Thoracic Aortic Aneurysm Development Is Dependent on Membrane Type-1 Matrix Metalloproteinase Activity and Abundance

TL;DR

This study shows that membrane type-1 matrix metalloproteinase (MT1-MMP) from aortic fibroblasts plays a key role in the development of thoracic aortic aneurysms.

Contribution

The study demonstrates that fibroblast-derived MT1-MMP contributes to TAA development via TGF-β signaling.

Findings

MT1-MMP deficiency reduced aortic dilatation and collagen changes in TAA models.

Fibroblast-specific MT1-MMP knockout attenuated TAA-induced aortic changes.

MT1-MMP abundance correlates with TGF-β activation in aortic fibroblasts.

Abstract

Thoracic aortic aneurysm (TAA) results from dysregulated remodeling of the extracellular matrix mediated by matrix metalloproteinase (MMP) activity. Previous studies identified elevated membrane type-1 MMP (MT1-MMP) abundance and activity during TAA development and suggested aortic fibroblasts as a potential key source. Herein, we extended our understanding of the role of MT1-MMP during TAA development using various MT1-MMP transgenic mouse strains. MT1-MMP deficient (MT1-MMP+/−) mice exhibited reduced MT1-MMP abundance, activity, and collagen volume fraction following TAA induction, concomitant with reduced aortic dilatation. TAA tissue from wild-type and MT1-MMP+/− mice showed a similar reduction in thin collagen fibers, while the MT1-MMP+/− mice displayed no change in thick collagen fibers. The role of fibroblast-derived MT1-MMP was examined using a conditional fibroblast-specific tamoxifen-inducible MT1-MMP knockout strain (FbMT1KO). TAA-induced changes in aortic diameter and MT1-MMP abundance were attenuated in FbMT1KO mice. Using aortic fibroblasts isolated from multiple mouse strains expressing different levels of MT1-MMP, a significant correlation between MT1-MMP abundance and TGF-β activation was observed. Importantly, treatment with MT1-MMP activity-neutralizing antibody or TGF-β neutralizing antibody resulted in the attenuation aortic dilatation. Together, these findings suggest that fibroblast-derived MT1-MMP is required for TAA development, in part through its ability to induce TGF-β signaling.

Figures

Full text

1. Introduction

Thoracic aortic aneurysm (TAA) is characterized as an abnormal dilatation of the thoracic aorta mediated through dynamic remodeling of the aortic extracellular matrix (ECM), leading to a weakened vessel wall and an increased susceptibility to dissection or rupture [1]. The underlying mechanisms driving this process remain elusive and likely comprise a complex interplay between susceptibility genetics and exposomics, accounting for multiple environmental influences (physical, chemical, biological, and social). TAAs carry significant morbidity and mortality [2,3], but therapy other than surgical or endovascular intervention has been largely ineffective or focused solely on managing risk factors and not aimed at the underlying mechanisms driving the disease [4,5,6,7,8]. Therefore, understanding the molecular mechanisms underlying TAA formation and progression is critical for developing effective non-invasive therapeutic strategies for this disease.

TAA development is accompanied by disruption of elastin filaments and disordered collagen deposition [1], which strongly associates with increased abundance and activity of ECM proteolytic systems, particularly matrix metalloproteinases (MMPs) [9]. Several studies involving human aneurysm specimens and murine aneurysm models have revealed that increased abundance and activity of the membrane type-1 MMP (MT1-MMP) correlates with aortic dilation, suggesting that MT1-MMP may play an important role in aneurysm formation [10,11,12,13]. MT1-MMP is speculated to contribute to TAA development in multiple ways. First, MT1-MMP can directly degrade ECM structural proteins and modulate the aortic pericellular microenvironment due to its gelatinolytic and collagenolytic activity [14,15]. Moreover, MT1-MMP plays an indispensable role in the activation (removal of pro-domain) of MMP-2 (mediated by an MT1-MMP/pro-MMP-2/TIMP-2 complex [15,16]), MMP-13, and, indirectly, MMP-9 [17,18,19]. Mice deficient in MMP-2 or TIMP-2 have shown resistance in abdominal aneurysm growth [20,21], suggesting MT1-MMP-dependent activation of MMP-2 is essential during aneurysm development. Lastly, MT1-MMP is capable of releasing/activating latent ECM-bound growth factors that in-turn activate cognate signaling mechanisms, such as the transforming growth factor-β (TGF-β) pathway [17,22,23,24,25,26].

Previous studies from this laboratory have demonstrated the elevated abundance of MT1-MMP in a calcium chloride-induced murine TAA model and demonstrated that MT1-MMP colocalized with cell-specific markers for aortic fibroblasts within the thoracic aortic wall [27]. Importantly, this past study identified that the time-dependent change in aortic diameter was directly associated with increased MT1-MMP activity, suggesting that MT1-MMP abundance and activity was essential for aortic dilation and TAA development. Accordingly, the present study was designed to further examine the causal effect of MT1-MMP in TAA development, by testing the hypothesis that loss of MT1-MMP/TGF-β pathway activation, especially in aortic fibroblasts, prevents TAA development.

2. Material and Methods

2.1. Transgenic Animals

MT1-MMP Knockout Mice. Heterozygous offspring of MT1-MMP total-body knockout mice (MT1-MMP^+/−^), originally described by Zhou et al. [28], and their wild-type littermates (FVB background), were used to examine the role of MT1-MMP in TAA development. The MMT1-MMP^/−^ mice have normal physiology, breed normally, and live for a normal length of time. This strain is regularly maintained in the heterozygous state as the homozygous total knockout offspring (MT1-MMP^−/−^) typically die within 21 days postpartum as a result of significant malformations [29]. As such, homozygous animals are unsuitable for TAA studies.

2.2. MT1-MMP Fibroblast-Specific Tamoxifen-Inducible Knockout Mice

To assess the role of MT1-MMP in aortic fibroblasts during TAA development, we generated a Cre-lox-based cell-specific conditional MT1-MMP knockout mouse strain (FbMT1KO) by crossing two transgenic strains: (1) a fibroblast-specific tamoxifen-inducible Cre-recombinase strain (Col1A2-Cre(ERT2)) and (2) a floxed MT1-MMP strain, both on a C57BL/6 background. The fibroblast-specific, tamoxifen-inducible, Cre-recombinase mice (Col1A2-Cre(ERT2)) were constructed by inserting a 6.4 kb fragment of the murine pro-alpha2(I) collagen (Col1A2) gene (containing the promoter/enhancer region described by Bou-Gharios et al. [30] and Denton et al. [31], which strongly directs transgene expression in fibroblasts) upstream of human β-globin intron 2, the Cre-ERT2 ORF, and an SV40-polyA sequence into a pGL3 vector (Promega, Madison, WI, USA) (Taconic-Artemis, Germantown, NY, USA) (Supplemental Figure S1). The ~10 kb transgene was microinjected into ES cells from C57BL/6NTac mice. At the same time, floxed-MT1-MMP (MT1-MMP^(fl/fl)^) mice were constructed on a C57BL/6NTac background by homologous recombination, inserting loxP sites flanking exons II–IV (Taconic-Artemis, Germantown, NY, USA) (Supplemental Figure S1B). These two strains were bred to develop the Col1A2-Cre(ERT2)^+^ × MT1-MMP^(fl/fl)^ mice (FbMT1KO) and the Col1A2-Cre(ERT2)^−^ × MT1-MMP^(fl/fl)^ mice (control). Upon treatment with tamoxifen (75 mg/kg/day intraperitoneally for 5 consecutive days), Cre-recombinase translocates to the nucleus in fibroblasts expressing the Col1A2-Cre(ERT2) construct, where Cre removes the intervening DNA between loxP sites. Loss of exons II–IV results in loss of MT1-MMP function and a frameshift yielding a premature stop codon in exon V. The resulting Col1A2-Cre(ERT2) × MT1-MMP^(fl/fl)^ mice and their derivatives were maintained in our local colony and genotypes were regularly confirmed by PCR.

2.3. FbMT1-MMP Overexpressing Mice

This transgenic strain was designed to enhance MT1-MMP expression in murine fibroblasts. It was constructed by inserting the 6.4 kb murine Col1A2 promoter/enhancer fragment, along with the full-length human MMP14 cDNA and an SV40-polyA sequence, into a pGL3 vector (Promega) (Taconic-Artemis, Germantown, NY, USA) (Supplemental Figure S1C). The vector was confirmed by sequencing and microinjected into wild-type FVB embryos. Integration-positive animals were identified by PCR and maintained in our local colony.

2.4. Animal Studies

All animals were treated and cared for in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice were housed in temperature- and humidity-controlled rooms equipped with automated timers to provide a 12:12 light cycle, in hanging solid bottom cages maintained in HEPA-filtered ventilated racks. Nesting material was provided as enrichment. Cages were changed at least weekly.

All mice were 8–14 weeks of age when inducted into an experimental protocol and comprised both sexes (~50% distribution). Sample sizes for all in vivo studies were determined a priori using the Stata statistical software package v.12 (StataCorp LLC, College Station, TX, USA). The primary outcome variable measured in vivo was the aortic diameter. For studies using the heterozygous MT1-MMP^+/−^ mice and the FbMT1KO mice, sample sizes were determined based on past results observed for the wild-type groups (FVB or C57BL/6 controls) in response to TAA induction. For both experimental groups, we hypothesized a conservative 20% reduction in aortic diameter due the reduced abundance of MT1-MMP in the MT1-MMP^+/−^ or FbMT1KO animals at each time point compared to their respective controls. The variance in each group was also conservatively estimated to be 15%. Accordingly, for the Power Analysis, the risk of encountering a Type I error (a) was set to 0.05 and the power was set at 0.80 (defining the risk of Type II error at 0.20) using a 2-way ANOVA model with interactions. The results determined that a minimum sample size of 5 animals per group would be required. The graph representing the relationship between total sample size and power is shown in Supplementary Figure S5A. For the studies examining the effects of antibody treatment on TAA development a priori sample sizes were also estimated to be a minimum of 5 animals per group to reach statistical significance. Again, based on wild-type TAA development, a single time point post-TAA induction was examined following antibody treatment (MT1-MMP-InhAb or TGF-β-NAb) and change in aortic diameter from baseline was compared to vehicle controls. The risk of a Type I error (a) was set of 0.05 and the power was set to 0.80, and a one-sample mean comparison t-test model was utilized. The graph representing the relationship between total sample size and power for this study is shown in Supplementary Figure S5B. For all other biochemical and histological studies, samples sizes were determined based on the availability of tissue specimens. Mouse numbers in some experimental groups were increased to accommodate the need for downstream analyses.

For all surgical procedures, anesthesia was established with isoflurane. Pain was controlled with buprenorphine, which was administered prior to first incision, then post-operatively every 8 h up to 48 h. Any animal not recovering well or possessing any evidence of acute distress in the early post-operative period was immediately euthanized using methods consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association. While humane endpoints were pre-determined as acute hemodynamic stress or severe weight loss, all animals completed the experimental protocol as designed and none had exhibited any severe complications requiring removal from study or early euthanasia.

2.5. TAA Induction and Digital Micrometry

Murine TAAs were induced in approximately equal numbers of male and female, 10–14 weeks old, as previously described [27]. Briefly, mice were anesthetized (2% isoflurane), the descending thoracic aorta exposed via a left thoracotomy, and digital images of the aorta were obtained using a 3-megapixel color camera (PAXcam3, Midwest Information Systems, Villa Park, IL, USA) linked to a laptop computer running PAX-it image management software v.2.1 (Midwest Information Systems, Villa Park, IL, USA). A sponge soaked in 0.5 M CaCl_2_ was placed in direct contact with the periadventitial surface for 15 min. The chest was irrigated liberally with normal saline and closed in layers, and mice were then allowed to recover. The overall mortality rate was less than 20% and mainly due to pulmonary complications. No cases of aortic rupture were observed. Aortic diameter was measured in each mouse at baseline (time of TAA induction) and again at terminal surgery. A calibrated digital caliper was used to measure outer aortic diameter. Terminal aortic size in each animal was expressed as a percent change from its own respective baseline measurement.

Animals were randomized to be studied at either 2, 4, 8, or 16 weeks following TAA induction. At the designated terminal time point, the animals underwent thoracotomy, exposure of the descending thoracic aorta, and diameter measurement. The animals were euthanized with an overdose of inhaled isoflurane (5%) injection and exsanguination, consistent with AVMA guidelines on euthanasia.

2.6. Immunoblotting Analysis

Immunoblotting was performed as described previously [32]. Briefly, TAA extracts (10 μg of total protein) were loaded onto a 4–12% SDS-PAGE and subjected to electrophoretic separation. The separated proteins were then transferred to a nitrocellulose membrane. Membranes were probed using antibodies to MT1-MMP (1:2000 dilution; MilliporeSigma, Burlington, MA, USA), phospho Smad-2 (1:1000; Cell Signaling Technology, Danvers, MA, USA), and β-actin (1:1000; Cell Signaling Technology). Membranes were then washed and incubated in horseradish peroxidase-conjugated secondary antibody (1:5000, Cell Signaling Technology). Immunoreactive signals were detected by chemiluminescence (Western Lighting, Perkin Elmer, Waltham, MA, USA). Protein abundance in the aorta extracts was analyzed using densitometric methods and values were expressed as a percentage of the abundance in unoperated control aortic extracts.

2.7. Histology

Approximately 1 cm of descending thoracic aorta (centered around the region of TAA induction) from wild-type FVB mice and MT1-MMP^+/−^ heterozygous mice, at baseline (control), 4 weeks, or 8 weeks post-TAA induction were harvested and fixed in 10% formalin for 48 h, followed by storage at 4 °C in 70% ethanol. The fixed aortic segments were bisected through the TAA induction site. Both segments were embedded in paraffin on end (TAA site facing up) and 5 µm sections were collected and stained with Hematoxylin and Eosin (H&E) to visualize the general structure or picrosirius red (PSR) to identify collagen content. The stained aortic sections were visualized on a Zeiss Axioskop 2 microscope (Carl Zeiss MicroImaging, Thornwood, NY, USA) and digital images were acquired (Axiocam MRc camera and AxioVision Software, v4.8, Carl Zeiss MicroImaging, Thornwood, NY, USA). All quantitative image analysis was performed using SigmaScanPro v5.0 (SPSS, Chicago, IL, USA) and Adobe Photoshop CS5 (Adobe, San Jose, CA, USA). For each study, analysis of histology sections was performed using at least three randomly chosen high-power fields from three different sections of each mouse aorta. Aortic medial wall thickness was measured from the internal elastic lamina to the adventitial border. Collagen volume fraction was measured from picrosirius red-stained sections illuminated with polarized light, and total collagen, thin collagen fibers (yellow-green birefringence), and thick collagen fibers (orange-red birefringence) fibers were measured.

2.8. MT1-MMP Quenched Fluorogenic Peptide Activity Assay

MT1-MMP activity was directly measured in aorta homogenates from each mouse using a MT1-MMP-specific quenched fluorogenic substrate (MCA-Pro-Leu-Ala-Cys(p-OMeBz)-Trp-Ala-Arg(Dpa)-NH2; MMP-14 Substrate I, Cat #444258, MilliporeSigma, Burlington, MA, USA), as previously described [27]. Briefly, aortic tissue was harvested from wild-type and MT1-MMP^+/−^ mice at 2, 4, 8, and 16 weeks post-TAA induction, and compared to an unoperated control. Aortic homogenates (15 μg) were incubated (37 °C) in the presence and absence of the MT1-MMP-specific quenched fluorogenic substrate. Fluorescence data was collected continuously for 20 h (Fluorostar Galaxy, 328/400 nm ex/em; BMG Labtechnologies, Cary, NC, USA), compared to a standard curve using active recombinant MT1-MMP, and normalized to the amount of GAPDH present in each sample (determined by ELISA assay (Cat# 3401, Bioo Scientific, Austin, TX, USA). MT1-MMP activity was expressed as ng of cleaved substrate/mg GAPDH/hr.

2.9. Aortic Fibroblast Isolation

Aortic fibroblasts were isolated using an established outgrowth procedure, as previously described [33]. Briefly, the descending thoracic aorta was excised from wild-type FVB, FbMT1-MMP, MT1-MMP^+/−^, and MT1-MMP^−/−^ mice (in the absence of TAA), rinsed in sterile saline, and cut longitudinally. The endothelial cell layer was removed by gently rubbing the luminal surface with a sterile swab. The aorta was then cut into approximately ten 2 mm pieces and carefully placed onto the surface of a dry tissue culture flask (T-75, Cat#13-680-65; BD Falcon, Fisher Scientific, Pittsburgh, PA, USA). The tissue was allowed to adhere in the absence of medium for 5–10 min, then 10 mL of fibroblast growth medium (Fibroblast Growth Medium, Promocell Cat#C39315, Heidelberg, Germany), containing Fibroblast Growth Supplement, (Promocell Cat#C23010; consisting of 1.0 ng/mL basic fibroblast growth factor, and 5.0 μg/mL insulin at final concentration) and 10% heat-inactivated fetal-calf serum (Invitrogen, Cat# 10082-147, Carlsbad, CA, USA), was carefully added to the flask. The flask was placed in a humidified 5% CO_2_ incubator, and the fibroblasts were allowed to grow out of the individual tissue chunks. Evidence of fibroblast outgrowth was typically observed within 7 days of plating. Fibroblasts were identified as spindle-shaped cells during log phase growth, and their identity was confirmed by staining log phase cells with phalloidin to observe cellular architecture and with cell-type-specific markers to verify purity (DDR2, prolyl-4-hydroxylase, and heavy chain myosin). The established cell lines were maintained in culture and split into new flasks when the cell density reached approximately 85% confluence. Fibroblasts in passages 3–6 were used for experimental studies.

2.10. Antibody Injections

To assess the role of MT1-MMP activity and downstream signaling during TAA development, mice were injected with one of two inhibitory antibodies: (1) an MT1-MMP activity-neutralizing antibody (MT1-MMP-InhAb; custom-made antibody designed using a peptide located within the catalytic domain of MT1-MMP (aa 90–120), Open Biosystems), or (2) a TGF-β neutralizing antibody that binds and sequesters all TGF-β isoforms (TGF-β -1, -2, -3) (TGF-β-NAb, Clone 1D11, R&D Systems, San Jose, CA, USA). Following measurement of baseline aortic diameters, TAA was induced in C57BL/6 mice. The following day mice were randomly assigned to receive intraperitoneal injections of either (1) MT1-MMP-InhAb (5 μg of MT1-MMP-InhAb × 3 injections/wk in 100 μL sterile PBS) or control (100 μL sterile PBS × 3 injections/wk), or (2) TGF-β-NAb (5 μg of TGF-β-NAb × 1 injection/wk in 100 μL sterile PBS) or control (5 μg of mouse IgG × 1 injection/wk in 100 μL sterile PBS), as previously described [34,35]. Terminal surgery was performed, and aortic diameter measurements were made 4 weeks following TAA induction.

2.11. Statistical Analysis

For all live-animal TAA studies, aortic diameters were determined within each animal and expressed as a percent change from its own baseline. Mean aortic diameter was compared within terminal study points using a one-sample mean comparison test versus 100% (baseline value), and over time between groups using ANOVA with Tukey’s wsd post hoc test for separation of means. For MT1-MMP abundance, MT1-MMP activity, and all histological findings, comparisons across time points were made using ANOVA with Tukey’s wsd post hoc test for separation of means. Comparisons within each time point were made between wild-type and MT1-MMP^+/−^ experimental groups using 2-tailed Student t-test. Correlation analysis was completed using pairwise linear regression. Slopes of regression lines between the wild-type and MT1-MMP^+/−^ were compared using analysis of covariance using a dummy variable to check for parallelism. Robust statistical tests were also considered for all variables to account for potential non-normal distributions (Shapiro–Wilk test). If variables were non-normally distributed, pairwise comparisons and post hoc separation of means were performed using the Kruskal–Wallis test and Mann–Whitney U-test, respectively. All statistical procedures were performed using the Stata statistical software package v.12 (StataCorp LLC, College Station, TX, USA). Results are presented as mean ± SEM. Values of p < 0.05 were considered to be statistically significant.

3. Results

3.1. Heterozygous Deficiency of MT1-MMP in TAA Development

3.1.1. Aortic Diameter Change in MT1-MMP+/− Mice

TAAs were induced in wild-type FVB and MT1-MMP^+/−^ mice. Mice were then randomly assigned for terminal study at 2, 4, 8, or 16 weeks post-TAA induction (n = 8 at 2 and 16 weeks post-induction; n = 13 at 4 and 8 weeks post-induction). Results were compared to unoperated control animals (wild-type FVB (n = 13) and MT1-MMP^+/−^ (n = 13)). As expected, there was a time-dependent increase in aortic diameter with TAA induction in both groups (Figure 1). In wild-type mice, aorta diameter was significantly increased over baseline starting at 2 weeks post-TAA induction and continued to dilate over time, reaching a maximum of approximately 50% over baseline by 16 weeks post-TAA induction, indicating dilation reached the level of a true aneurysm. In MT1-MMP^+/−^ mice, however, aortic diameter was significantly increased through 4 weeks post-TAA induction but plateaued at 8 and 16 weeks post-TAA at approximately 30% over baseline. This suggests that MT1-MMP deficiency caused significant attenuation of aortic dilation in the MT1-MMP^+/−^ group at the later time points of 8 and 16-week post-TAA induction.

3.1.2. MT1-MMP Protein Abundance

MT1-MMP abundance was subsequently determined in aortic samples by immunoblotting (representative immunoblot shown in Figure 2A). Normalized to unoperated wild-type controls, MT1-MMP abundance increased over time and was significantly different from control values at 8 and 16 weeks post-TAA induction (Figure 2B). MT1-MMP abundance in the MT1-MMP^+/−^ mice was lower than wild-type controls at baseline, but similarly increased over time, maintaining a decreased abundance compared to wild-type mice at every time point following TAA induction (Figure 2B).

3.1.3. MT1-MMP Activity

Using a quenched fluorogenic peptide cleavage assay, MT1-MMP activity was examined in aortic homogenates from TAA-induced wild-type FVB (n = 8) and MT1-MMP^+/−^ (n = 8) mice and compared to unoperated controls of each strain (Figure 3A). Results demonstrated that baseline MT1-MMP activity in unoperated MT1-MMP^+/−^ mice (n = 8) was significantly lower than that of wild-type FVB mice (n = 8). Moreover, following TAA induction, MT1-MMP activity in both wild-type and MT1-MMP^+/−^ mice increased from control levels at all time points tested. However, while MT1-MMP activity increased in the MT1-MMP^+/−^ mice due to the single remaining functional allele, the MT1-MMP activity values were significantly lower in the MT1-MMP^+/−^ mice at 4, 8, and 16 weeks post-TAA induction as compared to the wild-type animals, consistent with reduced MT1-MMP protein abundance. To determine whether a relationship existed between MT1-MMP activity and aortic dilatation in both groups, pairwise linear regression was performed (Figure 3B). Results demonstrated a significant positive correlation between MT1-MMP activity and the change in aortic diameter in wild-type mice (r = 0.7348, p < 0.01). More importantly, a positive correlation was also identified in the MT1-MMP^+/−^ mice (r = 0.6245, p < 0.01), showing a similar relationship but at a lower amplitude (Figure 3B). While the difference in slopes between the two regression lines (mouse strains) was not significant, the fact that the lines were not completely parallel is interesting and suggests there are likely other factors that also contribute to aortic dilation that may be dependent on the amount of MT1-MMP present. Taken together, these results suggest that MT1-MMP activity may be required for aortic dilatation.

3.1.4. Histology

Given the differences in aortic diameter, MT1-MMP abundance, and MT1-MMP activity between the wild-type control and the MT1-MMP^+/−^ mice, a subset of aortic specimens (wild-type FVB (n = 5) and MT1-MMP^+/−^ (n = 5)) were collected and processed for histological analysis. Images of picrosirius red-stained (PSR) sections were utilized for aortic medial wall thickness measurements as well as a determination of changes in collagen content in the presence and absence of TAA (Figure 4A). In unoperated control animals, wild-type FVB and MT1-MMP^+/−^ control mice showed similar aortic medial wall thickness. However, following TAA induction, while aortic wall thickness decreased in a time-dependent manner in both groups, there was a notable attenuation of the reduction in wall thickness at 8 weeks post-TAA in the MT1-MMP^+/−^ mice compared to the wild-type control animals (Figure 4B). To further examine MT1-MMP-dependent changes in the aortic wall, total collagen content was then measured in tunica media from the PSR sections imaged under polarized light. At baseline (in the absence of TAA), the MT1-MMP^+/−^ mice showed lower values of total collagen compared to the wild-type FVB mice (Figure 4C, left panel). This strengthens the suggestion that MT1-MMP plays a significant role in mediating overall collagen content, as has been previously described [11,36]. Following TAA induction, total collagen content in both groups was reduced at 4 and 8 weeks post-TAA induction compared to unoperated controls within their respective strains; however, no difference in total collagen content was observed between the two strains (Figure 4C, right panel). The PSR sections were further examined to differentiate the thinner/less crosslinked fibers, based on yellow-green birefringence, and the thicker/more highly crosslinked filaments based on orange-red birefringence. In the unoperated control animals, the yellow-green birefringence was similar in the baseline control strains, while the orange-red birefringence was significantly decreased (Figure 4D,E, left panels). Following TAA induction, levels of yellow-green birefringence were significantly reduced in both strains at 4 and 8 weeks compared to their respective control strains; however, no differences were observed between strains (Figure 4D, right panel). This suggests that MT1-MMP does not function to degrade newer/thinner/less crosslinked collagen fibers. Conversely, while the levels of orange-red birefringence decreased in the wild-type FVB mice at 4 and 8 weeks post-TAA induction compared to controls, the levels of orange-red birefringence in the MT1-MMP^+/−^ mice did not significantly change from control levels (Figure 4E, right panel). This suggested that older/thicker/more crosslinked collagen fibers are a more appropriate substrate for MT1-MMP activity.

Thus, taken together, these results suggest that the loss of MT1-MMPdue to genetic allelic reduction, results in the attenuation of medial wall thinning, which may be mediated in part by reduced MT1-MMP-dependent proteolysis of highly crosslinked collagen fibers within the aortic wall.

3.2. Fibroblast-Specific MT1-MMP in TAA Development

Previous efforts to identify a cellular source of MT1-MMP within the aorta during TAA development have demonstrated a significant colocalization of MT1-MMP and DDR2, a fibroblast/myofibroblast cell-specific marker [27,37]. To determine the role of fibroblast-derived MT1-MMP in TAA development, we developed a conditional fibroblast-specific tamoxifen-inducible MT1-MMP knockout mouse using a Cre-lox strategy. The resulting FbMT1KO mice were subjected to tamoxifen treatment (intraperitoneal injections, 75 mg/kg/day (tamoxifen in corn oil), for 5 consecutive days, followed by 10 days of rest), and the fibroblast specificity of MT1-MMP knockout was validated at both the protein and PCR level (Supplemental Figure S2). Accordingly, the deletion of fibroblast-specific MT1-MMP was tested under two conditions: (1) “early” deletion prior to TAA induction and (2) “late” deletion 4 weeks following TAA induction.

In the first condition involving “early” deletion, control mice (Col1A2-Cre(ERT2)-negative × MT1-MMP^fl/fl^) and FbMT1KO mice were treated with tamoxifen for 5 days. Ten days following the end of treatment, TAA was induced. Aortic diameter and aortic MT1-MMP protein abundance were measured at 4, 8, and 16 weeks post-TAA induction. Unfortunately, extensive scarring was observed at the site of TAA induction in the tamoxifen-treated FbMT1KO mice, which compromised the ability to accurately visualize and measure the change in aortic diameter over time (Supplemental Figure S3A). However, the time-dependent increase in MT1-MMP protein abundance during TAA progression observed in the tamoxifen-treated control mice was significantly inhibited in tamoxifen-treated FbMT1KO mice (Supplemental Figure S3B), suggesting that tamoxifen-induced knockout of MT1-MMP in aortic fibroblasts can, at minimum, recapitulate the effects of the MT1-MMP deficiency at the protein level. Importantly, these findings also suggest that MT1-MMP may play a role in the regulation of post-surgical wound healing and remodeling of scar tissue, which, when knocked out, may permit the overabundance of scar formation.

In the second condition involving “late” deletion, the role of MT1-MMP in TAA progression was examined by knocking out fibroblast-specific MT1-MMP in mice that had already undergone TAA induction. Accordingly, TAA was induced in control and FbMT1KO mice and, at 4 weeks post-TAA induction, tamoxifen treatment was initiated for 5 consecutive days. Mice then underwent terminal surgery at 4, 8, and 16 weeks post-TAA and aortic diameters were measured as before. Excessive scar tissue was not observed at any time point post-TAA in either treatment group (Figure 5A). Aortic diameter measurements increased significantly over baseline at 4 and 8 weeks post-TAA induction in both experimental groups, consistent with expectations of aneurysm development observed in previous studies [27,38]. Importantly, at 16 weeks post-TAA induction, while the control animals continued to dilate, reaching a maximal diameter increase of 67 ± 3% over baseline, aortic dilation in the FbMT1KO mice was arrested, remaining at 36 ± 5% over baseline (Figure 5B). Aortic homogenates were then examined for MT1-MMP abundance. As expected, MT1-MMP protein levels were increased over baseline at 16 weeks post-TAA in the tamoxifen-treated control mice, while no increase was observed in tamoxifen-treated FbMT1KO mice (Figure 5C). Taken together, these findings strongly suggest that fibroblast-derived MT1-MMP is critically required for TAA progression.

3.3. Inhibition of MT1-MMP/TGF-β Signaling in TAA Development

As mentioned, it is well-described that MT1-MMP plays an important role in the release of matrix-bound growth factors like TGF-β [17,22,23,24,25,26]; thus, the cell-specific correlation between MT1-MMP abundance and TGF-β activation was further explored in the aortic fibroblasts. Aortic fibroblasts were isolated from wild-type FVB, MT1-MMP overexpressing (FbMT1-MMP), MT1-MMP^+/−^, and MT1-MMP^−/−^ mice. Baseline levels of phosphorylated Smad-2 (pSmad-2), a marker of TGF-β activation, were examined by immunoblotting (Figure 6A). Compared with wild-type fibroblasts, pSmad-2 levels were elevated in FbMT1-MMP fibroblasts but progressively decreased in heterozygous MT1-MMP^+/−^ and knockout MT1-MMP^−/−^ fibroblasts (Figure 6B). Moreover, pairwise linear regression analysis revealed a direct and significant correlation between MT1-MMP abundance and pSmad-2 abundance (r = 0.5745, p < 0.01) (Figure 6C), supporting the idea that MT1-MMP plays a critical role in the release of ECM-bound TGF-β, which subsequently stimulates TGF-β signaling pathways.

Based on this observation, it is of great interest to examine whether the inhibition of MT1-MMP activity, as well as the neutralization/sequestration of MT1-MMP-released TGF-β ligands, could attenuate TAA progression. To address this, a custom-made antibody was designed to bind within the catalytic domain of MT1-MMP (aa 90–120). The inhibitory effect was validated in vitro by co-incubating with human recombinant MT1-MMP catalytic domain and assessing MT1-MMP activity using a commercially available quenched fluorogenic peptide substrate (Supplemental Figure S4, top panel). The IC_50_ of this MT1-MMP activity-neutralizing antibody (MT1-MMP-InhAb) was determined to be approximately 4.6 µg/mL (Supplemental Figure S4, bottom panel). To examine its effects in vivo, TAA was induced in C57BL/6 mice and, on the following day, mice began treatment with the MT1-MMP-InhAb (three injections/week ip for 4 weeks; 5 μg/mL in 100 μL of PBS) or PBS alone (Figure 7A). Aortic diameters were measured at TAA induction and again at terminal study, 4 weeks post-TAA induction. As shown in Figure 7B, mice injected with PBS exhibited significant aortic dilatation from baseline as expected, whereas TAA-induced mice injected with MT1-MMP-InhAb showed significant attenuation of aortic dilatation and failed to show aortic dilatation from baseline. These data suggest that this antibody-mediated inhibition of MT1-MMP activity is sufficient to attenuate aortic dilatation in TAA-induced animals.

Similarly, if MT1-MMP is releasing ECM-sequestered TGF-β as part of the TAA development process, blocking TGF-β signaling should also attenuate aortic dilation. Accordingly, TAA was induced in C57BL/6 mice and, on the following day, mice began treatment with a well-described TGF-β neutralizing antibody (TGF-β-NAb; Clone 1D11, 1 injection/week ip for 4 weeks; 5 mg/kg in 100 μL PBS) or a mouse IgG control antibody (1 injection/week ip for 4 weeks; 5 mg/kg in 100 μL PBS) (Figure 7A). Aortic diameters were measured at TAA induction and again at terminal study, 4 weeks later. As shown in Figure 7C, mice injected with either mouse IgG or TGF-β-NAb both showed significant dilation over baseline; however, aortic dilation was significantly attenuated at 4 weeks in the mice treated with TGF-β-NAb as compared with mice treated with the control IgG, indicating that sequestration of free TGF-β is likewise capable of attenuating aortic dilation in TAA-induced animals.

4. Discussion

A thoracic aortic aneurysm (TAA) is defined as a localized dilatation of the supra-diaphragmatic aorta to a cross-sectional diameter greater than 1.5 times its normal baseline value [9,11]. The mean age at diagnosis is approximately 65 years [1], but it is likely that the initiation of this process begins much earlier. There are numerous etiologies that contribute to the formation and progression of TAA disease, with the most common etiology being the least well-defined, resulting from non-genetic/non-syndromic idiopathic degeneration of the aortic vascular ECM. Aneurysm development usually proceeds by an asymptomatic process that results in a weakened aortic wall and is manifested as gross dilatation that progresses to rupture if left unattended [9,39]. Current treatment options are limited and consist primarily of surgical reconstruction or endovascular intervention [40,41,42,43]. Once diagnosed, a “watch and wait” surveillance program is initiated until the risk of aortic rupture outweighs the risk of the surgical repair. During this period, patients are typically treated with β-blockers to attenuate elevated blood pressure and myocardial dP/dt (rate-rise time), dampening the force of the cardiac pulse wave moving through the aorta. While recent advancements in stent-grafting have significantly decreased the early mortality and postoperative complications associated with open surgical repair, endovascular aortic repair (EVAR) is not without limitations and has not shown a proven benefit over open repair beyond 4 years [44]. Most importantly, neither of these options are aimed directly at the underlying cellular and molecular mechanisms responsible for this devastating disease. Due to the paucity of non-surgical options for the treatment of TAA [45,46], further diagnostic and therapeutic advancement is critical and especially relevant for those patients who have not yet reached surgical criteria. As highlighted by Rentschler et al., ideal medical therapy does not need to completely arrest aneurysm development; slowing the aneurysm growth rate, by even 30–50%, could delay the need for surgical intervention beyond the patient’s lifespan [7].

The present study was designed to follow up on previous observations to demonstrate that MT1-MMP plays a causative role in aneurysm formation and progression [27,47]. Accordingly, the present study examined TAA development in total-body and tamoxifen-inducible fibroblast-specific MT1-MMP knockout mice and demonstrated that aortic dilation can be attenuated through targeting MT1-MMP abundance, activity, or MT1-MMP-mediated release of ECM-bound TGF-beta ligands. The unique findings of this study were three-fold. First, there is a direct and positive relationship between MT1-MMP abundance, activity, and aortic dilatation. Inhibition or partial genetic deletion of MT1-MMP was sufficient to attenuate TAA formation and progression. Moreover, specifically deleting fibroblast-derived MT1-MMP was also sufficient to attenuate TAA development. Second, MT1-MMP-mediated pericellular proteolysis regulates the abundance of thicker/more crosslinked collagen fibers within the medial aortic wall, as well as the release of ECM-bound TGF-β. Together, this suggests that altering MT1-MMP activity is sufficient to attenuate aortic structural remodeling during TAA development, slowing aortic dilation. Furthermore, these results suggest that TGF-β release occurs downstream of MT1-MMP activity, validating the hypothesis that MT1-MMP activity and TGF-β mediated signaling are required for aortic remodeling during TAA development. Lastly, these results highlight MT1-MMP as a critical target for therapeutic advantage and argue that non-surgical therapy may be beneficial for the treatment of TAA disease. Strategies aimed at inhibiting changes in MT1-MMP abundance, activity, or functional consequences (TGF-β release) should be further explored in animal models and clinical trials.

4.1. Altering MT1-MMP Abundance or Activity

Previous work from this laboratory has suggested that MT1-MMP abundance increases over time during TAA development, suggesting that it plays a critical role in the pathogenesis [27]. To directly test this, we surgically induced TAA in MT1-MMP^+/−^ mice and compared the results to wild-type FVB littermates. As anticipated, in the wild-type animals, upon induction, the aorta dilated over time, reaching a true aneurysm by 16 weeks. Interestingly, TAA development was attenuated in mice carrying the partial genetic deletion of MT1-MMP. Upon analysis of tissue homogenates, we observed that MT1-MMP abundance and MT1-MMP activity increased over time in both strains, but the protein levels and activity were reduced in the heterozygous mice, starting from a lower level at baseline. A significant positive correlation between MT1-MMP activity and aortic diameter was identified in wild-type FVB mice, such that higher MT1-MMP activity was associated with a larger aortic diameter. In the MT1-MMP heterozygous deficit animals, a similar relationship was identified, although at a lower amplitude; no difference between slopes of the two regression lines was observed. While MT1-MMP abundance and activity both increased with TAA development, the overall reduction in MT1-MMP in the heterozygous mice resulted in a slower aortic dilation rate. These data suggest that strategies aimed at reducing MT1-MMP abundance and/or activity may be sufficient to slow aortic dilatation. Indeed, when MT1-MMP activity was inhibited using a highly specific antibody directed against a peptide contained within the catalytic domain, aortic dilation was attenuated. Importantly, our previous work identified concomitant changes in fibroblast phenotype and function during TAA development [33,38] and demonstrated that MT1-MMP colocalized with these DDR2^+^ fibroblasts [27]. This suggested that alterations in fibroblast phenotypes may be responsible for the increase in aortic MT1-MMP abundance and activity and thereby may play a direct role in mediating TAA formation and progression. To examine this further, we generated a conditional fibroblast-specific tamoxifen-inducible MT1-MMP knockout mouse (FbMT1KO). We then induced TAA, and, at 4 weeks post-TAA induction, started tamoxifen treatment for 5 consecutive days. We followed the animals through 16 weeks post-TAA induction and discovered that aortic diameter was stabilized in the FbMT1KO mice. Taken together, these results suggest that aortic fibroblast-derived MT1-MMP is both necessary and sufficient to mediate aortic dilation during TAA development.

4.2. Pericellular Proteolysis and TGF-Beta Release

MT1-MMP is a multifunctional protein that acts upon a diverse set of pericellular matrix proteins at the cell–matrix boundary, plays an indispensable role in activating other MMPs, and plays an important role in regulating/releasing ECM-bound growth factors. Accordingly, we examined the role of MT1-MMP in mediating ECM degradation during TAA development in the wild-type and MT1-MMP^+/−^ deficient mice. Histological sections were used to measure wall thickness and collagen volume fraction. First, we identified that, as TAA development progressed in both strains, aortic wall thickness decreased. While aortic dilatation is often accompanied by increased wall thickness secondary to atherosclerosis and intimal hyperplasia, typically observed in abdominal aortic aneurysms or some descending aortic aneurysm with advanced atherosclerotic disease, we do not typically observe increased aortic wall thickness in our mouse model of TAA. The model we utilize involves briefly exposing the periadventitial surface of the descending thoracic aorta to a calcium chloride solution, which induces a process, absent from acute chemical injury, that initiates a series of events that culminate in aneurysmal dilation. During TAA development, we observe changes in the aortic structural elements and the cellular constituents consistent with what is observed in human aneurysmal disease (activation of proteolytic enzymes, changes in cellular signaling, and alterations in smooth muscle cell and fibroblast phenotype and function). However, it is notable that this occurs in the absence of atherosclerosis and intimal hyperplasia. Interestingly, in the MT1-MMP^+/−^ deficient mice, the decrease in wall thickness occurred at a slower rate as compared to wild-type control animals due to the genetic allelic reduction in MT1-MMP^+/−^ mice, resulting in an attenuation of medial wall thinning during TAA development. When total collagen content was measured, the MT1-MMP^+/−^ deficient mice showed a significant decrease compared to controls. We then further examined the picrosirius red-stained collagen content under polarized light and separated yellow-green birefringence, representing thinner/less crosslinked collagen fibers, from orange-red birefringence, representing thicker/more crosslinked collagen fibers. The results demonstrated that a decrease in the yellow-green fibers was evident and similar in both strains; however, the decrease in orange-red fibers in the MT1-MMP^+/−^ deficient mice was attenuated. These results suggest that MT1-MMP plays a role in regulating/managing crosslinked collagen within the aortic wall. Although thick fibers constitute a smaller portion of the total collagen content, the preservation of thicker collagen filaments may contribute to the decreased post-TAA aortic wall thinning, and may prevent/attenuate aortic structural remodeling, reducing the rate of aortic dilation in MT1-MMP^+/−^ mice.

Since MT1-MMP has been implicated in the release of ECM-sequestered TGF-β ligands [25], we examined the ability of primary aortic fibroblasts isolated from multiple mouse strains, genetically manipulated to express different levels of MT1-MMP (wild-type, over-expressers, heterozygous deficient, homozygous deficient), to activate TGF-β signaling in vitro under normal (unstimulated) growth conditions. Protein lysates were examined for MT1-MMP and phosphorylated Smad-2 (pSmad-2) abundance by immunoblotting. As expected, MT1-MMP protein levels in each isolated cell line varied according to its genetic abundance. Importantly, pSmad-2 displayed a similar profile and a significant positive relationship between MT1-MMP and pSmad-2 abundance was identified, indicating that MT1-MMP may indeed regulate local ECM-sequestered TGF-β ligands as well as the baseline activation of the TGF-beta signaling pathway. Thus, these results may explain the baseline differences in aortic collagen volume fraction observed between the MT1-MMP^+/−^ mice and wild-type controls. Furthermore, based on our previous work demonstrating alterations in TGF-β signaling during TAA formation and progression [32], these results suggest that MT1-MMP may also play an important upstream role in TGF-β release during TAA development. Indeed, when TAA was induced in wild-type C57BL/6 mice treated with the TGF-β-neutralizing antibody, aortic dilation was attenuated, similar to findings from Habashi et al. and Holm et al. in a murine model of Marfan Syndrome [35,48]. Taken together, these results suggest that targeting MT1-MMP activity can alter TGF-β release, preventing TGF-β pathway activation, which is sufficient to slow aortic dilation.

4.3. Therapeutic Strategies Targeting MT1-MMP Activity and Functional Outcomes

Results from the present study have demonstrated that suppression of MT1-MMP protein production through genetic ablation (MT1-MMP^+/−^ or FbMT1KO) was sufficient to reduce MT1-MMP activity and attenuate TAA development. Furthermore, targeting MT1-MMP activity using an MT1-MMP activity-neutralizing antibody or targeting the downstream consequences of MT1-MMP function (release of ECM-sequestered TGF-β ligands) was likewise capable of attenuating aortic dilation. These results argue that MT1-MMP should be identified as a high-value target for the treatment of TAA disease. Notably, while these results focused on MT1-MMP and TGF-β ligand inhibition within the aorta at the site of aneurysm, the potential for site-specific off-target effects of these systemically delivered antibodies remains a concern. Therapeutic strategies designed to inhibit MMP activity as a means to limit ECM remodeling had lost favor clinically years ago because of reported severe off-target effects. Thus, targeting therapeutics to the site of the aneurysm has proven to be a tremendous challenge in the field. While antibody therapy in general has recently gained traction within the medical community, a new resurgence of MMP-directed antibody therapy is being explored under various conditions [49]. This is particularly important for therapeutic strategies targeting common proteolytic mechanisms that are found throughout the body. So, while these results highlight MT1-MMP as a critical target for therapeutic advantage and argue that non-surgical therapy may be beneficial for the treatment of TAA disease, additional strategies aimed at the long-term inhibition of MT1-MMP abundance, activity, or functional consequences (TGF-β release) should be further explored in preclinical animal models and clinical trials.

4.4. Limitations

While these novel results suggest that MT1-MMP plays an essential role in TAA development, this study is not without limitations. First, we have detailed the results from a total-body knockout of MT1-MMP and a conditional–inducible MT1-MMP knockout. These animals, while internally controlled within each experiment, were generated on different background mouse strains (FVB: total-body knockout; C57BL/6: conditional–inducible knockout). While we do not believe that the differences in genetic background between these two sets of studies altered the response to TAA induction surgery in any way (the control animals in both studies dilated to a similar degree), it is possible that the differences in genetic background could affect baseline aortic ECM structure and composition, which could affect the rate or trajectory of TAA development. Second, our surgical mouse model of TAA involves the use of CaCl_2_ applied directly to the periadventitial surface of the descending thoracic aorta. While it is understood that no animal model perfectly replicates clinical TAA disease based on its complexities, this model has many benefits, including portability to any mouse strain (as well as transgenics and knockouts) without regard to genetic background or murine lipid status. Moreover, because aneurysm formation occurs over a prolonged period of time (16 weeks), it facilitates discovery of the mechanistic underpinnings driving the pathological process during TAA development. Importantly, while this mouse model recapitulates many of the key hallmarks of clinical TAA disease, it lacks the natural contribution of other key comorbidities such as atherosclerosis and hypertension. With all of this in mind, these results would likely be generalizable for other segments of the thoracic aorta, such as ascending TAA, but may fail to adequately represent abdominal aortic disease. Lastly, we present results using two different antibody-mediated therapies focused independently on MT1-MMP activity neutralization or TGF-β ligand sequestration. It must be recognized that, due to systemic delivery, these antibodies can target epitopes throughout the body; thus, off-target effects can be difficult to manage. That said, both antibodies were able to alter TAA development without inducing obvious deleterious effects on the treated animals, further suggesting that MT1-MMP activity and TGF-β signaling are key requirements for TAA formation and/or progression. Issues related to the optimal route of delivery, effective antibody dosing, the precise site of action, and the duration of the inhibitory effects will need to be explored in subsequent studies. Furthermore, additional studies designed to definitively prove the serial relationship between MT1-MMP activity and the activation of the TGF-β signaling pathway should also be explored.

5. Conclusions

The unique findings of this study demonstrate that aortic fibroblast-derived MT1-MMP plays an obligate role in TAA formation and progression, and by suppressing its protein expression or inhibiting its activity, aortic dilation can be attenuated. These results support our previous studies identifying MT1-MMP as a turn-key mediator of TAA development and argue for the need to develop rationally designed targeted therapeutics, whether they be cellular or pharmacologically based, that focus on the functional regulation of MT1-MMP for the treatment of TAA disease.