Jump to
Abstract
Objective— Motivated by the central roles that vascular endothelial growth factor (VEGF) and transforming growth factor (TGF)-β play in the assembly and maintenance of the vasculature, we examined the impact of systemic VEGF or TGF-β signal inhibition on endothelial activation as detected by leukocyte-endothelial interactions.
Methods and Results— VEGF or TGF-β inhibition, accomplished using adenovirus expression of soluble Flt1 (Ad-sFlt1) or soluble endoglin (Ad-sEng), resulted in a significant increase in the number of leukocytes rolling along the mesenteric venous endothelium and a significant decrease in rolling velocity in Ad-sEng mice. Neutralization of VEGF or TGF-β resulted in endothelial surface expression of P-selectin and impaired peripheral vasodilatation. Neither inhibition of VEGF nor TGF-β was associated with platelet or leukocyte activation, as detected by the activation markers platelet P-selectin and the active integrin αIIbβIII, or by leukocyte expression of L-selectin. Soluble vascular cell adhesion molecule (VCAM)-1 and E-selectin were increased in sEng-expressing mice, indicating higher levels of these adhesion receptors.
Conclusions— VEGF or TGF-β neutralization leads to impaired endothelium-mediated vasodilatation and elevated expression of surface adhesion molecules, resulting in increased leukocyte adhesion. These results indicate an essential role for both VEGF and TGF-β in maintaining the endothelium in a nonactivated state and have implications for therapeutic approaches that neutralize VEGF or TGF-β.
A key property of the endothelium, which lines the entire vasculature, is maintenance of the nonthrombogenic and noninflammatory vascular surface. During localized infection or injury, the endothelium becomes activated, resulting in expression of adhesion molecules that prevent bleeding and mediate the inflammation that is associated with rolling of immune cells as well as their attachment and extravasation at the site of injury.1 These events are regulated by interplay between cytokines and adhesion receptors on both the endothelial cells (ECs) and the circulating blood cells. Under nonpathological conditions the endothelium is equipped with several mechanisms to prevent thrombus formation and immune cell adhesion, including production of nitric oxide (NO),2 which acts as an antithrombotic factor, vasodilator, and modulator of leukocyte adhesion. Acute endothelial dysfunction is often characterized by impaired vascular autoregulation, whereas chronic endothelial dysfunction leads to inflammation and altered expression of cell adhesion molecules.3
Emerging clinical and experimental observations suggest that vascular endothelial growth factor (VEGF) is required for the maintenance of specialized EC stability and function.4,5 Selective inhibition of VEGF in the kidney leads to thrombotic glomerular injury.6 Preeclampsia, a disease mediated in part by high levels of the circulating soluble Flt1 (sFlt1), which binds both VEGF and placental growth factor (PlGF), is associated with increased thrombosis7,8 and leukocyte adhesion to endothelium.9 Interestingly, an increased incidence of thrombotic events has been reported in patients treated with the VEGF-neutralizing drugs bevacizumab (Avastin)6,10 and sunitinib.6,11
Transforming growth factor (TGF)-β is also implicated in the pathogenesis of preeclampsia and its associated systemic inflammation.12 Circulating levels of soluble endoglin (sEng), a TGF-β1 inhibitor, are increased in preeclampsia. sEng also binds TGF-β3, which is expressed primarily by mesenchymal cells, macrophages, and endothelial cells.13 Genetic knockout of TGF-β1 leads to a variety of abnormalities, the most prominent of which is multi-organ inflammation.14 Some of the clinical features of preeclampsia can be recapitulated by experimental VEGF and TGF-β1 inhibition in rodents, which results in the formation of microthrombi in the kidney,15 pancreas,16 and choroid plexus.4 Although observations of endothelial dysfunction associated with experimental and clinical VEGF neutralization support a requirement for VEGF in vascular maintenance, the mechanism(s) by which VEGF maintains EC homeostasis in vivo is unknown. Similarly, though observations suggest a role for TGF-β1 signaling during inflammation,14,17 the effect of systemic neutralization of TGF-β1 in healthy nonpregnant mice has not been investigated.
In light of the known roles of VEGF and TGF-β in vessel integrity,18,19 we sought to investigate the effect of their neutralization on the endothelial-blood interface in vivo. Mesenteric venules in mice are very accessible to visualization using intravital microscopy, and previous studies demonstrating both TGF-β1 and VEGF expression in mesenteric vessels indicate that this vascular bed is appropriate for these investigations. Expression of Ad-sFlt1 or Ad-sEng led to increased numbers of rolling leukocytes, as well as elevated expression of P-selectin at the EC surface. Peripheral blood flow studies revealed impaired endothelial-dependent vascular autoregulation, suggesting that VEGF or TGF-β1 neutralization leads to systemic vascular endothelial dysfunction.
Materials and Methods
An expanded Methods section is provided in the supplemental materials (available online at http://atvb.ahajournals.org).
Animals
Animal experiments were carried out at the Immune Disease Institute and at the Schepens Eye Research Institute in accordance with IACUC guidelines. For intravital microscopy, 3- to 4-week-old Balb/c mice were injected intravenously on day 0 with 1×1010 viral particles (V.P.) Ad-CMV-null control (Ad-null), 2.5×109 V.P. Ad-CMV-sFlt1 (Ad-sFlt1), or 2.5×109 VP Ad-CMV-sEng (Ad-sEng) (Q · Biogene). For vascular autoregulation studies, 6- to 8-week-old CD-1 mice were used (Jackson Laboratories, Bar Harbor, Maine).
Analysis of Platelet and Leukocyte Concentration and Activation
Leukocyte number was determined using a Z1 particle counter (Coulter). Platelets in whole blood were analyzed by fluorescent activated cell sorting (FACS) for the levels of P-selectin and αIIbβIII in its active conformation using anti–P-selectin-FITC (BD-Pharmingen) and JONA-PE (Emfret Analytics) antibodies, respectively. To determine leukocyte activation, L-selectin (BD-Pharmingen) levels on Mac-1 (anti–Mac-1; BD-Pharmingen)–positive leukocytes were detected by FACS after RBC lysis.
Analysis of Circulating Adhesion Molecules
Plasma levels of circulating soluble P-selectin (sP-selectin), soluble E-selectin (sE-selectin), and soluble vascular cell adhesion molecule (sVCAM)-1 were assayed by enzyme-linked immunosorbent assay (ELISA), according to the manufacturer’s instructions (R&D Systems), and circulating von Willebrand factor (vWF) was determined as described.20
Intravital Microscopy
Phase contrast and fluorescence intravital microscopy were used to measure the velocity of leukocyte and expression of surface P-selectin as described.20,21
Vascular Autoregulation
After 7 days of adenoviral expression, blood flow rates were recorded noninvasively in the tail vein using a CODA6 BP device (Kent Scientific), according to manufacturer’s instructions. This device measures tail vein blood flow by recording the volume (ml) that enters the tail vein between the systolic and diastolic measurements. Each cycle measurement corresponds to 20 seconds. After baseline values were determined, mice were injected retroorbitally with saline as a control or with 10 μg/kg acetylcholine (ACh; Calbiochem), an endothelial-dependent relaxant, and recordings were taken until values returned to baseline levels. Values were averaged in 5-cycle intervals pre- and postinjection and normalized to baseline flow values per animal.
FACs
Bovine aortic endothelial cells (BAECs) were grown to confluence in DMEM supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 1× insulin-transferrin-sodium selenium (ITS), maintained in a humidified atmosphere of 5% CO2 / 95% air, and routinely used between passages 4 to 8.
For analysis of surface P-selectin, BAECs were grown in DMEM containing 1% FBS for 24 hours before addition of VEGF (10 ng/mL) or TGF-β1 (1 ng/mL) for 48 hours. Inhibitors of signaling pathways that were used include the p38 inhibitor PD169316 (Sigma P9248), the JNK inhibitor SP600125 (Sigma S5567), and an ERK inhibitor (Sigma A6355). BAECs collected by trysinization and centrifugation at 100g for 4 minutes. BAECs from each well of a 12-well plate were resuspended in 350 μL of FBS buffer (10% FBS in PBS) containing 7 μL of FITC-labeled P-selectin antibody (BD Biosciences cat. No 553744). The cells were gently rotated in the dark for 20 minutes, washed and resuspended in FBS buffer, then analyzed by FACS using a FACSCAN flow cytometer.
Statistical Analysis
The values are presented as mean±SEM. Statistical significance was calculated using a student unpaired t test. Probability values <0.05 were considered statistically significant.
Results
Effect of VEGF and TGF-β Neutralization on Leukocyte Number and Rolling
To determine the activation state of the endothelium during early stages of VEGF or TGF-β neutralization, intravital microscopy was used to observe the behavior of leukocytes in mesenteric venules. After 7 days of Ad-sFlt1, Ad-sEng or Ad-null (control) expression, there was a statistically significant increase in the number of rolling leukocytes on the endothelium of Ad-sFlt1– (120±19 leukocytes/min, n=10) and Ad-sEng–expressing mice (58±13 leukocytes/min, n=10) when compared to Ad-null mice (22±8 leukocytes/min, n=7; Figure 1A).
Figure 1. A, TGF-β or VEGF neutralization increased the number of rolling leukocytes on mesenteric endothelium. Scale=50 μm. *P<0.05, ***P<0.01. B, Doppler measurement of blood velocity (μm/s) and shear rate (s−1) showed similar values in each group. C, Quantification of leukocyte rolling velocity revealed decreased velocity in sEng-expressing mice.
Because blood velocity and shear stress may influence leukocyte rolling on the endothelium, we measured mesenteric shear rates on the endothelial surface using a Doppler flowmeter; veins with average shear rate of 150 s−1 were analyzed. There was no significant difference in shear rates among Ad-sFlt1, Ad-sEng, and control mice (Ad-null; Figure 1B). The velocity of rolling leukocytes is critical to inflammation because rolling slows before stable adhesion to the endothelium and extravasation through the vessel wall.22 Quantification of leukocyte rolling velocity revealed no change in the Ad-sFlt1 mice, however median rolling velocity decreased in Ad-sEng–expressing mice (8.14 μm/s) when compared to control Ad-null–injected mice (11.72 μm/s).
Effects of VEGF and TGF-β Neutralization on Activation of Leukocytes and Platelets
Because increased leukocyte rolling number may also result from an elevated number of circulating leukocytes, the concentration of leukocytes in the blood was examined. VEGF neutralization led to an increase in the number of peripheral leukocytes (Ad-sFlt1: 15.9±1.5×106 /mL, n=24) compared to control (Ad-null: 11±0.8/×106 /mL, n=23); leukocyte concentration was not significantly altered in sEng-expressing mice (Ad-sEng: 14.5±1.9×106 /mL, n=20; Figure 2A). Although there was a 40% increase in leukocyte concentration in Ad-sFlt1 mice, the number of rolling leukocytes was increased by 500% so that the increased rolling cannot be solely attributable to increased concentration.
Figure 2. A, Neutralization of VEGF, but not TGF-β, increased circulating leukocyte numbers. **P<0.01. B, Activation status of leukocytes, as detected by L-selectin surface expression, revealed no change in sFlt1- or sEng-expressing mice. C and D, Platelet counts decreased in sEng-expressing mice, whereas FACS analysis revealed no change in surface platelet P-selectin or αIIbβIII in its active conformation. ***P<0.0001.
Examination of the surface expression of L-selectin on Mac-1–positive leukocytes, as a reflection of their activation state, revealed no change in the levels of L-selectin, indicating that the leukocytes were not activated (Figure 2B). In light of the ability of activated platelets to induce Weibel-Palade body secretion and leukocyte rolling via endothelial P-selectin expression,20 we examined the concentration and activation state of platelets. Overexpression of Ad-sFlt1 did not alter platelet concentration (Ad-sFlt: 1.26±0.06×109/mL; n=24, versus Ad-null: 1.21±0.07×109/mL; n=23), whereas Ad-sEng–expressing mice exhibited decreased platelet numbers when compared to Ad-null mice (Ad-sEng: 0.78±0.08×109/mL; n=20; Figure 2C). FACS analysis revealed no change in the surface levels of P-selectin (FITC) (Ad-sFlt: 7.4±0.2, n=10; Ad-sEng: 6.7±0.2, n=6; Ad-null: 7±0; n=5) or activated αIIbβ3 (JONA-PE; Ad-sFlt: 8.3±0.9, n=10; Ad-sEng: 8.2±1.2, n=6; Ad-null: 7.8±0.8; n=5) on circulating platelets in the sFlt1 and the sEng-overexpressing mice compared to control mice (P>0.05; Figure 2D), indicating that platelets were not activated.
Leukocyte rolling can also be mediated by endothelial surface expression of P-selectin. Injection of anti–P-selectin–conjugated beads (or IgG-conjugated beads as a control) to visualize P-selectin on the endothelial surface revealed a significant increase in P-selectin in both Ad-sFlt1 (367.2±63.0/μm2, n=12) and Ad-sEng-expressing (789.1±195.0/μm2, n=8) mice compared to Ad-null mice (85.9±21.5/μm2, n=12; Figure 3A and 3B). There was virtually no adherence of control-IgG conjugated beads to the venules in any treatment group (Figure 3A). The significant increase in levels of P-selectin on the endothelial cell surface in Ad-sEng mice was accompanied by an increase in sP-selectin (90.12±2.1 ng/mL) when compared to Ad-null–injected mice (65.8±3.9 ng/mL; Figure 3C) that was not seen in the sFlt1-expressing mice (65.1±2.4 ng/mL).
Figure 3. A, VEGF and TGF-β neutralization increased mesenteric venule P-selectin surface expression. Scale=100 μm. B, Quantification of A. C, Circulating sP-selectin levels increased in sEng-expressing mice compared to control (Ad-null). ***P<0.001.
Role of Mitogen Activated Protein Kinase Signaling in the Modulation of Endothelial Surface P-Selectin
We examined the role of mitogen activated protein (MAP) kinase signaling in increased endothelial surface P-selectin using flow cytometry of endothelial cells in vitro. Addition of VEGF (10 ng/mL) or TGF-β1 (1 ng/mL) to BAECs significantly decreased surface P-selectin (control: 59.18%; VEGF: 49.70%; TGF-β1: 52.90%; Figure 4). The effect of p38, JNK, and ERK inhibition on VEGF- or TGF-β–induced changes in surface P-selectin levels was determined. BAECs were grown in reduced serum±VEGF or TGF-β in the absence or presence of 1 μmol/L PD169316, 1 μmol/L SP600125 or 10 μmol/L 3-(2-aminoethyl)-5-((4-ethoxyphenyl) methylene)-2,4-thiazolidinedione hydrochloride, specific inhibitors of p38, JNK, and ERK, respectively. PD169316 significantly attenuated TGF-β1–induced changes in BAEC surface P-selectin (63.45%), whereas SP600125 led to a significant reduction of VEGF-induced changes in BAEC surface P-selectin (58.90%; Figure 4).
Figure 4. Role of MAP kinase signaling in endothelial surface P-selection. FACs analysis revealed that VEGF (10 ng/mL) or TGF-β1 (1 ng/mL) significantly decreased surface P-selectin on BAECs. Pharmacological inhibition of p38 prevented TGF-β–induced changes, whereas inhibition of JNK blocked VEGF modulation of BAEC surface P-selectin.
Effects of VEGF and TGF-β Neutralization on Other Markers of Endothelial Activation
Endothelial activation leads to surface expression of P-selectin and release of vWF into the plasma, and can be succeeded by expression of E-selectin and VCAM-1 as inflammation progresses.3,22 We therefore examined the plasma levels of sE-selectin, sVCAM-1, and vWF in Ad-sFlt1– and Ad-sEng–expressing mice. When compared to Ad-null mice (31.06±2.473 ng/mL), sE-selectin levels in Ad-sFlt1–expressing mice (28.11±3.323 ng/mL) were unchanged (Figure 4A). However, sE-selectin levels were significantly increased in Ad-sEng–expressing mice (54.37±3.069 ng/mL; Figure 5A). Similarly, sVCAM-1 was elevated in Ad-sEng–expressing mice (587.7±20.53 ng/mL), compared to control Ad-null mice (462.7±16.84 ng/mL; Figure 5B). In contrast, circulating sVCAM-1 was slightly decreased in Ad-sFlt1–expressing mice compared to control (396.6±11.63 ng/mL; Figure 5B). No changes were apparent in circulating vWF in Ad-sFlt1–expressing (0.437±0.028 ng/mL, n=3) or Ad-sEng-expressing mice (0.506±0.032 ng/mL, n=6) compared to Ad-null mice (0.403±0.046 ng/mL, n=7; Figure 5C), though a trend of increasing vWF in Ad-sEng–expressing mice was noted. TNF-α levels were not increased in either Ad-sEng– or Ad-sFlt1–expressing mice compared to Ad-null (Figure 5D). Controls for the TNF-α assay were culture media from nonstimulated and LPS-stimulated macrophages, which released more than 1500 pg/mL TNF-α.
Figure 5. Effect of VEGF and TGF-β neutralization on circulating adhesion molecules. E-selectin (A), VCAM-1 (B), vWF (C), and TNF-α (D). Neutralization of TGF-β increased circulating E-selectin and VCAM-1. Neutralization of VEGF decreased circulating VCAM-1. **P<0.001; ***P<0.0001.
Effect of VEGF and TGF-β Neutralization on Endothelial Autoregulation
To indirectly determine the effect of VEGF and TGF-β neutralization on endothelial NO formation, we measured the peripheral blood flow autoregulatory response.23 Flow rates in the tail-vein were measured in response to intravascular injection of ACh, an endothelium-dependent vasodilator that increases local endothelial release of NO. As expected, tail vein flow rates were transiently increased after ACh in Ad-null mice (1.619±0.107 mL/cycle, n=5; Figure 6) as compared to both baseline flow rates (preinjection) or to saline injection (data not shown). In contrast, there was no change in blood flow rates in response to ACh in Ad-sFlt1- (0.852±0.112 mL/cycle, n=4) or Ad-sEng-expressing mice (0.960±0.032 mL/cycle; n=4; Figure 6). This impaired autoregulation suggests that VEGF or TGF-β signaling is necessary to maintain normal vascular autoregulation.
Figure 6. Effect of TGF-β or VEGF neutralization on endothelial autoregulation. Blood flow rates in mice tail vein were measured in response to ACh. In Ad-null mice, ACh increased tail-vein blood flow rates 11 to 15 cycles postinjection, whereas blood flow rates were unchanged in Ad-sFlt1– or Ad-sEng–expressing mice. *P<0.01.
Discussion
The quiescent vascular endothelium maintains a nonthrombogenic and noninflammatory surface. At the same time, the endothelium plays an important regulatory role in the inflammatory response by expressing factors that modulate the surface adherence of leukocytes and platelets.24 There is increasing evidence that VEGF and TGF-β are required for endothelial stability in vivo.4,5,15 In this study, we demonstrate that VEGF and TGF-β are important modulators of endothelial activation. Systemic VEGF neutralization resulted in increased numbers of rolling leukocytes, as well as an elevated number of peripheral blood leukocytes. These results are consistent with the previous observation that administration of VEGF attenuates leukocyte-endothelial interactions in an experimental model of endothelial activation25 and with the systemic inflammation associated with preeclampsia where circulating sFlt1 is elevated.15 Similarly, patients with age-related macular degeneration treated intravitreally with the anti-VEGF drug Avastin display an enhanced local inflammatory response and increased leukocyte infiltration of choroidal neovessels.26
In contrast to these descriptions of the antiinflammatory actions of VEGF, several reports describe that VEGF treatment of endothelial cells in culture leads to induction of endothelial adhesion factors.27–29 Transgenic overexpression of VEGF in the epidermis of mice can induce a psoriasis-like phenotype in a mouse model of inflammation.30 Whereas circulating levels of VEGF are reportedly in the range of 10 to 70 pg/mL,31–33 tissue culture studies use 1 to 25 ng/mL VEGF. The transgenic overexpression of VEGF likely leads to relatively high VEGF levels and may explain why the effects are more similar to the proinflammatory actions seen in the tissue culture studies.
Neutralization of TGF-β led to an increase in the number of leukocytes rolling on the EC surface and a decreased leukocyte velocity. Consistent with these findings, TGF-β1 treatment of cultured ECs led to decreased leukocyte adhesion,34,35 and increased TGF-β1 in tumor microvessels is associated with decreased leukocyte-endothelial interactions.36,37 Similarly, TGF-β1 null mice display marked immune cell infiltration in the heart, lung, liver, stomach, the CNS, and pancreas,38 and mice lacking the TGF-β1 activators, thrombospondin-1 or the integrin β6 subunit, display a similar phenotype with multi-organ inflammation and infiltration.39,40
Because neither platelet nor leukocyte activation was detected with VEGF and TGF-β neutralization, activation appears to be at the level of the endothelium. Increased levels of TNF-α were not detected in mice overexpressing sEng or sFlt1, indicating the observed endothelial activation is not attributable to adenoviral-induced inflammation. Rather, these results reflect a role for basal levels of endothelial VEGF and TGF-β signaling in the maintenance of endothelial quiescence.
TGF-β neutralization also led to a decreased number of circulating platelets, suggesting that circulating TGF-β may contribute to the maintenance of platelet numbers. Previous accounts demonstrate both inhibitory and stimulatory effects of TGF-β1 on megakaryopoiesis, depending on growth factor concentration and presence of other growth factors.41,42 Interestingly, the adhesive interactions of megakaryocytes with the vessel wall also impacts megakaryopoiesis in the bone marrow, with P-selectin but not E-selectin deletion resulting in expansion of megakaryocyte progenitors and immature megakaryoblasts, and with enhanced release of TGF-β1,43 suggesting a role for TGF-β1 in platelet production. Finally, we have previously observed microthrombi in the microvasculature of the choroid plexus of mice expressing sFlt1, raising the possibility that the reduced platelet number could also be attributable to increased consumption or clearance.
Under normal conditions, leukocytes interact minimally with the endothelium. However, on endothelial activation, surface expression of adhesion molecules results in leukocyte-endothelial interaction, leading to leukocyte extravasation. Neutralization of VEGF or TGF-β led to increased endothelial surface P-selectin levels and increased leukocyte rolling. Changes in the plasma concentration of soluble adhesion molecules reflect altered cell surface turnover and proteolytic cleavage. P-selectin is found in both ECs and platelets and activation of either can contribute to an increase in soluble P-selectin in plasma.44 The elevated sP-selectin observed with TGF-β neutralization was likely derived from ECs as P-selectin levels on platelets were unchanged. On the other hand, despite increased EC surface P-selectin in sFlt-1–expressing mice, sP-selectin levels were unchanged. This finding can be explained by the reported role of VEGF in the shedding of E-selectin.27 Despite increased endothelial P-selectin in mice overexpressing sFlt1 and sEng, we did not note any thrombosis in the mesenteric venules, however we have previously reported microthrombosis in the choroid plexus of mice expressing sFlt1.4
TGF-β1 has been reported to suppress neutrophil recruitment in vitro via decreased endothelial E-selectin,35,45 and induction of E-selectin is blocked by pretreatment of ECs with TGF-β1.46 The observed increase in sE-selectin in sEng-expressing mice corroborates these observations and provides insight into the mechanism by which baseline circulating TGF-β contributes to maintaining the endothelium in a quiescent state. The presence of circulating TGF-β1 (<5 ng/mL in mice47 and ≈170 pg/mL in humans48) suggests a role for a basal level of TGF-β1 in maintaining endothelial homeostasis.
Numerous studies describe the role of the MAP kinase pathways (p38, JNK, ERK) in VEGF or TGF-β signaling in endothelial cells, however the relative contribution of MAPKs in maintaining an antiinflammatory endothelium has not been described. We demonstrate that basal levels of VEGF limit P-selectin on the endothelial surface via the JNK pathway, but not the p38 or ERK pathway, whereas TGF-β1 acts through the p38 pathway, but not ERK or JNK.
Endothelial production of NO plays a central role in maintaining the endothelium in a nonthrombogenic and nonactivated state.49 NO synthesis in the endothelium is mediated by the endothelial nitric oxide synthase (eNOS). VEGF and TGF-β receptors interact with eNOS in the caveolae of normal ECs to regulate eNOS enzymatic activity.50,51 Imbalance of NO leads to deficient arteriolar dilation so that the impaired EC autoregulation observed on VEGF and TGF-β neutralization is likely mediated via reduced eNOS expression. Consistent with our observations, experimental systemic NO inhibition has been shown to lead to endothelial activation and leukocyte adhesion, effects that were reversible with VEGF administration.52 Recent evidence suggests infusion of ACh can also lead to vasodilation by inducing formation of prostaglandins and P450 eicosanoids,53,54 though the vasodilatory effects of P450 eicosanoids are potentially through induction of NO release from ECs.55 Because both VEGF and TGF-β1 induce ECs to release NO, and because previous studies demonstrate that NO prevents leukocytes adhering to the EC surface,56 we speculated that the increased endothelial P-selection observed in the sFlt1- and sEng-expressing mice may be attributable to reduced endothelial NO production. However, addition of the NO inhibitor L-NAME did not block the ability of VEGF or TGF-β1 to limit EC surface P-selectin in vitro (data not shown), suggesting that NO production downstream of VEGF or TGF-β1 is not the mechanism limiting P-selectin at the endothelial membrane.
Preeclampsia, which is characterized by high levels of circulating sFlt-1 and sEng, is associated with an increase in cytokine levels, endothelial cell damage,57 platelet activation,58 as well as elevations in P-selectin,59,60 ICAM-1, VCAM-1, and E-selectin.61 In addition, plasma from patients with preeclampsia significantly increases VCAM-1 expression62 and leukocyte adhesion to endothelial cells in vitro.9,63 In our model of systemic adenoviral expression of sEng and sFlt, concentrations of sEng and sFlt are similar to those reported in the serum of preeclampsia patients,64,65 and thus our observations support the increasing evidence that inhibition of VEGF and TGF-β1 mediates much of the pathology in preeclampsia.
Endothelial dysfunction is the basis of many cardiovascular pathologies. Our data suggest that basal circulating levels of VEGF and TGF-β contribute to vascular stability by maintaining the endothelium in a quiescent state. Understanding the molecular mechanisms that underlie endothelial dysfunction after VEGF or TGF-β neutralization in vivo will provide insight into the role of these growth factors in vascular homeostasis and may provide strategies to circumvent side effects associated with therapies targeting VEGF or TGF-β.
Acknowledgments
We thank Christine Bagley for her editorial assistance with the preparation of the manuscript.
Sources of Funding
This work was supported by NIH grant EY05318 (to P.A.D.), POI HL066105 (to D.D.W.), and Knights Templar Eye Foundation award (to T.E.W.).
Disclosures
None.
Footnotes
-
T.E.W., V.S.D., and A.S.R.M. contributed equally to this study.
-
Received July 16, 2008; revision accepted May 11, 2009.
References
This Issue
Jump to
Article Tools
Share this Article
- Inhibition of VEGF or TGF-β Signaling Activates Endothelium and Increases Leukocyte Rolling
