Fibulin-4 deficiency results in ascending aortic aneurysms: a potential link between abnormal smooth muscle cell phenotype and aneurysm progression

Abstract

Rationale

Loss of fibulin-4 during embryogenesis results in perinatal lethality due to aneurysm rupture, and defective elastic fiber assembly has been proposed as an underlying cause for the aneurysm phenotype. However, aneurysms are never seen in mice deficient for elastin, or for fibulin-5, which absence also leads to compromised elastic fibers.

Objective

We sought to determine the mechanism of aneurysm development in the absence of fibulin-4 and establish the role of fibulin-4 in aortic development.

Methods and Results

We generated germline and smooth muscle cell (SMC)-specific deletion of the fibulin-4 gene in mice (Fbln4^GKO and Fbln4^SMKO, respectively). Fbln4^GKO and Fbln4^SMKO aortic walls fail to fully differentiate, exhibiting reduced expression of SM-specific contractile genes and focal proliferation of SMCs accompanied by degenerative changes of the medial wall. Marked upregulation of ERK1/2 signaling pathway was observed in the aneurysmal wall of Fbln4^GKO and Fbln4^SMKO mice and both mutants developed aneurysm predominantly in the ascending thoracic aorta. In vitro, Fbln4^GKO SMCs exhibit an immature SMC phenotype with a marked reduction of SM-myosin heavy chain and increased proliferative capacity.

Conclusion

The vascular phenotype in Fbln4 mutant mice is remarkably similar to a subset of human thoracic aortic aneurysms caused by mutations in SMC contractile genes. Our study provides a potential link between the intrinsic properties of SMCs and aneurysm progression in vivo and supports the dual role of fibulin-4 in the formation of elastic fibers as well as terminal differentiation and maturation of SMCs in the aortic wall.

Introduction

Aortic aneurysms are characterized by abnormal enlargements of the aorta caused by global or localized weakness of the vessel wall. Traditionally, aneurysm formation is believed to involve defects in synthesis or assembly of extracellular matrix (ECM) proteins or accelerated degradation of vascular ECM (reviewed in ^{1, 2}). For example, mutations in the fibrillin-1 (FBN1) or type III alpha 1 collagen gene (COL3A1) genes are responsible for Marfan syndrome and vascular Ehlers-Danlos syndrome ^{3, 4}, respectively. Mutations in Fbn1 profoundly affect the formation of microfibrils that surround an elastin core and are required for proper assembly of elastic fibers ⁵. Genetically-engineered mice provide further evidence that the vascular ECM is the critical determinant of aneurysm development, including an allelic series of Fbn1 mutant mice, and mice homozygous for null alleles of the genes encoding lysyl oxidase, a crosslinking enzyme for elastin and collagen, and ECM proteins such as fibulin-4 (Fbln4) and biglycan (Bgn) ^6-9. In addition to the disruption of ECM in the vessel wall, alterations of SMC intracellular contractile proteins have recently been shown to be responsible for subsets of thoracic aneurysms in humans ^10-12. Two heterozygous mutations in MYH11 (SM-MHC) were identified in kindreds presenting with thoracic aortic aneurysms and/or aortic dissections (TAADs) and patent ductus arteriosus. Heterozygous missense mutations in ACTA2 (SM α-actin) were also found in 14% of inherited TAADs ¹². Thus, the pathogenesis of aortic aneurysm formation may involve abnormalities in homeostasis of vascular ECM proteins or abnormal SMC development, or both.

Fibulin-4 belongs to the fibulin family of ECM proteins consisting of seven known members characterized by tandem repeats of calcium-binding epidermal growth factor (cbEGF)-like motifs and a C-terminal fibulin module ¹³. In the aorta, fibulin-5 is expressed at one magnitude higher than fibulin-4 and its absence leads to compromised elastic fibers with aggregates of elastin, but does not develop aneurysms. ^{14, 15}. Fbln4-null (Fbln4^-/-) mice, in contrast, exhibit a more severe phenotype with perinatal lethality due to rupture of aortic aneurysms and marked disruption of elastic fibers ⁸. Genetic mutations in the FBLN4 have also been identified in two human cutis laxa patients, both of which exhibited aortic aneurysms ^{16, 17}. Defective elastogenesis was proposed to be an underlying cause of aortic aneuryms in Fbln4^-/- mice, however, elastin-null mice (Eln^-/-) do not develop aortic aneurysms, rather they exhibit stenosis of the aorta due to subendothelial proliferation of SMCs ¹⁸. It is not clear, therefore, whether the phenotypic differences between Fbln4^-/- and Fbln5^-/- mice are due solely to the difference in severity of elastic fiber defects. It is possible that fibulin-4 has additional function(s) that act independent of, or in concert with, elastogenesis in the developing aorta. Recently, hypomorphic Fbln4^R/R mice were shown to contain increased phosphorylated (p)-Smad2, enhanced production of CTGF and collagen fibers, and increased proliferation of SMCs in the aortic wall ¹⁹. However, the precise mechanism of upregulation of TGF-β in Fbln4^R/R mice or whether it is a primary cause of the vascular phenotype is unclear, especially since SMC-specific ablation of TGF-β receptor type 2 causes aortic aneurysms by down-regulating elastogenic genes ²⁰.

To determine the role of fibulin-4 in aneurysm formation, we generated mice with a germline deletion or vascular cell-specific deletion of Fbln4. SMC-specific, but not endothelial cell (EC)-specific knockout of Fbln4, resulted in the formation of large aneurysms exclusively in the ascending aorta. We found that SMCs failed to fully differentiate in the absence of fibulin-4 in the aorta as evidenced by down-regulation of SMC-specific contractile protein genes and an immature phenotype of Fbln4-null SMCs accompanied by focal hyperproliferation of SMCs and degenerative changes affecting the integrity of the vessel wall. These data highlight a novel role for fibulin-4 in the development of SMCs, and together with its role in the formation of elastic fibers, fibulin-4 functions to protect the aortic wall against aneurysm formation in vivo.

Materials and Methods

Mouse

Generation of conditional knockout mice is described in the expanded Materials and Methods sections online.

Mechanical stretch

Two or 4 × 10⁵ cells were plated on 6-well silicon elastomer-bottomed culture plates covalently coated with type I collagen (Flexcell). After allowing cells to adhere to the plates (6-12 h), the media was changed to serum free and incubated for 24 h to synchronize the cell cycle prior to initiation of stretch stimulation. Cells were subjected to unidirectional cyclic strain (0.2 Hz and 20% strain) for 24 h in a humidified incubator with 5% CO₂ at 37°C using FX-4000T Flexcell Tension Plus system (Flexcell International). Cells incubated in static condition served as controls.

Standard protocol was used for immunostaining, Western blotting, qPCR, gelatin zymography, histological and electron microscopic analysis and is available online.

Results

Germline and SMC-specific deletions of Fbln4 in vivo

To investigate aneurysm development in the absence of fibulin-4, we generated a conditional allele of Fbln4 using Cre-loxP technique in ES cells (Online Figure IA). Fbln4^neo-loxP mice were mated with transgenic mice expressing the heat stable Flp-recombinase to generate mice homozygous for Fbln4^loxP/loxP (Online Figures IB, C). Fbln4^neo-loxP mice were also mated with CAG-Cre transgenic mice to delete Fbln4 in germline, and subsequently homozygous mice for null alleles were generated (termed Fbln4^GKO). In concordance with the previous reported data ⁸, all Fbln4^GKO mice died in or after parturition due to aneurysm rupture (Online Figure ID) and quantitative RT-PCR (qPCR) confirmed absence of Fbln4 in the aorta of Fbln4^GKO mice (Online Figure IE).

Fbln4 is ubiquitously expressed during embryogenesis and is detected in both ECs and SMCs in the aorta (Online Figures IIA, IIB). To determine the cell type responsible for the aneurysm phenotype, we generated EC-specific and SMC-specific knockout mice for Fbln4 by mating Fbln4^loxP/loxP mice with transgenic mice expressing Cre-recombinase under a control of the Tie2 promoter and SM22α promoter, respectively ^{21, 22}. EC-specific knockout mice (Fbln4^loxP/GKO;Tie2-Cre, termed Fbln4^ECKO) were healthy and indistinguishable from control littermates. Histological finding of the aorta was normal (Online Figure III). In contrast, SMC-specific knockout mice (Fbln4^loxP/GKO;SM22-Cre, termed Fbln4^SMKO) began to die spontaneously at approximately 2 months of age despite absence of embryonic or neonatal lethality. Aortae harvested from adult Fbln4^SMKO mice exhibited large aneurysms exclusively in the ascending aorta (Figure 1A). Aneurysms were observed in Fbln4^SMKO mice with complete penetrance. The ductus arteriosus was closed and the ligamentous remnant was markedly elongated. qPCR confirmed a significant reduction of Fbln4 transcripts in the ascending aorta where the aneurysms developed, as well as in the descending thoracic aorta where only elongation was evident (Figure 1B). These data indicate that loss of SMC-derived fibulin-4 is responsible for the aneurysm formation and that fibulin-4 derived from other cell types compensated for aneurysm formation during embryonic life and perinatal lethality in Fbln4^SMKO mice.

Figure 1.

Aneurysms and tortuous aorta in Fbln4^SMKO mice. A. Ascending thoracic aortic aneurysm in Fbln4^SMKO mouse at P90. B. qPCR analysis of Fbln4 transcripts in the ascending (Asc) and descending (Dsc) thoracic aortae of Fbln4^SMKO mice compared with corresponding controls (CTLs). Aortae from 4 different sibling pairs of CTL and Fbln4^SMKO were analyzed in which CTL was considered as 1. Data represent means ± SD. NS, not significant.

Compromised elastic fibers, increased wall thickness and disarray of SMCs in the Fbln4^GKO aorta

To better understand the pathological processes of aneurysm formation in the absence of Fbln4, we performed a developmental analysis of the ascending aorta in Fbln4^GKO embryos. Consistent with a previous observation ⁸, elastic fibers were clearly detected in the aortic wall of wild-type embryos at E14.5, whereas elastic fibers were decreased in the aorta from Fbln4^GKO embryos (Online Figures IVb, d). Despite compromised elastic fibers in the Fbln4^GKO aorta, at this timepoint, the overall morphology of SMCs was indistinguishable from the wild-type (Online Figures IVa, IVc). From E15.5, elastic fibers increased progressively in the wild-type aorta and SMCs became spindle-shaped with elongated nuclei (Online Figures IVe-f, IVi-j). In the Fbln4^GKO aorta, however, elastic fiber assembly remained defective and SMCs appeared immature with round nuclei (Online Figure IVg-h, IVk-l). Although not reported previously, eosin-positive focal degeneration and marked proliferation of SMCs were observed in aneurysmal lesions of the Fbln4^GKO aorta at E17.5 (Online Figure IVo) and E18.5 (arrow in Figure 2b). The degenerative area was devoid of elastic fibers and did not stain with Alcian Blue-PAS, indicating no accumulation of polysaccharides or mucins (Figure 2d and Online Figure IVp, data not shown). Only a small numbers of neutrophils and mononuclear cells were detected around degenerative lesions and some infiltrating through the intima (Online Figure VAa). In advanced lesions with dissection, more infiltrates were observed within the vessel wall (Online Figure VAb). No difference in proMMP-2 or MMP-2 levels was observed between wild-type and Fbln4^GKO descending thoracic aortae (Online Figure VB). Interestingly, this focal degeneration of the aortic wall was not detected in embryos younger than E17.5. Since Fbln4^GKO mutants die in or after parturition, we further characterized the aortic wall at E18.5. A significant upregulation of p-ERK in ECs and SMCs was observed in the Fbln4^GKO aorta (Figure 2f) and GRP78, a marker of the endoplasmic reticulum stress response, was upregulated in SMCs surrounding the degenerative area (Figure 2h). Increased apoptotic cells were also observed in both ECs and SMCs in degenerative areas (data not shown) as well as in advanced aneurysmal lesions (Figure 2j). These observations suggest that aneurysm development in Fbln4^GKO mice involved not only defective elastic fibers but also abnormal SMCs with degenerative medial wall changes.

Figure 2.

Various pathological changes associated with aneurysms in the Fbln4^GKO aorta at E18.5. (a-d) Representative photos of HE and Hart's staining of transverse sections of the ascending aorta from wild-type (a, c) and Fbln4^GKO (b, d) embryos (n=5 per genotype). Note focal degeneration of the medial wall in the mutant (arrow in b) devoid of elastic fibers (d). Bars indicate. (e-j) Characterization of aneurysmal lesions (n=3 per genotype). (e, f) p-ERK1/2 levels are increased in the Fbln4^GKO aorta (f) compared to wild-type (e). Asterisk in f indicates degeneration of the medial wall. (g, h) Accumulation of GRP78-positive cells around hyaline degeneration in the Fbln4^GKO aorta (arrows in h). (i, j) TUNEL staining shows numerous apoptotic cells within the aneurysmal wall, including the endothelial cells in the Fbln4^GKO aorta (j). Bars in a-d indicate 20 μm, e-h indicate 40 μm and i-j indicate 100 μm.

Defects in terminal differentiation of SMCs in Fbln4^GKO aorta

Histological abnormalities in SMCs of the Fbln4^GKO aorta led us to further evaluate the SMC phenotype in Fbln4^GKO embryos. Specifically, we examined expression of SMC differentiation markers in wild-type and Fbln4^GKO descending thoracic aortae at E16.5. In the wild-type aorta, SM α-actin and SM-MHC, early and late SMC differentiation markers, respectively, were expressed throughout the aortic wall (Figures 3Aa and 3Ac). In the Fbln4^GKO aorta, the intensity of both markers was not homogenous along the longitudinal axis of the aorta, particularly for SM-MHC (Figures 3Ab and 3Ad). At E18.5, SM α-actin and SM-MHC levels were increased in both wild-type and mutant aortae. The mutant aorta, however, still exhibited reduced marker expression compared to the wild-type aorta (Figures 3Ai-3Al). To confirm these findings, tissue extracts from the descending thoracic aorta devoid of aneurysm lesions from Fbln4^GKO embryos were compared with those of wild-type embryos by Western analysis (Figure 3B). Consistent with tissue immunostaining, we observed significant decreases in SM-MHC and α-SM-actin expression. In addition, a marked increase in MAPK signaling pathways, including p-ERK1/2 and p-MEK1/2, was observed in mutant aortae (Figure 3B). Other components of MAPK pathways were only moderately increased in the mutants.

Figure 3.

SMCs exhibit terminal differentiation defects in Fbln4^GKO aorta. A. Representative transverse sections of the descending aorta from wild-type (n=3) and Fbln4^GKO (n=3) embryos immunostained with SM α-actin and SM-MHC at E16.5 and E18.5. Corresponding DAPI staining is shown. At E16.5, when aneurysm formation is not evident, reduced SM α-actin and SM-MHC staining is already observed in the mutant (b, d) compared with the wild-type aorta (a, c). At E18.5, strong staining for SM α-actin and SM-MHC is observed in the wild-type aorta (i, k) whereas expression of both proteins is reduced in the mutants (j, l). Note comparative expression of SM-MHC in small vessels near the aorta in the wild-type and mutant (arrows in k and l). Bars indicate 100 μm. B. Representative Western blots of descending thoracic aorta extracts from E18.5 wild-type (n=3) and Fbln4^GKO (n=3) embryos. For α-SM-actin, the experiments were done 2 times using 4 aortae per genotype. SM-MHC is reduced and stress-mediated kinase pathways p-ERK1/2 and p-MEK1/2 are increased in the Fbln4^GKO aorta. p-JNK and p-p38 are moderately increased.

Disrupted wall structure and alterations of SMC gene expression in Fbln4^GKO aorta

It has been shown that the aorta lacking elastin exhibits increased SMC proliferation and changes in SMC organization ¹⁸. We therefore examined aortic ultrastructure at the electron microscopic level in E18.5 embryos. In the wild-type aorta, near continuous elastic laminae were observed throughout the aortic wall, with those closer to the lumen being more completely formed at this time (Figures 4Aa, 4Ac). SMCs maintained close contacts with neighboring cells and developing elastic fibers. In contrast, discontinuous aggregates of elastin were observed in the Fbln4^GKO aorta (Figure 4Ab, arrow). Higher magnification showed loose cell-cell associations and very few cell-elastin contacts with increased pericellular spaces in the Fbln4^GKO aorta, exhibiting an immature vessel wall (Figure 4Ad).

Figure 4.

Disorganized elastic laminae and altered expression of SMC marker genes in Fbln4^GKO aortae. A. Electron micrographs of E18.5 aortae from wild-type and Fbln4^GKO embryos. Elastic fibers (arrow in a) are being formed the wild-type aorta. In the Fbln4^GKO aorta, numerous elastin globules are seen in the mutant vessel (arrow in b). At higher magnification, disarray of SMCs and an increased pericellular ECM is observed (d). Bars indicate 5 μm. B. qPCR analysis of SMC-specific genes in the descending thoracic aortae devoid of aneurysm lesions from E18.5 wild-type (designated as 1, white column) and Fbln4^GKO embryos (black column). N=4 using 12 embryos per genotype. Bars indicate means ± SD.

To examine if the absence of fibulin-4 affected SMC-specific gene expression in the aortic wall, we performed qPCR analysis using E18.5 descending thoracic aorta from wild-type and Fbln4^GKO mice. SMC-specific genes, including Acta2 and Myh11, were downregulated in the mutant aorta (Figure 4B), consistent with the data obtained from Western blot and immunostaining analyses (Figure 3). Interestingly, transcription factors involved in SMC differentiation, including Myocd (myocardin), Mkl1 (myocardin-like 1/MAL/MRTF-A) and Mkl2 (myocardin-like 2/MRTF-B) were also downregulated. Because the lack of intact elastic fibers has been shown to affect SMC phenotype in the Eln^-/- aorta, and recombinant tropoelastin induced actin polymerization without changing actin expression in Eln^-/- SMCs ²³, the immature phenotypes of Fbln4^GKO SMCs may be secondary to the loss of intact elastic fibers. To examine this possibility, we evaluated the expression profile of SMC marker genes in Eln^-/- aortae. Transcripts for Myocd, Mkl1, Mkl2, Myh11 or Acta2 were not downregulated in Eln^-/- aortae (Online Figure VI). Rather, Myocd, Myh11 and Acta2 were upregulated compared to wild-type aortae. Although we cannot completely distinguish a direct or elastic fiber-mediated role of fibulin-4 on differentiation of SMCs in the current experimental system, the distinct transcriptional profile of the Fbln4^GKO aorta suggest a unique and/or additional role of fibulin-4 in development of the aortic wall.

Loss of fibulin-4 results in an immature phenotype in Fbln4^GKO SMCs

To further investigate the effect of loss of fibulin-4 on SMC phenotype, we isolated SMCs from the aorta of wild-type or Fbln4^GKO embryos between E17.5 and E18.5. Morphological analysis with phalloidin staining revealed that stress fibers were decreased and cortical actin staining more prominent in Fbln4^GKO SMCs compared to wild-type SMCs (Figure 5A). SM-MHC protein content was markedly reduced in Fbln4^GKO SMCs by Western analysis (Figure 5Ac). Separation of SM1 (204 kDa) and SM2 (200 kDa) by SDS-PAGE revealed that loss of SM-MHC was predominantly due to loss of SM1, the predominant myosin heavy chain in the developed aorta. SM α-actin was highly expressed in SMCs and we did not detect any appreciable difference in SM α-actin content between the genotypes by Western blot analysis (data not shown). Next, we examined the capacity of Fbln4^GKO cells to proliferate in the absence of serum, a characteristic feature of embryonic-type SMCs ²⁴. Wild-type and Fbln4^GKO SMCs were allowed to adhere to the culture dish and the medium was changed to a serum-free medium and grown for 24h. Immunostaining with anti-phosphorylated histone H3 (PH3), a mitosis marker, showed a significant increase in mitotic cells in Fbln4^GKO SMCs (Figures 5Bb, 5Bd), whereas positive cells were virtually absent in wild-type SMCs (Figures 5Ba, 5Bc). Proliferation assays performed following stimulation with 20% FBS also exhibited a marked increase in the proliferation of Fbln4^GKO cells (Figure 5C).

Figure 5.

Fbln4^GKO SMCs display an embryonic phenotype. A. Actin fibers detected by fluorescein-conjugated phalloidin in wild-type (a) and Fbln4^GKO (b) SMCs. Note an overall decrease in actin fibers in Fbln4^GKO SMCs. Five captured images were analyzed for each of the three independent experiments. Bars indicate 50 μm. (c) Representative Western blots showing a marked reduction of SM-MHC protein in Fbln4^GKO SMCs. B. Immunostaining of wild-type (a, c) and Fbln4^GKO SMCs (b, d) grown in serum-free media with anti-PH3 antibody (a, b). The experiment was done twice in duplicates. A significant increase in the number of proliferating cells was observed in Fbln4^GKO SMCs. Bars indicate 100 μm. C. Proliferation assays. The experiment was done twice in triplicate. Bars indicate means ± SD. D. a. Cyclic stretch assays. Two × 10⁵ cells (group 1) or 4 × 10⁵ cells (group 2) of wild-type and Fbln4^GKO SMCs were subjected to stretch stimulation followed by Western blot analysis with anti-SM-MHC and anti-GAPDH. The experiments were done five times. b. Representative Western blot showing p-ERK1/2 and t-ERK1/2 levels upon cyclic stretch stimulation in wild-type and Fbln4^GKO SMCs. c. Densitometric analysis of p-ERK1/2 levels normalized to t-ERK1/2 and GAPDH. Bars are means ± SD (n=5). No statistical significant difference is observed in p-ERK levels.

Compromised contractile forces resulting from the mutations of genes encoding for SMC-specific contractile proteins or cardiac sarcomeric proteins are proposed to cause familial thoracic aortic aneurysms and familial hypertrophic or dilated cardiomyopathy, respectively ^{25, 26}. Therefore, it is possible that Fbln4-null SMCs with reduced expression of SMC contractile proteins may exhibit an abnormal response to mechanical strain compared to wild-type SMCs. To examine this possibility, we performed in vitro stretch assays. Wild-type and Fbln4^GKO SMCs were plated on a type I collagen pre-coated elastomer well and subjected to cyclic stretch (0.2 Hz, 20% strain) for 24 h without serum. Cells incubated in a static condition served as controls. Changes in SM-MHC and activation of downstream signaling pathways were evaluated by Western analysis. As Figure 5Da shows, wild-type SMCs maintained SM-MHC levels after the stretch stimulation, however, Fbln4^GKO cells showed a marked decrease in SM-MHC levels in both static and stretch conditions compared to wild-type cells (p<0.005). Although a trend indicating a decrease in SM-MHC expression was observed after stretch stimulation in mutant cells by densitometric analysis, no statistical significance was obtained (p=0.056). Cyclic stretch showed a moderate increase in p-ERK in Fbln4^GKO SMCs compared to wild-type cells (Figure 5Db), however, statistical significance was not obtained under the given condition (p=0.088, Figure 5Dc). These data suggest that absence of fibulin-4 results in an immature SMC phenotype with compromised level of SM-MHC and altered proliferative response.

Differentiation defect of SMCs and marked upregulation of p-ERK1/2 in the Fbln4^SMKO aorta

Finally, we asked if postnatal aneurysms in the Fbln4^SMKO mice share the same underlying pathology as those of Fbln4^GKO mice. Histological analysis of Fbln4^SMKO ascending aorta at P90 revealed focal lesions with a thickened aortic wall, which primarily due to an increase in medial SMC layers (Figures 6Aa, 6Ad). Disarray of SMCs was particularly prominent in the outer two-thirds of the aortic wall in which disorganized elastic fibers were also prominent (Figures 6Ab, 6Ae). Inflammatory infiltrates, including neutrophils and macrophages, were not prominently seen in Fbln4^SMKO aortae. An overall increase of collagen fibers was also detected in the mutant aorta (Figures 6Ac, 6Af). Surprisingly, in contrast to the dramatic changes seen in the ascending aorta, the descending thoracic aorta appeared normal without hyperproliferation of SMCs (Online Figure VIId). Elastic laminae were more organized compared to those in the ascending aorta, although discontinuities were more frequently observed in mutants compared to the wild-type aorta (Online Figure VIIe). Masson-Trichrome staining revealed significant increases in collagen fibers in the mutant descending thoracic aorta (Online Figure VIIf).

Figure 6.

SMC-specific deletion of Fbln4 recapitulates the aortic phenotype in Fbln4-null embryos. A. Histological evaluation of focal proliferative lesions in the Fbln4^SMKO aorta at P90. Representative photos of transverse sections of ascending aorta from control (a-c) and Fbln4^SMKO (d-f) mice stained with HE (a, d), Hart's (b, e) or Masson-Trichrome (c, f). B. qPCR analysis of SMC specific genes in ascending (stripe column) and descending (black column) thoracic aortae from Fbln4^SMKO mice and corresponding controls (white column). Four sibling pairs were used. Bars indicate means ± SD. C. Representative immunostaining of the ascending aorta from control (a, b) and Fbln4^SMKO (c, d) mice stained with anti-SMemb (P25) and SM-MHC (P90) antibodies. Note strong expression of SMemb and reduced expression of SM-MHC in the mutant aorta. D. (upper panel) Representative Western blot of the ascending aorta before (P25) and after (P90) the formation of aneurysms. (lower graph) A ratio between SM-MHC to GAPDH in P25 and P90 ascending aortae is determined densitometrically and plotted. Each data point represents a measurement from an ascending aorta at P25 or P90.

To examine if the differentiation defect of SMCs is also present in the Fbln4^SMKO aorta, qPCR analysis was performed using ascending and descending thoracic aortae from Fbln4^SMKO and control mice at P90. Transcripts for SMC contractile genes were significantly downregulated in the diseased ascending aorta as well as in the descending aorta where no aneurysms were observed (Figure 6B). In addition, Myocd was downregulated in both ascending and descending thoracic aortae. Immunostaining of the ascending aorta at P25 with anti-SMemb, a marker of embryonic SMCs ²⁷, showed faint staining in the wild-type aorta, whereas persistent expression was detected in the mutant aorta (Figures 6Ca and 6Cc). In contrast, immunostaining with anti-SM-MHC detected strong expression in P90 wild-type ascending aorta, whereas a marked reduction of SM-MHC was observed in the mutant aorta (Figures 6Cb and 6Cd). The reduction of SM-MHC levels was further confirmed by Western blot analysis using ascending aortae harvested from P25 and P90 Fbln4^SMKO mice (Figure 6D). At both time points, before or after aneurysm development, reduction of SM-MHC was observed in the mutant aorta

Finally, we examined signaling pathways responsible for the proliferative changes of the aneurysmal wall. Upregulation of p-ERK1/2 was observed in the Fbln4^SMKO aorta at P25 and P90 by Western blot analysis (Figure 7A). Immunostaining of the P90 ascending aorta revealed strong upregulation of p-ERK1/2 in the mutant SMCs compared to control vessels (Figure 7B). Phosphorylated p38 level was moderately increased in the mutant aorta (data not shown). Levels of p-Smad2/3, p-Smad1/5/8, and CTGF, a downstream target of TGF-β, were comparable between control and Fbln4^SMKO aortae examined at P25 and 3-months of age (Online Figure VIII Aa, B). In addition, we did not detect any difference in p-Smad2/3 levels between wild-type and Fbln4^GKO aortae by Western blot or immunostaining (Online Figure VIII Ab, data not shown). Taken together, these observations in the Fbln4^GKO and Fbln4^SMKO aortae suggested that 1) alteration of SMC-specific genes was not simply a consequence of aneurysm formation in the mutant aorta, and 2) the ERK1/2 pathway played a pivotal role in the focal proliferative response of abnormal SMCs in the aneurysmal wall.

Figure 7.

Increased ERK1/2 signaling in the Fbln4^SMKO aorta. A. Representative Western blot of ascending aortae from control and Fbln4^SMKO mice at P25 and P90. An increase in p-ERK1/2 is already evident at P25. B. Representative immunostaining of P90 ascending aorta from control (a, b) and Fbln4^SMKO (c, d) mice stained with p-ERK1/2. Increased level of p-ERK1/2 is observed in SMCs of the Fbln4^SMKO aorta. Note strong signal in perivascular adipose tissue from both control and the mutants (arrow in a and c). Bars in (a, c) indicate 500 μm and in (b, d) indicate 50 μm. C. A model illustrating pathological changes in Fbln4 mutant aorta. Absence of fibulin-4 in the aortic wall affects formation of elastic fibers and terminal differentiation of SMCs. Fbln4-null SMCs exhibit an increase in ERK1/2 phosphorylation, possibly mediated by external factors, including mechanical stimuli or soluble factors. Proliferation of SMCs, apoptosis of vascular cells (illustrated by deformed cells) and further destruction of vascular ECM by secondary inflammatory cell response, together contribute to the aneurysm development.

Discussion

Potential contribution of abnormal SMCs to aneurysm development in Fbln4 mutant mice

Deletion of fibulin-4 in a global or SMC-specific manner resulted in a differentiation defect of SMCs with marked down-regulation of SMC contractile genes. Contractile forces in SMCs are generated by cyclic interactions of myosin heavy chain and actin filaments induced by MLCK-mediated phosphorylation of myosin light chains (Reviewed in ^{28, 29}). The mutations in components of the contractile apparatus in SMC such as ACTA2 and MYH11 lead to a focal proliferation of SMCs with disarray, disruption of elastic fibers, and eventual loss of SMCs ^10-12. It is also shown that patients with ACTA2 disease haplotype exhibit reduced aortic compliance and increased pulse velocity, altering aortic stiffness ¹⁰. Although aortic aneurysms were not observed in genetic ablation of the SM2 isoform of SM-MHC in mice, the mutants did develop an enlarged thin bladder and hydronephrosis and died within 4 weeks of age ³⁰. A concomitant reduction of SM1 was observed in the SM2^-/- bladder and an appropriate ratio of SM1 to SM2 was shown to be critical for normal contractility of SMCs. Absence of Acta2 also exhibited a significant decrease in vascular contractility and impaired maintenance of blood pressure and blood flow ³¹. Thus, quality and quantity of contractile proteins significantly affect contractile functions of SMCs and eventually affect mechanical properties of the aortic wall. Taken together, we suggest that Fbln4 mutant SMCs with reduced expression of contractile proteins and highly synthetic phenotype, together with the abnormal ECM, alter vessel wall stability and results in maladaptation of the vessel wall, providing a potential link to an aneurysm phenotype (Figure 7C).

Fibulin-4 provides a permissive environment for differentiation and maintenance of SMCs

SMC differentiation is largely governed by the myocardin family of transcriptional co-activators (reviewed in ³²). Myocardin, Mkl1 and Mkl2 serve as a potent co-factor of SRF and upregulate a variety of SMC contractile genes ^33-36. The reduced expression of Myocd, Mkl1, and Mkl2 in the Fbln4^GKO mutant aorta indicates a potential mechanism of extracellular regulation of SMC differentiation genes by fibulin-4. Fibulin-4 is located in microfibril bundles which tether elastic fibers to SMCs via integrin-mediated binding at regions of the cell membrane occupied by intracellular membrane-associated dense plaques, anchoring sites for actin filaments ^{37, 38}. Thus, loss of fibulin-4 may disrupt the interaction between elastic fibers and SMCs, leading to the alterations in actin cytoskeleton organization. The actin dynamics between filamentous (F) and globular (G) actin are known to regulate the activity of Mkl1: G-actin binds to cytoplasmic Mkl1 and inhibits translocation of Mkl1 to the nucleus thereby disrupting the interaction between Mkl1 and SRF and downregulating the transcription of SMC contractile genes ³⁹. Our observation that actin filaments were less prominent in Fbln4^GKO SMCs compared to wild-type cells is consistent with this hypothesis. A similar disruption of the elastin-SMC connection was reported in a mouse deficient for the polycystic kidney disease 1 (Pkd1) gene, which developed descending aortic aneurysms with hyperproliferative SMCs ⁴⁰. Ultrastructural studies of Fbn1^C1039G/+ aortae also revealed disruption between SMC and elastic fiber connections, leading to activation of SMCs ⁴¹. Taken together, fibulin-4 is critical for the maintenance of the permissive and stable environment for SMC differentiation.

Elastic fibers and protection of aortic aneurysms

Our present study, together with other reports, confirmed that a loss of fibulin-4 profoundly affects formation of elastic fibers ^{8, 19}. Various experimental manipulations that disrupt elastic fibers have successfully recapitulated pathological changes of aortic aneurysms in rodents ^{42, 43}. Genetic or pharmacological stabilization of elastin or inhibition of its degradation was shown to attenuate development of abdominal aortic aneurysms ^44-46. Although elastic fibers or insoluble elastin have been shown to suppress proliferation and migration of SMCs ²³, a causal relationship between disrupted elastic fibers and aneurysm formation during development has not been fully established. Interestingly, mice deficient in the fibrillin-1 gene (Fbn1^mgN/mgN) that die from aneurysm rupture within two weeks of age, exhibit maturation defects of the aortic wall and persistent expression of embryonic SMC markers ⁴⁷. Thus, components of elastic fibers affect SMC phenotypes directly or indirectly during development, a time when SMCs have a high degree of plasticity.

Signaling pathways involved in proliferation of SMCs in the Fbln4 mutant aorta

Hyperproliferation of SMCs is a critical component involved in the alteration of the Fbln4 mutant aorta and is predominantly mediated by the ERK1/2 signaling. A cue for the activation of ERK1/2 may be provided by mechanical stimuli such as increased hemodynamic force and changes in aortic ECM components, paracrine factors or by undefined mechanisms. Cyclic stretch activates MAPK pathway (both ERK and p38) and increases import of nuclear protein as well as transcription of nuclear pore complex ⁴⁸. Although the causal relationship between cyclic strain and ERK1/2 phosphorylation was not established in Fbln4^GKO SMCs using an in vitro system, the observation that aneurysms developed exclusively in the ascending aorta, the region of highest systolic pressure, in Fbln4^SMKO mice suggests that mechanical strain may influence SMC function in vivo. Angiotensin II activates the ERK pathway and is also a known mediator of the stretch response in SMCs through local production of angiotensinogen or local conversion of angiotensin I to angiotensin II ⁴⁹. AT1 receptor blockade inhibits gene transcription and proliferation of SMCs induced by stretch stimulation as well as the hypertrophic response of cardiac myocytes induced by pressure overload ^50-52. Upregulation of IGF-1 and phosphorylation of IGF-receptor is also induced by cyclic stretch and lead to the proliferation of venous SMCs ^{53, 54}. Interestingly, a high level of angiotensin II and IGF-1 was observed in primary SMC cultures prepared from a patient with a MYH11 mutation, suggesting the possible involvement of these soluble factors in the proliferative response of SMCs ¹¹. It is also possible that TGF-β activates ERK1/2 pathways by a non-Smad dependent mechanism ⁵⁵. Thus, Fbln4 mutant mice give us the opportunity to test pharmacological interventions to block specific signaling pathway and attenuate aneurysm formation in vivo.

Limitation of this study

Our current study revealed a critical role for fibulin-4 on differentiation of SMCs and protection from aneurysm development. Since fibulin-4 is a secreted protein, our conditional knockout phenotypes may be partially masked by fibulin-4 supplied by adjacent tissues or remaining production of fibulin-4 within the aortic wall. Fibulin-4 plays a critical role in elastic fiber assembly, therefore, the temporal and causal relationship between defective elastic fibers and alteration of SMC phenotype in Fbln4-null aorta remain to be further investigated. In addition, potential cellular functions of fibulin-4 such as the effect on apoptosis or collagen fiber formation, may be explored.

Non-standard Abbreviations and Acronyms

CTGF

connective tissue growth factor

CTL

control

ECM

extracellular matrix

ECKO

endothelial cell-specific knockout

Eln

elastin

ERK

extracellular signal –regulated kinase

Fbln4

fibulin-4

Fbln5

fibulin-5

IGF

insulin growth factor

JNK

Janus kinase

MEK

mitogen-activated protein kinase kinase and extracellular signal-regulated kinase kinase

p38

p38 MAP kinase

SMC

smooth muscle cell

SMKO

smooth muscle cell-specific knockout

SM-MHC

smooth muscle myosin heavy chain

TAAD

thoracic aortic aneurysms and/or aortic dissection

TGF-β

transforming growth factor-beta

BMP

bone morphogenetic protein