|Home | About | Journals | Submit | Contact Us | Français|
β1 integrin has been shown to contribute to vascular smooth muscle cell differentiation, adhesion and mechanosensation in vitro. Here we showed that deletion of β1 integrin at the onset of smooth muscle differentiation resulted in interrupted aortic arch, aneurysms and failure to assemble extracellular matrix proteins. These defects result in lethality prior to birth. Our data indicates that β1 integrin is not required for the acquisition, but it is essential for the maintenance of the smooth muscle cell phenotype, as levels of critical smooth muscle proteins are gradually reduced in mutant mice. Furthermore, while deposition of extracellular matrix was not affected, its structure was disrupted. Interestingly, defects in extracellular matrix and vascular wall assembly, were restricted to the aortic arch and its branches, compromising the brachiocephalic and carotid arteries and to the exclusion of the descending aorta. Additional analysis of β1 integrin in the pharyngeal arch smooth muscle progenitors was performed using wnt1Cre. Neural crest cells deleted for β1 integrin were able to migrate to the pharyngeal arches and associate with endothelial lined arteries; but exhibited vascular remodeling defects and early lethality. This work demonstrates that β1 integrin is dispensable for migration and initiation of the smooth muscle differentiation program, however, it is essential for remodeling of the pharyngeal arch arteries and for the assembly of the vessel wall of their derivatives. It further establishes a critical role of β1 integrin in the protection against aneurysms that is particularly confined to the ascending aorta and its branches.
The assembly of the vascular wall during development occurs through the rapid investment of endothelial tubes by mesenchymal cells and their progressive differentiation into smooth muscle cells (Drake et al., 1998; Hungerford and Little, 1999; Owens, 1995). During this process, vascular smooth muscle cells (vSMCs) secrete and organize layers of extracellular matrix proteins that quickly intertwine with smooth muscle cells to generate a highly integrated tissue able to respond to and regulate intravascular pressure (Gasser et al., 2006; Jones et al., 1979; Li et al., 2003; Wagenseil and Mecham, 2009). The various extracellular matrix (ECM) components are critical to the integrity of the wall, as the smooth muscle cells themselves are not sufficient to comply with the constant mechanical stress imposed by the pulsatile blood flow (Wagenseil and Mecham, 2009). Human mutations in fibrillin 1 (FBN1) or type III α1 collagen (COL3A1) genes cause Marfan and Ehlers-Danlos syndromes respectively and result in aortic aneurysms, an abnormal enlargement of the aorta caused by thinning of the vessel wall (Dietz et al., 1991; Dietz and Pyeritz, 1995; Pope et al., 1975). Furthermore, mouse models that lack Fbn1, fibulin 4 (FBLN4) and biglycan (BGN) also led to the development of aneurysms in the aorta (Heegaard et al., 2007; Maki et al., 2002; McLaughlin et al., 2006; Pereira et al., 1997). Likewise, genetic mutations in smooth muscle contractile proteins were noticed to be responsible for hereditary vascular anomalies (Guo et al., 2007; Pannu et al., 2007; Zhu et al., 2006). For example, missense mutations in α actin (ACTA2) were found to be associated with 14% of inherited aortic dissections (Guo et al., 2007). More recently, heterozygous mutations in smooth muscle myosin heavy chain (MYH11) were identified in kindreds with a wide variety of vascular anomalies, including patent ductus arteriosus, thoracic aortic aneurysms and aortic dissections (Pannu et al., 2007; Zhu et al., 2006). Thus, it is clear that the pathobiological processes that lead to the development of aneurysms include abnormalities in either vascular ECM proteins, vSMC or in a combination of both.
The integrin family of heterodimeric transmembrane receptors connects the extracellular matrix to the actin cytoskeleton and is thought to participate in vSMC differentiation, ECM organization and mechano-sensing in vivo (Hynes, 2002; Li et al., 2003; Sun et al., 2005; Xiao et al., 2007). While many of the integrins have been shown to contribute to vSMC function in vitro, β1 integrin is considered to play a particularly significant role as it partners with multiple α subunits and thus constitutes a highly represented receptor (Martinez-Lemus et al., 2003). Furthermore, in vitro studies demonstrated that β1 integrin is important for the differentiation of stem cells into smooth muscle and in the organization of collagen and fibronectin fibers (Li et al., 2003; Xiao et al., 2007). Though in vitro and ex vivo studies are informative, these studies cannot fully recapitulate the in vivo context.
Global inactivation of β1 integrin result in peri-implantation embryonic lethality around embryonic day 5.5 (E5.5), several days prior to smooth muscle cell differentiation in the embryo (Fassler and Meyer, 1995; Stephens et al., 1995). Consequently, cell-specific deletion experiments through Cre-lox systems have been critical to elucidate the specific contributions of this gene within a cellular compartment. To determine the role of β1 integrin in differentiated smooth muscle cells, we independently crossed two different β1 integrin flox alleles to the sm22αCre recombinase mouse (Holtwick et al., 2002). The use of the sm22α ensured that deletion occurred in a population of cells that were committed to the smooth muscle fate and had initiated their differentiation program. Mutant mice were embryonic lethal at late gestation and exhibited interrupted aortic arch, aneurysms and failure to assemble ECM proteins. Interestingly, the phenotype was mostly confined to the branches of the aortic arch and compromising the brachiocephalic and carotid arteries. The defects did not extend to the dorsal aorta. Deletion of β1 integrin in neural crest using the wnt1Cre confirmed that absence of the protein did not affect migration of smooth muscle precursors to the aortic arch. These findings establish a critical period in development (E15.5 to E18.5) that requires β1 integrin for the assembly of the vascular wall. The data also highlights an essential contribution of β1 integrin in the protection against aneurysms that is confined to the ascending aorta and emerging/proximal carotid arteries.
β1e3 (Raghavan et al., 2000), sm22αCre (Holtwick et al., 2002) and wnt1Cre (Danielian et al., 1998) mice were purchased from Jackson Labs. We also used the previous published β1fl mouse (Potocnik et al., 2000). Cre recombinase expressing mouse lines were independently bred to the β1fl and β1e3 transgenic lines. Additionally, β1e3; sm22αCre+ and β1e3; wnt1Cre mice were crossed to the ROSA26R (Soriano, 1999) line, also obtained from Jackson Labs. Genotyping for these lines was performed as previously described (Turlo et al., 2010; Zovein et al 2010). The vaginal plug in conjunction with developmental staging was used to determine the embryonic age of the embryos evaluated. Experiments were evaluated and approved by the Animal Research Committee at the University of California, Los Angeles.
β-galactosidase staining was performed as described previously (Turlo et al., 2010). Immunostaining for β1 integrin on the β1fl dorsal aorta was done on vibratomed sections without an unmasking step. Tissue was dissected and placed in 2% paraformaldehyde overnight. Sections were obtained after the tissue was embedded in agarose and vibratomed into 300–500μm sections. Specimens were then rinsed in 1X PBS and placed in blocking solution (1x PBS pH7.8, 5% donkey serum, 0.3% triton) for 1h. Next, sections were incubated in blocking buffer with primary antibody overnight following standard immunostaining techniques as previously described (Zovein et al., 2010). All other immunostaining was performed on histological sections also following protocols described (Zovein et al. 2010, for detailed protocols visit: http://www.mcdb.ucla.edu/Research/Arispe/index.php). Unmasking of epitopes was done in 10mM sodium citrate at pH 6.0 and heated to 100°C for 15 minutes. Sections were blocked in 5% donkey serum in 1X PBS with 0.3% triton. Antibodies used in immunostaining were β1 integrin (1:100, Millipore, MAB 1997), FITC-α-smooth muscle actin (1:200, Sigma, F3777), myh11 (1:200, ABD Serotec), calponin (1:200, Abcam), tropoelastin (1:200, abcam), fibronectin (1:200, abcam), laminin (1:200, Sigma), collagen IV (1:200, abd serotec) and sm22α (1:200, Abcam). Samples were imaged using a 10x, 40x or 100x objective in a Zeiss LSM 510 META confocal microscope.
For Western blots on E15.5 and P(0) animals, the whole thoracic aorta was extracted in RIPA buffer (Lee et al., 2006). Tissue for Western blot of E10.5 embryos was collected from the pharyngeal arch region of the embryo. Tissue was first snap frozen and subsequently extracted in RIPA buffer. Total protein concentrations were obtained using DC protein assay (Bio-Rad) on the Bio-Rad Molecular imager Chemi Doc XRS+ using Image Lab Software v3. Equal total protein levels (generally 5–10μg) were loaded per lane. Antibodies used for Westerns included: β1 integrin (1:1000, Millipore, AB1952), α-tubulin (1:1000, Sigma, T5168), GAPDH (1:5000, Millipore, MAB374) and α-smooth muscle actin (1:5000, Sigma, A2547).
While the expression of β1 integrin in mature vSMCs is well established (Mechtersheimer et al., 1994), the initiation of β1 integrin expression in this cell population has not been characterized. Thus, we crossed mice carrying β1 integrin loxP flanked allele (β1fl) that includes a lacZ reporter to mice with the sm22αCre (sm) transgene (Fig. 1A) (Holtwick et al., 2002; Potocnik et al., 2000). Excision of the DNA between the loxP sites concurrently results in genetic deletion and reporter activation providing a faithful read-out of β1 integrin promoter activity. We chose to use the sm22αCre as this gene is expressed at the onset of smooth muscle differentiation (Yang et al., 2010). From this cross we found β-galactosidase (β-gal) positive cells are present as early as E9.5 in the dorsal aorta and heart (Fig. 1B). By E15.5 all smooth muscle cells of the aortic arch and carotids are positive for β-gal (Fig. 1C). This expression persists into maturity and it includes all smooth muscle cells in the vessel wall suggesting a constitutive need for this gene in vSMC even after development has ceased (Fig. 1D). In addition to vSMCs, staining for β-gal in smR26R and smβ1fl/wt embryos confirmed sm22αCre also targets early cardiomyocytes, where the β1 integrin promoter is active as early as E8.5 (Fig. 1B, C and Fig. S1). Overall, we found that expression of β1 integrin essentially mirrors the sm22α promoter, as per ROSA26R β-gal evaluations (data not shown).
In addition to functioning as a reporter of β1 integrin promoter activity, the β1fl allele also promotes deletion of β1 integrin. Generation of smβ1fl/fl mice revealed that these mice died embryonically or shortly after birth. No live animals with the mutation were found after postnatal day 5 [P(5)] (Table I). At the neonate stage, smβ1fl/fl animals were easily recognized among their wild type littermates, as per their pale appearance (Fig. S3A). Previous work by our lab revealed differences in phenotypes depending on the β1 integrin loxP construct used (Turlo et al., 2010; Zovein et al., 2010). To gain a more comprehensive insight into β1 integrin deletion, we crossed the sm22αCre mouse to a second β1 integrin loxP allele (β1e3, Fig. 1A) (Raghavan et al., 2000). In this model, no Cre positive smβ1e3/e3 mice survived past parturition (Table I). Western blot analysis of β1 integrin in the dorsal aorta for both alleles, smβ1fl/fl at P(0) and smβ1e3/e3 at E15.5, confirmed that both models resulted in significant reduction of β1 integrin protein levels (Fig. 1E). This was further verified by immunofluorescence of β1 integrin in the dorsal aorta and carotid where β1 integrin expression was maintained in the endothelium (white arrows) of smβ1fl/fl and smβ1e3/e3 while a significant reduction to complete loss of protein was noted in the smooth muscle (Fig. 1F–G and Fig. S2).
Cre-recombinase activity was mapped using the ROSA26 reporter early in development (E 9.5) and found to target both cardiomyocytes and vascular smooth muscle cells (Fig. 1B–D, Fig. S1). Examination of mutant mice (smβ1fl/fl) at P(0), revealed cyanosis that was consistent with multiple sites of hemorrhage in the epicardium (Fig. S3E [black arrows]). These hemorrhages were present in over 80% of E18.5 smβ1fl/fl animals and were also seen in embryonic smβ1e3/e3 knockouts (Fig. S3F, S4A [black arrows], S4D [black arrows] and S4E). Histological evaluation of the smβ1e3/e3 hearts denoted further anomalies (Fig. S4A and S4D). Incomplete ventricular septum was found in all of the smβ1e3/e3 animals examined for this defect (E15.5, n=5; Fig. S4A [red arrows]). In contrast control β1e3/e3 littermates exhibited fusion of the ventricular septum (Fig. S4A; n=4). By E18.5, the ventricular septal defects resolved (Fig. S4D) and the hemorrhages in the epicardium were less prominent (Fig. S4D [black arrow]).
The ventricular septal defect at E15.5 and the overall size of the smβ1e3/e3 hearts suggested a significant developmental delay. This data corroborates a previous report indicating deletion of β1 integrin impairs proliferation in developing cardiomyocytes, albeit that study was done with a distinct Cre line (Ieda et al., 2009). We confirmed reduced proliferative capacity in the smβ1e3/e3 particularly within the ventricular septum (Fig. S4B). As a result of this reduced proliferation, the hearts of the smβ1e3/e3 were smaller than the β1e3/e3 controls at E15.5 and E18.5 (Fig. S4A and S4C).
In addition to the cardiac defects, smβ1fl/fl and smβ1e3/e3 embryos exhibited several anomalies in the large vessels emanating from the heart. Both mice showed frequent aneurysms in the carotids and occasional defects in the aortic arch near the left common carotid branch (Fig. 2A–B, S3B–D). Aneurysms were noted at both E15.5 and 18.5 with the defect being more pronounced at later developmental stages (Figure 2A). In addition, poor extension of the carotids between aorta and internal and external carotid branch point was a commonly noted phenotype (Fig. 2A [E18.5, green arrow]). At E15.5 more than 60% of the smβ1e3/e3 were found to have either aortic arch or carotid aneurysms or an interrupted aortic (Fig. 2B). By E18.5, 79% of animals showed defects (13% interrupted aortic arch, 66% carotid/aortic arch aneurysms) (Fig. 2B). The interrupted aortic arch phenotype occurred in our early analysis of the smβ1e3/e3 mice, but it was later lost as the mixed background became more homogeneous. Of the smβ1fl/fl animals analyzed, none exhibited interrupted aortic arch. The differences in phenotype between the β1fl and β1e3 models may be attributed to the differential efficiency in β1 integrin deletion. Though both alleles showed a significant decrease in β1 integrin protein, low levels of β1 integrin in smooth muscle cells were noted in the smβ1fl/fl carotids, but complete loss was documented when using the smβ1e3/e3 (Fig. S2, 1F and G). Although the descending aorta was examined thoroughly, no dilation or thinning of the wall was ever observed in either deletion model.
The remarkable restriction of the vascular defects to the aortic arch and carotids/brachiocephalic led us to examine β1 integrin expression in this region versus the descending aorta. Western blot analysis of β1 integrin levels in the aortic arch and descending dorsal aorta of wild type E18.5 embryos demonstrated the aortic arch to have elevated levels of the protein both in late embryos and in adult mice (Fig. 2D and data not shown). For this reason we focused our attention on the etiology and molecular basis of the aneurysms of the aortic arch and carotids. Further, although both the smβ1fl/fl and the smβ1e3/e3 had similar phenotypes, here we describe the β1e3 line for sake of space. Results from the β1fl line can be found in the supplemental material (Fig. S9, S11, S12 and S14).
In order to understand the cellular mechanisms that resulted in vascular aneurysms, we performed histological analysis at the onset of the defects (E15.5) and a late stage (E18.5) when the vascular defects were more pronounced. Our systematic analysis of the dissected vessels showed frequent, multiple dilations in the carotids and aorta (Figure 2A). Histological assessment confirmed increased diameter and abnormal compaction of the vessel wall (Fig. 2C). To further understand the nature of the defect, we considered possible alterations in the number of smooth muscle cells. Thus, we first evaluated neural crest migration during investment of cells into the pharyngeal arch. This was a likely possibility as β1 integrin has been shown to be required from migration in vitro. Smooth muscle cells of the aortic arch and the base of the carotids derive from neural crest cells that migrate from the dorsal portion of the neural tube to populate the vascular media (Hiruma et al., 2002; Majesky, 2007; Serbedzija et al., 1992). Analysis of the β-gal stained smβ1fl/wt indicates that deletion using the sm22αCre occurs around E9.5, as neural crest cells reach the pharyngeal arches (Fig. 1B). The result is dependent on both sm22αCre activity and β1 integrin expression. Evaluation of sm22α (red) by immunofluorescence in conjunction with α-smooth muscle actin (green) at E10.5 demonstrates that both proteins are present and further suggests that the sm22α promoter is active in smooth muscle/progenitor by this time (Fig. S5). It appears both at the gross and histological level that Cre recombinase becomes active at the pharyngeal arch artery around E9.5 and that the vessels are equally populated by β-gal positive cells both the smβ1e3/wt; R26R controls and smβ1e3/e3; R26R knockouts (Fig. S1 and data not shown). From this data, we concluded that deletion of β1 integrin in sm22αCre expressing cells does not affect the initial populating of the aortic arch and carotids.
After migration, neural crest – derived progenitors differentiate into vSMCs, proliferate and organize into multiple concentric layers. We examined proliferation at both E15.5 and E18.5 and found no significant differences (Fig. S6A). Furthermore, assessment of apoptosis by evaluation of cleaved caspase 3 staining revealed no significant differences between both β1e3/e3 and smβ1e3/e3 (Fig. S6B). Combined, our findings indicated that the aneurysms did not result from abnormal apoptosis or proliferation.
A caveat in the interpretation of the vascular phenotype in these mice is that β1 integrin was also deleted in the heart, as per early expression of sm22α in this tissue. In order to determine whether the vascular defects were primary or secondary to deletion of β1 integrin in the heart, we used a third knock out model that bypasses heart deletion and impacts the population of vSMC affected by the deletion, namely neural crest cell derivatives. Wnt1Cre targets neural crest cells including vSMC progenitors that populate the aortic arches (Fig. 3A) (Danielian et al., 1998) and unlike the sm22αCre, the wnt1Cre is not active in the heart (Fig. 3A). Previous reports of neural crest deletion of β1 integrin, using the HtPaCre crossed to the β1fl described a Hirshsprungs-like phenotype with insufficient innervation of the colon due to reduced migration (Breau et al., 2006; Pietri et al., 2003). The animals died of malnutrition at a later postnatal stage. We crossed the Wnt1Cre to the β1fl/wt and evaluated reporter expression by β-gal activity. At E9.5 the entire pharyngeal arch region is targeted by the Wnt1Cre, as shown by cells positive for β-gal (Fig. 3A). In contrast, recombination in the smβ1fl/wt in the same region and developmental stage occurs only in differentiating cells (compare Fig. 3A and and1B).1B). Consistent with the known contribution of neural crest cells to the aorta; no β-gal positive cells were detected beyond the aortic arch or posterior to the branch of the left common carotid artery in the adult (Fig. 3A).
Evaluation of the ncβ1fl/fl revealed an intact aortic arch that was grossly wild type at E15.5 (Fig. S7). Importantly, the animals survived beyond birth as previously reported (Breau et al., 2006; Pietri et al., 2003) (Table I). In contrast, crosses between Wnt1Cre (nc) and the β1e3 resulted in fully penetrant embryonic lethality by E12.5 (Table I). At E11.5 we found a single viable ncβ1e3/e3 embryo, thus we evaluated the phenotype at E10.5. At this time, point a significant decrease in β1 integrin was noted in the pharyngeal arches (Fig. 3E). Deletion of β1 integrin in neural crest cells revealed pharyngeal arch artery aneurysms that accumulate blood (Fig. 3B [white arrow], 3D [black arrow]).
In order to assess migration of neural crest progenitors β1e3 mice, we crossed the ncβ1e3/wt to the R26R reporter mouse. Whole mount and histological examination for E9.5 and E10.5 of both ncβ1e3/wt; R26R controls and ncβ1e3/e3; R26R knockouts show the location of β-gal positive cells and the composition within the pharyngeal arches, to be equivalent (Fig. 3C–D). While aneurysms are present in ncβ1e3/e3; R26R by E10.5, this dilation is completely surrounded by β-gal positive cells, indicating that neural crest cells were able to migrate albeit with significantly reduced levels of β1 integrin (Fig. 3D [black arrow]). Immunofluorescence of dilated pharyngeal arch arteries showed cells expressing α-smooth muscle actin (green) and sm22α (red), again confirming neural crest migration and differentiation (Fig. S8). In addition, the pathological vascular dilations were present in the context of a wild type heart, indicating that the vascular defects were independent of the heart abnormalities.
The ncβ1e3/e3 data showed that a migration defect was unlikely to contribute to the observed phenotype and immunofluorescence staining indicated the defects in the carotids were not due to proliferation or apoptosis. Now that proliferation and heart defects were discarded as contributing significantly to the phenotype, we turned our attention to the structure of the tunica media. We evaluated serial sections of the carotid and brachial arteries as well as the aortic arch to better identify the extent and location of the disorganized media (Fig. 4B and 4D). The thoracic dorsal aorta was sectioned independently of the arch. Histological sections encompassing the carotids and branchial arteries deleted for β1 integrin showed defects in the organization of the vascular wall (β1e3/e3 Fig. 4A–4D and β1fl/fl Fig. S8A–S8D). However, sections across the aortic arch showed large portions of organized medial smooth muscle and ECM. In some of the smβ1e3/e3 we were able to identify defects in the arch at the base of the carotid branches (Fig. 2A red arrows and S10). Whole mount arches of P(0) smβ1fl/fl exhibited this defect most clearly (Fig. S3D). Sections of the descending aorta were undistinguishable between smβ1e3/e3 and β1e3/e3 controls.
Deletion of β1 integrin in ES cells blocks acquisition of cardiac and smooth muscle cell fate (Fassler et al., 1996; Xiao et al., 2010). In order to determine the effect of β1 integrin deletion on vSMC maturation after commitment, we evaluated arteries from smβ1e3/e3 and β1e3/e3 embryos at E15.5 and E18.5 (Fig. 5A–C and S11). By E15.5 α-smooth muscle actin, sm22α, calponin and smooth muscle myosin heavy chain were present and expressed at approximately equivalent levels in mutant and control mice. Indicating that loss of β1 integrin did not alter the differentiation. In contrast, by E18.5 the expression of sm22α, calponin and smooth muscle myosin heavy chain were significantly reduced in the smβ1e3/e3 when compared with the β1e3/e3 in the carotid. Evaluation of the dorsal aorta also showed that calponin levels are reduced in the smβ1e3/e3 though smooth muscle myosin heavy chain showed a less significant difference in expression when compared to wild type descending dorsal aorta. α-Smooth muscle actin appeared less affected by β1 integrin deletion in both the carotid and the dorsal aorta. This data indicates smooth muscle cells are able to complete their differentiation program in the absence of β1 integrin, but could not sustain adequate levels for a subset of proteins associated with the fully differentiated phenotype.
The sm22α cre recombinase enabled the evaluation of β1 integrin in the maintenance of established differentiation (Fig. S5 and and5).5). The wnt1 cre recombinase provided the means to determine if differentiation was affected by β1 integrin deletion prior to commitment to the smooth muscle fate (Fig. S8). This data further demonstrates that β1 integrin is not required for the onset of differentiation, but is necessary for the maintenance of wild type expression levels later in development.
As previously reported, ECM proteins are critical to support the integrity of the vascular wall (Wagenseil and Mecham, 2009). β1 integrin has been shown to participate in organization and cross-linking of ECM proteins (Li et al., 2003). Thus, we explored whether synthesis and/or organization of ECM in the carotid and dorsal aorta was compromised in the absence of β1 integrin. For this, we evaluated fibronectin, collagen, elastin and laminin by immunohistochemistry at E15.5 and E18.5 in β1e3/e3 and smβ1e3/e3 carotids and aorta (Fig. 6A–C and S14). Because elastin, in particular, exhibits autofluorescence, we also performed secondary controls for each of the primary antibodies analyzed (Fig. S15) as well as DAB immunohistochemical analysis (data not shown).
Levels of extracellular matrix proteins were not found to be distinct between mutant and control littermates, whereas organization of ECM was clearly abnormal when in the absence of β1 integrin. The defects were apparent by E18.5 and carotids and brachiocephalic artery to the exclusion of the descending aorta (Fig. 6). In particular, we found that the organization of ECM lamellar structures is impaired in the absence of β1 integrin. Review of the β1e3/e3 and ncβ1e3/e3 animals at E10.5 showed collagen IV and laminin are deposited and localize normally within some portions of the pharyngeal arch arteries, but that collagen IV is reduced and laminin appears disorganized in regions of aneurysms (Fig. S13).
β1 integrin has been shown to be critical for fibrillogenesis of fibronectin in vitro (Wu et al., 1995), but a comparable effect in vivo has not been reported to date. Our data shows that the contribution of β1 integrin is essential for the assembly of the ECM, particularly in ascending aorta, carotids and brachiocephalic arteries. These defects lead to the development of aneurysms and likely contribute to the lethality observed.
Our findings indicate that deletion of β1 integrin early in smooth muscle differentiation affects the structure of the ECM in areas of high stress. Whereas, deletion of this protein in smooth muscle progenitors reveals contributions to aortic arch remodeling (Fig. 7A). This phenotypic series is informative, in that it demonstrates the value of different and overlapping Cre recombinases as well as variations in loxP constructs. It is also instructive on the role of β1 integrin in pharyngeal arch remodeling. Further, assessing the timing of deletion in relation to the outcomes together with previous published work on the aortic arch enables detailed mapping of vessel assembly, an important aspect of vascular development (Fig. 7B) (Hiruma et al., 2002; Lindahl et al., 1997; Serbedzija et al., 1992).
Integrins have been shown to function in migration, regulate differentiation, participate in the assembly of extracellular matrix and serve as cellular mechanosensors (Abraham et al., 2008; Breau et al., 2006; Carlson et al., 2008; Ingber, 2006; Li et al., 2003; Wu et al., 1995; Xiao et al., 2010; Xiao et al., 2007). In this study we describe the role of β1 integrin in differentiated vSMCs and in a subset of vSMC progenitors. We discuss the contribution of different loxP flanked alleles and evaluate the outcome of β1 integrin deletion during the assembly of the vascular wall.
Deletion of β1 integrin in smooth muscle resulted in a range of aortic arch defects; however, the interrupted aortic arch was seen exclusively in smβ1e3/e3 embryos. We predict that this is due to differences in recombination between the two models used. The short distance between loxP sites in the β1e3 model yields a more efficient recombination than in the β1fl transgenic mouse and the β1fl allele may create a pseudo-plasmid enabling expression after recombination (Turlo et al., 2010). This may also explain the difference between phenotypes in the wnt1Cre (nc) mice. The persistence of some β1 integrin positive smooth muscle cells in the media of the smβ1fl/fl at E18.5 provides support to this concept. Regardless, the range of phenotypes in the different models highlights the significance of timing in relation to phenotypic outcomes and helps to identify distinct contributions of β1 integrin at progressive developmental stages. The embryonic lethality of the smβ1e3/e3 mice reported here uncovers critical contributions of β1 integrin in aortic arch remodeling and vascular wall assembly that were not revealed by deletion of β1 integrin using the PDGFRβCre (Abraham et al., 2008). In that model live knockouts were found as late as P(10) and inactivation of β1 integrin resulted in discrete dilations in small caliber vessels combined with adhesion defects, but no abnormalities were observed in large vessels or aortic arch remodeling (Abraham et al., 2008). Along the same lines, deletion of β1 integrin using the wnt1Cre resulted in greatly diverse phenotypes (β1fl versus β1e3 model). While no vascular defects were noted in the ncβ1fl/fl mouse; the ncβ1e3/e3 model resulted in fully penetrant lethality by E12.5 due to hemorrhage in the aortic arches. This range of outcomes provides a useful temporal framework in which to identify essential roles of β1 integrin in vascular development.
The smooth muscle of the aortic arch and carotids originates from the neural crest cells (Majesky, 2007). These progenitors migrate, differentiate and proliferate as the pharyngeal arches undergo extensive remodeling (Schwartz, 1997; Serbedzija et al., 1992). Previous work by Breau and colleagues indicated that inactivation of β1 integrin in neural crest (using the Ht-PA-Cre) resulted in a Hirschsprungs-like phenotype due to migration defects in enteric neural crest cells (Breau et al., 2006), but did not result in embryonic lethality. In contrast, we found that deletion of β1 integrin using Wnt1Cre in the β1e3 model (ncβ1e3/e3) was embryonic lethal due to defects in vessel wall structure, but migration of vSMC progenitors was not affected. This was a rather surprising result, as β1 integrin has been shown to impair migration in other cell types in vivo. Thus it appears that smooth muscle cells, or more likely their progenitors, utilize other receptors for migration during development.
Initiation of differentiation towards the smooth muscle fate was not blocked by lack of β1 integrin, but maintenance of some smooth muscle cell specific proteins was affected. We evaluated three early (α smooth muscle actin, SM22α, and SMMHC) and one late (calponin) proteins that are indicative of vSMC differentiation. Our findings indicate that while α-smooth muscle actin was not affected by absence of β1 integrin; SMMHC, SM22α and calponin, in particular, were reduced in smβ1e3/e3 at later time points, when compared to controls. It is thought that integrins recognize the ECM and signal to maintain contractile protein expression (Zargham et al., 2007). Our data supports such a concept with some qualifiers, as not all contractile proteins were affected equally by absence of this integrin.
While β1 integrin expression was found throughout the entire length of the aorta and major branches; incidence of aneurysms upon deletion was restricted to the aortic arch and its immediate branches. Similarly isolated defects of specific portions of the aorta and its branches occurred in the deletion of TGFβ, with locations of the defect depending on whether the deletion was partially penetrant in all mesenchymal vSMCs or completely penetrant within the neural crest (Choudhary et al., 2009). In our case, Western blots demonstrated an increase in β1 integrin expression in the aortic arch over the descending dorsal aorta at E18.5 (Figure 2D). This may explain why these areas are more vulnerable to the loss of β1 integrin. They may require greater amounts of β1 integrin as they experience high shear-stress, as indicated by computer modeling of flow in the carotid and vascular branches in general (Huo et al., 2008; Meng et al., 2007; Shojima et al., 2005). Further, the dilations seen in the ncβ1e3/e3 may also correspond to areas of high shear-stress based the wall shear stress diagrams in the pharyngeal arches of the chick (Wang 2009). The data indicates the cells migrate and differentiate, but they are unable to support flow in the pharyngeal arches, while the dorsal aorta remains intact. These defects may occur earlier in the ncβ1e3/e3 mouse as the cre recombinase targets a large population of smooth muscle progenitors prior to differentiation and some non-vSMCs as well.
Our early analysis of these dilations focused on the cellular component that poses physical resistance to these forces (ie., smooth muscle). However, we found no significant changes in smooth muscle cell number or ability to differentiate. Thus, we turned our evaluation to the extracellular matrix components. In fact early studies indicate that resistance to flow in the great vessels relies heavily on ECM proteins (Berry et al., 1975). This was supported by the fact that induced SMC death by approximately 50% had little to no impact on vessel diameter (Clarke et al., 2006), indicating that the ECM plays a critical role in the vessel wall by providing strength and resistance to repeated mechanical pressure.
Lessons from KO mice have indicated that the assembly of a compliant vascular wall requires the organization of a complex lattice of interacting proteins that include the adequate combination of fibronectin, fibulin, fibrillin, collagen, elastin and laminin (Wagenseil and Mecham, 2009). Our data further reveals that β1 integrins are required for the assembly and appropriate organization of those proteins into lamellar structures. These lattices flank concentric layers of smooth muscle cells; these, in turn, bind tightly to the organized ECM enabling the formation of an integrated tissue specialized in distention and contraction. Absence of β1 integrins imposes a developmental arrest in ECM assembly leading to structural defects in the organization of the ECM lattices and resulting in the development of aneurysms. Immunostaining of ECM at E15.5 carotids revealed that matrix is laid down at similar levels and patterns in smβ1e3/e3 and β1e3/e3 embryos. The discrepancy between the two groups becomes noticeable only at E18.5 highlighting a previously unsuspected, but critical contribution of β1 integrin in ECM architecture during vessel differentiation.
Prior to the identification of integrins, early observations showed an alignment between fibronectin and the cytoskeleton predicting the existence of a molecule that linked the two (Hynes and Destree, 1978). Subsequent studies have served to confirm the relationship between integrins, assembly of fibronectin as well as collagen fibrils (van der Flier et al., 2010; Wu et al., 1995; Li et al., 2003). Our findings suggest that, in addition to fibronectin, β1 integrins contribute to the organization of several other matrix molecules. An alternative interpretation is that the assembly of the fibronectin fibers provides the early patterning or scaffold that serves as the foundation for additional assembly of other matrix proteins.
The defects noted in smβ1e3/e3 mice are reminiscent of structural failure due to absence of specific matrix proteins. For example, it has been shown that mutations in fibrillin-1 can result in large vessel aneurysms (Mariko et al., 2011; Pereira et al., 1997). Fibrillin-1 contributes to the construction of elastic fibers, which participates in the elastic recoil that aids in the maintenance of blood flow (Wagenseil and Mecham, 2009). A similar phenotype is also shared by mice deficient in fibulin-4. These mice display ascending aortic aneurysms also thought to originate from defects in elastin assembly (Huang et al., 2010). Here deletion of β1 integrin prevents the organization ECM proteins leading to aneurysms and indicate that this molecule is also critical to the functional compliance of the ECM in the vessel wall. Overall these findings uncover a critical role for β1 integrin in vascular matrix assembly and protection against the development of aneurysms in the aortic arch and carotids.
In deleting β1 integrin using an sm22αCre line, we sought to evaluate the consequences of loss of function at the onset of vascular organization. Contrary to expectations, β1 integrin was not required for smooth muscle cell proliferation, migration or differentiation. Instead we found that β1 integrin was specifically required in smooth muscle cells for aortic arch remodeling and for the assembly of ECM proteins within the vessel wall. Similar phenotypes have been identified in mice lacking specific ECM molecules. Evaluation of the ECM in mutant mice indicates similar level of deposition, but lack of lamellar organization. Overall these findings support an essential role for β1 integrin in matrix assembly that is critical for vascular function. A nuance of the phenotype is that the effect was restricted to the ascending aorta, aortic arch and carotid arteries. The reason for this regional restriction is unclear our speculation is that the aortic arch and carotids require higher demands in ECM assembly that are incompatible with the absence β1 integrin.
Whole mount and histological sections of β-galactosidase stained smR26R embryos at E8.5, E9.5 and E10.5. Arrows indicate activity in the aorta and pharyngeal arches. (H=heart).
Immunofluorescence of β1 integrin (red), α-smooth muscle actin (green) and topro (blue) in the carotids of E15.5 and E18.5 β1e3/e3 and smβ1e3/e3 embryos and E18.5 β1fl/fl and smβ1fl/fl embryos (arrow=smooth muscle cell positive for β1 integrin).
A) Live neonates illustrating lack of perfusion in the smβ1fl/fl compared to β1fl/fl littermate. B) Plastinated (blue) carotids showing branch point dilations in smβ1fl/fl neonates against β1fl/fl control. C) Intact and isolated aortic arches from E18.5 β1fl/fl and smβ1fl/fl embryos (white arrows=carotid aneurysm, red arrow=aneurysm at base of pulmonary artery, A=atria, V=ventricle, DA=aorta). D) Isolated P(0) arches with aneurysm at carotid branch (white arrows) and in aortic arch (red arrows). E) Hearts from E18.5 β1fl/fl and smβ1fl/fl embryos (black arrows=hemorrhages). F) Frequency of aorta/carotid aneurysms and hemorrhages in the epicardium of E18.5 β1fl/fl and smβ1fl/fl embryos.
A) Whole mount and histological section of hearts from β1e3/e3 and smβ1e3/e3 embryos at E15.5 (black arrows=hemorrhages and ventricular septal defect). B) Immunofluorescence of ki67 (red) and topro (blue) in the ventricular septum of E15.5 hearts from β1e3/e3 and smβ1e3/e3 embryos. C) Quantification of width of compact layer in the right and left ventricular wall, the ventricular septum and the total area of heart sections for E15.5 β1e3/e3 (n=4) and smβ1e3/e3 (n=4). D) Whole mount and histological section of hearts from β1e3/e3 and smβ1e3/e3 embryos at E18.5 (black arrows=hemorrhages). E) Frequency of epicardial hemorrhage at E18.5 in the β1e3/e3 and smβ1e3/e3 animals.
Immunofluorescence of α-smooth muscle actin (green), sm22α (red), and topro (blue) in the pharyngeal arch arteries of E10.5 β1e3/e3 and smβ1e3/e3 embryos (H=heart, NT=neural tube, arrows indicate sm22α/α-sma double positive cells).
A) Immunofluorescence of ki67 (ki67=red), α-smooth muscle actin (green) and topro (blue) in the carotid of E15.5 and E18.5 β1e3/e3 and smβ1e3/e3 embryos (white arrow=positive for ki67). B) Immunofluorescence of cleaved caspase 3 (CC3=red), α-smooth muscle actin (green) and topro (blue) in the carotid of E15.5 and E18.5 β1e3/e3 and smβ1e3/e3 embryos.
Aortic arches in E15.5 β1fl/fl and ncβ1fl/fl embryos (arrows indicate branch points).
Immunofluorescence of α-smooth muscle actin (green), sm22α (red), and topro (blue) in the pharyngeal arch arteries of E10.5 β1e3/e3 and ncβ1e3/e3 embryos. (large white arrow=wild type pharyngeal arch vessel, red arrow=aneurysm, white box=wild type area of focus, red box=aneurysm area of focus, NT=neural tube, H=heart, small white arrows=α-sma/sm22α positive cells).
A) Schematic of a β1fl/fl aortic arch with lines (red) representing serial sections through the right (left) and left (right) carotids through the aortic arch. B) Trichrome stained serial sections of the carotids and aorta from a representative β1fl/fl embryo at E18.5. C) Schematic of an smβ1fl/fl aortic arch with lines (red) representing serial sections through the right (left) and left (right) carotids through the aortic arch. D) Trichrome stained serial sections of the carotids and aorta from a representative smβ1e3/e3 embryo at E18.5.
Immunofluorescence aortic arch cross-section diagramed approximately from β1e3/e3 and smβ1e3/e3 with collagen IV (red), α-smooth muscle actin (green) and topro (blue). (white arrows=gaps in media)
Expression of sm22α, calponin and smooth muscle myosin heavy chain (SMMHC) represented in red in α-smooth muscle actin (green) positive cells of the carotid (A) and dorsal aorta (B) of β1fl/fl and smβ1fl/fl at E18.5. (topro=blue).
Expression of A) fibronectin, B) collagen IV and C) elastin represented in red with α-smooth muscle actin (green) positive cells of the carotid at E15.5 and E18.5 and the dorsal aorta at E18.5. (topro=blue).
A) Whole mount β1e3/e3 and nc β1e3/e3 at E10.5 indicating pharyngeal arch artery (black arrow) and aneurysm (red arrow). B) Immunofluorescence of histological section from β1e3/e3 and ncβ1e3/e3 of collagen IV (red), α-smooth muscle actin (green) and topro (blue) with wild type pharyngeal arch arteries (white arrow & box) and aneurysm (red arrow and box). (NT=neural tube, H=heart). C) Immunofluorescence of wild type pharyngeal arch arteries from β1e3/e3 and ncβ1e3/e3 with collagen IV (red), α-smooth muscle actin (green) and topro (blue). D) Immunofluorescence of wild type pharyngeal arch arteries from β1e3/e3 and ncβ1e3/e3 with laminin (red), α-smooth muscle actin (green) and topro (blue). E) Immunofluorescence of an aneurysm in the pharyngeal arch arteries of ncβ1e3/e3 with collagen IV (red), α-smooth muscle actin (green) and topro (blue). F) Immunofluorescence of an aneurysm in the pharyngeal arch arteries of ncβ1e3/e3 with laminin (red), α-smooth muscle actin (green) and topro (blue).
Expression of laminin represented in red in α-smooth muscle actin (green) positive cells of the carotid and the dorsal aorta at E18.5 in the A) β1fl/fl and smβ1flfl and B) β1e3/e3 and smβ1e3/e3 (topro=blue).
Control for autofluorescence of ECM molecules with dorsal aorta from E18.5 β1e3/e3 animals stained with primary A) fibronectin, B) collagen IV and C) elastin antibodies followed by a Cy3 conjugated secondary (left), or with Cy3-conjugated secondary alone (right). Expression of ECM represented in red in α-smooth muscle actin (green) positive cells of the carotid and the dorsal aorta at E18.5. (topro=blue).
The authors would like to acknowledge Jason Scapa, Analila Valencia, Liman Zhao and Ana Marie Marquez for their technical assistance. We also recognize Dr. Karen Lyons for insightful conversations on this work. We would like to thank TPCL for their support in histological processing of our samples. This work was supported from funds from the National Institutes of Health (RO1HL 085618). Onika Noel is a trainee supported by the Vascular Biology Training Grant (T32 HL69766).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.