CB2-deficiency is associated with a stronger

Author's Personal Copy Life Sciences 215 (2018) 96–105

Contents lists available at ScienceDirect

Life Sciences journal homepage: www.elsevier.com/locate/lifescie

CB2-deficiency is associated with a stronger hypertrophy and remodeling of the right ventricle in a murine model of left pulmonary artery occlusion

T

⁎

Georg Daniel Duerra, ,1, Andreas Feißta,1, Katharina Halbacha, Luise Verfuertha, Christopher Gestricha, Daniela Wenzelb, Andreas Zimmerc, Johannes Breuerd, Oliver Dewalda a

Department of Cardiac Surgery, University Clinical Center Bonn, Germany Institute of Physiology I, Life&Brain Center, University of Bonn, Germany c Institute of Molecular Psychiatry, Life&Brain Center, University of Bonn, Germany d Department of Pediatric Cardiology, University Clinical Center Bonn, Germany b

A R T I C LE I N FO

A B S T R A C T

Keywords: Endocannabinoids Pulmonary hypertension Cardiac remodeling

Aims: Pulmonary hypertension (PH) leads to right ventricular (RV) adaptation and remodeling and has deleterious long-term effects on RV function. The endocannabinoid receptor CB2 has been associated with protective effects in adaptation and remodeling of the left ventricle after ischemia. Therefore, we investigated the role of CB2 receptor in RV adaptation after occlusion of the left pulmonary artery (LPA) in a murine model. Main methods: C57/Bl6 (WT)- and CB2 receptor-deficient (Cnr2−/−)-mice underwent paramedian sternotomy and LPA was occluded using a metal clip. Right heart hemodynamic study (Millar®) preceded organ harvesting for immunohistochemistry and mRNA analysis 7 and 21 days (d) post-occlusion. Key findings: LPA occlusion led to higher RV systolic pressure in Cnr2−/−-hearts, while hemodynamics were comparable with WT-hearts after 21d. Cnr2−/−-hearts showed higher macrophage infiltration and lower interleukin-10 expression after 7 d, but otherwise a comparable inflammatory mediator expression profile. Cardiomyocyte-hypertrophy was stronger in Cnr2−/−-mice, presenting with higher tenascin-C expression than WT-hearts. Planimetry revealed higher collagen area in Cnr2−/−-hearts and small areas of cardiomyocyte-loss. Surrounding cardiomyocytes were cleaved caspase-3- and TUNEL positive in Cnr2−/−-hearts. This was associated by maladaptation of myosin heavy-chain isoforms and lower reactive oxygen scavenger enzymes induction in Cnr2−/−-hearts. We found comparable morphological changes in both lungs between the two genotypes. Significance: LPA occlusion led to increased systolic pressure and adaptation of RV in CB2-deficient mice. CB2 receptor seems to modulate RV adaptation through expression of contractile elements, reactive oxygen scavenger enzymes, and inflammatory response in order to prevent cardiomyocyte apoptosis.

1. Introduction Pulmonary hypertension (PH) is a severe disease with different underlying pathomechanisms, leading to an enhanced vascular resistance associated with augmented blood pressure in the pulmonary arteries. PH is defined by an increase of mean pulmonary arterial pressure (PAP) > 25 mmHg in rest state measured by right heart catheterization [1]. PH can be classified as pre- or post-capillary upon hemodynamic evaluation. Aetiology-based classification describes primary PH, i.e. of unknown origin, in contrast to secondary PH caused by underlying disease, e.g. lung fibrosis or sarcoidosis [2]. An age-standardized death rate between 4.5 and 12.3 per 100,000 people has been

reported in the USA. Over the past decades, our understanding of the underlying pathomechanisms evolved mostly on a progressive obliteration of small pulmonary arteries. Several therapeutic options are aiming to improve endothelial function of the pulmonary arterioles, e.g. epoprostenol with its derivates, endothelin receptor antagonists and phosphodiesterase (PDE)-5 inhibitors [3]. However, none of these therapies is curative and the overall prognosis of PH remains poor. Therefore, further research is needed for better understanding of underlying pathomechanisms. In this regard, a significant impact of the endocannabinoid system (ECS) on the regulation of pulmonary vasculature has been reported. The endocannabinoid anadamide has been identified as a mediator in

⁎

Corresponding author at: Department of Cardiac Surgery, University Clinical Centre Bonn, Sigmund-Freud-Str. 25, D-53105 Bonn, Germany. E-mail address: [email protected] (G.D. Duerr). 1 These authors contributed equally to this paper. https://doi.org/10.1016/j.lfs.2018.11.003 Received 29 August 2018; Received in revised form 28 October 2018; Accepted 2 November 2018 Available online 04 November 2018 0024-3205/ © 2018 Elsevier Inc. All rights reserved.


G.D. Duerr et al.

(Z-fix, 4%, Anatech, Battle Creek, USA). For mRNA isolation the hearts were dissected free from great vessels and atria, and the right ventricle was separated from the left ventricle (LV). Right ventricles were then immediately stored in RNA later (Qiagen, Hilden, Germany) at 4 °C for a maximum of 7 d until further processing.

hypoxic pulmonary vasoconstriction ex vivo while acting via the endocannabinoid receptor CB2 or via degradation by fatty acid amide hydrolase (FAAH) [4,5]. The available data investigated effects on the lungs, but provided no insights on effects of ECS on the myocardial adaptation of the right ventricle. Previous studies have shown cardioprotective effects of the endocannabinoid-CB2 axis in the left ventricle using experimental models of ischemia [6,7]. Currently, several animal models have been established to induce PH where peripheral hemodynamics was altered by pharmaceutial means (e.g. moncrotaline, exposure to cigarette smoke) or hypoxia [4,8–12]. Large animal models have been established where PH was induced by different means: in a swine model, postcapilary PH was created surgically [13]. Systolic volume overload was induced in a goat model of PA banding and adjusted to 0.7 RV-to-aortic pressure ratio [14]. In dogs, the LPA was anastomosed with the descending thoracic aorta, and adjustable banding of the anastomosed PA was performed. The band was gradually loosened over weeks leading to pulmonary artery pressure (PAP) overload [15]. With respect to small animal models, lung embolism was induced via clipping of one femoral vein to form deep vein thrombosis in rats [16]. Partial PA ligation procedure, and PA half-closed clip procedure led to RV fibrosis in rats [17,18]. To date, only very few data exist on RV remodeling after PA banding in transgenic mice. Aim of this study was to establish a new murine model of RV overload with characteristics of PH by complete occlusion of the LPA via metal clip, thus increasing pulmonary vascular resistance. We further examined the role of the endocannabinoid receptor CB2 in this model.

2.3. Millar-Catheter and Fulton's index Right ventricular hemodynamic parameters were measured using Millar® pressure catheter (Millar Instruments, Houston, TX, USA) 21 d after surgery. Comparable to the initial surgery procedure, animals were anesthestisized with isoflurane, buprenorphine and carprofen at above-described doses. A 1.0 F-Catheter was inserted through the right jugular vein, passed through the tricuspid valve and introduced into the right ventricle as described [22]. Measurements were performed after stabilization of hemodynamic parameters via a cardiac pressure analysis program (PVAN 3.5, Millar Instruments). Following this experiment, whole hearts where excised and after removal of the atrioventricular valves we separated the RV free wall and the LV including ventricular septum and weighed them independently. Fulton's index (right ventricular weight/weight of left ventricle plus interventricular septum) as a marker of PH was calculated [4]. 2.4. Histology and immunhistochemistry The zinc-paraformaldehyde fixated organs were embedded in paraffin and each 250 μm a set of ten 5 μm slices was mounted on glass slides. As previously published, basic histological evaluation using hematoxylin and eosin (HE), and quantitative planimetry of collagen using picrosirius red (SR) staining (Sigma, Taufkirchen, D) were performed. [21] We further evaluated the area and circumference of the alveoles as well as the wall thickness and the luminal circumference of the peribronchial arteries of the right lung using HE staining. For immunohistochemical staining, Vectastain Elite Kits and diaminobenzidine (AXXORA, Lörrach, D) were used with following antibodies: MAC2 clone 3/38 rat antibody for macrophages (AXXORA), human/mouse cleaved caspase-3 antibody (Asp 175; R&D Systems, Minneapolis, MN, USA), and tenascin-C rabbit anti-chicken polyclonal antibody (Chemicon, Temecula, CA, USA). Macrophage density was evaluated by manual cell count on four pictures (magnification: 400×) taken from the RV of one slide per mouse.

2. Methods 2.1. Study animals Throughout the entire experimental process protocols approved by local government authorities and according to the EU directive 2010/ 63/EU were adhered to. All mouse experiments were performed on 18–25 g and 8–12 week-old mice. C57/Bl6 wild type (WT)-mice (Charles River, Sulzfeld, Germany) and homozygote CB2 receptor-deficient (Cnr2−/−)-mice were used, as published previously [19]. All mice were sacrificed by cranio-cervical dislocation. 2.2. Mouse model

2.5. TUNEL staining Mice were treated with buprenorphine s.c. (0,1 mg/kg KG Temgesic; Remedix, Friedrichsdorf, D) and carprofen s.c. (5 mg/kg KG, Rimadyl®; Zoetis, Berlin, D) for intra- and perioperative analgesia. For left paramedian thoracotomy, mice were anesthetized with isoflurane (0.8–1.5 Vol% Forene, Abbvie, Wiesbaden, D). In order to get a clear view on the left pulmonary hilus, a partial excision of pericardial fat was performed. By discretely lifting the heart, the left pulmonary hilus was exposed and the LPA was identified. The latter was then completely occluded by a fine metal clip (Horizon™ Clips, Teleflex Medical GmbH, Fellbach, D). The success of total LPA occlusion was visually confirmed in every mouse. After chest closure with Prolene 6-0 suture (Ethicon, Johnson & Johnson Medical GmbH, Norderstedt, D) mice recovered from the surgical procedure. For the first postoperative week, buprenorphine and carprofen were used twice a day for postoperative analgesia at abovedescribed doses. Thereafter, mice were examined once a day and treated only if symptoms of pain were encountered. After 7 or 21 d mice were sacrificed by cranio-cervical dislocation and the heart and lungs were excised. Sham animals underwent left paramedian thoracotomy without implantation of the clip and were allowed to recover for 7 d, because we showed, that surgery-related systemic inflammation was not measurable after this time point [20,21]. For histological experiments whole hearts where excised, blood was rinsed out with ice-cold cardioplegic solution and hearts were fixated in zinc-paraformaldehyde

Nuclei of cells with DNA-fragmentation where stained with TUNEL in Situ Cell Death Detection Kit (POD, Roche, Mannheim, Germany) according to the manufacturer's protocol. The fluorescence signal was converted and then stained with AEC + high sensitivity substrate chromogen, which forms a red end product at the site of target nucleic acid (DAKO, Glostrup, Denmark). Slides were counterstained with Quick hemalaun kit and Eosin (Vector) [23]. TUNEL positive cardiomyocyte count was evaluated by manual cell counting on three pictures (magnification: 400×) taken from one slide of the RV per mouse; cardiomyocytes were identified according to their cross-striation appearance and only spherical, cigar-shaped TUNEL positive nuclei were counted. 2.6. mRNA isolation and Taqman® RT-qPCR mRNA-expression was determined using Taqman® real-time quantitative PCR system (RT-qPCR; Applied Biosystems, Foster City, CA, USA) as previously published [7]. Following a standard protocol total RNA was isolated via phenole/chloroform extraction (Trizol; Invitrogen-Life Technologies, Darmstadt, D). First-strand cDNA was synthesized using the High capacity cDNA transcription kit™ (Applied Biosystems) and random hexameric primers following manufacterer's 97


G.D. Duerr et al.

recommendation. The extent of mRNA-expression was evaluated using Taqman® realtime quantitative PCR system (RT-qPCR, Applied Biosystems), FAM TAMRA®-gene expression assays (Applied Biosystems) and 15 ng of each cDNA sample in a triplet measurement. Result analysis was performed with ABI Prism 7900HT Sequence Detection System and SDS 2.4 Software (Applied Biosystems), again according to manufacturer's protocol. Each target gene expression was normalized to respective sham group and the housekeeping gene GAPDH, as previously described [7].

indicates that clip-occlusion of the LPA leads to hypertrophic response in both genotypes after 21 d. RV cardiomyocyte circumference and area were significantly higher after 7 and 21 d in Cnr2−/−-hearts when compared to their respective WT-groups. Interestingly, atrial natriuretic factor (ANF)-mRNA was significantly upregulated in both genotypes after 7 d when compared to respective shams, but we found no significant difference between the genotypes (Fig. 3I). Therefore, Cnr2−/−-hearts experienced a stronger hypertrophic response in the right ventricle 21 d after occlusion of the LPA.

2.7. Statistical analysis

3.4. Increased collagen deposition in Cnr2−/−-hearts

Statistical analyses were performed with Prism 5.0 (Graphpad Software, La Jolla, Ca, USA). Normal distribution of the data was tested and data presented as mean ± SEM. One-way ANOVA with Student Newman-Keuls post hoc testing was performed. When comparing two groups, unpaired t-test with Welch's correction was chosen as indicated. Differences with P ≤ 0.05 were considered statistically significant.

Collagen staining of the right ventricle revealed small areas of cardiomyocyte loss in Cnr2−/−-hearts after 21 d, when compared to their respective WT group (Fig. 4A–D). Planimetric evaluation of the total collagen area in the RV showed significantly higher collagen deposition in Cnr2−/−- when compared to WT-mice at day 7 (Fig. 4E). Collagen deposition went along with development of hypertrophy, since 7d Cnr2−/−-heart and both 21d groups had significantly larger collagen area in the RV than the respective sham groups. Expression of TGF-β mRNA was not different between both groups, while tenascin C (TNC) mRNA was significantly upregulated in absence of CB2 after 7 d (Fig. 4F, G). Therefore, Cnr2−/−-heart presented with more collagen deposition and small areas of cardiomyocyte loss after 21 d.

3. Results 3.1. LPA occlusion induces pulmonary hypertension Basic histology using hematoxylin and eosin (HE) staining showed low cellularity in morphologically intact myocardium after 7 and 21 d in WT. In contrast, Cnr2−/−-mice presented with a stronger cellular infiltration of interstitial space (Fig. 1A–F). RV pressure measurements with Millar® catheter revealed significantly higher systolic RV pressure in Cnr2−/−-mice 21 d after LPA-clip when compared to their respective shams (Fig. 1G). At the same time, RV pressure was also significantly higher in Cnr2−/−- when compared to WT-mice. Diastolic RV pressure was not significantly different between the genotypes. Fulton's index, as a marker of RV hypertrophy, was significantly higher in Cnr2−/−-mice after LPA-clip when compared to their shams (t-test with Welch's correction, Fig. 1H). These results show, that occlusion of the LPA induces characteristics of PH and RV hypertrophy after 21 d in Cnr2−/−-mice.

3.5. Increased cardiomyocyte apoptosis in Cnr2−/−-hearts Measurement of the mRNA expression of apoptotic mediators showed a significantly stronger up regulation of cleaved caspase-3 in Cnr2−/−-hearts 7 d after LPA-clip when compared to their shams (Fig. 5A). The observed loss of cardiomyocytes was further investigated using staining for apoptosis. We found a strong cleaved caspase-3 signal in the RV of Cnr2−/−-hearts already after 7 d compared to WT (Fig. 5B, C). To further corroborate this finding, we performed TUNEL staining and found significantly higher number of TUNEL positive cardiomyocyte nuclei in CB2-deficient RV when compared to WT after 7 d, and to respective shams (Fig. 5D–F). These data revealed an increased apoptotic loss of cardiomyocytes due to hypertrophy in Cnr2−/−-hearts when compared to WT-hearts.

3.2. CB2-deficient mice were associated with stronger inflammatory response In the next step we identified the cells infiltrating the myocardium using macrophage staining and found a significantly higher macrophage density in Cnr2−/−- when compared to WT-mice after 7 d (Fig. 2A–C). The macrophage density was comparable between the genotypes after 21 d. The mRNA-expression of macrophage-associated chemokine CCL2 was comparable between the genotypes, while the 7 d Cnr2−/−-mice showed significantly higher expression than their respective shams (Fig. 2D). We found a comparable mRNA-expression of CCL2-receptor CCR2 and pro-inflammatory cytokine TNF-α between the genotypes (Fig. 2E, F). The anti-inflammatory cytokine IL-10 mRNA was significantly upregulated in WT after 7 d when compared to Cnr2−/−-hearts (Fig. 2G). Taken together, Cnr2−/−-mice revealed a stronger cellular inflammatory response after 7 d with comparable mediator expression profiles to the WT-mice.

3.6. Impaired adaptation of cardiomyocytes due to CB2-deficiency in hypertrophic RV In the next step we measured expression of protective enzymes associated with myocardial adaptation. Both heme oxygenase (HMOX)-1 and glutathione peroxidase (GPX)-1 mRNA expression were significantly less induced in Cnr2−/−-mice when compared to a robust induction of them in WT samples after 7 d (Fig. 6A & B). The expression of myosin heavy chain (MHC)-isoform α was downregulated in both genotypes after 7 d and 21 d, while only the WT groups were significantly different from their respective shams (Fig. 6C). Interestingly, WT mice experienced a significant upregulation of MHC-β after 7 d when compared to their shams, but we found no induction of it in respective Cnr2−/−-hearts. In the light of our previously published data [7], it seems that maladaptation of the cardiomyocytes represents the mechanistic explanation for the increased apoptosis in Cnr2−/−-hearts after PA-clip.

3.3. CB2-deficiency is associated with a stronger hypertrophy of the RV Mosaic pictures of representative picrosirius red stained slides of whole hearts show a stronger hypertrophic response in Cnr2−/−-hearts after 21 d than in respective WT-hearts (Fig. 3A–D). Measurement of the RV wall thickness confirmed this finding by a significantly thicker free wall in Cnr2−/−-hearts (Fig. 3E). In order to confirm these results, we also measured RV cardiomyocyte circumference and RV cardiomyocyte area. Both parameters showed comparable results between the genotypes (Fig. 3F–H). The significant difference between 7 d WThearts, as well as 21 d WT- and Cnr2−/−-hearts vs. respective shams

3.7. Changes in lung morphology after occlusion of the LPA In the last step we investigated the lung morphology changes in this model. Basic histology of the right and left lung revealed a comparable morphology between both lungs after 21 d in WT mice (Fig. 7A, B). We measured alveolar septal thickness in the right lungs and found no difference between the groups after 21 d (WT: 4,49 ± 0,12 vs. Cnr2−/− 4,28 ± 0,16 μm). At the same time, Cnr2−/−-mice 98


G.D. Duerr et al.

Fig. 1. CB2-deficiency leads to enhanced cellular infiltration of the right ventricular myocardium and higher right ventricular pressure compared to WT after LPA occlusion. Representative HE-stainings of WT and CB2−/−-mice show higher cellular infiltration in CB2-deficient hearts after 7 and 21 d compared to WT and their shams (A–F). Millar® pressure catheter measurements show significantly higher systolic RV pressure in Cnr2−/−-mice 21 d after LPA-clip when compared to their respective shams and to WT-mice at the same time (G). (H) Fulton's index for RV hypertrophy (right ventricular weight/weight of LV plus interventricular septum) was higher in CB2−/−-mice 21 d after LPA-clip compared to sham (t-test with Welch's correction). n = 8–10/group; Scale bar in A–F: 50 μm; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days.

therapeutic option exists. Several animal models have been developed to simulate human pathology of PH utilizing different surgical manipulations or pharmaceutical means [8,9,11,12,24]. Since large animal models and rat models [14,15,25] could not provide broad use of transgenic lines, we developed a mouse model of LPA occlusion based on half-clip model in rats. [17] Our study provides evidence, that occlusion of the LPA is leading to characteristics of a PH and RV adaptation in absence of the CB2 receptor. The CB2 receptor seems to act protectively in this pathology, since its absence is associated with cardiomyocyte loss and apoptosis. In accordance with our previously published data [6,7,26], this seems to be associated to maladaptation of contractile elements and scavenger enzymes for reactive oxygen species, and going along with an increased cellular inflammatory response. In most animal models, PH is induced with drugs (e.g. monocrotaline) or hypoxia, thereby acting through changes in the pulmonary microvasculature, but without affecting the greater arterial pulmonary vessels [27–29]. Although it has been reported that some drugs (e.g. Sildenafil®) were able to partially resolve the induced pathology, there is still a big discrepancy in results observed in animal models and outcome in patients [30]. This leaves room for controversy regarding the drug-based animal models of PH [11,31], and offers potential to our model for further research on pathomechanisms of PH. Our study demonstrates that the occlusion of the LPA leads to characteristics of PH

presented with discrete rarefication and thinning of alveoles in the hyperperfused right lung (Fig. 7C, D). We therefore evaluated the right alveolar area and circumference as markers of alveolar rarification in pulmonary hypertension. We found significantly lower right alveolar area and circumference in WT-mice after 7 d and 21 d than in respective shams (Fig. 7E, F). The decrease in alveolar area and circumference was slower in Cnr2−/−-mice reaching significantly lower values after 21 d, which were then comparable with the respective WT-group areas. We also measured the wall thickness of peribronchial arteriols, which usually become thicker during development of pulmonary hypertension. Basic histology of left and right lung with a focus on peri-bronchial arterioles showed no visible morphological differences after 21 d in both genotypes (Fig. 7G–J). Measurement od right arteriolar wall thickness, as well as luminal circumference, confirmed these finding with no significant differences after 7 and 21 d in both genotypes (Fig. 7K, L). Taken together, morphological changes in the alveoli or in peribronchial arterioles of the overperfused right lungs were not different from the respective left lungs or between the both genotypes. 4. Discussion Despite our growing understanding of the underlying pathophysiology of PH, overall prognosis remains poor as to date no curative 99


G.D. Duerr et al.

Fig. 2. Absence of CB2 leads to enhanced macrophage infiltration and expression of inflammatory mediators in the right heart after LPA occlusion. MAC-2 staining of representative right ventricular sections after 7 d PA-occlusion demonstrates predominant interstitial macrophage infiltration (black arrows) in Cnr2−/−-hearts (A & B). (C) Quantification of MAC-2+ cells confirms higher macrophage infiltration after 7 d only in CB2-deficient right ventricles. Gene expression of macrophage-related chemokine (D) CCL2 and its receptor (E) CCR2, (F) TNF-α and (G) IL-10. n = 8–10/group; Scale bars in A & B: 50 μm; RT-qPCR using Taqman®, mRNA expression is related to shams and GAPDH using comparative ΔΔCt-method; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days.

in Cnr2−/−-mice after 21 days and is associated with a significantly higher RV-pressure than in WT-mice and respective shams, while having no major impact on other hemodynamic parameters. Also, enhanced RV- and cardiomyocyte size, as well as RV hypertrophy-marker Fulton's index in Cnr2−/−-mice substantiate this hypothesis. In contrast, our surgical approach did not lead to sufficient RV afterload to induce significant development of PH in WT-mice. Our data are in accordance with findings of other groups showing similar values for RVpressure in murine and rat models of surgically or hypoxia-induced PH [4,18,32,33]. The ECS seems to be involved in hypoxic pulmonary vasoconstriction [4,5]. Studies reported controversial data on vasodilatory effects of the endocannabinoid anandamide on preconstricted rat and isolated human pulmonary arteries [34,35]. A potential therapeutical benefit of the ECS-mediated modulatory effects on immune cell recruitment has been suggested in this context [36]. At the same time, the cardioprotective effects of the endocannabinoid-CB2 axis have been suggested [37–39], and mechanistically dissected in our previous work in murine models of LV ischemia and reperfusion [6,7,40]. The ECS is also activated in human hypertrophic cardiomyopathy [6,7,26]. These studies also showed that ECS and CB2 receptor are involved in modulation of macrophage function. In the present study, we found significantly higher macrophage density only after 7 days in Cnr2−/−-hearts, being

accompanied by a significantly lower induction of the anti-inflammatory IL-10 expression. At the same time, we found a comparable expression profile of other inflammatory mediators in both genotypes. Still, the lack of anti-inflammatory effects of IL-10 causing a postponed resolution of inflammatory response and a higher macrophage density could be sufficient enough to induce damage to the cardiomyocytes in Cnr2−/−-hearts in this model. The above-mentioned studies support this explanation in addition to the fact, that IL-10 is an important factor in limiting inflammatory reaction after myocardial damage [41]. This probably triggered the increased loss of Cnr2−/−-cardiomyocytes in the RV during its adaptation to the increased afterload by inducing hypertrophy. Our present data show, that the cardiomyocyte hypertrophy was morphologically significantly stronger in Cnr2−/−-hearts and thus confirms their principal capacity to hypertrophy, which was indicated by a significant increase in ANF. Enhanced collagen deposition in the heart has already been described in animal models of PH [14,24]. At the same time this adaptation was malfunctional in our model, as revealed by small areas of dense collagen deposition indicating replacement fibrosis due to cardiomyocyte loss [7,42]. The significant increase in tenascin C expression seems to be linked to this increased collagen deposition, due to its function during early tissue remodeling [43,44]. Our previous work using Cnr2−/−-mice in a model of repetitive 100


G.D. Duerr et al.

Fig. 3. The pressure induced right ventricular hypertrophy in our LPA occlusion mouse model is more intense in absence of CB2. (A–C) Mosaic pictures of representative picrosirius red stained, complete WT and Cnr2−/−-hearts at 10× magnification show LPA occlusion-induced RV hypertrophy in (C) WT – and even more predominant – in (D) CB2-deficient hearts. This finding is corroborated by the measurement of RV wall thickness (E). On a cellular level, cardiomyocyte hypertrophy is demonstrated by means of cardiomyocyte circumference (F) and area (G), which have been measured in picrosirius red stained sections of the RV (H). The mRNA expression of ANF confirms hypertrophic stimulus after LPA occlusion in both genotypes (I). n = 8–10/group; Scale bars in I: 50 μm; RT-qPCR using Taqman®, mRNA expression is related to shams and GAPDH using comparative ΔΔCt-method; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

loss was associated with impaired upgregulation of antioxidative enzymes and maladaptation of contractile elements in Cnr2−/−-hearts [6,7]. Both mechanisms are crucial to preserve myocardial integrity and function in conditions of myocardial stress. Induction of antioxidative enzymes attenuates the burden of reactive oxygen species

ischemia and reperfusion also showed small areas of cardiomyocyte loss being associated with apoptosis [6,7,40]. The cleaved Caspase-3 and TUNEL staining, as well as the mRNA expression of Caspase-3, in the present study confirm this interpretation. We previously showed in murine I/R models that LV cardiomyocyte 101


G.D. Duerr et al.

Fig. 4. CB2-deficiency leads to micro-scar formation and adverse right ventricular remodeling after LPA occlusion. Representative picrosirius red stainings show interstitial fibrosis (arrows in C & D) in WT- and Cnr2−/−-hearts, and delineated areas of dense collagen deposition in micro-infarctions (dotted lines in D), which can only be seen in Cnr2−/−-hearts. Planimetric analysis of picrosirius red stained right ventricular sections (A–D) reveals increased total collagen area after 7 and 21 d in Cnr2−/−-hearts. (E) Gene expression of remodeling-related transforming growth factor (TGF)-β (F) and (G) Tenascin-C (TNC). n = 8–10/group; Scale bars in A–D: 100 μm; RT-qPCR using Taqman®, mRNA expression is related to shams and GAPDH using comparative ΔΔCtmethod; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

associated with local inflammation and tissue remodeling [4,8–11]. We can only speculate about the mechanisms behind these findings, e.g. compensatory widening of the bronchial arteries in the left lung, which we were not able to assess.

(ROS), and repetitive I/R is leading to increased production of ROS [21,45]. We therefore measured the expression of antioxidative mediators and found a preserved mRNA-induction of HMOX-1 and GPX-1 in WT-mice. The lack of their significant upregulation in Cnr2−/−-mice underlines their dependence on ECS and its role in cardioprotection. In damaged myocardium, ATP-consumption is reduced in order to preserve cardiomyocyte integrity by a switch of expression of the more ATP-consuming α-MHC toward the energetically more efficient β-isoform [46]. In our PA-clip model, Cnr2−/−-mice were unable to switch their expression from α-MHC to the less ATP-demanding β-isoform. This results in “overworking” of the cardiomyocytes under poor antioxidative capacity and thus favouring cardiomyocyte apoptosis, as described in other studies [47]. Animal models of PH provided insights in morphological changes of the lungs, e.g. intimal proliferation and obstruction of muscular arteries [11,15,25], as it is widely recognized that structural alterations in the vascular wall contribute to all forms of pulmonary hypertension [48]. We therefore expected morphological changes in the “overperfused” right lungs in our model due to the increase in PA pressure. Interestingly, when compared to published data on sham mice, the moderate RV pressure increase through LPA occlusion in our model led even to a decreased alveolar size and no thickening of peribronchial arterioles in both genotypes after 21 d indicating absence of major morphological changes in the non-occluded right lung. This is in contrast to other models where PH was induced with drugs or hypoxia, which is

5. Conclusion The occlusion of the LPA for 21 d led to induction of characteristics of PH and myocardial hypertrophy of the RV in CB2-deficient murine heart. This animal model led to a stronger cardiomyocyte hypertrophy in absence of the CB2 receptor, but at the cost of subsequent loss of cardiomyocytes. The Cnr2−/−-hearts were unable to modulate their expression profile of myosin heavy chain isoforms and reactive oxygen scavenger enzymes, and their cellular inflammatory response. Our data provide evidence for the cardioprotective role of the endocannabinoidCB2 receptor axis beyond the pathologies in the left ventricle and suggest a role for ECS in the development of PH. The present study opens perspective for further pharmacological studies using synthetic CB2 agonists. 6. Limitations of the study In our previous investigations, we reported that CB2-deficient hearts experienced postponed anti-inflammatory M2a macrophages activation in response to ischemia. The IL-10 data from the present study seem to 102


G.D. Duerr et al.

Fig. 5. Increased cardiomyocyte apoptosis in Cnr2−/−-hearts due to LPA occlusion. (A) RT-qPCR shows significantly increased mRNA expression of caspase-3 7 days after PA-clip in Cnr2−/−-hearts. Representative slides of cleaved Caspase-3 stained sections present (B) only few apoptotic cardiomyocytes in WT-hearts, when compared to (C) numerous cleaved Caspase-3 positive cardiomyocytes in Cnr2−/−-hearts after 7 d (arrows). (D) Quantification of TUNEL positive nuclei confirms significantly more apoptotic cardiomyocytes in CB2-deficient RV compared to WT after 7 d. Representative slides of TUNEL staining reveal higher number of TUNEL positive cardiomyocyte nuclei in (F) Cnr2−/−- compared to (E) WT-hearts after 7 d (arrows). Insets in (E) and (F) are magnifications of the slides that point out the morphology of TUNEL positive cardiomyocyte nuclei. n = 8–10/group; Scale bars in B, C, E, F: 50 μm; RT-qPCR using Taqman®, mRNA expression is related to shams and GAPDH using comparative ΔΔCt-method; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days.

apoptosis in this model. One minor limitation of our study is the lack of littermate breeding.

indicate, that a similar mechanism is involved in the RV in response to LPA occlusion, but we did not investigate the macrophage subtypes here. This should be investigated in future studies. Also, our data do not show direct evidence for a link between the observed molecular changes, e.g. maladaptation of MHC-isoforms and lower oxygen scavenger enzymes induction in Cnr2−/−-hearts, and the increased cardiomyocyte apoptosis. Again, future studies should investigate potential benefits of an antioxidative therapy on the extent of

Conflict of interest statement The authors declare that there are no conflicts of interest.

Fig. 6. Antioxidative capacity and adaptation of contractile elements is impaired in Cnr2−/−-hearts. RT-qPCR shows mRNA expression of (A) heme oxygenase (HMOX)-1, (B) glutathione peroxidase (GPX)1, (C) myosin heavy chain (MHC) isoform α, and (D) β-MHC isoform. n = 8–10/group; RT-qPCR using Taqman®, mRNA expression is related to shams and GAPDH using comparative ΔΔCt-method; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days.

103


G.D. Duerr et al.

Fig. 7. Occlusion of the LPA does not lead to significant changes in lung morphology. (A–D) Representative sections of the left and right lung focussing on alveoles of WT- and CB2-deficient mice. Panel (D) gives the impression of alveolar thinning and rarefication and larger alveolar lumina of the right lung in Cnr2−/−-mice (arrows) compared to WT (B) after 21 d. Assessment of the alveolar area (E) and alveolar circumference (F) in the right lung cannot confirmed this finding. Representative slides of left and right lung tissue with focus on peri-bronchial arterioles (G–J, black arrows) as well as assessment of arteriolar wall thickness (K) and arteriolar luminal circumference (L) do not show differences between both genotypes. n = 8–10/ group; Scale bars in A–D and G–J: 50 μm; brackets indicate P ≤ 0.05 between the genotypes; * indicate P ≤ 0.05 vs. respective sham; d, days.

Acknowledgements

Funding

We thank Christine Peigney for her expert assistance in histology and Michaela Matthey for help with lung perfusions and analysis of Fulton index.

This work was supported by BONFOR grant from Medical School, University of Bonn (A.F.) and by the Research Unit FOR926 from the Deutsche Forschungsgemeinschaft (DFG): O.D., DW (Subproject 8, DE801/2-2), B.L. (Central Project 1, Lu 775/4-2), and A.Z. (Central Project 2, Zi 361/5-2).

104


G.D. Duerr et al.

References

[25] L. Meng, et al., Original rat model of high kinetic unilateral pulmonary hypertension surgically induced by combined surgery, J. Thorac. Cardiovasc. Surg. 146 (5) (2013) 1220–1226 (e1). [26] G.D. Duerr, et al., Myocardial hypertrophy is associated with inflammation and activation of endocannabinoid system in patients with aortic valve stenosis, Life Sci. 92 (20−21) (2013) 976–983. [27] M. Rabinovitch, N. Chesler, R.C. Molthen, Point:Counterpoint: chronic hypoxiainduced pulmonary hypertension does/does not lead to loss of pulmonary vasculature, J. Appl. Physiol. (1985) 103 (4) (2007) 1449–1451. [28] P. McLoughlin, I. McMurtry, Counterpoint: chronic hypoxia-induced pulmonary hypertension does not lead to loss of pulmonary vasculature, J. Appl. Physiol. (1985) 103 (4) (2007) 1451–1453 (discussion 1453-4). [29] K.R. Stenmark, I.F. McMurtry, Vascular remodeling versus vasoconstriction in chronic hypoxic pulmonary hypertension: a time for reappraisal? Circ. Res. 97 (2) (2005) 95–98. [30] T. Itoh, et al., A combination of oral sildenafil and beraprost ameliorates pulmonary hypertension in rats, Am. J. Respir. Crit. Care Med. 169 (1) (2004) 34–38. [31] N.R. Bauer, T.M. Moore, I.F. McMurtry, Rodent models of PAH: are we there yet? Am. J. Phys. Lung Cell. Mol. Phys. 293 (3) (2007) L580–L582. [32] D.M. Tabima, T.A. Hacker, N.C. Chesler, Measuring right ventricular function in the normal and hypertensive mouse hearts using admittance-derived pressure-volume loops, Am. J. Physiol. Heart Circ. Physiol. 299 (6) (2010) H2069–H2075. [33] H.G. Zimmer, et al., Right heart catheterization in rats with pulmonary hypertension and right ventricular hypertrophy, Basic Res. Cardiol. 83 (1) (1988) 48–57. [34] M. Baranowska-Kuczko, et al., Endothelium-dependent mechanisms of the vasodilatory effect of the endocannabinoid, anandamide, in the rat pulmonary artery, Pharmacol. Res. 66 (3) (2012) 251–259. [35] M. Baranowska-Kuczko, et al., Mechanisms of endothelium-dependent relaxation evoked by anandamide in isolated human pulmonary arteries, Naunyn Schmiedeberg's Arch. Pharmacol. 387 (5) (2014) 477–486. [36] C. Turcotte, et al., Impact of cannabis, cannabinoids, and endocannabinoids in the lungs, Front. Pharmacol. 7 (2016) 317. [37] N. Defer, et al., The cannabinoid receptor type 2 promotes cardiac myocyte and fibroblast survival and protects against ischemia/reperfusion-induced cardiomyopathy, FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology 23 (7) (2009) 2120–2130. [38] F. Montecucco, et al., CB(2) cannabinoid receptor activation is cardioprotective in a mouse model of ischemia/reperfusion, J. Mol. Cell. Cardiol. 46 (5) (2009) 612–620. [39] F. Mach, F. Montecucco, S. Steffens, Cannabinoid receptors in acute and chronic complications of atherosclerosis, Br. J. Pharmacol. 153 (2) (2008) 290–298. [40] J.C. Heinemann, et al., CB2 receptor-mediated effects of pro-inflammatory macrophages influence survival of cardiomyocytes, Life Sci. 138 (2015) 18–28. [41] N.G. Frangogiannis, et al., IL-10 is induced in the reperfused myocardium and may modulate the reaction to injury, J. Immunol. 165 (5) (2000) 2798–2808. [42] A. Elsasser, et al., The extracellular matrix in hibernating myocardium—a significant factor causing structural defects and cardiac dysfunction, Mol. Cell. Biochem. 186 (1–2) (1998) 147–158. [43] G.D. Duerr, et al., Cardioprotective effects of osteopontin-1 during development of murine ischemic cardiomyopathy, Biomed. Res. Int. 2014 (2014) 124063. [44] N.G. Frangogiannis, et al., Active interstitial remodeling: an important process in the hibernating human myocardium, J. Am. Coll. Cardiol. 39 (9) (2002) 1468–1474. [45] O. Dewald, et al., Downregulation of peroxisome proliferator-activated receptoralpha gene expression in a mouse model of ischemic cardiomyopathy is dependent on reactive oxygen species and prevents lipotoxicity, Circulation 112 (3) (2005) 407–415. [46] P. Razeghi, et al., Metabolic gene expression in fetal and failing human heart, Circulation 104 (24) (2001) 2923–2931. [47] B. Julien, et al., Antifibrogenic role of the cannabinoid receptor CB2 in the liver, Gastroenterology 128 (3) (2005) 742–755. [48] L.A. Shimoda, S.S. Laurie, Vascular remodeling in pulmonary hypertension, J. Mol. Med. (Berl) 91 (3) (2013) 297–309.

[1] M. Humbert, et al., Definition and classification of pulmonary hypertension, Handb. Exp. Pharmacol. 218 (2013) 3–29. [2] N. Galie, et al., 2015 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension, Rev. Esp. Cardiol. (Engl Ed) 69 (2) (2016) 177. [3] D. Montani, et al., Targeted therapies in pulmonary arterial hypertension, Pharmacol. Ther. 141 (2) (2014) 172–191. [4] D. Wenzel, et al., Endocannabinoid anandamide mediates hypoxic pulmonary vasoconstriction, Proc. Natl. Acad. Sci. U. S. A. 110 (46) (2013) 18710–18715. [5] Z. Liu, et al., CB2 receptor activation ameliorates the proinflammatory activity in acute lung injury induced by paraquat, Biomed. Res. Int. 2014 (2014) 971750. [6] G.D. Duerr, et al., Impaired border zone formation and adverse remodeling after reperfused myocardial infarction in cannabinoid CB2 receptor deficient mice, Life Sci. 138 (2015) 8–17. [7] G.D. Duerr, et al., The endocannabinoid-CB2 receptor axis protects the ischemic heart at the early stage of cardiomyopathy, Basic Res. Cardiol. 109 (4) (2014) 425. [8] X. Jiang, et al., Effect of simvastatin on 5-HT and 5-HTT in a rat model of pulmonary artery hypertension, Cell. Physiol. Biochem. 37 (5) (2015) 1712–1724. [9] M.A. de Raaf, et al., Serotonin transporter is not required for the development of severe pulmonary hypertension in the Sugen hypoxia rat model, Am. J. Phys. Lung Cell. Mol. Phys. 309 (10) (2015) L1164–L1173. [10] N.F. Voelkel, et al., Inhibition of 5-lipoxygenase-activating protein (FLAP) reduces pulmonary vascular reactivity and pulmonary hypertension in hypoxic rats, J. Clin. Invest. 97 (11) (1996) 2491–2498. [11] N.F. Voelkel, R.M. Tuder, Hypoxia-induced pulmonary vascular remodeling: a model for what human disease? J. Clin. Invest. 106 (6) (2000) 733–738. [12] J.G. Gomez-Arroyo, et al., The monocrotaline model of pulmonary hypertension in perspective, Am. J. Phys. Lung Cell. Mol. Phys. 302 (4) (2012) L363–L369. [13] J. Aguero, et al., Characterization of right ventricular remodeling and failure in a chronic pulmonary hypertension model, Am. J. Physiol. Heart Circ. Physiol. 307 (8) (2014) H1204–H1215. [14] L.A. Miana, et al., Reversible pulmonary trunk banding VIII: intermittent overload causes harmless hypertrophy in adult goat, Ann. Thorac. Surg. 95 (4) (2013) 1422–1428. [15] G. Yaginuma, et al., A new experimental model of pulmonary hypertension in dog using an adjustable pulmonary artery band, Nihon Kyobu Geka Gakkai Zasshi 38 (11) (1990) 2194–2201. [16] Y.Q. Ji, et al., Varied response of the pulmonary arterial endothelium in a novel rat model of venous thromboembolism, Chin. Med. J. 126 (1) (2013) 114–117. [17] M. Hirata, et al., Novel model of pulmonary artery banding leading to right heart failure in rats, Biomed. Res. Int. 2015 (2015) 753210. [18] N. Rai, et al., Effect of riociguat and sildenafil on right heart remodeling and function in pressure overload induced model of pulmonary arterial banding, Biomed. Res. Int. 2018 (2018) 3293584. [19] N.E. Buckley, et al., Immunomodulation by cannabinoids is absent in mice deficient for the cannabinoid CB(2) receptor, Eur. J. Pharmacol. 396 (2–3) (2000) 141–149. [20] T.O. Nossuli, et al., A chronic mouse model of myocardial ischemia-reperfusion: essential in cytokine studies, Am. J. Physiol. Heart Circ. Physiol. 278 (4) (2000) H1049–H1055. [21] O. Dewald, et al., Development of murine ischemic cardiomyopathy is associated with a transient inflammatory reaction and depends on reactive oxygen species, Proc. Natl. Acad. Sci. U. S. A. 100 (5) (2003) 2700–2705. [22] D. Wenzel, et al., beta(2)-adrenoceptor antagonist ICI 118,551 decreases pulmonary vascular tone in mice via a G(i/o) protein/nitric oxide-coupled pathway, Hypertension 54 (1) (2009) 157–163. [23] G.D. Duerr, et al., Cardiomyocyte specific peroxisome proliferator-activated receptor-alpha overexpression leads to irreversible damage in ischemic murine heart, Life Sci. 102 (2) (2014) 88–97. [24] A. Klinke, et al., Protective effects of 10-nitro-oleic acid in a hypoxia-induced murine model of pulmonary hypertension, Am. J. Respir. Cell Mol. Biol. 51 (1) (2014) 155–162.

105