Cardiovascular Agents Affect the Tone of Pulmonary Arteries and Veins in Precision-Cut Lung Slices

Introduction Cardiovascular agents are pivotal in the therapy of heart failure. Apart from their action on ventricular contractility and systemic afterload, they affect pulmonary arteries and veins. Although these effects are crucial in heart failure with coexisting pulmonary hypertension or lung oedema, they are poorly defined, especially in pulmonary veins. Therefore, we investigated the pulmonary vascular effects of adrenoceptor agonists, vasopressin and angiotensin II in the model of precision-cut lung slices that allows simultaneous studies of pulmonary arteries and veins. Materials and Methods Precision-cut lung slices were prepared from guinea pigs and imaged by videomicroscopy. Concentration-response curves of cardiovascular drugs were analysed in pulmonary arteries and veins. Results Pulmonary veins responded stronger than arteries to α1-agonists (contraction) and β2-agonists (relaxation). Notably, inhibition of β2-adrenoceptors unmasked the α1-mimetic effect of norepinephrine and epinephrine in pulmonary veins. Vasopressin and angiotensin II contracted pulmonary veins via V1a and AT1 receptors, respectively, without affecting pulmonary arteries. Discussion Vasopressin and (nor)epinephrine in combination with β2-inhibition caused pulmonary venoconstriction. If applicable in humans, these treatments would enhance capillary hydrostatic pressures and lung oedema, suggesting their cautious use in left heart failure. Vice versa, the prevention of pulmonary venoconstriction by AT1 receptor antagonists might contribute to their beneficial effects seen in left heart failure. Further, α1-mimetic agents might exacerbate pulmonary hypertension and right ventricular failure by contracting pulmonary arteries, whereas vasopressin might not.


Introduction
Treatment of acute and chronic heart failure is based on the therapy with cardiovascular agents that aim at improved ventricular contractility, enhanced coronary perfusion and reduced myocardial oxygen consumption. Importantly however, cardiovascular agents interact with the pulmonary vascular bed and thereby also influence myocardial function: First, contraction of pulmonary arteries (PAs) enhances right ventricular afterload and worsens right ventricular failure. Second, contraction of pulmonary veins (PVs) increases pulmonary capillary pressure and causes hydrostatic pulmonary oedema and deterioration of gas exchange. Thus, it is clinically important how PAs and PVs respond to cardiovascular agents. However, the differential effects of cardiovascular drugs along the pulmonary vascular bed are only incompletely defined. Most previous studies focused on PAs [1][2][3][4][5], probably due to their central role in pulmonary hypertension and right ventricular failure. Recently, PVs are receiving growing attention and their relevance in the regulation of total pulmonary vascular resistance is becoming evident [6]. Therefore, and due to completely different responses of PAs and PVs [7], simultaneous studies of both vessels are of great clinical interest; however, they are rare [4]. Further, pulmonary vessels differ from systemic vessels in their response to hypoxia, hypercapnia and acidosis [8], thus results from systemic vessels may not be applicable to the low pressure pulmonary vascular bed.
The aim of this study was to investigate the effects of adrenoceptor agonists, vasopressin and angiotensin II on PAs and PVs. We have chosen the model of precision-cut lung slices (PCLS), because it permits simultaneous studies of PAs and PVs. Further, guinea pigs (GPs) were chosen, because previous studies on airway pharmacology suggest that GPs may be a reasonable proxy of human lung tissue [9]. Our results indicate that GPs' PAs and PVs respond significantly different to adrenoceptor agonists, vasopressin and angiotension II. These findings suggest that differential effects of cardiovascular drugs along the pulmonary vascular tree might influence the success of heart failure therapy.

Guinea pigs (GPs)
Female Dunkin Hartley GPs (400650 g) were obtained from Charles River (Sulzfeld, Germany) and held under standard conditions. All animal care and experimental procedures were performed according to the rules of the University Hospital Aachen (Aachen, Germany) and the Directive 2010/63/EU of the European Parliament. They were approved by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen (LANUV, approval-ID: 8.87-51.05.20.10.245).

Precision-cut lung slices (PCLS)
PCLS from GPs (n = 39) were prepared as described before [9]. In brief, intraperitoneal anaesthesia was performed with 95 mg kg 21 pentobarbital (Narcoren; Garbsen, Germany) and its depth was monitored by missing reflexes. Afterwards, the abdomen was opened and the GP exsanguinated. Further, the trachea was cannulated, the diaphragm opened and the lungs filled with low melting point agarose (final concentration: 1.5%), containing 1 mM isoproterenol. To solidify the agarose, the lungs were covered with ice. The lobes were removed; tissue cores prepared and cut into 300 mm thick slices with a Krumdieck tissue slicer (

Identification of the vessels, histology
Pulmonary vessels were identified using the following criteria: PAs accompany the airways and PVs lie aside. After staining with haematoxylin and eosin (HE) PAs show a wrinkled inner lining and a thick media wall [7], as it is illustrated in Fig. 1A-D. Thus, after termination of the experiments PCLS were fixed in 4% formalin and embedded in paraffin. Sections (4 mM) were cut and HE-stained. Images were taken by a microscope (Leica, DM 600 B).

Measurements and Imaging
At the beginning of the study, all agents were investigated with regard to the onset of their maximal contractile or relaxant effect. According to these results, the duration of exposure was defined for all agents, i.e. for (nor)epinephrine, isoproterenol, phenylephrine, procaterol, denopamine, CL 316243 and A 61603 3 minutes, for vasopressin 5 minutes and for angiotensin 10 minutes. The slices were exposed to different drugs on day one and two after preparation. Concentration-response curves were performed on pulmonary arteries and veins and cross sectional area of the vessels was calculated. Control experiments were performed on consecutive sections. Pulmonary vessels were imaged and digitised using a digital video camera (Leica Viscam 1280 or Leica DFC 280). The images were analysed with Optimas 6.5 (Media Cybernetics, Bothell, WA).

Statistics
Statistics was conducted using SAS software version 9.1 (SAS Institute, Cary, North Carolina, USA) and GraphPad Prism version 5.0 (GraphPad Software, La Jolla, USA). Homoscedasticity of values was evaluated. Changes of the vessel area are expressed as percentage of its initial area. All values are shown as mean 6 SEM. Paired observations were analysed using the one sample t-test or the Wilcoxon signed rank test. Unpaired observations were compared using the Mann-Whitney Test. When the effect of increasing concentrations on the vessels was roughly linear, data were analysed using a linear mixed model analysis (LMM). In case of sigmoidal concentration-response curves; the standard logistic regression model was used to calculate and compare EC 50 values. For LMM and logistic regression the AIC-criterion was used to select the preferred model. All p-values were adjusted for multiple comparisons by the false discovery rate [10]. P-values ,0.05 were considered as significant. For all experiments, (n) indicates the numbers of animals.

Stimulation of aand b-receptors in PAs and PVs
The endogenous vasoactive compounds norepinephrine and epinephrine [(nor)epinephrine] represent cardiovascular agents, that are worldwide most commonly used to restore circulation in cardiac failure and shock.
In PVs, isoproterenol, procaterol, norepinephrine and epinephrine caused relaxation, with the following EC 50 values: isoproterenol 0.26 mM, procaterol 0.12 mM, norepinephrine 30 mM and epinephrine 1 mM (Fig. 3A). Simultaneous treatment with epinephrine and the pure b 2 -agonist procaterol did not enhance epinephrineinduced relaxation (Fig. 3B). Further, combined treatment with norepinephrine and procaterol did not alter the maximal relaxant effect of norepinephrine; however EC 50 values were shifted leftwards to lower concentrations, i.e. EC 50 values were 8.6 nM for simultaneous treatment with procaterol and norepinephrine instead of 30 mM for norepinephrine alone. Pre-treatment with 100 nM prazosine had no effect alone (Table 1) or on epinephrineinduced relaxation (Fig. 3B), but enhanced the effect of norepinephrine (Fig. 3C). When PVs were pre-treated with the b 2antagonist ICI 118551 (10 mM or 100 nM) the effect of (nor)epinephrine was reversed and it became contractile ( Fig. 3B/C). At 1 mM epinephrine showed some relaxation despite the presence of ICI 118551 (Fig. 3B). Similar results were obtained for (nor)epinephrine after pre-treatment with 1 mM propanolol (not shown). Neither ICI 118551 nor propanolol did affect the basal tone of PVs ( Table 1). The a 1 -agonist A 61603 contracted PVs, whereas the a 1agonist phenylephrine had only a slight contractile effect at 10 mM that was reversed at higher concentrations (Fig. 3D). However, after pre-treatment with 100 nM ICI 118551 PVs responded to phenylephrine comparable to A 61603 (Fig. 3D). In addition, the a 2 -agonist clonidine, the b 1 -agonist denopamine and the b 3 -agonst CL 316243 had no effects in PVs (not shown).

Effects of angiotensin II on PAs and PVs
Angiotensin II induces vasoconstriction via the G-protein coupled AT 1 receptor. Angiotensin II contracted PVs (EC 50 : 0.1 nM; Fig. 6B-D), but not PAs (Fig. 6A). Pre-treatment with indomethacin (10 mM) or L-NAME (100 mM) did not alter the response of pulmonary vessels to angiotensin II (Fig. 6A-C). In order to analyse, whether angiotensin-induced contraction is specific to AT 1 binding, PCLS were pre-treated with the AT 1 antagonist losartan. Losartan (1 mM) had no effect alone, but abolished the effect of angiotensin II (Fig. 6D, Table 1).

Discussion
(Nor)epinephrine and vasopressin are clinically relevant cardiovascular agents that are daily applied in the treatment of acute haemodynamic instability. Further, angiotensin-converting enzyme inhibitors and AT 1 antagonists are used to treat chronic heart failure. Commonly, their effects on systemic circulation are well controlled; whereas their pulmonary vascular effects are rarely assessed, especially in PVs. However this entity is of relevance, especially in pulmonary hypertension and right heart failure. This study compared the pulmonary vascular effects of (nor)epinephrine, vasopressin and angiotensin II in PCLS of GPs. Adrenergic agonists contracted PAs and PVs, but their contractile effect on PVs was unmasked only in the presence of b 2 -inhibition. Vasopressin and angiotensin II only contracted PVs.  Alpha -and b-adrenergic stimulation (Nor)epinephrine interacted with a 1 /b 1/2 -adrenoceptors of PAs and PVs. (Nor)epinephrine-related contraction of PAs was inhibited by prazosine and thus most likely caused by activation of a 1 -adrenoceptors. In addition, inhibition of ß 1 -adrenoceptors abolished epinephrine-induced contraction, suggesting their involvement in a 1 -mediated contraction. This somewhat surprising conclusion is supported by two findings: First, a 1 /b 1/2 -agonists such as (nor)epinephrine contracted PAs stronger than phenylephrine and A 61603, which selectively act on a 1 -adrenoceptors in PAs. Second, the selective b 2 -agonist procaterol relaxed PAs, whereas the b 1/2 -agonist isoproterenol did not, suggesting neutralization of b 1 -mediated contraction by b 2 -mediated relaxation. Concurrent activation of a 1 /b 1 -adrenoceptors was before reported in PVs [11], albeit with regard to ectopic activity. Previously, isoproterenol that binds stronger on b 1 -adrenoceptors than on b 2 -adrenoceptors [12] was observed to contract vessels [13]. Of note, the b 1 -agonist denopamine failed to contract PAs or to enhance phenylephrine-induced contraction. These findings suggest a complex and indirect activation of b 1 -adrenoceptors by epinephrine that remains to be further elucidated.
Interestingly, the contractile potency of epinephrine decreased above 1 mM indicating the additional activation of b 2 -adrenoceptors and subsequent vasorelaxation. This conclusion is supported by the effects of procaterol that relaxed PAs and prevented epinephrine-and norepinephrine-induced contraction. Further, treatment of PAs with ICI 118551 enhanced norepinephrinerelated contraction. Thus, vasoconstriction appears to be partially masked by stimulation of b 2 -adrenoceptors. In contrast, ICI 118551 did not alter epinephrine-induced contraction. Possibly, epinephrine at 1 mM mainly acts as a relatively pure a 1 -agonist, whereas above 1 mM it competes with ICI 118551 for b 2adrenoceptors.
Thus, adrenergic agents interact in a complex manner with a 1 / b 1/2 -adrenoceptors of PAs (Table 2). Contraction is mainly mediated by a 1 -adrenoceptors, but is antagonised by b 2 -adrenoceptors and aggravated by b 1 -adrenoceptors.
The pulmonary venous vascular bed contributes up to 40% to total pulmonary vascular resistance [6]. In PVs, b 2 -adrenoceptors appear to be dominant and mask the activation of a 1adrenoceptors (Table 2). This is concluded from the observation that PVs contracted to (nor)epinephrine and phenylephrine only in   the presence of b 2 -blockers. Given alone, (nor)epinephrine was relaxant, whereas phenylephrine was not, which is in line with its weak b 2 -affinity. Pre-treatment with prazosine only enhanced the relaxant effect of norepinephrine. Probably, maximal relaxation of epinephrine was already reached, as it stimulates b 2 -adrenoceptors more potently than norepinephrine [12]. Further, combined treatment with procaterol and epinephrine did not alter epinephrine-induced relaxation in PVs, probably due to similar binding affinities in respect of the b 2 -receptor [14]. However, as expected, simultaneous treatment with procaterol and norepi-   nephrine was superior compared to norepinephrine alone and is in line with the high binding affinity of procaterol to the b 2 -receptor [14]. Our results may help to put previous findings into perspective. In perfused feline lung lobes, norepinephrine and phenylephrine enhanced pulmonary vascular resistance; while epinephrine and isoproterenol had the opposite effect [15]. These results might reflect the net effect of a 1 -dependent pulmonary arterial contraction versus b 2 -dependent pulmonary venous relaxation. Though, in perfused rat lungs norepinephrine and phenylephrine reduced the pulmonary perfusion pressure [5], likely due to b 2mediated relaxation [16].
Beta 2 -mediated pulmonary venous relaxation is of clinical interest, as mixed b 1/2 -blockers are widely-used. Commonly, a 1 / ß 1/2 -agonists, such as (nor)epinephrine are applied in heart failure or shock. However, if these patients are pre-treated with mixed b 1/2 -blockers, circulation support with (nor)epinephrine might worsen gas exchange, due to pulmonary venoconstriction, as (nor)epinephrine mainly activate a 1 -adrenoceptors, while ß 2adrenoceptors are still blocked. In pulmonary hypertension, (nor)epinephrine might increase right ventricular afterload and aggravate right ventricular failure. If these findings could be confirmed in humans, this should be considered in the therapy of heart failure.

Vasopressin
Vasopressin has various physiological functions, including V 1a receptor-mediated regulation of blood pressure and V 2 receptormediated control of body water [17]. Further, its relevance in resuscitation is increasingly discussed. Here, vasopressin only contracted PVs. Inhibition of endothelial NO-synthase (eNOS) tended to enhance this contractile effect and the NO-donor SNAP reversed it in part. Further, L-NAME given alone also contracted PVs. This indicates the critical role of eNOS in PVs as opposed to arteries, similar to observations in human [18] and porcine PVs [19]. Moreover, indomethacin attenuated the effect of vasopressin, indicating its partial action through the release of contractile prostanoids such as thromboxane.
In contrast to our results, vasopressin relaxed PAs in isolated perfused rat lungs [3] and isolated canine pulmonary vessels in dependence to eNOS [1,4]. In dogs, it contracted PAs [2]. Vasopressin contracts vessels via V 1a receptor-mediated phospolipase C activation and IP 3 -signalling [17] and opposing to our results, the involvement of relaxant prostaglandins was shown [20]. Thus, the effect of vasopressin in pulmonary vessels may strongly depend on the studied species [21]. Best to our knowledge, pulmonary venous contraction due to vasopressin was not yet reported.
Interestingly, human data indirectly support our results [22]: application of vasopressin led to enhanced pulmonary capillary wedge pressures (PCWP). Vice versa, application of the V 1a/2 antagonist conivaptan decreased PCWP [23], whereas the V 2 antagonist tolvaptan did not, but increased vasopressin plasma levels [24]. Thus, vasopressin antagonists that do not block V 1a receptors might be problematic. In patients with heart failure, vasopressin plasma levels are increased up to 0.28 nM [25], concentrations that contracted PVs in our in vitro model. Hence, vasopressin-related contraction of PVs might enhance pulmonary hydrostatic pressures, left ventricular preload and wall stress. Taken together, our findings suggest that V 1a antagonists might reduce pulmonary complications in heart failure and further, that vasopressin might not worsen right ventricular afterload (Table 2).

Angiotensin II
Angiotensin II is the key peptide of the renin-angiotensin system and mainly produced in the pulmonary arterial vascular bed [26]. In the present study, angiotensin II only contracted the PVs; neither inhibition of NO-nor prostanoid synthesis altered this response.
In contrast to our results, angiotensin contracted endothelium denuded, isolated PAs of GPs and dogs [4,27,28], whereas canine PVs relaxed [4] or failed to respond [28]. Moreover, indomethacin enhanced the contractile effect of angiotensin in PAs [4]. In line with our results, angiotensin contracted rat PVs [29]. Further, in patients with atrial and ventricular septum defects, angiotensin increased left atrial and pulmonary venous pressures, but did not alter the pulmonary arterial resistance [30]. In addition, only the extent of left to right shunts increased. According to our results, pulmonary venoconstriction might be a reasonable explanation for these observations. Prevention of pulmonary venous contraction by AT 1 antagonists or ACE-inhibitors might contribute to their beneficial effect in heart failure.

PCLS from GPs
Thus, our findings are in line with clinical studies and suggest that PCLS from GPs resemble human pulmonary vascular pharmacology reasonably well, as already indicated for airway pharmacology [9]. This study was performed in vitro and thus excludes factors that affect vascular responses in vivo such as shear stress or embolism. In vivo, the PA can be accessed by catheterization, whereas the access of PVs is more difficult. The PCWP relates to left atrial pressure and to large PVs, but small PVs are not reflected [31]. Further, vascular pressure represents a product of vascular tone and filling. Hence, it is influenced by ventricular contractility. Complementary to in vivo studies, PCLS allow exclusively studying the vascular tone of pulmonary vessels.
For the first time, this study systematically compared the effects of clinically relevant cardiovascular agents simultaneously on PAs and PVs. Our results indicate that PAs and PVs are contracted by a 1 -agonists, while relaxation, which occurs predominantly in PVs, is mediated by b 2 -adrenoceptors. Of note, b 1 -adrenoceptors contribute to adrenergic contraction in PAs. Further, vasopressin and angiotensin target predominantly PVs and thus raise the hypothesis that activation of V 1a receptors and AT 1 receptors favours pulmonary oedema. Thus, clinically, the application of vasopressin in left heart failure should be faced with caution and conversely suggests a beneficial role of V 1a and AT 1 antagonists. In contrast to vasopressin, (nor)epinephrine may increase right ventricular afterload, but not pulmonary oedema, except in patients pre-treated with b 1/2 -inhibitors. Taken together, both vascular beds exhibit important differences in their responses to cardiovascular drugs. In conclusion, successful restoration of circulation in heart failure and shock should take into account the differential effects of cardiovascular agents on PAs and PVs.