethics Committee (PPL70/7875, 16 September 2013).

Animals
Male and female wild-type (WT) and Cx43 heterozygous (Cx43 +/− ) mice (C57BL6, 5 to 9 months old) were used in this study. The generation of Cx43 +/− mice was originally carried out by replacing exon 2 of the Cx43 gene with the neomycin resistance gene, which was previously described [24].
Mice were grouped under standard laboratory conditions. All mice had access to a commercial diet and water ad libitum.

Genotyping
DNA was extracted from ear notch tissues derived from WT and Cx43 +/− mice. Tissues were suspended in 300 µL TNES buffer [10 mM tris(hydroxymethyl)aminomethane (Tris), 0.4 M sodium chloride (NaCl), 100 mM ethylenediaminetetraacetic acid (EDTA), and 0.6% sodium dodecyl sulphate (SDS)] to which 1.5 µL proteinase K (Fisher Scientific, Loughborough, UK) was added. Samples were then incubated overnight at 55 • C. The next day, 84 µL of 5 M NaCl was added to each sample and samples were centrifuged for 10 min. The supernatant was collected and transferred to fresh tubes. DNA was precipitated by adding 200 µL ice cold 100% ethanol to each tube and vortexing. Samples were then centrifuged for 10 min, the supernatant was discarded, and the pellet was retained. Excess salt was removed by adding 200 µL ice cold 75% ethanol and samples were centrifuged again for 10 min. The supernatant was decanted gently and the pellet was allowed to air dry. The pellet was then re-suspended in 15 µL nuclease-free water and stored at −20 • C until use. A polymerase chain reaction (PCR) was then carried out to amplify the Cx43 (GJA1) and neomycin resistance (neo r ) genes. The primers used were as follows: neo r forward 5 -GATCGGCCATTGAACAAGATG, melting temperature (Tm) = 56.

Induction of Hypoxia
For induction of hypoxia, male WT and Cx43 +/− mice were placed in a hypobaric chamber for 14 days, which is previously described [19]. The pressure was adjusted to 550 mbar (equivalent to 10% v/v O 2 ) slowly over two days to allow mice to acclimate. The temperature was maintained at 20-22 • C. Control littermates were kept in a normoxic environment.

Tissue Preparation
Mice were euthanized by injection of phenobarbitone (60 mg/kg i.p.). After death was confirmed, the chest walls were opened using the mediastinal approach and the hearts and lungs were dissected freely.

Wire Myography Studies
Pharmacological experiments were carried out in third generation intra-lobar pulmonary arteries (IPAs;~300 µm internal diameter), which was previously described [67,68]. IPAs were mounted on a wire myograph (Danish Myo Technology, DMT) in freshly prepared Krebs-Henseleit Solution (composition (mmol/L) NaCl 119, KCl 4.7, CaCl 2 2.5, MgSO 4 1.2, NaHCO 3 25, KH 2 PO 4 1.2, and D-glucose 5.5) at (37 • C) and gassed with 95%O 2 / 5% CO 2 . All chemicals required for Krebs-Henseleit solution were purchased from Fisher Scientific except CaCl 2 which was purchased from VWR International Ltd. (Lutterworth, Leicestershire, UK). Following equilibration for one hour, IPAs from normoxic mice were placed under pressures of 12-15 mmHg and IPAs from hypoxic mice that were placed under pressures of 30-35 mmHg to mimic the in vivo environment described previously [67]. Arteries were initially constricted with potassium chloride (KCl, 60 mM), which were then washed out. These processes were repeated two times before contractile or relaxation experiments were carried out. For contractile experiments, cumulative concentration response curves (CCRCs) to 5-HT (1 nM-300 µM) or ET-1 (0.1 nM-0.1 µM) were constructed. For relaxation experiments, vessels were pre-constricted with phenylephrine (PE; 3 µM) and CRCs to MCh (0.1 nM-30 µM) or isoprenaline (1 nM-30 µM) were constructed. In a subset of relaxation experiments, vessels were incubated with 37,43 Gap27 (100 µM) for 30 min prior to MCh or isoprenaline-induced relaxation responses. 37,43 Gap27 has previously been shown to have an IC 50 of 31.5 ± 4.1 µM and to produce a maximum effect at 100 µM [69]. We previously used 37,43 Gap27 at 100 µM [70][71][72]. In another subset of relaxation experiments, the L-NAME (100 µM) was used to inhibit eNOS. In these experiments the L-NAME was applied for 30 min prior to pre-constriction with PE and the concentration of PE was reduced to 30 nM. It should be noted that 30 nM PE in the presence of L-NAME produced a similar contractile response to 3 µM PE alone. Changes in isometric tension were recorded on LabChart 7 software (AD Instruments Pty Ltd., Bella Vista, New South Wales, Australia). PE, L-NAME, MCh, ET and 5-HT were purchased from Sigma-Aldrich Company Ltd. (Gillingham, Dorset, UK).

Quantitative Real Time PCR (qPCR)
Main and branch pulmonary arteries (1st and 2nd order) were homogenized and RNA was extracted using the Nucleospin RNA kit (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany) as per the manufacturer instructions. RNA was then reverse transcribed to cDNA using the Precision nanoScript TM 2 Reverse Transcription Kit (Primerdesign Ltd., Chandler's Ford, UK), according to the manufacturer's protocols. qPCR was performed using Precision PLUS 2x qPCR master mix (Primerdesign Ltd.) II with taqman primers described in Table 4. β 2 -microglobulin (B2M) (assay ID: HK-DD-mo-300, Primerdesign Ltd.) was used as the endogenous control and the sequences for the B2M were kept confidential by Primerdesign Ltd. qPCR reactions were run in a real time PCR thermo-cycler machine (Viia TM 7 Real Time PCR System, ThermoFisher Scientific, Loughborough, UK) using the following conditions: 50 • C for 2 min and 95 • C for 10 min followed by 40 cycles of 95 • C for 15 s and 60 • C for 1 min. Gene expression was analyzed by the 2 ∆∆Ct method. Samples from at least six mice were used for each group and reactions for each sample were run in triplicate.

Immunofluorescence Staining
Sagittal sections (7 µM) were cut from lung embedded at an optimal cutting temperature (OCT) compound using a cryostat (Cryostar TM NX70 Cryostat from, Thermofisher Scientific). Sections were fixed in ice cold (−20 • C) methanol for 20 min and rehydrated in phosphate buffer saline (PBS, pH = 7.4). Sections were permeablised in PBS containing 0.1% (v/v) Triton-X100 before being blocked in 5% (w/v) skimmed milk in PBS solution for 30 min at room temperature. Sections were stained with primary antibody (rabbit polyclonal anti-Cx43, 1:100; Sigma-Aldrich Company Ltd., Gillingham, Dorset, UK) and incubated at 4 • C overnight. The next day, sections were washed in PBS for 30 min at room temperature and incubated with secondary antibody (goat anti-rabbit conjugated to Alexa flour 488, 1:500; Fisher Scientific) at 4 • C for 2 h. Nuclei were counter-stained with DAPI (1:1000; Thermofisher. Loughborough, UK). Mounting medium was applied on tissue sections and cover slides were applied. Slides were then examined under the LSM 800 Carl ZEISS confocal microscope, (Königsallee, Germany) for immunoreactivity.

Statistical Analysis
All data were shown as mean ± S.E.M. Data for cumulative concentration response curves (CRCs) were analyzed using GraphPad Prism 6 software (La Jolla, CA, USA). Data were fitted to a logistic equation, CRCs were generated, and EC 50 values were derived. Global differences in CRCs were compared by two-way ANOVA with the Bonferroni's post hoc test. Changes in the logarithm of median effective concentration (Log EC 50 ), maximal contractile responses (E max ), and maximal relaxation responses (R max ) between two different groups were analyzed by using the Student's t-test.