We have previously shown that VEGF is a potent stimulus for in vitro growth and proliferation of HIMEC 2829. Additional studies have demonstrated that in vitro activation of HIMEC with TNF-α/LPS will result in inflammatory activation, expression of cell adhesion molecules and increased leukocyte binding which is mediated by activation of NF-κB and mitogen activated protein kinases 30. Because MAPK activation has been linked to proliferation of various cell types, including endothelial cells, we performed in vitro growth studies to assess the effect of VEGF and TNF-α/LPS on HIMEC. HIMEC monolayers stimulated with VEGF for 24 hr resulted in a significant increase in cell number, while TNF-α/LPS had essentially no effect and was similar to unstimulated cell growth (Figure 1). As expected, VEGF stimulation of HIMEC resulted in activation of all three members of MAPK family members (p44/42 MAPK, JNK, p38 MAPK; Figure 2), which was similar to that achieved by stimulation with TNF-α/LPS 30. This data demonstrates that MAPK activation is a common pathway involved in both the inflammatory and angiogenic activation of HIMEC, but also suggested that growth and proliferation would involve additional intracellular signaling mechanisms.

CsA exerts potent inhibitory effects on TNF-α/LPS induced activation of HIMEC, by inhibiting p38 MAPK activation 30. We began experiments investigating the effect of CsA on VEGF induced angiogenic activation in HIMEC by assessing endothelial stress fiber assembly and cytoskeletal architectural rearrangement, an early step in angiogenesis. HIMEC rapidly demonstrated stress fiber assembly following 1–10 ng/ml VEGF stimulation (Figure 3). However, higher dosages of VEGF (i.e. 50 ng/ml) resulted in a reduction in stress fiber assembly, suggesting that the effects of this growth factor follow a bimodal distribution, where higher concentrations result in diminished biologic effects. CsA decreased actin stress fiber formation following VEGF which was similar to the effect of the p38 MAPK inhibitor SB203580. The p44/42 MAPK inhibitor PD098059 failed to affect VEGF induced stress fiber assembly. VEGF (10 ng/ml) induced stress fiber assembly in HIMEC was abolished by receptor blockade using neutralizing monoclonal antibodies, suggesting that this phenomenon was mediated through VEGFR2.

Angiogenesis involves multiple events in endothelial cells, including migration and proliferation. HIMEC migration demonstrated a bimodal distribution in response to increasing dosages of VEGF, where 1–10 ng/ml resulted in a higher response compared to 50 ng/ml (Figure 4 left panel). CsA significantly inhibited VEGF induced migration of HIMEC, which was similar to the p38 MAPK inhibitor SB203580 (Figure 4 right panel). FK506, an additional calcineurin inhibitor and potent immunosuppressive agent, exhibited a similar effect compared to CsA, which exhibited a low level increase in baseline migration, which was similar to that observed in VEGF stimulated cells. Rapamycin, an immunosuppressive agent which functions through a distinct mechanism, interacting with m-TOR (mammalian target of rapamycin), demonstrated a stimulatory effect on resting cells and a diminished inhibition of VEGF induced HIMEC migration compared to CsA and FK506. Taken together with the previous experiments, these data suggest that CsA is a potent inhibitor of both stress fiber formation and cell migration, which is potentially mediated through effects on p38 MAPK activation.

Cell cycle re-entry and DNA replication in endothelial cells is a requisite step in angiogenesis. Experiments were performed on HIMEC proliferation assessing ³H thymidine uptake in response to low (10 ng/ml) and high (50 ng/ml) dose VEGF. Proliferation in response to VEGF (10 ng/ml) revealed a complete inhibition of increased thymidine uptake in HIMEC treated with CsA and the COX-2 inhibitor NS398 (Figure 5 right panel). In contrast to previous studies in stress fiber assembly and migration, the p44/42 MAPK inhibitor PD098059 was more effective than the p38 MAPK inhibitor SB203580 in blocking thymidine uptake in HIMEC stimulated with 10 ng/ml VEGF. At high dosages of VEGF (i.e. 50 ng/ml), CsA continued to block proliferation, which was partially inhibited by SB203580 and PD098059 (Figure 5 left panel).

Experiments evaluating endothelial growth in a wounded monolayer, with cell expansion across a leading edge, again demonstrated a potent angiogenic effect of VEGF compared to cells which were unstimulated (Figure 6). CsA treated HIMEC monolayers were completely unresponsive to VEGF, and grew at rates similar to untreated cells. In these experiments, the most potent inhibitor of monolayer expansion was the p44/42 MAPK inhibitor PD098059, while SB203580 failed to block this phenomenon.

In vitro-angiogenesis of HIMEC was assessed by tube formation assays. HIMEC seeded onto Matrigel™ in complete growth medium display formation of robust tube-like structures within 8 hr (data not shown), with further maturation after 16 hr (Fig 7 upper panel). CsA (0.1 μM), PD098059 (10 μM) and NS398 (10 μM) exhibited a marked inhibitory effect on the formation of tube-like structures by HIMEC, visible by the disruption of these capillary-like microvessels and cells remaining coherent in spherical clusters (Fig 7 lower panel). These findings indicate that activation of p44/42 MAPK and COX-2 is required for endogenous in vitro-tube formation in HIMEC.

The ability of VEGF to induce MAPK phosphorylation was assessed by immunoblot analysis using phospho-specific antibodies. VEGF stimulation of HIMEC leads to a marked phosphorylation and activation of extracellular signal regulated kinase (ERK 1/2, p44/42 MAPK), SAPK/JNK and p38 MAPK. ERK 1/2 phosphorylation by VEGF in HIMEC was time and dose dependent, reaching its maximum at 30 min of stimulation by 50 ng/ml of VEGF (Figure 8). PD098059 and CsA abolished p44/42 MAPK phoshorylation in HIMEC whereas SB203580 did not. Equal loading of protein was assured by western blotting after stripping and reprobing membranes with non-phospho-specific-MAPK antibodies along with Coomassie staining of the polyacrylamide gel after separation of proteins. Phosphorylation of p38 MAPK by VEGF in HIMEC was maximal at 5 min, and SB203580, a specific inhibitor of p38 MAPK resulted in inhibition of the p38 MAPK activity but not its phosphorylation. CsA partially inhibited p38 MAPK activation in HIMEC in response to VEGF (Figure 9). Control experiments were performed which demonstrated no effect of SB203580, PD098059 on the non-VEGF stimulated HIMEC (data not shown).

VEGF stimulation of HIMEC leads to dephosphorylation of NFATp and its nuclear translocation from the cytosol as detected by western blot analysis using specific anti-NFAT antibodies. Stimulation of HIMEC with 50 ng/ml VEGF resulted in a complete NFATp translocation to the nucleus within 15 min which was sustained for 2 hr. Nuclear NFATp declined to undetectable levels after 5 hr (Figure 10). Pre-treatment of HIMEC with 0.1 μM of CsA prevented NFATp translocation to the nucleus in VEGF stimulated cells, as shown by western blot analysis of nuclear extracts (Figure 11). Proinflammatory activation of HIMEC using a combination of TNF-α/LPS did not result in a nuclear translocation of NFATp, demonstrating the specificity of this signaling pathway in angiogenesis.

Having shown that VEGF activates NFATp in HIMEC, we next determined the binding of the transcription factor to an NFAT site of the human COX-2 promoter (nucleotides -117 to -91), NF-κB (nucleotides -277 to-211 containing the NFκB site of the human COX-2 promoter) and AP-1 (nucleotides -82 to -58 containing the NFAT-AP1 site of the human COX-2 promoter) 22. Figure 12 shows the electrophoretic mobility shift assay analysis of the nuclear extracts from VEGF activated HIMEC which demonstrate a marked increase in NFAT sequences of the COX-2 promoter bound to nuclear proteins of VEGF activated but not TNF-α/LPS stimulated HIMEC. The NFAT DNA-binding activity in HIMEC was inhibited by 0.1 μM of CsA pre-treatment (Figure 12 left panel). Next we studied the effect of VEGF on NF-κB activation in HIMEC. Figure 12 (middle panel) demonstrates that VEGF fails to activate NF-κB in HIMEC. This was in contrast to TNF-α/LPS activation of HIMEC which resulted in marked NF-κB activation. CsA had no effect on TNF-α/LPS induced binding of NF-κB (Figure 12 middle panel). VEGF stimulation of HIMEC also resulted in increased AP-1 activity, as shown in Figure 12 (right panel), AP-1 probe efficiently and specifically bound to nuclear proteins and this binding was not sensitive to CsA pre-treatment

To visualize the nuclear translocation of NFAT, the p65 subunit of NF-κB and the AP-1 components c-fos and c-Jun, we performed immunofluorescence staining of VEGF stimulated HIMEC using specific antibodies (i.e. anti-NFATp, p65, c-fos and c-Jun). As shown in Figure 13, in unstimulated HIMEC, NFATp was detected in the cytoplasm and following VEGF activation was translocated to the nucleus, whereas TNF-α/LPS activation did not result in nuclear translocation of NFAT. CsA pre-treatment resulted in abrogation of the NFATp nuclear translocation (Figure 13). Thus, the immunofluorescence results confirmed the data obtained from western blot analysis. Interestingly, VEGF did not lead to translocation of the NF-κB-subunit p65, whereas TNF-α/LPS activation of HIMEC resulted in p65 translocation from cytosol to the nucleus and was unaffected by CsA (Figure 14). VEGF stimulation of HIMEC also resulted in nuclear translocation of c-fos and c-Jun, which was also not affected by CsA (Figure 15).

Activation of HIMEC with VEGF also resulted in enhanced COX-2 expression at the mRNA level as detected by RT-PCR using COX-2 specific primers. Enhanced COX-2 expression was time dependent, increased from 3–6 hr which was maximal, declining from 12 to 24 hr. Pre-treatment of HIMEC with 0.1 μM of CsA, 10 μM NS398 and 10 μM PD098059 completely abolished VEGF induction of COX-2, whereas SB203580 did not exert an inhibitory effect on expression of this gene product (Figure 16). There was no detectable change in the level of COX-1 mRNA after VEGF stimulation. β-actin gene expression was used as an internal control in these experiments.

The effect of various cytokines on COX-2 protein expression was determined. As demonstrated in Figure 17 both TNF-α/LPS and VEGF were strong activators of COX-2 expression in HIMEC. Enhanced COX-2 protein expression in HIMEC following VEGF activation was time dependent, which increased at 3 hr and declined after 24 hr, as detected by western blotting using a specific anti-COX-2 antibody. In marked contrast, levels of COX-1 protein expression remained unchanged following VEGF stimulation of HIMEC (Figure 18). Consistent with the gene expression data, pre-treatment of HIMEC with 0.1 μM of CsA completely abolished VEGF induction of COX-2 protein expression (Figure 18).

VEGF and anti-human VEGF antibodies (VEGFR1 and VEGFR2) were from R&D Systems (Minneapolis, MN). Cyclosporine A (CsA) was from Alexis (San Diego, CA). Rapamycin, FK506 and the MAPK inhibitors (PD098059, SB203580) were obtained from Calbiochem (La Jolla, CA). The selective COX-2 inhibitor NS398 was from Cayman Chemical Co. (Ann Arbor, MI). Antibodies against the MAPK superfamily members (p44/42 MAPK, p38 MAPK, and c-jun NH₂-terminal MAPK (SAPK)) were from New England Biolabs (Beverly, MA). Antibodies to NFAT, p65 subunit of NF-κB and COX-1/-2 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Fluorescein-conjugated phalloidin was from Molecular Probes, Inc. (Eugene, OR). Fluorescein-conjugated streptavidin was from Pierce (Rockford, IL). Endothelial Cell Growth Supplement (ECGS) was from Upstate Biotech Inc. (Lake Placid, NY). MCDB-131 medium, porcine heparin, PSF (penicillin/streptomycin/fungizone), and soybean trypsin inhibitor were from Sigma (St. Louis, MO). Fetal bovine serum (FBS) was from Bio Whittaker (Walkersville, MD). Collagenase type II was from Worthington Biochemical Corporation (Lakewood, NJ), and bovine serum albumin (BSA, Fraction V) was obtained from Fisher Scientific (Fair Lawn, NJ). Human plasma fibronectin was from Chemicon International (Temecula, CA). Oligonucleotide and primers were from Operon (Alameda, CA).

Macroscopically normal small intestinal specimens for HIMEC isolation were obtained from patients undergoing scheduled bowel resection. The use of human tissues was approved by the Institutional Review Board of the Medical College of Wisconsin. HIMEC were isolated from surgical specimens and maintained as described earlier 28. HIMEC cultures were recognized by microscopical features and modified lipoprotein uptake (Dil-ac-LDL, Biomedical Technology, Inc., Stoughton, MA) and expression of Factor VIII-associated antigen. All experiments were carried out using primary endothelial cell cultures between passages 8–12.

Chemotaxis assay was carried out as described earlier 29. Briefly, using polycarbonate filters (8 μm pore size, Becton Dickinson Labware, Franklin Lakes, NJ) 5 × 10⁵ cells were added to the upper chamber, and chemotaxis buffer (1000 μl) containing VEGF (1–50 ng/ml) was filled into the lower compartment of the 12-well plates. After 3 hr of incubation at 37°C, cell culture inserts were removed, wiped and stained with DiffQuik (Baxter Scientific, McGraw, IL). Migrated HIMEC adherent to the lower side of the membrane were counted (ten random high-power-fields (HPF; 40×) per condition in a blinded fashion). In inhibition studies, resuspended cells were incubated with neutralizing anti-VEGFR2 antibody, isotype control antibody, CsA (0.1 μM), FK506 (50 nM), Rapamycin (20 nM), SB203580 (5 μM), PD098059 (10 μM) or the selective COX-2 inhibitor, NS398 (10 μM), for 30 min at 37°C. Cell viability was >95% as assessed by trypan blue exclusion. Each condition was assessed in triplicate.

3 × 10⁴ HIMEC per well were seeded onto fibronectin-coated 24-well plates using growth medium without ECGS as described earlier 29. After pre-treatment with the indicated inhibitors (CsA 0.1 μM, PD098059 10 μM, SB230580 5 μM, NS398 10 μM) for 30 min at 37°C, cells were stimulated with VEGF (50 ng/mL) for 24, 48, and 72 hr, or left untreated. Following complete detachment, cells were re-suspended and counted in a Coulter Counter (Coulter, Brea, CA). In parallel experiments, cell viability was assessed by trypan blue exclusion and was greater than 95%. Each condition was assessed in triplicate.

Cellular DNA synthesis was assessed by ³H-thymidine uptake in HIMEC as described earlier 29. HIMEC were pulsed with ³H-thymidine (1 μCi/ml; Amersham, Arlington Heights, IL), washed twice on ice with 5% (v/v) trichloroacetic acid prior to fixation. DNA was then released from precipitated material by alkaline lysis in 0.5 N NaOH, and supernatants were quantified in a beta-counter. Each condition was assessed in triplicate.

To assess HIMEC migration in response to angiogenic stimuli, a microscopic wounding assay was performed as described earlier 52. In brief, a HIMEC confluent monolayer was scraped along a straight line, and the remaining monolayer was then incubated with growth medium (without ECGS), and cells were pretreated for 30 min at 37°C with or without CsA (0.1 μM), SB203580 (5 μM) or PD098059 (10 μM). Then, cells were stimulated by addition of VEGF (10 ng/ml) or left untreated. The migration of HIMEC across the demarcation line was monitored using an inverted microscope. At each time point (24, 48, and 72 hr), 10 random fields using an ocular grid were counted in a blinded fashion. Data were expressed as cells/mm², and each condition was assessed in triplicate.

Endothelial tube formation was assessed using Matrigel™, a solubilized extracellular basement membrane matrix extracted from the Engelbreth-Holm-Swarm mouse sarcoma, as described previously 53. HIMEC resuspended in complete growth medium which were seeded at a density of 5 × 10⁴ cells per well. Where indicated, the growth medium was supplemented with CsA (0.1 μM), PD098059 (10 μM), SB203580 (5 μM) and NS398 (10 μM). Control wells contained no inhibitors. Endothelial tube formation on Matrigel™ after 16 hr was assessed by inverted phase contrast microscopy and photographed with an inverted tissue culture microscope. Five high power fields per condition were examined and experiments were repeated in two independent HIMEC cultures.

Confluent HIMEC monolayers in 35-mm culture dishes (one dish per condition) were pre-treated with various inhibitors for 30 min or left untreated before VEGF activation (50 ng/ml) for different time periods (1, 5,10, 15, 20, 30, 60, and 120 min) as described previously 54. The homogenates were centrifuged and supernatants (cytosolic fraction) were removed and protein concentrations were determined using a Bradford Assay (Bio-Rad, Hercules, CA). Equal amounts of protein were separated by SDS-PAGE and transferred to nitrocellulose membranes 30. The membranes were blocked for 3 hr at room temperature in 3% (w/v) BSA, 3% (w/v) nonfat dry milk in Tris-buffered saline (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) containing 0.1% (v/v) Tween 20(TBS-T), then were incubated with specific primary antibodies to NFATp, COX-2, phosphorylated and non-phosphorylated MAPK (ERK1/2, p38 MAPK and JNK) at 4°C overnight as specified. Detection was by secondary antibody coupled to horseradish peroxidase (HRP) and ECL™ (Amersham Pharmacia Biotech; Arlington Heights, IL).

Nuclear protein extraction was performed as described previously 30. In brief, HIMEC monolayers were lysed in hypotonic lysis buffer and cell nuclei were collected and frozen immediately in liquid nitrogen until further usage. Protein concentrations were determined using a Bradford assay (Bio-Rad, Hercules, CA). The samples were then incubated on ice with ³²P-labeled double stranded oligonucleotide for 30 min and DNA-protein complexes were separated by SDS-PAGE. Dried gels were exposed to X-ray film to detect DNA binding of NFAT, NF-κB and AP-1. The synthetic oligonucleotides 22 used as probes in electrophoretic mobility shift assays (EMSAs) were as follows: NFAT; 5'-tcgaCAAGGGGAGAGGAGGGAAAAATTTGTGGC-3' (nucleotides -117 to -91 containing the NFAT site of the human COX-2 promoter); NF-κB; 5'-gatcAGTGGGGACTACCCCCT-3' (nucleotides -277 to-211 containing the NF-κB site of the human COX-2 promoter) and for AP-1; 5'-tcgaCAAAAGGCGGAAAGAAACAGTCATTTC-3' (nucleotides -82 to -58 containing the NFAT-AP1 site of the human Cox-2 promoter).

F-Actin polymerization was assessed in subconfluent HIMEC seeded on fibronectin-coated glass chamber slides (LabTek; Nalge Nunc, Naperville, IL) as described previously 29. Cells were cultured in MCDB-131 without FBS for 12 hr and pretreated for 30 min at 37°C with or without CsA (0.1 μM), SB203580 (5 μM) or PD098059 (10 μM). HIMEC were then stimulated with 10–50 ng/ml recombinant human VEGF (1–15 min), and fixed with 3.7% (v/v) formaldehyde in PBS for 20 min at room temperature. Cells were washed and permeabilized with Triton X-100 (0.1% (v/v) in PBS) for 10 min, blocked with 2.5% (w/v) BSA/PBS, and stained with fluorescein-phalloidin (Molecular Probes, Eugene, OR). After washing, slides were air dried and mounted with Fluoromount-G (Southern Biotechnology, Birmingham, AL) and examined with a fluorescence microscope (Olympus BX-40) using a fixed shutter speed to allow for comparison of fluorescence intensity. In some experiments, stress fiber assembly was blocked by pre-incubation of cells with 10 μg/ml VEGFR2 antibody, (30 min at 37°C).

For immunofluoresence staining to demonstrate translocation of NFAT, p65 (NF-κB subunit), c-Jun and c-Fos (AP-1 subunit), cells were grown as above, left untreated or pre-treated with CsA (0.1 μM) before VEGF (50 ng/ml) or TNF-α/LPS (TNF-α 100 units/ml, LPS 1 μg/ml) for 30 min. Cells were then fixed with 3% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 15 min at room temperature and washed three times (5 min each) with washing buffer (PBS, 0.01% (v/v) Nonidet P-40 [NP-40]). After blocking for 30 min as above, the cover slips were incubated with the anti-NFATp, p65, c-Jun, phospho-c-Jun and c-Fos antibodies for 60 min at room temperature. Unbound antibody was removed by rinsing three times with washing buffer and the cover slips were incubated for 30 min with a fluorescein-conjugated secondary antibody (Santa Cruz), washed three times, mounted and visualized as above.

For determination of COX-1 and COX-2 mRNA expression, confluent HIMEC cultures were serum starved for 12 hr and stimulated with VEGF (50 ng/ml) with or without inhibitors for 3, 6 and 12 hr. Total RNA was extracted using TRIzol (Life Technologies, Rockville, MD) and treated with Deoxyribonuclease I, Amplification Grade (Life Technologies) according to the manufacturer's instruction. The complementary DNA (cDNA) was generated from 1 μg of total RNA with oligo (dT) primer using Superscript First-Strand Synthesis System for RT-PCR (Life Technologies) according to the protocol and in a total volume of 80 μl. Four μl of cDNA solution were used for polymerase chain reaction (PCR) in a total volume of 40 μl containing 1.5 mM MgCl₂, 0.2 mM dNTP mix, 0.5 μM of sense and anti-sense primers, respectively, and 1 IU of Taq polymerase (Life Technologies). PCR amplifications were performed 30 cycles for COX-1 (94°C for 1 min, 56°C for 1 min, 72°C for 1 min), 35 cycles for COX-2 (94°C for 1 min, 54°C for 1 min, 72°C for 1 min) and 25 cycles for β-actin (94°C for 1 min, 60°C for 1 min, 72°C for 1 min), using COX-1, COX-2 and β-actin specific primers (below) and followed by final extension for 7 min at 72°C. Twenty μl of PCR product were visualized on 1.2% agarose gels stained with ethidium bromide. RNA solution without reverse transcription was used as negative control (no RT), and β-actin served as an internal control. Primer sequences used are shown in table 1.

Statistical Analysis was performed using Statview 4.5 and superANOVA software for the Macintosh. When single comparisons were made, t-tests were used, applying paired or unpaired analysis as appropriate. When multiple comparisons between groups were performed one way or two-way analysis of variance was used as appropriate followed by the Student-Newman Keuls test. P ≤ 0.05 was considered significant.