2.2.1 Histology and immunohistochemistry.

Hearts of all animals were collected at the end of the experiment and first total heart weight was determined. Then right ventricle was removed and left ventricle was weighted again followed by fixation in 4% formalin buffered in PBS pH 7.6 and processed by dehydration in ascending alcohol series followed by xylol and paraffin embedding. For immunohistochemistry, paraffin sections were deparaffinated and rehydrated using xylol followed by descending alcohol series. Haematoxylin/eosin and Sirius Red stain was performed using standard routine protocols. For immunohistochemistry endogenous peroxidase was blocked by incubation with 3% H₂O₂ for 20 min followed by incubation in blocking solution (1% bovine serum albumin (BSA) in Tris(hydroxymethyl)aminomethan-buffered saline supplemented with 0.05% Tween 20 (all purchased from Sigma Aldrich, Deisenhofen, Germany)) for 10 min at room temperature to prevent unspecific binding. The sections were incubated with the following primary antibodies were diluted in blocking solution at 4°C overnight: a rabbit polyclonal antibody directed against collagen I (Biotrend Chemikalien GmbH, Köln, Germany); a goat polyclonal antibody directed against human collagen IV (Southern Biotech, Birmingham, AL, USA), a mouse monoclonal antibody directed against alpha smooth muscle actin (SMA) (DAKO, Glostrup, Denmark), a rat monoclonal antibody directed against CD31(DAKO, Glostrup, Denmark); a goat polyclonal antibody directed against P-Smad 2/3 (sc-11769; Santa Cruz Biotechnology, Dallas, TX, USA); a rabbit polyclonal antibody directed against p44/42 MAPK (ERK1/2; Thr202/Tyr204) (Cell Signaling Technology Europe B.V., Leiden NL). After washing with Tris-buffer pH 7.6 supplemented with 0.05% Tween 20, sections were incubated for 30 minutes at room temperature with appropriate biotinylated secondary antibodies diluted 1:500 in blocking solution: a horse anti mouse IgG (BA-2001); a rabbit anti-goat IgG (BA-5000); a goat anti-rabbit IgG (BA-1000) (all purchased from Vector laboratories). Bound secondary antibodies were detected using ABC-Kit and DAB-Immpact as a substrate (both from Vector laboratories). Finally, nuclei were counter-stained using hemalaun (Merck KGaA, Darmstadt, Germany) and sections were covered with Entellan (Merck KGaA, Darmstadt, Germany) and a coverslip.

2.2.2 Quantification of Sirius Red stain, immunohistochemistry and grading of histopathological changes.

HE stained heart sections were used to analyze LVH. For this purpose, the total and lumen area of the median sectioned hearts were measured using CellSens software (Olympus Deutschland GmbH, Hamburg Germany). From this, the ratio of total to lumen area was also calculated. In addition, wall thickness of intramyocardial arterioles was measured at 10 different representative sites and averaged. Sirius Red was used for analysis of cardiac fibrosis. Fibrosis was determined using 4 different stainings. Quantification of Sirius Red and α-smooth muscle actin (SMA) stained fibrotic areas, was performed by computer-assisted automatic color recognition using Tissue Studio from Definiens (Definiens, Carlsbad, Califonia, US) with MatLab (The MatchWorks inc., Natick, Massachusetts, US) after the sections were digitized with a Slidescanner (Zeiss Z1, Zeiss, Oberkochen, Germany). Sirius Red stained area was expressed as percentage of the total cardiac tissue section area while the percentage of SMA-positive cells of all cells was determined.

In addition, collagen I and collagen IV-positive areas were scored semi-quantitatively. For each visual field, the degree of fibrosis was scored from 0–4. For this purpose, sections were analyzed with a 10x10 eyepiece grid and the least squares were determined with a positive stain. Fields of view with less than 6 squares were classified as score 0, with 6–10 as score 1, with 11–20 as score 2, with 21–30 squares as score 3 and further scaling was assigned in equal steps. Alpha-SMA was stained to detect myofibroblasts in heart tissue. For evaluation, all preparations were digitised and the percentage of α-SMA-positive cells of the total number of cardiac cells was analysed.

CD31 staining was used to detect cardiac capillary number per cardiomyocytes. For this purpose, the number of cardiomyocytes and capillaries was determined in 10 fields of vision with transversely sectioned cardiomyocytes at a magnification of 400x, followed by calculation of the ratio of capillaries per myocyte.

2.2.3 Analysis of PAR2 effects using isolated neonatal cardiac rat fibroblasts.

Hearts from 3-day-old postnatal rats were isolated and digested using the gentleMACS Dissociation kit (Miltenyi Biotech GmbH, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. For P3 fibroblast enrichment, cells were pre-plated in DMEM-F12/Glutamax TM-I (Life Technologies, Darmstadt, Germany) with 10% fetal bovine serum (FBS, Biowest, Nuaille, France) and penicillin/streptomycin (100 mg/ml) (Life Technologies). After 1 hour, non-adherent cells were discarded, and fibroblasts were cultured in the same medium for 1 day. Cells were then trypsinized and seeded in p12 wells (1 ml) or p24 wells (0.5 ml) at a density of 2x10^5 cells/ml. After 24 hours, cells were starved in medium containing 0.1% FBS for 24 hours before stimulation with Ang II (1 μM), in the absence or presence of PAR2 inhibitors. PAR2 was inhibited with either a mouse anti-PAR2 monoclonal antibody (1 μg/ml, clone SAM11, Santa Cruz Biotechnology Inc, Santa Cruz, CA, USA) [20] or the PAR2 inhibitor AZ3451 [21] (2 μM, Tocris Bioscience, Bristol, UK). Controls either remained untreated or received an isotype control antibody (1 μg/ml mouse IgG2a, 10-458-C025, Exbio Antibodies, Vestec, Czech Republic). Cells were then harvested and RNA isolated for expression analysis. 3–5 replicates were performed per condition. For immunofluorescence staining, fibroblasts were fixed with 4% formalin for 15 min and permeabilized with ice-cold methanol at 4°C for 10 min after washing in PBS. Fibroblasts were then incubated for 1 hour at 37°C with rabbit monoclonal anti-P-ERK1/2 (4370L, Cell signalling Technology Europe BV, Leiden, The Netherlands) or rabbit polyclonal anti-P-Smad3 (9520, Cell signalling Technology Europe BV) diluted in 1% BSA in PBS pH 7.6. After washing with PBS pH 7.6, the coverslips were incubated with donkey anti-rabbit IgG Alexa Fluor 568 (Life Technologies GmbH, Darmstadt, Germany) diluted 1:500 in 1% BSA in PBS pH 7.6. After another washing step, the cells were coverslipped with TrueView DAPI (Vector Laboratories Inc.). Finally, 5 images per preparation were captured with a 20x objective on a confocal laser scanning microscope (Zeiss 710, Zeiss, Oberkochen, Germany) and the percentage of positive cells was determined using QuPath software version 0.2.3 [22]. Detailed parameters used for evaluation of P-ERK1/2 and P-Smad3-positive cardiac rat fibroblasts are listed in S1 File. Since numbers of positive cells varied from experiment to experiment, these data were normalized to the level of untreated controls.

2.2.4 Quantitative real-time PCR and in situ hybrization.

We analysed levels of RNA expression of various activators of PAR2 as well profibrotic proteins. Real time PCR of PAR2 was also performed to prove effectiveness of knockout procedure. For expression analysis, RNA was isolated from fresh frozen heart or isolated cardiac fibroblasts using the RNeasy Fibrous tissue mini Kit (Qiagen, Venlo, Netherlands). First-strand cDNA was synthesized with TaqMan reverse transcription reagents (Applied Biosystems, Darmstadt, Germany) using random hexamers as primers. Reactions without multiscribe reverse transcriptase were used as negative controls for genomic DNA contamination. PCR was performed with a Step One Plus Sequence Detector System FastSYBR Green Universal PCR Master Mix (Applied Biosystems), as described previously [23]. All samples were run in triplicate. Specific mRNA levels in mice were calculated and normalized to 18S (fw primer: TTGATTAAGTCCCTGCCCTTTGT; rv primer: CGATCCGAGGGCCTCACTA) as housekeeping gene and shown as fold change expression levels using the ΔΔct method. The Primers used for expression analyses of mouse collagen I, FGF23, cardiac troponin, BNP and PAR2 as well as rat collagen 1, collagen 4, SMA, fibronectin and TGF-ß are shown in S1 Table.

Formalin-fixed paraffin-embedded sections of 4μm thickness were stained with RNA-Scope method to get information about the localization as well as the quantity of the PAR2-RNA expression using the PAR2-specific probe Mm-F2rl1 (417541; ACD, Hayward, CA, USA). The hybridization protocol was performed as described previously [24].

4.1.1. LVH and PAR2.

Our first finding showed reduced cardiac hypertrophy in PAR2 deficient mice, linked to lower levels of cardiac FGF23 and cardiac troponin. This aligns with other studies indicating PAR2’s role in hypertrophy. In cultured neonatal rat cardiomyocytes, PAR2 induced hypertrophic growth in a MEK1/2 and p38-dependent manner, and cardiomyocyte-specific overexpression of PAR2 in mice induced heart hypertrophy [18,26]. Treatment of renin-overexpressing mice, as a model of hypertension, with the PAR2 antagonist FSLLRY for 4 weeks resulted in reduced left ventricular wall thickness [27]. However, the effect of PAR2 on hypertrophy depends on level of stimulation since low level stimulation of PAR2 resulted in anti-hypertrophic response [28]. Because we found reduced LVH in PAR2-deficient mice, AngII treatment appears to be more consistent with stronger stimulation. However, at the endpoint of our study we did not find reduced ERK1/2 activation in PAR2-deficient mice; instead, ERK1/2 phosphorylation was increased in healthy PAR2-deficient mice but equally low in both AngII-treated groups. It is likely that ERK1/2 phosphorylation is not permanently increased in the animal model because only a transient increase in ERK phosphorylation was detected in AngII-stimulated cardiomyocytes [26]. In our study, mainly interstitial cells such as fibroblasts and endothelial cells or possibly inflammatory cells are P-ERK1/2 positive. Our in situ hybridization analyses for PAR2 confirm that these cells also carry the PAR2 receptor. Therefore, we also performed cell culture experiments with primary cardiac fibroblasts. Although these experiments showed no significant differences between untreated and PAR2 inhibitor-treated cells, most P-ERK1/2-positive fibroblasts were again found in the samples in which PAR2 was inhibited and not stimulated with AngII. Next, we investigated TGF-β signalling as another potential pathway influenced by PAR2. AngII treatment increases the expression of the cytokine TGF-β, which can mediate hypertrophy in AngII-induced hypertrophy, as TGF-β-deficient mice exhibited reduced LVH compared to control animals [29]. One way PAR2 affects TGF-β signalling is by interacting with TGF-β receptor 1. Absence or downregulation of PAR2 was associated with downregulation of caveolin-1, leading to decreased internalization of pro-fibrotic PAR1 and TGF- β receptor [13]. Smad2/3 phosphorylation, as a marker of TGF-β activation, was variable without any significant differences between groups in the AngII animal experiment. Thus, these data do not help us to understand the cardiac changes in PAR2-deficient mice. However, studies of ERK1/2 and TGF-β pathways have always been performed in cell culture systems and are therefore difficult to translate to a complex, long-term animal experiment. However, our own experiments in vitro using primary cardiac fibroblasts showed no significant differences between groups. In contrast, Tgfb2 mRNA expression was significantly increased by both PAR2 inhibitors used, but only in unstimulated fibroblasts. Interestingly, in PAR2-deficient mice, cardiac vascular smooth muscle cells responded to AngII stimulation with hypertrophy, in contrast to cardiomyocytes. An important role of PAR2 for VSMC is supported by our observation that VSMC express PAR2 at higher levels than cardiomyocytes. However, previous studies in which PAR2 was overexpressed in VSMC and led to proliferation of these cells also suggested a pro-hypertrophic effect for this cell type [30]. Since vascular hypertrophy is again mediated by TGF-β, PAR2-induced TGF-β receptor internalization might explain the unexpected findings.

4.1.2. Cardiac fibrosis and PAR2.

The second finding in our study was, that PAR2 deficiency enhanced cardiac fibrosis after treatment with AngII. Staining for collagen IV, Sirius Red, and alpha-SMA revealed significantly greater left ventricular fibrosis in PAR2-/- mice compared with wt mice after treatment with AngII, suggesting an anti-fibrotic effect of the PAR2 receptor in vivo. In our in vitro experiments using cardiac fibroblasts we observed significantly increased expression of the TGF-ß target genes Col1a1 and Fn1 in the unstimulated cells which was in line with the increased TGF-ß upregulation under these conditions. However, in AngII-stimulated fibroblasts, the PAR2 inhibitor did not further increase the expression of these matrix genes, but rather tended to decrease their expression. The role of PAR2 in the development of fibrosis is still unclear, as both pro- and anti-fibrotic effects have been described. However, PAR2 has shown pro-fibrotic effects in most studies. There are numerous studies showing that cardiac hypertrophy is usually associated with fibroblast activation and fibrosis [31,32], and that both hypertrophy and fibrosis can be reduced by, for example, RAAS blockade [33]. It is therefore unusual that these two processes appear to be separate in our study. In most cases, the two processes of cardiac hypertrophy and fibrosis are coupled, i.e. a reduction in hypertrophy is usually associated with a reduction in fibrosis. However, inhibition of hypertrophy alone may also lead to adverse cardiac effects in the setting of sustained volume overload [34]. If the absence of PAR2 only affects hypertrophy, an activation of fibroblasts would be conceivable, as previous studies in this model with PAR2-deficient mice have shown that blood pressure is not significantly reduced even in PAR2-deficient mice [35]. This means that the fibroblasts may continue to be stimulated by the elevated blood pressure. In contrast to our observation that cardiac fibrosis was significantly higher in PAR2-deficient mice than in wt animals when exposed to the same treatment, the study by Matsuura et al. showed exactly the opposite [36]. Here, significant interstitial fibrosis occurred in the left atrium of wt mice as early as 2 weeks after AngII infusion, which was significantly reduced in PAR2-deficient mice receiving the same treatment [36]. It is unclear why we obtained such contrasting results in the same AngII-induced animal model. As in our experiment, cardiac PAR2 expression after AngII treatment tended to be lower than in healthy controls [36]. However, the experiment was terminated earlier and we examined fibrosis in the left ventricle rather than the left atrium. Data obtained by the same research group in spontaneously hypertensive rats support the pro-fibrotic effect of PAR2 in hypertensive heart injury. Spontaneously hypertensive rats treated with rivaroxaban, which inhibits activated factor X and may therefore reduce the activation of PAR2, also developed less cardiac fibrosis. When PAR2 was inhibited with the PAR2 antagonist FSLLRY in the hypertensive renin transgenic mouse model collagen-3 expression was reduced [27], also supporting the pro-fibrotic effect of PAR2 in hypertension in the heart. It has been shown that PAR2 is expressed on fibroblasts cultured in vitro and that activation of the receptor leads to increased proliferation of these cells [16,37] and induces a pro-fibrotic phenotype in atrial fibroblasts expressing increased collagen and fibronectin [37]. Pro-fibrotic effects of PAR2 were also described in different models of kidney diseases [38] and injury models in other organs [39,40]. In contrast, anti-fibrotic effects of PAR2 are rarely found. For example, PAR2 deficient mice developed increased fibrosis as shown by excessive glomerular collagen deposition in a model of diabetic nephropathy [41]. Furthermore, PAR2 appears to protect against age-related cardiac fibrosis. Endomyocardial biopsies from elderly patients with heart failure and preserved ejection fraction showed that reduced cardiac PAR2 expression was associated with increased myocardial fibrosis [13]. These findings were confirmed in 1-year-old PAR2-deficient mice, which also showed increased age-dependent alpha-smooth muscle actin expression and collagen 1 deposition compared to wt controls [13].

Our study is constrained by the limited ability to investigate PAR2-mediated signalling. The regulation of the PAR2 receptor and its influences in the heart appear to be diverse and partly contradictory. Due to the large number of different activating proteases, receptor-receptor interactions and feedback loops, a variety of effects of the PAR2 receptor are already known.

Since the activation of the receptor results in various, even opposing effects, a precise understanding of the regulation is of crucial importance. Since some of the results in our study showed substantial alterations in cardiac tissue after a very short period of time, the PAR2 receptor may be a potent target for influencing cardiac remodelling with pharmaceuticals. Further studies are needed to elucidate the exact interaction of PAR2 in AngII-induced cardiac injury.