Laminar Shear Inhibits Tubule Formation and Migration of Endothelial Cells by an Angiopoietin-2–Dependent Mechanism
Objective— Fluid shear stress plays a role in angiogenesis. Laminar shear stress (LS) promotes endothelial cell (EC) quiescence, whereas oscillatory shear stress (OS) promotes EC turnover and dysfunction, which could lead to pathological angiogenesis. We hypothesized that LS inhibits EC migration and tubule formation, 2 functions important in angiogenesis, by inhibiting the secretion of proangiogenic factors.
Methods and Results— Human umbilical vein ECs (HUVECs), human microvascular ECs (HMECs), or bovine aortic ECs (BAECs) were subjected to either LS (15 dyn/cm2) or OS (±5 dyn/cm2) for 24 hours and used in Matrigel tubule formation or scratch migration assays. Exposure of HUVECs, HMECs, but not BAECs, to LS inhibited tubule formation compared with OS. LS also inhibited migration of HUVECs and BAECs compared with OS. Angiopoietin-2 (Ang2), a known angiogenic protein, was found to be downregulated by LS both in cultured ECs and mouse aortas. Using Ang2 siRNA, Ang2 knockdown blocked OS-mediated migration and tubule formation and the LS-inhibited tubule formation was partially rescued by recombinant Ang2.
Conclusions— Our data suggests that Ang2 produced by OS in ECs plays a critical role in migration and tubule formation, and may play an important role in diseases with disturbed flow and angiogenesis.
Fluid shear stress, the dragging force created by flow through blood vessels, is sensed by the endothelium and plays an important role in normal physiological responses as well as disease pathologies. In particular, shear stress is thought to play a role in angiogenesis, or the formation of new blood vessels from preexisting blood vessels.1–3 At the cellular and molecular level, unidirectional laminar shear stress (LS) is thought to promote endothelial cell (EC) quiescence; laminar sheared ECs are antiproliferative, antiapoptotic, and antithrombotic.4 However, oscillatory shear stress (OS), a type of disturbed shear stress implicated in diseases such as atherosclerosis, is thought to promote EC dysfunction; oscillatory sheared ECs are proproliferative, promigratory, prothrombotic, and secrete growth factors that stimulate smooth muscle cell proliferation and migration.4 The secretion of growth factors from dysfunctional ECs exposed to OS could also play a role in blood vessel remodeling and angiogenesis.
There are several diseases associated with both angiogenesis and disturbed flow, such as atherosclerosis, aortic valve calcification, and arterial occlusion. In atherosclerosis, atherosclerotic plaques preferentially occur in areas of the arterial system exposed to disturbed flow and angiogenesis in the plaque is thought to promote the progression of atherosclerosis.5,6 Disturbed flow is found in the aortic valve sinus and angiogenesis in the valve leaflet is associated with calcification.7,8 In arterial occlusion, disturbed flow is found in the postocclusive site and angiogenesis is important for the subsequent ischemia.9,10 The detailed mechanisms resulting in pathological angiogenesis and neovessel formation remain uncertain. In these physiological and pathophysiological scenarios, fluid shear stress may provide a driving force for angiogenesis.
Here, we hypothesized that LS inhibits angiogenesis by downregulation of the secretion of proangiogenic factors or cytokines. To examine this hypothesis, we carried out gene array and protein array studies specific for angiogenic factors and cytokines, respectively. From these and subsequent functional studies, we demonstrate that LS inhibits angiogenesis through the downregulation of Angiopoietin-2 (Ang2), a secreted protein that is required for postnatal angiogenesis.11
Materials and Methods
For expanded methods and results, please see the supplemental materials, available online at http://atvb.ahajournals.org.
Cell Culture and Shear Studies
Human umbilical vein endothelial cells (HUVECs), bovine aortic ECs (BAECs), and human microvascular EC line (HMEC-1) were grown to confluence and exposed to unidirectional LS (5 or 15 dyn/cm2), OS (±5 or ±15 dyn/cm2 at 1 Hz frequency), or static control (ST) for 24 hours using a cone-and-plate device as described by us.12
Matrigel Tubule Formation Assay
After shear, cells were trypsinized and resuspended in reduced serum media. Resuspended cells were added to a growth factor reduced Matrigel (BD Bioscience) coated 96-well plate and incubated for 6 hours (HUVECs, HMECs) or 22 hours (BAECs) at 37°C. Tubule formation was quantified microscopically by measuring tubule length using NIH ImageJ.
Scratch Migration Assay
After shear, cell monolayers were scratched with a 200-μL pipette tip. The monolayer was washed once and the medium was replaced with M199%-2% FBS (HUVECs) or DMEM-0.5% FBS (BAECs). After 6 hours, the amount of cells migrated into the scratch area were quantified microscopically using NIH ImageJ.
Preparation of Cell Lysates and Immunoblotting
After shear, EC lysates were prepared and analyzed by Western blot as previously described.13
Real-Time Quantitative Polymerase Chain Reaction
Real-time polymerase chain reaction (PCR) for Ang2 and TSP-1 mRNA quantification was carried out as previously described.12
Aorta En Face Immunostaining
Male C57/BL6J mice (8 weeks of age, n=5) were euthanized by CO2 inhalation, pressure perfused with saline (0.9% NaCl), and pressure fixed with 10% formalin. The aortic arch and thoracic aorta were isolated and cleaned of adventitia. Finally, a 2-step immunohistochemical protocol was followed as previously described.14 Images were taken using a Zeiss LSM 510 META confocal microscope.
Ang2 Knockdown by siRNA and Ang2 Addition Experiments
Subconfluent (50%) HUVECs were transfected for 4 hours with 50 nmol/L of Ang2 siRNA (siAng2; MWG Biotech) or nonsilencing siRNA (NS) using Oligofectamine (Invitrogen) in Opti-MEM media (Gibco). For Ang2 addition experiments, recombinant human Ang215 was added to CM at concentrations of 50, 200, 500, and 800 ng/mL during Matrigel tubule formation.
Data are reported as average±SEM obtained from at least 3 independent studies. Statistical significance (P<0.05) was assessed by Student t test using a Microcal Origin statistical package.
LS Inhibits Tubule Formation and OS Does Not Compared With ST in ECs
To determine whether shear stress regulates the tubule forming capability of ECs, we performed a Matrigel tubule formation assay. Static culture conditions (ST), cells cultured under no shear stress, were used as a control for our shear system. However, the majority of arterial endothelial cells in vivo are continuously exposed to shear stress, and static effects are not physiologically relevant. LS is a more appropriate control, representing a healthy “normal” state, which we will compare to OS, our disease state.
HUVECs that were preconditioned under static or OS (±5 dyn/cm2) formed more, longer tubules than those preconditioned under LS (15 dyn/cm2; Figure 1A). The low level of OS (±5 dyn/cm2), which is typically found in atheroprone areas of human arteries,5 also promoted greater tubule formation compared with that of static cells (Figure 1A). HUVECs that were exposed to 2 different levels of LS (5 and 15 dyn/cm2) and OS (±5 and ±15 dyn/cm2) were compared with each other to investigate the magnitude-dependence and directional-dependence of the shear stress effect. LS exposure at both 5 and 15 dyn/cm2 significantly inhibited tubule formation of HUVECs compared with both low and high levels of OS (±5 and ±15 dyn/cm2) and ST (Figure 1A). This suggests that the unidirectional component of the shear stress plays a greater role than the magnitude of the shear stress in inhibiting tubule formation. Based on this finding, we used 2 typical arterial levels of shear conditions: atheroprotective 15 dyn/cm2 LS and proatherogenic ±5 dyn/cm2 OS for the rest of the studies.5
The inhibitory effect of LS on tubule formation was also observed in ECs obtained from microvascular origins (HMEC-1, Figure 1C). However, ECs obtained from aortas (BAECs), which are less likely to participate in angiogenesis in vivo, showed no significant differences in tubule formation between LS and OS (Figure 1D), suggesting that shear stress does not play a role in tubule formation by aortic ECs.
We next examined whether ECs would produce a secreted factor(s) which regulates shear-mediated tubule formation. We added conditioned media (CM) collected from sheared cells (LS, OS, and ST for 1 day) to static HUVECs in the Matrigel tubule formation assay. LS CM obtained from HUVECs significantly inhibited tubule formation of static HUVECs whereas OS and ST CM did not (Figure 1B). Similarly, LS CM from BAECs inhibited tubule formation of static HUVECs compared with OS and ST CM (Figure 1E). These results suggest that aortic ECs may not form capillary-like structures by themselves, but they still could produce factors promoting tubule formation in response to OS.
LS Inhibits Migration and OS Does Not Compared With ST
To investigate whether shear stress would regulate EC migration we performed the scratch migration assay. We found that HUVECs and BAECs preconditioned with LS had inhibited migration into the denuded zone compared with ST whereas OS promoted migration with similar results as ST (Figure 2A and 2B).
Next, we investigated whether a secreted factor(s) in response to shear stress regulates migration. For this study, we added shear CM to static HUVECs or BAECs in a scratch migration assay. LS CM significantly inhibited migration compared with those of both ST and OS (Figure 2C and 2D). This suggests that ECs produce a secreted factor(s) which regulates shear-mediated migration.
Shear-Sensitive Angiogenic Genes and Proteins
To identify the genes and the secreted proteins that are responsible for shear-mediated tubule formation and migration, we performed a gene array containing 128 human angiogenic genes and a protein array of 68 human cytokines (supplemental Tables I and II). Of interest, thrombospondin 1 (TSP-1) was upregulated 14.6-fold by OS over LS. From the protein array, Ang2 was upregulated 3.4-fold by OS compared with LS.
The gene and protein array results for Ang2 were confirmed using real-time PCR and Western blot, whereas TSP-1 was used as an additional control. The CM and cell lysates of HUVECs exposed to LS expressed 2- to 3-fold less Ang2 protein than that of OS (Figure 3A and supplemental Figure IA). The molecular mass of Ang2 is known to be 68 kDa (the upper band), whereas the lower band (≈61 kDa) seems to correspond to Ang2 (443) an alternative splice variant of Ang2.16 Real time PCR showed that LS exposure significantly downregulated Ang2 mRNA level in HUVECs by almost 7-fold below that of OS (Figure 3B). Whereas OS significantly increased Ang2 mRNA levels above that of static, Ang2 protein levels in OS and static cells were similar to each other (supplemental Figure IA and IB). In this study, the well-known effect of LS on increased endothelial nitric oxide synthase (eNOS) protein level and its phosphorylation at Ser-1177 (P-eNOS)17 was used as a control for shear experiments (supplemental Figure IA).
TSP-1 protein was downregulated in the CM and cell lysate of HUVECs exposed to LS (supplemental Figure IC). Static cells and OS-exposed cells expressed similar amounts of TSP-1 protein. Real-time PCR data showed that TSP-1 mRNA level was ≈2.5-fold higher in OS exposed HUVECs than that of LS (supplemental Figure ID), which was not as high as the 14.6-fold change found on the array, which may reflect differences between the array and PCR assays.
Ang2 Is Upregulated at Sites of Disturbed Flow In Vivo
The aortic arch has been shown to be exposed to both unidirectional laminar shear stress in the greater curvature and disturbed, oscillatory shear stress in the lesser curvature.14 Using en face immunostaining of the aortic arch in mice, we found that Ang2 was downregulated in the atheroprotected greater curvature compared with the atheroprone lesser curvature14 (Figure 3C and 3D), suggesting that Ang2 is downregulated by laminar shear stress and upregulated by disturbed oscillatory shear stress in vivo. Ang1 which is known to be unchanged by shear stress18 was used as a control and found to be unchanged by shear stress in vivo (Figure 3F and 3G). Both Ang1 and Ang2 were found only in the endothelium as depicted in orthogonal images of Z-stacks showing stained endothelial cells on top of the internal elastic lamina, with negative smooth muscle cells underneath (supplemental Figure II).
Knocking Down Ang2 Inhibits OS-Mediated Migration
We next examined whether Ang2 was responsible for shear-mediated migration of HUVECs using Ang2 siRNA. Ang2 protein level was significantly and specifically reduced in HUVECs by siAng2 but not by nonsilencing siRNA as determined by Western blot (Figure 4A through 4C). siAng2 did not knock down Ang1 (Figure 4A), showing its specificity for Ang2. Silencing Ang2 significantly inhibited OS-mediated migration (Figure 4D), suggesting that Ang2 plays an important role in migration. Ang2 knockdown had no significant effect on LS-mediated migration (Figure 4D).
Knocking Down Ang2 Inhibits OS-Mediated Tubule Formation, and LS-Inhibited Tubule Formation Can Be Partially Rescued by Recombinant Ang2
Next, we examined whether secreted Ang2 in response to OS was responsible for OS-induced tubule formation. Silencing Ang2 inhibited OS-mediated tubule formation but not LS-mediated tubule formation (Figure 5A), suggesting the specific role of Ang2 in OS-mediated tubule formation. We next tested whether the inhibitory effect of LS on tubule formation is attributable to a lack of Ang2 secreted into the CM. To test this hypothesis, recombinant Ang2 protein was added to LS CM in the Matrigel assay. As shown in Figure 5B, recombinant Ang2 partially rescued the inhibitory effect of LS on tubule formation in a dose-dependent manner. In contrast, recombinant Ang2 protein added to OS shear CM did not have an effect on tubule formation (Figure 5B), suggesting that Ang2 level in the OS CM already reached a maximum effective level. Ang2 protein is found physiologically at a concentration of ≤300 ng/mL19 and therefore physiologically relevant levels of Ang2 can partially rescue LS-inhibited tubule formation. The fact that Ang2 only partially rescues tubule formation suggests that other factors may also be playing a role.
In this study we found that: (1) LS inhibits tubule formation in HUVECs and HMECs, but not BAECs, however CM from laminar sheared BAECs can inhibit tubule formation of HUVECs, (2) Preconditioning with LS inhibits migration in HUVECs and BAECs, whereas OS does not inhibit migration and this is mediated through secreted protein, (3) Ang2 is downregulated by LS both at the gene and protein level in vitro, (4) Ang2 is downregulated at sites of LS and upregulated at sites of OS in vivo, (5) Knockdown of Ang2 inhibits OS-mediated tubule formation and migration, and (6) Addition of recombinant Ang2 partially rescues LS-inhibited tubule formation. Collectively, these findings suggest that Ang2 secreted by ECs in response to OS, acts as a promigratory and proangiogenic molecule that could play an important role in diseases with altered shear stress.
Fluid shear stress is thought to play a role in angiogenesis and arteriogenesis. In angiogenesis, endothelial proliferation and migration are 2 important responses. It is well established that LS causes a reduction in the rate of EC proliferation.20 Contrastingly, disturbed flow causes an increase in the rate of EC proliferation.20 The effect of shear stress on migration appears to be conflicting at first glance. Consistent with our finding, Tardy et al have also shown that ECs exposed to 48 hours of LS continually rearranged their relative position with no net migration, whereas ECs exposed to 48 hours of OS had a 2-fold increase in cell motility.20 However, Hsu et al reported that LS increases migration of ECs into a scratched area as compared with cells exposed to static and disturbed flow.21 These discrepancies are likely attributable to differences in experimental conditions: our cells were preconditioned with shear stress for 1 day, then scratched to see what the cell was programmed to do, whereas in Hsu et al’s experiments static cultured cells were scratched and then sheared. Our unpublished results indeed confirm Hsu et al’s result, suggesting that endothelial migration responses are different depending on when shear stress is applied.
To identify a molecular mechanism for the shear-mediated migration and tubule formation, we performed a gene and protein array from which we identified Ang2 and TSP-1 as being downregulated by LS compared with OS. Both Ang2 and TSP-1 have been previously identified as shear-sensitive in vitro,18,22 however it has not been studied how shear stress affects tubule formation and migration through expression of Ang2 and TSP-1. The role of TSP-1 in angiogenesis however is controversial, with most groups claiming it is antiangiogenic and some claiming that it is proangiogenic in certain settings.23,24 In our hands, a TSP-1 inhibitory peptide did not show any significant effect on tubule formation induced by OS (data not shown). The downregulation of TSP-1 by LS may be more important in regulating inflammation as TSP-1 has been shown to bind CD36 and promote monocyte binding.22
Angiopoietin-1 (Ang1) and Tie 2, the shared receptor of Ang1 and 2, were both on the gene array. However, Ang1 was not significantly changed by shear stress. Interestingly, Tie 2 was significantly upregulated by LS compared with OS (supplemental Table IA). This could be attributable to a compensatory mechanism where either lack of the ligand causes increased expression of the receptor or, conversely, too much of the ligand causes receptor desensitization which decreases the expression of the receptor. In fact, it has recently been shown that Tie 2 is internalized on binding with Ang2 or Ang1, and this internalization could act as feedback to downregulate mRNA expression of Tie 2.25
Our gene array found that VEGF was unchanged by shear (supplemental Table IB). VEGF has been shown to be important for the angiogenic potential of Ang2,11 however Bureau et al found that Ang2 can induce angiogenesis in a VEGF-independent manner.26 VEGF-C was significantly upregulated by LS compared with OS on our gene array. VEGF-C has been shown to be important in lymphangiogenesis.27
We identified Ang2 as being downregulated by LS compared with OS, both in vitro and in vivo. Ang2 is a 496-aa-long protein with 6 potential N-glycosylation sites and a secretion signal peptide.16 Ang2 knockout mice show that Ang2 is not required for embryonic angiogenesis but is requisite for postnatal angiogenesis.11 Recently, it has been published that after inhibition of the PI3K/Akt pathway Ang2 expression is rapidly induced in ECs by the transcription factor FOXO1.28 LS is known to stimulate the phosphorylation of Akt17,29 which inhibits Ang2 expression by phosphorylating and inhibiting FOXO1,28 suggesting a possible mechanism for the downregulation of Ang2 in response to LS.
Ang2 is thought to be involved in several disease pathologies including cancer,30 and Ang2 has been found in neovessels of advanced human atherosclerotic plaques.31 Inhibiting Ang2 may prove to be an effective treatment for these diseases. We were able to show that knocking down Ang2 using siRNA blocked OS-mediated migration and tubule formation, and this effect was partially recaptured by addition of Ang2 to LS shear CM. These results indicate that regulating Ang2 could be used as an effective strategy for patients with diseases, such as atherosclerosis or ischemia that involve pathological angiogenesis.
In summary, we showed that LS inhibited proangiogenic responses whereas OS did not, and this was through an Ang2-dependent manner. Elucidating the role of shear stress in regulating angiogenic signaling molecules such as Ang2 is important for understanding how these forces play a role in blood vessel remodeling and angiogenesis.
Sources of Funding
This work was supported by funding from NIH grants HL75209, HL70531, and UO1HL80711 (H.J.) and CA107783 (R.P.H.).
R.P.H. may be entitled to royalties derived from RayBiotech, INC.
Original received March 19, 2007; final version accepted July 19, 2007.
Ueda A, Koga M, Ikeda M, Kudo S, Tanishita K. Effect of shear stress on microvessel network formation of endothelial cells with in vitro three-dimensional model. Am J Physiol Heart Circ Physiol. 2004; 287: H994–H1002.
Sho E, Sho M, Nanjo H, Kawamura K, Masuda H, Dalman RL. Hemodynamic regulation of CD34+ cell localization and differentiation in experimental aneurysms. Arterioscler Thromb Vasc Biol. 2004; 24: 1916–1921.
Traub O, Berk BC. Laminar shear stress: mechanisms by which endothelial cells transduce an atheroprotective force. Arterioscler Thromb Vasc Biol. 1998; 18: 677–685.
Zarins CK, Giddens DP, Bharadvaj BK, Sottiurai VS, Mabon RF, Glagov S. Carotid bifurcation atherosclerosis. Quantitative correlation of plaque localization with flow velocity profiles and wall shear stress. Circ Res. 1983; 53: 502–514.
Heil M, Schaper W. Influence of mechanical, cellular, and molecular factors on collateral artery growth (arteriogenesis). Circ Res. 2004; 95: 449–458.
Gale NW, Thurston G, Hackett SF, Renard R, Wang Q, McClain J, Martin C, Witte C, Witte MH, Jackson D, Suri C, Campochiaro PA, Wiegand SJ, Yancopoulos GD. Angiopoietin-2 is required for postnatal angiogenesis and lymphatic patterning, and only the latter role is rescued by Angiopoietin-1. Dev Cell. 2002; 3: 411–423.
Sorescu GP, Sykes M, Weiss D, Platt MO, Saha A, Hwang J, Boyd N, Boo YC, Vega JD, Taylor WR, Jo H. Bone morphogenic protein 4 produced in endothelial cells by oscillatory shear stress stimulates an inflammatory response. J Biol Chem. 2003; 278: 31128–31135.
Platt MO, Ankeny RF, Jo H. Laminar shear stress inhibits cathepsin L activity in endothelial cells. Arterioscler Thromb Vasc Biol. 2006; 26: 1784–1790.
Suo J, Ferrara DE, Sorescu D, Guldberg RE, Taylor WR, Giddens DP. Hemodynamic shear stresses in mouse aortas: implications for atherogenesis. Arterioscler Thromb Vasc Biol. 2007; 27: 346–351.
Kim I, Kim JH, Ryu YS, Jung SH, Nah JJ, Koh GY. Characterization and expression of a novel alternatively spliced human angiopoietin-2. J Biol Chem. 2000; 275: 18550–18556.
Boo YC, Sorescu G, Boyd N, Shiojima I, Walsh K, Du J, Jo H. Shear stress stimulates phosphorylation of endothelial nitric-oxide synthase at Ser1179 by Akt-independent mechanisms: role of protein kinase A. J Biol Chem. 2002; 277: 3388–3396.
Harfouche R, Hussain SN. Signaling and regulation of endothelial cell survival by angiopoietin-2. Am J Physiol Heart Circ Physiol. 2006; 291: H1635–H1645.
Tardy Y, Resnick N, Nagel T, Gimbrone MA Jr, Dewey CF Jr. Shear stress gradients remodel endothelial monolayers in vitro via a cell proliferation-migration-loss cycle. Arterioscler Thromb Vasc Biol. 1997; 17: 3102–3106.
Good DJ, Polverini PJ, Rastinejad F, Le Beau MM, Lemons RS, Frazier WA, Bouck NP. A tumor suppressor-dependent inhibitor of angiogenesis is immunologically and functionally indistinguishable from a fragment of thrombospondin. Proc Natl Acad Sci U S A. 1990; 87: 6624–6628.
Ferrari do Outeiro-Bernstein MA, Nunes SS, Andrade AC, Alves TR, Legrand C, Morandi V. A recombinant NH(2)-terminal heparin-binding domain of the adhesive glycoprotein, thrombospondin-1, promotes endothelial tube formation and cell survival: a possible role for syndecan-4 proteoglycan. Matrix Biol. 2002; 21: 311–324.
Bogdanovic E, Nguyen VP, Dumont DJ. Activation of Tie2 by angiopoietin-1 and angiopoietin-2 results in their release and receptor internalization. J Cell Sci. 2006; 119: 3551–3560.
Bureau W, Van Slyke P, Jones J, Han RN, Ward NL, Stewart DJ, Dumont DJ. Chronic systemic delivery of angiopoietin-2 reveals a possible independent angiogenic effect. Am J Physiol Heart Circ Physiol. 2006; 291: H948–H956.
Jeltsch M, Kaipainen A, Joukov V, Meng X, Lakso M, Rauvala H, Swartz M, Fukumura D, Jain RK, Alitalo K. Hyperplasia of lymphatic vessels in VEGF-C transgenic mice. Science. 1997; 276: 1423–1425.
Daly C, Pasnikowski E, Burova E, Wong V, Aldrich TH, Griffiths J, Ioffe E, Daly TJ, Fandl JP, Papadopoulos N, McDonald DM, Thurston G, Yancopoulos GD, Rudge JS. Angiopoietin-2 functions as an autocrine protective factor in stressed endothelial cells. Proc Natl Acad Sci U S A. 2006; 103: 15491–15496.
Boyd NL, Park H, Yi H, Boo YC, Sorescu GP, Sykes M, Jo H. Chronic shear induces caveolae formation and alters ERK and Akt responses in endothelial cells. Am J Physiol Heart Circ Physiol. 2003; 285: H1113–H1122.
Etoh T, Inoue H, Tanaka S, Barnard GF, Kitano S, Mori M. Angiopoietin-2 is related to tumor angiogenesis in gastric carcinoma: possible in vivo regulation via induction of proteases. Cancer Res. 2001; 61: 2145–2153.
Calvi C, Dentelli P, Pagano M, Rosso A, Pegoraro M, Giunti S, Garbarino G, Camussi G, Pegoraro L, Brizzi MF. Angiopoietin 2 induces cell cycle arrest in endothelial cells: a possible mechanism involved in advanced plaque neovascularization. Arterioscler Thromb Vasc Biol. 2004; 24: 511–518.