Author + information
- Received April 16, 2015
- Revision received June 8, 2015
- Accepted June 12, 2015
- Published online August 25, 2015.
- David G. Guzzardi, BSc∗,
- Alex J. Barker, PhD†,
- Pim van Ooij, PhD†,‡,
- S. Chris Malaisrie, MD§,
- Jyothy J. Puthumana, MD‖,
- Darrell D. Belke, PhD∗,
- Holly E.M. Mewhort, MD∗,
- Daniyil A. Svystonyuk, BSc∗,
- Sean Kang, BSc∗,
- Subodh Verma, MD, PhD¶,
- Jeremy Collins, MD†,
- James Carr, MD†,
- Robert O. Bonow, MD‖,
- Michael Markl, PhD†,#,
- James D. Thomas, MD‖,
- Patrick M. McCarthy, MD§,# and
- Paul W.M. Fedak, MD, PhD∗,§∗ ()
- ∗Department of Cardiac Sciences, Libin Cardiovascular Institute of Alberta, University of Calgary, Calgary, Canada
- †Department of Radiology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois
- ‡Department of Radiology, Academic Medical Center, Amsterdam, the Netherlands
- §Division of Cardiac Surgery, Department of Surgery, Bluhm Cardiovascular Institute, Northwestern University, Chicago, Illinois
- ‖Division of Cardiology, Department of Medicine, Bluhm Cardiovascular Institute, Northwestern University, Chicago, Illinois
- ¶Division of Cardiac Surgery, Li Ka Shing Knowledge Institute of St. Michael’s Hospital, University of Toronto, Toronto, Canada
- #Department of Biomedical Engineering, McCormick School of Engineering, Northwestern University, Chicago, Illinois
- ↵∗Reprint requests and correspondence:
Dr. Paul W.M. Fedak, Department of Cardiac Sciences, University of Calgary, Room 880, 1403-29 Street NW, Calgary, Alberta T2N 2T9, Canada.
Background Suspected genetic causes for extracellular matrix (ECM) dysregulation in the ascending aorta in patients with bicuspid aortic valves (BAV) have influenced strategies and thresholds for surgical resection of BAV aortopathy. Using 4-dimensional (4D) flow cardiac magnetic resonance imaging (CMR), we have documented increased regional wall shear stress (WSS) in the ascending aorta of BAV patients.
Objectives This study assessed the relationship between WSS and regional aortic tissue remodeling in BAV patients to determine the influence of regional WSS on the expression of ECM dysregulation.
Methods BAV patients (n = 20) undergoing ascending aortic resection underwent pre-operative 4D flow CMR to regionally map WSS. Paired aortic wall samples (i.e., within-patient samples obtained from regions of elevated and normal WSS) were collected and compared for medial elastin degeneration by histology and ECM regulation by protein expression.
Results Regions of increased WSS showed greater medial elastin degradation compared to adjacent areas with normal WSS: decreased total elastin (p = 0.01) with thinner fibers (p = 0.00007) that were farther apart (p = 0.001). Multiplex protein analyses of ECM regulatory molecules revealed an increase in transforming growth factor β-1 (p = 0.04), matrix metalloproteinase (MMP)-1 (p = 0.03), MMP-2 (p = 0.06), MMP-3 (p = 0.02), and tissue inhibitor of metalloproteinase-1 (p = 0.04) in elevated WSS regions, indicating ECM dysregulation in regions of high WSS.
Conclusions Regions of increased WSS correspond with ECM dysregulation and elastic fiber degeneration in the ascending aorta of BAV patients, implicating valve-related hemodynamics as a contributing factor in the development of aortopathy. Further study to validate the use of 4D flow CMR as a noninvasive biomarker of disease progression and its ability to individualize resection strategies is warranted.
Bicuspid aortic valves (BAVs) are associated with an increased predisposition towards dilation of the ascending aorta that could increase the rates of aortic complications such as aortic dissection, rupture, and/or sudden death (1,2). Although several dilation patterns have been proposed (2), considerable debate remains as to whether they are due to an inherent aortic wall defect (i.e., genetic aortopathy) or are secondary to valve-related changes in regional hemodynamics and shear stress (i.e., acquired etiology). A genetic etiology for BAV aortopathy is widely accepted and may prompt aggressive resection strategies to remove diseased tissues at risk of future complications (3,4). Increasingly, valve-related hemodynamics are believed to contribute to disease progression (5). Greater understanding of the pathophysiology of BAV aortopathy may facilitate improved surgical resection strategies, development of best practices and effective clinical guidelines, and in so doing, optimize clinical outcomes (6).
Flow-sensitive cardiac magnetic resonance imaging (CMR) with full volumetric coverage of the ascending aorta (4-dimensional [4D] flow CMR) can measure and visualize complex aortic 3-dimensional (3D) blood flow patterns, such as flow jets, vortices, and helical flow. Using 4D flow CMR, we previously observed that normally functioning BAVs are associated with disturbed flow patterns in the ascending aorta, with regional increases in wall shear stress (WSS) (7), a parameter known to be associated with vessel wall remodeling (8). We further established that the location of BAV cusp fusion is associated with different patterns of ascending aorta dilation (9). Recent studies have provided significant associative evidence that the pattern of cusp fusion corresponds with the expression of aortopathy (10,11), thus aligning with previous imaging findings implicating altered outflow patterns and the regional expression of elevated WSS with BAV morphology. To further investigate these findings, in this study, we measured aortic WSS by 4D flow CMR in healthy normal volunteers and BAV patients to detect nonphysiological values, and for the first time, correlated valve-related changes in WSS to regional tissue architecture and remodeling in paired BAV aortic wall tissue samples.
With internal review board approval and informed consent, 20 BAV patients referred for ascending aortic surgery were enrolled. Patients with previous ascending aortic surgery or evidence of other forms of connective tissue disease were excluded. Healthy age-matched controls (n = 10) with tricuspid aortic valves were enrolled to compute regionally resolved 95% confidence interval values for physiologically normal aortic WSS (12); these controls had no evidence of cardiovascular disease and did not undergo surgery. The degree of aortic stenosis was graded based on absolute systolic peak velocity by continuous-wave Doppler ultrasound (mild: 2 to 3 m/s; moderate/severe: ≥3 m/s), and aortic regurgitation was graded based on regurgitant fraction (mild: <30%; moderate/severe: ≥30%) (13).
Participants received pre-operative CMR at 1.5-T or 3-T (Magnetom Aera, Espree, Avanto, Skyra, Siemens Healthcare, Erlangen, Germany) to assess presence and significance of suspected BAV. 4D flow CMR provided complete volumetric coverage of the thoracic aorta for quantification of temporally resolved 3D blood flow velocities. Data were acquired during free breathing using respiratory and prospective electrocardiographic gating (14), with imaging parameters as described previously (12). Velocity encoding ranged from 150 to 400 cm/s based on the severity of valve stenosis. If the glomerular filtration rate was >30 ml/min, gadopentetate dimeglumine, gadofosveset trisodium, or gadobenate dimeglumine was administered intravenously, and the flip angle was set to 15°; otherwise, 7° was used. Patient-specific WSS heat maps of the BAV aorta were computed relative to a map of the population average for healthy age-matched controls as described in detail previously (12,15,16). WSS regions outside the healthy 95% confidence intervals were classified as abnormal. Intra-aortic regions of normal, depressed, and elevated WSS were mapped onto 3D visualizations of patient-specific aortas (Figure 1).
Aortic wall samples were collected as permitted by the extent of ascending aorta resection; surgeons were blinded to WSS heat maps. Samples were labeled according to pre-operative zonal designations relative to the position of the right pulmonary artery (zones 1, 2, and 3 correspond to regions proximal, adjacent, and distal to the right pulmonary artery, respectively) (Figure 1), and according to circumferential location (greater curvature, lesser curvature, anterior or posterior wall). Tissue samples were divided in 2 for histology and protein analysis, flash-frozen in optimal cutting temperature freezing compound (VWR International, Radnor, Pennsylvania) and liquid nitrogen within 15 min of resection, and then stored at −80°C until use. Paired samples of each patient’s aorta from regions of normal and elevated WSS were compared.
Tissue designated for histology was thawed, sectioned circumferentially, and then fixed in 10% buffered neutral formalin (VWR International) for a maximum of 1 week and paraffin-embedded. Sections of 6 μm were mounted for Verhoeff-Van Gieson staining of medial elastin fibers. Chromatic analysis of elastin abundance was performed (Aperio ePathology, Leica Biosystems, Buffalo Grove, Illinois). Total elastin content was computed as the mean percent area of elastin staining relative to the total area analyzed of 6 representative fields of view (40× magnification) as selected by blinded observer. Distance between intact elastin fibers and their thicknesses were measured (ImageScope Viewing Software, Leica Biosystems) by blinded observer with a mean value computed for each sample taken from an average of 131 measurements per circumferentially sectioned slide (as described by Bauer et al. ).
Tissue designated for protein quantification was homogenized in 20 mmol/l Tris-HCl buffer (pH 7.5, 0.5% Tween 20, 150 mmol/l NaCl, and Roche complete protease inhibitor 1:100) and centrifuged at 10,000 g for 10 min at 4°C. Supernatant containing 1.0 μg of protein was assayed in duplicate for concentration of transforming growth factor (TGF)-β-1, matrix metalloproteinases (MMPs), and tissue inhibitors of MMPs (TIMP) using the multiplex fluorescent bead assay (Eve Technologies, Calgary, Alberta, Canada).
All group data are presented as mean ± SD if normally distributed or as median (interquartile range [IQR]) if non-normally distributed. Normality was assessed using the Shapiro-Wilk test. Normally and non-normally distributed data were compared using a paired Student t test or the Wilcoxon signed rank test, respectively (both 2-tailed). Cohort-averages between 2 groups were compared using the Student t test when normally distributed or the Mann-Whitney test when non-normally distributed. Statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software, La Jolla, California), with p < 0.05 considered statistically significant.
Patient demographics are summarized in Table 1. Healthy controls were used to define physiologically normal WSS values and were age-matched to BAV patients (50 ± 14 years vs. 48 ± 15 years, respectively; p = 0.83). BAV patients were predominantly male with type 1 fusion of the right and left coronary cusps. Surgery was primarily indicated for BAV dysfunction; all patients had moderate/severe stenosis or regurgitation. Mean aortic diameter was 4.4 ± 0.5 cm at the sinus of Valsalva and 4.7 ± 0.6 cm for the ascending aorta. Most patients underwent aortic valve replacement (1 patient underwent BAV repair). All patients had ascending aortic resection, with most requiring concomitant root replacement and a few undergoing hemiarch resection using deep hypothermic circulatory arrest. Tissue corresponding to regions of normal and elevated WSS were primarily collected from zone 2 of the BAV aorta and from the anterior wall and greater curvature, respectively. Samples collected from patients are summarized in Online Table.
We quantified the total elastin content and architecture by histology image analysis. Aortic wall from regions of normal and elevated WSS demonstrated significantly decreased elastin content and architecture consistent with aortopathy (Figures 1A to 1C). Aortic wall exposed to elevated WSS showed reduced elastin abundance compared with aortic wall subjected to normal WSS within the same aorta (p = 0.01) (Figure 2A). Similarly, the cohort-averaged percent area of elastin was significantly decreased in regions of elevated WSS among BAV aortas (36.61 ± 16.87% vs. 49.12 ± 16.53%; p = 0.04) (Figure 2B). Compared with aortic wall subjected to normal WSS, regions of elevated WSS had elastin fibers that were significantly thinner among patient pairs (p = 0.00007) (Figure 2C). The absolute mean thickness of elastin fibers also was significantly decreased in regions of elevated WSS compared with regions of normal WSS (2.72 ± 0.40 μm vs. 3.25 ± 0.27 μm; p = 0.00002) (Figure 2D). The median distance between elastin fibers was significantly greater in regions of elevated WSS compared to normal WSS (p = 0.001) (Figure 2E). Similarly, cohort-averaged median distance between elastin fibers was significantly increased in aortic wall subjected to elevated WSS (28.34 μm [IQR: 22.03 to 36.32 μm] vs. 20.39 μm [IQR: 18.35 to 24.43 μm]; p = 0.01) (Figure 2F).
Regions of elevated WSS had significantly higher concentrations of TGFβ-1 protein (p = 0.04) (Figure 3) compared with paired regions of normal WSS in the same aortas.
Aortic wall subjected to normal and elevated WSS within each patient’s aorta were profiled for MMP and TIMP protein concentrations (Table 2). The absolute levels of MMP and TIMP protein expression in the samples were highly variable. MMP-2 and TIMP-2 were the most highly expressed MMP and TIMP in the aorta. Compared with regions of normal WSS, aortic wall exposed to elevated WSS demonstrated significantly increased relative concentrations of MMP-1 (p = 0.03), MMP-3 (p = 0.02), and TIMP-1 (p = 0.04), and there was a trend (p = 0.06) for increased concentrations of MMP-2.
BAV is an inheritable disorder, and a genetic theory for the associated aortopathy is widely held, positing that the aorta has an inherent genetic weakness and is prone to dilation and rupture from an underlying dysregulation of extracellular matrix (ECM) in the aortic medial layer (1,5). This perspective has encouraged more aggressive approaches towards aortic resection using strategies similar to those applied in patients with proven genetic aortopathies, such as Marfan syndrome. Recently, we and others have shown that valve-related hemodynamics may play an important role in disease progression, where altered flow in the aorta is consistent with patterns of aortic dilation (9–11,18,19). Recognizing that hemodynamics may be altered in different regions of the ascending aorta, regional differences have been documented in key ECM proteins and cell phenotypic expressions, focused primarily on the convexity versus the concavity of the BAV aorta (20,21). These translational tissue studies suggest that valve-related hemodynamics may influence disease progression.
There is mounting evidence that both theories may coexist (5,22). Valve-related hemodynamics may exacerbate disease progression in genetically susceptible aortas (6). Given the absence of data showing a mechanistic link between cusp fusion patterns, altered aortic flow, and expression of disease, considerable debate surrounds the role of hemodynamics in mediating BAV aortopathy (5). Recently, using 4D flow CMR, we observed that even normally functioning BAVs are associated with disturbed ascending aortic flow and regional WSS increases, beyond the hemodynamic derangements accounted for by measures of stenosis and regurgitation (7). We further established that the location of BAV cusp fusion was associated with different patterns of ascending aortic dilation (9), suggesting that valve-related hemodynamics may influence the expression of BAV aortopathy.
Despite well-documented changes of WSS in BAV patients, its role in the underlying aortopathy is unclear. In the current study, distinct regions of increased WSS were uniformly observed in BAV patients despite varied cusp fusion patterns and types or degrees of valve dysfunction. The significant relationship between WSS derived from 4D flow CMR and regional aortic tissue remodeling provides evidence, for the first time, of the role of valve-related hemodynamics in BAV aortopathy (Central Illustration).
WSS corresponds to changes related to aortopathy
To validate the influence of regional WSS on tissue remodeling, we compared the medial matrix architecture between areas of high and normal WSS within each BAV aorta. This analysis is advantageous because each patient serves as his or her own control. Degeneration of the aortic media, particularly its elastic laminae, is the sine qua non of BAV aortopathy, with a hallmark of elastic fiber fragmentation (23). Bauer et al. (17) also observed medial elastin degeneration characteristic of BAV aortopathy, documenting thinner elastic laminae and increased distances between laminae. In patients with BAV stenosis undergoing aortic valve replacement, Girdauskas et al. (24) documented qualitative differences in aortic histology at the level of aortotomy with different systolic transvalvular flow patterns.
We observed disrupted medial elastin fiber architecture in the aortic wall of our patient cohort, consistent with BAV aortopathy, coupled with the novel finding that elastin fiber fragmentation and architectural derangement was significantly increased in areas of high versus normal WSS within the same aortas. Areas of normal WSS also had evidence of medial matrix degradation consistent with BAV aortopathy, although less severe than corresponding areas of high WSS. Given that histological abnormalities are diffuse within the BAV aorta and our study population had relatively mild levels of aortic dilation (most <5.0 cm), the increased elastin degradation observed in the areas of high WSS relative to adjacent areas with normal WSS is a striking observation that provides strong evidence implicating valve-related hemodynamics in medial matrix degradation in BAV aortopathy.
TGFβ is strongly implicated in the mechanotransduction of WSS upstream of flow-induced vascular remodeling (25). In animal models of human aortopathy, TGFβ signaling is implicated in vascular disease and progression as a critical mediator of aortopathy (26). Forte and colleagues (20) showed that TGFβ and TGFβ receptor-2 are increased in BAV aorta, suggesting a role for TGFβ in mediating disease. In this study, we provide novel data linking aortic TGFβ protein expression to local increases in WSS, with ascending aortic TGFβ concentration significantly elevated in regions of high aortic WSS compared with adjacent regions with normal WSS. These data strongly implicate WSS as contributing to TGFβ expression in the BAV aorta.
MMPs can directly degrade elastic ECM components. Supporting a mechanistic role for valve-related hemodynamics in MMP expression, Ikonomidis et al. (27) showed altered MMP profiles stratified by cusp fusion pattern in the BAV aorta. Although many MMPs are expressed in the aorta, MMP-2 is highly implicated in aortic aneurysms and elastic matrix degeneration of remodeled arteries (28). Animal models provide proof of concept for WSS as a trigger for aortic wall MMP activation. For example, flow-mediated induction of MMP-2 expression was demonstrated in rabbit carotid arteries in association with medial matrix degeneration and vessel dilation (29). Furthermore, Atkins et al. (30) modeled BAV-related elevated WSS in an ex vivo porcine aorta system, and found increased MMP-2 expression and activity. We previously reported evidence of MMP-2 elevations in the BAV aorta (31). The cumulative evidence by meta-analysis of all such studies confirms that MMP-2 is consistently elevated in human BAV aortopathy (32).
In the current study, we observed that aortic wall MMP-2 was the most highly expressed MMP in the BAV aorta, with a trend (p = 0.06) for greater expression in areas of high WSS as compared to adjacent areas with normal WSS in the same patients, and with significant increases in MMP-3 in the same areas. Importantly, the elevated levels of MMP-2, MMP-3, and TGFβ areas of increased WSS were associated with fragmented medial elastin fibers. This observation also strongly implicates WSS in mediating BAV aortic disease progression. In addition to a direct role in degrading elastic tissue, it is noteworthy that MMP-2 and MMP-3 can activate latent TGFβ and increase its activity. These factors may work synergistically to induce medial matrix remodeling.
The observed TIMP changes are complex, and the substantial variability between patients and regions studied suggests that the regulatory control of TIMP expression is less influenced by WSS than specific MMPs and TGFβ. TIMP-1 levels were increased in areas of high WSS, but no significant difference was observed between groups with respect to other TIMP species. The lack of increased compensatory TIMP-2 expression in areas of high WSS may result in heightened MMP-2 activity. Further study is required to better define and understand the influence of specific MMPs and TIMPs in BAV aortopathy and their expression and activities with respect to valve-related hemodynamics.
In our study, 4D flow CMR was used as a research tool to obtain WSS maps in individual patient aortas to better understand whether blood flow patterns play a role in the expression of BAV aortopathy. Given the spatial and temporal resolution of 4D flow CMR, underestimation of WSS is known to occur (7); however, relative WSS values (i.e., high/low WSS) are retained between subjects when consistent imaging parameters are employed. Here, we used similar parameters to those of a prior study that demonstrated the robustness of WSS measurements (33). Additionally, same-patient samples ensured that the relative WSS measurements were obtained in the same subject and same exam. Therefore, resolution was not deemed a significant factor for the study protocol to detect elevated WSS. Although the concentrations of specific proteins of interest were examined, further studies should investigate TGFβ and MMP-2 activities and other associated factors such as shifts in cell phenotype that may also mediate matrix remodeling. Additional comparison to patients with tricuspid valves and dilated aortas with areas of increased WSS should also be explored.
Regions of increased WSS correspond with ECM dysregulation and elastic fiber degeneration in the ascending aorta of BAV patients, implicating valve-related hemodynamics as a mediator of aortopathy. WSS as assessed by 4D flow CMR may serve as a noninvasive biomarker of regional aortic disease in patients with BAV aortopathy (deemed sufficiently severe to warrant operative intervention). Its utility in prediction of disease progression, particularly in earlier and less severe BAV aortopathy, awaits careful study, as does the efficacy of a targeted surgical approach incorporating regional aortic WSS.
COMPETENCY IN MEDICAL KNOWLEDGE: Measurements made using 4D flow CMR suggest that valve-related hemodynamics contribute to the development of aortopathy in patients with BAV.
TRANSLATIONAL OUTLOOK: Validation studies are needed to confirm the utility of 4D flow CMR as a noninvasive marker of disease progression, predictor of aortic dissection or rupture, and guide to the extent of surgical resection based on detailed measurements of regional aortic wall shear stress.
The authors gratefully acknowledge the recruitment and organizational efforts of Colleen Clennon (Northwestern University), the staff of Northwestern’s Center for Translational Imaging, Drs. Amy Bromley (University of Calgary) and Yong-Xiang Chen (University of Calgary) for their histological expertise, and Thomas Kryton for histological database management (University of Calgary).
For a supplemental table, please see the online version of this article.
Funded by Melman Bicuspid Aortic Valve Program, Bluhm Cardiovascular Institute (Dr. Fedak), American Heart Association grant 14POST20460151 (Dr. van Ooij), National Institutes of Health (NIH) grant K25HL119608 (Dr. Barker), and NIH grant R01HL115828 (Dr. Markl). Dr. Carr has a research agreement with Siemens. Dr. Thomas is a consultant for Abbott; and has received honoraria from Edwards Lifesciences, GE, and Abbott. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bicuspid aortic valve
- cardiac magnetic resonance imaging
- extracellular matrix
- interquartile range
- matrix metalloproteinase
- transforming growth factor
- tissue inhibitor of matrix metalloproteinase
- wall shear stress
- Received April 16, 2015.
- Revision received June 8, 2015.
- Accepted June 12, 2015.
- American College of Cardiology Foundation
- Fedak P.W.,
- Verma S.,
- David T.E.,
- et al.
- Girdauskas E.,
- Borger M.A.,
- Secknus M.A.,
- et al.
- Barker A.J.,
- Markl M.,
- Burk J.,
- et al.
- Mahadevia R.,
- Barker A.J.,
- Schnell S.,
- et al.
- Kang J.W.,
- Song H.G.,
- Yang D.H.,
- et al.
- Della Corte A.,
- Bancone C.,
- Dialetto G.,
- et al.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Bissell M.M.,
- Hess A.T.,
- Biasiolli L.,
- et al.
- Della Corte A.
- Girdauskas E.,
- Rouman M.,
- Disha K.,
- et al.
- Holm T.M.,
- Habashi J.P.,
- Doyle J.J.,
- et al.
- Chung A.W.,
- Au Yeung K.,
- Sandor G.G.,
- et al.
- Atkins S.K.,
- Cao K.,
- Rajamannan N.M.,
- Sucosky P.