Alterations in endothelial cell (EC) signaling could serve as a marker of effective antiangiogenic therapy. We determined the effect of an antiangiogenic tyrosine kinase inhibitor, SU6668, on tumor EC signaling in liver metastases in mice. In vitro immunofluorescence verified that pretreatment of ECs with SU6668 before exposure to VEGF decreased in vitro phosphorylation of Erk and Akt. Using double-fluorescence immunohistochemistry, phosphorylated Erk and Akt were constitutively expressed in ECs in liver metastases in untreated mice, but SU6668 blocked activation of these signaling intermediates. Determining the activation status of the Erk and Akt signaling pathways in tumor ECs may serve as a surrogate marker for the effectiveness of antiangiogenic regimens.
VEGF3 is the angiogenic factor most closely associated with aggressive disease in numerous solid malignancies (1, 2, 3, 4, 5). VEGF is known to induce intracellular signaling in ECs, activating the MAP kinase/Erk and PI-3 kinase/Akt pathways (6, 7, 8). These two pathways are protogenic survival pathways in ECs (8, 9, 10, 11, 12). It is now well documented that VEGF is not only an angiogenic factor, but is also a survival factor for normal and tumor endothelium (13, 14, 15, 16). An important antiangiogenic strategy is to neutralize the activity of VEGF. It may be possible to determine the effect of such therapies by examining EC signaling in biopsy specimens. Tumor angiogenesis is a dynamic process where ECs in different areas of a tumor may undergo different processes (i.e., proliferation and apoptosis) concurrently. Changes in EC intracellular signaling presumably occur before the alterations in the architecture in the tumor microvasculature (15, 17). In addition, alterations in EC signaling may indicate an abrogated mitogenic response to angiogenic factors. In the current study, we sought to determine the feasibility of examining EC signaling as a surrogate marker for antiangiogenic therapy designed to target VEGF. Our goal was to demonstrate that in the presence of antiangiogenic tyrosine kinase inhibitor, VEGF would be unable to activate the EC signaling pathways that mediate angiogenesis (EC proliferation and/or survival).
Materials and Methods
ECs and Culture Conditions.
Neonatal HDMECs (Cascade Biologicals, Inc., Portland, OR) were cultured and maintained in growth-supplemented medium 131 as recommended by the supplier. For immunohistochemical analyses, HDMECs were plated on two chamber slides (Becton Dickinson Labware, Franklin Lakes, NJ) at a concentration of 3 × 105 cells/chamber. Adherent cells were then serum-starved 12 h before treatment in MEM supplemented with 1% fetal bovine serum. For Western blotting, cells were starved overnight in MEM supplemented with 5% fetal bovine serum.
Reagents and Antibodies.
SU6668 and SU5416 were provided by SUGEN, Inc. (South San Francisco, CA). PD98059 (2-(2′-amino-3′-methoxyphenyl)-oxanaphthalen-4-one] and wortmannin were purchased from New England Biolabs (Beverly, MA). SU6668 was chosen for use in in vitro studies because it is more soluble than SU5416. hrVEGF was obtained from R&D Systems (Minneapolis, MN). Hoechst dye 3342 MW 615.9 (Hoechst, Warrington, PA) was used to counterstain nuclei, and Prolong (Molecular Probes, Eugene, OR) was used as mounting medium. Antibodies were purchased as follows: polyclonal rabbit antihuman phospho-p44/42 MAP kinase (Erk1/2), polyclonal rabbit anti-p44/42 MAP kinase (Erk1/2), polyclonal rabbit antihuman phospho-Akt (Thr308), and polyclonal rabbit anti-Akt from New England Biolabs and Alexa Fluor 488 goat antirabbit IgG (H+L) and Alexa Fluor 594 goat antimouse IgG (H+L) from Molecular Probes. Antibodies for CD31/PECAM-1 staining were purchased as described previously (15).
Western Blot Hybridization.
Cells were pretreated for 30 min with either medium alone (control), PD98059 (50 μm), wortmannin (200 nm), or SU6668 (1 μm) before hrVEGF (20 ng/ml) stimulation for 15 min. Blots were done as described previously (18) using a primary antibody dilution of 1:1000. Relative expression of the phosphorylated forms of the proteins was determined using NIH Image 1.61 software.
Immunofluorescence Double Staining of HDMECs Grown in Vitro.
HDMECs were treated with medium only (control condition), with 20 ng/ml of hrVEGF for 15 min, with 1 μm SU6668 for 30 min, or with 1 μm SU6668 for 30 min followed by 20 ng/ml of hrVEGF for 15 min. Immediately after treatment, the medium was discarded and cells were washed briefly with TBS. Cells were fixed in 2% paraformaldehyde for 10 min at room temperature. After washing for 10-min with TBS, fixed cells were blocked with blocking buffer (1× TBS, 0.1% Triton X-100, 0.5 m NaCl with 3% BSA) for 30 min and incubated with primary antibody (antihuman CD31/PECAM) diluted in antibody buffer (1 × TBS, 0.1% Triton X-100 with 3% BSA) overnight at 4°C. Primary antibody was removed by washing with TBS for 10 min, and the fixed cells were treated with blocking buffer for 10 min. The secondary antibody (Texas Red-conjugated) was diluted in antibody buffer (1:400) and added to the chamber slides for 1 h at room temperature. After a 1-h incubation, the slides were washed three times for 3 min each time with TBS, and fixed cells were treated with blocking buffer for 30 min. The second primary antibody (pErk, 1:100, or pAkt, 1:100 to 1:400) was added, and the mixture maintained overnight at 4°C. On the third day, the second primary antibody was removed by rinsing with TBS and the second secondary antibody (fluorescent green-conjugated) was added for 1 h at room temperature. Finally, the fixed cells were washed one last time for 30 min with TBS, and their nuclei were counterstained and mounted. This technique was repeated four times with similar results. For all immunohistochemical studies, the primary antibody was omitted as negative control.
Immunofluorescence Double Staining in Colon Cancer Liver Metastases.
We used harvested liver metastases from a previous study (15) to develop an in vivo immunohistochemical assay to evaluate Erk and Akt phosphorylation in the tumor endothelium. The assay uses a double-labeling technique in which ECs in liver metastases are labeled red (anti-CD31 antibody), and cells with activated signaling intermediates are labeled green (anti-pErk or anti-pAkt antibody). Colocalization of both colors leads to yellow, which represents ECs in which the Erk or Akt pathway has been activated. As described in a previous report, mice that had been implanted with human colon cancer cells were given daily 200-μl i.p. injections of either control vehicle [30% PEG-300 (w/v) in 0.1 m sodium phosphate buffer (pH 8.2), SU6668 (60 mg/kg in control vehicle), or SU5416 (12mg/kg in control vehicle)]. Specimens from seven mice/treatment group were analyzed. We examined viable ECs that did not have the morphology of apoptotic ECs. Tissue preparation and CD31/PECAM stains were done as described previously (15). The second primary antibody (anti-phospho-Erk, 1:100; or anti-phospho-Akt 1:400) was added and incubated overnight at 4°C. The second primary antibody was removed by rinsing with TBS, the second secondary antibody (fluorescent green-conjugated; 1:400) was added, and the sections were incubated for 1 h at room temperature. Finally, the sections were washed with TBS for 30 min, the nuclei were counterstained, and the sections were mounted for microscopic analysis. The double-staining procedure was repeated on four separate occasions with reproducible results.
In Vitro Activation of Erk and Akt in HDMECs as Determined by Western Blotting.
We pretreated cells with rhVEGF in the presence or absence of the MAP kinase inhibitor PD98059, the PI-3kinase-specific inhibitor wortmannin, the VEGFR2 tyrosine-kinase inhibitor SU6668, or medium alone (control). After pretreatment, cells were stimulated with 20 ng/ml of rhVEGF. VEGF did induce activation of both Erk and Akt (Fig. 1). Pretreating the HDMECs with PD98059 effectively inhibited the activation of Erk after VEGF stimulation. All signaling inhibitors blocked the total amounts of Akt, as compared with total Erk (the same membranes were reprobed for both signaling pathways). Wortmannin blunted the activation of Akt and, to some extent, activation of Erk. The most significant decrease in phosphorylation was seen after SU6668 treatment: SU6668 abrogated the activation of Erk (∼79% decrease) and Akt (∼62% decrease) in HDMECs that had been stimulated with VEGF.
Effect of VEGF and Angiogenic Inhibitors on HDMEC Signaling as Determined by Immunofluorescence Staining.
We developed an in vitro immunohistochemical assay for activated Erk and activated Akt in HDMECs. Our immunohistochemical results were consistent with our Western blotting results. The basal level of pErk noted in the control condition increased significantly after stimulation with VEGF (visualized as brighter green or as a burst of yellow in the double-stained EC; Fig. 2,A). In the absence of VEGF stimulation, SU6668 treatment led to a decrease in the basal level of pErk; in the presence of VEGF, SU6668 produced an almost complete abrogation of pErk (Fig. 2,A). We did observe some minor heterogeneity in cellular signaling in some cells that were more confluent and of higher passage. This is likely attributable to stress in cells grown to confluence, because ECs are highly sensitive to changes in in vitro conditions. Localization analyses revealed that VEGF led to pErk staining in the nuclei with some staining in the cytoplasm and pAkt staining in the nuclei (Fig. 2, A and B). Pretreatment of HDMECs with SU6668 led to complete abrogation of any basal or stimulated levels of pAkt in ECs in vitro (Fig. 2 B).
In Vivo Inhibition of Intracellular Tumor EC Signaling.
Previously we showed that antiangiogenic therapy targeting the tyrosine-kinase receptor (KDR) for VEGF (SU5416) or the receptors for VEGF, platelet-derived growth factor, and fibroblast growth factor (SU6668) inhibited the vascularity, proliferation, and growth of colon cancer liver metastases in mice and significantly increased apoptosis of ECs and tumor cells. In the control mice, tumor vessels and some tumor cells near those vessels showed increased staining for pErk. In contrast, mice treated with SU6668 (Fig. 3,A) or SU5416 (not shown) had decreased levels of pErk in both tumor cells and tumor endothelium (Fig. 3,A). Similar results were noted when tumors were double-stained for CD31 and pAkt (Fig. 3 B). Findings were relatively homogenous in tumor specimens, in contrast with the in vitro studies.
In this report, we assessed whether markers of intracellular endothelial signaling could be used to measure the effectiveness of antiangiogenic therapy targeting the VEGF system receptor. Results from our in vitro model verified that VEGF activated the MAP kinase/Erk and PI-3 kinase/Akt pathways in HDMECs. We also showed that SU5416 and SU6668 inhibited EC signaling in liver metastases. The tumor tissue for our in vivo experiment came from a previous study of ours in which mice were implanted with human colon cancer cells in the spleen, leading to liver metastases (15). In that study, two tyrosine-kinase inhibitors were examined; one targeting the receptor for VEGF (SU5416) and the other targeting the receptors for VEGF, basic fibroblast growth factor, or platelet-derived growth factor (SU6668). Both were found to inhibit tumor angiogenesis and growth and to induce apoptosis of ECs and tumor cells. These results, in combination with those of the present study, demonstrate that antiangiogenic therapies that target VEGF in this murine model significantly increase EC apoptosis and decrease EC signaling. VEGF protects ECs from apoptosis in vivo by activating appropriate intracellular survival pathways. The activation of these pathways can be documented by immunohistochemistry and could potentially serve as a surrogate marker of the effectiveness of antiangiogenic therapies.
We theorized that VEGF-directed antiangiogenic therapies would block important EC signaling cascades, and that this would serve as early surrogate markers of their effectiveness. To prove this hypothesis, we sequentially devised in vitro and in vivo models. Using Western blotting, pretreating the HDMEC with PD98059 effectively inhibited the activation of Erk after VEGF stimulation. Wortmannin prevented activation of Akt and, to some extent, activation of Erk. This is a finding consistent with previous studies demonstrating that wortmannin inhibits VEGF-induced Erk activation (in HUVECs) in a dose-dependent manner (8) and implicating PI-3 kinase as a modulator of MAP kinase activation by VEGF. Once we confirmed that VEGF induced activation of the signaling pathways expected, we then proceeded with the development of an in vitro immunohistochemical technique to determine the feasibility of examining signaling activation ECs, in situ. In this in vitro assay, we used immunohistochemical techniques to assess VEGF-stimulated HDMECs pretreated with specific inhibitors of signaling molecules. In the absence of VEGF stimulation, SU6668 treatment led to a decrease in the basal level of activated Erk; in the presence of VEGF, SU6668 produced an almost complete abrogation of Erk activation. Pretreatment of HDMECs with SU6668 also led to complete abrogation of any basal or stimulated levels of activated Akt in EC in vitro. In previous studies (11), VEGF was proven to ameliorate the time-dependent increase in apoptosis of serum-starved HDMECs through activation of the MAP kinase/Erk pathway together with inhibition of stress-activated protein kinase/c-jun-NH2-kinase. Other reports found that VEGF acts as a survival factor for HUVECs through its effects on the VEGFR2 (Flk-1/KDR) leading to the activation of the PI-3 kinase/Akt signaling pathway, which in turn plays a critical role in preventing HUVEC apoptosis (8, 9, 10, 19).
For our in vivo model, we used harvested liver metastases from a previous study of colon cancer metastasis in a murine model and examined viable ECs that did not have the morphology of apoptotic ECs. The antiangiogenic agents SU6668 and SU5416 block important signaling cascades in tumor ECs. The altered signaling events observed after treatments are likely to be upstream and earlier mediators of alterations in either EC proliferation (Erk), survival (Akt), or both. SU6668 antiangiogenic effects could arise from any or all of the potential receptor systems targeted by the compound. SU6668 treatment effects on ECs and tumor cells were similar to the effects seen with SU5416 treatments, pointing to VEGF as the main angiogenic and EC survival factor in this murine model.
In summary, our results demonstrate that antiangiogenic therapies that target VEGF, in this murine model, significantly increase EC apoptosis and decrease EC signaling. VEGF protects ECs from apoptosis by activating appropriate intracellular survival pathways acting in vivo as a tumor EC survival factor and an angiogenic factor. The inhibition of activation of these pathways after antiangiogenic therapies can be documented by immunohistochemistry and could potentially serve as a surrogate marker of the effectiveness of these therapies. Additional studies are necessary to confirm these results in human tissues.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
These studies were supported, in part, by NIH Grant T-32 09599 (to C. C. S.), the Jon and Susie Hall Fund for Colon Cancer Research (to L. M. E.), the Gillson Longenbaugh Foundation (to L. M. E.), and Core Grant CCSG CA16672.
The abbreviations used are: VEGF, vascular endothelial growth factor; EC, endothelial cell; MAP, mitogen activated protein; Erk, extracellular signal-regulated kinase; HDMEC, human dermal microvascular EC; hrVEGF, human recombinant VEGF; TBS, Tris-buffered saline; VEGFR2, VEGF receptor-2; HUVEC, human umbilical vascular EC; Akt, protein kinase B; PI-3, phosphatidyl inositol 3-kinase.
We thank Christine Wogan of the Department of Scientific Publications at the M. D. Anderson Cancer Center for editorial assistance.