Temporal inhibition of ADAM17 in fibroblasts reduces stiffness and promotes vascularization following myocardial infarction

Abstract

Aims

Myocardial infarction (MI) triggers a complex remodeling that, if uncontrolled, leads to heart failure. Increased levels of ADAM17 (disintegrin and metalloproteinase-17) in ischemic injury has been reported, but its direct role in scar formation and subsequent recovery from MI has not been identified. We investigated the role of ADAM17 in the function of homeostatic fibroblasts (FBs) vs. activated myofibroblasts (myoFBs) in scar formation, and recovery following MI.

Methods and Results

Human myocardial specimens showed upregulated ADAM17 in the infarct tissue, colocalized to myofibroblasts. We generated two inducible genetic mouse models with Adam17 knockdown in FBs (Adam17^FB−KD) or myoFB (Adam17^myoFB−KD) and subjected them to MI. Loss of ADAM17 in FBs impaired infarct formation and increased mortality due to left ventricular (LV) rupture in males and females. In contrast, ADAM17 loss in myoFBs limited infarct expansion, LV dilation and dysfunction up to 4-wks post-MI. Macrophage infiltration was suppressed in both genotypes. Ex vivo and in vitro experiments revealed that loss of ADAM17 in myoFB resulted in scar tissue with reduced stiffness due to suppressed activation of epidermal growth factor receptor and the Yes-associated protein (YAP) pathway. This promoted VEGFR signaling, endothelial cell (EC) proliferation, and vascularization in the infarcted myocardium, limiting infarct expansion. RNAseq analyses showed drastic changes in extracellular matrix (ECM) genes in Adam17^KD FB and myoFBs in hypoxia. In vitro co-culture of myoFB and ECs confirmed that the ECM deposited by Adam17-deficient myoFB promotes EC proliferation and sprouting. Pharmacological inhibition of ADAM17 before MI was ineffective, but short-term ADAM17 inhibition after MI (targeting the myoFBs), limited infarct expansion, LV dilation and dysfunction up to 4-weeks post-MI.

Conclusion

Short-term inhibition of ADAM17 after MI optimizes the compliance of the infarct tissue, promoting vascularization, limiting infarct expansion, preventing long-term adverse LV remodeling, dysfunction, and heart failure. Targeting the homeostatic FB vs. myoFB also highlights the critical timing of ADAM17 inhibition as its presence is essential for the initial healing of the infarcted heart, but inhibition of its persistent upregulation reduces scar stiffness and improves the outcome post-MI.

Graphical Abstract

Graphical Abstract.

Time of primary review: 27 days

See the editorial comment for this article ‘Dual roles of ADAM17: from fibroblast foe to myofibroblast friend’, by P. Ranjan and S.K. Verma., https://doi.org/10.1093/cvr/cvag001.

1. Introduction

Ischemic heart disease remains a predominant cause of mortality worldwide. Obstruction of the coronary arteries leads to acute myocardial infarction and heart failure. Following myocardial infarction (MI), a series of time-dependent cellular events dictate the integrity of infarct tissue formation, the structural remodeling of the myocardium, and the subsequent clinical outcomes. Fibroblasts (FBs), also referred to as homeostatic or quiescent FBs, are the principal cell source of extracellular matrix (ECM) proteins, play a critical role in maintaining physiological cardiac structure and function, and can contribute to myocardial remodeling in pathological conditions.^1–3 Upon ischemic injury, FBs in the myocardium become activated and differentiate into myofibroblasts (myoFBs) which possess enhanced ECM synthesis capabilities and are the prominent contributors to the formation of the infarct (scar) tissue.^4–7 A balanced fibrotic response is essential to ensure sufficient repair of the damaged ischemic myocardium without excess ECM deposition that would cause left ventricle (LV) dilation and dysfunction, and eventually heart failure. Hence, the process of FB activation to myoFBs, and subsequently, the extent of activity of myoFBs are key determining factors for constructive remodeling of the ischemic myocardium.

In addition to their direct contribution to scar formation, FBs can regulate myocardial remodeling through their interaction with other cell types in the myocardium, including endothelial cells (EC), inflammatory cells, and cardiomyocytes, either through direct cell-cell contact or through paracrine effects.^8–10 FB-derived growth factors can promote EC proliferation and lumen formation,^11–13 while the stiffness of the scar tissue can influence EC function.^14–17

A disintegrin and metalloproteinase-17 (ADAM17) is a transmembrane sheddase with a broad range of substrates in different cell types,^18,19 and diverse roles in various cell function.²⁰ ADAM17 is expressed in many cell types in the cardiovascular system, including vascular smooth muscle cells,^21–23 endothelial cells,²⁴ cardiomyocytes,²⁵ fibroblasts,²⁶ as well as inflammatory cells.^27–29 ADAM17 can mediate ectodomain shedding of several membrane-bound proteins, thereby regulating their biological functions, intercellular signaling, and paracrine communications.^30–33

In patients with acute MI, elevated ADAM17 levels were detected at 1 day after MI which was decreased at 14 days after MI.³⁴ The increase in ADAM17 levels in patients with acute MI were detected at the site of plaque rupture that caused the MI,³⁵ while another study reported increased ADAM17 levels in the peripheral blood mononucleated cells.³⁶ Studies in animal models have reported increased ADAM17 mRNA expression in rats following MI.³⁷ Cardiomyocyte-specific loss of ADAM17 exacerbated LV dilation and dysfunction due to suppressed angiogenic response,²⁵ but improved LV remodeling and function in diabetic cardiomyopathy by limiting ADAM17-mediated ACE2 shedding.³⁸

In this study, we elucidate the FB-specific functions of ADAM17 in post-MI events by delineating its contribution to the role of FB subtypes and the time course of their contribution to the structural and functional changes post-MI. We report that loss of ADAM17 in homeostatic FBs prior to the onset MI impairs infarct formation and increases the rate of cardiac rupture and mortality. Conversely, ablation of ADAM17 in myoFBs, after MI, results in the formation of infarct tissue with optimal ECM compliance conducive to EC proliferation and vascularization, which limits infarct expansion, LV dilation and dysfunction. Reduced activity of epidermal growth factor receptor (EGFR) in Adam17^myoFB-KD hearts, likely due to reduced bioavailability of EGFR ligands, which are shed by ADAM17,^18,39 resulted in decreased nuclear translocation of YAP and reduced transcription of pro-fibrotic genes. Furthermore, we demonstrate that short-term pharmacological inhibition of ADAM17 after MI is sufficient to exert long-term beneficial effects in cardiac structure and function, identifying an effective therapeutic window for ADAM17 inhibition for optimal post-MI recovery.

2. Methods

Extended methods with additional details are included in the Supplementary File.

2.1. Study approval

This study was conducted in accordance with the guidelines and regulations approved by the University of Alberta. All animal experiments were performed according to the guidelines outlined in ARRIVE (Animal Research: Reporting of in vivo Experiments), Canadian Council of Animal Care (CCAC) and the NIH Guide for the Care and Use of Laboratory Animals. All studies were approved by the University of Alberta Animal Care and Use Committee (ACUC, AUP 396). Human heart specimens (Figure 1E and F) were collected according to the protocol approved by the Mazankowski Alberta Heart Institute and the University of Alberta Hospital (Human Explanted Heart Program, HELP, Pro00011739). Written informed consent was obtained from the patients before procurement of the explanted hearts, and this investigation conformed to the principles outlined in the Declaration of Helsinki.

Figure 1.

ADAM17 (A distintegrin and metalloproteinase-17) is upregulated in ischemic cardiomyopathy (ICM) in patients, and in mice after myocardial infarction (MI). (A-D) Single-nucleus RNA sequencing analysis of publicly available data (Kuppe et al. PMID: 35948637) of human non-failing control (NFC, n = 4), and the infarct zone (IZ; n = 6), border zone (BZ; n = 3) and remote non-ischemic zones (RZ; n = 5) from patients with ICM. Uniform Manifold Approximation and Projection (UMAP) were performed for dimensionality reduction, followed by clustering analysis (Seurat framework). Heatmap illustrates the mean expression levels of the top 30 DEGs in different cardiac cells showing cell specific marker genes. (B) Stacked bar plot shows the distribution of different cardiac cells in NFC and ICM (IZ, BZ and RZ) samples. (C) FeaturePlot and Pseudobulk analysis of ADAM17 in total cells (C) and in fibroblasts (D) in NFC and ICM (IZ, BZ and RZ). (E) Representative immunoblot and protein quantification of ADAM17 protein levels in NFC and ICM patients (n = 7 NFC, 13 ICM). (F) Representative co-immunostaining staining for ADAM17 (red) and periostin (POSTN, green) in NFC and ICM specimens. Representative immunoblot and protein quantification for ADAM17 levels (G) and representative co-immunostaining staining for ADA M17 (red) and periostin (POSTN, green) in mouse sham and MI (1 wk post-MI) hearts. A.U. = Arbitrary units. Student’s test (E) and one-way ANOVA with Bonferroni post-hoc test (G) were performed. Average values represent Mean±SEM. * P < 0.05 vs. NFC or corresponding sham group.

2.2. Human heart specimens

The single-nucleus RNA sequencing data on human samples are based on the publicly available dataset published by Kuppe et al.⁴⁰ The raw data were obtained from the European Genome-phenome Archive Zenodo data archive (https://ega-archive.org/datasets/EGAD00001008952). The dataset includes samples from non-failing controls (n = 4), and from the ischemic zone (IZ), border zone (BZ), and remote zone (RZ) of patients with ischemic heart failure (n = 14) who underwent heart transplantation.⁴⁰

The specimens used for the Western blot and Immunofluorescence staining (Figure 1E, 1F) were collected from patients with ischemic cardiomyopathy (median age 66 years; range: 57–68 years) with symptomatic heart failure (New York Heart Association class III/IV) who underwent heart transplantation.^37,41,42 Ischemic regions were identified based on angiogram results and ECG data. Non-ischemic myocardium was used as the control.

2.3. Experimental animals and surgical procedures

Adam17 ^flox/flox mice⁴³ and periostin-Cre^Esr1 (Postn^MerCreMer; Postn^Cre/+)⁴⁴ mice were purchased from Jackson Laboratory, and Tcf21-Cre^Esr1/Rosa26R^tdTomato (Tcf21^MerCreMer; Tcf21^Cre/+) mice were a kind gift from the Tallquist laboratory.^45,46  Adam17^flox/flox mice were cross-bred with Tcf21^Cre/+ or Postn^Cre/+ mice to generate mice with inducible ADAM17 knockdown in the homeostatic fibroblasts (Adam17^f/f/Tcf21^Cre/+/Rosa26R^tdT or Adam17^FB-KD), or in myofibroblast (Adam17^f/f/Postn^Cre/+, or Adam17^myoFB-KD), and littermate controls (Adam17^f/f). Postn^Cre/+ and Tcf21^Cre/+ mice were used as controls for Cre-recombinase expression. Inducible Cre-recombinase was activated by intraperitoneal injection of tamoxifen for 5 days, as before.⁴⁷ ADAM17 knockdown in homeostatic FBs and myoFBs were confirmed (see Supplementary material online, Figure S2). Wildtype mice (WT, C57Bl6J) were used in experiments with a pharmacological ADAM17 inhibitor. Myocardial infarction (MI) was induced in anesthetized (2% isoflurane) male or female mice by ligation of the left anterior descending artery (LAD) as before,^25,48 with 99% recovery from surgery and 3% mortality (1–2 days post-MI) due to non-LV rupture causes. Mice were euthanized after being anesthetized by Ketamine/Xylazine (100 mg/Kg; 10 mg/Kg) cocktail and excision of the heart.

2.4. Echocardiography and myocardial strain analysis

Cardiac structure and function were assessed by non-invasive transthoracic echocardiography on anesthetized mice (1–1.5% isoflurane) (Vevo 3100 VisualSonics) as before.^25,48,49 Strain analyses were performed using the Vevo LAB software. In vivo infarct size was measured at parasternal long axis by LV wall motion as previously described.^42,50

2.5. Primary adult cardiac fibroblast isolation, culture, and treatments

Primary cardiac fibroblasts (cFBs) were isolated from adult WT mice using a Langendorff perfusion system as before.⁴⁸ Cardiac FBs at passage 1 were used as homeostatic FBs (cFB) with minimized phenotypic switching. Myofibroblasts (myoFBs) were generated by activating cFBs with TGFβ1 (1 ng/mL, 24 h) and kept in the presence of TGFβ1 to sustain their activated state (as myoFB). Hypoxia was used to simulate the post-MI condition in the infarct tissue. Ischemia was generated by nutrient deprivation plus hypoxia. Adam17 knockdown was achieved by Adam17-siRNA Silencer™ (or Negative Control siRNA) using Lipofectamine RNAiMAX Transfection Reagent. MyoFBs were randomized into 4 groups: (ⅰ) Normoxia (21% O₂)+ saline; (ⅱ) Normoxia + TGFβ1; (ⅲ) Hypoxia (1% O₂)+ saline; (ⅳ) Hypoxia + TGFβ1.

Cardiac FB activation was determined by co-immunostaining with α-smooth muscle actin (αSMA) and vimentin. Cell proliferation was measured using 5-bromo-2′-deoxyuridine (BrdU) incorporation assay (serum-free solution). YAP nucleus translocation was assessed using co-immunofluorescence staining for YAP and DAPI.

2.6. Immunostaining, protein and collagen cross-linking analyses

Formalin-fixed hearts were used for Trichrome staining, and OCT-cryosections for immunofluorescence staining as before.^25,48 Flash-frozen LV tissue (inf, peri, non), or cultured FBs, were processed for protein extraction (CellLytic M buffer), or RNA (Trizol) as before.^42,48 Band intensity in Western blots was quantified (ImageQuant TL) and values were first normalized to the corresponding loading control for each sample, and then normalized to the value of the control samples within each blot (Adam17^f/f sham, or saline) if needed. Hydroxyproline assay was performed to measure the amount of insoluble and soluble collagen content as before.^41,48

2.7. Bulk RNA-sequencing

Total RNA was extracted from primary cardiac fibroblasts (cFBs) or myofibroblasts (myoFBs) with intact or knockdown Adam17, cultured under normoxic or hypoxic conditions, using the RNeasy Plus Universal Mini Kit (QIAGEN, 73404), according to the manufacturer’s instructions. RNA quality assessment, library preparation, and sequencing were performed by the SBME-Seq Core facility (University of British Columbia, BC). All RNA samples had RIN scores between 8.9 and 10 and were subjected to mRNA sequencing with the depth of 20 million paired-end reads per sample. Analyses and figure generation were performed by R (v4.3.0) using standard Bioconductor and CRAN packages. The raw and processed sequencing data have been deposited in the NCBI Gene Expression Omnibus (GEO) under accession number GSE306012.

2.8. Biomechanical measurements

Stiffness of the infarcted myocardium was measured at 1-week post-MI ex vivo using Mach-1 Biomomentum mechanical tester equipped with a single-axis 1.5N load cell.⁵¹ Data were fitted with Hayes model and analyzed using Automated Indentation and Thickness Batch Analysis software, and further mapped using the Analysis-MAP software. Regional stiffness for each heart was normalized to tissue thickness and data are reported as instantaneous modulus at 10% strain of indentation.

2.9. Micro-computed tomography (micro-CT) imaging of coronary vasculature

Freshly excised hearts were perfused with contrast agent Microfil MV-122 Yellow (Flow Tech Incorporation, MA, USA). Images were acquired and reconstructed by MILabs softwares.⁵²

2.10. Co-culture of myofibroblasts and endothelial cells

Primary human coronary endothelial cells (hECs, from ATCC) were seeded on top of myoFBs that were treated with Adam17-siRNA (or control siRNA), in a 50–50 culture medium mixture for the two cell types (DMEM/F12 and VCBM). After 3 days, conditioned media were collected to analyze the level of cleaved VEGFR2 (Human sVEGFR2/KDR DuoSet ELISA kit, Bio-Techne, DY357 & DY008B), cells were harvested for immunoblotting, or fixed for immunofluorescent staining.

2.11. In vivo ADAM17 inhibitor administration, and ADAM17 enzymatic activity

ADAM17-selective inhibitor (ADAM17i, PZ0416 from Sigma), was suspended in 2% Tween-80/0.5% methylcellulose as before.⁴⁷ ADAM17i (100 mg/kg) or placebo (2% Tween-80/0.5% methylcellulose) was administrated to wildtype mice following MI by oral gavage twice daily according to two regimens (Figure 2): (ⅰ) ADAM17i (or placebo) was delivered starting 2 days prior to MI and continued for 4 days after MI induction; (ⅱ) ADAM17i (or placebo) delivery started on day 3 after MI induction and was continued for four days. ADAM17 enzymatic activity was measured in myocardial tissue post-sham or MI, as before.⁴⁷

Figure 2.

ADAM17 loss in homeostatic fibroblasts (FB) increases the rate of post-MI rupture. (A) Schematic presentation of induction of Tcf21-derived ADAM17 knockdown in homeostatic FBs. Mice received Tamoxifen (TMX) at 10 weeks of age (i.p. injection, 5 consecutive days), and MI induction was performed by LAD ligation at 12 weeks of age. (B) Survival rate and (C) cardiac rupture rates post-MI in indicated genotypes (n = 21-23; Kaplan-Meier survival curves, log-rank test). (D) Long axis heart cross-sections in the indicated groups at 3 days and 1 week post-sham/MI. (E) Infarct size in the indicated genotypes at 3-days and at 1-week post-MI. (F) Insoluble, soluble and total collagen content in the left ventricle of mice at day 3 post-sham or post-MI in each genotype (n = 3–5 hearts/group/genotype). (G) Representative immunoblots and averaged protein content for pro-lysyl oxidase, lysyl oxidase (LOX), ADAMTS2, α-smooth muscle actin (αSMA), collagen type I (COL I) and collagen type III (COL III) in the indicated groups at day 3 post-sham or post-MI (n = 4–5/group/genotype). Ad17^f/f = Adam17^f/f, Ad17^FB-KD = Adam17^FB-KD. INF = infarct, PERI = peri-infarct, NON = non-infarct. A.U. = Arbitrary units. One-way ANOVA (E) and two-way ANOVA with Bonferroni post-hoc test (F, H) were performed. Average values represent Mean±SEM. * P < 0.05 vs. corresponding sham group, #P < 0.05 vs. corresponding Adam17^f/f group.

2.12. Statistics

All statistical analyses were performed using IBM SPSS Statistics 21 software or Prism (for log-tank test). Data were tested for normal distribution by the Shapiro-Wilks Normality Test. Comparison between two groups was made using Student’s two-tailed t-test. Comparisons among multiple groups were made using two-way ANOVA followed by Bonferroni or Tukey post-hoc test. Log rank test was performed to compare the Kaplan-Meier survival curves. Each n value indicates the number of mice used (in vivo), or the number of independent experiments performed (in vitro experiments). Average values are presented as Mean±SEM (standard error of the mean). Summarized snRNAseq data show quartile plots and error bars represent the highest and the lowest values (Figure 1C, D and see Supplementary material online, Figure S1C-J) and one-way ANOVA followed by Bonferroni post-hoc tests was performed. Statistical significance was recognized at P < 0.05.

3. Results

3.1. ADAM17 levels are increased in the fibroblast of ischemic myocardium in patients and in mice

To assess if the cell-specific expression of ADAM17 is different in the heart with ischemic cardiomyopathy (ICM) compared with non-failing control (NFC), we analyzed the publicly available single-nuclei RNA sequencing (snRNA-seq) dataset by Kuppe et al.⁴⁰ Uniform Manifold Approximation and Projection (UMAP) plots were generated, and major cardiac cells were clustered and separated into NFC and ICM groups (Figure 1A), and further categorized into ischemic zone (IZ), border zone (BZ), and remote zone (RZ) (see Supplementary material online, Figure S1A). A heatmap of differentially expressed genes (DEGs) across all cardiac cells shows the cell-specific marker genes identified among the DEGs (Figure 1A). Verification Cell clusters were further verified by Feature Plots showing major curated marker genes as reported in the literature (see Supplementary material online, Figure S1A, middle panel).^53,54 The distribution of cardiac cells in NFC, and in IZ, BZ, and RZ of ICM revealed a decrease in cardiomyocyte population in the ischemic and border zones, but an increase in fibroblasts, endothelial cells, and myeloid cells in the ischemic zone comparing to non-failing controls (Figure 1B) and in the replicate samples from each zone (see Supplementary material online, Figure S1A, right panel). Comparing the expression of ADAMs that have been shown to be involved in heart disease showed that Adam17 increased in the total cell population (Figure 1C) and in the fibroblasts (Figure 1D) in the ischemic zone, Adam9 and Adam15 showed minimal changes in all regions (see Supplementary material online, Figure S1C, D, I, J), Adam10 decreased in total cells (see Supplementary material online, Figure S1E) but did not change in fibroblasts (see Supplementary material online, Figure S1F), and Adam12 increased in total cells and in fibroblasts in the ischemic zone (see Supplementary material online, Figure S1G,H).

We further examined the changes in Adam17 expression levels at the protein level using specimens from explanted hearts from patients with ICM and NFC. Western blot analysis for ADAM17 shows a significant increase in the ischemic myocardium (Figure 1E), while co-immunostaining further demonstrates increased ADAM17 levels colocalization with periostin-positive myoFBs in the fibrotic region of the ischemic myocardium (Figure 1F). Consistent with these findings in human specimens, ADAM17 protein levels showed a significantly higher level in the infarct and peri-infarct regions of mouse myocardium after MI as shown by Western blot (Figure 1G), and by co-immunostaining showing strong colocalization with myofibroblasts (periostin-positive staining) (Figure 1H).

3.2. ADAM17 deficiency in homeostatic fibroblasts impairs scar formation following myocardial infarction

ADAM17 knockdown in homeostatic fibroblasts was confirmed at DNA and protein levels, Adam17^FB-KD mice show a significant reduction in ADAM17 levels (46.3%) in the infarct region compared with Adam17^f/f mice (see Supplementary material online, Figure S2A and B). Following MI induction (Figure 2A), male Adam17^FB-KD mice exhibited a significantly compromised post-MI survival (52.0% vs. 80.9% in Adam17^f/f mice; Figure 2B), which we confirmed to be due to an increased rate of LV rupture between days 3 and 7 post-MI (Figure 2C). Heart function assessment by echo-ultrasound showed comparable cardiac dysfunction between the two genotypes at 3 days post-MI (prior to LV rupture), however, the surviving Adam17^FB-KD mice showed decreased infarct size (Figure 2D and E) and decreased LV dilation at 1 week post-MI, but no improvement in functional parameters (Table 1). Mice expressing Tcf21-derived Cre-recombinase with intact Adam17 (Tcf21-Cre^Esr1/Rosa26^tdT, or Tcf21^Cre/+) did not show increased rate of LV rupture (see Supplementary material online, Figure S3A and B), nor any changes in cardiac structure or function improvement (Table 2), compared with their littermate control mice, confirming that the observed increased rate of post-MI rupture in Adam17^FB-KD mice due to Adam17 loss in the homeostatic fibroblasts. Similar to the male mice, female Adam17^FB-KD mice showed an increase in the rate of LV rupture compared with the female Adam17^f/f mice (22.2% vs. 0%). However, female Adam17^FB-KD mice showed comparable functional and structural changes compared with the parallel female Adam17^f/f mice (Table 3).

Table 1.

Echocardiography parameters from MALE Adam17^f/f/R26^tdT (Adam17^f/f) and Adam17^f/f  Tcf21^Cre/+/R26^tdT (Adam17^FB-KD) mice following sham, 3-days and 1-week after myocardial infarction

	Sham		3d post-MI		1 wk post-MI
	Adam17 f/f	Adam17 FB-KD	Adam17 f/f	Adam17 FB-KD	Adam17 f/f	Adam17 FB-KD
Parameter (unit)	n = 12	n = 12	n = 10	n = 9	n = 9	n = 9
HR (bpm)	472±11	444±11	496±7	496±6a	511±20	459±12
EF (%)	67.3±1.2	68.2±1.5	28.2±1.6a	29.2±1.9a	18.6±1.3a,c	23.6±1.3a
FS (%)	37.4±1.4	39.7±0.9	7.3±0.7a	8.6±1.6a	7.0±1.3a	6.7±0.8a
EDV (μL)	56.7±2.2	53.2±2.5	88.1±6.2a	75.6±2.9a	139.3±9.8a,c	108.9±6.8a,c,b
ESV (μL)	21.8±1.6	19.9±1.2	72.0±5.9a	63.4±2.6a	120.6±10.0a,c	93.2± 6.8a,c,b
GLS (%)	−18.6±0.8	−19.6±0.6	−4.1±0.6a	−3.9±0.6a	−2.3± 0.4a	−3.5±0.5a
GCS (%)	−31.4±1.3	−34.5±1.4	−5.4±0.5a	−6.7±0.8a	−1.7± 0.3a	−3.1±0.8a
LA (mm)	1.75±0.04	1.74±0.02	2.75±0.11a	2.65±0.17a	2.83±0.15a	2.90±0.11a
LVAWd (mm)	0.84±0.03	0.81±0.02	0.71±0.07a	0.50±0.02a,b	0.55±0.05a	0.50±0.02a
LVAWs (mm)	1.29±0.04	1.30±0.04	0.81±0.09a	0.56±0.02a,b	0.63±0.06a	0.56±0.03a
LVIDd (mm)	3.77±0.07	3.97±0.10	4.38±0.11a	4.28±0.09a	5.05±0.17a,c	4.5±0.1a,c,b
LVIDs (mm)	2.3±0.1	2.3±0.1	3.6±0.1a	3.5±0.1a	4.8±0.2a,c	4.2±0.1a,c
LVPWd (mm)	0.74±0.02	0.75±0.01	0.60±0.06	0.50±0.02a	0.46±0.06a	0.34±0.02a,c
LVPWs (mm)	1.27±0.03	1.25±0.03	0.71±0.07a	0.65±0.06a	0.54±0.09a	0.40±0.02a,c

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with corresponding sham.

^b P < 0.05 compared with corresponding Adam17  ^f/f.

^c P < 0.05 compared with corresponding 3d-MI group.

Table 2.

Echocardiography parameters from MALE WT/R26^tdT (WT) and Tcf21^Cre/+/R26^tdT (Tcf21^Cre/+) mice following sham, or myocardial infarction at 3-days and 1-week

	Sham		3d post-MI		1 wk post-MI
	WT	Tcf21 Cre/+	WT	Tcf21 Cre/+	WT	Tcf21 Cre/+
Parameter, (unit)	n = 3	n = 6	n = 3	n = 6	n = 3	n = 5
HR (bpm)	424±30	443±21	485±15	513±6	496±49	533±9a
EF (%)	64.4±2.6	67.7±1.1	29.2±3.1a	34.6±2.3a	22.1±2.8a	24.7±2.5a,b
FS (%)	38.5±4.8	35.9±1.9	17.0±2.3a	11.7±2.4a	7.0±1.3a	9.8±1.9a
EDV (μL)	56.4±4.1	55.5±2.1	73.5±3.6	64.4±2.8	101.0±4.5a,b	92.1±6.1a,b
ESV (μL)	19.8±3.8	22.5±1.4	58.9±4.6a	53.6±2.5a	83.2±3.4a,b	76.4±6.6a,b
GLS (%)	−21.4±0.7	−18.9±1.3	−3.1±0.3a	−4.3±0.5a	−2.4±0.7a	−3.5±1.0a
GCS (%)	−31.7±3.4	−30.7±1.8	−6.8±1.8a	−7.5±2.6a	−2.6±1.2a	−3.8±1.0a
LA (mm)	1.55±0.02	1.67±0.03	2.10±0.06a	2.27±0.11a	2.26±0.06a	2.46±0.10a
LVAWd (mm)	0.88±0.04	0.86±0.05	0.67±0.05a	0.61±0.01a	0.55±0.02a	0.53±0.02a
LVAWs (mm)	1.35±0.05	1.32±0.04	0.76±0.05a	0.68±0.01a	0.59±0.06a	0.60±0.02a
LVIDd (mm)	4.02±0.06	3.85±0.09	4.61±0.17	4.20±0.13	5.41±0.07a,b	5.21±0.14a,b
LVIDs (mm)	2.37±0.16	2.27±0.09	4.02±0.20a	3.81±0.22a	5.13±0.06a,b	4.69±0.21a,b
LVPWd (mm)	0.85±0.04	0.77±0.03	0.86±0.07	0.75±0.08	0.53±0.08	0.56±0.05
LVPWs (mm)	1.23±0.03	1.29±0.06	1.15±0.07a	0.90±0.10a	0.66±0.07a	0.79±0.07a

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with corresponding sham.

^b P < 0.05 compared with corresponding 3 days post-MI group.

Table 3.

Echocardiography parameters from FEMALE Adam17^f/f/R26^tdT (Adam17^f/f) and Adam17^f/f  Tcf21^Cre/+/R26^tdT (Adam17^FB-KD) mice following sham, 3-days and 1-week after myocardial infarction (MI)

	Sham		3d post-MI		1 wk post-MI
	Adam17 f/f	Adam17 FB-KD	Adam17 f/f	Adam17 FB-KD	Adam17 f/f	Adam17 FB-KD
Parameter (unit)	n = 8	n = 9	n = 7	n = 9	n = 6	n = 6
HR (bpm)	410±15	427±10	507±6a	468±16	453±20	483±20
EF (%)	66.8±1.6	71.6±1.4	26.8±2.6a	27.2±1.9a	20.9±2.6a	27.3±1.2a
FS (%)	36.0±2.6	40.5±2.2	14.9±2.8a	10.6±2.5a	7.5±2.0a	8.0±3.0a
EDV (μL)	50.0±2.3	46.5±2.2	59.9±2.8a	62.1±4.4a	86.9±11.2a,c	70.3±4.5a
ESV (μL)	21.3±1.5	16.2±1.1	46.9±2.7a	48.9±4.3a	75.4±11.2a,c	55.0± 3.8a
GLS (%)	−17.7±0.6	−20.4±0.7	−3.6±0.7a	−4.4±0.6a	−3.3± 1.2a	−4.4±0.6a
GCS (%)	−31.1±1.0	−34.4±1.4	−3.9±0.7a	−4.4±0.9a	−1.6± 0.5a	−6.1±1.1a
LA (mm)	1.56±0.02	1.65±0.02	2.05±0.10a	2.12±0.09a	2.36±0.13a	2.10±0.12a
LVAWd (mm)	0.86±0.02	0.87±0.02	0.63±0.04a	0.62±0.03a	0.50±0.03a	0.53±0.04a
LVAWs (mm)	1.31±0.02	1.40±0.03	0.72±0.04a	0.71±0.04a	0.55±0.03a	0.57±0.04a
LVIDd (mm)	3.56±0.07	3.55±0.10	3.96±0.12	4.22±0.13a	4.87±0.29a,c	4.54±0.17a
LVIDs (mm)	2.18±0.11	2.09±0.09	3.58±0.20a	3.91±0.17a	4.60±0.33a,c	4.24±0.21a
LVPWd (mm)	0.76±0.04	0.83±0.03	0.70±0.06	0.67±0.06	0.45±0.04a	0.61±0.12
LVPWs (mm)	1.18±0.05	1.28±0.03	0.84±0.08a	0.82±0.07a	0.57±0.05a	0.74±0.16a

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with corresponding sham.

^b P < 0.05 compared with corresponding Adam17  ^f/f. None detected.

^c P < 0.05 compared with corresponding 3d-MI group.

To investigate the underlying mechanism for the increased cardiac rupture rate in male Adam17^FB-KD mice, we examined collagen homeostasis, synthesis and assembly in these mice at 3 days post-MI (the time point prior to the onset of LV rupture). Collagen cross-linking was found to be compromised in the infarct and peri-infarct myocardium in Adam17^FB-KD compared with Adam17^f/f mice as detected by the lower levels of insoluble and total collagen content (Figure 2F). Consistent with the reduced collagen cross-linking, enzymes involved in collagen processing and cross-linking, lysyl oxidase (LOX) and its proenzyme (pro-LOX), and ADAMTS2 (metalloproteinase with thrombospondin motif 2), were significantly decreased in the infarct region, while αSMA (a marker of FB activation), and collagen type-III, but not Collagen type-I levels were decreased in the peri-infarct and infarct myocardium, respectively in Adam17^FB-KD compared with parallel Adam17^f/f mice (Figure 2G). Although collagen type I and type III are often presumed to follow the same expression pattern, it has been reported that upregulation of collagen type III is often an early response to wound healing,⁵⁵ and its suppressed upregulation in Adam17^FB-KD could contribute to the impaired scar formation in these mice. Interestingly, expression of fibronectin, tissue inhibitor of matrix metalloproteinase (TIMP)-1, TIMP3 and Smad2/3 were not altered, while EGFR phosphorylation was reduced at 3-days post-MI (see Supplementary material online, Figure S4A and B). At 1 week post-MI, the surviving Adam17^FB-KD mice showed similarly reduced fibronectin, decreased phosphorylation of Smad2/3 and EGFR phosphorylation (see Supplementary material online, Figure S4C and D), decreased TIMP1 and increased TIMP3 levels (see Supplementary material online, Figure S4C and D) at 1 week post-MI indicating suppressed activation of the pro-fibrotic pathways at 1 week (but not at 3-days) post-MI in these mice. Inflammation is another contributor to scar formation. We assessed the inflammation severity in male and female mice at 1-week post-MI. Adam17^FB-KD hearts showed a significant reduction in macrophage (CD68⁺) and neutrophil (Ly6G⁺) population in the infarct and peri-infarct myocardium compared with Adam17^f.f hearts in male mice (see Supplementary material online, Figure S5). Similarly, macrophage and neutrophil populations were significantly lower in the infarct and peri-infarct regions of female Adam17^FB-KD hearts (see Supplementary material online, Figure S6).

In vitro, exposure of primary adult cardiac fibroblasts (cFBs) to hypoxia (Figure 3), or hypoxia plus nutrient deprivation (see Supplementary material online, Figure S7), as simulations of in vivo ischemic injury, showed that activation of cFBs lacking Adam17 (Ad17^KD) to myofibroblasts was markedly suppressed, as evident by the significantly lower expression of αSMA, compared with Adam17-intact (Ad17^WT) cFBs (Figure 3B and C and S7), and also shown by the colocalization of αSMA with vimentin by surface plot analysis (Figure 3D). There was no difference between the genotypes under normoxic conditions, nor nutrient deprivation alone (Figure 3B–D, S7). Further, proliferation of cFB in hypoxia, assessed by BrdU incorporation, was significantly lower in Ad17^KD cFBs in hypoxic conditions, but not in normoxic conditions, compared with Ad17^WT cFBs (Figure 3E and F). To further explore the transcriptional changes of cFBs with Adam17KD in response to normoxia or hypoxia, we performed bulk RNA sequencing (see Supplementary material online, Figure S8). Principal-component (PC) analysis showed that Ad17^KD cFBs clustered in close proximity to Ad17^WT cFBs under normoxia, whereas under hypoxia, clusters of Ad17^KD and Ad17^WT cFBs were far apart (see Supplementary material online, Figure S8A). Differential expression (DE) analysis to identify genes with significant differential expressions (P value < 0.05 and log2^{fold change} [log2^FC], log2^FC >0.26, or log2^FC < −0.26) showed only 56 downregulated and 175 upregulated genes in Ad17^KD cFBs compared with Ad17^WT cFBs in normoxic condition, but 1593 downregulated and 1889 upregulated genes under hypoxia (see Supplementary material online, Figure S8B and C). DEGs were further clustered into main features and functions of FBs in response to ischemic injury, including collagen fibril organization, extracellular matrix (ECM) production and turnover, fibroblast activation, cell proliferation and mechanotransduction (see Supplementary material online, Figure S8D and E). Consistent with the phenotypic data, genes associated with collagen fibril organization (Loxl1, 2, 3, 4), ECM production (Ccn2, Col1α1, Col1α2, Tgfb1), FB activation (Cthrc1, Ddah1, Postn, Tnc) and integrin-mediated mechanotransduction (Itga1, Itga3, Itga5, Itgb5) were significantly downregulated in Ad17^KD cFBs compared with Ad17^WT cFBs under hypoxia (see Supplementary material online, Figure S8D and E). These data provide evidence that loss of ADAM17 in homeostatic FBs alters their gene expression profile towards a decrease in activation and proliferation, that could lead to impaired scar formation and, subsequently, cardiac rupture.

Figure 3.

Loss of ADAM17 in primary adult cardiac fibroblasts (cFBs) suppresses their activation to myofibroblasts (myoFBs), and their proliferation in hypoxic conditions in vitro. (A) Confirmation of ADAM17 knockdown in cFBs with Adam17siRNA (Adam17^KD) compared with controls treated with scrambled siRNA (Adam17^WT) in normoxic or hypoxic condition. (B&C) Representative images of co-immunofluorescence staining (and quantification) for α-smooth muscle actin (αSMA, marker of FB activation, green) and vimentin (general FB marker, red) in normoxia (21% O₂) and hypoxia (1% O₂) conditions; and the corresponding 3D surface plots to illustrate colocalization of αSMA and vimentin (yellow) in each group Scale bar = 100μm. (D) Representative immunofluorescence staining for BrdU (Bromodeoxyuridine) and averaged quantification of percent BrdU⁺ cFBs in indicated groups, in normoxic and hypoxic conditions in vitro. Magnified images are shown on the bottom row (n = 3 independent cFB isolation and culture, 2 wells/group/experiment). Scale bar = 100μm. Two-way ANOVA with Bonferroni post-hoc test was performed. Average values represent Mean±SEM. * P < 0.05 vs. corresponding normoxia group, # P < 0.05 vs. corresponding Adam17^WT group.

Figure 4.

Loss of ADAM17 in myofibroblasts (myoFB) limits infarct expansion, preserves cardiac function up to 4 weeks after myocardial infarction (MI). (A) Schematic resentation of induction of Postn-derived Adam17 loss by injection of tamoxifen (TMX) 5 days after MI (i.p injection, 5 consecutive days). (B) Survival rate post-MI (Kaplan-Meier survival curves followed by log-rank test). (C) Infarct size at 1 week and 4 weeks post-MI in indicated genotypes. (D) Trichrome—stained heart cross-section at 1 week and 4 weeks post-sham/MI. (E) Echocardiography parameters (EF: ejection fraction; EDV: End-diastolic volume; ESV: End-systolic volume; FS: fractional shortening; LVIDd: left ventricle internal diameter at diastole; LVIDs: left ventricle internal diameter at systole) following after 1 wk or 4 wks of sham/MI surgery (n = 7-9 mice/group/genotype). (F) Representative immunoblots and averaged protein quantification for collagen type I (COL I), collagen type III (COL III), Fibronectin (FBN1), and α-smooth muscle actin (αSMA) at 1 week post-sham/MI (n = 4–5/group/genotype). Ad17^f/f = Adam17^f/f, Ad17^myoFB-KD = Adam17^myoFB-KD. A.U. = Arbitrary units. Student’s t-test (C) and two-way ANOVA with Bonferroni post-hoc test (E, G) were performed. Average values represent Mean±SEM. * P < 0.05 vs. corresponding sham, # P < 0.05 vs. corresponding Adam17^f/f, § P < 0.05 vs. corresponding 1 wk-MI.

3.3. Loss of ADAM17 in myofibroblasts improves structural and functional remodeling following myocardial infarction

Myofibroblasts (myoFBs) are responsible for scar formation, but if left unchecked, their highly synthetic ability can promote excess fibrotic deposition and infarct expansion post-MI.^5,6,56 Given the required balance in the function of myoFBs in post-MI remodeling, ADAM17 knockdown was induced in myoFBs after MI (Figure 4A). ADAM17 knockdown in myofibroblasts was confirmed at DNA and protein levels, Adam17^myoFB-KD mice show a significant reduction in ADAM17 (35.3%) in the infarct region compared with Adam17^f/f mice (see Supplementary material online, Figure S2C and D). Post-MI survival was not impacted by Adam17 knockdown in myoFBs (74.2% LV rupture in Adam17^myoFB-KD vs. 80.0% in Adam17^f/f mice; Figure 4B); however, the infarct size was significantly reduced in male Adam17^myoFB-KD mice at 1 week and 4 weeks after MI compared with the parallel Adam17^f/f mice (Figure 4C and D). Consistent with the improved structural remodeling, echocardiography and strain analyses showed alleviated systolic dysfunction and LV dilation in male Adam17^myoFB-KD mice as detected by greater ejection fraction, reduced end-diastolic and systolic volumes at 1-week post-MI, while these beneficial outcomes persisted until 4 weeks post-MI (Figure 4E; Table 4). Male mice expressing only Periostin-derived Cre-recombinase with intact Adam17 expression (Postn-Cre^Esr1, or Postn^Cre/+) showed comparable LV rupture rate (see Supplementary material online, Figure S9A), and did not exhibit improvements in post-MI infract size nor function (see Supplementary material online, Figure S9B–D, Table 5). Assessing myocardial inflammation showed a significant reduction in macrophage infiltration, but not neutrophil infiltration, in male Adam17^myoFB-KD compared with male Adam17^f/f mice post-MI (see Supplementary material online, Figure S10). Similar experiments in female mice show that Adam17 knockdown in myoFBs did not alter the post-MI cardiac dysfunction and structural remodeling (Table 6). Unlike the males, female Adam17^myoFB-KD mice did not show reduced infarct expansion or LV dilation (Table 6), and post-MI inflammation (CD68⁺ and Ly6G⁺ cells) was comparable between genotypes (see Supplementary material online, Figure S11). This could perhaps be due to the generally smaller infarct size and less severe post-MI remodeling in female mice such that the beneficial impact of Adam17 knockdown in myoFBs did not have a drastic impact on the females. In addition, no LV rupture was detected in female Adam17^myoFB-KD or the parallel female Adam17^f/f mice post-MI which is consistent with previous report on lack of post-MI LV rupture in female mice.⁴⁸

Table 4.

Echocardiography parameters from MALE Adam17^f/f and Adam17^f/f/Postn^Cre/+ (Adam17^myoFB-KD) mice following sham, or myocardial infarction at 1-week and 4-weeks

	Sham		1 wk-MI		4 wk-MI
	Adam17 f/f	Adam17 myoFB-KD	Adam17 f/f	Adam17 myoFB-KD	Adam17 f/f	Adam17 myoFB-KD
Parameter (unit)	n = 9	n = 9	n = 8	n = 8	n = 8	n = 8
HR (bpm)	502±11	478±13	459±10	456±10	504±10	506±13
EF (%)	69.1±1.3	69.2±1.9	21.0±1.9a	32.2±3.1a,b	16.8±0.8a	28.6±1.7a,b
FS (%)	36.6±1.9	35.6±3.0	6.3±0.7a	11.5±3.0a	4.0±0.3a	6.5±0.8a
EDV (μL)	49.5±2.3	51.6±4.1	124.9±8.4a	81.6±8.1a,b	138.4±9.1a	103.5±4.1a,b
ESV (μL)	18.4±1.9	22.7±3.3	104.6±7.6a	62.4±8.0a,b	119.8±8.9a	81.5±5.4a,b
GLS (%)	−20.8±1.1	−17.9±1.2	−2.7±0.3a	−4.7±0.7a	−1.8±0.4a	−4.1±0.6a
GCS (%)	−30.0±1.3	−31.1±0.7	−4.7±0.6a	−9.7±1.4a,b	−2.7±0.6a	−6.7±1.4a
LA (mm)	1.52±0.02	1.59±0.02	2.42±0.16a	1.92±0.06b	2.69±0.07a	2.39±0.08a,c
LVAWd (mm)	1.00±0.03	0.92±0.04	0.57±0.02a	0.60±0.04a	0.36±0.02a	0.53±0.05a
LVAWs (mm)	1.53±0.04	1.45±0.06	0.63±0.03a	0.66±0.04a	0.43±0.03a	0.66±0.08a,b
LVIDd (mm)	3.58±0.07	3.56±0.09	4.62±0.17a	4.18±0.18a	5.90±0.13a,c	5.01±0.11a,c,b
LVIDs (mm)	2.28±0.08	2.20±0.12	4.23±0.18a	3.38±0.21a,b	5.64±0.13a,c	4.54±0.11a,c,b
LVPWd (mm)	0.76±0.03	0.75±0.02	0.48±0.04a	0.52±0.05a	0.32±0.04a	0.61±0.08b
LVPWs (mm)	1.22±0.05	1.19±0.04	0.53±0.04a	0.54±0.06a	0.37±0.05a	0.66±0.08a,b

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with the corresponding sham.

^b P < 0.05 compared with corresponding Adam17  ^f/f.

^c P < 0.05 compared with corresponding 1 week-MI group

Table 5.

Echocardiography parameters from MALE WT and Postn^Cre/+ mice following sham, or myocardial infarction at 1-week

	Sham		1 wk-MI
	WT	Postn Cre/+	WT	Postn Cre/+
Parameter, unit	n = 7	n = 8	n = 5	n = 6
HR (bpm)	496±14	473±3	450±6	455±14
EF, %	61.5±2.3	65.5±2.1	25.9±2.5a	24.3±2.2a
FS,%	34.2±2.5	35.7±2.2	7.2±0.9a	5.8±0.7a
EDV, μL	53.6±3.7	53.9±2.7	125.9±10.7a	140.3±9.2a
ESV, μL	23.3±1.9	21.3±1.6	104.6±10.6a	120.1±8.1a
GLS, %	−16.1±1.8	−19.4±1.3	−1.9±0.2a	−2.0±0.3a
GCS, %	−27.7±1.6	−32.5±1.3	−2.6±0.4a	−3.8±1.9a
LA, mm	1.54±0.01	1.53±0.02	2.42±0.16a	2.82±0.15b
LVAWd, mm	0.83±0.05	0.79±0.02	0.58±0.05a	0.56±0.04a
LVAWs, mm	1.30±0.04	1.35±0.05	0.70±0.05a	0.67±0.03a
LVIDd, mm	3.77±0.18	3.57±0.13	5.86±0.24a	6.09±0.36a
LVIDs, mm	2.36±0.19	2.07±0.10	5.51±0.27a	5.70±0.42a
LVPWd, mm	0.80±0.03	0.76±0.03	0.42±0.03a	0.56±0.04a
LVPWs, mm	1.20±0.04	1.23±0.05	0.52±0.04a	0.68±0.06a

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with the corresponding sham.

^b P < 0.05 compared with corresponding WT.

Table 6.

Echocardiography parameters from FEMALE Adam17^f/f and Adam17^f/f/Postn^Cre/+ (Adam17^myoFB-KD) mice following sham, 1-week and 4-weeks of myocardial infarction

	Sham		1 wk post-MI		4 wks post-MI
	Adam17 f/f	Adam17myoFB-KD	Adam17 f/f	Adam17myoFB-KD	­Adam17f/f	Adam17myoFB-KD
Parameter (Unit)	n = 7	n = 9	n = 5	n = 7	n = 4	n = 5
HR (bpm)	483±14	483±28	422±12	427±12	484±25	481±13
EF (%)	64.7±1.1	62.4±1.9	26.1±1.4a	30.1±2.9a	20.1±1.6a	26.4±2.2a
FS (%)	32.5±1.9	30.2±1.5	8.3±1.3a	8.4±0.8a	5.4±1.2a	5.6±0.5a
EDV (μL)	41.0±2.4	45.8±1.6	89.4±6.6a	85.5±5.7a	107.2±8.6a	103.9±4.5a
ESV (μL)	18.8±1.8	22.1±1.6	70.7±6.0a	64.2±4.4a	92.3±6.4a,c	84.5±4.3a,c
GLS (%)	−17.0±1.5	−17.1±1.4	−4.2±0.7a	−6.1±0.8a	−2.5±0.3a	−4.9±0.9a
GCS (%)	−29.4±2.2	−29.5±1.7	−4.0±0.8a	−5.9±1.7a	−1.3±0.7a	−3.8±1.0a
LA (mm)	1.57±0.05	1.60±0.06	2.43±0.10a	2.17±0.07a	2.71±0.14a	2.31±0.06a,b
LVAWd (mm)	0.82±0.03	0.81±0.04	0.45±0.05a	0.43±0.02a	0.44±0.05a	0.50±0.06a
LVAWs (mm)	1.26±0.03	1.22±0.03	0.55±0.07a	0.54±0.03a	0.54±0.07a	0.59±0.07a
LVIDd (mm)	3.46±0.09	3.51±0.05	4.96±0.20a	5.04±0.12a	5.61±0.17a,c	5.34±0.27a
LVIDs (mm)	2.15±0.12	2.28±0.09	4.58±0.21a	4.64±0.17a	5.30±0.17a	5.02±0.33a
LVPWd (mm)	0.73±0.03	0.69±0.04	0.42±0.20a	0.32±0.05a	0.22±0.02a	0.32±0.06a
LVPWs (mm)	1.12±0.05	1.05±0.06	0.50±0.08a	0.39±0.06a	0.27±0.03a	0.37±0.08a

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with the corresponding Sham group.

^b P < 0.05 compared with corresponding Adam17  ^f/f.

^c P < 0.05 compared with the corresponding 1wk-MI group.

Examination of the myocardial tissue for scar formation showed reduced protein levels of key ECM proteins, collagen type I, collagen type III, Fibronectin, and αSMA in the infarct and peri-infarct myocardium of Adam17^myoFB-KD compared with Adam17^f/f hearts at 1week after MI (Figure 4F). Adam17^myoFB-KD hearts showed decreased TIMP1 (a marker for fibrosis) but not TIMP3 protein level in infarct and peri-infarct myocardium (see Supplementary material online, Figure S12). In vitro, loss of ADAM17 in myoFBs significantly suppressed cell proliferation in normoxic and hypoxic conditions (see Supplementary material online, Figure S13A and B). Bulk-RNA sequencing on myofibroblasts with intact Adam17 or Adam17 knockdown under normoxia or hypoxia was performed, and principal component analysis revealed that clusters of these 4 myoFB groups were far apart from each other (see Supplementary material online, Figure S14A). Compared with Ad17^WT myoFBs, myoFBs with Adam17KD showed 100 downregulated and 145 upregulated genes in normoxia (see Supplementary material online, Figure S14B), 480 downregulated and 490 upregulated genes under hypoxia (see Supplementary material online, Figure S14C). In contrast to the differential expressions observed in cFBs which were primarily related to ECM production and organization, Adam17KD in myoFBs additionally caused a significant shift in expression profiles towards ECM mineralization and sprouting angiogenesis (see Supplementary material online, Figure S14D). Specific genes that were upregulated within these pathways included genes associated with ECM turnover (Timp2, Timp3, Mmp13, Mmp16), and sprouting angiogenesis (Bmp6, Ccl2, Vegfd) (see Supplementary material online, Figure S14E). Conversely, genes associated with ECM mineralization (Cilp, Ptn, Enpp1) were significantly downregulated in Ad17^KD compared with Ad17^WT myoFBs (see Supplementary material online, Figure S14E). These results indicate that ADAM17 deficiency in myoFB reduces their proliferation, and modifies the expression of ECM-regulatory and pro-angiogenic molecules toward reduced ECM accumulation and potentially increased vascularization.

3.4. Adam17 knockdown in myofibroblasts increased vascularization in the infarct tissue in vivo, and endothelial tubular structure in vitro

We investigated the mechanism underlying the limited infarct expansion and attenuated LV dilation and dysfunction post-MI in Adam17^myoFB-KD mice, and we found greater coronary density in the infarct and peri-infarct regions in Adam17^myoFB-KD compared with Adam17^f/f mice at 1 week post-MI as detected by immunofluorescence staining for CD31 (endothelial protein) (Figure 5A), and the 3-dimentional visualization of functional coronary arteries by micro-CT scan in each genotype post-sham or MI (Figure 5B). This revealed that while coronary density was reduced in both genotypes post-MI, significantly greater arterial volume was present in the LV wall in Adam17^myoFB-KD compared with Adam17^f/f mice at 1 week post-MI (Figure 5A and B).

Figure 5.

Greater coronary density in Adam17^myoFB-KD hearts after myocardial infarction (MI), and more tubular structures formed by endothelial cells (ECs) co-cultured with Adam17-deficient myoFBs. (A) Co-immunostaining for CD31 (ECs) and DAPI (nuclear) in infarct (INF), peri-infarct (PERI), and non-infarct (NON) regions in each genotype at 1wk post-sham/MI (n = 2–4 hearts/group/genotype; each data point shows averaged values from 6 to 8 sections/heart). (B) Representative micro-CT images visualizing the coronaries at 1 wk post-sham/MI (red = coronaries in anterior LV wall, blue = other coronaries), and averaged quantification per heart. (C) Co-immunostaining for vascular endothelial (VE)-cadherin and vimentin (myoFB) on co-culture of human coronary ECs with myoFB, with intact Adam17 or Adam17-deficiency in myoFBs (TGFβ1 simulates the pro-fibrotic infarct environment). (D) Levels of soluble VEGFR2 (vascular endothelial growth factor receptor-2, an ADAM17 substrate), in the conditioned media from co-cultures. (E) Immunofluorescent staining and quantification of phospho-VEGFR2 (pVEGFR2) in co-culture of ECs with myoFB (Ad17-intact or -deficient myoFBs). Two-way ANOVA with Bonferroni post-hoc test (A, C, D), and Student’s t test (B) were performed. Average values represent Mean ± SEM. *P < 0.05 vs. corresponding sham or saline group, #P < 0.05 vs. corresponding Ad17^f/f or myoFB^Ad17WT/EC group.

To further determine how ADAM17 knockdown in myoFBs impacts the function of endothelial cells (EC), we utilized a co-culture of primary endothelial cells (hECs) and myoFBs to investigate the EC-FB interactions. ECs with intact ADAM17 expression were added to myoFBs with intact or knocked down Adam17 expression. The siRNA-mediated reduction of ADAM17 protein level persisted for at least 6 days (144 h) (see Supplementary material online, Figure S15A). Co-immunofluorescent staining for ECs (VE-cadherin) and FBs (vimentin) showed that ECs form tubular structures when cultured on myoFBs, and that compared with the co-cultures with intact ADAM17 expression, Adam17 knockdown in myoFBs promotes a significant increase in the length and the number of branches of EC-formed tubular structures in pro-fibrotic conditions (+TGFβ1) in normoxia (Figure 5C; Supplementary material online, Figure S15C) and in hypoxia (see Supplementary material online, Figure S15B and C). When cultured in the absence of myoFBs, ECs formed a monolayer (and not tubular structures), and TGFß1 did not alter their morphology or survival, in normoxic or hypoxic conditions (see Supplementary material online, Figure S15D).

Next, we investigated the mechanism responsible for the greater ECs proliferation and growth in the presence of Ad17^KD myoFBs, and found that the reduced ADAM17 levels (and its sheddase activity) in the co-culture of myoFB^Ad17KD and EC resulted in decreased shedding of vascular endothelial growth factor receptor (VEGFR)-2, a well-known substrate for ADAM17.^57,58 The levels of soluble VEGFR2 in the conditioned media were significantly lower in the co-cultures with myoFB^Ad17KD (Figure 5D), suggesting less shedding of this protein in the absence of ADAM17 (and a greater presence of this receptor in the cell membrane). Consistently, higher levels of phosphorylated VEGFR2 were detected in the co-cultures of Ad17^KD myoFB and ECs (Figure 5E). These results indicate that reduced ADAM17 sheddase activity in myoFB stabilizes VEGFR2 on the endothelial cell (EC) membrane promoting its activation, which led to increased EC proliferation and vascular formation in Adam17^myoFB-KD myocardium.

3.5. Myofibroblast ADAM17 regulates infarct tissue compliance and vascularization via the EGFR-YAP axis

Excess collagen deposition contributes to a more dense ECM and greater tissue stiffness.⁵ Increased tissue stiffness transmits mechanical stimuli to the residing endothelial cells impacting vascular morphogenesis, EC differentiation, and cell dysfunction.^59,60 To investigate if the deficiency of ADAM17 in myoFBs alters the biomechanical properties of the infarcted myocardium, regional myocardial stiffness was measured using the indentation method (Figure 6A). The infarct tissue in Adam17^myoFB-KD mice exhibited a significantly lower instantaneous modulus value, a measure of tissue stiffness (23.68±0.87 kPa vs. 33.7±0.89 kPa in Adam17^f/f; Figure 6B). The YAP-mediated pathway is a key driver of mechano-transduction signaling,⁶¹ with a reciprocal function in regulating fibrosis development upon organ injury.⁶² Activation of the EGFR pathways, or increased mechanical stress secondary to increased tissue stiffness dephosphorylate YAP allowing its nuclear translocation and transcriptional induction of pro-fibrotic genes that can further elevate tissue stiffness.^62–65 Consistent with attenuated myocardial stiffness in Adam17^myoFB-KD hearts, reduced activation of EGFR (due to reduced ADAM17-mediated release of its ligands), and higher phosphorylated YAP (unable to translocate to the nucleus) were detected in the infarct and/or peri-infarct regions of Adam17^myoFB-KD compared with the parallel Adam17^f/f mice (Figure 6C). The direct impact of ADAM17 loss in myoFB on YAP nuclear translocation was confirmed in vitro as immunostaining for YAP showed reduced nuclear localization in Ad17^KD myoFBs compared with control myoFBs (Figure 6D). In addition, stiffness of ECM has been linked to expression of proteins such as integrin β1, a mechanosensory and a cell membrane receptor for ECM proteins, collagen Iα1, and smooth muscle (SM)-22α, a mesenchymal marker upregulated in ECs in response to stiff matrix,⁶⁶ all of which were significantly decreased in Adam17^KD myoFB-EC co-cultures when stimulated with pro-fibrotic conditions (see Supplementary material online, Figure S16A and B). Consistent with our in vivo data (Figure 6C), phosphorylation of EGFR was reduced in myoFB^Ad17KD/EC^WT co-cultures compared with myoFB^Ad17WT/EC^WT cocultures (see Supplementary material online, Figure S16A and B) due to decreased bioavailability of EGFR agonists (substrates for ADAM17 sheddase activity). Taken together, downregulation of ADAM17 in myoFBs attenuates ECM stiffness, which promotes EC proliferation through increased activation of VEGFR2 (in ECs) and suppresses activation of EGFR and proliferation of myoFBs.

Figure 6.

ADAM17 knockdown in myofibroblasts (myoFB) mitigates post-myocardial infarction (MI) tissue stiffness. (A) Representative mapping images of instantaneous modulus on the LV tissue at 1-week post-MI, and (B) Violin plot of instantaneous modulus data points for each group. n = 35–45 points/heart; n = 3–4 hearts/genotype. (C) Representative immunoblots and averaged protein quantification of phosphorylated and total EGFR (epidermal growth factor receptor) and YAP (yes-associated protein) in indicated groups at 1 week post-sham/MI (n = 4–5/group/genotype). A.U. = Arbitrary units. *P < 0.05 vs. corresponding sham, # P < 0.05 vs. corresponding Adam17^f/f. (D) Representative immunostaining and quantification of YAP nuclear localization in myoFBs (activated by TGFβ1) in normoxia (21% O₂) or hypoxia (1% O2) in vitro. 3D surface plots show the overlay of YAP and DAPI (i.e. nuclear translocation of YAP). n = 3 independent FB isolations and cultures, 2–3 wells/group/experiment. * P < 0.05 vs. corresponding normoxia, # P < 0.05 vs. corresponding Adam17^WT. Student’s t test (B) and two-way ANOVA with Bonferroni post-hoc test (C, D) were performed. Average values represent Mean±SEM.

3.6. Identifying the therapeutic window for pharmacological inhibition of ADAM17 following MI

Our study revealed opposing effects of targeting ADAM17 in homeostatic fibroblasts vs. myoFBs in post-MI recovery. Hence, we explored the potential therapeutic window for ADAM17 inhibition as a treatment for optimal outcomes post-MI recovery. A selective pharmacological ADAM17 inhibitor (PZ0416) with minimal impact on other ADAMs or other metalloproteinases⁴⁷ was administered as before (twice daily by gavage) to obtain a plasma concentration of 4 μg/mL that we previously reported to be sufficient to reduce tissue ADAM17 activity.²² We employed two therapeutic approaches using wildtype mice: (ⅰ) ADAM17 inhibition started 2 days before MI induction and continued for 4 days after MI (Figure 7A–D); (ⅱ) ADAM17 inhibition started 3 days after MI induction for 4 days (days 3–7 post-MI) (Figure 7E–H). We confirmed that the PZ0416 significantly decreased the rise in ADAM17 activity following MI (see Supplementary material online, Figure S17A). In the first approach, inhibition of ADAM17 before and shortly after MI did not alter post-MI survival, infarct size, nor LV dilation and dysfunction (Figure 7B–D, Table 7). In contrast, short-term inhibition of ADAM17 after the initial phase of post-MI scar formation was completed (3d post-MI), significantly reduced infarct expansion and alleviated LV dilation and dysfunction up to 4 weeks post-MI compared with placebo-treated mice (Figure 7H, Table 8). Therefore, the initial upregulation of ADAM17 post-MI is essential for optimal scar formation, but its inhibition following this initial phase effectively mitigates the adverse post-MI remodeling and dysfunction, and exerts long-term protective effects.

Figure 7.

Identifying a therapeutic window for pharmacological inhibition of ADAM17 (by PZ0416, ADAM17i) following myocardial infarction (MI) with long-term beneficial effect. (A) The timeline for administration of placebo (vehicle) or ADAM17i in wildtype (WT) mice starting 2 days before, and continued until 4 days after MI induction. (B) Representative trichrome-stained sections of hearts at 1 week post-MI. Scale bar = 1 mm. (C) Post-MI survival rate in the indicated groups. (D) Parameters of left ventricle systolic function and dilation: Ejection fraction (EF), infarct size, end-diastolic/systolic volume (EDV, ESV) before (pre-MI) and 1 week after MI in each group. (E) The timeline for short-term administration of placebo or ADAM17i in WT mice, from 3 days post-MI to 1 week post-MI. (F) Representative trichrome-stained heart sections after 1 week or 4 weeks of MI, in placebo and ADAM17i groups. Scale bar = 1mm. (G) Survival rate post-MI in indicated groups. (H) EF, infarct size, EDV, ESV before MI (preMI), and 1-week, 2-weeks, and 4-weeks post-MI in each group. The black line for each data set shows Mean±SEM. Log-rank test (C, G), and Two-way ANOVA with Bonferroni post-hoc test (D, H) were performed. * P < 0.05 vs. corresponding preMI group, # P < 0.05 vs. corresponding placebo group.

Table 7.

Echocardiography parameters from MALE wildtype (WT) mice with placebo or ADAM17 inhibitor (PZ0416) administered from 2 days before until 4 days after sham or MI

	Sham		1 wk-MI
	Placebo	ADAM17 inhibitor	Placebo	ADAM17 inhibitor
Parameter (unit)	n = 6	n = 7	n = 6	n = 7
BP (bpm)	463±24	476±15	507±14	509±12
EF (%)	66.3±1.1	64.2±2.6	25.3±2.1a	24.4±3.4a
FS (%)	30.1±1.7	36.7±2.6	9.0±2.2a	7.9±2.6a
EDV μL	60.2±2.8	61.7±3.7	102.3±7.8a	112.1±10.9a
ESV, μL	27.6±1.1	26.1±2.3	84.5±9.0a	97.0±12.9a
GLS, %	−19.1±1.1	−17.5±1.0	−2.4±0.4a	−3.3±1.2a
GCS, %	−29.4±1.3	−28.5±1.0	−4.1±1.2a	−4.3±1.7a
LA, mm	1.59±0.03	1.67±0.03	2.36±0.12a	2.62±0.23a
LVAWd, mm	0.78±0.02	0.77±0.02	0.47±0.05a	0.48±0.04a
LVAWs, mm	1.15±0.04	1.18±0.03	0.52±0.04a	0.52±0.04a
LVIDd, mm	3.84±0.10	4.00±0.11	5.22±0.22a	5.57±0.29a
LVIDs, mm	2.55±0.13	2.65±0.13	4.88±0.27a	5.24±0.35a
LVPWd, mm	0.76±0.02	0.74±0.01	0.53±0.11	0.49±0.08
LVPWs, mm	1.13±0.03	1.12±0.03	0.65±0.14a	0.61±0.12a

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with corresponding sham.

Table 8.

Echocardiography parameters from MALE wildtype (WT) mice following short-term treatment with ADAM17 inhibitor (PZ0416), or placebo, administered on days 3–7 after MI

	Sham		1 wk post-MI		4 wks post-MI
	Placebo	ADAM17 inhibitor	Placebo	ADAM17 inhibitor	Placebo	ADAM17 inhibitor
Parameter (unit)	n = 7	n = 7	n = 6	n = 6	n = 5	n = 5
HR (bpm)	468±17	469±19	499±15	499±19	535±4	515±15
EF (%)	63.5±1.8	61.6±0.7	20.3±1.4a	34.7±3.4a,b	15.5±1.4a	31.0±1.4a,b
FS (%)	35.6±1.9	33.5±1.2	5.8±0.8a	13.1±1.1a,b	5.0±0.9a	12.1±1.7a,b
EDV (μL)	51.5±3.0	57.5±3.4	128.0±4.4a	87.0±6.8a,b	153.3±7.4a,c	101.4±6.0a,b,c
ESV (μL)	23.1±2.5	25.3±2.1	113.0±5.82a	66.4±6.1a,b	138.0±8.5a	81.4±6.5a,b
GLS (%)	−17.7±1.2	−18.7±1.2	−2.1±0.2a	−7.3±1.0a,b	−1.7±0.2a	−7.4±1.2a,b
GCS (%)	−29.2±1.4	−28.4±1.0	−2.5±0.3a	−8.4±1.5a,b	−1.5±0.2a	−8.2±1.3a,b
LA (mm)	1.62±0.02	1.65±0.05	2.68±0.11a	2.02±0.09a,b	2.84±0.09a	2.04±0.10a,b
LVAWd (mm)	0.91±0.04	0.84±0.04	0.45±0.02a	0.56±0.04a	0.41±0.03a	0.43±0.03a
LVAWs (mm)	1.36±0.05	1.26±0.06	0.50±0.03a	0.63±0.03a	0.45±0.03a	0.51±0.03a
LVIDd (mm)	3.80±0.17	3.94±0.08	5.90±0.14a	4.60±0.26a,b	6.39±0.13a	4.94±0.21a,b
LVIDs (mm)	2.49±0.16	2.62±0.09	5.68±0.12a	4.19±0.32a,b	6.19±0.13a	4.42±0.25a,b
LVPWd (mm)	0.76±0.03	0.74±0.04	0.48±0.04a	0.60±0.04a,b	0.32±0.04a	0.77±0.06a,b
LVPWs (mm)	1.15±0.04	1.18±0.05	0.38±0.03a	0.79±0.09a,b	0.38±0.06a	1.03±0.09a,b

EDV/ESV, end diastolic/systolic volume; EF, left ventricular ejection fraction; FS, fractional shortening; GCS, global circumferential strain; GLS, global longitudinal strain; HR, heart rate; LA, left atrium; LVAWd and LVAWs, left ventricular anterior wall thickness during diastole or systole; LVIDd and LVIDs, left ventricular internal diameter (diastole and systole); LVPWd/LVPWs, left ventricle posterior wall thickness (during diastole or systole).

^a P < 0.05 compared with corresponding sham.

^b P < 0.05 compared with corresponding placebo.

^c P < 0.05 compared with corresponding 1 wk-MI group.

4. Discussion

Disintegrin and metalloproteinases are a family of transmembrane proteinases with diverse functions. ADAMs 9, 10, 12, 15 and 17 are reported to be involved in cardiac physiology and pathology to varying degrees.^20,67–69 Our screening of Adam expression in human myocardium showed a marked increase in Adam17 RNA expression in the infarct zone in patients with ICM. Reports from other labs have also suggested involvement of ADAM17 in heart disease in humans^34–36 and animal models.^25,38,70 However, given the critical role of ADAMs in normal physiology, a long-term or general inhibition of ADAM17 would not be a feasible therapy. Hence, in this study, we investigated a targeted approach to identify the therapeutic potential of ADAM17 inhibition in limiting the adverse myocardial remodeling following myocardial infarction. Infarct formation is the first step following MI to replace the lost cardiomyocytes, and it is critical for recovery and survival,^71,72 but if uncontrolled, the infarct scar expands, replaces the viable cardiomyocytes, leading to ventricular dilation, impaired systolic function and eventually heart failure.⁷³ In this study, we report that limiting ADAM17 activity in homeostatic FBs interrupted the initial ECM deposition that is critical for reparative infarct formation and led to LV rupture. However, ADAM17 knockdown in myoFBs, after the initial infarct scar tissue is formed, reduced the stiffness of the infarct tissue, promoted endothelial proliferation and vascularization in the ischemic myocardium, thereby limiting infarct expansion, LV dilation and dysfunction. Female mice generally develop a smaller infarct, less LV dilation and dysfunction and minimal-to-no LV rupture post-MI,⁷⁴ however, we detected 22% LV rupture incidence in female Adam17^FB-KD mice compared with 0.0% in females of all other genotypes, which demonstrates the important role of ADAM17 in cFBs for the initial infarct tissue formation in both sexes. Hence, interventions that impact these aspects of post-MI remodeling are usually less profound in female mice. Nonetheless, female Adam17^myoFB-KO mice exhibited improved overall LV remodeling and function post-MI.

We found that the increase in ADAM17 expression in the ischemic myocardium is localized to periostin-positive activated FBs (myoFBs), suggesting a cell-specific function of ADAM17 in ischemic cardiomyopathy. FBs and myoFBs are key cell types in formation of the infarct tissue and a delicate balance in their function is required to ensure optimal remodeling following injury. Insufficient activation of homeostatic FB (to myoFB) will result in inadequate ECM synthesis, poor infarct formation and ventricular rupture. On the other hand, prolonged and uncontrolled proliferation and activation of myoFBs lead to excess ECM accumulation, infarct expansion and greater stiffness, adverse myocardial remodeling and cardiac dysfunction.^9,75 By targeting the homeostatic FB vs. myoFB, the timing for ADAM17 inhibition was also controlled, demonstrating that ADAM17 is required for the initial reparative scar formation, but afterwards, its inhibition limits myoFB excess activity resulting in scar tissue with reduced stiffness and optimal for EC growth and vascularization. ADAM17 is an ectodomain sheddase of several membrane-bound proteins whereby it can regulate cell functions as well as intercellular communications.^30–33 Although ADAM10 and ADAM17 can have overlapping sheddase substrates, Adam17 loss in myoFB did not result in a compensatory upregulation in Adam10 expression, suggesting that the expression of these two ADAMs may not be inter-dependent in FBs or myoFBs. ADAM17 is a strong sheddase of the ligands that activate EGFR.^18,39 Hence, ADAM17 can increase the bioavailability of EGFR ligands, activating EGFR in myoFBs leading to their excess activity and excess ECM production.^76–78

The integrity and density of the ECM produced by myoFBs can modulate tissue remodeling by impacting the function and interaction between the residing cells.^8-10 We found that reduced ADAM17 in myoFB leads to formation of a markedly less stiff infarct tissue, due to reduced activation of the EGFR, and subsequently the YAP pathways, which is an important mediator of mechanotransduction and cell response to matrix stiffness.⁶³ YAP pathway can become activated secondary to EGFR activation, or in response to increased mechanical stress, leading to YAP dephosphorylation and its nuclear translocation where it induces transcription of pro-fibrotic genes that can further increase the stiffness of the ECM and the tissue.^62–65 The decreased EGFR activation in Adam17^myoFB-KD myocardium reduced YAP activation and suppressed the subsequent excess transcription of ECM proteins which in turn ameliorated tissue stiffness in Adam17^myoFB-KD mice, further suppressing YAP activation.

The reduced ECM stiffness in Adam17^myoFB-KD infarct tissue and the increased vascularization limited infarct expansion, demonstrating the impact of myoFB on EC function. Tissue compliance is a determinant of EC morphology and function.^14–17 Moreover, FB-derived cytokines and growth factors, such as fibroblast growth factor (FGF),¹¹ CXCL12,¹² can impact EC proliferation and function. VEGFR2, whose activation promotes EC proliferation, is a well-known substrate for ADAM17 sheddase activity.¹³ Reduced ADAM17-mediated shedding of VEGFR stabilizes this receptor, promoting its activation and EC proliferation. The increased expression and activity of ADAM17 following myocardial infarction leads to uncontrolled activity of myoFB, excess deposition of ECM, increased stiffness and mechanical stress within the infarcted myocardium thereby limiting EC proliferation and microvascular density.^15,79,80 Modifying tissue stiffness to improve recovery from an infarction has been explored as a therapeutic approach.⁸¹ We demonstrate that genetic knockdown of ADAM17 in myoFB, or short-term pharmacological inhibition of ADAM17 during the ‘proliferative phase’ of post-MI remodeling, markedly reduces tissue stiffness, increases vascularization, and limits infarct expansion, LV dilation and dysfunction.

We have identified a therapeutic window for ADAM17 inhibition post-MI for long-term beneficial effects. While pharmacological inhibition of ADAM17 prior to MI induction did not offer any beneficial effects, short-term ADAM17 inhibition after the initial infarct formation was sufficient to limit the activity of myoFB, limited infarct expansion and LV dilation, and preserved cardiac function which lasted (at least) 4 weeks after MI. Therefore, this study identifies a therapeutic window for ADAM17 inhibition shortly after ischemic injury which will set the myocardium on the right path for optimal remodeling and prevents progression to heart failure.

Translational Perspective.

ADAM17 levels are increased in ischemic injury, which is associated with adverse outcomes. However, ADAM17 has diverse functions in different cell types, and therefore, its general or long-term inhibition is not clinically feasible. In this study, we identify the cell type and the timeline when inhibition of ADAM17 improves cardiac recovery post-MI. Our study shows, for the first time, that the initial increase in ADAM17 activity post-MI is essential for survival, but its short-term inhibition after MI is highly effective for long-term protective outcomes. This offers a highly feasible therapeutic approach in targeting ADAM17 post-MI.

Supplementary material

Extended Methods Supplementary material is available at Cardiovascular Research online. Extended Supplementary material online, Figures S1-S8.

Authors’ contributions

Z.K. and Y.L. conceived the research idea, Y.L. performed most of the experiments, prepared all the figures, and wrote the first draft of the manuscript. R.A.R. performed the HPRO assay and snRNAseq analyses with guidance from S.E.; F.W. performed all animal surgeries; H.L. and Y.L. performed the experiments and data analyses for infarct tissue stiffness assessment with guidance from L.W. S.E. provided expertise for snRNAseq experiments, analyses and interpretation; M.D.T. provided Tcf21-MerCreMer mice and guidance on characterization; G.Y.O. provided human heart specimens and input on data interpretation. Z.K. and Y.L. finalized the manuscript text and figures. All authors read the manuscript, provided critical input, and have agreed to the submission.

Funding

This work was supported by Canadian Institute for Health Research Project Grant to Z.K. (PJT-469620), G.Y.O. (PJT-462950), and Grant-in-aid from Heart and Stroke Foundation to Z.K. (G-22-0032063). Y.L. and R.A.R. are supported by Motyl Endowment Graduate Scholarship. Z.K. holds a Canada Research Chair (Tier 1) in Cardiovascular Extracellular Matrix.

Data availability

All data included in this manuscript will be made available upon reasonable request to the corresponding author.