MRTF-A promotes angiotensin II-induced inflammatory response and aortic dissection in mice

Aortic dissection (AD) is a major cause of acute aortic syndrome with high mortality due to the destruction of aortic walls. Although recent studies indicate the critical role of inflammation in the disease mechanism of AD, it is unclear how inflammatory response is initiated. Here, we demonstrate that myocardin-related transcription factor A (MRTF-A), a signal transducer of humoral and mechanical stress, plays an important role in pathogenesis of AD in a mouse model. A mouse model of AD was created by continuous infusion of angiotensin II (AngII) that induced MRTF-A expression and caused AD in 4 days. Systemic deletion of Mrtfa gene resulted in a marked suppression of AD development. Transcriptome and gene annotation enrichment analyses revealed that AngII infusion for 1 day caused pro-inflammatory and pro-apoptotic responses before AD development, which were suppressed by Mrtfa deletion. AngII infusion for 1 day induced pro-inflammatory response, as demonstrated by expressions of Il6, Tnf, and Ccl2, and apoptosis of aortic wall cells, as detected by TUNEL staining, in an MRTF-A-dependent manner. Pharmacological inhibition of MRTF-A by CCG-203971 during AngII infusion partially suppressed AD phenotype, indicating that acute suppression of MRTF-A is effective in preventing the aortic wall destruction. These results indicate that MRTF-A transduces the stress of AngII challenge to the pro-inflammatory and pro-apoptotic responses, ultimately leading to AD development. Intervening this pathway may represent a potential therapeutic strategy.


Introduction
Aortic dissection (AD) is a serious aortic disease which often results in a lethal outcome [1]. Although establishing preventive and therapeutic strategies is an urgent issue for AD, it is hampered by the elusive molecular pathogenesis of AD. Several genetic disorders are known to predispose individuals to aortic aneurysm and AD [2]. The best characterized genetic disorders for familial AD are mutation of FBN1 that causes Marfan syndrome, and mutations of TGFB2, TGFBR1, TGFBR2 and SMAD3 that cause Loeys-Dietz syndrome, underscoring the importance of TGF-β pathway for the homeostasis of aortic walls. Another group of culprit genes for familial AD are those encoding contractile proteins in smooth muscle cells (SMCs), including smooth muscle α-actin (ACTA2), myosin heavy chain (MYH11), and those encoding the regulators of SMC contraction, including myosin light chain (MYLK) and cyclic GMPdependent protein kinase (PRKG), indicating the importance of SMC cytoskeleton in pathogenesis of AD. Currently it is unclear how abnormalities in SMC cytoskeletal proteins result in AD development. Studies in human AD samples and animal AD models indicate that inflammatory response plays a central role in the pathogenesis of AD [3,4]. We also reported the involvement of inflammatory response in mouse models [5][6][7] and in human AD tissue [8]. Although these studies demonstrated the importance of inflammation, it remains elusive how inflammatory response is initiated or whether and how it is related to SMC cytoskeleton.
Recently, myocardin-related transcription factor (MRTF) cofactors were proposed to transduce neurohumoral and mechanical stimuli to stress response [9]. Specifically, neurohumoral and mechanical stimuli to cells promote Rho family GTPases-dependent actin polymerization that liberates MRTF to enter the nuclei and activate serum response factor (SRF)-dependent gene expression [9]. A series of studies have shown that MRTF-A, a member of MRTF family, regulates cardiovascular inflammation and remodeling in response to mechanical and humoral stimuli [10][11][12]. Therefore, MRTF-A is an attractive candidate that links insults on aortic wall, actin cytoskeleton, and inflammatory response in the context of AD pathogenesis. Current study was designed to test if MRTF-A is involved in the molecular pathogenesis of AD, using angiotensin II-induced mouse model of AD [13] with genetic deletion [11] or pharmacological inhibition [14] of MRTF-A.

Animal experiments
All animal experimental protocols were approved by the Animal Experiments Review Boards of Kurume University. Mice with genetic deletion of Mrtfa (MRTF-A-KO) was created as previously described [11] and backcrossed to C57BL6/J for more than 5 generations. All animal experiments ware done in male mice at the age of 10-11 weeks, as AD predominantly affects men [15]. C57BL/6J mice were purchased from Charles River Laboratories. We used wild type (WT) C57BL/6J mouse as a control for MRTF-A-KO mouse to facilitate the experiments and minimize the number of mice to sacrifice, because the reproductivity of MRTF-A-KO line was low in our hands. Angiotensin II (AngII, 1 μg/kg/min, Peptide Institute #4001, Osaka, Japan) was subcutaneously infused using osmotic minipumps (Durect Alzet #1007D, Cupertino, CA) for 4 days. Implantation of the minipumps filled with saline did not cause discernible aortopathy (S1 Fig). Mice were administered twice daily with 300 mg/kg CCG-203971 (Cayman Chemical #15075, Ann Arbor, MI) in 50 μL dimethylsulfoxide (DMSO) or 50 μL DMSO alone. Administration of CCG-203971 or DMSO was started 12 hours before starting AngII infusion and continued throughout the observational period. Systolic blood pressure was measured by the tail-cuff method (BP-98A, Softron, Tokyo, Japan).
At the end of experimental periods, mice were killed by pentobarbital overdose. For the histological analysis, we excised aortae 4 days after AngII infusion. For the expression analysis we obtained aortae with or without 1 day AngII infusion. The aortic samples were obtained after phosphate-buffered saline (PBS) perfusion at the physiological pressure from the root of aorta to just bellow subclavian artery branch, frozen in liquid nitrogen, and stored at -80˚C until analysis. We used 4% paraformaldehyde in PBS for histological analysis. For the expression analysis, perfusion was performed with PBS. For quantification of the tear area, 200 μl of 1% Evans blue solution was administered from orbital venous plexus before pentobarbital overdose, followed by PBS perfusion to wash out excessive Evans blue.

Quantification of area of tear and hematoma
Excised aortae was opened, pinned and photographed under natural light and 488 nm excitation light to obtain the fluoresce of Evans blue. We measured the area of tear and hematoma using Image-Pro PLUS software version 6 (Media Cybernetics, Rockville, MD, USA). We also measure the area of ascending aorta defined as an area from the root of aorta to just bellow subclavian artery branch.
We defined aortic dissection (AD) as the lesion with the intramural hematoma connected to the intimal tear. When the intimal tear was visualized by Evans blue perfusion, intramural hematoma was always associated with the intimal tear. Our collaborators also reported that the intramural hematoma in this model was associated with the disruption of intima-medial elastic lamellae by propagation-based phase-contrast synchrotron imaging [16]. As for the incidence of AD, the presence of AD was determined by the presence of intramural hematoma in the macroscopic images.

Expression analysis
For protein and mRNA expression analysis, aortic samples were pulverized using SK Mill (Tokken, Kashiwa, Japan), and the proteins were extracted with RIPA buffer. After resolving the proteins using the NuPAGE electrophoresis system (Invitrogen, Carlsbad, CA), western blotting was performed using antibodies to Myc (0.5 μg/mL, Cell Signaling Technology #9402, Danvers, MA), Stat3 (0.5 μg/mL, Cell Signaling Technology #12648), phospho-Stat3 (P-Tyr705, 0.5 μg/mL, Cell Signaling Technology #9145), MRTF-A (0.56 μg/mL, Proteintech #21166-1-AP, Rosemont, IL) and Gapdh (0.2 μg/mL, Merck Millipore #MAB374, Temecula, CA). As we noticed in preliminary study that protein expression levels were highly variable within a experimental group, we used 16 to 18 samples in a given group for western blotting. Images of whole membranes are shown in S2 Fig. For mRNA expression analysis, we used RNeasy kit (Qiagen, Hilden, Germany) to isolate total RNA. We performed transcriptome analysis using the SurePrint G3 Mouse Gene Expression v2 8x60K Microarray Kit (Agilent, Santa Clara, CA). The full dataset of the transcriptome analysis is available at Gene Expression Omnibus (accession # GSE138484). We performed the gene annotation enrichment analysis for the selected genes using Database for Annotation, Visualization and Integrated Discovery (DAVID) [17] with the Gene Ontology terms set to GOTERM_BP_FAT, GOTERM_CC_FAT and GOTERM_MF_FAT. The full lists of the DAVID analysis and corresponding gene descriptions are available as S1 and S2 Tables. mRNA expressions of Il6, Ccl2, Tnf and Actb were measured by quantitative real time polymerase chain reaction (qRT-PCR) using commercially available probes (primer set ID MA104898, MA108953, MA163499, MA050368: Takara Bio, Shiga, Japan).

Statistical analysis
All data are expressed as medians and interquartile ranges using box plots. Statistical analysis was performed with Mann-Whitney test for the comparisons of 2 groups. Kruskal-Wallis test was performed for multiple groups. Post-test analysis was performed by Dunn's multiple-comparison test. P < 0.05 was considered significant.

Role of MRTF-A in AngII-induced AD
We administered AngII (1 μg/kg/min) to WT mice that caused ascending aortic dissection (AD) in 4 days after AngII infusion ( Fig 1A) without significant changes in systolic blood pressure (Fig 1B), as previously reported [18]. Our collaborators have demonstrated that this AD model is characterized by the progressive disruption of intima-medial elastic lamellae that is associated with the intramural hematoma by three dimensional imaging [16]. Consistently, our histological observation showed intramural hematomas and tears defined as disruption of elastic lamella at 4 days of AngII infusion ( Fig 1C). For the quantitative analysis of AD, we measured the areas of intramural hematoma and intimal tear in the ascending aorta as visualized by Evans blue perfusion ( Fig 1D).
To evaluate the localization of MRTF-A in mouse aorta, we performed immunofluorescence staining for MRTF-A and smooth muscle α-actin (SMA), a marker of smooth muscle cells (Fig 2A). MRTF-A staining was detected mainly in the medial layer and overlapped with part of SMA-positive smooth muscle cells in the aortic wall. AngII infusion induced MRTF-A protein expression mainly in the medial layer as demonstrated by immunofluorescence staining ( Fig 2B). AngII-induced increase in MRTF-A expression was confirmed by western blotting ( Fig 2C). To clarify the role of MRTF-A in AD model, we compared the AD phenotype in WT mice and MRTF-A knockout (MRTF-A-KO) mice ( Fig 2D, Table 1). The incidence of AD in MRTF-A-KO mice (0.0%, n = 9) was significantly lower compared to WT mice (55.6%, n = 18; p<0.05). Quantitative analysis of hematoma and tear areas also demonstrated the significant suppression of AD phenotype in MRTF-A-KO mice (Fig 2E).

Role of MRTF-A in gene expression during AD development
To clarify the role of MRTF-A in AngII-induced aortic pathology, we performed transcriptome analysis of aortic tissues from WT and MRTF-A-KO mice with and without AngII. We performed the analysis at the first day of AngII administration before AD development to avoid the consequence of the catastrophic aortic wall destruction due to AD. We defined ADrelated genes as those with significant changes by AngII (P < 0.05, fold change > 2 or < 0.5). Among the AD-related genes, we further defined MRTF-A-regulate genes, as those with significant changes between WT and MRTF-A-KO ( Fig 3A). Within the AngII-sensitive genes, approximately 6,000 genes were upregulated and 700 genes were downregulated ( Fig 3B). The effect of Mrtfa deletion was a partial suppression of the AngII-induced changes. In the absence of AngII challenge, approximately 350 genes were upregulated and 250 genes were downregulated by Mrtfa deletion (Fig 3C). These findings suggested that MRTF-A mediated part of AngII-induced changes in gene expression. We performed the gene annotation enrichment analysis for the selected genes using Database for Annotation, Visualization and Integrated Discovery (DAVID) [17]. The DAVID analysis revealed that AngII-induced, MRTF-A-dependent genes were highly enriched for inflammatory response, angiogenesis and cell death ( Table 2). AngII-suppressed, MRTF-dependent genes were enriched for myofibril and muscle organ development (Table 3), although the enrichment scores were not as high as the AngIIinduced genes.

Role of MRTF-A in aortic wall inflammation and apoptosis
As the transcriptome analysis suggested that MRTF-A regulates the inflammatory response that plays a central role in AD pathogenesis [3,4], we performed quantitative analysis of inflammatory cytokines by qRT-PCR in aortic tissue (Fig 4). The mRNA expressions of Il6 and Ccl2 were induced by 1 day of AngII infusion in WT aorta, while this response was not observed in MRTF-A-KO mice (Fig 4A). AngII challenge did not significantly alter Tnf expression, although MRTF-A-KO aorta showed lower Tnf expression in the presence of AngII. AngII-induced and MRTF-A-dependent inflammatory response was also demonstrated at the protein level by the western blotting for activated (phosphorylated) Stat3 (Fig 4B and 4C). At the tissue level, infiltration of CD45-positive inflammatory cells was demonstrated near the site of intimal tear at day 4 of AngII infusion (Fig 4D). We next focused on apoptosis that may participate in AD pathogenesis [19]. We performed TUNEL staining to detect the double strand break of DNA, a marker of apoptotic cells, in WT or MRTF-A-KO aorta before, 1 day and 4 days after AngII infusion (Fig 5A). TUNEL-positive cells were not observed in either WT or MRTF-A-KO aortic tissue before AngII infusion. WT aortic tissue exhibited TUNEL-positive cells in the intimal and medial layers at day 1, and at the site of intimal tear at day 4 of AngII infusion, whereas MRTF-A-KO aorta did not. Double fluorescence labeling with cell type markers revealed the TUNEL staining colocalized with endothelial cell marker Cd31 and smooth muscle cell marker SMA (Fig 5B). To better characterize the MRTF-A-dependent inflammation and apoptosis, we examined the protein expression of Myc, a well-characterized regulator of apoptosis [20], as Myc mRNA expression was induced by AngII in an MRTF-A-dependent manner in our transcriptome analysis. Western blotting revealed that Myc protein was induced by AngII infusion in WT aorta, and to the lesser extent in MRTF-A-KO aorta (Fig 6A and 6B), suggesting that Myc induction by AngII was partly mediated by MRTF-A. Double fluorescence labeling revealed the colocalization of TUNEL staining and Myc expression (Fig 6C), consistent with the notion Table 2 that Myc regulates apoptosis pathway [20]. Tissue localization of TUNEL positive cells in the aortic medial layer was closely correlated with the Il6 expression in the adventitial layer ( Fig 6D).

Pharmacological MRTF-A inhibition in AD
The results thus far indicated that MRTF-A promotes AD development. Recently, CCG-203971, a specific inhibitor of MRTF-A-dependent transcription [14], has been reported to prevent the adverse tissue remodeling during inflammatory conditions [21]. To test the effect of the acute MRTF-A inhibition and to explore the clinical implication of our findings, we examined the effect of pharmacological MRTF-A inhibition on the AD model by treating WT mice with CCG-203971. The difference in the prevalence of AD between CCG-203971-treated group and vehicle-treated group did not reach statistical significance (Fig 7A, Table 1). CCG-203971 significantly suppressed the area of intramural hematoma by AD, whereas its effect on the tear area did not reach statistical significance (Fig 7B). AngII infusion caused induction of Il6 and Ccl2, which was not observed in mice treated with CCG-203971 ( Fig 7C). On the other hand, CCG-203971 did not significantly alter the appearance of TUNEL-positive cells at day 1 of AngII infusion (Fig 7D). These results indicated that pharmacological inhibition of MRTF-A by CCG-203971 partially suppressed AD phenotype.

Discussion
The main finding in this study was that MRTF-A mediates inflammatory and apoptotic responses, and promotes AD development by AngII challenge. The inflammatory and apoptotic responses occurred at day 1, as indicated by the gene expression and histological analyses. Because the appearance of aortic wall tear, as detected by the Evans blue perfusion, occurred at day 4 of AngII infusion, inflammatory and apoptotic responses likely precede AD development. The involvement of MRTF-A in AD pathogenesis was further supported by the fact that CCG-203971, an inhibitor of MRTF-A, prevented the AngII-induced AD development. As recent studies indicated the involvement of inflammatory response [5,6,22,23] and apoptosis [19,24] in AD pathogenesis, these data suggested that MRTF-A promotes AD by regulating inflammatory and apoptotic responses.  Proinflammatory molecules, including secreted cytokines and chemokines, and intracellular signaling molecular including Stat3 and NFκB play critical roles in AD pathogenesis [4-6, 23, 25]. [19] These molecules are activated by, and regulate the expression of cytokines such as Il6 and chemokines such as Ccl2 to orchestrate the behavior of resident cells including smooth muscle cells and infiltrating inflammatory cells [5,23,26]. Consistent with this notion, we found MRTF-A regulate expression of Il6 and Ccl2, and infiltration of Cd45-positive inflammatory cells in the AD model. On the other hand, a previous report described that Cd45-positive cells appeared only sparsely at the late stage of AngII-induced aortopathy [18]. This apparent discrepancy may be explained by the fact that we observed specifically at the site of intimal tear whereas the previous report seems to show the aortic tissue without apparent tear. These findings suggest that spacio-temporal analysis of cellular dynamics would be required to understand the molecular pathogenesis of AD.
Apoptosis of aortic cells, mainly SMCs, is observed both in human AD and mouse model of AD [24], and tightly coupled with the phenotype of AD in mouse model [19,27,28]. Causative involvement of apoptosis in aortopathy was demonstrate by the suppression of SMC apoptosis and thoracic aortic aneurysm by a caspase inhibitor in a mouse model of Marfan syndrome [29]. These findings support the notion that pro-inflammatory and pro-apoptotic responses are the central biological process in pathogenesis of AD. Consistent with this notion, we observed TUNEL-positive apoptotic cells in aortic media appeared at the same time with adventitial expression of Il6. On the other hand, while the localizations of apoptotic cells and Il6 expression were close to each other, they did not overlap. Therefore, a molecular and cellular network is likely to underlie the manifestation of apoptosis and inflammation as a stress response in AD pathogenesis [19]. The findings we obtained by MRTF-A inhibition suggest that pro-inflammatory and pro-apoptotic function is the basis of the MRTF-A-mediated AD development.
The role of MRTF-A in adverse cardiovascular remodeling is consistent with previous reports [10][11][12]30]. Mechanistically, MRTF-A is regulated by dynamics of actin cytoskeleton [31] to transduce neurohumoral and mechanical stimuli to the stress response [12]. In this regard, it is noteworthy that genetic abnormalities of SMC cytoskeletal and contractile proteins predispose the patients to AD [32]. Although it is unclear how mutations in cytoskeletal and contractile protein genes lead to AD development, current results suggest that abnormality in actin-dependent regulation of MRTF-A might be one of the mechanisms for the AD susceptibility.
The biological process under the control of MRTF-A seems context-dependent. MRTF-A is reported to mediate pro-inflammatory signal in various cell types including vascular SMCs [10], macrophages [33], and glomerular mesangial cells [34]. In the context of diseases, MRTF-A is pro-inflammatory in atherosclerosis [11,35], liver injury [36], and inflammatory bowel disease [37]. On the other hand, MRTF-A exerts anti-inflammatory effect in pulmonary arterial SMCs [38]. MRTF-A is neuroprotective in ischemic cerebral injury possibly by inhibiting neuronal apoptosis [39,40]. MRTF-A also prevents myocardial cell apoptosis [41]. Within the context of anticancer drug doxorubicin effect, MRTF-A can either cell-protective [42] or cell-toxic [43]. Of note, MRTF-A promotes atherogenesis by promoting macrophage survival and proliferation, and inflammatory response [35], exemplifying the complex role of MRTF-A and the relationship between inflammation and apoptosis. blotting for pStat3, Stat3 and Gapdh. Representative images (B) and quantitative analysis (C) of western blotting are shown. The samples were ascending aortae from WT or MRTF-A-KO mice with or without AngII administration for 1 day. The numbers in parenthesis indicate the number of biological replicates. � P < 0.05, �� P < 0.01, ��� P < 0.001, N.S. not significant. (D) Macroscopic and histological images are shown with EVG staining and Cd45 immunohistochemical staining along with the isotype control. Dotted line in the macroscopic image demarcates the aortic tissue. Aortic samples were obtained at day 4 of AngII infusion. Black and white arrowheads indicate the intimal tear and Cd45-positive staining, respectively. Bar 0.1 mm.
https://doi.org/10.1371/journal.pone.0229888.g004   Our data demonstrated that acute pharmacological inhibition of MRTF-A was effective in preventing AD. This finding suggests that MRTF-A plays an acute role in AD development, probably by promoting inflammation and apoptosis as discussed above. While MRTF-A protein was induced by AngII at day 4, AngII-induced Il6 and Ccl2 expressions were suppressed at day 1 by Mrtfa deletion or CCG-203971. Therefore, AngII is likely to activate MRTF-A both at the expression and functional levels. Whether MRTF-A is the trigger of AD initiation or the promoter of AD development awaits further study to delineate the chronological and functional sequence of the molecular events during AD development. Although our findings suggest that MRTF-A inhibition is a potential therapeutic opportunity for AD, the AD-inhibiting effect of CCG-203971 was not as complete as Mrtfa deletion. This may be due to the insufficient inhibition of MRTF-A by CCG-203971 during AD development. Alternatively, whole body Mrtfa deletion may developmentally alter the function of cells and tissue [44][45][46], which may affect AD pathogenesis. In addition, prevention of AD in the clinical situation is highly challenging. The annual incidence of AD in the general population is only 6 in 100,000, and prediction of AD is currently difficult, if not impossible [47]. Those who with genetic predisposition to AD, including Marfan syndrome, Loeys-Dietz syndrome and other forms of familial thoracic aortic aneurysm/dissection, may benefit from MRTF-A inhibition for preventing the development of life-threatening AD. However, even for the high risk population of AD, care should be taken for prophylactic chronic MRTF-A inhibition, because the function and its (patho)physiological role of MRTF-A seem context-dependent. Indeed, MRTF-A may play a protective role in neuronal and myocardial ischemia [30,39,40,48,49]. Another potential opportunity is the acute inhibition of MRTF-A after AD development to prevent further tissue destruction. For the MRTF-A inhibition therapy to realize, the role of MRTF-A needs to be clarified in the context of complex molecular and pathological changes during tissue destruction after AD onset.
In conclusion, our data demonstrated that MRTF-A promoted AD development possibly by activating proinflammatory and proapoptotic pathway. Further study for clarifying the cell type-and disease context-dependent function of MRTF-A would be essential for understanding the pathogenesis of AD, and for the development of preventive, diagnostic and therapeutic strategies for this fatal disease. Red rectangles indicate the area that were used in corresponding figures. MW; molecular weight marker, LC; loading control. Detection of Myc band: Myc antibody revealed multiple bands in western blotting. True band was determined as the main band in lysate of proliferating smooth muscle cells. Bands with identical molecular weight were also observed in aortic tissue lysate with AngII stimulation, but not in that without AngII. Because Myc was induced by AngII, we concluded that the AngII-dependent bands was a true band of Myc, as indicated by arrowheads. (TIFF) S1 Table. Gene annotation enrichment analysis. The table contains Entrez Gene IDs in the comparison groups shown in Fig 3B and 3C, and the results of the gene annotation enrichment analysis. (XLSX) S2 Table. Genes in annotation clusters. The table contains Entrez Gene IDs and descriptions of genes in the annotation clusters obtained by the gene annotation enrichment analysis of the comparison groups shown in Fig 3B and 3C. (XLSX)