1525
Journal of Cell Science 113, 1525-1534 (2000) Printed in Great Britain © The Company of Biologists Limited 2000 JCS0737
VEGF nuclear accumulation correlates with phenotypical changes in endothelial cells Wenlu Li and Gilbert-A. Keller* Department of Pharmacokinetics and Metabolism, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA *Author for correspondence (e-mail:
[email protected])
Accepted 25 February; published on WWW 6 April 2000
SUMMARY Vascular endothelial growth factor (VEGF) is a multifunctional cytokine that plays a prominent role in normal vascular biology and pathology. In an experimental wound model, the mechanical disruption of monolayers of cultured endothelial cells resulted in two phenotypically distinct cell subpopulations in which VEGF was internalized by alternative endocytotic pathways and delivered to different subcellular compartments. In the cells away from the wound, VEGF was internalized via the classical receptor-mediated endocytosis pathway and accumulated in the endosomal compartment, whereas in the cells situated at the edges of a wound, VEGF was rapidly taken up and translocated to the nucleus. VEGF internalization and subsequent nuclear accumulation only
occurred for a short period of time after the wounding and was specifically abolished by antibodies that bind to the KDR binding site of VEGF. In the cells with VEGF nuclear accumulation, the levels of wound healing related proteins, such as Factor VIII (FVIII), tissue factor (TF) and tissue plasminogen activator, rapidly and dramatically increased compared to the cells that internalized VEGF via the classical endocytotic pathway. The increase in FVIII and TF was abolished when the nuclear transport is blocked. These data suggest that nuclear VEGF accumulation may be involved in modulating the levels of the proteins of the coagulation and fibrinolysis pathways.
INTRODUCTION
detectable levels as granulation tissue forms. These observations suggest that the decrease in VEGF expression may be responsible for the defective wound healing in diabetic rats (Frank et al., 1995). VEGF has been shown to induce conversion and expression of several endothelial proteins involved in the coagulation cascade and fibrinolysis including tissue factor (TF), von Willebrand factor (vWf), tissue plasminogen activator (tPA) and plasminogen activator inhibitor-1 (PAI-1) (Brock et al., 1991; Pepper et al., 1991; Dvorak et al., 1995; Clauss et al., 1996; Mandriota and Pepper, 1997; Zucker et al., 1998). Using an in vitro wounding model, we demonstrated that VEGF was internalized and transported to the endosomes in the cells away from the wound whereas, in the cells facing the wound, VEGF accumulated in the nucleus. In these cells, the levels of different wound healing related proteins, such as integrin β3, TF, Factor VIII (FVIII) and tPA, increased rapidly and dramatically. These results suggest that VEGF nuclear accumulation may play a role in the stimulation of coagulation and fibrinolysis pathways and affect vascular endothelial cellular physiology in ways that are independent of its growthpromoting effects and permeability properties.
Among the various molecules known to affect endothelial cell physiology, VEGF is a major multifunction cytokine that regulates angiogenesis and vasculogenesis (Ferrara and DavisSmyth, 1997). In addition to being a potent and specific mitogen for endothelial cells (EC), VEGF also serves diverse functions such as mediating EC migration, modulating the vascular permeability and promoting the fenestration in the EC of capillaries (Clauss et al., 1990; Roberts and Palade, 1995). Five VEGF isoforms of 121, 145, 165, 189 and 206 amino acid residues have been identified (Ferrara and Keyt, 1997; Poltrak et al., 1997). The predominant isoform is the 165/165 homodimer, a 45 kDa protein that signals through two endothelial cell transmembrane receptor tyrosine kinases, VEGFR1 (Flt-1) and VEGFR2 (KDR/Flk-1) (de Vries et al., 1992; Terman et al., 1994; Walterberger et al., 1994), and Neuropilin-1 (Soker et al., 1998). Like bFGF, VEGF165 also binds with low affinity to heparan sulfate proteoglycan on the cell surface (Ferrara and Henzel, 1989). It has been demonstrated that, following ischemia or infarction, VEGF expression is upregulated in cardiac myocytes (Banai et al., 1994). In rats, the levels of mRNAs for VEGF and its receptors are substantially elevated as early as 1 hour after coronary artery ligation (Hashimoto et al., 1994). In contrast to normal mice, the rise of VEGF in congenitally diabetic db/db mice is not sustained and VEGF expression plummets to barely
Key words: VEGF, Internalization, Nucleus, Coagulation, Fibrinolysis
MATERIALS AND METHODS Reagents Human recombinant VEGF165 was expressed in E. coli (Walter et al., 1996). mAbs to VEGF, TF, FVIII, PAI-1 and the goat anti-tPA
1526 W. Li and G.-A. Keller antibody were generated and characterized at Genentech. Antibodies to integrin β1 and β3 were purchased from Transduction Laboratories; to von Willebrand factor (vWf) from DAKO; to SPARC from Zymed; and to c-fos from Oncogene Science. The anti-c-myc antibody was a gift from Dr M. Bishop (University of California in San Francisco). Recombinant bFGF was purchased from Clonetics. Cell cultures Bovine adrenal cortex endothelial cells (ACE) were cultured in low glucose DMEM medium (Gibco) containing 10% bovine calf serum (BCS) (Hyclone). All the experiments described below were carried out with cell cultures at passages 4-18. Control SKBR3 breast cancer cells were cultured in high glucose DMEM medium containing 10% fetal bovine serum (Hyclone). Fluorescent VEGF and bFGF Cy3 (indocarbocyanine) is a fluorescent probe with spectral properties similar to those of rhodamine. VEGF165 and bFGF were conjugated with Cy3 as recommended by the manufacturer (Research Organic, Inc.). Internalization experiments To study the internalization and intracellular transport of VEGF, confluent ACE cells were grown on glass coverslips and incubated with 1 µg/ml of Cy3-VEGF165 in medium with or without 10% BCS at 37°C, 5% CO2 for 3 minutes to overnight, washed, fixed in phosphate-buffered 3% formaldehyde and mounted in bufferedglycerol containing 0.1% p-phenylediamine as an anti-bleaching agent. Immediately before adding Cy3-VEGF, experimental wounds were created by scratching a confluent monolayer with a sharp instrument (Todaro et al., 1965). In control experiments with wounded and unwounded monolayers, cells were incubated with 5 ng/ml Cy3bFGF or Cy3 alone. Competing and blocking experiments Either 0.1 mg/ml non-labeled VEGF165, 0.2 mg/ml humanized monoclonal anti-VEGF antibody (mhu α-VEGF) or a monoclonal anti-VEGF antibody was added to samples prior to 1 µg/ml Cy3VEGF. The samples were treated, fixed and processed as described above. Immunofluorescence assays and quantification For indirect immunofluorescence, cells were fixed in phosphatebuffered 3% formaldehyde, permeabilized in 1% Triton X-100 and immunolabeled with primary antibodies (10 µg/ml) followed by FITC-conjugated secondary antibodies (Jackson ImmunoResearch). Confocal images were collected with a Molecular Dynamics Laser Scanning Confocal Microscope. Projection images were generated with ImageSpace software from serial section data. For quantification of the immunolabeling, images were recorded with a Hamamatsu color chilled 3CCD camera with a 10× objective lens. Several fields were recorded from each set of experimental conditions. ‘Experimental conditions’ refers to time post-challenge, combined with the presence/absence of VEGF treatment. Each field was measured and the data were recorded using Openlab software (Improvision). For each field, the response variable calculated for analysis was the increase in intensity over background (the average intensity of the area surrounding a wound), multiplied by area (µm2), normalized by the wound length for that field: Intensity/µm wound = (MI – MB) × area (µm2)/length of wound (µm), where MI is mean intensity of increased area along a wound and MB is mean intensity of background. Means and standard deviations were calculated based on data from at least four fields in each set of experimental conditions. Comparisons between the average response for VEGF-treated and non-VEGF-treated control fields were made using a two-sample t-test
(two-sided). A critical level of 0.05 was used in the determination of statistical significance. Western blot analysis ACE cells were cultured to confluence in 10 cm Petri dishes and treated with or without 1 µg/ml VEGF for 3 hours. Cells were then lysed with 0.5 ml 2× SDS-PAGE sample buffer. Protein concentration was determined with Bio-Rad DC Protein Assay with BSA as a standard. 40 µg total protein/sample were loaded on a 4%-20% Novex Gradient mini-gel. After transfer to nitrocellulose membranes, the samples were probed with 1C7, a monoclonal antibody to FVIII and alkaline phosphatase-conjugated secondary antibody. After incubation in BCIT/NBT (Zymed), the membranes were scanned with a Microtek ScanMaker 5 and quantified using NIH Image 1.60 software. Thymidine incorporation assay Cells were incubated in the presence of 2 µCi [3H]thymidine/ml (Amersham) for various amounts of time, fixed in phosphate-buffered 3% formaldehyde and processed for autoradiography (Wimber and Quastler, 1963).
RESULTS VEGF is selectively transported to the nucleus in cells facing an experimental wound To follow the internalization and intracellular movement of VEGF, confluent monolayers of ACE cells were incubated with 1 µg/ml of Cy3-VEGF165 for periods of time varying from 0 minutes to overnight. Live cells were observed by fluorescence microscopy or fixed and examined by confocal microscopy. Within minutes, Cy3-VEGF165 was detected in vesicular organelles corresponding to the endosomal compartment in all the cells of the monolayer. No fluorescence was detected in the nuclei (Fig. 1A). Two different fluorescence patterns were observed when an experimental wound was produced by scratching the confluent monolayer with a fine forceps. In the cells away from the wound, Cy3-VEGF165 was contained in the endosomes whereas in the cells facing the wound, the fluorescence was localized to the nucleus and no other organelle was labeled (arrows, Fig. 1B). The appearance of fluorescence in the nucleus could be detected as early as 3 minutes after the addition of Cy3-VEGF165 to the medium and remained visible for up to 6 hours in cells in serum-free medium. However, in the presence of serum, the nuclear fluorescence for Cy3-VEGF165 lasted for only 2 hours then faded and disappeared. When Cy3-VEGF165 was added to disrupted monolayers at time points ranging from 10 minutes to overnight, Cy3-VEGF165 targeted to the nucleus in the cells facing the wound only during the first 2 hours postwounding. Cy3-VEGF165 added 3 hours after the wounding was not transported to the nucleus but localized to the endosomes. When unwounded monolayers were treated with Cy3-VEGF165 first and then wounded, fluorescence remained in the endosomes and was not observed in the nuclei. These results show that the ability of ACE to internalize and target VEGF to the nucleus is transitory and limited. We also observed that the number of cells competent to transport VEGF to the nucleus gradually diminished with the number of passages. After circa 20 passages, ACE cells internalized and transported VEGF to the endosomes exclusively.
VEGF internalization and EC phenotypes 1527
Fig. 1. Cy3-VEGF165 is internalized through two distinct pathways depending on the localization of ACE cells in a wounded monolayer. (A,B) Confocal projection images; (C) a low magnification image of regular fluorescence microscopy. All images were collected from the cells incubated with Cy3-VEGF for 1 hour at 37°C. (A) In a confluent monolayer of ACE cells, internalized Cy3VEGF165 is sequestered in vesicular organelles scattered throughout the cytoplasm. No fluorescence is detected in the nucleus. (B,C) In a monolayer that has been wounded by scratching with a fine forceps, Cy3VEGF165 is internalized and transported to the nuclei in the cells bordering the wound (arrows) whereas it is transported to the endosomes in the cells away from the wound. * indicates the wounding area. Bars, 10 µm.
The KDR binding site is essential for VEGF internalization and transport to the nucleus We carried out a series of experiments to rule out that VEGF internalization and nuclear transport in the cells at the wound edge resulted from transitory changes in the plasma membrane permeability induced by the disruption of the monolayer. In SKBR-3 cells, a breast cancer cell line that does not express VEGF receptors, no measurable uptake and no nuclear translocation of Cy3-VEGF165 was detected either in cells facing or away from the wound (not shown). Cy3-bFGF, an 18 kDa growth factor that undergoes nuclear translocation in bovine aortic endothelial cells as well as neurons and astrocytes (Baldin et al., 1990; Walicke and Baird, 1991), only localized to the endosomes in all cells. No nuclear fluorescence was seen either in the wound edge or in the resting cells (not shown). To further examine the specificity of the VEGF internalization and subsequent nuclear transport, mAbs recognizing different Table 1. Results of Cy3-VEGF nuclear translocation blocking experiments Competing/ blocking molecule
Binding site on VEGF
Cy3-VEGF nuclear labeling
Blocking of VEGF nuclear translocation
− cold VEGF mhu-α-VEGF A461* 2E3* 5F8*
N/A N/A KDR site KDR site Flt site Heparin
+++ − − − ++ +
N/A Yes Yes Yes No Partially
N/A, not applicable. *A461, 2E3 and 5F8 are monoclonal antibodies to VEGF developed at Genentech.
epitopes of VEGF were used in blocking experiments. Incubation with anti-VEGF antibodies that specifically block the KDR/flk-1 binding site (Kim et al., 1992) abolished Cy3VEGF165 internalization and nuclear transport in the cells facing the wound. In contrast, a mAb directed to the Flt-1 binding site did not prevent the internalization and nuclear transport (Table 1). As expected, a large excess of non-labeled VEGF165 completely abolished Cy3-VEGF165 accumulation in both nuclei and endosomes. These results demonstrate that the VEGF internalization that leads to nuclear transport is a wellregulated process in which the KDR binding site is essential. VEGF nuclear accumulation is not associated with the incorporation of [3H]thymidine or the induction of c-myc and c-fos Since cells at the periphery of a subconfluent monolayer and cells facing an experimental wound are often in a proliferative state (Todaro et al., 1965), and the addition of VEGF to ACE cells stimulates DNA synthesis (Keyt et al., 1996), we incubated a wounded monolayer with VEGF165 and [3H]thymidine for different time intervals and processed the cells for autoradiography and fluorescence microscopy. Quantitative analysis showed that, although 10-20% of cells away from the wound contained autoradiographic grains in their nuclei after a 5 hour or overnight incubation, there was no difference in the incorporation of [3H]thymidine between the cells in the region close or away from the wound (Fig. 2A). Furthermore, as some growth factors mediate their biological effects via the expression of the immediate early genes (Evan and Littlewood, 1993), we immunolabeled wounded monolayers with monospecific antibodies to c-myc and c-fos after treatment with Cy3-VEGF165. As shown in Fig. 2B,C, c-
1528 W. Li and G.-A. Keller
Fig. 2. [3H]thymidine incorporation and c-myc, c-fos immunofluorescence labeling in wounded ACE monolayers. (A) [3H]thymidine incorporation. The ACE monolayer was wounded and incubated with VEGF165 and [3H]thymidine at 37°C for 5 hours. The white line indicates the edge of the wound. The dark spots correspond to the cells that have incorporated [3H]thymidine in their nuclei. (B,C) Confocal projection images; ACE monolayers were wounded and incubated with 1 µg/ml Cy3-VEGF165 at 37°C for 1 hour. After fixation and permeabilization, cells were immunolabeled with anti-c-myc or anti-c-fos antibody and an FITC-conjugated secondary antibody. (B) c-myc; (C) c-fos. * indicates the wounding area; arrows, Cy3-VEGF165 nuclear labelling. Bars, 10 µm.
myc and c-fos distributed to the cytosol and nuclei of the cells in serum-free medium and there was no apparent difference in their levels between the cells with Cy3-VEGF nuclear labeling. Taken together, these results indicate that VEGF nuclear accumulation does not lead to the proliferation of ACE. VEGF nuclear accumulation correlates with the activation of wound healing proteins In situ, blood vessel disruption is followed by a series of acute and simultaneous reactions in which several molecular systems are involved, including the secretion of extracellular matrix proteins and the activation of a coagulation cascade and fibrinolysis system (Clark, 1988; Carmeliet and Collen, 1997). Since VEGF nuclear accumulation occurred exclusively in the cells facing the experimental wound, we hypothesized that this event might be associated with the healing of vascular injury, a multi-step process in which EC play a crucial role. To test this idea, we immunolabeled Cy3-VEGF treated ACE cells with antibodies to the major molecules Fig. 3. VEGF nuclear translocation correlates with the modulation of the levels of major wound healing proteins. VEGF internalization experiments were carried out in serumfree medium except for (E). Following incubation with Cy3VEGF165 for 1 hour at 37°C, cells were immunolabeled with specific primary antibodies and an FITC-conjugated secondary antibody. (A-E) Confocal microscopy projection images. (A) Integrin β3; (B) von Willebrand factor; (C)TF; (D) FVIII; (E) tPA, in the presence of serum; (F) PAI-1. Note that the immunolabeling shows distinct patterns. vWf and tPA are contained in vesicular structures, whereas integrin β3 and FVIII appear to be localized to the cytoplasm and the plasma membrane. In contrast, PAI-1 immunolabeling is restricted to the cells without VEGF nuclear accumulation away from the wound. * indicates the wounding area. Bars, 10 µm.
involved in wound repairing. As shown in Fig. 3A, the level of integrin β3, a subunit of αvβ3 (Brook et al., 1994), increased dramatically in the cells at the edge of the wound. Similarly, intense labeling for vWf, the receptor for FVIII (Fig. 3B), TF,
VEGF internalization and EC phenotypes 1529 Fig. 4. Tissue factor is specifically enhanced in cells with VEGF nuclear accumulation. ACE cells were incubated without (A,B) or with 1 µg/ml Cy3-VEGF165 (C,D) at 37°C for 30 minutes or 3 hours, then immunolabeled with anti-TF mAb 6B4.2E9 and FITC-antimouse IgG. All images are single sections of confocal microscopy. (A) Control, 30 minutes; (B) control, 3 hours; (C) Cy3VEGF165 treated, 30 minutes; (D) Cy3-VEGF165 treated, 3 hours. (E) Quantification of fluorescence intensity of TF immunolabeling in the 3 hour samples. 6-7 fields with more than 10,000 µm wounds from each sample were measured and analyzed as described in Materials and Methods. t-test: along wound, P0.05. * indicates the wounding area. Bars, 10 µm.
Molecular system
Protein detected
Extra-cellular Integrin β1 matrix Integrin β3
Intensity of fluorescence (wound versus center)
Correlation to VEGF nuclear translocation
No difference Increased along a wound
No Yes not by VEGF
Coagulation cascade
vWf Increased along a wound Tissue factor Increased along a wound Factor VIII Increased along a wound
Fibrinolysis
tPA
Early gene product
Yes Yes Yes
PAI-1
Yes need serum factors Decreased along a wound Yes
c-myc c-fos
No difference No difference
E
90
80 Intensity of Fluorescence
Table 2. Correlation of VEGF nuclear translocation to wound healing proteins and early gene products determined by immunofluorescence assays
70 Along wound Background
60
50
Increased along a wound
No No
the receptor for Factor VII /VIIa and initiator of coagulation system (Fig. 3C), FVIII, a cofactor in the coagulation cascade (Fig. 3D) and tPA, the major activator of fibrinolysis system (Fig. 3E), was also observed exclusively in the cells facing the wound. Importantly, the increase of immunolabeling for vWf, TF, FVIII, tPA and integrin β3 strictly correlated with the nuclear accumulation of VEGF (Fig. 3A-E). In sharp contrast, the labeling for PAI-1, the major regulator for tPA, decreased significantly in the cells with VEGF nuclear accumulation, but remained constant in the cells with VEGF endosomal accumulation (Fig. 3F). The levels of integrin β1 and SPARC, a secreted matricellular protein thought to be involved in late stages of vascular wound healing (Reed and Sage, 1996), remained unchanged (Table 2). Control experiments showed that, under normal culture conditions, the intensity of immunolabeling for integrin β3, TF, FVIII, vWf and tPA in ACE cells was low and similar to that in the cells away from the wound in VEGF treated monolayers.
40 Control
VEGF165 Sample
The TF and FVIII are enhanced specifically in cells with nuclear VEGF Further characterization indicated that the cellular responses depicted in Fig. 3 were not all triggered by the accumulation of VEGF in the nucleus. The changes in the integrin β3 levels were in response to the disruption of the endothelial monolayer, as non-VEGF treated control samples displayed similar enhancements. Also, VEGF alone was not sufficient to activate tPA as the immunolabeling for tPA showed little increase in cells incubated with VEGF in serum-free medium (Table 2). However, the dramatic changes of both TF and FVIII appeared to be in response to VEGF treatment. As shown in Fig. 4A,B, the TF level in control samples was increased in cells along the wound within the first 30 minutes but declined and was almost undetectable 3 hours after the wounding. The level of TF was higher than in the untreated cells in the presence of VEGF (Fig. 4C,D). The increase in TF cellular level was only observed in the cells with VEGF nuclear
1530 W. Li and G.-A. Keller
A
B
600
Intensity of fluorescence
500
400 Along wound
300
Background
200
D
250
100
200 Control
VEGF165 Sample
C
Integrated Density (%)
0
150
100
50
Fig. 5. VEGF165 enhances the level of FVIII in cells with VEGF 0 nuclear accumulation. (A) Quantification of FVIII immunolabeling in 1 2 3 4 ACE cells. ACE monolayers were wounded and incubated with or without Cy3-VEGF165 for 3 hours, prior to immunolabeling. About Sample 5,000 µm wounds from four fields of each sample were measured and analyzed. t-test: along wound, P=0.02; background, P>0.05. (B) Low magnification micrograph showing increased FVIII restricted along the wound in cells with VEGF nuclear accumulation. * indicates the wounding area. Bar, 10 µm. (C) Western blot of FVIII in unwounded and wounded ACE monolayers. A single band above 300 kDa was detected in all samples. (D) Quantification of the western blot analysis. The integrated density of the same size area from each sample was compared to the blank control. Sample 1: blank control; the monolayer was not wounded and not treated by VEGF165; sample 2: not wounded but treated by VEGF165; sample 3: wounded, not treated by VEGF165; sample 4: wounded and treated by VEGF165.
accumulation and did not occur in cells containing VEGF in their endosomes. Confocal microscopy showed that TF was located in the perinuclear region of the cells, corresponding to the Golgi apparatus area (Fig. 4C,D). Quantification of the intensity of TF immunolabeling at the 3 hour time point showed that the level of TF in cells along a wound is significantly higher in VEGF-treated samples compared to the non-VEGF-treated control (P=0.001; Fig. 4E). There was no statistically significant difference in the undisturbed area of the monolayers. Quantification of the immunofluorescence labeling showed that FVIII, another indispensable coagulation protein, was also significantly increased in the cells with VEGF nuclear accumulation in the 3 hour sample (P=0.02, Fig. 5A). We carried out a western blot analysis to further compare the protein levels of FVIII between wounded and unwounded monolayers (Fig. 5C,D). As shown in Fig. 5C, a single band above 300 kDa, corresponding to FVIII, was detected in all samples. The level of FVIII increased by 16% when VEGF was added to unwounded monolayers (Fig. 5D, samples 2 and 1) and increased by 35% in wounded but non-VEGF-treated
monolayers (compare samples 3 and 1). However, when a wounded monolayer was treated with VEGF, the FVIII level nearly doubled compared to the non-wounded control (samples 4 and 1) and was 43% higher than the level from the wounded but non-VEGF-treated sample (samples 4 and 3). These two different approaches indicated that FVIII is significantly enhanced when wounded ACE cells were treated with VEGF. In contrast, when mhu-α-VEGF was used to block the internalization and nuclear translocation of VEGF (Table 1), TF and FVIII levels in cells along a wound were dramatically decreased to the levels of blank controls while there was no significant difference in the undisturbed area of the monolayers (Fig. 6). DISCUSSION With a few exceptions (Brock et al., 1991), the biological responses of wounded EC to growth factors have been followed over time periods of several hours to days (Peters et al., 1993; Brook et al., 1994; Kaji et al., 1996; Mandriota and
VEGF internalization and EC phenotypes 1531 A
Intensity of immunofluorescence labeling
700
600 Along wound Background
500
400
300
200
100
0 Sample
1
2
3
mhu-α-VEGF VEGF165
-
+
+ +
B
Intensity of immunofluorescence labeling
response to the wounding, the cells facing the wound become competent to internalize VEGF by an alternative route that is different from the clathrin-coated endocytosis pathway and results in its rapid transport to the nucleus. Secondly, concomitant with VEGF nuclear accumulation, these cells undergo a phenotypical shift that is characterized by a dramatic modulation in the activation or expression of the major wound healing proteins. Integrin β3, TF, FVIII, vWf and tPA cellular levels increase while the integrin β1 level remains unchanged and the PAI-I level decreases. Thirdly, in the cells away from the wound or in cells under standard culture conditions, neither the VEGF nuclear accumulation nor the phenotypical shift takes place.
400 350 Along wound Background
300 250 200 150 100 50 0 Sample
mhu-α-VEGF VEGF165
1
2
3
-
+
+ +
Fig. 6. The levels of TF and FVIII in cells along a wound decreased when VEGF nuclear translocation was blocked by mhu-α-VEGF. ACE cells were treated for 3 hours with Cy3-VEGF in the presence or absence of mhu-α-VEGF and processed for immunolabeling as described in Materials and Methods. 7,000-10,000 µm of wounds from 7 different fields of each sample were measured and analyzed. P values represent the t-test results of sample 2 versus sample 3. (A) TF, P
