| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Genetics and Pathology, Uppsala University, Uppsala, Sweden;
Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden; and
Department of Human Anatomy, University of Liverpool, Liverpool, UK
1Correspondence: Department of Genetics and Pathology, Uppsala University, Rudbeck Laboratory, Dag Hammarskjöldsv 20, SE-75185, Uppsala, Sweden. E-mail: lena.welsh{at}genpat.uu.se
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4- and 5-chain laminins, fibronectin, collagen IV, and HSPGs. In sprouting angiogenesis, laminins were produced by stalk cells, as well as the leading tip cell, and deposited along the sprout length, including tip cell filopodia. In embryonic stem cells deficient in laminins, due to lamc1 (laminin 
1) deletion, vascular development and organization were largely unaffected. However, the frequency of vessels with wide lumens was increased 4-fold. Laminin-deficient vessels were moreover characterized by increased fibronectin levels and enhanced endothelial cell proliferation. We conclude that laminins are dispensable for vascular development but that they regulate lumen formation in the absence of flow and vascular tone.—Jakobsson, L., Domogatskaya, A., Tryggvason, K., Edgar, D., Claesson-Welsh, L. Laminin deposition is dispensable for vasculogenesis but regulates blood vessel diameter independent of flow.
Key Words: angiogenesis • sprouting • basement membrane • VEGF
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Angiogenesis is initiated by vascular endothelial growth factor (VEGF), which activates vascular endothelial growth factor receptor 2 (VEGFR-2) on the EC (1)
. VEGFR-2 activation triggers the release of proteases that degrade the BM, accompanied by detachment of smooth muscle cells. As a result, ECs participate in a program of active angiogenesis, proliferating and migrating into the surrounding tissue. Although breakdown of the BM is considered vital for generation of new blood vessels, little is known about the deposition and assembly of BM components during blood vessel formation. Moreover, a number of BM-derived peptides (for a review, see ref. 2
) are known to possess antiangiogenic properties in vitro as well as in vivo, but their relevance for vessel stability and remodeling is unknown. Maturation of the newly formed vessel is thought to involve thickening of the BM and recruitment of supporting cells (pericytes) by the secretion of platelet-derived growth factor (3)
.
Most molecules residing within the BM, such as laminins (LMs), collagens, fibronectin, and heparan sulfate proteoglycans (HSPGs), have been associated with diverse processes such as cell differentiation, attachment, migration, polarization, guidance, and survival (4)
. LMs constitute a family of heterotrimeric glycoproteins, with each family member containing one -, β-, and -chain. To date 5 (1–5), 3 β (β1–3), and 3 (1–3) chains are known to combine into at least 16 mammalian isoforms (including the major splice variants), which show developmental stage-specific and cell type-specific expression patterns.
The LM -chains are highly bioactive with binding sites for integrins, dystroglycans, and HSPGs. In addition, the -chains are more restricted in their expression than the β- and -chains. The only LM -chains present in the vascular BM are the 4- and 5-chains, which are expressed and deposited during mouse development, at least from embryonic day (E) 11–13 and onwards (5
, 6)
; for a review, see ref. 7
. Deletion of LM-5 results in embryonic lethality at 17.5 days postcoitum with brain and limb defects, enlarged and disorganized placental vessels, and aberrant renal glomeruli vascularization (8
, 9)
. The LM-4 null mice are viable and fertile but exhibit prenatal vascular bleedings, a phenotype that is gradually lost during development (10)
. Many of the LM-chain specific knockouts result in embryonic lethality (gene targeting of 1 results in death at E7, β1 at E5.5, and 1 at E5.5) before onset of vasculogenesis (11
, 12)
.
To overcome problems with implantation and early embryonic lethality, we have studied embryonic stem cells with targeted deletions of lamc1 (the gene encoding LM-1). By in vitro differentiation of wild-type and mutant stem cells we have been able to monitor the differential expression of the LM chains and isoforms and to study the role of LMs in vasculogenic and angiogenic processes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, were kind gifts from Dr. Andras Nagy and Dr. Janet Rossant (Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada), respectively. The lamc1–/– ES cells were established as described by Smyth et al. (12)
.
Embryoid bodies
Culturing of ES cell clones and generation of embryoid bodies were as described previously (14)
. At day 4, the embryoid bodies were placed in 8-well glass culture slides (BD Falcon; BD, Franklin Lakes, NJ, USA) in the presence or absence of LM111, LM411 (15)
, or LM511 (16)
at a concentration of 20 µg/ml. Alternatively, embryoid bodies were placed in a polymerized collagen I gel, as described previously (14)
, with the addition of 32.4 µg/ml LM111, LM411, or LM511. Medium with or without growth factors was changed at day 8.
Proliferation
Bromodeoxyuridine (BrdU) was added to the cultures at day 14, and cultures were processed for immunostaining 24 h later. Proliferative cells were visualized using the anti-BrdU nuclease kit (Amersham Pharmacia Biotech, Uppsala, Sweden) in combination with Alexa 488 goat anti-mouse highly cross-absorbed IgG (Molecular Probes, Eugene, OR, USA).
Immunostaining of embryoid bodies
Embryoid bodies on glass slides and in collagen I gels were processed for whole-mount immunohistochemical staining as described previously (14)
. The following primary antibodies were used for detection: rat anti-mouse CD31 antibody (PharMingen, San Diego, CA, USA); fluorescein isothiocyanate (FITC) -conjugated mouse monoclonal anti--smooth muscle actin antibody, goat anti-mouse VEGFR-2 (R&D, Minneapolis, MN, USA); rabbit anti-NG2, goat anti-collagen IV (Chemicon, Temecula, CA, USA); IgM mouse anti-mouse HepSS-1 antibody (Seikagaku, Tokyo, Japan); rabbit antiserum to LM-4 (clone377) (kindly provided by Dr. Lydia Sorokin, Universitätsklinikum Münster, Münster, Germany; ref. 17
); rabbit antiserum to LM-5 (kindly provided by Dr. Jeffrey H. Miner, Washington University School of Medicine, St. Louis, MO, USA; ref. 18
); rabbit anti-pan LM (Sigma, St. Louis, MO, USA); rabbit anti-fibronectin (Dako, Copenhagen, Denmark). Secondary antibodies: Alexa 568 goat anti-rat IgG, Alexa 488 goat anti-rat IgG, Alexa 555 donkey anti-goat IgG, Alexa 488 donkey anti-rabbit, Alexa 555 goat anti-mouse IgM (Molecular Probes); Cy5 donkey anti-rabbit IgG (Jackson ImmunoResearch, West Grove, PA, USA); biotinylated goat anti-rat IgG antibody, horseradish peroxidase-conjugated streptavidin (Vector, Burlingame, CA, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 19)
, embryonic stem cells differentiate into ECs that form capillary-like structures within the embryoid body, faithfully mimicking different stages in vasculogenesis and angiogenesis (Fig. 1
A) (for a review, see ref. 20
). We analyzed the presence and composition of the endothelial BM in this model.
|
Immunostaining of day 10 embryoid bodies showed LM-4 (a component of LM411 and LM421) deposition on the surface of the endothelium (Fig. 1B
). This agrees with previous observations of murine LM-4 in the vascular BM at E13. Immunoblotting showed expression of LM-4 as early as day 6 of differentiation in embryoid bodies (Supplemental Fig. 1). LM-4 preferentially colocalized with the EC protein CD31/PECAM (platelet-EC adhesion molecule) -1. LM-5 (a component of LM511 and LM521), which is present in the mature vascular BM, outlined the vascular structures but was also widely distributed throughout the embryoid body. In accordance, staining with the general LM-pan antibody recognizing the 1-, β1-, and 1-chains, and thus 11 out of 16 known isoforms, showed abundant and extensive expression of LMs. Collagen IV, another major component of the vascular BM, displayed an expression pattern similar to that of LM-4 but with a slightly wider distribution. Fibronectin, which is known to bind and activate integrin vβ3, thereby affecting angiogenesis, was detected in the EC BM, but also elsewhere (Fig. 1B
).
To examine the composition of the vascular BM during sprouting angiogenesis, embryoid bodies were cultured in a 3-D matrix of collagen I. Treatment with VEGF induced invasion of ECs into the surrounding matrix, as described previously (14)
(Fig. 2
). At day 12, embryoid bodies showed intense sprouting of capillary-like structures composed of CD31-positive ECs, surrounded by NG2 and/or SMA-positive pericytes (Fig. 2)
. Fibronectin, collagen IV, and LM were all present in the vascular BM stretching from the stalk to the tip cells, embedding both ECs and pericytes (Figs. 2
and 5
and data not shown). Immunostaining to detect heparan sulfate (HS) revealed the presence of HSPGs in the vascular BM. We conclude that vascular structures in embryoid bodies are embedded in an in vivo-like vascular BM, thereby constituting a relevant model for the study of the contribution of LMs to EC development and organization.
|
|
|
|
Vascular development proceeds in the absence of LM deposition
The LM-1 chain has been detected previously in embryoid bodies differentiated for 48 h (12)
. Inactivation of the lamc1 gene encoding the LM-1 chain in the mouse leads to embryonic death at E 5.5 (21)
. Embryoid bodies formed from lamc1–/– ES cells still supported formation of ECs and their organization into vascular structures. The resulting vascular plexus were less branched than seen in wild-type cultures, and ECs were more spread out (Fig. 3
A, compared with Figs. 1A
and 6A
, B). As expected, no reactivity against LM-1 (Supplemental Fig. 2A) was found, whereas diffuse, predominantly intracellular patches of other LM chains were detected (Fig. 3B
). Immunostaining with antibodies for LM-4 and LM-5 and using the LM-pan antibody consistently demonstrated a complete absence of deposited LMs in embryoid bodies lacking LM-1 expression (Fig. 3B
). Also, polymerized collagen IV was severely reduced in the vascular BM and instead found evenly dispersed over the culture (Fig. 3B
, compare confocal images of lamc1 –/– cultures with wild-type in Fig. 3C
). However, increased levels of fibronectin were detected in the absence of LM-1 expression (Fig. 3B, C
and Supplemental Fig. 3).
|
We next asked whether LM-1 deficiency would affect angiogenic sprouting (Fig. 4
). Although there was a close to complete lack of LM expression and deposition in angiogenic sprouts from LM-1-deficient embryoid bodies, the number and length of sprouts were indistinguishable from those of wild-type embryoid bodies (Fig. 4)
. Furthermore, the extent of pericyte coating was similar in wild-type and lamc1–/– angiogenic sprouts. Collagen IV was expressed, but its perivascular deposition was absent in the sprouts of LM-1-deficient embryoid bodies. In contrast, perivascular deposition of fibronectin was clearly seen and vascular BM also contained HSPGs, as detected by HepSS1 immunostaining. We conclude that, despite the dramatic effects on the BM composition with both LM and collagen IV deficiency in the absence of LM-1, no obvious vascular phenotype was recorded with regard to pericyte investment and ability of EC sprouting.
LMs are produced and deposited from the stalk to the tip
To investigate which part of the angiogenic sprout is responsible for the deposition of LMs within the vascular BM, we analyzed sprouting tips by high-resolution confocal microscopy. Angiogenic sprouts formed during development are known to be composed of proliferating stalk ECs, headed by a growth-arrested tip cell, which guides the path of the sprout (22)
. Immunostaining of wild-type embryoid bodies revealed the presence of LMs spanning from the stalk cells to the extending filopodia of the leading tip cell (Fig. 5
and Supplemental Movie 1). This was true also when staining for LM-4 (Supplemental Fig. 4). In the LM-1-deficient cells, no extracellular LM deposition was observed, while intracellular reactivity was evident in stalk cells as well as the tip cells indicative of LM production in both cell populations (Fig. 5
and Supplemental Movie 2). This implies that the ECs throughout the sprout produce the vascular BM.
EC differentiation and organization in the presence of exogenous LMs
Addition of purified LMs has been shown to rescue or modulate different processes in the embryoid body model (23
24
25
26)
. Figure 6
A shows that addition of purified LM111, LM411, or LM511 had no appreciable effect on vascular development in differentiating wild-type embryoid bodies. In contrast, addition of purified, recombinant LMs to the LM-1-chain-deficient cultures enhanced EC differentiation, as shown by the additional branch points and slender vessel formation (Fig. 6B
). The most prominent effect was seen with LM111 and slightly less in LM411- and LM511-treated cultures, which is consistent with the extent of the BM rescue as determined by LM deposition (Supplemental Fig. 2B). As judged from immunostaining using the LM-pan antibody, none of the treatments resulted in complete restoration of LM deposition.
Vascular stability in the absence of a LM-containing vascular BM
LMs have been suggested to influence vascular stability and to serve as a path finding track both during developmental angiogenesis in the zebra fish and during vessel regeneration in the kidney (27
, 28)
. To investigate the effects on vessel stability, survival, and regeneration, we compared angiogenic sprouting in wild-type and LM-1-deficient embryoid bodies over time. VEGF was added to the cultures for the first 10 days, followed by withdrawal of VEGF between days 10 and 14, and reintroduction of VEGF again at day 14 until day 18 (Fig. 7
). At day 10, sprouts of ECs of wild-type as well as in LM-1-null embryoid bodies invaded the gel. Two days after withdrawal of VEGF, signs of disassembly of the sprouts were evident (data not shown). At day 14, the vascular sprouts were almost completely dissolved (one or two short sprouts remained per embryoid body; Fig. 7B
). The kinetics of the retraction of the vessels was similar in wild-type and lamc1–/– cultures, suggesting that LM does not enhance EC survival in this setting. In conjunction with EC dissociation, the pericyte coat was also dissolved to a large extent. The kinetics of this process followed that of the EC degradation, i.e., with no apparent difference between wild-type and LM-1-deficient embryoid bodies. Reintroduction of VEGF at day 15 promoted restoration of vascular structures (Fig. 7)
. These structures failed to invade the collagen I matrix and were found close to the embryoid body "core." We did not observe signs of regrowth of ECs in vascular BM sleeves, as has been shown to occur after interruption of therapeutic VEGF neutralization (28)
. It has been previously shown that tumor vessels lacking mural cell coverage are more sensitive to VEGF depletion than vessels of normal composition (29)
. Consistent with this report, the few remaining EC sprouts after VEGF withdrawal were completely covered with SMA-positive cells, to an extent never observed with continuous VEGF treatment (Fig. 7B
).
|
LMs control the vascular lumen diameter
The assembly of a proper BM with deposited LMs has been shown to be a key event in polarization of several cell types, and apical-basal polarization of ECs is a prerequisite for lumen formation. To evaluate the importance of LM deposition for proper lumen formation we analyzed angiogenic sprouts formed from wild-type and lamc1–/– embryoid bodies by confocal microscopy. Lumen-containing vessels were observed in wild-type as well as in lamc1–/– embryoid bodies (Fig. 8
A). The number of vessels with a diameter exceeding 36 µm was approximately 4x higher in LM-1-null embryoid bodies compared to wild-type from day 14 and onwards (Fig. 8B
). This finding is consistent with the larger vascular lumens observed in both LM-4 and LM-5 knockout animals (8
, 10)
. Addition of purified, recombinant LMs did not affect sprout features in wild-type or lamc1–/– embryoid bodies (Supplemental Fig. 5 and data not shown). Therefore, in order to analyze rescue of lumen diameter, we used chimeric embryoid bodies to allow directed delivery of endogenous LMs.
|
In embryoid bodies composed of a mixture of lamc1–/– and vegfr-2–/– ES cells, a partial normalization of the lumen diameters was observed (Fig. 8B
). VEGFR-2 is a prerequisite for the expansion, differentiation and migration of ECs. Thus, ECs in this chimeric setting are derived from the lamc1–/– ES cells and lack LM expression, whereas non-ECs (i.e., pericytes) are either of vegfr-2–/– origin (producing LMs) or lamc1–/– origin. In cultures initially containing 90% lamc1–/– cells and 10% vegfr-2–/–, a patchy pattern of extracellular LMs was observed in the embryoid body sprouts, although at considerably lower levels than in wild type (Fig. 8A
).
To examine whether the lumen enlargement was caused by differences in cell morphology and/or proliferation, the number of ECs of sprouts with certain diameter from both conditions was counted. No difference in the number of ECs/sprout was recorded when comparing vessels of similar size derived from wild-type and lamc1–/– embryoid bodies. However, a correlation was found between increased vessel width and increased number of ECs/sprout length (Supplemental Fig. 6
). We, therefore, investigated possible differences in EC proliferation in wild-type and lamc1–/– embryoid bodies during the initial steps of lumen formation. Quantification of cells that had incorporated BrdU over a 24 h period revealed increased EC proliferation in the absence of LMs (Figs. 8C, D)
. Despite large variation within each group, a clear difference in the mean value was noted. This finding suggests that the tendency to increased proliferation in absence of LMs paralleled by increased fibronectin retention may be decisive for the luminal expansion. The increase in lumen size in the absence of LM-4 in gene targeted mice has been inferred to be due to a weakening of the vessel wall and consequent widening of the lumen. Our data, however, suggest that LMs either provide or restrict signals to control vascular lumen formation in the absence of blood pressure and circulation.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4- and LM-5-deficient mice. We conclude that the increase in lumen diameter is independent of blood flow, which is lacking in embryoid bodies. Rather, control of lumen diameter by LM deposition in the BM may involve integrin ligation and integrin-dependent signal transduction in ECs, similar to that suggested for epithelial cells (30)
. That EC integrin-matrix interactions indeed may be altered in the absence of LM deposition was indicated by increased EC proliferation and the more spread-out morphology of ECs in the lamc1–/– embryoid bodies (see Figs. 6
and 8
).
The presence of LM-4 and LM-5 chains in the vascular BM of embryoid bodies at all stages of differentiation implies an important role for these LMs in EC biology. Previously, the LM-5 chain has not been detected in vascular BM until E13 and has, therefore, not been considered vital for vasculogenesis. The increase of the LM-5 chain over time in the vascular BMs suggests a function in stabilization rather than early differentiation of ECs.
LM-4 is the first -chain detected in the vascular BM during development, and gene inactivation results in a relatively mild embryonic vascular phenotype with enlarged vascular lumens and bleedings. The vascular phenotype decreases after birth, and animals are fertile and unaffected under nonpathological conditions. This surprisingly mild phenotype might be explained by compensatory activity of other LM chains during early development. Notably, no -chains were detected in vascular BMs at birth, but increasing levels of the LM-5 chain was evident at later stages (10)
.
Gene inactivation of LM-1, which is a component of 10 of the 16 known LMs, leads to embryonic death at E5.5 due to defects in Reichert’s membrane and disturbed implantation (12
, 21)
. To study developmental processes at later stages, Li and co-workers employed ES cells with targeted deletion of the lamc1 gene (24)
. None of the known LMs were deposited in epithelial BMs in the LM-1–/– embryoid bodies. Furthermore, collagen IV failed to assemble due to lack of retention to the BM. In this study, however, both LM and collagen IV assembly was rescued by exogenous LM111. Overall, these data agree with our results on the features of vascular BMs in lamc1–/– embryoid bodies. Our observations deviate slightly in that polymerization of collagen IV, although considerably reduced, was still evident in the absence of the LM-1 chain. Furthermore, increased levels of fibronectin in the LM-1-deficient embryoid bodies were found. This increase might be a result of compensatory up-regulation of fibronectin, a mechanism not previously described, or more likely due to increased retention of fibronectin due to lack of competition with the LM-5 chain for binding to integrins vβ3. The critical role of fibronectin in mammalian biology is demonstrated by the lethality of fibronectin-null mice at E11 with defective notochord and somite formation as well as vascular malfunctions (31)
.
An elegant study by Parsons et al. (27)
showed that inactivation of the LM-1 gene in zebra fish leads to a pronounced vascular phenotype. The dorsal aorta in lamc1–/– zebra fish formed normally, but no sprouting was seen from intersomitic vessels, a defect possibly related to the defective perinotochordal BM organization. This finding clearly demonstrates the importance of extracellular LMs in patterning of the vasculature in this model. Nevertheless, due to potential contribution of maternally derived LMs at early embryonic development, it did not address the role for LMs in early EC differentiation.
Although deficiency in LM deposition affects the morphology of early ECs in the embryoid body model (Fig. 6)
, we conclude that EC differentiation still occurs. This report provides the first evidence that LMs are dispensable for vasculogenesis in the mammalian system. Fine-tuning of vessel morphogenesis appears to involve LMs, as both LM111 and LM511 noticeably rescued the LM-1-deficient phenotype in the embryoid body cultures (see Fig. 6
) and resulted in more branch points and slender vessels. Furthermore, collagen IV seems not to be required for early vascular development since it was essentially absent in the BM of LM-1-deficient embryoid bodies. In agreement with this observation, the collagen IV knockout mouse shows normal composition of the BM until E9.5 but is lethal around E11 and has a mild vascular phenotype with irregular protrusion of capillaries into the neural layer (32)
.
The degree of BM and pericyte coating has been correlated with vessel stability. Hence, tumor vessels often show a deficient perivascular coat that has been associated with the higher vessel turnover in this condition. In this study, however, even a markedly compromised vascular BM lacking LMs and collagen IV did not affect vessel stability and EC survival. Furthermore, withdrawal of VEGF led to vessel disintegration with similar kinetics in wild-type and lamc1–/– embryoid bodies (Fig. 7)
. It is interesting that angiogenic sprouts remaining after several days of VEGF deficiency were densely coated with pericytes, indicating that pericytes rather than the BM itself are critical in EC survival and vessel stability (14
, 28
, 33)
. Alternatively, it is also possible that fibronectin becomes more efficiently retained in the vascular BM in the absence of LM-1 and may compensate by mediating survival signals to ECs (34)
.
In conclusion, although our data strongly indicate that LMs are not required for vasculogenesis per se, we concur that fine-tuning of the 3-D aspect of the vascular tube involves LM-dependent signal transduction and that in vivo conditions of flow and mechanical forces would further compromise vascular function under conditions of LM-deficiency.
|
ACKNOWLEDGMENTS |
|---|
5 and LM-4 antiserum, respectively. This study was supported by grants to L.C.W. from the Swedish Science Council (project K2005-32X12552-08A) and the Swedish Cancer Foundation (project 3820-B05-10XBC). Received for publication September 9, 2007. Accepted for publication November 8, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. De Smet, I. Segura, K. De Bock, P. J. Hohensinner, and P. Carmeliet Mechanisms of Vessel Branching: Filopodia on Endothelial Tip Cells Lead the Way Arterioscler. Thromb. Vasc. Biol., May 1, 2009; 29(5): 639 - 649. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
