Role of PAR1-Egr1 in the Initiation of Thoracic Aortic Aneurysm in Fbln4-Deficient Mice

Introduction

Thoracic aortic aneurysms (TAAs) are characterized by an abnormal enlargement of the aortic lumen with silent and progressive dilatation, which may lead to dissection and/or rupture with fatal consequences. Although mortality from TAAs has been gradually declined owing to the development of technologies in medical care, the incidence increases because of associated risk factors such as hypertension and atherosclerosis that are influenced by a modern lifestyle.¹ TAAs are often associated with heritable diseases with syndromic features such as Marfan syndrome and Loeys-Dietz syndrome, which exhibit a marked activation of the TGF-β (transforming growth factor beta) signaling.² In addition, there are heritable TAAs without syndromic features but with underlying alterations in the contractile apparatus of vascular smooth muscle cells (SMCs).³ Most recently, dysfunction of mechanosensing in the aortic wall in response to hemodynamics has been proposed to be a key driver of pathogenesis of TAAs.⁴

We previously established a mouse model of postnatal TAA by deleting the fibulin-4 gene (Fbln4) in vascular SMCs (Fbln4^SMKO, termed SMKO).⁵ Fbln4 is a secreted glycoprotein and a component of elastic fibers, where it is localized to microfibrils.⁶ In SMKO aortas, elastic fibers fail to form normal elastic lamina-SMC connections during the early postnatal period, leading to a compensatory upregulation of mechanoresponsive molecules, such as Egr1 (early growth response 1), ACE (angiotensin-converting enzyme), Thbs1 (thrombospondin-1), and a local elevation of Ang II signaling.^7,8 We also showed that serine/threonine phosphatase Ssh1 (slingshot 1) causes dephosphorylation of cofilin (active form) and disruption of actin filaments.⁸ Furthermore, inhibition of Thbs1 sufficiently prevented the development of ascending aortic aneurysms and improved the integrity of elastic fibers and restored actin filaments.⁹ However, the precise molecular pathways involved in the initiation of aneurysms driven by the altered mechanosensing are not fully understood.

Remodeling of extracellular matrix (ECM) by matrix proteases plays a vital role in cardiovascular homeostasis. Protease activated receptors (PARs) are tethered-ligand receptors that are activated by proteolytic cleavage of their extracellular domains.¹⁰ PARs are expressed on the surface of endothelium, smooth muscle cells, platelets, neutrophils, macrophages, and leukemic white cells¹¹ and regulates platelet aggregation, cell shape, adhesion, cell proliferation, chemokine production, and migration via the G-protein pathways.¹² PAR1 (protease-activated receptor 1) was identified >20 years ago as a thrombin receptor and 3 additional PARs have been identified so far: PAR2, PAR3, and PAR4.^13,14 PAR1 ligands are high-affinity serine proteases, including thrombin, plasmin, factor Xa, and APC (activated protein C), known as canonical activators,¹⁵ and noncanonical activation by matrix metalloproteinases (MMPs).¹⁶ PAR1 also acts as a sensor for altered proteases in the extracellular microenvironment.¹⁷ More recently, PAR1 has been shown to be critical for tissue remodeling such as angiogenesis and atherosclerosis.^18,19

Here, we show that PAR1 is markedly upregulated in SMKO and human TAAs, and PAR1-mediated signals control Egr1 expression, which is causal for aneurysm development in vivo. Mechanistically, PAR1 and its ligands, thrombin and MMP-9, are induced by increased mechanical stress and loss of Fbln4 as early as postnatal day (P)1 and generate abnormal microenvironment containing dysregulated protease in SMKO aortas. Pharmacological inhibition by thrombin inhibitor (Dabigatran) or factor Xa inhibitor (Rivaroxaban) ameliorated aneurysm phenotype in SMKO mice. Taken together, PAR1 upregulates the mechanoresponsive Egr1-Thbs1 pathway during aneurysm initiation. Our study demonstrates the synergistic and feed-forward interactions between mechanical stress and protease activation, leading to the development of aneurysms in SMKO mice.

Materials and Methods

All the data that support the findings of this study are available from the corresponding author upon reasonable request.

Mice

SMKO mice were generated previously and there were no phenotypic differences for aneurysm formation and incidence between the male and female.⁵Egr1 null mice were purchased from The Jackson Laboratory (B6N;129-Egr1^tm1Jmi/J, stock number: 012924). Fbln4^+/+, Fbln4^lxp/+ or Fbln4^KO/+ mice containing SM22α-Cre transgene were used as control in this study. Comparisons of the phenotype were performed between animals on the same genetic background and both males and females (≈1:1 ratio) were used in the study. All mice were kept on a 12 hours/12 hours light/dark cycle under specific pathogen free condition and all animal protocols were approved by the Institutional Animal Experiment Committee of the University of Tsukuba.

Histology, Immunohistochemistry, and Morphometric Analysis

Mouse or human aortas were harvested and perfusion-fixed with 4% paraformaldehyde and embedded in paraffin. Five-micrometer sections were stained with hematoxylin and eosin (HE), Hart’s (Elastic fibers) or Masson trichrome (Collagens). Images were digitally captured with Leica DM2000 microscope (Leica Microsystems, DM2000). Immunohistochemistry was done as previously described^8,9 and morphometric analysis was performed with NIH image J software (https://imagej.nih.gov/ij/index.html) as described previously.^8,9

Western Blot Analysis

Mouse or human aortas were harvested without perivascular adipose tissues. For mouse aorta, P30 thoracic aortas were divided into ascending parts (from the aortic root to the left subclavian artery) and descending parts, P1 thoracic aortas were used entirely. Aortas were minced in liquid nitrogen by pestle and dissolved in RIPA Lysis Buffer (Sigma-Aldrich, #R0278) containing 1% protease inhibitor (Sigma-Aldrich, #P8340) and 1% phosphatase inhibitor (Wako, #67-24381). The lysates were mixed with 3×SDS sample buffer with 2-mercaptoethanol (Wako, #133-14571) and boiled at 95°C for 5 minutes and then were subjected to SDS-PAGE. Proteins were transferred to a PVDF membrane Immobilon-P Transfer Membranes (Millipore, IPVH00010) and immunoblotted with indicated antibodies (Table I in the Data Supplement). Membranes were incubated with secondary antibody of anti-mouse (Bio-Rad, #170-6516) or anti-rabbit (Bio-Rad, #170-6515) and detected with Chemiluminescence kit (Santa Cruz Biotechnology, #sc-2048) or SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, #RF232643).

IPA Analysis

Ingenuity Pathway Analysis (QIAGEN) was carried out for searching upstream regulators. The software predicted the upstream genes of specific target through Grow Tool. We set as default for data resources, confidential level, relation types, node types, and mutation and set for tissue and cell lines as endothelial cell, smooth muscle cells and cardiomyocyte and disease for cardiovascular disease, connective tissue disorders. After prediction, the predicted pathway was re-designed by PathDesigner (Communication Infrastructure Corporation).

Human Thoracic Aortic Tissues

The participation of patients undergoing cardiac surgery was in accordance with the research protocol approved by the Clinical Ethics Committee of University of Tsukuba Hospital (approved number #H27-217). Each patient was provided a written informed consent for the collection of aortic tissue samples. The presence of thoracic aortic aneurysm was diagnosed and documented before surgery by computed tomography. The diagnosis of thoracic aortic aneurysm was confirmed at the time of surgery by experienced cardiothoracic surgeons, and clinical phenotype diagnosis was confirmed by standard histopathology. CTRL samples were obtained from punched aortic wall tissues of patients of coronary artery disease and aortic valve stenosis.

RNA Extraction and qPCR

RNA was purified from human aortas, ascending aortas of P30 mice, thoracic aortas of P1 pups and rat vascular SMCs using RNeasy Mini Kit (QIAGEN, #74104). Five hundred nanograms of total RNA was subjected to reverse transcription reactions by iScript Reverse Transcription Supermix (Bio-Rad, #170-8841). iTaq Universal SYBR Green Supermix (Bio-Rad, #1725121) was used for amplicon detection and gene expression was normalized to the expression of housekeeping gene GAPDH. PCR reactions were carried out in triplicate in a CFX96 Real-time PCR Detection System (Bio-Rad, #1855195) with one cycle of 3 min at 95°C, then 39 cycles of 10 seconds at 95°C and 30 seconds at 55°C. Levels of mRNA were determined using the ddCt method and expressed relative to the mean dCt of controls. Primer sequences are provided in Table II in the Data Supplement.

Gelatin Zymography

P1 CTRL and SMKO aortas were pulverized and homogenized in Tris buffer (10 mmol/L Tris-HCl, pH7.4, 150 mmol/L NaCl, 10 mmol/L CaCl₂) containing 0.1% Triton-X 100. Fifteen micrograms of protein extracts were loaded on 10% gelatin SDS-PAGE gels. After electrophoresis, gels were washed in 2.5% of TritonX-100 for 30 minutes 4× and incubated in enzyme activating buffer at 37°C for 72 hours. After incubation, gels were rinsed with water and stained with CBB solution.

Isolation and Primary Culture of Mouse SMCs

Primary mouse SMCs were isolated and cultured from P30 CTRL and SMKO ascending aortas. Ascending aortas were minced and incubated with DMEM media (Thermo Fisher Scientific, #41965039) supplemented with 20% (v/v) FBS (Thermo Fisher Scientific, #26140079), 250 U/mL collagenase type I (Affymetrix/USB, #AAJ13820MC), and 13.5 U/mL of elastase (Affymetrix/USB, #15475) for 3 hours at 37°C with gentle shaking. Cells were pelleted and re-suspended in DMEM media containing 20% FBS, 0.1 μg/mL of rhEGF (WAKO, #05907873), 1 μg/mL of rhFGF (WAKO, #06405381), and 1× Antibiotic-Antimycotic (Thermo Fisher Scientific, #15240062) in a 24-well dish. From the second passage, primary mouse SMCs were cultured in DMEM/F12 with 20% FBS and 1% of Antibiotic-Antimycotic.

Cell Culture and Thrombin

Rat vascular SMCs (Lonza, R-ASM-580, isolated from the aorta of adult male Sprague-Dawley rats) were cultured in DMEM/F12 media (Thermo Fisher Scientific, #11320033) with 20% FBS and 1% of Antibiotic-Antimycotic. Rat vascular SMCs were cultured with DMEM/F12-serum-free media for 24 hours, then re-cultured in DMEM/F12-serum-free media with 25 Unit/mL of thrombin from bovine plasma (Sigma-Aldrich, #9002044) for indicated time points (0, 0.5, 1, 3, and 6 hours). After re-culturing, cells were scraped and used for Western blot analysis.

Transfection of siRNA

Rat vascular SMCs were transiently transfected with scramble (Scr) siRNA (Thermo Fisher Scientific, #12935110), PAR1 siRNA (F2r RSS303633 [Thermo Fisher Scientific, #4331182]), Fbln4 siRNA (Efemp2 [Thermo Fisher Scientific, #RSS308656]) by using Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific, #13778030). Two days after transfection, cells were used in each experiment.

Stretch Assay

Cyclic stretch was performed using a uniaxial cell stretch system (Central Workshop, Tsukuba University) in the presence of 20% FBS. About 4×10⁵ primary mouse SMCs were plated on a silicon chamber with Attachment Factor (Thermo Fisher Scientific, #S006100) as described previously.^9,20 Cyclic stretch was performed with a frequency of 1.0 Hz (60 cycles/minute) and an elongation of 20% for 8 hours. After stretching, cell lysates and condition medium were harvested, condition medium was condensed by Amicon Ultra Centrifugal Filters (Millipore, # UFC200324).

Thrombin and MMP-9 Activity Assay

Thrombin and MMP-9 activity were measured using SensoLyte 520 Assay Kit (Anaspec, #AS-72129 for thrombin and #AS-71155 for MMP-9). Aortic tissues were excised and homogenized in assay buffer. Fifteen micrograms of protein extracts were used for the detection of active form of enzyme. Activity reactions were carried out in duplicate in Multi-label plate reader Wallac 1420 ARVOsx (PerkinElmer).

Dabigatran and Rivaroxaban Treatment In Vivo

SMKO and control pups were divided into 2 groups: vehicle control and inhibitor treatment. Dabigatran (30 µg/g body weight; Combi-Blocks, QB-6987) or factor Xa inhibitor, Rivaroxaban (10 µg/g body weight; Chemscene LLC, CS-0555) or saline was administered orally to P1 pups as described previously.^21–23 The treatments were continued from P1 to P30 every day and at P30, pups were sacrificed and the aortas were harvested for evaluation of the aneurysm phenotype.

Statistical Analysis

All experiments are presented as mean±SEM. Statistical analysis was performed using Prism 8 (Graph Pad, ver. 8.4.0). Shapiro-Wilk test was used for the normality test. When the data followed normal distribution, statistical significance was determined by either unpaired Student t-test, 1-way or 2-way ANOVA followed by Tukey multiple comparison test. If the normality assumption was violated, nonparametric tests were conducted. Mann-Whitney U test was used in Figure 2C (Egr1 and Thrombin) and Figure 3B (PAR3), Kruskal-Wallis test with Dunn multiple comparisons was used in Figure 1D (total vessel area), Figure 1E (Thbs1), and Figure IC in the Data Supplement (total vessel area). P<0.05 denotes statistical significance.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Results

Deletion of Egr1 Prevented Aneurysm Formation in Smooth Muscle Cell-Specific Knockout Aorta

We have previously shown that the deletion of Thbs1 prevents aneurysm formation in SMKO,⁹ and Egr1 is known to regulate the promotor activity of Thbs1 with the subsequent transcription and protein synthesis.²⁴ Egr1 is a zinc-finger transcription factor that responds to various stimuli, including mechanical stress, cell proliferation, and differentiation.^25,26 Based on these observations, we hypothesized that Egr1 contributes to the pathogenesis of aortic aneurysm in SMKO mice by inducing Thbs1. To test this hypothesis, we generated SMKO mice on an Egr1-null background (termed DKO: SMKO;Egr1 knockout). Generation of DKO mice was confirmed by genotyping, and Fbln4^+/+;Egr1^+/+, Fbln4^loxp/+; Egr1^+/+, Fbln4^KO/+; Egr1^+/+ mice were used as controls (CTRL) in the following experiments (Figure 1A). Aneurysms were examined at 1 month of age in comparison to respective CTRL littermates. Aneurysms were prevented in DKO mice (6 out of 14; 42.8%), and some of SMKO;Egr1^+/− mice also showed amelioration of the aneurysm (6 out of 23; 26.0%, Figure 1B; Table 1) while none of SMKO; Egr1^+/+ showed improvement of the aneurysm phenotype. Histologically, elastic fibers in rescued DKO aortas were much organized compared to SMKO aortas, whereas collagen levels were comparable between SMKO and rescued DKO aortas (Figure 1C). Morphological analysis revealed that the internal elastic lamina (IEL) perimeter and outer perimeter were smaller in DKO compared with SMKO aortas; however, wall thickness and total vessel areas remained unchanged (Figure 1D). Thbs1 was significantly downregulated in all rescued DKO aortas compared to SMKO aortas (Figure 1E), although some of the nonrescued DKO aortas showed decreased expression of Thbs1 (Figure I in the Data Supplement). These data indicated that Egr1 is involved in the pathogenesis of aortic aneurysm in SMKO mice.

Upregulation of Egr1, PAR1, and Thrombin in Human TAAs

To explore the upstream signaling(s) of Egr1-Thbs1, we conducted the Ingenuity pathway analysis. F2rl3 (PAR4),²⁷Elk3,^26,28 and F2r (PAR1)^27,29 were identified as potential upstream regulators of Egr1 (Figure II in the Data Supplement). To investigate if these genes are involved in the pathogenesis of human TAA, we examined transcription levels of these genes and other genes previously described to be involved in the pathogenesis of human TAA.^30,31 TAA samples were obtained from nonsyndromic sporadic patients with TAA who underwent surgery. Nonaneurysmal CTRL samples were obtained from aortic wall punch biopsies of patients undergoing coronary artery bypass surgery. There were no differences regarding sex, age, metabolic rate, blood pressure, and cardiac functions between CTRL and TAA patients (Table 2). Several genes in the angiotensin signaling pathway, vascular SMC contractile markers and synthetic markers were upregulated in TAAs compared with CTRL aortas, whereas the transcript level of FBLN4 was decreased in TAAs (Figure 2A). ELK3 is a member of the ERK-regulated TCF (ternary complex factor) subfamily and acts with the transcription factor SRF (serum response factor) to activate mitogen-induced transcription.^32,33 Interestingly, SRF was markedly downregulated in TAAs, whereas its target genes of vascular SMC contractile markers, including ACTA2, MYH11, and SM22α, were upregulated in TAAs (Figure 2A). PAR1 was also upregulated in TAAs (Figure 2A). Since PAR1 is known to be involved in cardiovascular diseases,^11,19 we focused on PAR1 for further analyses. Consistent with the transcript levels, protein expressions of PAR1 and Egr1 were significantly increased in TAA samples compared with CTRL (Figure 2B and 2C), and the cleaved form of thrombin (prethrombin and active thrombin) was also increased (Figure 2B and 2C). PAR1 was negatively correlated with heart rate (Pearson r=−0.445, P=0.010 in Figure III in the Data Supplement) and Egr1 level was positively correlated with heart rate (Pearson r=0.403, P=0.022) and systolic blood pressure (Pearson r=0.435, P=0.012) as shown in Figure IV in the Data Supplement. No correlations were found between thrombin levels and clinical features (Figure V in the Data Supplement). Histological analysis showed that PAR1 was detected in the entire aortic wall, including intima, medial layers, and adventitial layers in TAAs, and most strongly in the endothelial layer of vasa vasorum (asterisks in Figure 2D and Figure VI in the Data Supplement), whereas CD41 was rarely observed in the aortic wall (Figure VI in the Data Supplement). These data suggest that the upregulation of PAR1 in human TAAs is not predominantly because of platelet-derived PAR1. Taken together, these data demonstrated that the elevation of Egr1, PAR1, and thrombin were associated with nonsyndromic sporadic TAA in humans.

PAR1 and Its Ligands Thrombin and MMP-9 Are Highly Activated in SMKO Aortas Before Aneurysm Formation

To understand the mechanistic relevance of PAR1 and Egr1 in aortic aneurysm formation, we returned to the aneurysmal mouse model (SMKO) and examined the expression of PAR1 in the initial stage of aortic aneurysm formation. As we reported previously, the aortic wall began to expand in SMKO aortas at P7, and the aneurysm established at P30.⁷ We defined this period as a critical therapeutic time window, in which aneurysm formation can be prevented by inhibiting Ang II-mediated signaling(s) caused by local upregulation of ACE,⁷ and subsequent downstream signaling, Ssh1-cofilin.⁸ The transcript level of Egr1 was highly increased in SMKO aortas at P1, and PAR1 was significantly increased in SMKO aortas among PARs at P1 before aneurysms formed (Figure 3A). Western blot analysis showed a marked upregulation of PAR1, PAR3 and Egr1 in SMKO aortas at P1 (Figure 3B). Next, we examined thrombin expression in SMKO aortas. Active thrombin was converted from prothrombin and prethrombin through proteolytic cleavage by factors Xa and Va (Figure 3C). Western blot analysis showed that thrombin and prethrombin were highly increased in SMKO aortas at P1 (Figure 3D), and thrombin activity was significantly increased in SMKO aortas compared with CTRL aorta (Figure 3E). In addition, MMP-9, a known ligand of PAR1, was activated in SMKO aortas at P1 (Figure 3F and 3G). These data strongly suggest that PAR1 is activated at the initial stage of aortic aneurysm and may serve as a trigger for subsequent aneurysmal changes in SMKO aortas.

PAR1 Is Markedly Increased in SMCs During Aneurysm Formation in Smooth Muscle Cell–Specific Knockout Aortas

To determine whether PAR1 is continuously expressed during aneurysm development, we evaluated the transcription and protein levels of Egr1 and PARs in CTRL and SMKO aorta after aneurysms were formed. Western blots showed that Egr1, PAR1, and thrombin, but not PAR3, were upregulated in the ascending aortas of P30 SMKO mice (Figure 4A). Both Egr1 and PAR1 transcripts were also upregulated in SMKO aortas at P90 (Figure VII in the Data Supplement). Immunofluorescence staining revealed the thrombin expression in SMCs in addition to ECs and adventitial cells in SMKO aortas, indicating that thrombin was most likely derived from the vascular wall (Figure 4B; Figure VIII in the Data Supplement). We also confirmed the upregulation of thrombin activity in P30 SMKO aorta (Figure 4C). These data suggest that PAR1 was initially activated in ECs and adventitial cells, then expanded to SMCs in the ascending aorta of SMKO during aneurysm development. The ascending aorta is under complex mechanical stimuli because of shear stress and pulsating pressure that may contribute to the increased susceptibility to aneurysm formation.³⁴ In addition, SMC-specific Fbln4 knockout mice, but not the endothelial cell-specific Fbln4 knockout mouse developed aneurysms.⁵ Therefore, we further focused on the role and regulation of PAR1 in SMCs with respect to mechanical stretch. Primary mouse SMCs isolated from CTRL and SMKO aortas were subjected to cyclic stretch (1.0 Hz; 20% strain) for 8 hours. As Figure 4D shows, PAR1 was increased by mechanical stretch as previously reported in human pulmonary ECs.^9,35 Furthermore, prothrombin was markedly increased in conditioned media of CTRL and SMKO, although prethrombin and active thrombin were undetectable in this condition. These data suggest that the precursor of thrombin is produced by SMCs under mechanical stretch.

• Download figure
• Download PowerPoint

Thrombin-Induced Upregulation of Egr1 and Thbs1 Is Mediated by PAR1

To examine whether Egr1-Thbs1 is regulated by PAR1 and if deletion of Fbln4 enhances this signaling pathway in SMCs, we performed small interfering RNA (siRNA)-mediated knockdown (KD) of PAR1 and Fbln4 in rat vascular SMCs. Scr (as CTRL), PAR1, or Fbln4 siRNA was administered to rat SMCs, and the efficiency of KD was confirmed by quantitative polymerase chain reaction (Figure 5A). We then examined the response of PAR1KD or Fbln4KD cells to thrombin in the absence of serum. In CTRL cells, thrombin (25 U/mL) treatment induced Egr1 expression at 1 and 3 hours and subsequently induced Thbs1 at 3 and 6 hours. Knockdown of Fbln4 did not show a statistically significant difference in thrombin-induced Egr1 and Thbs1 expression compared with CTRL cells. In contrast, induction of Egr1 and Thbs1 by thrombin was significantly suppressed in PAR1KD cells (Figure 5B). These data indicated that the thrombin-induced upregulation of Egr1 and Thbs1 was mediated partly in a PAR1-dependent manner.

• Download figure
• Download PowerPoint

We finally asked if we could prevent aneurysm formation by pharmacologically inhibiting thrombin. We treated SMKO mice with Dabigatran (thrombin inhibitor, n=4) and Rivaroxaban (factor Xa inhibitor, n=5) from P1 to P30 and examined the formation of aneurysm at P30. Neither drug affected body weight during treatment (Figure IXA in the Data Supplement). Dabigatran-treated (2 out of 4) and Rivaroxaban-treated (2 out of 5) SMKO mice showed amelioration of aneurysm phenotype (Figure IXB in the Data Supplement). Although thrombin activity in the aortas was decreased significantly in drug-treated groups compared with untreated SMKOs, the levels were still higher than those of controls (Figure IXC in the Data Supplement). Taken together, these results suggested that the increased thrombin activity may be one of the contributing factors for the aortic aneurysm initiation in SMKO mice.

Discussion

In this study, we reported that the genetic deletion of Egr1 negatively impacted the formation of aortic aneurysms in SMKO mice, and PAR1 was upstream of Egr1-Thbs1 in TAAs in SMKO mice. Consistently, PAR1 and Egr1 are both upregulated in human TAAs. In SMKO aortas, PAR1 was abundant and activation of thrombin and MMP-9 was evident before the aneurysm formation. In vitro analysis revealed that thrombin- and mechanical stretch induced expressions of Egr1 and Thbs1 in a PAR1-dependent manner. Furthermore, the loss of Fbln4 increases MMP-9 activity in SMKO aortas, all of which tips the balance for the proteolytic cleavage of PAR1 in the initiation of aortic aneurysm formation (Figure 6).

• Download figure
• Download PowerPoint

Mechanical Stress Responsive Factor Egr1 Contributes to the Aneurysm Formation in SMKO Aorta

Egr1 is involved in the response to stress in various organs and emerges in a variety of pathological conditions.³⁶ Egr1 contains a high conserved DNA-binding domain composed of zinc fingers that binds to the prototype target GC-rich consensus sequence GCG(G/T)GGGCG. We and others have shown that mechanical stretch induces Egr1 in vascular SMCs^9,20,25, and pressure overload upregulates Egr1 in the ascending aorta in vivo.⁸ In the context of aortic aneurysms, several studies have suggested a role of Egr1 in the pathogenesis of intraluminal thrombus formation in human abdominal aortic aneurysm (AAA),³⁷ CaCl₂-induced AAA mice,³⁸ and angiotensin II-induced AAA in apolipoprotein E-deficient mice.³⁹ Based on our results using DKO (SMKO; Egr1^−/−) mice in this study and genetic deletion of Thbs1 (SMKO; Thbs1^−/−) in the previous study,⁹ we concluded that the inhibition of mechanotransduction pathway mediated by Egr1-Thbs1 was sufficient to prevent aneurysm formation in SMKO mice. However, the rescue efficiency of aneurysm phenotype in SMKO; Egr1^−/− (42.8%) mice was lower than that of SMKO; Thbs1^−/− (78.9%) mice. There are 3 other members of EGR family, Egr2, Egr3, and Egr4, all of which can be induced by growth factor and/or mechanical stimuli and bind to the same GC-rich consensus sequence. Indeed, we observed that Egr2 was upregulated in DKO mice, and the mechanical stretch-induced upregulation of Thbs1 was suppressed by Egr1 KD⁹ as well as Egr2 KD but not by Egr3 KD (data not shown). Our data are consistent with other studies showing the overlapping regulation of target genes by Egr1 and other EGRs, and a compensatory upregulation of EGR family members in the Egr1 null mice.^40,41 Therefore, the remaining level of Thbs1 by other EGRs in DKO might have supported the aneurysm formation.

Canonical Activation of PAR1 by Thrombin and MMP-9 in the Preaneurysm Lesions

A cleavage of the extracellular N-terminal domain of PAR1 by thrombin occurs at a canonical R₄₁-S₄₂ site, which is distinct from the MMP cleavage site, resulting in conformational changes in the transmembrane domain and subsequent intracellular signal transduction.^13,16 MMP-1 and MMP-13 cleave PAR1 at a noncanonical site, D₃₉-P₄₀, and S₄₂-F₄₃, respectively.^42,43 On the other hand, MMP-9 has been shown to cleave conventional PAR1 site (R₄₁-S₄₂) in activated microglia,⁴⁴ and we observed a marked upregulation of MMP-9 activity in the P1 aortas of SMKO mice. In our experiments, thrombin induced Egr1 and Thbs1 via PAR1, and mechanical stretch also induced PAR1, Egr1, and Thbs1⁹, whereas temporal deletion of Fbln4 in rat SMC did not affect mechanical signals. These data support the notion that mechanical signal transduction mediated by PAR1-Egr1-Thbs1 was triggered by the cleavage mediated by MMP-9 and thrombin.

Our previous study showed that Thbs1 was upregulated in ECs and SMCs underneath ECs in the SMKO aortas.⁹ Similarly, we observed thrombin expressed in ECs and SMCs during aneurysm development. This observation may suggest that PAR1 propagates signals derived from ECs to SMC layers. Current study, however, failed to identify signals and a mode of signal transduction from ECs to SMCs in SMKO aortas. Vascular ECs can communicate with SMCs via gap junctions comprised of connexin (Cx) protein family, including Cx-37, Cx-40, and Cx-43.⁴⁵ Interestingly, Cx-43 hemichannels are controlled by thrombin-induced cytosolic Ca²⁺,^46,47 and Cx-43 promotor activity is regulated by PAR1 through the binding of SP-1 and AP-1 transcription factors in melanoma cells.⁴⁸ Several studies showed that Cx-43 is involved in vascular injury.^49,50 Therefore, it will be interesting to examine the interactions between ECs and SMCs possibly mediated by PAR1 and Cx-43 in aortic aneurysms and other vascular diseases.

Loss of Fbln4 Induces Activation of MMP-9 in the Aorta

MMPs are responsible for the degradation of the ECM in aortic aneurysms and upregulation of MMP-9 also play a crucial role in Marfan syndrome.^51,52 We observed the activation of MMP-9 in the initial stage of aneurysm development in SMKO mice, as well as in the mouse model with reduced Fbln4 expression (Fbln4^R/R).^7,53 There are possible explanations for the mechanism by which Fbln4 deficiency initiates excessive MMP-9 activation. First, TGF-β signaling has been shown to upregulate MMP-9 expression in vivo and in vitro.^54,55 TGF-β is secreted as an inactive form, then tethered onto microfibrils via latent TGF-β binding proteins.⁵⁶ Absence of Fbln4 may affect microfibril assembly and disrupt tethering of the inactive TGF-β, increasing the bioavailability of TGF-β. MMP-9 is also known as an activator of TGF-β by proteolytic cleavage of the latent TGF-β-binding proteins. Therefore, increased TGF-β signaling in SMKO aortas possibly induces transcription of MMP-9, in turn mediates TGF-β activation, creating a positive feed-forward loop. Second, LOX (lysyl oxidase) has been shown to enhance elastin synthesis and suppress MMP-9 activity.⁵⁷ LOX is a copper-dependent enzyme that catalyzes cross-linking of elastin molecules and is secreted as an inactive proenzyme, containing a N-terminal propeptide domain followed by the catalytic domain. The propeptides of the proenzyme are eventually cleaved by proteases, producing mature LOX. Since LOX activity is regulated in a Fbln4-dependent manner,⁵⁸Fbln4 deficiency may cause inactivation of LOX and mediate an increase of the MMP-9 activity.

Less is known about the relationship between mechanical stretch and MMP-9 in aneurysm models. Recently, using a rat abdominal aortic aneurysm and dissection model, mechanical loading has been shown to induce MMP-9 expression via the stretch-activated channel⁵⁹ or via p-ERK1/2 and inflammatory mediators.⁶⁰ Cyclic stretch upregulates Nox1 NADPH oxidase and ROS in a MEF2B-dependent manner, leading to augmentation of MMP-9 activity.⁶¹ In addition, Egr1 has also been shown to directly upregulate MMP-9 transcription in response to TNFα in nonvascular cells⁶² or in collaboration with snail and SP-1,⁶³ in nonstretch condition. Taken together, our study connects mechanical stress to MMP-9, which acts as a ligand for PAR1 and induces downstream signaling involving Egr1-Thbs1, which again forms a positive feed-forward loop and generates a microenvironment with dysregulated proteases.

Therapeutic Potential of Inhibiting PAR1 for Human TAA

Our current study does not directly address whether PAR1 inhibition can rescue the aneurysmal phenotype in SMKO mice. Since PAR1 null mice mostly die at around embryonic day 9 to 10,⁶⁴ pharmacological approach may be more suitable to block activation of PAR1 in SMKO mice. As we showed here, single administration of thrombin inhibitor (Dabigatran) or factor Xa inhibitor (Rivaroxaban) did not completely rescue aneurysm formation in SMKO mice, which correlated with the remaining thrombin activity. A higher dose of Dabigatran or Rivaroxaban, or combination of these drugs may be effective for prevention of aneurysms; however, it may increase risk of aortic dissection in Marfan patients, where blood clot formation was often observed.^65,66 The potential use of PAR1 inhibitors with a combination of MMP-9 inhibitors and/or low doses of thrombin inhibitors may be effective for the prevention of aneurysm formation. It is of note that parenteral administration of factor Xa/IIa inhibitor (enoxaparin) and FXa inhibitor (fondaparinux) effectively inhibits PAR2-, but not PAR1-, mediated Smad2/3 signaling and MMP-2 expression in angiotensin II-induced aortic dilatation in ApoE^−/− mice.⁶⁷ The potential difficulty is that PARs form heterodimers: PAR1-PAR4, PAR1-PAR3, PAR1-PAR2, and PAR3 acts as an allosteric regulator of PAR1, enhancing the Gα₁₃ coupling.⁶⁸ Thus, inhibitors have to target both PARs effectively. Furthermore, PAR2 deletion in cardiac fibroblasts upregulates PAR1 expressions, exhibiting a compensatory relationship among PARs.⁶⁹ Nonetheless, these reports provide the potential combinatorial strategies to block multiple pathways that function as the initial signal for aneurysm formation. Further investigation is necessary to establish an optimal drug protocol for patients with TAA.

Footnotes