Homeostatic Role of Decorin in Right Ventricular Pressure Overload and Pulmonary Hypertension Induced Remodeling

Visual Abstract

Highlights

• •RV pressure loading induces RV profibrotic signaling and fibrosis associated with RV dysfunction.
• •RV decorin protein levels are decreased in patients with chronic RV pressure loading.
• •RV decorin protein levels are decreased in several animal models of mechanical RV pressure loading and pulmonary arterial hypertension.
• •Decorin protein expression levels positively correlate with cardiac output.
• •Human cardiac fibroblasts overexpressing decorin show diminished collagen-1 secretion in response to mechanical or chemical profibrotic stress, whereas decorin knockout human cardiac fibroblasts show increased collagen-1 secretion in response to stress.

Summary

Right ventricular (RV) pressure loading induces RV profibrotic signaling and fibrosis associated with RV dysfunction. RV decorin protein levels are decreased in patients with chronic RV pressure loading. RV decorin protein levels are also decreased in 4 animal models of mechanical RV pressure loading and pulmonary arterial hypertension. Human cardiac fibroblasts overexpressing decorin show diminished collagen-1 secretion in response to mechanical or chemical profibrotic stress while decorin knockout human cardiac fibroblasts show increased collagen-1 secretion in response to stress. Downregulation of decorin may play a key role in upregulating transforming growth factor-β1 profibrotic signaling and fibrosis that contribute to RV dysfunction in RV pressure loading.

Right ventricular (RV) failure (RVF) determines morbidity and mortality in diverse diseases that pressure load the RV such as congenital heart diseases and pulmonary hypertension (PH). The RV is particularly vulnerable to pressure loading, which leads to progressive RV hypertrophy and ultimately a vicious cycle of worsening RV dilatation, wall stress, fibrosis, dysfunction, and failure.¹ RVF stems from several pathophysiological mechanisms including fibrosis, which contributes to increased myocardial stiffness and myocyte dysfunction.² We, and others, have shown that RV fibrosis is associated with RV dysfunction and worse clinical outcomes in acquired and congenital heart disease.3, 4, 5

Interstitial fibrosis in RV pressure loading is triggered by mechanotransduction, where increased pressure and mechanical wall stress trigger molecular signaling that promotes fibrosis, apoptosis, and a variety of molecular and cellular processes collectively known as remodeling.^6,7 We previously showed that altered RV geometry from pressure loading triggers a cascade of molecular signals in regions of increased wall stress leading to RV fibrosis, apoptosis, and dysfunction; mediated, in part, by the β1-integrin and the transforming growth factor (TGF)-β1 pathways.^8,9 These and other studies focused predominantly on upregulation of pro-hypertrophic and profibrotic signaling.¹⁰ Much less is known on downregulation of molecules that potentially have homeostatic or protective roles in RV mechanotransduction and hence may be harnessed for therapeutic benefit.

Decorin is a matrix proteoglycan synthesized principally by fibroblasts, whose high affinity to collagen maintains appropriate interfibrillar distance and regulates collagen fibrillogenesis.¹¹ Decorin has roles in matrix assembly, cell proliferation, and extracellular matrix (ECM) signaling.¹¹ Moreover, decorin sequesters multiple ECM growth factors, especially TGF-β1,¹² which we have previously shown to be strongly upregulated in RV pressure loading.¹⁰ Decorin plays a homeostatic role in protecting against excessive fibrosis in the lung, liver, joints, and cornea.13, 14, 15 In the left ventricle (LV), decorin suppresses TGF-β1 during inflammation¹⁶ and may attenuate rejection in cardiac transplantation¹⁷; however, its role in RV injury, particularly in response to RV pressure loading, is poorly characterized. Accordingly, the objective of this study was to investigate decorin expression in the pressure-loaded RV and to explore its role in relation to collagen-1 secretion. We hypothesized that RV pressure loading leads to downregulation of RV myocardial decorin and that decorin expression is inversely related to RV function in vivo and profibrotic signaling and fibrosis in cardiac fibroblasts.

Methods

Animal experiments

All animal experiments were approved by the Sickkids Research Institute Animal Care Committee (protocol # 59576) in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 85-23, revised in 2011 by the National Research Council). To adopt a nonbiased approach, we first screened for downregulated transcription of ECM genes in pulmonary artery banding (PAB) pressure-stressed rabbit RV myocardium using messenger RNA gene arrays incorporating 46 matrix genes (RT2 profiler Fibrosis Array, Qiagen).

To confirm protein translation of gene transcription, we used Western blots to assay decorin protein in the same rabbit RV myocardium. Next, to assess gene translation in a different species from the species where ECM gene transcription was assayed and based on the results of the screening matrix gene transcription, we measured decorin protein expression in RV myocardium from 3 distinct rat models of RV pressure loading: PAB and 2 models of pulmonary arterial hypertension using either sugen + hypoxia (SuHx) or monocrotaline (MCT).¹⁸ For the PAB model, male Sprague-Dawley rats weighing 200 to 250 g were randomized to undergo sham surgery or PAB to generate RV pressure loading.¹⁸ Anesthesia was induced by intraperitoneal pentobarbital sodium (60 mg/kg). Animals were intubated and mechanically ventilated with a volume-controlled respirator and oxygen-enriched room air. Positive end-expiratory pressure was maintained at 4 cmH₂O. The pulmonary artery (PA) was accessed through a left thoracotomy. A silk thread was then tied around the PA and an 18-gauge needle. Removal of the needle produced a fixed PA constriction proportional to the needle diameter. The thorax was then closed in layers, and subcutaneous buprenorphine (15 μg/kg) administered for pain relief. Then rats were maintained for 6 weeks and animal growth with a fixed PAB results in progressive PA constriction and RV pressure loading.18, 19, 20 For the SuHx PH model, male Sprague-Dawley rats weighing 200 to 250 g received a single 20-mg/kg subcutaneous injection of the vascular growth factor inhibitor SU5416 followed by 3-weeks’ exposure to hypoxia (10% O₂) and then 3-weeks normoxia.²¹ For the monocrotaline PH model, 200- to 250-g rats were injected subcutaneously once with MCT (60 mg/kg) and then maintained until the terminal experiment at 5 weeks after MCT injection. These 3 RV pressure-overloading models result in comparable RV remodeling and dysfunction.^18,20

Assessment of cardiac function

Echocardiography

Rats were imaged under 2% isoflurane sedation with a 12-MHz phased-array transducer (Vivid E9, GE Healthcare). Two-dimensional frame rates were 275 frames/second. Two-dimensional, M-mode, color Doppler, and conventional pulsed Doppler images were obtained with simultaneous electrocardiogram on the monitor. Digital data were stored for offline analysis (EchoPac, version 8.0, GE Healthcare). RV functional parameters included fractional area change, tricuspid annular plane systolic excursion, and RV myocardial performance index.^22,23

Hemodynamic measurements

RV and LV hemodynamics were measured immediately following echocardiography. A 2F conductance catheter (Millar Instruments, Inc) was inserted into the RV through the apex.²⁴ The maximal rate of ventricular positive and negative pressure development (+dp/dt, −dp/dt), and myocardial contractility (elastance, Ees) were derived from steady-state measurements. We further measured stroke volume and cardiac output using echo and cuvette-calibrated conductance catheter measurements.^18,24 Thereafter, the animal was euthanized by en bloc excision of the heart under deep anesthesia.

Histology

A 5-mm slice of the heart was cross-cut. Tissue was immersion-fixed in 10% formalin for at least 24 hours, dehydrated in graded ethanol, and left overnight in xylene. After paraffin embedding, 4-μm sections were microtome-sectioned (Leica Microsystems A/S).

RV fibrosis

The sections were subjected to histochemical staining with Picrosirius red to visualize interstitial collagen, as previously described.⁸

Immunofluorescence staining for detection of the cell types expressing decorin in the RV

Cryostat sections (5 μm) were prepared, air-dried, and fixed in 4% paraformaldehyde/phosphate-buffered saline (PBS) for 15 minutes. The sections were then incubated with antibody against decorin (MBS9607199, MyBioSource.com), Vimentin (Proteintech Group, Inc), CD31 (BD Pharmingen Technical), and alpha-smooth muscle actin (α-SMA) (Sigma) at 4°C overnight. Sections were then incubated with appropriate fluorescein-conjugated secondary antibodies for 60 minutes at room temperature. Negative controls were performed for all immunological staining by omission of the primary antibody. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI).

Design-based stereological quantification of immunoreactive decorin

Paraffin-embedded tissue blocks of RV free-wall were sectioned according to established principles for serial uniform random sectioning,²⁵ with a minimum of 3 levels per tissue block obtained 250 mm (50 sections) apart. Sections were stained for immunoreactive decorin by automated immunohistochemical staining (Leica Biosystems) using a modification of protocol F (elimination of the post-primary step). Sections were pretreated using heat-mediated antigen retrieval (BOND Epitope Retrieval Solution 1, Leica Biosystems) for 20 minutes, followed by incubation at room temperature with primary antibody for 30 minutes and detection using a horseradish peroxidase–conjugated compact polymer system (Powervision+, Leica Biosystems). Volume fraction (Vv) of decorin-immunoreactive tissue, relative to total parenchymal volume, was quantified using the Area Fraction Fractionator point-counting probe (Stereo Investigator 2020, MBF Bioscience). Sufficient sampling was conducted to obtain coefficients of error (CE) (Gundersen m=1 method) <0.1 and CE²/coefficients of variation (CV)² ratios < 0.5.

Western blot analyses

Protein was extracted from RV free-wall tissue in in vivo experiments and from cultured fibroblasts with lysis buffer and diluted 1:1 with 2× sodium dodecyl sulfate sample buffer (Invitrogen Canada Inc). Equal amounts of protein (20 μg) were loaded in each lane of 10% Tris-Glycinegel. Proteins were separated by electrophoresis and transferred from the gel to an immune-Blot polyvinyl difluoride membrane with an electroblotting apparatus. Membranes were incubated with 5% nonfat dry milk for 1 hour to decrease nonspecific sites and incubated overnight at 4°C with primary antibodies to decorin (ab35378, polyclonal sheep antibody, Abcam), collagen-1, TGF-β1, Smad2/3, and phosphorylated Smad 2 and 3 (Cell Signaling Technology). Samples were then washed and incubated with peroxidase-conjugated secondary antibody and detected by an electrochemiluminescence detection kit (Bio-Rad).

Immunofluorescence staining for cells

After treatment or mechanical stretch, the cells were washed twice with PBS, fixed in 4% paraformaldehyde in PBS for 15 minutes, and blocked in 10% normal bovine serum albumin in PBS for 15 minutes. Then, they were incubated with primary antibodies recognizing collagen-1 at 4°C overnight, followed by 60-minute incubation with the appropriate secondary antibody conjugated with red tetrarhodamine isothiocyanate (Sigma). Nuclei were counterstained blue with DAPI. Secondary antibodies alone were used as an additional control. The cells were washed twice with PBS and mounted onto slides with cover slips.

Decorin overexpression experiments in human cardiac fibroblasts

Given reduced in vivo RV decorin expression in RV pressure loading, concomitant with activated expression of the profibrotic signaling components TGF-β1, phosphorylated Smad-2/3 and their ratio to total Smad-2/3 from in vivo experiments, we next studied the possible effects of biochemical and mechanical stress on confluent cultures of decorin deletion or overexpression in CRISPR/Cas genetically edited human cardiac fibroblasts.

Commercially available primary human cardiac fibroblasts (Cell Applications Inc) were genome edited using Cas9/adeno-associated virus (AAV) to express decorin and green fluorescent protein (GFP) simultaneously connected by a t2A peptide that would be cleaved after translation. Therefore, the modified fibroblasts could be identified by the presence of GFP without the likelihood of changing the post-translation modification or function of decorin. The cassette was inserted in the C-C chemokine receptor type 5 (CCR5) locus. CCR5 is recognized as a nonessential gene and bi-allelic knockout of CCR5 is not known to have any detrimental health outcomes other than increased resistance to HIV infection and susceptibility for West Nile virus infection.²⁶ Previous studies have successfully knocked-in therapeutic proteins in the CCR5 locus of human hematopoietic stem and progenitor cells using CRISPR/Cas9 and phenotypically corrected Mucopolysaccharidosis type I.²⁷ Thus, decorin was overexpressed in the human cardiac fibroblasts by the insertion of an expression cassette consisting of the spleen focus forming virus promoter, decorin complementary DNA, t2A sequence coding a self-cleaving peptide, citrine (GFP) and an SV40 polyA tail (Figure 1A). The constructs were packaged into AAV6 constructs as previously described.^28,29 Cells were cultured using the recommended media. Y-27632, an inhibitor of the Rho-associated coiled coil kinase (ROCK) protein was added (10 mM) to the media. Previous studies with airway and intestinal stem cells have reported improved editing outcomes when cells were treated with ROCK inhibitors.^30,31 We used a previously reported single guide RNA (sgRNA) (5ʹ-GCAGCATAGTGAGCCCAGAA-3ʹ) to induce double-stranded breaks in the CCR5 locus.²⁷ The sgRNA was chemically modified with 2′ O-Methyl 3′ phosphorothioate at the 5ʹ and 3ʹ ends. This modification was previously reported to enhance editing efficiencies in stem cells.³² The decorin construct was flanked by sequences homologous to the active site of the sgRNA on either side (left and right homology arms). Human cardiac fibroblasts were thawed and cultured for 5 days. Cas9 and sgRNA were complexed at a molar ratio of 1:2.5 (6 mg Cas9:3.2 mg sgRNA). Cas9/sgRNA complexes were electroporated using a Lonza-4D electroporator with the CA-137 program. Immediately after electroporation, AAV6 containing the decorin construct was added at a multiplicity of infection of 500,000 particles per cell. Cells were cultured in the presence of ROCK inhibitor for 2 days after electroporation and then cultured in regular media. Cells were analyzed for GFP expression after day 4 and day 10.

Figure 1.

Decorin Overexpression in CRISPR/Cas Genetically Edited Human Cardiac Fibroblasts

(A) Construct expressing decorin and green fluorescent protein (GFP) concurrently. LHA and RHA refer to the left and right homology arms that target the insertion of the construct into DSB created by Cas9/single guide RNA complex in the C-C chemokine receptor type 5 (CCR5) locus. (B) Approximately 30% of the edited cardiac fibroblasts were GFP+ after 10 days. These cells were enriched using fluorescence-activated cell sorting (FACS). On day 14, >85% of sorted cells were GFP+. FSC-A = forward scatter area.

Decorin knockout

Decorin was knocked out in commercially available primary human cardiac fibroblasts (Cell Applications Inc) using a combination of 3 sgRNAs (sgRNA1: AAGAUGGCAUUGACAGCGGA, sgRNA2: UCGCGGUCAUCAGGAACUUC, and sgRNA3: AGGAAACUUGUGCAAGCAGA). Fibroblasts were electroporated with the sgRNA mixture complexed to Cas9 at a molar ratio of 1:2.5. Cas9/sgRNA complexes were electroporated using the same protocol described for knocking in decorin. The percent of alleles with insertions and deletions (INDELs) were quantified after 4 days by amplifying the target locus using polymerase chain reaction followed by Sanger sequencing. The INDELs were quantified using the Synthego ICE bioinformatics tool.³³ INDELs were observed in >70% of alleles.

Stress stimulation of decorin-genetically modified human cardiac fibroblasts

Biochemical stimulation

Decorin knockin and knockout human fibroblasts were seeded as 1 × 10⁵ cells/cm² into 6 wells on glass coverslips in culture media and allowed to grow until confluence. Cells were serum-starved for 1-hour and subjected to: 1) no treatment; or 2) angiotensin II (Ang II) (100 nM) as it is released after mechanical stimulation and induces fibrosis signaling in cardiac fibroblasts.⁸ After 24 hours, cells were washed and fixed in 4% paraformaldehyde in PBS for 15 minutes, or cold methanol for 20 minutes.

Mechanical stretch of decorin-genetically modified human cardiac fibroblasts

To further study the role of decorin in response to mechanical stress, we mimicked the mechanical stretch induced by RV hypertension in primary human cardiac fibroblast cultures (Flexcell 4000).⁸ Cells were plated at a density of 25 × 10⁴ cells/well in Dulbecco’s modified Eagle’s medium (DMEM) and maintained at 37°C in humid air with 5% CO₂. When cells were ∼80% confluent, the culture medium was changed to serum-free DMEM for 2 hours and then changed to 2% serum DMEM at the start of cell stretching. Stretched cell cultures were exposed to cyclic 20% equiaxial stretch at 1.2 Hz for 24 hours in a Flexcell FX-4000 strain unit equipped with loading posts.⁸ Non-stretched human cardiac fibroblasts were cultured on stretch chambers in the same incubator as controls.

Oxygen consumption rate in cultured RV cardiac fibroblasts

As RV fibroblast metabolism has been shown to be an important factor in pulmonary hypertensive RVs, we measured the oxygen consumption rate (OCR) (Agilent Seahorse XF96 Analyzer) in cultured RV cardiac fibroblasts to explore if modulation of energy metabolism is a possible mechanism for decorin’s effects in the RV fibroblast.¹ An extracellular flux analyzer was used to assess the OCR of nontreated rat RV cardiac fibroblasts vs decorin (5 nM)-treated cells. OCRs (in pmole/min) were measured after addition of oligomycin (Oligo, 2 μM), FCCP (2 nM), and RAA (rotenone and antimycin A, 1 μM each) and basal respiration, Oligo (proton leak), FCCP (maximal respiration), spare respiratory capacity, non-mitochondrial oxygen, ATP production, coupling efficiency, and spare respiratory capacity were calculated and expressed per 3 × 10⁴ cells.

Human RV biopsies

Studies using human samples comply with the Declaration of Helsinki and were approved by the local ethical committee (REB # 1000061028). All patients or their legal guardians gave informed consent before the study. We obtained human pressure-loaded RV myocardial specimens from the Hospital for Sick Children biobank. These patients had undergone surgery for tetralogy of Fallot (TOF) or repair of ventricular septal defect with pulmonary stenosis with informed written consent for biobanking and analysis of RV tissue excised at surgery. After excluding tissues from 4 patients with known genetic abnormalities (trisomy 18, trisomy 21, 22q11 del, Noonan syndrome), 8 RV myocardial specimens were obtained at time of surgery, and snap-frozen at −80°C.

Analysis of published RNA-sequencing data

The CLARA web portal³⁴ was used to explore several whole-heart RNA-sequencing (RNA-seq) datasets that have been previously analyzed and published. Murine gene expression was evaluated using the Integrated Mouse-Heart dataset, and human gene expression was evaluated using the Tucker et al. dataset.³⁵ Graphing of feature plots and analysis of marker genes was conducted through CLARA’s built-in tools, via the “Gene expression” and “Marker gene expression” tabs, respectively. Chamber-specific differences in gene expression were evaluated in the Tucker et al. dataset via the “Gene expression by condition” tab. Processing and analysis of single-cell RNA-seq data from murine cardiac fibroblasts was performed as described by Dewar et al.³⁶ Feature plots were created with R Statistical Software (version 4.3.2), using the Seurat version 4 package.³⁷

Statistics

Results are presented as mean ± SEM, unless otherwise specified. The Shapiro-Wilk test was performed to assess whether data were normally distributed. Two groups were compared using Student’s t-test, and >2 groups were compared using analysis of variance or Kruskal-Wallis test with Tukey's or Dunn's post hoc test for multiple pairwise comparisons for normal and non-normally distributed data, respectively. Linear regression (Pearson’s correlation coefficient) was used to evaluate the association between 2 continuous variables. A P value <0.05 was considered statistically significant, and GraphPad Prism 8.0 (GraphPad Software) was used for all analyses. The data underlying this article will be shared on reasonable request to the corresponding author.

Results

Decorin gene transcription in the rabbit PAB model

The PAB model in rabbits achieves two-thirds to near systemic RV pressures.38, 39, 40

In silico gene screening demonstrated that among the 46 matrix genes screened in 3 PAB rabbits, decorin transcription was the most strongly downregulated (−10.2-fold) (Table 1) when compared with shams.

Table 1.

RV Gene Transcription in RV Pressure Loading

Downregulated Genes	Gene Description	Fold Change
Dcn	Decorin	−10.3
Jun	Jun oncogene	−5.8
Bcl2	B-cell CLL/lymphoma 2	−3.8
Egf	Epidermal growth factor	−3.1
Ilk	Integrin-linked kinase	−3.0
Acta2	Smooth muscle alpha-actin	−2.8
Hprt1	Hypoxanthine phosphoribosyltransferase 1	−2.7
Rplp1	Ribosomal protein, large, P1	−2.6
Il13	Interleukin 13	−2.5
Plg	Plasminogen	−2.4
Il13ra2	Interleukin 13 receptor, alpha 2	−2.3
Faslg	Fas ligand (TNF superfamily, member 6)	−2.3
Mmp13	Matrix metallopeptidase 13	−2.3
Hgf	Hepatocyte growth factor	−2.1
Tnf	Tumor necrosis factor (TNF superfamily, member 2)	−2.0

Of 46 genes, depicted are genes whose transcription was downregulated >2.0-fold in pulmonary artery banding PAB rabbits vs sham controls.

TNF = tumor necrosis factor; RV = right ventricular.

RV decorin expression in RV pressure-loading models

To confirm screening results of downregulated decorin gene transcription, we assessed decorin protein levels by Western blot. These were markedly decreased vs sham controls in the same PAB rabbits (Figure 2A). We next evaluated whether decorin was also decreased in RV pressure loading in other species and conditions including PAB rats and MCT and SuHx-induced pulmonary arterial hypertension. Picrosirius red staining indicated a substantial increase in cardiac fibrosis in all pulmonary arterial hypertension models (Figure 2B). Echo and hemodynamic data from all models confirmed that RV systolic pressure was significantly increased (Table 2). Concomitantly, decorin protein levels were significantly decreased (Figure 2A) and immunostaining showed decreased decorin-positive labeling in all models compared with the sham group (Figure 2C). Stereology-based analysis confirmed decreased decorin protein contents in the RV free-wall in SuHx PH and PAB rats compared with sham controls (Figure 2D). These findings demonstrate a consistent downregulation of decorin gene transcription and protein expression in various RV pressure-loading models, including PAB rabbits, PAB rats, and models of pulmonary arterial hypertension induced by MCT and SuHx. The decrease in decorin is associated with increased cardiac fibrosis and elevated RV systolic pressure, highlighting the potential role of decorin in the pathophysiology of RV pressure overload and pulmonary arterial hypertension. To further identify the cell types producing decorin in the RV, we performed immunofluorescent co-staining of RV tissue from sham rats with decorin and several cell-specific markers. The staining showed that decorin was predominantly expressed in cardiac fibroblasts compared with endothelial cells and smooth muscle cells (Figure 2E). These experimental results are consistent with published RNA-seq data using the CLARA web portal, which revealed that decorin is predominantly expressed by cardiac fibroblasts in both murine and human cardiac tissue, with sporadic expression in other cell types. Single-cell RNA-seq analysis further confirmed strong decorin expression in nearly all captured murine cardiac fibroblasts (Figure 2F).

Figure 2.

Downregulation of DCN Levels in Animal Models of RV Pressure Loading

Decorin (DCN) was downregulated in the right ventricular (RV) free-wall tissue in 4 animal models of chronic RV pressure loading. (A) Representative Western blots detecting indicated DCN and their quantitative assessments in the RV of sham and (a) RV pressure-loading models of rabbit pulmonary artery banding (PAB), (b) rat monocrotaline (MCT), (c) rat sugen plus hypoxia (SuHx), and (d) rat PAB compared with sham controls. Data from n = 3-7 animals per group are presented as mean ± SEM. Unpaired Student’s t-test was performed for statistical analysis. (B) Representative Picrosirius red (PSR) staining of rat hearts (a) rat sham, (b) rat MCT, (c) rat SuHx, and (d) rat PAB. PSR staining demonstrates a substantial increase in cardiac fibrosis in all pulmonary arterial hypertension models (B). (C) Immunostaining detection of DCN-positive labeling (brown) in (a) rat sham, (b) rat MCT, (c) rat SuHx, and (d) rat PAB. Scale Bar = 40 μm. (D) Quantification of immunodetection of RV free-wall DCN levels by stereology in rat SuHx pulmonary hypertension (PH) and rat PAB models. Data from n = 7-8 animals per group are presented as mean ± SEM. Unpaired Student’s t-test was performed for statistical analysis. (E) Representative immunofluorescence staining for DCN expression in various cardiac cell types of sham rat RV tissue. (a) Positive labeling of DCN (red) and vimentin (green), merged show all fibroblasts colocalized with DCN. (b) Positive labeling of DCN (green) and CD31 (red), arrow shows CD31 colocalizing with DCN. (c) Positive labeling of DCN (green) and α-smooth muscle actin (SMA) (red), arrow shows α-SMA colocalizing with DCN. Scale bar = 40 μm. (F) Published RNA sequencing data for DCN expression in various cardiac cell types. (A.) DC expression in cells from the "mouse integrated" dataset in CLARA. (B) DCN expression in cells from the "human heart" dataset in CLARA. (C.) DCN expression in cardiac fibroblast data published by Dewar et al.³⁶ ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. GAPDH = glyceraldehyde 3-phosphate dehydrogenase.

Table 2.

Hemodynamic and Echocardiographic Data of Rats With RV Pressure Loading

	Sham (n = 5)	MCT (n = 5)	SUHx (n = 3)	PAB (n = 5)
Body weight at terminal experiment, g	479 ± 46	366 ± 56	449 ± 42	428 ± 80
Hemodynamic data
Heart rate, beats/min	289 ± 33	305 ± 33	275 ± 45	308 ± 36
RV systolic pressure, mm Hg	23.3 ± 2.5	71.4 ± 13.4a	75.1 ± 10.7a	68.3 ± 21.1a
RV end-diastolic pressure, mm Hg	1.5 ± 0.82	3.7 ± 2.2	7.0 ± 3.6	7.9 ± 4.1a
RV +dP/dt, mm Hg/s	1162 ± 436	2016 ± 352	2938 ± 478a	2575 ± 946a
RV −dP/dt, mm Hg/s	−777 ± 190	−1790 ± 438a	−2748 ± 166b	−1978 ± 902a
RV Ees, mm Hg/μL	0.070 ± 0.025	0.049 ± 0.054c	0.29 ± 0.16	0.31 ± 0.18
Stroke volume, mL/min	0.35 ± 0.20	0.13 ± 0.06b	0.09 ± 0.08b	0.06 ± 0.02b
Cardiac output, mL/min	104 ± 61	40 ± 21b	27 ± 27b	17 ± 6.4b
Echocardiographic data
RVAd (4 chamber), mm2	19.6 ± 3.6	42.2 ± 18.0	53.3 ± 11.5	60.9 ± 22.0b
RVAs (4 chamber), mm2	11.4 ± 2.2	33.6 ± 15.8	43.3 ± 11.5a	44.6 ± 18.5b
RV FAC, %	41.9 ± 3.1	21.7 ± 5.3a	19.5 ± 4.8a	28.0 ± 10.8
TAPSE, mm	2.6 ± 0.27	1.1 ± 0.40b	1.0 ± 0.34b	1.4 ± 0.59b
RV TDI s', cm/s	7.0 ± 1.4	3.2 ± 0.45a	4.3 ± 0.58	4.3 ± 1.3
RV TDI e’, cm/s	6.0 ± 1.0	3.6 ± 0.55	6.0± 3.0	6.1 ± 1.7
RV E/e’	12.5 ± 2.3	18.9 ± 3.8	13.7 ± 6.0	13.2 ± 5.3
RV MPI	0.14 ± 0.09	0.58 ± 0.15b	0.68 ± 0.19b	0.56 ± 0.14b

Values are mean ± SD for sham (n = 5), MCT (n = 5), SuHx (n = 3), and PAB (n = 5) rats. Data were subjected to one-way analysis of variance with Tukey post hoc analysis for normally distributed variables and the Kruskal-Wallis test with post hoc comparison using the Dunn multiple comparison test for non-normally distributed variables.

+dP/dt = maximal rate of ventricular positive pressure change; −dP/dt = maximal rate of ventricular negative pressure change; e’ = peak early diastolic annular velocity; E/e’ = ratio of tricuspid E wave velocity to e’ velocity; Ees = end-systolic elastance (slope of the end-systolic pulmonary vein relationship); MPI = myocardial performance index; RV = right ventricle; RVAd = right ventricular end-diastolic area; RVAs = right ventricular end-systolic area; RVFAC = right ventricular fractional area change; s’ = peak systolic annular velocity; SAX = short axis; SuHx = SU5416 combined with hypoxia; TAPSE = tricuspid annular plane systolic excursion; TDI = tissue Doppler imaging.

^aP <0.05.

^bP < 0.01 vs Sham.

^cP < 0.05 vs PAB.

Decorin expression is decreased in children with RV outflow obstruction

To evaluate clinical translation of the animal model findings, human RV outflow tract biopsy specimens obtained from patients undergoing surgery for TOF or ventricular septal defect with pulmonary stenosis repair were assayed for decorin protein content. Patient clinical characteristics are shown in Table 3. Patients were classified as having mild, moderate, or severe RV outflow obstruction based on the RV outflow tract pressure gradient by continuous wave Doppler echocardiography. Decorin levels were lower in RV samples from TOF patients with severe RV outflow tract gradients compared with patients with mild or moderate RV outflow gradients (Figure 3).

Table 3.

Clinical Characteristics of TOF Patients With RV Outflow Tract Obstruction and Controls

Patient	Age	Gender	Diagnosis	RVOT Pressure Gradient, mm Hg
1	3 y	Male	TOF	52
2	2 mo	Male	TOF	75
3	4 mo	Female	TOF	82
4	12 y	Female	TOF	92
5	4 mo	Female	TOF	62
6	13 y	Male	TOF	14
7	16 y	Male	vPS+VSD	10
8	8 mo	Male	vPS+VSD	24

RV = right ventricle; RVOT = right ventricular outflow tract; TOF = tetralogy of Fallot; vPS = valvar pulmonary valve stenosis; VSD = ventricular septal defect.

Figure 3.

Reduced DCN Levels in RV of Patients With Severe RV Pressure Loading

DCN expression levels in right ventricular (RV) myocardium are lower in tetralogy of Fallot patients with severe RV pressure overloading (right ventricular outflow tract pressure gradient >70 mm Hg) compared with those of patients with moderate (30-70 mm Hg) and mild (<30 mm Hg) RV pressure loading. Abbreviations as in Figure 2.

Echo, hemodynamic and fibrosis analysis in rat RV pressure-loading models

Systolic and diastolic function derived from echo were consistently impaired in all models compared with controls (Table 2). At the terminal experiment, RV systolic pressures were significantly increased in all rat RV pressure-loading models compared with sham controls. RV end-diastolic pressure in PAB rats was markedly higher vs shams (7.9 ± 4.1 MM Hg vs 1.5 ± 0.82 mm Hg, respectively, P = 0.016). RV myocardial collagen content was significantly increased in PAB compared with shams (12.2% ± 3.8% vs 5.2% ± 0.97%, respectively, P = 0.004). We have previously shown that TGFβ1-Smad signaling pathways are markedly upregulated in these models in association with increased RV myocardial collagen deposition.7, 8, 9, 10^,18 Cardiac output indexed for body weight and stroke volume, as well as stroke volume indexed for body weight among MCT, SuHx (SU5416 combined with hypoxia), and PAB rats were inversely correlated with RV decorin protein contents (Figure 4).

Figure 4.

Correlation Between RV DCN Expression Levels and Cardiac Output in Rats With RV Pressure Loading

Correlation between DCN expression levels and cardiac output indexed for body weight derived from catheter measurements in rats with monocrotaline (n = 5), sugen plus hypoxia (n = 2), and pulmonary artery banding (n = 5). Linear regression analysis was performed for statistical analysis (Pearson’s correlation coefficient). RVCO = right ventricular cardiac output; other abbreviations as in Figure 2.

In vitro decorin gene editing

Overexpression of decorin in human cardiac fibroblasts

On day 10, ∼30% of modified fibroblasts were GFP+ (Figure 1B) compared with control fibroblasts treated with AAV6 but not Cas9, which contained <1% GFP+ cells (not shown). Thus, the 30% modified fibroblasts were likely to have a stably integrated cassette expressing decorin. The GFP+ cells were sorted by fluorescence-activated cell sorting at day 10 and cultured for 4 days. On day 14, >85% of the fluorescence-activated cell sorted cells were GFP+ (Figure 1B). These cells were used for functional assays.

Decorin knock out

Decorin was knocked out in human cardiac fibroblasts using a combination of 3 sgRNAs as described previously. INDELs were observed in >70% of alleles.

Effect of decorin on in vitro chemical and mechanically stressed human cardiac fibroblasts

Biochemical or mechanical stimulation of human cardiac fibroblasts overexpressing decorin

In response to Ang II stimulation or cyclic mechanical stretch, human cardiac fibroblasts significantly upregulated their production of collagen-1, TGF-β1, and pSMAD /total SMAD2/3 compared with controls (Figures 5B and 5C). In contrast, human cardiac fibroblasts overexpressing the decorin gene did not increase expression of collagen-1, collagen-3, TGF-β1, or pSMAD 2,3/total SMAD2/3 in response to Ang II or mechanical stretch (Figures 5B and 5C). Matrix metalloproteinase (MMP-2 protein level was significantly increased in decorin transgenic fibroblasts in both untreated and treated cells. There was no significant change in MMP9 and collagen-3 expression (Figures 5D and 5E).

Figure 5.

Amelioration of Fibrosis Signaling and Collagen-1, MMP Secretion in Human Cardiac Fibroblasts Overexpressing DCN

CRISPR/Cas genetically manipulated human cardiac fibroblasts overexpressing DCN show ameliorated fibrosis signaling and collagen secretion in response to stress. Angiotensin II (Ang II) stimulation of human cardiac fibroblasts upregulates collagen-1 and transforming growth factor (TGF)-β1 expression levels in wild-type (WT) but not in cardiac fibroblasts overexpressing DCN. Ang II and mechanical stretch each increase pSmad2/3/Smad2/3 significantly in WT but not in cardiac fibroblasts overexpressing DCN. Matrix metalloproteinase (MMP)-2 expression level was significantly increased in DCN overexpressing fibroblasts in both untreated and treated cells. There was no significant change in collagen-3 and MMP-9 expression. (A) Schematic of the experimental design. (B) Representative Western blots detecting collagen-1, TGF-β1, pSmad2/3/Smad2/3, DCN, and GAPDH in DCN overexpressing transgenic human cardiac fibroblasts (DCN TG) and nonmodified control human cardiac fibroblasts (WT). (C) Quantification of collagen-1, TGF-β1, pSmad2/3, Smad2/3, DCN in WT, and DCN TG. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. (D) Representative Western blots detecting collagen-3, MMP-2, and MMP-9 in DCN TG human cardiac fibroblast and nonmodified control human cardiac fibroblasts (WT). (E) Quantification of collagen-III, MMP-2, and MMP-9 in WT and DCN TG. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. Data from n = 3-7 samples per group are presented as mean ± SEM. One-way analysis of variance with Tukey’s post hoc test was performed for statistical analysis. Abbreviations as in Figure 2.

Stressed human cardiac fibroblasts underexpressing decorin

Collagen-1 levels in human cardiac fibroblasts treated with Ang II, cyclic mechanical stretch, and in decorin knocked out human cardiac fibroblasts were increased compared with control. In addition, TGF-β1 levels were markedly higher in fibroblasts treated with Ang II or subjected to cyclic mechanical stretch. The expression of phosphorylated SMAD2/3 relative to total SMAD2/3 (pSMAD/total SMAD2/3) was also increased following Ang II treatment and in decorin knockout fibroblasts (Figure 6).

Figure 6.

Upregulation of Profibrotic Signaling and Collagen-1 Secretion in Human Cardiac Fibroblasts and in Genetically Edited DCN KO Human Cardiac Fibroblasts

Ang II stimulation upregulates collagen-1, TGF-β1, pSmad2/3/Smad2/3, and DCN expression levels in human cardiac fibroblasts and in genetically edited DCN knockout (DCN KO) human cardiac fibroblasts. (A) Schematic of the experimental design. (B) Representative Western blots detecting collagen-1, TGF-β1, pSmad2/3, Smad2/3, and GAPDH in DCN KO human cardiac fibroblasts and control human cardiac fibroblasts (WT). (C) Quantitative assessment of collagen-1, TGF-β1, pSmad2/3, Smad2/ in control and DCN KO human cardiac fibroblasts. ∗P < 0.05, ∗∗P < 0.01. Data from n = 5-6 samples per group are presented as mean ± SEM. One-way analysis of variance with Tukey’s post hoc test was performed for statistical analysis. Abbreviations as in Figures 2 and 5.

Immunodetection of collagen-1 fibers in human cardiac fibroblasts over- or underexpressing decorin

Following 24 hours of Ang II stimulation or mechanical stretch, cultured human cardiac fibroblasts significantly upregulated immune-detected collagen-1 fiber production. Collagen-1 production was almost completely suppressed in human cardiac fibroblasts overexpressing decorin and remained low after Ang II stimulation and mechanical stretch. In contrast, production of collagen-1 was upregulated at baseline in decorin knockout human cardiac fibroblasts and further augmented in cells stimulated with Ang II or mechanical stretch (Figure 7).

Figure 7.

Immunodetection of Collagen-1 Levels in Genetically Modified Human Cardiac Fibroblasts Where DCN Is Overexpressed or Knocked Out

Representative micrographs of cultured human cardiac fibroblasts immune-stained with anti-collagen-1 antibody further confirms effects of DCN overexpression (DCN TG) or knockout (DCN KO) on collagen deposition. Immunodetection of collagen-1 deposition (red) with (A) Ang II treatment and (C) mechanical stretch. Bars = 40x. Blue depicts nuclear staining by 4ʹ,6-diamidino-2-phenylindole (DAPI). Green depicts DCN expression. (B, D) Following Ang II stimulation and mechanical stretch, immune-detected collagen-1 fiber production increased in control and DCN KO human cardiac fibroblasts vs a smaller increase in DCN overexpressing cells (DCN TG). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Data from n = 3-5 samples per group are presented as mean ± SEM and analyzed using one-way analysis of variance with Tukey’s post hoc test performed for statistical analysis. Abbreviations as in Figures 2, 5, and 6.

Effect of decorin on RV fibroblast energy metabolism

Decorin treatment did not impact basal OCR, ATP production, proton leak, maximal respiration capacity, or spare respiratory capacity in RV fibroblasts vs nontreated control cells (Figure 8).

Figure 8.

OCR in Cultured Right Ventricular Cardiac Fibroblasts

Demonstrate differences in mitochondrial function as compared with control cardiac fibroblast at baseline (nontreated control) and after exposure to decorin (5 nM) for 24 hours. Oxygen consumption rates (OCR, in pmole/min) were measured after addition of oligomycin (Oligo, 2 μM), FCCP (2 nM), and rotenone and antimycin A (1 μM each) and basal respiration, Oligo (proton leak), FCCP (maximal respiration), spare respiratory capacity, non-mitochondrial oxygen, ATP production, coupling efficiency, and spare respiratory capacity were calculated and expressed per 3 × 10⁴ cells as shown in (A) mitochondrial respiration, (B) mitochondrial respiration profile, and (C) basal respiration, spare respiratory capacity, proton leak respiration, and ATP-linked respiration.

Discussion

RV pressure loading increases wall stress to trigger a mechanotransduction-mediated cascade of β1integrin and TGF-β1 molecular signaling that ultimately result in RV ECM remodeling, increased fibrosis, and dysfunction.7, 8, 9, 10^,18

To date, most research has focused on profibrotic signaling molecules that may constitute clinically relevant clinical targets for antagonism or blockade. In contrast, little is known on the modulation and downregulation of intrinsic protective or homeostatic factors. The results of the current study show that decorin, which physiologically inhibits excessive collagen deposition and fibrosis, is among the most downregulated ECM molecules in the pressure-stressed RV, in several animal models and in clinical human RV pressure-stressed samples; in association with increased TGF-β1 signaling and fibrosis. These results were confirmed at both transcriptional and translational (Western blot, immunohistochemistry) levels. Moreover, we show that mechanical stress of fibroblasts downregulates decorin expression and that genetically overexpressing decorin in human cardiac fibroblasts ameliorates fibrosis signaling, collagen secretion, and upregulates MMP-2 expression. In addition, we confirmed that decorin was predominantly expressed in cardiac fibroblasts compared with endothelial cells and smooth muscle cells, which was consistent with our in vivo RV rat data and human RNA-seq data.

Taken together, these results suggest that the RV fibroblast is the dominant cell type secreting decorin, which is suppressed in RV pressure loading. Collectively, these results provide proof of concept that administration or enhancement of decorin may reduce fibrosis in the pressure-stressed and pulmonary hypertensive RV.

Decorin binds receptors with high affinity for growth factors, particularly the TGFβ family, insulin-like growth factor, epidermal growth factor receptor, and cytokines, such as tumor necrosis factor-alpha. Through these pathways, decorin plays an important homeostatic role in maintaining the ECM. As the ECM affects cardiomyocyte hypertrophy and function, and thereby the ventricles’ contractility and compliance, decorin may play an important role in preserving RV cardiac function and its reduction may enhance progression to RVF.⁴¹

Although there is heterogeneity according to cell type, most studies show a downregulatory effect of TGF-β on decorin.⁴² In LV myocardial infarction, decorin deficiency adversely affects collagen fibril formation, thereby changing scar structure and its mechanical properties.⁴³ In rats with systemic hypertension, recombinant decorin overexpression by AAV vector gene delivery inhibits TGF-β/Smad signaling and attenuates activation of p38-MAPK signaling, leading to reduction of hypertension-induced cardiac fibrosis and hypertrophy and improved cardiac function.⁴⁴ In the pressure-loaded LV from aortic banding, decorin levels are significantly reduced when the chemokine receptor type 5 is knocked out.⁴⁵ In contrast, decorin expression has been found to be elevated in other aortic banding LV pressure-load studies.⁴⁶ Taken together, these studies suggest that appropriate expression of decorin and collagen fibril formation may ameliorate excessive ventricular dilatation.⁴³ As RV dilatation is a prognostic factor in PH,⁴⁷ our results showing downregulation of decorin together with RV dilatation may be important in terms of RV remodeling.

Few studies have investigated decorin in the context of RV disease, and there are scant mechanistic data on decorin in RV dysfunction and specifically in RV pressure loading. Our results are consistent with observations in LV pathologies, as decorin was downregulated in RV pressure loading, in association with increased fibrosis. A previous study in a mouse PAB model, which focused on the role of chemokines in RV pressure loading, found that decorin was mildly, but not statistically significantly decreased after PAB, although a higher molecular weight decorin, presumed by the investigators to reflect glycosylated decorin, was increased compared with controls.⁴⁸ In that study, stimulating cardiac fibroblasts in vitro, with various chemokines, increased glycosylated decorin levels.⁴⁸ That study differed from ours in the species used (mouse vs rat) and the duration and severity of PAB (1 week vs 6 weeks). The mildly decreased (or unchanged) decorin levels after 1 week of RV pressure loading in that study may correspond to the significant decrease we found after the substantially longer duration of RV pressure loading of 6 weeks in both rabbit and rat PAB models. Their results may also reflect the impact of early inflammatory response to surgery and/or PAB with increased chemokine secretion impacting specific decorin isoforms, a question not investigated in the current study. Of note, the same investigators found similar results after aortic banding to induce LV pressure loading.⁴⁵ In keeping with our in vivo results in the human pressure-loaded RV, decorin has been found to be downregulated in LV biopsies from failing human hearts undergoing implantation of LV mechanical support.¹⁷ After 2 weeks of mechanical circulatory support, which is hypothesized to allow LV reverse remodeling,⁴⁹ there was an increase in decorin, concomitantly with a decrease in profibrotic signaling downstream of TGF-β, as reflected by decreased p-SMAD2.¹⁷ Moreover, TGF-β stimulation of LV-derived fibroblasts decreased their collagen production when exposed to decorin.¹⁷ These in vitro results are consistent with our in vitro observations in human cardiac-derived fibroblasts, where decorin overexpression reduced fibrosis signaling and collagen production, whereas collagen production was increased in decorin knockout cells.

Decorin has also been found to be protective after experimental LV infarction.⁴³ After ligation of the left anterior descending coronary artery, decorin-null mice had more heterogeneous collagen fibril size and less organized collagen vs wild-type controls. This was associated with more LV dilatation and dysfunction and RV hypertrophy in the decorin-null animals.⁴³ This suggests that decorin’s protective/homeostatic role is not specific for pressure-load remodeling.

The relative paucity of capillaries in the pressure-loaded and pulmonary hypertensive RV, often referred to as capillary rarefication, is thought to be an important mechanism contributing to RV dysfunction.¹⁸ Therefore, the role of decorin in angiogenesis is of interest. In diabetic rats, overexpression of cardiac decorin was associated with upregulated vascular endothelial growth factor expression and angiogenesis, leading to increased capillary density.⁵⁰ However, in vitro and in vivo, decorin has been shown to have anti-angiogenic properties, possibly through its interaction with thrombospondin.⁵¹ We previously showed RV capillary rarefication in the same RV pressure-loading models used in the current study.¹⁸ This suggests that decorin does not inhibit angiogenesis in RV pressure loading. Our data further suggest that decorin does not modulate fibroblast function through modulation of mitochondrial respiratory metabolism (Figure 8).

We are not aware of prior published literature evaluating decorin in human pressure-loaded RVs. Thus, our results from human RV biopsies from children with RV pressure loading, although from too small a sample to be conclusive, are supportive of the animal and in vitro data in that the patients with the most severe RV outflow tract obstruction had the lowest levels of decorin. These results are in keeping with our in vivo results in the rabbit and rat PAB and PH models. Decorin levels correlated with cardiac output in our animal RV pressure loading models. Cardiac output is an important determinant of mortality in severe RV pressure loading, such as severe PH, and some consider it the sine qua non of RV failure. Cardiac output has multiple determinants; however, myocardial function, and hence ECM remodeling, plays a role.⁵² Thus, relief of RV pressure loading may lead to increased cardiac output through direct hemodynamic effects, but also through improved ECM remodeling. To date, reducing RV fibrosis to improve RV function, by medications such as angiotensin receptor inhibitors has shown promise in experimental models,¹⁰ but not in human trials of RV failure.⁵³ Enhancing intrinsic antifibrotic molecules, such as decorin, may therefore constitute an attractive therapeutic approach. Decorin deficiency in other human diseases has been related to epithelial to mesynchymal transition and promotion of metastatic cancer.⁵⁴ Consequently, decorin treatment has been proposed in cancer.⁴⁷ The protective effects of decorin for fibrosis have been reported in other organs. Decorin knockout diabetic mice develop severe nephropathy with increased levels of TGF-β and collagen-1.⁵⁵ Therapeutically, bleomycin-induced lung fibrosis can be ameliorated by decorin treatment.⁵⁶

Our use of a transgenic human cardiac fibroblast model with reduced collagen secretion by decorin overexpressing fibroblasts, suggests that a transgenic approach, possibly via a viral vector, may be a therapeutic option.⁵⁷ Statins, which enhance decorin expression in cardiac fibroblasts, might be an another clinically effective therapy for attenuating fibrosis in RV with pressure overloading. Simvastatin upregulated decorin expression in Ang II–induced human cardiac myocyte-derived exosomes leading to suppressed migration of human cardiac fibroblasts and less collagen deposition in vitro.⁵⁸ In rat myocardial infarction models, simvastatin treatment reduces TGF-β1 and Smad3 with reduced cardiac fibrosis and improved cardiac function in vivo.⁵⁹ Furthermore, a novel potential delivery system of decorin to the target tissue using injectable hyaluronic acid microrods has recently been reported in a rat model of ischemia-reperfusion myocardial infarction.⁶⁰ In that study, injecting polymeric hyaluronic acid microrods loaded with decorin to the infarcted heart successfully delivered decorin to the target tissue and attenuated fibrosis and cardiomyocyte hypertrophy with improved cardiac function and ventricular remodeling.⁶¹ This novel methodology has the benefit of being injectable, cell-free, and avoids systemic side effects due to its local delivery. Other delivery methods could also be contemplated such as delivery with a viral vector.

Study limitations

We studied wild-type rats and transgenic in vitro fibroblasts. Although a germline decorin knockout mouse is commercially available, the widespread lesions, including reduced dermal tensile strength and healing deficiencies are disadvantageous for the PAB model.

Investigating whether overexpressing decorin can reverse the RV remodeling would have strengthened the study. Consequently, an overexpressing decorin knockin rodent model would have been relevant to study but was beyond our resources. Similarly, exogenous administration of human recombinant in vivo to PAB rats was not feasible because of very high costs. These warrant study in future investigations. Our in vitro transgenic human cardiac fibroblast experiments provide mechanistic data that augment the in vivo results. We used only male animals for in vivo studies as not to double the number of animals. Given that sex is an important biological variable, and based on our current results, it would be important to study female rats. Finally, the small number of RV samples from patients with TOF was a limitation.

Conclusions

We found that decorin transcription and protein expression are downregulated in 4 distinct models of RV pressure stress in association with increased fibrosis signaling, fibrosis, and RV dysfunction. These results correspond to decreased decorin expression in human RV biopsies with severe RV outflow obstruction. In vitro, decorin overexpression ameliorated collagen-1 secretion in response to mechanical and chemical stretch, whereas decorin knockdown increased collagen-1 secretion in human cardiac fibroblasts. Our results suggest that decorin plays an important role in the profibrotic RV remodeling and dysfunction that occur in response to RV pressure loading. Therefore, augmenting decorin may represent a novel therapeutic approach in RV pressure loading and failure.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: PH and other diseases characterized by severe RV pressure loading carry high morbidity and mortality driven by RV failure. RV pressure loading triggers TGF-β1 signaling and fibrosis, which contribute to RV dysfunction. Decorin is a matrix proteoglycan that sequesters TGF-β1 and collagen-1, maintaining matrix composition and countering excess fibrosis. Here, we show that RV decorin is significantly downregulated in 4 animal models of RV pressure loading and pulmonary hypertension; and in pressure-loaded human RV tissue, correlating with reduced cardiac output. Furthermore, overexpressing decorin in human cardiac fibroblasts ameliorates collagen deposition in response to mechanical or Ang II stress, while knocking out decorin in these cells augments collagen deposition in response to profibrotic stress.

TRANSLATIONAL OUTLOOK: Our results suggest that downregulation of decorin contributes to RV fibrosis and dysfunction in PH and congenital heart disease and that restoring decorin to the RV myocardium may be a novel approach to treat RV failure.

Funding Support and Author Disclosures

This work was supported in part by the Heart and Stroke Foundation of Canada and by the Canadian Institutes of Health Research (FRN 16226). Dr Connelly holds the Keenan Chair for Research Leadership, Keenan Research Centre for Biomedical Science, Toronto. All other authors have reported they have no relationships relevant to the contents of this paper to disclose.