Central Roles of α5β1 Integrin and Fibronectin in Vascular Development in Mouse Embryos and Embryoid Bodies
Vascular development and maturation are dependent on the interactions of endothelial cell integrins with surrounding extracellular matrix. Previous investigations of the primacy of certain integrins in vascular development have not addressed whether this could also be a secondary effect due to poor embryonic nutrition. Here, we show that the α5 integrin subunit and fibronectin have critical roles in blood vessel development in mouse embryos and in embryoid bodies (EBs) differentiated from embryonic stem cells (a situation in which there is no nutritional deficit caused by the mutations). In contrast, vascular development in vivo and in vitro is not strongly dependent on αv or β3 integrin subunits. In mouse embryos lacking α5 integrin, greatly distended blood vessels are seen in the vitelline yolk sac and in the embryo itself. Additionally, overall blood vessel pattern complexity is reduced in α5-null tissues. This defective vascular phenotype is correlated with a decrease in the ligand for α5 integrin, fibronectin (FN), in the endothelial basement membranes. A striking and significant reduction in early capillary plexus formation and maturation was apparent in EBs formed from embryonic stem cells lacking α5 integrin or FN compared with wild-type EBs or EBs lacking αv or β3 integrin subunits. Vessel phenotype could be partially restored to FN-null EBs by the addition of whole FN to the culture system. These findings confirm a clear role for α5 and FN in early blood vessel development not dependent on embryo nutrition or αv or β3 integrin subunits. Thus, successful early vasculogenesis and angiogenesis require α5-FN interactions.
Blood vessels in vertebrate embryos can develop through either of 2 processes, vasculogenesis or angiogenesis.1,2⇓ In vasculogenesis, blood vessels are generated from mesodermally derived angioblasts, whereas in angiogenesis, vessels arise as sprouts from preexisting vessels by bridging or by intussusception.2 In the mouse, vessels are formed by the migration of angioblasts in the embryo and in the blood islands in the extraembryonic tissues. During vessel development, the endothelial precursors and differentiated cells are regulated by a number of environmental cues, including growth factors (such as fibroblast growth factors and vascular endothelial growth factors), cytokines, proteoglycans, extracellular adhesive glycoproteins, and interactions with the extracellular matrix.3–6⇓⇓⇓
Endothelial cell interactions with the extracellular matrix are mediated in large part by the integrin family of adhesion receptors, heterodimeric transmembrane glycoproteins, consisting of α and β subunits. At the cell surface, integrins can play adhesive as well as signaling functions.7,8⇓ Ligand specificity and signaling ability of specific integrins are determined by their heterodimeric composition. Endothelial cells have been shown to express a variety of integrins, including the following: α1β1, α2β1, and α3β1, which are laminin and collagen receptors; α5β1, αvβ1, and αvβ5, which are receptors for fibronectin (FN); α6β1, a laminin receptor; and αvβ3, a receptor for FN, vitronectin, osteopontin, von Willebrand factor, laminin, and collagen.4,7,9⇓⇓
A number of inhibition experiments in vivo and in vitro have indicated a role for endothelial-FN interactions in vascular development.10–12⇓⇓ Moreover, knockouts of FN have shown that it is essential for the organization of heart and blood vessels.13,14⇓ In the absence of FN, no blood vessels form in the vitelline yolk sac, whereas aortic endothelial cells in the embryo proper are scattered and disorganized. Furthermore, ablation of the α5 integrin in mice results in extensive vascular as well as mesodermal defects and early embryonic lethality,15,16⇓ and Kim et al17 have reported that antibody or peptide blockade of the α5β1-FN interaction interferes with angiogenesis. The αv integrins, in particular, αvβ3, have previously been implicated in a number of angiogenic functions through peptide- or antibody-blocking experiments.18–21⇓⇓⇓ Surprisingly, however, mouse embryos lacking all αv integrins display extensive vasculogenesis, angiogenesis, and organ development, leading to questions about the primacy of the αv integrins in vascular development.22 Mouse knockouts of the β3 integrin are also viable and fertile, with normal developmental angiogenesis and postnatal neovascularization of the retina.23
In the present study, we have addressed the role of the α5 integrin and its major ligand, FN, in mouse vascular development in greater detail with the use of whole embryos and quantifiable embryoid bodies (EBs) in in vitro assays. We report that α5-null embryos display marked decreases in the complexity of the vasculature that can be correlated with decreased FN matrix assembly and organization in α5-null endothelial basement membranes. In addition, in embryonic stem (ES) cells preferentially differentiated toward an endothelial lineage, primitive vessel formation is significantly reduced in α5-null and FN-null EBs compared with wild-type, β3-null, or αv-null EBs. Notably, vascular phenotype could be partially restored in FN-null EBs by the addition of whole FN to the culture system. These results strongly support a critical role for α5 integrin–FN interactions in the normal cellular processes involved in generating the embryonic vasculature.
The Methods section can be accessed online (please see http://www.ahajournals.org).
Swollen Vessels and Reduced Vessel Complexity in α5-Null Embryos
The heads of α5-null and wild-type embryos were of similar sizes at the stages observed. To obtain a more detailed examination of blood vessels in the α5-null embryos, whole-mount immunohistochemistry using an antibody recognizing platelet and endothelial cell adhesion molecule (PECAM)-1, a marker for endothelial cells, was performed at E8.5 and E9.5 stages (where E indicates embryonic day) in α5-null and wild-type littermates (Figure 1). PECAM-1 staining highlighted the abnormally swollen blood vessels in the α5-null vitelline membranes at E8.5 (Figure 1A and 1B) and E9.5 stages. Cross sections through the yolk sacs stained for PECAM-1 showed that the enlarged vessels in α5-null embryos were due to separation of the endodermal and mesodermal layers of the yolk sacs (Figure 1C and 1D). However, PECAM-1 staining also revealed a lining of endothelial cells around the walls of the dilated vessels. The staining also revealed a decrease in the complexity of the vascular network of the primary perineural plexus in α5-null embryos (Figure 1E through 1H). The cranial plexus mainly consisted of large vessels that branched less frequently in the null embryos compared with age-matched wild-type littermates.
Decreased FN Expression In Vivo in the Absence of α5
Because development and maintenance of the endothelium involves attachment to adhesive glycoproteins and the basement membrane,4,9⇓ the abnormal vessel patterning in the α5 knockouts led us to a closer examination of the endothelial basement membrane. It has been shown previously that abundant levels of FN are present in blood islands and the capillary plexus,13 whereas laminin, collagen, and other extracellular matrix molecules are produced by endothelial cells later in vasculogenesis.3
There was no reduction in the amount of mesoderm in the α5 compared with the wild-type yolk sacs.16 Staining of E8.5 yolk sac blood vessels and dorsal aortas for FN showed that less FN is deposited/retained in the matrix (Figure 2). In yolk sacs, FN was decreased in particular at the endoderm–endothelial basement membrane interface (Figure 2A and 2B), whereas laminin and collagen IV expression were similar in the 2 strains (Figure 2C through 2F). The expression of entactin was also unchanged (data not shown). Similarly, in the embryo, dorsal aortic expression of FN was decreased in null relative to wild-type embryos (Figure 2G and 2H, arrowheads), whereas FN expression in epithelial basement membranes, such as those surrounding the neural tube and the hindgut, remained equally strong in the null embryos compared with wild-type embryos.
These results extend earlier descriptions of the defects in α5-null embryos and show reduced FN deposition. FN-null embryos show similar or more severe defects,13,14⇓ confirming a key role for α5-FN interactions in vessel development in vivo. However, these results could be, in part, a secondary consequence of other defects in these embryos, such as nutritional deficiencies arising from vascular or other defects. To analyze the roles of α5 and FN in more detail without these attendant complications, we turned to an in vitro system.
Primitive Vessel Formation in Wild-Type and Integrin-Deficient EBs
To determine more clearly the role of the different integrins, we used a well-established model of early vascular plexus formation, the formation of EBs from ES cells.24,25⇓ EB development was monitored at 3, 4, 5, 7, 11, and 15 days after seeding by using planimetry. In view of the severe phenotype observed in the α5-null embryos, we anticipated difficulties in performing the EB assays. As predicted from the in vivo data, α5-deficient EBs were difficult to grow because the ES cells lacked cohesion and because EBs tended to break down as they grew larger (from day 11 onward). Despite this, cells that formed EBs did not differ significantly in size from wild-type cells (Table).
EB diameters were not significantly different at any time point for the ES cells lacking any of the various integrins or when endothelial growth-promoting factors were omitted from the methylcellulose-containing medium (Table). All EBs began to pulsate at day 8 to day 9 of culture, indicating the development of cardiomyocytes.
By use of confocal laser scanning microscopy of EBs stained for the expression of the endothelial marker PECAM-1, no vascular network was visible at 11 days in EBs cultured in the absence of endothelial growth-promoting factors (please see online Figure IA, which can be accessed at http://www.ahajournals.org). However, a complex lattice of PECAM-1–positive cells was visible in wild-type EBs from day 7 of culture under endothelial growth-promoting conditions, with large lacuna-like structures visible from day 11 (please see online Figure IB and Figure 3A). The diameters of these early vessel-like structures varied somewhat among EBs (wild type, 50±7 μm; αv null, 30±4 μm [P<0.05 compared with wild type]; and β3 null, 40±7 μm [all n=10]).
For wild-type, αv-null, α5-null, β3-null, and FN-null EBs, individual z series were combined and projected in 2D (Figure 3), and the percent area occupied by PECAM-1–positive cells was measured by drawing around the area bounded by PECAM-1–positive staining with the use of a hand-held mouse (Table). These data confirmed that there were no differences in EB diameter regardless of whether specific integrins were present or not. However, confocal sectioning of the EBs revealed marked differences in the occupation of the body by PECAM-1–positive cells. In wild-type EBs, >50% of the total area of the EB was occupied by PECAM-1–positive cells in contrast to 22% for α5-null cells (Table). Montages of confocal slices through these EBs (Figures 4 [wild type] and 5 [α5-null]) and 3D reconstruction of volume-rendered 3D images (data not shown) indicated that specification of cells occurred in both cases but that in α5-null EBs, the ability of the cells to form tubes and, therefore, lacuna-like structures appeared to be inhibited, with “islands” of PECAM-1–positive cells making only occasional contact (Figures 3B and 5⇓). Compared with wild-type cells, cells within the islands tended to be densely packed in layers, and the percent area occupied by PECAM-1–positive staining was reduced significantly (Table). In contrast to this, wild-type EBs exhibited a complex weblike pattern of PECAM-1–positive cells (Figures 3A and 4⇓). Similarly, EBs lacking αv or β3 also developed extensive PECAM-1–positive structures (Figure 3C and 3D and Table).
FN-Null EBs Contain Endothelial Cells, but These Do Not Organize Into Islands or Form Vascular Structures
EB assays were also performed by using FN-null ES cells. FN-null, like α5-null, EBs exhibited PECAM-1 staining but with no distinct pattern or organization within the EB (Figure 3E). In rescue experiments, in which FN-null EBs were cultured in the presence of 100 μg/mL whole FN, partial rescue of the null phenotype (an increase in PECAM-1–positive cells) occurred in all EBs examined (26±3.5% [rescued] versus 6±0.7% [(FN null], P<0.05; Table and Figure 3F.
The establishment and regulation of blood vessel growth is critical for normal mammalian embryonic development and for pathological processes in the adult, such as wound repair and tumorigenesis. Numerous studies have illustrated the importance of cell–extracellular matrix interactions (in particular, via the integrin family of adhesion/signaling receptors) in the mechanisms behind vascular development.4,26–33⇓⇓⇓⇓⇓⇓⇓⇓ A major focus of these investigations has been on the roles of the αvβ3 and αvβ5 integrins. Indeed, much evidence has suggested the importance of the αv integrins in angiogenesis19 (see reviews21,34⇓).
Recent studies have pointed to a more central role for β1 integrins in vascular development. In the absence of β1 integrins, fewer blood vessels form in a teratoma assay, and formation of a complex vasculature is delayed.35 The basement membrane also lacks laminin 1 in EBs.36 Senger et al37 have shown that angiogenesis induced by vascular endothelial growth factor can be inhibited by antibodies against α1β1 and α2β1. Knockouts of the αv22 and β323 integrins have shown much milder vessel defects than anticipated: αv-deficient mice display a complex embryonic vasculature, whereas β3-null mice are viable and fertile and show no vessel defects in perinatal retinal neovascularization. In the present study, we have shown that α5-null embryos exhibit a lower complexity of blood vessel formation correlating with reduced FN matrix assembly in vivo. Because there may be concerns that these integrin-deficient embryos could be nutritionally limited, we have reinforced these data by performing ES cell differentiation, a technique in which these limitations do not apply. In EB assays in vitro, α5-null and FN-null EBs are both unable to form any significant primitive vasculature. It is of interest that some vascular phenotype can be restored in FN-null EBs by the addition of whole FN to the culture system. Taken together, these data suggest that α5β1-FN interactions are necessary for basic cellular processes involved in normal vessel development and that the endothelial functions of α5 can be separated from those of αv and β3.
Using whole-mount PECAM staining of embryos and EBs, we have shown that the initial generation of endothelial cells occurs normally in the absence of α5 integrin, similar to the situation seen in FN-null embryos.38 Contrary to the extreme picture in the FN-deficient embryos, the α5-null endothelial cells do appear to organize themselves into vessels in the yolk sac and the embryo proper and into islands of endothelial cells in EBs. Because FN has been shown to be necessary for normal tube formation in the yolk sac, the initial vessel formation seen in the α5 knockout is probably due to the function of other FN receptors, such as the αv integrins, which, as we have shown (K.L. Goh, unpublished data, 2001), are expressed in α5-null primary endothelial cells. However, the α5-null blood vessels seen in the embryo are not completely normal, inasmuch as they are enlarged and, as illustrated by the head vessels, lack the complexity of pattern seen in the wild-type control vessels. Less capillary branching, ie, angiogenesis, seems to occur in the α5-null cranial plexus. This effect is also seen in α5-null EBs, in which (although islands of endothelial cells form) their ability to contact one another to form a network of tubes appears to be limited.
Preliminary analyses of α5-null cells in vivo and in EBs and of α5-null endothelial cells cultured from embryos suggest that cell proliferation and survival are somewhat reduced in culture (authors’ unpublished data, 2001). Similar results have been observed in α5-null teratocarcinomas,39 although they were not evident at early stages in the embryos.16 As for cell adhesion, α5-null endothelial cells show reduced adhesion to FN, as expected, but are normally adherent to other matrix proteins (laminin, vitronectin, and collagen IV; authors’ unpublished data, 2002).
In the α5-null yolk sacs, the distended blood vessels are accompanied by a separation between the endodermal and mesodermal layers, which is also seen in a more severe form in the FN-null yolk sacs.38 This separation is interesting because in situ differentiation of endothelial cells occurs primarily from mesodermal cells in contact with the endoderm,5,40,41⇓⇓ and separation of the endoderm from the mesoderm has been shown previously to result in the absence of a vascular network.42 The endoderm is of importance because it is thought to be the main source of basic fibroblast growth factor (bFGF), a factor required for normal vasculogenesis.2 Interestingly, treatment with bFGF has been shown to cause a significant increase in the surface expression of the α2β1, α3β1, α5β1, α6β1, and αvβ5 integrins in microvascular endothelial cells.43 In contrast, the levels of expression of α1β1 and αvβ3 were decreased in bFGF-treated cells. The addition of transforming growth factor-β1 and bFGF results in a synergistic induction of α5, with no significant changes in the expression of β1.44 Thus, it is possible that the failure of normal signaling from the endoderm could contribute to the vessel defects seen in the α5-null embryos.
Another possibility suggested by our data is that endothelial cells play an active role in organizing and assembling the FN matrix and that the failure to organize the matrix appropriately in the endothelial basement membrane could lead to defective endothelial cell adhesion and migration and, hence, to defects in vessel remodeling and angiogenesis. It has been previously shown that the profile of the subendothelial basement matrix changes as vascular development proceeds in the embryo, with FN being the earliest and most abundantly expressed matrix molecule (Risau and Lemmon3 and the present study). Moreover, the assembly of an FN matrix has been shown to influence a number of cellular functions, including the organization of intracellular cytoskeletal structures and changes in signaling pathways; eg, assembly of a native FN matrix has been shown to induce rapid formation of actin stress fibers and colocalization of α5β1 integrin, focal adhesion kinase, vinculin, and paxillin to regions of cell-matrix contact45 and is required for Rho GTPase activation and cell-cycle progression.46 In addition, and reinforcing the importance of FN as a primary matrix molecule in vessel formation, FN-null EBs exhibit a more severe defect in endothelial cell organization than that seen in α5-null EBs: PECAM-1–positive cell content is markedly reduced compared with other EBs, and no islands of endothelial cells are observed. It is indeed noteworthy that a partial vascular phenotype could be restored by the addition of whole mouse FN to the culture system. A possible reason that only a partial rescue was observed may be the difficulty of access and/or inadequate concentration of FN available to the growing EB.
All the experiments performed have suggested that α5 may be a critical player in organizing the FN matrix underlying the endothelial cells during periods of blood vessel development in the embryo or in angiogenesis in teratomas, contributing to the normal assembly of the endothelial basement membrane. This matrix-organization function of α5β1 may explain in part the observations in β1-null teratomas, in which diffuse patterns of FN matrix, irregular basement membranes, and a poor vasculature have been detected.35,36⇓
Our results may serve as an explanation for discrepancies in the literature concerning the role of integrins in angiogenesis. As mentioned previously, an important role for angiogenesis has been suggested for αvβ3 and αvβ5 integrins.47 However, examination of β1-null teratomas, which display abnormally developed vasculature, by Bloch et al35 revealed that αvβ3 and αvβ5 integrins were unchanged. The recently described mild vascular phenotypes of αv knockouts,22 in which 20% survive to birth, and β3 knockouts,23 which are viable and fertile, also raise the question of the necessity for αv integrins in embryonic vasculogenesis and angiogenesis. Our results suggest that it is probable that α5β1, perhaps along with α1β1 and α2β1, has critical functions in regulating early vessel formation, independent of αv and β3 integrins.
In conclusion, we present a detailed look at the critical involvement of α5 integrin in the cellular processes involved in vascular development in vivo and in ES cell cultures. Using α5-deficient mice and EBs, we have revealed the importance of α5 integrin and FN interaction in vascular development and have shown that EB vasculogenesis is not strongly dependent on either αv or β3.
S.E.F. was supported by a Fulbright Scholarship and a British Heart Foundation Travel Fellowship. K.L.G. was supported by a predoctoral fellowship from the Howard Hughes Medical Institute (HHMI), and R.O.H. is an investigator of HHMI. We are grateful to Denise Crowley for histology and to Joe McCarty and Julie Lively for helpful comments. We thank Joy Yang and Daniela Taverna for the α5-knockout mice. We also thank Bio-Rad UK and The University of Oxford for additional assistance with confocal microscopy.
Received August 7, 2001; revision accepted March 4, 2002.
- ↵Luscinkas FW, Lawler J. Integrins as dynamic regulators of vascular function. FASEB J. 1994; 8: 929–938.
- ↵Short SM, Talbott GA, Juliano RL. Integrin-mediated signaling events in human endothelial cells. Mol Biol Cell. 1998; 9: 1969–1980.
- ↵George EL, Georges-Labouesse EN, Patel-King RS, Rayburn H, Hynes RO. Defects in mesoderm, neural tube, and vascular development in mouse embryos mutant for fibronectin. Development. 1993; 119: 1079–1091.
- ↵George EL, Baldwin HS, Hynes RO Fibronectins are essential for heart and blood vessel morphogenesis but are dispensable for initial specification of precursor cells. Blood. 1997; 90: 3073–3081.
- ↵Yang JT, Rayburn H, Hynes RO. Embryonic mesodermal defects in α5 integrin-deficient mice. Development. 1993; 119: 1093–1105.
- ↵Goh KL, Yang JT, Hynes RO. Mesodermal defects and cranial neural crest apoptosis in α5 integrin-null embryos. Development. 1997; 124: 4309–4319.
- ↵Drake CJ, Cheresh DA, Little CD. An antagonist of integrin αvβ3 prevents maturation of blood vessels during embryonic neovascularization. J Cell Sci. 1995; 108: 2655–2661.
- ↵Friedlander M, Brooks PC, Shaffer RW, Kincaid CM, Varner JA, Cheresh DA. Definition of two angiogenic pathways by distinct alpha v integrins. Science. 1995; 270: 1500–1502.
- ↵Friedlander M, Theesfeld CL, Sugita M, Fruttiger M, Thomas MA, Chang S, Cheresh DA. Involvement of integrin αvβ3 and αvβ5 in ocular neovascular disease. Proc Natl Acad Sci U S A. 1996; 93: 9764–9769.
- ↵Hodivala-Dilke KM, McHugh KP, Tsakiris DA, Rayburn H, Crowley D, Ullman-Cullere M, Ross FP, Coller BS, Teitelbaum S, Hynes RO. Beta3-integrin deficient mice as a model for Glanzmann thrombasthenia showing placental defects and reduced survival. J Clin Invest. 1999; 103: 229–238.
- ↵Vittet D, et al. Embryonic stem cells differentiate in vitro to endothelial cells through successive maturation steps. Blood. 1996; 88: 3424–3431.
- ↵Vittet D, et al. Targeted null mutation in the vascular endothelial-cadherin gene impairs the organization of vascular-like structures in embryoid bodies. Proc Natl Acad Sci U S A. 1997; 94: 6273–6278.
- ↵Hirashima M, Kataoka H, Nishikawa S, Matsuyoshi N, Nishikawa S. Maturation of embryonic stem cells into endothelial cells in an in vitro model of angiogenesis. Blood. 1999; 93: 1253–1263.
- ↵Madri JA, Williams SK. Capillary endothelial cell cultures, phenotypic modulation by matrix components. J Cell Biol. 1983; 97: 153–165.
- ↵Ingber DE, Folkman J. Mechanochemical switching between growth and differentiation during fibroblast growth factor-stimulated angiogenesis in vitro: role of extracellular matrix. J Cell Biol. 1984; 109: 317–330.
- ↵Hynes RO, Hodivala-Dilke KM. Insights and questions arising from studies of a mouse model of Glanzmann thrombasthenia. Thromb Haemost. 1999; B2: 481–485.
- ↵Bloch WE, et al. β1 integrin is essential for teratoma growth and angiogenesis. J Cell Biol. 1997; 139: 265–278.
- ↵Aumailley M, Pesch M, Tunggai L, Gaill F, Fassler R. Altered synthesis of laminin 1 and absence of basement membrane component deposition in β1 integrin-deficient embryoid bodies. J Cell Sci. 2000; 113: 259–268.
- ↵Senger DR, Claffey KP, Benes JE, Perruzzi CA, Sergiou AP, Detmar M. Angiogenesis promoted by vascular endothelial growth factor, regulation through α1β1 and α2β1 integrins. Proc Natl Acad Sci U S A. 1997; 94: 13612–13617.
- ↵George EL, Baldwin HS, Hynes RO Fibronectins are essential for heart and blood vessel morphogenesis but are dispensable for initial specification of precursor cells. Blood. 1997; 90: 3073–3081.
- ↵Taverna DT, Hynes RO. Reduced blood vessel formation and tumour growth in α5-integrin-negative teratocarcinomas and embryoid bodies. Cancer Res. 2001; 61: 5255–5261.
- ↵Pardanaud P, Yassine F, Dieterlen-Lievre F. Relationship between vasculogenesis, angiogenesis and hematopoiesis during avian ontogeny. Development. 1989; 105: 473–485.
- ↵Palis J, McGrath KE, Kingsley PD Initiation of hematopoiesis and vasculogenesis in murine yolk sac explants. Blood. 1995; 86: 156–163.
- ↵Klein S, Giancotti FG, Presta M, Albeda SM, Buck CA, Rifkin DB. Basic fibroblast growth factor modulates integrin expression in microvascular endothelial cells. Mol Biol Cell. 1993; 4: 973–982.
- ↵Collo G, Pepper MS. Endothelial cell integrin α5β1 expression is modulated by cytokines and during migration in vitro. J Cell Biol. 1999; 112: 569–578.
- ↵Bourdoulous S, Orend G, MacKenna DA, Pasqualini R, Ruoslahti E. Fibronectin matrix regulates activation of RHO and CDC42 GTPases and cell cycle progression. J Cell Biol. 1998; 143: 267–276.