| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Cancer Biology and Angiogenesis, Department of Pathology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USA
1Correspondence: Department of Pathology, Research North, Beth Israel Deaconess Medical Center, 99 Brookline Ave., Boston, MA 02215, USA. E-mail: dsenger{at}caregroup.harvard.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: angiogenesis • collagen • laminin • ß1 integrins • VE-cadherin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. The process whereby sprouting ECs reorganize into new capillary networks within the interstitial matrix has been studied most extensively in the embryo, where capillary development starts with transition of endothelial precursor cells to a spindle-shaped morphology (3)
. Coincident with this spindle-shaped transition, EC precursors align into solid, precapillary cord-like structures (4
, 5)
that are interconnected to form a polygonal network (3
, 6)
. Solid precapillary cords also form during angiogenesis in the adult (7)
. During maturation, solid vascular cords form hollow lumens for the transport of blood (3
, 7)
and ECs are again sequestered from the interstitial matrix through establishment of a continuous basal lamina.
Interestingly, interstitial collagen I provokes ECs in vitro to initiate a morphogenetic program that imitates the early stages of capillary formation in vivo. After addition of collagen I to monolayer cultures, ECs retract and realign to form solid precapillary cords organized in a polygonal pattern (8
9
10
11
12)
. Furthermore, expression of collagen I by isolated EC clones closely correlates with spontaneous capillary morphogenesis (13
, 14)
. In ECs isolated from the microvasculature, collagen I-induced capillary morphogenesis is mediated by the 
1ß1 and 2ß1 integrins (15)
and expression of these integrins is selectively induced by the angiogenic cytokine VEGF (16)
. Moreover, antagonism of the 1ß1 and 2ß1 integrins inhibits angiogenesis in vivo (16
, 17)
.
In contrast to ECs, fibroblasts, which also express integrins 1ß1 and 2ß1, do not respond to collagen I with increased actin polymerization, changes in cell shape, or cellular alignment into cords (15)
; this may relate to the fact that fibroblasts normally reside within interstitial collagens. On the other hand, ECs encounter collagen I only during the sprouting and invasive stages of angiogenesis. Thus, collagen I is appropriately situated to serve as a stimulus for EC reorganization. The findings summarized above indicating that collagen I drives capillary morphogenesis clearly suggest the hypothesis that collagen I is an activator of EC signal transduction pathways that regulate EC morphogenesis. However, at the molecular level, the underlying mechanisms by which collagen I stimulates ECs to regulate cell shape and multicellular organization have largely been unexplored. Additionally, the possibility that basement membrane laminins on which quiescent ECs normally reside and interstitial collagens in which sprouting angiogenesis and cord formation occurs provoke distinctly different signaling in ECs has not been examined. Therefore, we designed experiments to address these questions with isolated human dermal microvascular ECs in culture.
Because it seemed probable that dynamic changes in cytoskeletal organization are pivotal for EC morphogenesis, we hypothesized that collagen I controls the activities of signaling molecules that regulate the actin cytoskeleton, and we tested the involvement of Src kinases and Rho GTPases. Src kinases are known to regulate the cytoskeleton and cellular processes relevant to vascular morphogenesis including cell adhesion and motility (18)
. Src kinases have been implicated in angiogenesis, especially in regulating vascular permeability (19)
; however, the function of Src kinases in regulating morphogenesis of vascular ECs has not been explored. Rho GTPases are also critical for regulating cytoskeletal organization, and Rho family members have been implicated in cell movement and morphogenetic processes involving changes in cell shape and polarity (20
, 21)
.
Experiments described here confirmed the hypothesis that collagen I activates key regulators of the actin cytoskeleton in microvascular ECs and identified activation of both Src and Rho as previously unrecognized, critical mechanisms by which collagen I ligation of ß1 integrins drives microvascular EC reorganization into precapillary cords. In sharp contrast to collagen I, basement laminin-1 does not activate Src or Rho, consistent with findings that laminin-1 does not provoke cord formation. Thus, these experiments identified key signaling molecules by which collagen I drives morphogenesis of microvascular ECs and defined major differences between signaling provoked by collagen I and basement membrane laminin-1.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Cells were used at the fourth to seventh passage.
Dominant negative Src (dn-Src) (K295R/Y527F) (22)
, dominant active Src (da-Src)(Y527F), and wild-type Src (wt-Src) in the pLNCX retroviral vector were kindly provided by Dr. Joan Brugge (Department of Cell Biology, Harvard Medical School, Boston, MA, USA). Retrovirus was prepared by transfecting PT67 retroviral packaging cells (Clontech, Palo Alto, CA, USA) and stable transfected clones expressing 1 x 105 c.f.u./mL were selected as a source of retrovirus. For retroviral transduction, microvascular ECs at passage 4 were grown to 30% confluence on a 10 cm dish and transduced separately with retroviruses encoding dn-Src, da-Src, wt-Src, and green fluorescent protein (GFP) (23)
. VEGF (20 ng/mL) was added to enhance proliferation and survival during the transduction procedure, which was repeated three times on consecutive days before subjecting cells to selection with 200 µg/mL G418. This procedure provided >90% transduction as monitored by fluorescence associated with cells transduced with GFP control retrovirus.
Stimulation of microvascular ECs with collagen I and laminin-1: incubation with Src and Rho inhibitors and ß1 integrin-blocking antibody
Replicate wells or flasks of microvascular ECs were grown to confluence in standard medium (see above). Twenty-four hours before stimulation with matrix proteins, the culture medium was removed and replaced with basal medium (EBM-2, Clonetics, San Diego, CA, USA) containing 2% fetal bovine serum. Acid-solubilized rat tail collagen I (BD Biosciences, Bedford, MA, USA) was neutralized according to the manufacturer’s instructions and diluted in serum-free EBM-2 medium to a concentration of 250 µg/mL unless indicated otherwise. Mouse laminin-1 (BD Biosciences) was diluted in serum-free EBM-2 to the same concentration as collagen I. For experiments, culture medium was gently removed from the cells and replaced with the collagen I-containing or laminin-1-containing serum-free medium. Where indicated, the following inhibitors were added: PP2 (Calbiochem, San Diego, CA, USA), a potent and selective inhibitor of the Src kinase family of protein tyrosine kinases, and recombinant exoenzyme C3 transferase (Cytoskeleton, Denver, CO, USA), a potent and selective inhibitor of Rho. For experiments with PP2, cells were preincubated with drug 30 min before the experiment. For C3, which is not as cell permeable, cells were preincubated overnight. For experiments with ß1 integrin-blocking monoclonal antibody (clone P4C10; Gibco, Carlsbad, CA, USA), cells were preincubated with purified antibody (20 µg/mL) for 30 min and antibody was included throughout the experimental interval. A matched purified isotype monoclonal antibody (MOPC 21; Sigma, St. Louis, MO, USA) served as control.
Phase and time-lapse phase photomicroscopy
Standard phase contrast images were collected with a Leica DC200 digital camera and associated software. For time-lapse experiments, microvascular ECs were cultured to confluence in a 25 cm2 flask. Two hours before stimulation with collagen I or laminin-1, the cell medium was changed to EBM-2 medium containing 2% fetal bovine serum. Subsequently, serum-free medium containing collagen I, collagen I with inhibitors, or laminin-1 was added to the cells; the flask was purged with 5% CO2, then sealed tightly. The cells were visualized using a phase-contrast inverted microscope (model Diaphot 300; Nikon, Inc., Melville, NY, USA) equipped with a stage heated to 37°C and images were recorded at 10 min intervals using a CCD (charge-coupled device) camera (Dage-MTI, Michigan City, IN, USA), a frame grabber (Scion, Frederick, MD, USA), IP LAB software (Scanalytics; Billerica, MA, USA), and a Macintosh computer.
Immuno-fluorescence microscopy
For F-actin staining, cells seeded on glass coverslips were fixed for 30 min in 3.7% buffered formaldehyde, then washed with phosphate-buffered saline (PBS). The cells were stained for F-actin with a solution of 1% BSA in PBS containing 0.1% Triton X-100 and Texas Red-labeled phalloidin (1:40 dilution; Molecular Probes, Eugene, OR, USA) for 45 min. Coverslips were washed three times with PBS and mounted with Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA, USA). For VE-cadherin staining, cells on coverslips were fixed for 10 min in acetone:methanol (1:1) at -20°C, air dried, and incubated with goat antibody against VE-cadherin (SC-6458; Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 1:200 for 1 h. After three washes with PBS, coverslips were incubated with fluorescein-conjugated donkey anti-goat IgG (Santa Cruz) at 1:200 for 45 min and washed again three times with PBS. Finally, coverslips were mounted and sealed for observation. Fluorescence images were collected with a Leica DC200 digital camera and associated software.
Western blot
Cell lysates were prepared by adding ice-cold lysis buffer (20 mM Tris-HCL, pH 7.5, 150 mM NaCl, 1% NP-40, 1 mM CaCl2, 1 mM MgCl2, 10% glycerol, protease inhibitor cocktail (#P8340, Sigma), 1 mM EDTA, and 1 mM EGTA) to cells and shaking for 30 min at 4°C. Cells were scraped and centrifuged for 10 min (14,000 rpm) at 4°C. Protein concentrations of the supernatants were determined by DC Protein Assay kit (BioRad, Richmond, CA, USA). Equal amounts of protein were loaded for electrophoresis on 4–12% tris-glycine sodium dodecyl sulfate (SDS) polyacrylamide gradient gels (Invitrogen, Carlsbad, CA, USA), then transferred to PVDF membranes (Millipore Corp., Bedford, MA, USA). The membrane was incubated with 0.5 µg/mL rabbit anti-Src [PY418] phospho-specific antibody (Biosource International, Camarillo, CA, USA) in Tris-buffered saline containing 3% BSA and 0.1% (v/v) Tween-20 for 1 h. The membrane was washed three times in Tris-buffered saline containing 0.1% (v/v) Tween-20, followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody for 1 h. Immunoreactive bands were detected by enhanced chemiluminescence (Renaissance from NEN, Boston, MA, USA).
Assays for Rho and Rac activity
Active GTP-Rho was measured with an established method (24)
involving affinity precipitation of cell lysates with glutathione-Sepharose beads coated with a glutathione S-transferase fusion protein containing the GTP-Rho binding domain of Rhotekin. Bound active Rho and total cell lysates were separately subjected to SDS polyacrylamide electrophoresis (25)
, transferred to PVDF membrane, and stained with rabbit polyclonal antibody to RhoA (#SC-179, Santa Cruz Biotechnology), followed by staining with secondary antibody and visualization of the stained bands with chemiluminescence as described above. Rac activity was measured by affinity precipitation of GTP-Rac with PAK-agarose beads (Rac activation assay kit, Upstate Biotechnology, Lake Placid, NY, USA) according to the manufacturer’s instructions. Bound active Rac and total cell lysates were subjected to electrophoresis and blotting as described above. Rac was detected by staining with mouse anti-Rac antibody (Upstate Biotechnology).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
As illustrated in Fig. 1
A, adherent dermal microvascular ECs responded to addition of collagen I by retracting and realigning into polygonal arrays of cord-like structures within 5 h. This morphogenetic process closely imitates an early organizational step in the formation of vascular plexuses during embryonic development in vivo (3
, 6)
; as we showed earlier, the marked shape changes induced by collagen I are not due to induction of apoptosis or cell death (15)
. In sharp contrast to collagen I, equivalent concentrations of laminin-1 had no significant effects on cell shape (Fig. 1A
).
|
To identify differences in signaling associated with collagen I vs. laminin-1 stimulation, we began by investigating tyrosine phosphorylation in cells stimulated by these two matrices. Collagen I strongly induced tyrosine phosphorylation in microvascular ECs compared with laminin-1 (Fig. 1B
, upper panel). Within 1 h collagen I strongly induced tyrosine phosphorylation of a band corresponding to 
85 kDa; this band was absent or undetectable in cells stimulated with laminin-1. Immunoprecipitation analyses identified this prominent tyrosine phosphorylated band as cortactin (not shown). Since cortactin is a prominent substrate for tyrosine phosphorylation by Src (26)
, we next investigated the possibility that collagen I induces Src activation in microvascular ECs. As shown in Fig. 1B
(lower panel), collagen I but not laminin-1 induced phosphorylation of Src on tyrosine residue 418 by 1 h, with a substantial further increase by 3 h. Phosphorylation of this residue is indicative of Src kinase activation (27
, 28)
; thus, these experiments indicate that collagen I induces a sustained activation of Src in microvascular ECs and that this activity is not shared by laminin-1.
The morphological changes induced by collagen I in microvascular ECs involves rearrangement of actin cytoskeleton (15)
, and so we considered the possibility that collagen I regulates the activities of Rho family GTPases which are critically involved in cytoskeletal regulation (20
, 21)
. With Rho activity assays, we established that Rho was markedly activated by collagen I with detectable increases by 30 min and a sustained increase through 3 h (Fig. 1C
). In marked contrast, stimulation of microvascular ECs by laminin-1 suppressed Rho activity (Fig. 1C
). Moreover, as determined with Rac activity assays, collagen I provoked a rapid and transient increase in Rac activity; this was followed by a marked decline: by 3 h Rac activity was less than before collagen I stimulation. In contrast, laminin-1 provoked a gradual but sustained increase in Rac activity (Fig. 1D
). These experiments established that stimulation of microvascular ECs with collagen I provokes a sustained increase in Rho activity, but laminin-1 provokes a decline in Rho activity. Moreover, laminin-1 provokes a sustained activation of Rac whereas collagen I provokes a transient increase in Rac activity, followed by a decline to sub-baseline levels. Thus, Rho and Rac are regulated differently by collagen I and laminin-1 in microvascular ECs.
Inhibition of Src and Rho suppresses capillary morphogenesis induced by collagen I
Because collagen I stimulation of microvascular ECs provokes a sustained activation of Src and Rho whereas laminin-1 does not, we investigated the importance of Src and Rho activities for collagen I-induced morphogenesis and formation of precapillary cords. We used time-lapse phase microscopy to determine whether inhibitors of Src and Rho, individually and in combination, block collagen I-induced morphogenesis. As shown in Fig. 2
A, collagen I induced disruption of the monolayer within 2 h, and cell reorganization into cords continued during the ensuing 3 h. In contrast, laminin-1 did not provoke significant shape changes during this interval. Videos constructed with images collected at 10 min intervals indicated that collagen I-induced reorganization of microvascular ECs into cords involves not only cell retraction but also motility and cell realignment. As shown in Fig. 2A
, the Src family kinase inhibitor PP2 and the Rho inhibitor C3 transferase both suppressed the morphological changes provoked by collagen I. Most significantly, the combination of PP2 and C3 abolished collagen I-induced cord formation (Fig. 2A
). Furthermore, as shown in Fig. 2B
(high-power views of the panels shown in Fig. 2A
), both PP2 and C3 blocked cell retraction, which is critical to the process of cord formation. Thus, these experiments implicate Src and Rho activity as critical for the mechanism by which collagen I induces EC morphogenesis and formation of precapillary cords.
|
To further test the function of Src in capillary morphogenesis, we transduced microvascular ECs with the retroviruses encoding wild-type Src (wt-Src) and a dominant-negative Src mutant (dn-Src). Transduction with either wt-Src or dn-Src did not affect cell morphology significantly in the absence of collagen I. However, as shown in Fig. 3
, human dermal microvascular ECs transduced with dn-Src failed to form typical cord-like structures in response to collagen I stimulation, consistent with experiments involving the Src inhibitor PP2 (Fig. 2)
. Furthermore, dermal microvascular ECs transduced with wt-Src formed cords more rapidly than cells transduced with control retrovirus encoding GFP. Thus, these experiments using retroviral transduction independently establish the importance of Src for collagen I-induced reorganization of microvascular ECs into precapillary cords.
|
Inhibition of Src and Rho suppresses collagen I induction of actin stress fibers in microvascular ECs
Time-lapse images presented in Fig. 2B
indicated that collagen I-induced capillary morphogenesis involves cell retraction and that retraction is inhibited by blocking either Src or Rho. These findings suggested the possibility that Src and Rho mediate important cytoskeletal changes in response to collagen I. Therefore, we investigated the functions of Src and Rho in regulating actin stress fibers induced by collagen I that are required for the formation of precapillary cords (15)
. As shown in Fig. 4
A, collagen I induced prominent actin stress fibers traversing the body of the cell; Src inhibitor PP2 and the Rho inhibitor C3 both substantially reduced collagen I-induced stress fiber formation. As reported by us earlier, laminin-1 did not induce stress fibers in these experiments (15)
; as shown in Fig. 1
, laminin-1 did not activate Src or Rho. Thus, these experiments implicate both Src and Rho as important to the mechanism by which collagen I selectively induces actin stress fibers in microvascular ECs. Because the formation of stress fibers is critical for the formation of precapillary cords (15)
, these experiments define a specific contribution of Src and Rho to the mechanism by which collagen I initiates capillary morphogenesis.
|
Activation of Src by collagen I disrupts VE-cadherin from cell–cell contacts
The videos constructed from time-lapse microscopy (see Fig. 2
for sample images) also indicated that collagen I initiated the formation of precapillary cords by provoking the appearance of evenly distributed gaps between cells in the monolayer. The Src inhibitor PP2 suppressed the early formation of gaps; experiments with the Rho inhibitor C3 also suggested inhibition, although C3 was less effective than PP2. Thus, these findings suggested that collagen I activation of Src mediates disruption of cell–cell contacts, thereby facilitating motility, cell retraction, and the realignment of microvascular ECs into precapillary cords.
Because VE-cadherin localization to regions of cell–cell contact is indicative of the integrity of EC junctions (29)
, we investigated the consequences of collagen I stimulation for VE-cadherin. As shown in Fig. 4B
, typical VE-cadherin staining indicative of intact cell–cell junctions was observed continuously along regions of cell–cell contacts in control cells grown to confluence. Stimulation with laminin-1 was without effect, but VE-cadherin staining was markedly reduced in cells stimulated with collagen I. The Src inhibitor PP2 effectively preserved VE-cadherin staining after collagen I stimulation whereas the Rho inhibitor C3 was less effective. These findings suggest an important role for Src activation in the mechanism whereby collagen I stimulation provokes disruption of VE-cadherin localization and cell–cell contacts, and are consistent with our findings that collagen I, not laminin-1, activates Src.
Transduction of microvascular ECs with retroviruses encoding Src mutants establishes a critical function for Src in mediating both collagen I induction of actin stress fibers and collagen I disruption of VE-cadherin
It is well established that Rho activation induces the formation of actin stress fibers (20
, 21
, 30
, 31)
, so our findings that collagen I induces Rho activation, actin stress fibers, and retraction of microvascular ECs (Figs. 1
, 2
, 4)
are consistent with the known function of Rho. However, a role for Src activation in the formation of actin stress fibers was less expected. To further test our findings presented in Fig. 4A
with the Src inhibitor PP2, we used microvascular ECs transduced with dn-Src and wt-Src. In the absence of stimulation by collagen I, cells transduced with dn-Src or wt-Src contained actin stress fibers similar to cells transduced with control vector encoding GFP (not shown). However, in the presence of collagen I stimulation, expression of dn-Src markedly reduced formation of actin stress fibers compared with control (Fig. 5
A). Transduction of cells with wt-Src resulted in increased formation of actin stress fibers in response to collagen I stimulation. Thus, our findings with cells transduced by dn-Src or wt-Src (Fig. 5A
) and experiments involving the Src inhibitor PP2 (Fig. 4A
) are consistent: they establish that activation of Src is critical to the mechanism by which collagen I induces actin stress fibers in microvascular ECs.
|
To test our findings with the Src inhibitor PP2 implicating Src activation as important to the mechanism by which collagen I disrupts VE-cadherin, we transduced microvascular ECs with retroviruses encoding dn-Src and wt-Src. In the absence of collagen I, confluent cells transduced with dn-Src, wt-Src, and control retrovirus encoding GFP revealed normal patterns of VE-cadherin distribution identical to untransduced cells, as shown in Fig. 5C
. However, marked differences among these cell populations were observed after stimulation with collagen I. As shown in Fig. 5B
, transduction with dn-Src preserved VE-cadherin staining at cell borders of collagen I-stimulated cells, similar to effects of the Src inhibitor PP2 (Fig. 4B
). In contrast, collagen I stimulation of cells transduced with wt-Src resulted in more extensive disruption of VE-cadherin localization compared with control. We also observed that in the absence of collagen I, cells transduced with a dominant active Src mutant (da-Src) exhibited marked reduction in VE-cadherin staining from cell borders (Fig. 5C
). Thus, these experiments all indicate that Src activity regulates the localization of VE-cadherin in microvascular ECs and they establish that collagen I activation of Src is important for the mechanism by which collagen I disrupts VE-cadherin localization and cell–cell contacts.
Collagen I activates Src and Rho in microvascular ECs through ligation of ß1 integrins
We had shown that collagen I stimulation of microvascular ECs to form precapillary cords involves binding of ß1 integrins by collagen I (15)
. To determine whether binding of ß1 integrins by collagen I is also critical to the mechanism by which collagen I activates Src and Rho in these cells, we used monoclonal antibody P4C10, which blocks collagen I binding to ß1 integrins. This antibody inhibited Src activation and Rho activation by collagen I, whereas isotype control antibody was without effect (Fig. 6
). We conclude that collagen I activation of Src and Rho in microvascular ECs requires ligation of ß1 integrins, consistent with earlier findings that collagen I stimulates EC morphogenesis and formation of precapillary cords through this integrin family.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 32
, 33)
. Collagen I-induced precapillary cords in vitro closely resemble embryonic and adult precapillary cords, which are the precursors to mature blood vessels with lumens in vivo (3
, 4
, 6
, 7
, 34)
. As determined here with time-lapse video microscopy, collagen I-induced capillary morphogenesis is a complex dynamic process that begins with EC retraction, motility, and organized realignment. Basement membrane laminin-1 does not provoke microvascular EC assembly into precapillary cords, underscoring distinct differences between the activities of collagen I and laminin-1 in regulating EC morphogenesis. Because collagen I but not laminin-1 drives microvascular ECs to reorganize into precapillary cords, we hypothesized there were critical matrix-dependent signaling events that control the initiation of morphogenesis. Our experiments confirmed this hypothesis and have defined several key and previously unrecognized events that distinguish collagen I signaling from laminin-1 signaling in microvascular ECs. First, we found that collagen I but not laminin-1 provoked a marked and sustained increase in Src activity, as suggested first by our finding that collagen I induces tyrosine phosphorylation of the Src substrate cortactin and later confirmed by identification of collagen I-induced phosphorylation of Src on tyrosine 418. Moreover, we found that the Src kinase inhibitor PP2 and transduction with dn-Src each inhibited induction of EC morphogenesis by collagen I and that microvascular ECs transduced with wt-Src responded to collagen I by forming precapillary cords more rapidly than ECs transduced with control retrovirus.
Besides sustained activation of Src, we found that collagen I provoked a sustained activation of Rho. In contrast, laminin-1 induced a sustained suppression of Rho activity.
Moreover, we found that the Rho inhibitor C3 transferase suppressed collagen I-induced formation of precapillary cords, linking Rho activation to the mechanism by which collagen I drives EC morphogenesis. Combined inhibition of Src and Rho was most effective at blocking cord formation, indicating that Src and Rho function cooperatively during this process.
Collagen I-induced reorganization of microvascular ECs into precapillary cords requires ligation of ß1 integrins (15)
; as shown here, activation of Src and activation of Rho by collagen I are also dependent on ß1 integrin ligation. Collagen I binds principally the 1ß1 and 2ß1 integrins on microvascular ECs (16)
; however, laminin-1 principally binds integrins 6ß1 and 6ß4 (our unpublished data). Several reports have suggested possible mechanisms by which ligation of these integrins transmits different signals in microvascular ECs and why laminin-1 provokes a sustained activation of Rac, whereas collagen I induces a brief activation of Rac, followed by sustained suppression of Rac activity to sub-basal levels. Firstly, collagen I decreases cyclic AMP-dependent protein kinase A (PKA) activity in microvascular endothelial cells whereas laminin-1 increases PKA activity (15)
. These findings, in combination with findings that Rac activation in carcinoma cells is supported by PKA (35)
, raise the intriguing possibility that collagen I suppresses and laminin-1 activates Rac activity through PKA. Secondly, laminin-1 binding to integrin 6ß4 has been shown to activate phosphoinositide 3-OH kinase in carcinoma cells (36)
, which in turn promotes activation of Rac (37)
. Finally, our observations that collagen I and laminin-1 exert distinct and, for the most part, opposing actions in regulating Rho GTPases in microvascular ECs are consistent with findings made with epithelial cells, where laminin-1 and collagen I act in opposite ways in regulating subcellular localization of the Rac exchange factor Tiam I (38)
.
Regarding the mechanisms by which Rho activation and Src activation initiate capillary morphogenesis, time-lapse phase microscopy described here indicate that collagen I-induced morphogenesis and formation of precapillary cords involve cell retraction and realignment into parallel arrays. In addition, cord formation requires actin polymerization and formation of prominent actin stress fibers (15)
, which is consistent with the importance of cell retraction for cord formation because stress fibers mediate cell retraction (31)
. Therefore, we defined the involvement of Src and Rho in mediating formation of actin stress fibers and found that both contribute critically. Our findings that Rho is important for stress fiber formation in microvascular ECs are consistent with the well-established functions of this GTPase (20
, 21)
. Our finding that Src activity was also required for stress fiber formation was unexpected but not without precedence. For example, in NIH3T3 fibroblasts, Src was found to cooperate with Rho effectors ROCK and mDia to produce prominent actin stress fibers (39)
. We found that transduction of microvascular ECs with wt-Src resulted in enhanced stress fiber formation in response to collagen I stimulation, but no differences between control and wt-Src-transduced cells were observed in the absence of collagen I. These observations are consistent with our finding that the morphogenetic response of wt-Src-transduced cells to collagen I was more rapid than that observed with controls. Moreover, they indicate that collagen I stimulation and the associated activation of Src are required for induction of stress fibers.
Time-lapse microscopy also indicated that disruption of cell–cell contacts is a prominent early event in the process whereby collagen I provokes microvascular ECs to form precapillary cords. Disruption of cell–cell contacts is likely critical for both cell motility and cell realignment into parallel arrays. Therefore, we investigated the involvement of Src and Rho in collagen I-induced disruption of cell–cell contacts. Our data indicate that compared with Rho, Src is significantly involved in the mechanism by which collagen I induces the early formation of gaps between microvascular ECs. We found that the Src inhibitor PP2 and transduction with a dn-Src mutant both inhibited the collagen I-induced disruption of the homotypic cell adhesion molecule VE-cadherin from regions of cell–cell contacts. Moreover, transduction with wt-Src accelerated collagen I-induced loss of VE-cadherin from the cell borders and transduction with da-Src mutant prevented the formation of normal VE-cadherin junctions, even in the absence of collagen I. Src activity also has been implicated in the disruption of stable cell–cell adhesion among epidermal keratinocytes (40)
; in colon carcinoma cells, Src activity disrupts normal localization of cadherins to regions of cell–cell contacts (41)
. Thus, Src activation appears to be critical for disruption of cell–cell adhesion and cadherin localization in both epithelial and endothelial cells. In contrast, Rac activation has been shown to support rather than disrupt epithelial cell–cell junctions (42
, 43)
. Thus, our finding that laminin-1 induces a sustained activation of Rac in microvascular ECs and that it does not disrupt cell–cell contacts is consistent with an established function for Rac in supporting cell–cell junctions.
Src activity has also been implicated as being important for survival of ECs during VEGF-driven angiogenesis and for mediating VEGF induction of microvascular permeability (19)
. The involvement of Src in regulating VEGF-mediated increases in microvascular permeability may relate to our findings that Src activity mediates disruption of VE-cadherin. Nonetheless, the biology of microvascular permeability is highly complex and likely involves more than disruption of cell–cell junctions (44)
. Consequently, Src may regulate microvascular permeability through multiple mechanisms.
In our view, the most intriguing aspect of findings presented here is that they suggest a model whereby interstitial collagen and basement membrane laminin differentially regulate various stages of angiogenesis. As summarized in Fig. 7
A, interstitial collagen I activation of both Src and Rho promotes formation of prominent actin stress fibers which mediate EC retraction and multicellular reorganization. Collagen I activation of Src disrupts VE-cadherin from cell junctions and promotes disruption of cell–cell contacts, thus facilitating EC morphogenesis. In sharp contrast to collagen I, basement membrane laminin-1 does not provoke EC morphogenesis or induce activation of Src or Rho. Rather, laminin-1 induces persistent activation of the GTPase Rac whereas collagen I suppresses Rac activity after a transient increase. The potential importance of these distinctions in matrix protein signaling for angiogenesis are summarized in Fig. 7B
. During the sprouting and proliferative stages of angiogenesis, the laminin-rich basal lamina is degraded, resulting in reduction of EC-laminin interactions. Our data predict that loss of laminin substratum results in reduced Rac activity and therefore a loss of Rac function in supporting integrity of cell–cell junctions during this phase. Moreover, upon degradation of basement membrane, sprouting ECs are exposed to underlying interstitial collagens and begin to invade it, resulting in activation of Src and Rho and initiation of capillary morphogenesis. Subsequently, as the newly formed capillary sprouts mature into new vessels with mature lumens, the intact basement membrane is reestablished. The continuous basement membrane sequesters ECs from interstitial collagens and thereby reestablishes normal activation levels for Rac, Rho, and Src. Thus, as summarized in Fig. 7
, data presented here support a model in which the laminin-rich basement membrane serves not only to maintain the integrity of the mature endothelium but also to sequester and thereby insulate ECs from interstitial collagens. In contrast, degradation of basement membrane exposes ECs to interstitial collagens and activates Src and Rho signaling pathways that drive cytoskeletal reorganization and sprouting morphogenesis.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 16, 2003. Accepted for publication November 21, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1ß1 and 2ß1 integrins. Proc. Natl. Acad. Sci. USA 94,13612-13617This article has been cited by other articles:
|
|
 |
|
  P. A. Singleton, J. G.N. Garcia, and J. Moss Synergistic effects of methylnaltrexone with 5-fluorouracil and bevacizumab on inhibition of vascular endothelial growth factor-induced angiogenesis Mol. Cancer Ther., June 1, 2008; 7(6): 1669 - 1679. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Zhang, N. E. Ramirez, T. E. Yankeelov, Z. Li, L. E. Ford, Y. Qi, A. Pozzi, and M. M. Zutter {alpha}2{beta}1 integrin expression in the tumor microenvironment enhances tumor angiogenesis in a tumor cell-specific manner Blood, February 15, 2008; 111(4): 1980 - 1988. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. B. Carter, G. Niu, M. D. Ward, G. Small, J. E. Hahn, and B. J. Muffly Mechanisms of HER2-Induced Endothelial Cell Retraction Ann. Surg. Oncol., October 1, 2007; 14(10): 2971 - 2978. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Takeda, A. R. Kazarov, C. E. Butterfield, B. D. Hopkins, L. E. Benjamin, A. Kaipainen, and M. E. Hemler Deletion of tetraspanin Cd151 results in decreased pathologic angiogenesis in vivo and in vitro Blood, February 15, 2007; 109(4): 1524 - 1532. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. H. Licht, O. T. Pein, L. Florin, B. Hartenstein, H. Reuter, B. Arnold, P. Lichter, P. Angel, and M. Schorpp-Kistner JunB is required for endothelial cell morphogenesis by regulating core-binding factor {beta} J. Cell Biol., December 18, 2006; 175(6): 981 - 991. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Ding, A. Lambrechts, M. Parepally, and P. Roy Silencing profilin-1 inhibits endothelial cell proliferation, migration and cord morphogenesis J. Cell Sci., October 1, 2006; 119(19): 4127 - 4137. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Qiao, F. Huang, R. P. Naikawadi, K. S. Kim, T. Said, and H. Lum Lysophosphatidylcholine impairs endothelial barrier function through the G protein-coupled receptor GPR4 Am J Physiol Lung Cell Mol Physiol, July 1, 2006; 291(1): L91 - L101. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Wollina, W.-D. Schmidt, C. Kronert, C. Nelskamp, A. Scheibe, and D. Fassler Some Effects of a Topical Collagen-Based Matrix on the Microcirculation and Wound Healing in Patients With Chronic Venous Leg Ulcers: Preliminary Observations International Journal of Lower Extremity Wounds, December 1, 2005; 4(4): 214 - 224. [Abstract] [PDF]
|
|
|
|
 |
|
