Defective Connective Tissue Remodeling in Smad3 Mice Leads to Accelerated Aneurysmal Growth Through Disturbed Downstream TGF-β Signaling.

Aneurysm-osteoarthritis syndrome characterized by unpredictable aortic aneurysm formation, is caused by SMAD3 mutations. SMAD3 is part of the SMAD2/3/4 transcription factor, essential for TGF-β-activated transcription. Although TGF-β-related gene mutations result in aneurysms, the underlying mechanism is unknown. Here, we examined aneurysm formation and progression in Smad3-/- animals. Smad3-/- animals developed aortic aneurysms rapidly, resulting in premature death. Aortic wall immunohistochemistry showed no increase in extracellular matrix and collagen accumulation, nor loss of vascular smooth muscle cells (VSMCs) but instead revealed medial elastin disruption and adventitial inflammation. Remarkably, matrix metalloproteases (MMPs) were not activated in VSMCs, but rather specifically in inflammatory areas. Although Smad3-/- aortas showed increased nuclear pSmad2 and pErk, indicating TGF-β receptor activation, downstream TGF-β-activated target genes were not upregulated. Increased pSmad2 and pErk staining in pre-aneurysmal Smad3-/- aortas implied that aortic damage and TGF-β receptor-activated signaling precede aortic inflammation. Finally, impaired downstream TGF-β activated transcription resulted in increased Smad3-/- VSMC proliferation. Smad3 deficiency leads to imbalanced activation of downstream genes, no activation of MMPs in VSMCs, and immune responses resulting in rapid aortic wall dilatation and rupture. Our findings uncover new possibilities for treatment of SMAD3 patients; instead of targeting TGF-β signaling, immune suppression may be more beneficial.

Introduction

Aortic aneurysms significantly increase the risk of tearing (dissection) or rupture of the aorta with life-threatening consequences. As for several cardiovascular diseases, male gender, advanced age, and positive family history are among the main risk factors for developing aneurysms (Albornoz et al., 2006, Grubb and Kron, 2011). Multiple genes for thoracic aortic aneurysms have been identified (Gillis et al., 2013), but most of the underlying genetic and molecular interactions are not known. These genes mainly fall into three categories, based on their function; 1) extracellular matrix integrity and structure, 2) involvement in TGF-β signaling, and 3) cytoskeleton maintenance and mobility. Although their functions are quite different, defects in these genes all lead to aneurysm formation.

The extracellular matrix (ECM) is important for the integrity of the aortic wall and mutations in ECM genes such as Fibulin-4 and Fibrillin-1 hence lead to aneurysm formation. Genetic studies in Fibulin-4 and Fibrillin-1 mutant mice implicated dysregulation of the TGF-β pathway as an important hallmark in the pathogenesis of aneurysm formation, both in mice and humans (Neptune et al., 2003, Habashi et al., 2006, Hanada et al., 2007, Huang et al., 2010, Kaijzel et al., 2010, McLaughlin et al., 2006).

Several genes that are involved in cytoskeletal maintenance are mutated in aneurysmal disease. Examples are ACTA2 and MYH11, which function in the contractile apparatus of the smooth muscle cell, and mainly cause thoracic aneurysms when mutated (Guo et al., 2007, Zhu et al., 2006), underscoring the importance of cytoskeleton maintenance and mobility in aneurysmal disease.

The TGF-β pathway plays a relevant role in the etiology of aortic aneurysms. Twenty years ago, the first member of the TGF-β signaling pathway was linked to a genetic vascular disease, after finding mutations in the gene coding endoglin (McAllister et al., 1994). Loss of function mutations in this TGF-β binding protein were described to cause hereditary hemorrhagic telangiectasia type I. This was followed by the discovery of mutations in the receptors TGFβR1 and TGFβR2 (Loeys et al., 2005, Mizuguchi et al., 2004), SMAD3 (van de Laar et al., 2011), the ligands TGFβ2 (Lindsay et al., 2012, Boileau et al., 2012) and TGFβ3 (Bertoli-Avella et al., 2015, Rienhoff et al., 2013). These mutations lead to a spectrum of systemic disorders characterized by aneurysms and other cardiovascular and skeletal features known as Loeys-Dietz syndrome (LDS).

The SMAD (SMA/MAD homology) proteins are important regulators of the TGF-β signaling pathway and function as signaling transducers downstream of TGF-β receptors. The SMAD protein family consists of receptor SMADs (SMAD1–3, SMAD5, SMAD8), the co-effector SMAD4 and inhibitory SMADs (SMAD6 and SMAD7) (Massague, 2012, Massague et al., 2005). Activated SMAD2 and SMAD3 can form heteromeric (pSMAD2/4, pSMAD3/4) complexes in the nucleus where they form transcription-activating complexes capable of inducing or repressing the expression of several genes (Massague et al., 2005, Moustakas and Heldin, 2002) in a cell-type and SMAD complex-dependent manner.

We recently described a genetic disease characterized by aneurysms, dissections and cardiac abnormalities in combination with early-onset osteoarthritis (OA) known as aneurysm-osteoarthritis syndrome (AOS or LDS3; MIM 613795) caused by heterozygous mutations in the SMAD3 gene (van de Laar et al., 2011). Patients carrying heterozygous SMAD3 mutations present with extreme clinical variability in cardiovascular disease onset and progression (Van de Laar et al., 2012, Van der Linde et al., 2012, Van der Linde et al., 2013). The exact molecular mechanisms and contributing factors underlying this lack of genotype-phenotype correlation remain to be established, as well as the variable effect that genetic variants of SMAD3 can have on different tissues. SMAD3 mutations are suggested to lead to upregulation of the TGF-β pathway in the aortic wall as indicated by nuclear translocated and activated SMAD2 (pSMAD2) (van de Laar et al., 2011). Activated SMAD2 is also seen upon mutational hits in the TGFβR1/2 receptors, or the TGFβ2 ligand (Lindsay and Dietz, 2011, Lindsay et al., 2012, Loeys et al., 2005). Because pSMAD2 is considered to report on activation of the TGFβ pathway, these finding are referred to as the TGFβ paradox, as one would expect that mutations in genes involved the TGFβ pathway would hamper TGFβ signaling (Akhurst, 2012, Massague, 2012). However, it is unclear whether pSMAD2 is a functional marker for the downstream upregulation of the TGFβ pathway. Similarly, mutations in genes involved in build-up and integrity of the ECM lead to an upregulation of the TGF-β signaling pathway and aneurysm formation. For the ECM related gene mutations it is thought that this upregulation is due to release of TGF-β ligand from the ECM, caused by loss of ECM integrity, resulting in ECM remodeling and aortic stiffness (Gillis et al., 2013). It remains to be seen whether the same underlying mechanism is at work when comparing ECM- and TGF-β related gene deficiency in aneurysm formation.

The clinical heterogeneity in AOS patients makes it difficult to study SMAD3 mutational effects on aneurysm formation. Fortunately, due to the homogenous genetic background, genetically engineered mouse models are useful in pinpointing the specific molecular mechanism leading to disease. Smad3 knockout animals present with skeletal abnormalities and osteoarthritis (OA) and as such, they have been used as a model to study OA (Yang and Cao, 2001, Li et al., 2009). A cardiovascular phenotype in these animals was overlooked until the recent link of human SMAD3 mutations and aortic aneurysms was established (Regalado et al., 2011, Van de Laar et al., 2011, Ye et al., 2013). Here we describe the cardiovascular phenotype of the Smad3 knockout mice and reveal the underlying mechanism of aneurysm growth caused by a SMAD3 deficiency.

Materials and Methods

Experimental Animals

Smad3+/‐ animals were bred into in a C57BL6 background to obtain Smad3−/− and Smad3 experimental animals (backcross 6). The numbers of animals, as well as procedures used, are described in the results section and below, respectively. Animals were housed at the Animal Resource Centre (Erasmus University Medical Centre), which operates in compliance with the “Animal Welfare Act” of the Dutch government, using the “Guide for the Care and Use of Laboratory Animals” as its standard. As required by Dutch law, formal permission to generate and use genetically modified animals was obtained from the responsible local and national authorities. An independent Animal Ethics Committee of the Erasmus Medical Center (Stichting DEC Consult) approved these studies (permit number 140–12-05), in accordance with national and international guidelines.

Litter- and gender matched controls were used for each experiment when available.

Echocardiographic Measurements

Ascending aortic diameter was measured in M-mode, aortic root diameter was measured at the site of the sinus of Valsalva in B-mode. Aortic length was measured as the distance between the sinus of Valsalva and the brachiocephalic trunk. All mice were ventilated and anesthetized with 2.5% isoflurane and echocardiography of the ascending aorta was performed using a Vevo2100 (VisualSonics Inc., Toronto, Canada). Longitudinal echocardiographic measurements of the ascending aorta were performed on 6, 12, 18 and 26 week old Smad3 and Smad3 mice (n = 18, 8 male and 10 female per genotype).

Immunohistochemistry

For histological analysis mice were euthanized by CO2-inhalation. After opening thorax and abdomen, mice were fixed by perfusion fixation through the left ventricle, with PBS and formalin. Organs were weighed and inspected for macroscopic abnormalities. Organs and tissues were fixed in formalin. Aortas were dehydrated through the histokinette processor (Microm), and paraffin embedded, after which 5-μm sections were prepared. Aortas were stained with HE for general pathology, Resorcin-Fuchsin (RF; Elastin von Gieson) for elastin structure, Alcian Blue (AB) to evaluate the extracellular matrix (ECM), and Picrosirius Red (PR) to assess collagen accumulation.

For immunohistochemical analyses, thoracic aortic sections were boiled in 100 mM Tris-HCl [pH 9.0] with 10 mM EDTA at 300 W for 20 min for antigen exposure, and emerged in 3% H2O2 in methanol to inhibit endogenous peroxidase for pSmad2, α-SMA, pERK, CD31, MMP, CD3, MAC2 and Ki-67 staining. Slides were first blocked in 5% Protifar in PBS and 0.025% Triton, and incubated with the primary antibodies overnight at 4 °C; Anti-Human Smooth Muscle Actin (1:100 mouse, clone 1A4 Dako), pSmad2 (1:100 monoclonal Rabbit anti-pSmad2 (S465 | 467 (138D4) Cell Signaling), pERK (1:200 Rabbit Polyclonal anti-pErk (phosphor-p44/42 Mapk (Erk1/2) (Thr202/Tyr204) antibody) (#9101) Cell Signaling), MMP-9 (1:50 goat polyclonal anti MMP-9 (sc-6840) Santa Cruz Biotechnology, Ki-67 (1:200 Rat anti mouse Ki-67, clone TEC-3, DAKO), MAC-2 (1:500 Rat anti mouse MAC-2, CL8942AP, Cedarlane), CD31 (1:50 Rabbit pAb to CD31, ab28364, Abcam), CD3 (1:200 Polyclonal Rabbit anti human CD3, REF A0452, DAKO). The next day slides were incubated with biotinylated secondary antibodies (1:100 DAKO) and avidin-biotinylated complex (Vectastain Universal Elite ABC kit Vector Laboratories). DAB chromogen (DAKO Liquid Dab substrate-chromogen system) was used as substrate and slides were counterstained with hematoxylin. In total 8 Smad3 and 8 Smad3+/+ mice were examined for each staining.

Molecular Imaging for MMP Activation

We used vascular fluorescent mediated tomography (FMT) imaging with near-infrared fluorescent protease activatable probes as previously described (Kaijzel et al., 2010, Nahrendorf et al., 2011). As in aortic aneurysms MMP2 and MMP9 are most abundant, the probe is mostly cleaved by these two proteases. FMT imaging was performed using an FMT 2500 system (Perkin Elmer, Inc.) at 680/700 nm excitation and emission wavelengths, 24 h after tail vein injection of 2 nmol per 25 g bodyweight of the MMPsense™ 680 probe, or at 750/775 nm excitation and emission wavelengths, 6 h after injection of 2 nmol per 25 g bodyweight of the MMPsense™750 FAST near-infrared fluorescent probe (Perkin Elmer, Inc.), mice were imaged in a portable animal imaging cassette between optically translucent windows. The FMT 2500 quantitative tomography software was then used to calculate 3D fluorochrome concentration distribution of the fluorescent signal. After fluorescence imaging, aortas were harvested and fluorescence was quantified using the FMT 2500 or Odyssey imaging systems (LI-COR Inc.). Near-infrared images were obtained in the 680- and 750-nm channels, respectively. Relative fluorescence intensities were calculated for each Smad3−/− animals compared to its Smad3+/+ littermate control.

CT-scans

μCT imaging was performed using a Quantum FX μCT system (90 kV,160 μA,FOV 60 mm, 2 min 85 mGy scantime) (Perkin Elmer, Inc.). To visualize the vasculature, blood-pool contrast was administered via tail vein injection with 150 μl eXIA™160 Iodine based Radiocontrast (160 mg I/ml) per 25 g bodyweight (Binitio biomedical). A mouse imaging shuttle device was used to sequentially image the mice with both the FMT 2500 and Quantum FX, achieving accurate animal positioning to align both the fluorescence and μCT images. The optical and CT data sets were co-registered in 3 dimensions using the TrueQuant software (Perkin Elmer, Inc.). This multi-modal approach allows simultaneous monitoring of aneurysm growth and MMP- activity.

Isolation of SMCs, Characterization, Cell Culture and Proliferation Assay

Vascular SMCs were isolated from the media of the aortic arch of Smad3−/− and Smad3+/+ mice. The tissue was washed with PBS, cut into 5 mm pieces with the luminal side on 0.1% gelatin-coated cell culture dishes and incubated. After 7–10 days, smooth muscle-like cell outgrowth was observed. SMCs were maintained in DMEM (Lonza, Leusden, the Netherlands), supplemented with 10% fetal calf serum (HyClone, Thermo Scientific, Breda, the Netherlands), 100 U/ml penicillin and 100 μg/ml streptomycin (Sigma-Aldrich, Zwijndrecht, the Netherlands). For characterization, subconfluent SMCs and HUVECs were grown on coverslips and fixed in 2% paraformaldehyde. Cells were permeabilized with PBS/Triton (0.1%) and blocked with PBS + (0.5% BSA/0.15% glycine in PBS). Coverslips were incubated overnight with the primary antibody, anti-Human Smooth Muscle Actin (1:1000 mouse, clone 1A4 Dako). After washing, coverslips were incubated with the secondary antibody, alexa fluor 594 goat anti mouse IgG (1:1000, Life technologies). Coverslips were mounted in Vectashield with Dapi (Vector Laboratories). For the proliferation assay, cells were used at passage 5–11. Smad3+/+ and Smad3−/− VSMCs were seeded in triplicate in 6 cm dishes (5000 cells/well) and allowed to attach. The cells were counted with the coulter counter every day for a week.

RNA Isolation and Real-time PCR

Thoracic (arch and ascending) aortic tissue from Smad3+/+ and Smad3−/− mice, snap frozen and stored at − 80 °C, was used for this experiment. RNA was isolated with the RNeasy Fibrous tissue mini kit (Qiagen) for aortic tissue and the RNeasy mini kit (Qiagen) for VSMCs. cDNA was made with the iScript cDNA synthesis kit (Biorad) according to the manufacturer's protocol. Q-PCR was performed with 200 nM forward and reverse primers and iQ™ SYBR® Green Supermix (biorad) on the CFX384 system (Biorad); denaturation at 95 °C for 3 min, 40 cycles denaturation at 95 °C for 15 s, annealing/extension at 55 °C or 60 °C (as indicated below) for 30 s. B2M, Tbp1 and Ppia were used as a housekeeping gene. Relative gene expression levels were determined with the comparative ΔΔCt method. ΔΔCT = (CT(gene of interest, wildtype) − CT(reference gene, wildtype)) − (CT(gene of interest, knock-out) − CT(reference gene, knock-out)), where the fold change is calculated as 2ΔΔCT.

Primer sequences used for real-time PCR:

Western Blotting

Equal amounts of protein (Lowry protein assay) were separated by electrophoresis on a 15% SDS-polyacrylamide gel. Proteins were transferred to a PVDF membrane (Millipore) and blocked with 3% milk powder (ERK, β-catenin) or 5% BSA (pERK) in PBS containing 0.1% Tween-20. Next, blots were incubated overnight with the primary antibody (ERK, 1:2000, Rabbit anti p44/42 MAPK (Erk1/2), #9102S, Cell Signaling, pERK, 1:2000, Rabbit anti Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204), #9101L, Cell Signaling, Mouse anti β-catenin, BD Biosciences (1:5000)). After washing, blots were incubated with HRP-conjugated secondary antibodies. Detection was performed by chemo luminescence. Quantification of protein signals was performed with Fiji. Both Erk as well as pErk levels were corrected for protein content with the loading control β-catenin, after which pErk/Erk levels were calculated.

TGF-β Transcriptional Response Assay

TGF-β response in VSMCs was determined using the (CAGA)12 − MLP − Luciferase promoter reporter construct (Dennler et al., 1998). This construct contains 12 palindromic repeats of the SMAD3/4 binding element derived from the PAI-1 promoter and was shown to be highly specific and sensitive to TGF-β. The assay was performed as described previously (Hawinkels et al., 2014). In short, VSMCs were seeded in 1% gelatin-coated 24-well plates and allowed to attach overnight. Cells, at subconfluent density, were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, California, USA) according to the manufacturer's protocol, with the CAGA-luciferase reporter plasmid together with an SV40 renilla-luciferase plasmid to correct for transfection efficiency. After 6 h medium was changed to DMEM containing 10% FCS | PS and the cells were incubated for 24 h. Cells were serum starved overnight after which cells were washed, lysed and luciferase activity was determined with the dual-glo luciferase assay system according to the manufacturer's protocol (Promega). Luciferase activity was corrected for transfection efficiency with renilla activity. The relative increase in luciferase activity was calculated versus controls. Experiments were performed with three independent cell lines, in duplicate.

Statistical Analysis

For the animal experiments described power analysis was performed with α 0.05 and β 0.80, and taking into account a drop-out rate of 15% due to unforeseen circumstances, which was approved by the local ethical committee. All experiments described were performed blinded by using cell line and mouse numbers without genotypes. Normal distribution of the data was assessed using the Shapiro Wilk test. The unpaired 2-tailed Student t-test was performed to analyze the specific sample groups for significant differences. All results are expressed as mean ± SEM. However, for data with non-normal distribution, log-transformation of the data, followed by the Student t-test, was performed. For the differences between genotypes in aortic diameter, aortic root diameter, aortic length and aortic distensibility (Fig. 1b, c and d, S1 Fig. b), Two way ANOVA was performed in order to account for gender effect. Survival curve analysis (Fig. 1e) was performed with the log rank test (with time to death as the outcome). Cell growth curves (Fig. 7c) were fitted to a nonlinear exponential growth equation after which slopes of the curve were compared. A p-value < 0.05 was considered to indicate a significant difference between groups. In the figures p < 0.05 is shown with *, and p < 0.01 with **. All analyses were performed using IBM SPSS Statistics version 21.0 (SPSS Inc., Chicago, IL, USA) and Graphpad.

Fig. 1

Increased aneurysm size and early death in Smad3−/− mice. a) B-mode ultrasound images of a Smad3−/− female mouse and its Smad3+/+ littermate, age 6 weeks, illustrating aortic diameter measurements. Green line indicates heart, red line the aorta and yellow lines points of measurement. b) Quantification of aortic diameter, c) aortic root and d) aortic length for Smad3+/+ males (n = 10, 10 and 8 respectively), females (n = 9, 9 and 8 respectively) and Smad3−/− males (n = 10, 10 and 6 respectively), females (n = 8, 7 and 7 respectively) at age 6 weeks (*p < 0.05, **p < 0.001, respectively). e) Survival of Smad3−/− compared to Smad3+/+ mice. Smad3−/− females (n = 29) and, even more, males (n = 18), show a severely decreased survival (*p < 0.05, **p < 0.001, respectively). f) Representative macroscopic picture of a female Smad3−/− (right) and Smad3+/+ littermate (left) aorta, age 4 months, clearly showing huge dilatation of the Smad3−/− thoracic aortic wall, however with the same translucency as its Smad3+/+ control. g) Representative HE staining (left) of a Smad3−/− aorta with aneurysm, age 4 months, showing disruption of the vascular wall, compared to Smad3+/+ controls. Alcian Blue staining (middle) of Smad3−/− aortas showed no increase of ECM in the media compared to Smad3+/+ aortas. Picrosirius Red staining for collagen deposition in Smad3−/− and Smad3+/+ aortas showed no differences. Observations made for n = 8 per group. Bar = 50 μm, m = media, a = adventitia, lu = lumen.

Fig. 7

Increased Smad3−/− VSMC proliferation due to decreased TGF-β induced transcriptional activity. a) KI67 staining of Smad3−/− aortic walls showing increased proliferation of VSMCs in the medial layer. b) Curve depicting faster growth of Smad3−/− VSMCs (open symbols) compared to Smad3+/+ (closed symbols). On the y-axis, the relative % of cells, compared to the cells at day 1 of the experiment, is shown. c) Western blots for Erk, pErk and loading control β-catenin of Smad3−/−, Smad3+/+, Fibulin-4+/+ and Fibulin-4 VSMC protein extracts (left). Quantification (corrected for β-catenin levels) shows an increase in pErk/Erk levels in Smad3−/− and Fibulin-4 VSMCs relative to wildtype (right). d) Luciferase TGF-β transcriptional based assay showing a lower transcriptional response in Smad3−/− VSMCs relative to Smad3+/+, and no change in Fibulin-4 VSMCs compared to Fibulin-4+/+. Relative luciferase signal compared to wildtype and corrected for transfection efficiency, is shown. Experiment performed with three independent cell lines. Luc = luciferase, Ren = renilla.

Results

Increased Aneurysm Size and Early Death in Smad3−/− Mice

In order examine the dynamics of aneurysm formation and progression, we set-up a cross-sectional cohort study of Smad3−/− male and female mice together with their Smad3+/+ littermate controls to perform echocardiograms at 6, 12, 18 and 26 weeks of age. Already at the age of 6 weeks, both Smad3−/− male and female mice showed a significant increase in the diameter of the aortic root and ascending aorta compared to their littermate controls. Male Smad3−/− mice showed a 14% increase and female Smad3−/− mice showed a 21% increase in aortic root diameter (Fig. 1a, b and c, p < 0.01). In addition, aortic length, measured from aortic root to the first aortic branch, was significantly increased in Smad3−/− mice compared to their wild type littermates; male Smad3−/− mice showed a 15% increase and female Smad3−/− mice showed a 26% increase in aortic length (Fig. 1d, p < 0.01). Interestingly, we observed a correlation between aortic diameter and aortic length increase for Smad3−/− animals, indicating that the aorta is enlarged in both dimensions (S1 Fig. a). Strikingly, no significant difference in aortic distensibility was detected in these animals (S1 Fig. b), demonstrating no increase in aortic stiffness. Together these data show limited, though significant dilatations and elongation of Smad3−/− aortas at 6 weeks of age.

Smad3−/− mice died suddenly between 6 and 30 weeks of age without overt symptoms, as opposed to none of the Smad3+/+ mice (Fig. 1e). This effect was even more pronounced in male Smad3−/− mice, with 65% mortality for the male animals before 3 months of age (vs. 22% mortality for female mice), implicating a gender difference in life expectancy. Necropsy analysis showed the occurrence of a severe aneurysm consisting of a 2 to 5-fold increase in aortic diameter of the ascending aorta (Fig. 1f). We noticed that the aneurysmatic aortic wall remained translucent, which could be indicative for absence of large-scale ECM remodeling or collagen deposition. HE staining showed altered appearance of smooth muscle cells and apparent changes in aortic wall structure, such as a thicker aortic wall, including a thicker adventitial layer (Fig. 1g). Yet, staining of the aortic wall for Alcian Blue or Picosirius Red showed no differences in proteoglycan and collagen staining, respectively, between Smad3−/− and Smad3+/+ mice (Fig. 1g), which indicates that there is no evidence for increased ECM and collagen accumulation. This is also consistent with the absence of any change in aortic distensibility. Moreover, this is in sharp contrast to what is seen in the Fibulin4 mouse; in this mouse model, the ECM-involved Fibulin-4 gene is expressed at a 4-fold lower level than wildtype, resulting in stiff aortas that lose their translucency and show increased ECM accumulation and elastin disorganization in the aortic wall (Hanada et al., 2007, Moltzer et al., 2011). To better understand the early and sudden death, we next proceeded with the longitudinal studies at older age to investigate aneurysm formation over time.

Rapid Aneurysmal Growth in Smad3−/− Mice, Not Restricted to the Aorta

We performed longitudinal ultrasound studies on both female and male Smad3−/− mice as well as littermate controls (n = 8 and n = 10 per group, respectively) with baseline measurements starting at the age of 6 weeks with intervals of 6 weeks. Consistent with aggressive aneurysmal growth in Smad3−/− animals, there was a drop-out of 4 females and 7 males during the experiment due to sudden death, which, when possible to determine, all presented with a thoracic aneurysm or a cardiac tamponade (Fig. 2a). Of the surviving Smad3−/− animals (4 females and 3 males) 50% of the females and 33% of the males showed a steep increase in aortic diameter in this relative short time span (Fig. 2b and c). No significant increase in aortic diameter was seen in the matched Smad3+/+ littermate controls in this same time span (data not shown). These data show that Smad3−/− animals experience rapid aneurysmal growth, and concurrent early death. Again, this is in contrast to Fibulin-4 animals that are born with an aneurysm, which shows a slow but progressive growth with age (Te Riet et al., 2016).

Fig. 2

Rapid aneurysmal growth in Smad3−/− mice, not restricted to the aorta. a) Table depicting the fate of Smad3−/− mice before 18 weeks, and the possible cause (ND: not determined). b) Graphs depicting aortic diameter in time for Smad3−/− females (left, n = 4) and Smad3−/− males (right, n = 3). c) B-mode ultrasound images of a Smad3−/− female mouse and its littermate, illustrating the huge increase in aortic diameter within 12 weeks of time. Green line indicates heart, red line the aorta. d) Representative CT pictures of a Smad3−/− female (right) and its Smad3+/+ littermate (left), age 4 months, in frontal and side view. Arrows indicate the aortic aneurysm in frontal view (white), enlargement of the jugular vein (yellow), and vena cava inferior (blue), for the Smad3−/− mouse. Same sites are indicated in the Smad3+/+ animal. Bone abnormalities such as kyphosis are also apparent in the Smad3−/− mouse.

To investigate vascular changes throughout the circulatory system we performed μCT-scans with an Iodine based contrast agent to visualize soft tissues in Smad3−/− and Smad3+/+ mice (n = 3 per genotype). Remarkably, these μCT scans not only revealed a massive dilatation of the aorta and increased heart size, but also enlargements in different locations of the body such as in the jugular vein and the vena cava inferior (Fig. 2d, white, yellow and blue arrows, respectively). However, dilatation of the jugular vein might also be the consequence of cardiac failure. These data indicate that in mice, Smad3-related disease is a widespread and aggressive pathology not restricted to the aorta, similar as previously observed in human (AOS) patients (Aubart et al., 2014, Martens et al., 2013, Van de Laar et al., 2012, Regalado et al., 2011). The μCT scans of Smad3−/− animals also showed skeletal abnormalities, such as kyphosis, as previously described (Li et al., 2006) and illustrated in our Supporting video files, showing the 3D rotation for both a Smad3−/− and Smad3+/+ mouse (Supporting info files).

Smad3−/− Aortic Aneurysms Show Elastin Disruption, and Increased Immune Response, pSmad2 and pErk

Histomorphological analysis of five Smad3−/− aortic walls with aneurysms showed changes predominantly in the ascending aorta, consisting of focal disruption and thinning of the media (Fig. 3a). Three aortas with aneurysms showed marked cystic spaces in the medial wall (Fig. 3a), some of which contained erythrocytes, which was not seen in Smad3+/+ mice (data not shown). We stained for CD31, an endothelial marker, that aligns the endothelial wall of vessels (indicated with an arrow in Fig. 3a and b), demonstrating that these cystic spaces actually represented small capillaries in the adventitia of Smad3−/− aortic walls. Occurrence of these capillaries could indicate (neo)vasculogenesis, possibly as a response to vascular damage (Fig. 3b). In line with this observation, disruption of the media structure was seen in conjunction with focal adventitial inflammation, fibrosis and granulation tissue (Fig. 3a). Interestingly, immune infiltrations have not been observed in Fibulin-4 animals.

Fig. 3

Aneurysmal Smad3−/− aortas show an increased immune response, elastin disruption, pSmad2 and pErk activation. a) HE staining of a Smad3−/− aorta with aneurysm showing disruption of the vascular wall, adventitial inflammation, and cystic spaces (example indicated with *). b) CD31 staining shows the endothelial layer (indicated by arrow), but also alignment of the observed cystic spaces (example indicated with *), showing that these spaces represent vessels in the adventitia. c) Resorcin Fuchsin staining for elastin showed disrupted elastin structures in the medial layer in aneurysmal Smad3−/− compared to Smad3+/+ aortas. d) Smooth Muscle Actin staining of aneurysmal Smad3−/− aortic walls showed no apparent loss of VSMCs in the media comparable to Smad3+/+. e) pSmad2 and f) pErk showed increased staining in the aortic wall of aneurysmal Smad3−/− compared to Smad3+/+ mice. This suggests an increased TGF-β receptor activation. Observations made for n = 8 per group, age 3–4 months old. Bar = 50 μm, m = media, a = adventitia, lu = lumen. *; cystic space, arrow; endothelial layer.

Elastin staining clearly showed disruption in the elastin structure of the aortic wall (Fig. 3c). Strikingly, this differs from Fibulin-4 aortic walls, where accumulation and disorganization of elastin structure was observed (Hanada et al., 2007, Moltzer et al., 2011). At sites where the media was still present, SMA staining did not show apparent loss of VSMCs in Smad3−/− aortic walls (Fig. 3d). To investigate the impact of Smad3 deficiency on the TGF-β signaling pathway, we performed pSmad2 and pErk immunohistochemical staining on Smad3−/− and Smad3+/+ aortic walls. These are markers for canonical (pSmad2) and non-canonical (pErk) activation of the TGF-β signaling pathway via the TGF-β receptor. pSmad2 and pErk are activated in the Fibulin-4 mouse model, in other aneurysmal diseases such as Marfan's syndrome and other TGF-β pathway related aneurysmal diseases (Hanada et al., 2007, Loeys et al., 2005, Neptune et al., 2003, Renard et al., 2013, Renard et al., 2010). Both pSmad2 as well as pErk staining (Fig. 3e and f), showed an increase in the aorta of Smad3−/− animals, indicative of increased TGF-β receptor activation.

Inflammation-associated MMP Activation in Aneurysmal Smad3−/− Mice

An important process that is regulated by the TGF-β signaling pathway is MMP activation. In aneurysmal diseases such as Marfan's syndrome, as well as aneurysmal mouse models, including Fibulin-4 and Fibrillin-1 mutant mice, MMP activity is strongly increased (Lemaitre et al., 2003, Longo et al., 2002, Segura et al., 1998, Kaijzel et al., 2010, Chung et al., 2007). To examine MMP activation in Smad3−/− aortas, we used molecular imaging with MMPsense near-infrared probes to monitor MMP activity both in vivo and ex vivo. No significant difference in MMP activation between Smad3−/− (with no or minor aneurysm formation) and Smad3+/+ aortas was observed (Fig. 4a, b and c). Thus, also suggesting that an increase in MMP may not always correlate with functional upregulation of the TGF-β signaling pathway. Interestingly, only 25% of aneurysmatic Smad3−/− aortas did seem to show an increase in MMP activity (Fig. 4c and d). This is in contrast to Fibulin-4 animals that show increased MMP activity already at a young age in VSMCs of the aortic wall (Kaijzel et al., 2010). Strikingly, in Smad3−/− animals this activation was derived from invading immune cells and not from VSMCs in the medial layer, as evident from immunohistochemical staining for MMP, CD3 (T-cells) and MAC2 (macrophages) (Fig. 4e). Thus, since this MMP activation was not seen in the VSMCs, this indicates impaired TGF-β signaling downstream of Smad3 in VSMCs specifically, leading to a lack of MMP activation and therefore no ECM remodeling. Together with the finding that we do not see changes in ECM composition (staining Fig. 1g), this can explain the fast expansion and rupture of Smad3−/− aortas.

Fig. 4

Inflammation-associated MMP activation in aneurysmal Smad3−/− mice. a) Representative in vivo images of male Smad3−/− and Smad3+/+ aortas showing comparable MMP activation. b) Representative ex vivo images of male Smad3−/− and Smad3+/+ aortas showing similar MMP activation. c) Quantification of the fluorescent signal, a measure for MMP activity, does not show significant differences between Smad3−/− and Smad3+/+ aortas. Values for each Smad3−/− aorta are compared to its Smad3+/+ littermate. Both male and female aortas were used for quantification (n = 8 total). The two Smad3−/− aortas with highest relative fluorescence intensity also had an aneurysm. d) Ex vivo images of male Smad3−/− aortas with an aneurysm occasionally do show increased MMP activation. e) Immunohistochemical staining for MMP9, T cells (CD3) and macrophages (MAC2) in an aneurysmal Smad3−/− aortic wall, showing that the MMP9 signal is derived from immune cells and not VSMCs (performed for n = 5–7 animals per group, age 3–4 months old). bar = 100 μm, m = media, a = adventitia, lu = lumen.

Pre-aneurysmal Smad3−/− Animals Show Elastin Disruption and pErk Activation

As it seemed that specifically Smad3−/− aortas with an aneurysm showed increased MMP activity, which might be derived from immune cells infiltrating the aortic wall, we next wondered which histological changes may already occur in pre-aneurysmal Smad3−/− animals. HE staining of aortic walls derived from these Smad3−/− animals showed morphological changes in VSMC appearance; a more round, less flattened appearance compared to Smad3+/+ aortas (Fig. 5a). Interestingly, the elastin structure was disrupted at several places in pre-aneurysmal Smad3−/− aortic walls, whereas SMA staining at these same sites showed no apparent smooth muscle cell loss (Fig. 5b and c). Surprisingly, pSmad2 and pErk staining were both already increased in the aortas of these pre-aneurysmal Smad3−/− animals (Fig. 5d and e), although the increased pErk staining was much more pronounced. However, at these sites of increased structural wall damage and increased pErk staining no apparent immune infiltration was observed (Fig. 5a) In agreement, staining for CD31 and for MAC2 were negative (data not shown). Together this indicates that both the damage in the aortic wall and the canonical and non-canonical TGF-β signaling activation, activated pSmad2, and pErk, precede the structural damage in the medial layers that might trigger the immune response.

Fig. 5

Pre-aneurysmal Smad3−/− animals show elastin disruption, pSmad2 and pErk activation. a) HE staining of pre-aneurysmal Smad3−/− aortas showing altered appearance of VSMCs in the medial layer. b) Smooth Muscle Actin staining of pre-aneurysmal Smad3−/− aortic walls showed no apparent VSMC loss in the media comparable to Smad3+/+ aortas. c) Resorcin Fuchsin staining for elastin showed a somewhat disrupted elastin structure in the medial layer of pre-aneurysmal Smad3−/− aortas compared to Smad3+/+. d) pSmad2 staining is already somewhat increased in the aortic wall of pre-aneurysmal Smad3−/− compared to Smad3+/+ mice. e) pErk clearly showed increased staining in the aortic wall of pre-aneurysmal Smad3−/− compared to Smad3+/+ mice. Observations made for n = 5 per group, age 3–4 months old. Bar = 50 μm, m = media, a = adventitia, lu = lumen.

Smad3−/− VSMCs and Aortas Show No Downstream Transcriptional Activation of TGF-β Signaling

Although Smad3 is not expressed, we find an increase in nuclear located pSmad2 and pErk activation. As this increased activation normally affects downstream transcriptional activation, as was shown before in Fibulin-4 aortas (Ramnath et al., 2015), we next decided to investigate the transcription of genes that should normally be activated in response to an increase in TGF-β receptor activation. In particular, next to the aorta itself, we were interested in the transcriptional response in VSMCs isolated from the aorta, as there we can exclude the contribution of other cell types such as immune cells or fibroblasts that are clearly present in the aneurysmal aortas of Smad3−/− animals. We therefore isolated VSMCs from Smad3−/− and Smad3+/+ aortas, and confirmed the smooth muscle cell phenotype with SMA staining (Fig. 6a). Next, we determined the mRNA expression levels of FN-1, PAI-1, Smad7 and Smad6 (Fig. 6b–e), together with TIMP-1, LTBP-1, TGFβR1, TGFβR2 and TGFβ1 (S2 Fig. a–e) in aortic extracts as well as in aortic VSMCs of Smad3−/− and Smad3+/+ animals. We found a significant decrease in FN-1, PAI-1 and Smad7 transcription in Smad3−/− aortic extracts as well as VSMCs compared to Smad3+/+ (Fig. 6b and c). Transcription of Smad6 and TIMP-1 was decreased in Smad3−/− compared to Smad3+/+ VSMCs and aortas (Fig. 6d, e and S2 Fig. a), and no significant change in LTBP-1 mRNA levels was detected (S2 Fig. b). Interestingly, transcripts for TGFβR1, TGFβR2 and TGFβ1 were significantly decreased in Smad3−/− VSMCs compared to Smad3+/+, but not in the aortic arch tissue, which could imply that this decrease in TGFβ receptors and TGFβ1 occurs specifically in VSMCs and not in immune (or other types of) cells (S2 Fig. c, d and e). These findings are different from Fibulin-4 aortas where increased TGFβ receptor activation and downstream transcriptional activation of genes such as TGFβ2 and PAI-1 were observed (Ramnath et al., 2015).

Fig. 6

Smad3−/− VSMCs and aortas show no downstream transcriptional activation of TGF-β signaling. a) Pictures showing Smad3+/+ and Smad3−/− VSMCs isolated from the aorta, confirmed by SMA staining. Real-time PCR analysis on VSMC and aorta mRNA shows downregulated mRNA levels of b) Fn-1, c) Pai-1, d) Smad7, and e) Smad6, in Smad3−/− VSMCs and aortic extracts compared to Smad3+/+. Fold changes are shown for Smad3−/− relative to Smad3+/+, *p < 0.05, **p < 0.01. The mean of three independent experimental means is shown, n = 9–12 per group.

Increased Smad3−/− VSMC Proliferation due to Decreased Downstream Transcriptional Activation

Interestingly, in the Fibulin-4 mouse model VSMCs have a higher downstream TGF-β-activated transcriptional response, which is linked to their reduced proliferation rate, and which can be rescued by inhibiting TGF-β (Ramnath et al., 2015). In contrast, during the culturing of Smad3−/− VSMCs we noticed that they appeared to proliferate faster than Smad3+/+ VSMCs. We therefore performed Ki67 staining on the aorta, which specifically marks proliferative cells, and found an increase in VSMC proliferation in Smad3−/− aortic walls compared to Smad3+/+ (Fig. 7a) as judged from the positive KI-67 staining. We next set up a controlled VSMC proliferation experiment, where we started with the same amount of cells for each genotype and counted the cells each day for seven days. Indeed, Smad3−/− VSMCs proliferated faster than Smad3+/+ VSMCs (Fig. 7b). This is different from the Fibulin-4 VSMCs, which proliferate slower than their wild type control cells (Ramnath et al., 2015). As we observed prominent increased pErk staining in both pre-aneurysmal as well as aneurysmal Smad3−/− aortas, we next investigated Erk phosphorylation status in both Smad3−/− as well as Fibulin-4 VSMCs. In both Smad3−/− and Fibulin-4 VSMCs pErk was increased compared to their control cell lines (Fig. 7c), which hence does not explain the increased proliferation observed in Smad3−/− VSMCs. However, we did observe an impaired downstream TGF-β transcriptional response in Smad3−/− VSMCs (Fig. 6). We therefore next decided to measure the transcriptional response in isolated Smad3−/− and Fibulin-4 VSMCs with a luciferase-based reporter assay as previously described (Dennler et al., 1998, Hawinkels et al., 2014, Ramnath et al., 2015). In this assay, a plasmid containing the SMAD3/4 binding element derived from the PAI-1 promoter is transfected, that is highly specific and sensitive to TGF-β receptor activation. We measured the transcriptional response without adding exogenous TGF-β. Although we observed no significant change in transcriptional response for Fibulin-4 compared to Fibulin-4 VSMC, Smad3−/− VSMCs showed a significantly 5-fold lower response than Smad3+/+ VSMCs, indicating impaired downstream activation of TGF-β responsive genes (Fig. 7d, p < 0.05). These data imply that the impaired downstream TGF-β induced transcriptional activation affects cellular fate such as proliferation.

Discussion

In humans, SMAD3 mutations cause a syndrome with cardiovascular, craniofacial, cutaneous and skeletal anomalies. Aortic aneurysms and osteoarthritis characterize this syndrome known as AOS or LDS3. Although an increasing number of SMAD3 mutations have recently been reported (Aubart et al., 2014, Fitzgerald et al., 2014, Hilhorst-Hofstee et al., 2013, Zhang et al., 1996), the functional effect of these mutations remains to be explored. The molecular mechanism underlying AOS, resulting from SMAD3 mutations is largely unknown. In this study we investigated the vascular phenotype in Smad3 mutant mice and explored its effects on the TGF-β signaling cascade.

As the functional consequence of heterozygous SMAD3 patient mutations at the protein level is unclear; it might for example lead to a null or non-functional protein (e.g. not able to bind SMAD2), we decided to study the effect of Smad3 deletion in Smad3−/− mice to better understand its role in aneurysm formation. In Smad3−/− mice, Smad3 deficiency leads to aortic aneurysms, which mainly affect the aortic root and ascending aorta, but also extends to other major arteries. Aneurysm formation is highly dependent on the age of the animals, as has been observed in AOS patients with SMAD3 mutations, which display an age-dependent penetrance. As shown here, male Smad3−/− animals suffered a more severe aortic phenotype (both dilatation and elongation) and consequently, died earlier than female Smad3−/− animals. The resulting phenotype due to Smad3 deficiency in mice is very similar to the vascular phenotype described in patients with SMAD3 mutations offering an excellent opportunity for disease modeling. Despite the homogeneous genetic background aneurysms still present irregularly in the Smad3−/− mice, which suggests that other external factors, like for example blood pressure and subtle intrinsic variations in transcriptional activation can also determine the variability in the resulting phenotype.

Strikingly, histology of the aortic wall shows that, unlike other models for aneurysmal disease, there is no increase in ECM accumulation, nor excessive collagen staining, or loss of vascular smooth muscle cells. However, we observe increased aortic pSmad2 and pERK activation. Moreover, this activation is already apparent before Smad3−/− animals present with aneurysms, showing that this activation precedes aneurysm formation. Importantly, these changes in the Smad3−/− aortic wall are distinct from what was previously observed for Fibulin-4 animals, which have reduced expression of the ECM protein Fibulin-4, and as a result also show aneurysm formation. Instead, Fibulin-4 aortas show increased ECM remodeling, and increased collagen and elastin structures (Hanada et al., 2007, Moltzer et al., 2011). Yet, they also show increased pSmad2 and pERK activation. A comparison of the Smad3−/− and Fibulin-4 changes in aneurysm formation is shown in Fig. 8. These findings led us to hypothesize that TGF-β signaling downstream of Smad3 might be deregulated due to Smad3 deficiency.

Fig. 8

Comparison of aneurysm formation, and model of differences in TGF-β signaling between Fibulin-4 and Smad3 deficient VSMCs. a) Summary of differences between aneurysm formation in Fibulin-4 and Smad3 deficient animals. b) Under normal conditions the TGF-β receptor is activated by TGF-β ligand (orange diamond), leading to phosphorylation of Smad2 (pSmad2), which, in complex with Smad3, translocates to the nucleus and induces a transcriptional response (left). One of the genes transcribed is Smad7, a potent inhibitor of upstream TGF-β receptor signaling, thereby providing a regulatory negative feedback loop. In case of Fibulin-4 deficiency, more TGF-β ligand is released from the matrix, leading to increased TGF-β receptor stimulation and Smad2 phosphorylation, resulting in increased downstream transcriptional activation (right). In Smad3 deficient cells the TGF-β receptor can still be activated by TGF-β ligand, and Smad2 can still be phosphorylated. However, due to absence of Smad3, the Smad2/3 complex will not translocate to the nucleus together, thereby not activating the normal downstream transcriptional program. In case of Smad7 and TGF-β this results in reduced expression. As Smad7 is a potent inhibitor of the TGF-β signaling pathway, when absent or reduced, TGF-β signaling remains activated instead of inhibited. This together could lead to an altered transcriptional program in Smad3 deficient VSMCs (middle).

While AOS and LDS patients have heterozygous loss of function mutations affecting genes from the TGF-β signaling pathway (TGFΒR1/2, SMAD3, TGFB2, TGFB3), aortic tissues of these patients show a signature of upregulated TGF-β pathway signaling, as indicated by the overexpression of pSMAD2, pERK1/2 and CTGF (Loeys et al., 2005, Van de Laar et al., 2011, Lindsay et al., 2012). This phenomenon is known as the TGF-β paradox (for comprehensive review see (Massague, 2012)). We therefore examined TGF-β receptor activation in the aortic wall of Smad3−/− animals by investigating pSmad2 expression. We observed that while lacking Smad3, Smad2 can still be phosphorylated and transported to the nucleus. Thus, lack of Smad3 might alter downstream transcriptional activation in the nucleus (also see Fig. 8). This lack of Smad3 might also shift TGF-β activated signaling via Smad2, which has similar but not overlapping functions (Moustakas and Heldin, 2002, Zhang et al., 1996), also resulting in an altered transcriptional response. We reasoned that these changes lead to reduction in the downstream transcription of genes such as ECM components and MMPs. Indeed, we demonstrate here that transcriptional activation of multiple genes downstream of the TGF-β signaling cascade is absent. Again, this is in contrast to what was previously found for Fibulin-4 aortas and VSMCs where downstream TGF-β-induced transcription was activated (Ramnath et al., 2015). Moreover, we found that the increased proliferation of Smad3−/− VSMCs is not due to increased pErk, but rather due to the lack of downstream TGF-β induced transcriptional activation. Moreover, it shows that Smad3 normally plays a role in TGF-β mediated growth inhibition.

We also examined Smad7 as an important regulator of the TGF-β signaling pathway as it inhibits phosphorylation of Smad2 by the TGF-β receptors, thereby acting as a negative feedback loop in this pathway to prevent deregulation. Without Smad7, TGF-β receptor activation cannot be inhibited -explaining for example increased pSmad2- but downstream transcriptional activation remains impaired since Smad3 is lacking (See Fig. 8, left panel for a schematic overview). In agreement, we found no upregulation of Smad7 in Smad3−/− aortas, or of other downstream genes such as PAI-1.

When we now compare the findings between Smad3−/− and Fibulin-4 animals, although in both mouse models aneurysms are formed, the underlying cause and phenotypical consequences are quite different (Fig. 8). In Fibulin-4 aortas it is thought that TGF-β is released from the matrix, resulting in activation of TGF-β signaling pathway, which eventually also results in activated downstream transcription. However, in Smad3−/− animals, although there can be activation of TGF-β receptors, as also shown by increased pSmad2 signal, transcriptional activation downstream of Smad3 is impaired. One important downstream gene that is normally transcribed is SMAD7, which is a potent inhibitor of the TGF-β signaling pathway. We find decreased Smad7 transcripts, which might explain the continued activation of the TGF-β receptor. Furthermore, this altered signaling and transcriptional pattern might also lead to an altered ‘appearance’ of the VSMCs for the immune system, thereby attracting immune cells. In accord, Smad3−/− animals suffered from so-called ‘sterile’ infections in the aortic wall, as derived from the HE stained sections showing infiltrations of immune cells. As these infiltrations were only observed after dilatation of the aorta became apparent, this might suggest that this immune reaction is triggered after the aneurysm is formed. Indeed, Ye et al. showed that administration of anti-GM-CSF antibody to Smad3−/− mice reduced inflammation and also diminished aorta dilation (Ye et al., 2013). This would also be in agreement with the fact that we only observed MMP activity in aortas with an aneurysm that could be traced back to the immune cells adjacent to the adventitia of the aortic wall. Moreover, the fact that here the MMP activation is derived from immune cells rather than from VSMCs, and is only present at late stages of aneurysm formation, after the immune infiltrates are present, would argue that MMP activity is a good marker for aneurysm formation caused by ECM deficiencies, but not for those due to SMAD3 mutations. The longitudinal echocardiograms showed dilatations with a rapid increase of the aneurysm within a very short period of time. It could be that the increased upstream TGF-β receptor activation, together with the lack of collagen and ECM accumulation results in dilatations; the structural integrity fails progressively. This, together with a possibly altered appearance of the VSMC both in structure as well as transcriptional profile, could attract immune cells, which start cleaning up the detected ‘vascular damage’ at the expense of macrophage-induced deterioration of the already fragile aortic wall. This would then explain the rapid and aggressive aneurysmal growth.

Most genetic studies have been focusing on delineating the clinical phenotype associated with SMAD3 mutations (Hilhorst-Hofstee et al., 2013, Aubart et al., 2014, Fitzgerald et al., 2014, Wischmeijer et al., 2013, Van der Linde et al., 2012). Despite the increasing number of reported mutations, functional studies indicating the pathological effect of these mutations are lacking. Existing experimental data and molecular predictions suggest that SMAD3 mutations are mainly loss of function (Van de Laar et al., 2011, Aubart et al., 2014). Many reported mutations lead to frame shifts, deletions and likely nonsense mediated decay. Others are perturbing the heterodimer formation SMAD3/4 or leading to nonfunctional complexes. Yet, a dominant-negative effect of some mutations cannot be excluded.

The finding of an aortic phenotype in a mouse model with a complete Smad3 deficiency very similar to the human disease supports the idea that also lack of functional SMAD3 could cause the human clinical phenotype. A similar scenario has been described for the TGF-β2 mouse model (Boileau et al., 2012). Moreover, in aneurysmal diseases such as Marfan's syndrome (MFS), the efficacy of interventions that target the TGF-β signaling pathway is being explored. So far the effects on delay of aneurysmal growth are quite promising (Neptune et al., 2003, Ng et al., 2004, Habashi et al., 2006, Cohn et al., 2007). However, similar intervention strategies might not be beneficial in case of a Smad3 deficiency. Since downstream transcriptional activation is hampered in the absence of Smad3, inhibition of components in the TGF-β signaling pathway might in this case worsen the outcome as even less ECM would be generated, and alternative ‘escape’ pathways would be blocked.

In conclusion, Smad3 deficiency leads to aortic aneurysms and sudden death in the Smad3 knockout animal model. This phenotype is influenced by age and gender of the animals. Although Smad3 is absent, we observed increased nuclear translocation of pSmad2, and upregulated pERK signaling, inferring increased upstream TGF-β receptor activation. However, the downstream TGF-β-activated transcriptional response seemed impaired as derived from the absence of MMP activation and lack of amorphous ECM accumulation in Smad3−/− mouse aortas. Together our data stress the importance of identifying the molecular mechanism of aneurysmal disease, as the outcome, and therefore treatment options, can differ dramatically. At the same time, the Smad3−/− mouse proves to be an ideal model to start testing these different interventional options on.

The following are the supplementary data related to this article.

Supplementary figures

Supplementary videos

Video files supplement to Fig. 2d. Rapid aneurysmal growth in Smad3−/− mice, not restricted to the aorta. d) Representative μCT video files of CT pictures of a Smad3−/− female (3D CT rendering Smad3ko mouse) and its littermate Smad3+/+ control (3D CT rendering WT mouse) in 3D rotating view.

Funding Sources

This study was funded by ‘Stichting lijf en leven’ (project: dilating versus stenosing arterial disease, 2011–2015), and partially funded by an Erasmus Fellowship (2009) to AM Bertoli-Avella. Funders had no role in study design, data collection, data analysis, interpretation or writing of the manuscript.

Conflicts of Interest

None.

Author contributions

I.P. experimental design, data analysis, figure design, writing of manuscript.

N.V. performed experiments, data analysis, figure design.

J.H.T. data analysis, writing.

J.L.R data analysis, writing.

Y.R. performed experiments, data analysis.

P.M.H. performed experiments, data analysis.

B.S.T. performed experiments, data analysis.

M.V. performed experiments.

R.M.G.B.B.-O. performed experiments.

S.E.H statistical analysis, writing.

H.J.M.V supervision of statistical analysis.

R.K. experimental design, writing.

A.M.B-A. experimental design, writing.

J.E. experimental design, data analysis, writing.

	Forward primer	Reverse primer	Size (bp)	Temp (°C)
B2M	5′-CTCACACTGAATTCACCCCCA-3′	5′-GTCTCGATCCCAGTAGACGGT-3′	98	55/60
Tbp1	5′-TCACTCCTGCCACACCAGCTTC-3′	5′-TGACTGCAGCAAATCGCTTGGG-3′	156	60
Ppia	5′-CGCGTCTCCTTCGAGCTGTTTG-3′	5′-TCCGTAGATGGACCTGCCGC-3′	186	60
Ppia	5′-GTCTCCTTCGAGCTGTTTGC-3′	5′-ACCACCCTGGCACATGAATC-3′	138	55
Smad7	5′-CAAACCAACTGCAGGCTGTC-3′	5′-CCCCAGGGGCCAGATAATTC-3′	77	55
Smad6	5′-GCCACTGGATCTGTCCGATT-3′	5′-GGTCGTACACCGCATAGAGG-3′	188	55
Pai-1	5′-TCTTTTTCACATTACAGTGGCCTG-3′	5′-TTTGGGTGACTCTGTTAATTCATC-3′	102	55
Fn1	5′-ACGGACATCTGTGGTGTAGC-3′	5′-CGAGTCTGAACCAAAACCGC-3′	91	55
Timp-1	5′-TCGGACCTGGTCATAAGGGC-3′	5′-GCTTTCCATGACTGGGGTGT-3′	162	55
Ltbp1	5′-CCAAACATGGCAGGCAAGTC-3′	5′-TCCACAGACGTTGATCCCCT-3′	116	55
Tgfb1	5′-GTGGACCGCAACAACGCCATCT-3′	5′-GCAATGGGGGTTCGGGCACT-3′	109	60
Tgfbr1	5′-TCGTCCGCAGCTCCTCATCGT-3′	5′-ACACTGTAATGCCTTCGCCCCC-3′	70	60
Tgfbr2	5′-CGCACGTTCCCAAGTCGGATGT-3′	5′-TCGCTGGCCATGACATCACTGT-3′	118	60