The role of-arrestin 2-dependent signaling in thoracic aortic aneurysm formation in a murine model of Marfan syndrome

James W. Wisler, Emily M. Harris, Michael Raisch, Lan Mao, Jihee Kim, Howard A. Rockman, and Robert J. Lefkowitz Department of Medicine, Duke University Medical Center, Durham, North Carolina; Department of Dermatology, Duke University Medical Center, Durham, North Carolina; Department of Cell Biology, Duke University Medical Center, Durham, North Carolina; Department of Molecular Genetics, Duke University Medical Center, Durham, North Carolina; Department of Biochemistry, Duke University Medical Center, Durham, North Carolina; and Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina

MARFAN SYNDROME (MFS) is a systemic, autosomal dominant disease of connective tissue cause by mutations in the gene encoding fibrillin-1, a major component of the microfibril network of the ECM (13,15).This disorder is characterized by a variety of clinical manifestations, including elongated limbs, bony overgrowth, craniofacial abnormalities, subluxation of the lens of the eye, as well as myxomatous degeneration of the cardiac valves, particularly the mitral valve.The leading cause of death in MFS is secondary to complications related to the development of thoracic aortic aneurysms (TAAs) (6).
Work over the past two decades has demonstrated that TAA formation in MFS results from multiple dysregulated signaling pathways involving a complex network of signaling mediators and receptors.However, many of these pathways are poorly understood with regard to their contribution to TAA development.Consequently, an enhanced understanding of the pathogenic signaling mechanisms involved in TAA formation may stimulate the development of more efficacious, safe, or targeted therapeutic options to prevent or treat this deadly complication of MFS.
In addition to aberrant transforming growth factor-␤ (TGF-␤) signaling, Habashi et al. (19) identified that blockade of the Ang II (ANG) type 1a receptor (AT 1a R) inhibited TAA development in a murine model of MFS.The beneficial effect of AT 1a R blockade on aortic root dilation was independent of the hemodynamic effects of losartan, and identified AT 1a R-mediated signaling was pathogenic in MFS (19).It was subsequently shown that in addition to activation of TGF-␤ ligands and receptors, AT 1a R-mediated signaling contributed to TAA development via upregulation of proaneurysmal signaling mediators, such as matrix metalloproteinases (MMPs) (25).
The AT 1a R represents a prototypical G protein-coupled receptor (GPCR).The canonical signal transduction paradigm of GPCRs is characterized by an agonist-induced conformational change upon ligand binding to the receptor.This conformational change allows the receptor to interact with G proteins, as well as the non-G protein transducers, ␤-arrestins (␤arr) 1 and 2 (11).Originally identified as proteins that mediate receptor desensitization (4,30) and internalization (17,23,24), work over the past two decades has demonstrated that arrestins also mediate G proteinindependent signaling for multiple GPCR types [for review, see DeWire et al. (11)].␤arr2 has been shown to mediate AT 1a R signaling, including activation of MAPKs, such as ERK1/2, by serving as a scaffold for proteins that facilitate signaling to ERK1/2 (26).AT 1a R-mediated, ␤arr2-dependent ERK1/2 activation is both spatially and temporally distinct from ERK1/2 activated via G q proteins (3).This distinct signaling profile, engendered by ␤arrs, leads to unique physiologic (22) and pharmacologic (34) properties in both receptor-and tissue-specific manners.
Signaling within vascular smooth muscle cells (VSMCs) is thought to play a key role in the pathogenesis of TAA in MFS (5,29).In VSMCs, ANG signaling via the AT 1a R stimulates several cellular processes known to be involved in aneurysm formation, including cellular proliferation, migration, fibrosis, and protein synthesis (12,22,31,35).SMAD2 and -3 (31,35) as well as ERK1/2 (21) have been shown to mediate these cellular processes independent of one another.Aberrant signaling within VSMCs, including ERK signaling, has been linked to TAA development in MFS (18).␤arr2 has also been shown to have a role in AT 1a R-mediated cellular proliferation, migration, fibrosis, and protein synthesis, independent of G protein activation (12,22,33).Recent work has also identified ␤arr2 as a necessary signaling mediator for ANG-induced expression of several proaneurysmal mediators, including MMP-2 and -9, in the murine aorta (33).
MMP-2 and -9 have previously been identified as playing a key role in the pathogenesis of TAA formation in MFS via enhanced elastolysis of the elastic matrix with resultant impaired structural integrity of the aortic wall (8,9,36).Production of MMP-2 and -9 has been shown to be both AT 1a R (33,38) and ERK dependent (14,28) in aortic VSMCs and as noted above, is ␤arr2 dependent in the murine abdominal aorta (33).In addition, upregulation of MMPs, including MMP-2 and -9, has been identified in VSMCs from patients with MFS (29), as well as VSMCs in a murine model of MFS (5).Therefore, we hypothesized that AT 1a R-mediated, ␤arr2-dependent signaling would mediate TAA formation in MFS by regulating the expression of the proaneurysmal signaling mediators MMP-2 and -9 in the murine aorta.

MATERIALS AND METHODS
Study design.The objective of this work was to evaluate whether ␤arr2-dependent signaling contributes to TAA formation in MFS.This project was carried out using a murine model of MFS, as well as tissue harvested from these animals and primary aortic root VSMCs harvested from Sprague-Dawley rats.For in vivo study, a power calculation was performed a priori to determine the neces-sary sample size.To detect a 10% difference in aortic root diameter with an ␣ value of 0.05, a ␤ value of 0.20, and a value of 0.15, 13 animals were required for each group.Transthoracic echocardiography was planned to continue until 36 wk of age, after which time, tissue would be harvested for analysis.Data were collected and analyzed in an intention-to-treat format with the primary outcome measure specified as aortic root diameter, as assessed by transthoracic echocardiography.As noted below, ECG interpreters were blinded to genotype.
Materials.ANG was synthesized by GenScript USA (Piscataway, NJ) to Ͼ98% purity; quality control was by HPLC and mass spectrometry.For stimulation of primary cells, the concentration of ANG was 100 nM.Losartan (final concentration of 1 M) was purchased from Sigma-Aldrich (St. Louis, MO).The MEK1/2 inhibitor U0126 (final concentration of 5 M) and EGFR-specific inhibitor tyrphostin AG1478 (final concentration of 250 nM) were purchased from Calbiochem (San Diego, CA).Erlotinib (final concentration of 5 mg/ml) was purchased from Selleck Chemicals (Houston, TX).TGF-␤ neutralizing antibody was purchased from R&D Systems (Minneapolis, MN).
Cell culture.Primary aortic root VSMCs were harvested from 150 g male Sprague-Dawley rats in the following manner: the thoracic aorta was excised from the aortic root to the brachiocephalic artery.Individual aortic roots were kept in separate dishes, and VSMCs were obtained by enzymatic digestion, as described previously (22), and maintained in DMEM, supplemented with 10% FBS and 1% penicillin-streptomycin. Slow growing, early passage cells (Ͻ5) were used for assay.Cultured cells from individual animals were kept separate throughout culturing and experimentation.
Experimental animals.Fbn C1039G/ϩ mice (C57BL/6J background) were crossed with ␤arr2 Ϫ/Ϫ mice (C57BL/6J background) to generate Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ animals.Littermate controls were used for in vivo study.Research with animals carried out for this study was handled according to approved protocols and animal welfare regulations of Duke University Medical Center's Institutional Review Boards.Male Sprague-Dawley animals (150 g) were purchased from The Jackson Laboratory (Bar Harbor, ME).
Transthoracic echocardiography.Transthoracic two-dimensional echocardiography was performed using a Vevo 2100 echocardiograph (FUJIFILM VisualSonics, Toronto, ON, Canada).Images were analyzed using Vevo 2100 software (FUJIFILM VisualSonics).Immediately before imaging, animals were anesthetized with a subcutaneous injection of ketamine/xylazine (90 and 4.5 mg/kg, respectively).Serial ECGs consisted of apical and short-and long-axis images of the heart, as well as aortic root and ascending aorta, and were obtained beginning at 8 wk and then every 4 wk thereafter until 36 wk of age.Two independent readers, blinded to animal genotype, measured aortic root diameter on three different images at each time point for each animal.Average aortic root diameter was calculated, and a third independent reader, also blinded to animal genotype, read any image for which the difference between readers 1 and 2 fell outside of 1 SD of the average difference between readers 1 and 2. When a third read was obtained, all three measurements were averaged to determine the aortic root diameter.
Hemodynamic study.After anesthetizing the animal with a subcutaneous injection of ketamine/xylazine (90 and 4.5 mg/kg, respectively), a high-fidelity transducer was inserted into the right carotid artery.Blood pressure was recorded continuously with a pressure- recording system (ADInstruments, Houston, TX).Continuous pressures were recorded in the proximal aorta, as well as in the left ventricle.
Erlotinib infusion.Beginning at 8 wk of age, transthoracic ECG was obtained as above.After anesthetizing the animal with subcutaneous injection of ketamine/xylazine (90 and 4.5 mg/kg, respectively), a subcutaneous osmotic pump (Alzet 2004; Durect, Cupertino, CA), containing either vehicle or erlotinib (5 mg/ml), was placed in the subcutaneous tissue immediately lateral to the spine on the posterior of the animal.After 4 wk, the initially placed pump was removed and replaced with a second pump containing the same treatment (vehicle vs. erlotinib).A follow-up transthoracic ECG was obtained after 8 wk of treatment to assess aortic root diameter, as detailed above.
Immunoblotting.Immunoblotting was performed using a modification of the methods of Kim et al. (21).For immunoblotting of stimulated VSMCs, 80 -90% confluent VSMCs in six-well plates were starved in serum-free medium for at least 48 h.Cells were stimulated for 15 min [for phosphorylated (p)ERK1/2 and pSMAD2/3 detection] or 24 h (for MMP-2 and -9 and pSMAD2/3).After stimulation, directly adding 2ϫ SDS-sample buffer, followed by sonication with a microtip at 30% amplitude for 15 s, solubilized monolayers.Equal amounts of cellular extracts were separated on 4 -20% Tris-glycine polyacrylamide gels and transferred onto nitrocellulose membranes for immunoblotting.
siRNA transfection.Immediately before transfection, 80 -90% confluent primary VSMCs plated in 100 mm dishes were placed in DMEM, containing 10% FBS without antibiotics.DharmaFECT 2 Transfection Reagent (GE Dharmacon, Lafayette, CO) was diluted in serum-free medium (0.1/100 l medium) and allowed to equilibrate for 5 min.Meanwhile, siRNA was diluted to 50 nM in serum-free medium and also allowed to equilibrate for 5 min, according to the manufacturer's instructions.The tube containing diluted siRNA was then added drop-wise to the tube containing the transfection reagent and was mixed vigorously and allowed to equilibrate at room temperature for 20 min.The transfection reaction was then added drop-wise to 100 mm cell plates.After 48 -60 h, cells were split into six-well dishes with antibiotic-free media.Twenty-four hours after splitting, the media were removed and replaced with serum-free media containing 0.1% BSA.Cells were serum starved for 48 h and stimulated as above.Immunoblotting was carried out as described above.Fig. 3. Matrix metalloproteinase (MMP)-2 and -9 gene and protein expression is reduced in TAA tissue of Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice.A: gene expression of MMP-2 and -9 in proximal ascending aortic tissue is decreased in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice at 12-16 wk of age.In addition, expression of collagen 1a1 (Col 1a1) is decreased in proximal ascending aortic tissue from Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice compared with Fbn C1039G/ϩ mice, whereas no differences are observed in the expression levels of SMAD2, transforming growth factor (TGF)-␤1, or thrombospondin-1 (Thbs).B: representative Western blots of TAA tissue of 2 Fbn C1039G/ϩ and 2 Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice between ages 12 and 16 wk (left) and 36 wk of age (right).C: ERK phosphorylation (p) is reduced in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice between 12 and 16 wk of age, whereas no difference is observed between genotypes at 36 wk of age.NS, nonsignificant.D: no difference is observed in the level of pSMAD2/3 between genotypes at early or late time points.MMP-2 (E) and MMP-9 (F) protein expression is decreased at early time points in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice, whereas no difference in expression levels is observed in TAA tissue of animals at 36 wk of age.For quantitative RT-PCR (qRT-PCR) experiments: Fbn C1039G/ϩ , n ϭ 14, and Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ , n ϭ 10; for Western analysis: Fbn C1039G/ϩ , n ϭ 8, and Fbn C1039G/ϩ /␤arr2
In Fbn C1039G/ϩ mice, the rate of aortic dilation slowed over time from 0.11 Ϯ 0.01 mm/4 wk from age 8 to 16 wk to 0.06 Ϯ 0.02 mm/4 wk after 16 wk of age (P Ͻ 0.01; Fig. 2D).This finding is consistent with previous work, demonstrating a gradual decrease in the rate of aortic dilation in Fbn C1039G/ϩ mice (18).In contrast, the rate of aortic dilation in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ increased over time from 0.03 Ϯ 0.01 mm/4 wk from age 8 to 16 wk to 0.09 Ϯ 0.02 mm/4 wk after 16 wk of age (P Ͻ 0.05; Fig. 2D).We observed no significant difference in aortic root diameter between wildtype and ␤arr2 Ϫ/Ϫ mice at 8 wk of age, as measured by repeated-measures, two-way ANOVA with post hoc Bonferroni testing (Fig. 2F).
We performed Western blot analysis of TAA tissue at 36 wk of age to assess the level of MMP-2 and -9, pERK1/2, and pSMAD2/3 at a time point where aortic size was similar between the two genotypes.Interestingly, no significant differences were observed in the expression levels of MMP-2 or -9 nor were differences observed in the levels of pERK1/2 or pSMAD2/3 at this later time point (Fig. 3,  B-F).
ANG stimulates increased MMP-2 and -9 gene and protein expression in primary aortic root VSMCs.VSMC-derived MMP-2 and -9 have been identified previously to be upregulated in aortic samples from MFS patients, as well as in aortic samples in a murine model of MFS (5,29).Work from our lab (7) has previously demonstrated that ␤arr2 does not interact with or augment signaling via the type I or II TGF-␤ receptors upon TGF-␤ stimulation.Therefore, we hypothesized that ␤arr2 might exert its effects on TAA formation in MFS via regulation of AT 1a R signaling.To assess the direct effect of AT 1a R-mediated signaling on the expression of the proaneurysmal genes MMP-2 and -9, we performed qRT-PCR on primary aortic VSMCs stimulated with ANG for 24 h.ANG stimulates a 3.1-fold increase in MMP-2 gene expression (P Ͻ 0.001) and a 2.0-fold increase in MMP-9 gene expression (P Ͻ 0.05; Fig. 4, A and B), with no significant stimulation of SMAD2 or col 1a1 gene expression in response to ANG treatment (Fig. 4A).ANG-stimulated increases in MMP-2 (P Ͻ 0.01) and MMP-9 (P Ͻ 0.05) gene expression are inhibited by blockade of the AT 1a R with losartan (Fig. 4B).By Western blot analysis, we found that ANG stimulates a 1.7-fold increase in MMP-2 protein expression (P Ͻ 0.005; Fig. 4C) and a 1.6-fold increase in MMP-9 protein expression (P Ͻ 0.001; Fig. 4D) after 24 h of stimulation.

DISCUSSION
Here, we demonstrate that ␤arr2 contributes to TAA formation in a murine model of MFS.In particular, we observed that Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice display delayed aortic root dilation relative to Fbn C1039G/ϩ mice.We found that gene and protein expression of MMP-2 and -9 is markedly reduced in TAA tissue obtained between 12 and 16 wk of age from Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ .We hypothesized that ␤arr2 mediates its affects on TAA development via regulation of AT 1a R signaling in VSMCs.We show that ANG-induced MMP-2 and -9 expression requires ␤arr2 in primary aortic root VSMCs.This pathway is mediated by both ERK1/2 activation as well as EGFR transactivation.We also observed that levels of MMP-2 and -9 and pERK1/2 varied over time in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice.This finding suggests that the signaling pathways that contribute to aortic root dilation in MFS may vary over time.This idea is supported by previous work demonstrating that microRNA-29b contributes to aortic root dilation only at early time points in a murine model of MFS (27).The recognition of the influence of this signaling pathway identifies a contributory signaling mechanism to TAA pathogenesis in MFS that may potentially serve as a target for pharmacologic inhibition.
Our study is corroborated by recent work demonstrating that AT 1a R-dependent, ␤arr2-mediated signaling contributes to abdominal aortic aneurysm (AAA) formation in an ANG-infusion model of AAA (33).Although this mouse model is  distinct from the model evaluated in this study, the authors showed that ANG-induced AAA formation and MMP-2 and -9 expression in the abdominal aorta are attenuated by genetic ␤arr2 Ϫ/Ϫ .We have identified a similar ␤arr2-dependent sig- naling mechanism present in the thoracic aorta.In particular, we demonstrate that ␤arr2-dependent MMP-2 and -9 expression requires both ERK1/2 and EGFR activation.We propose that this pathway contributes to TAA formation via regulation of the expression of proaneurysmal mediators MMP-2 and -9.Beyond TAA development, AT 1a R-mediated, ␤arr2-dependent signaling in MFS has also been shown to contribute to the development of a primary dilated cardiomyopathy in a murine model of MFS (10), providing further corroborative evidence for a deleterious role of AT 1a R-mediated, ␤arr2-dependent signaling in MFS.
Habashi et al. ( 18) have previously proposed a mechanism by which ANG contributes to TAA formation in MFS through three distinct pathways: 1) via upregulation of TGF-␤ ligands and receptors; 2) by upregulation of TGF-␤ signaling, directly as well as through upregulation of thrombospondin-1, a potentiator of TGF-␤ signaling; and 3) by directly stimulating VSMC proliferation, decreasing apoptosis, increasing profibrotic signaling, and upregulating MMP-2 and -9 (25).We propose that this third mechanism of ANG contribution to TAA development in MFS is ␤arr2 dependent (Fig. 10).Previous work from our lab (21,22) has demonstrated that ␤arr2 mediates ANG-induced VSMC proliferation, as well as ANG-induced antiapoptotic signaling through regulation of Bcl-2-associated death promoter phosphorylation (1).In addition, recent work suggests that ␤arr2 mediates proaneurysmal signaling and MMP-2 and -9 expression in abdominal aortic tissue (33).
Importantly, this work underscores the significance of ERKdependent signaling in TAA development in MFS.ERKdependent signaling has been demonstrated previously to contribute to aortic dilation in MFS (18,20).ANG stimulates ERK1/2 activation via the AT 1a R and both G q proteins as well as ␤arr2.ERK activated via these different transducers is both spatially and temporally distinct (3) with unique functional outcomes (22,34).Whereas ␤arr2-dependent ERK activation appears to lead to TGF-␤-independent, proaneurysmal signaling, ANG-stimulated activation of TGF-␤ signaling has been reported previously to involve G q proteins (35).Interestingly, G protein-and ␤arr2-dependent ERK1/2 activation has been shown to require EGFR transactivation in VSMC (21), suggesting the EGFR could serve as a mediator of AT 1a R-mediated pathogenic signaling in MFS.This hypothesis is supported by our preliminary work demonstrating a reduction in aortic dilation in Fbn C1039G/ϩ mice treated with the EGFR inhibitor erlotinib (Fig. 7).
There are several limitations of this study.First, whereas we have identified a role for ␤arr2 in the pathogenesis of TAA in MFS, the complex interplay of intracellular signaling events and ECM deficiencies in MFS makes in vitro recapitulation in cell culture systems of in vivo signaling events difficult.We demonstrate that pERK and MMP-2 and -9 are reduced significantly in aneurysmal tissue from Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice relative to Fbn C1039G/ϩ mice.ERK has been established previously as a key mediator of TAA formation in this model of MFS (18,20).We used primary aortic root VSMCs and demonstrated that ␤arr2 mediates ANG-stimulated ERK activation and MMP-2 and -9 production in VSMCs from the aortic root.As noted previously, MMP-2 and -9 production has been shown to be elevated in VSMCs from patients with MFS (29), as well as in a murine model of MFS (5).Therefore, we believe that our in vitro work represents a novel and relevant finding with regard to the signaling pathways involved in MMP-2 and -9 production in the aortic root.Nevertheless, it is possible that ␤arr2-mediated ERK activation and MMP-2 and -9 production in other cell types, such as endothelial cells, adventitial fibroblasts, or inflammatory cells, could contribute to TAA formation in MFS.This is currently the subject of ongoing investigation in our laboratory.
Additionally, it should be noted that our in vitro signaling work was performed in VSMCs harvested from rats as opposed to mice.We chose to use VSMCs harvested from rats due to a priori concerns regarding the number of animals and cells necessary for study.Because we planned to perform in vitro assays in VSMCs harvested as close to the aortic root as possible, we expected that VSMC yields from mice would be far too low for our needs and require a large number of animals (100 -200).Therefore, we had significant ethical and technical concerns regarding the use of murine VSMCs for this portion of our study.Notably, previous work in our lab (21) and others' (16) has demonstrated that AT 1a R-mediated ERK activation involves similar pathways in both murine and rat aortic VSMC, suggesting that rat VSMCs could serve as a viable model system in which to recapitulate the signaling pathways of interest in our study.
Second, ␤arr2-dependent effects appear to predominate at early time points in this murine model of MFS.This occurs concomitantly with a decrease in the levels of pERK and MMP-2 and -9 expression in the aortic root.Interestingly, over time, both aortic root diameter and MMP-2 and -9 expression in Fbn C1039G/ϩ /␤arr2 Ϫ/Ϫ mice return to levels similar to Fbn C1039G/ϩ mice.The precise mechanisms underlying this phenomenon are unclear; however, it suggests that the key contributors to TAA formation in MFS may differ over time and result in disparate rates of dilation.Again, this hypothesis is supported by previous work demonstrating a role for mi-croRNA-29b solely in early TAA development in a murine model of MFS (27).This time-dependent pattern of ␤arr2 effect is also the subject of ongoing investigation in our laboratory.
In conclusion, this work identifies a previously unrecognized signaling cascade that contributes to TAA formation in MFS.It also emphasizes the idea that TAA development in MFS is the end result of an interconnected network of signaling involving multiple receptor types and signaling mediators (Fig. 10), as well as ECM-dependent processes (18,19,25).In addition, it identifies ␤arr2 as a potential mediator of TAA formation and broadens our understanding of the complex signaling pathways involved in aortic aneurysmal disease.This may, in turn, lead to the development of more targeted therapeutics for the prevention and/or treatment of TAAs in patients with MFS.This work should be viewed as hypothesis generating, in that it identifies multiple, potential targets of inhibition, such as AT 1a R, TGF-␤, MMP-2 and -9, ERK1/2, EGFR, or ␤arr2.This is supported by previous work demonstrating the benefit of blockade of AT 1a R ( 19), TGF-␤ ( 19), MMP-2 and -9 (9, 36), and ERK1/2 (20) in murine models of MFS, as well as our own preliminary data demonstrating reduced aortic dilation in this model of MFS with pharmacologic inhibition of the EGFR (Fig. 8).Of particular interest is the recognition that pharmacologic agents currently in clinical use for other indications, such as inhibitors of EGFR and MMPs, may have use in the prevention or treatment of TAA formation in MFS.

Fig. 7 .
Fig. 7. ANG-stimulated MMP-2 and -9 gene and protein expression is dependent on the EGF receptor (EGFR) in primary aortic root VSMCs.A: ANG-stimulated gene expression of MMP-2 and -9 is blocked by EGFR inhibition with the EGFR inhibitor AG1478.B: ANG-stimulated protein expression of MMP-2 and -9 is also blocked by EGFR inhibition with the EGFR inhibitor AG1478.C: ANG-stimulated pERK1/2 at 15 min is completely blocked by inhibition of EGFR with AG1478.D: Western blot of the effect of EGFR inhibition with AG1478 of ANGstimulated MMP-2 and -9 production representative of 4 -8 individual experiments.**P Ͻ 0.01, ***P Ͻ 0.001.

Fig. 8 .
Fig.8.Pharmacologic inhibition of the EGFR blocks thoracic aortic dilation in a murine model of MFS.A: Fbn1 C1039G/ϩ mice were treated with subcutaneous infusion of either vehicle (left) or erlotinib (right) for 8 wk, beginning at 8 wk of age.Aortic root diameter was followed by transthoracic echocardiography at 8 and 16 wk of age.No significant aortic dilation was observed in Fbn1 C1039G/ϩ mice treated with erlotinib, whereas Fbn1 C1039G/ϩ mice treated with placebo displayed significant aortic dilation.NS, nonsignificant.B: the rate of change of aortic root diameter was significantly less in Fbn1 C1039G/ϩ mice treated with erlotinib (right) compared with vehicle-treated mice (left).*P ϭ Ͻ0.05, **P Ͻ 0.01 by t-test.

Table 2 .
Hemodynamic measurements of study animals ␤-ARRESTIN2 PROMOTES AORTIC DILATION IN MARFAN SYNDROME AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00291.2015• www.ajpheart.orgStatistical analysis.Data are expressed as means Ϯ SE.Statistical analysis was performed using Student's paired t-test, one-way ANOVA with Bonferroni multiple comparisons post hoc testing, two-way ANOVA, or repeated-measures two-way ANOVA with post hoc testing (GraphPad Prism; GraphPad Software, La Jolla, CA) where appropriate.P Ͻ 0.05 was considered statistically significant.Normality of the data was confirmed by D'Agostino-Pearson omnibus test or Shapiro-Wilk test, where appropriate.Outliers were identified and excluded by use of Grubbs' test.