Targeting of Mitogen-Activated Protein Kinases and Phosphatidylinositol 3 Kinase Inhibits Hepatocyte Growth Factor/Scatter Factor–Induced Angiogenesis
Background— Hepatocyte growth factor/scatter factor (HGF/SF) can sufficiently and independently induce pathophysiological angiogenesis. However, the treatment strategies have mostly been unsuccessful. The present study is the first to evaluate the possible targeting of downstream signals for the inhibition of HGF/SF-induced angiogenesis.
Methods and Results— In a multichannel scratch assay with human endothelial cells (ECs), HGF/SF induced a strong and prolonged activation of MAPK and cell proliferation that was inhibited by PD98059 and LY294002/wortmannin, selective inhibitors of MAPK and PI3K signaling modules, respectively. Western blotting demonstrated a temporal relation between the activation of the two pathways. Chemical inhibition of the PI3K and MAPK signals inhibited HGF/SF-induced chemoinvasion of ECs in vitro and blocked the HGF/SF-induced neovascularization into a polymer scaffold in vivo, as quantified by vessel counts and the clearance of radioactive 133Xe.
Conclusions— These data indicate that MEK and PI3K inhibitors represent a promising approach to the clinical management of pathological conditions characterized by overt HGF/SF-induced angiogenesis.
Received January 15, 2003; revision received February 28, 2003; accepted March 1, 2003.
Angiogenesis, the formation of neovasculature from an existing vascular bed, involves a temporal series of discrete but overlapping events, including chemoinvasion, proliferation, and tubulogenesis by endothelial cells. Given the pathophysiological implications of angiogenesis, the molecular dissection of angiogenic signaling is clinically important.
Hepatocyte growth factor/scatter factor (HGF/SF) is a mesenchymally derived signal that acts through the c-met receptor as a morphogen and mitogen for epithelial cells1 and hepatocytes2 and is implicated in both developmental and adult processes.3 In addition to mediating tumor-stromal interaction in a paracrine loop,4 HGF/SF can induce vascular endothelial growth factor (VEGF)5 and promote angiogenesis.6 However, we observed that HGF/SF could independently and sufficiently induce an angiogenic response in the presence of a VEGF-receptor blockade, necessitating the development of alternative strategies to inhibit HGF/SF-induced pathophysiological angiogenesis.7
In an interesting study, c-met was shown to play a prominent role during Ras-mediated tumor growth and metastasis.8 Ras has emerged as a convergent molecular switch that integrates and propagates extracellular signals to downstream cascades, the best characterized of which are the phosphatidylinositol-3-kinase (PI3K) and the mitogen-activated protein kinase (MAPK) pathways. By acting through these pathways, HGF/SF has been implicated in inducing the expression of proangiogenic cytokines in carcinoma.9 The present study was therefore designed to elucidate and target the downstream signaling cascades in an attempt to inhibit the HGF/SF-induced angiogenesis in vitro and in vivo.
PI3K inhibitors, wortmannin and LY294002 ([2-(4-Morpholinyl)-8-phenyl-4H-1-benzopyran-4–1]) were obtained from Sigma and Calbiochem, respectively. PD98059 (2-(2-amino-3-methoxyphenyl)-4H-1-benzopyran-4-1), a MEK inhibitor, was from Tocris. Radioactive 133Xenon was purchased from Dupont Pharma. PTK787, a VEGF receptor antagonist, was a gift from Dr Jeanette Wood of Novartis.
Preparation of HGF/SF
Murine recombinant HGF/SF was used for the mouse experiments and was generated by using the NSO mouse melanoma cell line transfected by electroporation with a mouse HGF/SF cDNA. For experiments involving human endothelial cells, the NSO cells were transfected with a human HGF/SF cDNA to generate hrHGF/SF. The proteins were purified by elution through a heparin-sepharose column and a Mono-S column.
In Vitro Scratch Model
A confluent monolayer of synchronized human umbilical vein endothelial cells (HUVECS) was scraped with a multichannel wounder,10 thereby producing 11 parallel lesions, each 400 μm wide, on the monolayer. Coverslips were rinsed in PBS to dislodge any cellular debris and placed into a well containing the appropriate treatment. Cells were pretreated with the enzyme inhibitors for 1 hour before wounding of the monolayer. At 24 hours after the lesion was created, coverslips were washed with ice-cold PBS and fixed in 4% formaldehyde.
Recovery of the denuded area was quantified with a Leica Q500, semiautomated, computerized image analysis system. Images were grabbed with a Nikon Diaphot inverted microscope coupled to a CCD (JVC). For each coverslip, 4 fields of view were selected at random. The lesion area of each field of view was measured and then converted to give percent regeneration relative to T0 values (attained from coverslips fixed at time of injury).
Endothelial cells (ECs) were injured as described above and cultured for 24 hours in media supplemented with 1% FCS and the appropriate treatment. At the end of this period, they were washed in ice cold PBS and trypsinized, and cell growth was expressed as counts by use of the trypan blue exclusion method.
In Vivo Scaffold Implant Angiogenesis Assay
This assay has been extensively validated in our laboratory against established methods such as hemoglobin measurements, 113Sn-microsphere clearance, and histology.11 Male balb/c mice (Tucks) were anesthetized with 4% and maintained on 2% isoflurane, in a mixture of oxygen (0.8 L/min) and nitrous oxide (0.6 L/min). An incision was made 0.5 cm caudal from the base of the tail, and two bilateral subcutaneous air pockets were created up to the dorsal subscapular region. A sterile polyether polyurethane scaffold (160 mm3) was inserted into each pocket, and the incision was closed with silk sutures (Mersilk).
Administration of the drugs into the scaffold was started 24 hours after implantation and continued for 10 days. A Precision Glide 30-gauge needle (Sigma) was used to deliver the injection, and the total volume administered into the implant was kept constant at 40 μL. Appropriate vehicle-treated control groups were run alongside each experiment. All the in vivo procedures were approved and conformed to the UK Home Office guidelines for handling of experimental animals.
Assay for Functional Status of Neovasculature
On day 15, the animals were anesthetized with a combination of fentanyl citrate–fluanisone and midazolam (diluted 1:1 in 20 in saline). Vascularization was assessed as a function of the blood flow through the implants, by direct injection of 133Xe-containing saline into the scaffold and monitoring its clearance over a 6-minute period. Radioactivity was measured with a microprocessor scalar ratemeter (Nuclear Enterprise) linked to a collimated, low-energy X-ray/γ-ray NaI-crystal with an Al-entrance window, on an HG-type mount coupled to an NE 5289C preamplifier. The data were expressed as the percentage of 133Xe cleared at every 40 seconds, calculated according to the following formula: [initial count−count at t(s)] × 100% initial count.12
Macroscopic Vessel Count
After the 133Xe-clearance measurements, animals were killed by CO2 exposure and cervical dislocation, and the dorsal skin flap was everted. The gross angiogenic response was photographed with the use of a macro lens connected to a Nikon SLR camera. Prints were developed on Kodak plates. Angiogenesis was quantified as the in-growth of vessels in the sponge-granuloma tissue.
In Vivo Matrigel Angiogenesis Assay
Growth factor–reduced matrigel (BD), mixed with growth factor and/or drugs, was injected subcutaneously into male C57/BL6 mice. On day 9, the animals were killed by CO2 exposure, and the skin was everted. Gross response was recorded with a high-resolution digital camera (Canon). The implants were excised and cryofrozen in OTC for immunohistochemistry.
Immunohistochemistry and Confocal Microscopy
The matrigel sections (12 μm) were fixed in −20°C methanol and probed with antibody against von Willebrand factor, an endothelial cell marker. The signal was amplified with the use of a Texas red–labeled secondary antibody, and the images were captured with a Zeiss LSM510 confocal microscope. Propidium iodide was used to counterstain nuclei.
Phosphorylation of ERK1/2 and Akt
Equivalent amounts of protein per sample were electrophoretically resolved on 10% SDS-PAGE and transferred to a membrane. Proteins were detected through the use of antibodies directed to phospho ERK1/2 and phospho Akt (1:800 dilution, New England BioLabs). Polyclonal anti-Akt and anti-ERK1/ERK2 antibodies (Santa Cruz Biotechnology) were used at a 1:500 dilution to detect the level of total proteins. Membranes were incubated with a 1:2000 dilution of the appropriate horseradish peroxidase–conjugated secondary antibody (Amersham). The immunocomplexes were visualized with the use of enhanced chemiluminescence detection (Amersham Life Science). The Western blots shown were representative of at least three separate experiments, and each panel was taken from a single immunoblot.
Statistical significance was tested by using 1-way ANOVA followed by Dunnett’s or Friedman’s post hoc test. Bonferroni’s test was used to test for overall dose-response effects (Graphpad Prism 3 software). A value of P<0.05 was considered significant.
Effect of Pharmacological Inhibition of MAPK on HGF/SF-Induced Reendothelialization Response
As shown in Figure 1A, in comparison to the vehicle, HGF/SF (1 nmol/L) induced a strong phosphorylation of ERK1/2 as early as 10 minutes, and the response persisted until at least 10 hours after treatment. Treatment with PD98059 blocked the HGF/SF-induced phosphorylation of the ERKs. HGF/SF-induced a significant increase in the total recovery of the denuded area, or reendothelialization, as compared with the vehicle-treated cells, an effect that was blocked by PD98059 (Figure 1).
Regeneration of an injured area involves both chemokinesis and cell division.10 In the present study, treatment with HGF/SF induced a significant increase in the cell count, which was blocked by PD98059 (Figure 1), suggesting that the MAPK pathway is implicated in the mitogenic effect of HGF/SF on endothelial cells. Higher concentrations of PD98059 inhibited basal proliferation. The vehicle for PD98059, DMSO (0.01% to 0.5%) did not inhibit regeneration or proliferation.
Effect of Pharmacological Inhibition of PI3K on HGF/SF-Induced Reendothelialization Response
Both LY294002 and wortmannin produced a concentration-dependent inhibition of the HGF/SF-induced repair of the injured monolayer of HUVECs (Figure 2). Cell proliferation induced by HGF/SF was inhibited by LY294002 and wortmannin in a concentration-dependent manner (Figure 2). However, at concentrations higher than 10 μmol/L for LY294002, and 10 nmol/L for wortmannin, there was a reduction in the basal cell counts, suggesting that the basal recovery is also dependent on the PI3K activity. DMSO, used as a vehicle for LY294002, did not affect basal regeneration or cell counts at the final dilutions used.
Western blotting revealed that HGF/SF activated PI3K evident by the phosphorylation of the downstream substrate Akt, which is inhibited by LY294002 (Figure 2) Furthermore, the late-stage phosphorylation of MAPK was more susceptible to inhibition by LY294002 (1 μmol/L), whereas early phosphorylation was partially inhibited at a much higher concentration (20 μmol/L) (Figure 2G). Interestingly, the late-phase activation of MAPK by HGF/SF was not affected by PTK787, a VEGFR antagonist (Figure 2).
Induction of Angiogenesis by HGF/SF In Vivo and Effect of Pharmacological Inhibitors of MAPK and PI3K
As shown in Figure 3, HGF/SF induced a dose-dependent neovascular response into the scaffold, as is evident by the increase in the vessel counts and total 133Xe cleared in the polyether-polyurethane scaffold implant murine model. A maximal effect was reached at 30 ng/sponge of HGF/SF. Interestingly, we detected a progressive increase in phospho-Akt signal in the HGF/SF-treated group over the 10-day investigative period, which was stronger than the vehicle-treated group after day 2 time points (Figure 3H). To further implicate MAPK and PI3K in the HGF/SF-induced angiogenesis in vivo, PD98059 and wortmannin were administered into the implants before the growth factor. As shown in Figure 3, both PD98059 and wortmannin could block the HGF/SF-induced neovascularization at the higher doses used, as is evident from vessel counts and 133Xe clearance (Figure 3 and Table). At these doses, no gross phenotypic abnormalities were observed, nor was there any loss of body weight (Figure 3).
In the matrigel implant assay, gross morphological assessment of HGF/SF-treated gels revealed a significantly greater neovascular response than seen in the presence of PD98059 and LY294002 (Figure 4). However, this discrimination was lesser when quantified after immunolabeling against von Willebrand factor on endothelial cells. A possible reason for this discrepancy could be that a signaling blockade inhibits the HGF/SF-induced recruitment of endothelial cells, and this is evident from the lesser number of propidium iodide–positive infiltrating cells in the inhibitor-treated implants. Most likely, the individual migrating endothelial cells are too few to form a functional lumen. The coadministration of PD98059 and LY294002 did not show any additive or synergistic effects (Figure 4). The vehicle for the inhibitors did not affect the basal or HGF/SF-induced angiogenesis.
In the present study, we established that HGF/SF induces angiogenesis by activating the MAPK and the PI3K pathways and that targeting these signal transducers by using pharmacological inhibitors could be harnessed as a therapeutic approach for the inhibition of HGF/SF-induced pathological angiogenesis. Furthermore, we demonstrated that there is a temporal relation between the activation of the two pathways, with a prolonged activation of the MAP kinase pathway and the delayed phase being more susceptible to modulation by the PI3K pathway. By using a deconstructional approach, in which we replicated all the key steps in the angiogenic process in vitro, we demonstrated that HGF/SF could promote each of these steps, that is, chemoinvasion, proliferation, and tube formation by endothelial cells, spatiotemporally implicating the two signaling cascades in these steps for the first time.
HGF/SF, and its receptor c-met, have been implicated in contributing to tumorigenicity by altering cellular differentiation and by stimulating tumor neovascularization.13 Furthermore, HGF/SF has been implicated in inducing angiogenesis in diabetic retinopathies,14 psoriasis,15 and arthritis,16 both in cohesion with and independent of other angiogenic factors.6,17 Interestingly, however, research into the mechanisms of angiogenic action of HGF/SF and the development of modulating therapeutics against it has been overshadowed by the importance being placed on VEGF.
An interesting approach for inhibiting HGF/SF-induced angiogenesis would be by blocking the downstream signal transduction cascade. The optimal level for intervention in a signaling pathway remains debatable but should reflect a maximal therapeutic outcome with minimal toxicity to normal tissues. However, most of the studies elucidating the mechanisms of action of HGF/SF have been carried out with epithelial cells that undergo a scattering response on exposure to the growth factor3 and have implicated the Ras→MEK→MAPK and the PI3K pathways in the phenotypic outcome.18 To elucidate the angiogenic mechanisms of action of HGF/SF, we used a multiple scratch model to mimic the pathophysiological situation and amplify signals.11 Contact-inhibited endothelial cells were injured, switching them from a physiological “off’ state to a pathophysiological “on” state. Incubation with HGF/SF promoted a fast and prolonged activation of the MAPK pathway, indicated by the rapid change in the phosphorylation of ERK1/2, which persisted for 10 hours after injury. Indeed, HGF/SF promoted recovery of the denuded area and proliferation of the endothelial cells, effects that were blocked by PD98059 and LY294002, highly selective pharmacological inhibitors of the MEK and PI3K, respectively,19 implicating both these pathways in HGF/SF-induced angiogenesis.
The MAPK pathway is among the well-characterized cascades downstream of Ras, and mutation of the latter resulting in the permanent “switching on” is implicated in tumorigenesis. HGF/SF had been shown to activate Ras,20 which can interact with GTPase-activating proteins (GAP), Grb2 proteins, or Sos. However, therapeutic approaches targeting Ras-Grb2, Ras-Sos, or Ras-GAP have not been preclinically successful.21 Binding of Ras to GTP results in the activation of Raf kinase, which then phosphorylates MEK triggering the sequential phosphorylation of the MAPK module. In an interesting study, Sebolt-Leopold et al22 reported the suppression of the growth of colon tumors with a concomitant reduction in neovascularization after treatment with a MAPK inhibitor, PD184352, currently undergoing clinical trials. In the present study, we observed a significant reduction in the HGF/SF-induced functional angiogenesis in vivo. This was consistent with the observations of Milella et al,23 in which the MEK→MAPK transduction module was targeted by small molecule inhibitors, resulting in a profound inhibition of acute myeloid leukemia, with little effect on cells with low steady-state MAPK activity. However, we did observe that higher concentrations of PD98059 inhibited the basal cell proliferation in vitro, suggesting that basal recovery by endothelial cells after an injury is dependent on MAPK activity.
In the temporal sequence of angiogenesis, endothelial cell proliferation is preceded by the invasion and migration of the cells toward an angiogenic cue. We found that both PD98059 and LY294002 could block the chemoinvasion of endothelial cells promoted by HGF/SF without altering the basal response (see Data Supplement Figure I). HGF/SF promotes cell adhesion to the extracellular matrix, followed by invasion mediated by integrins through a PI3K-dependent mechanism.24 Although PI3K inhibition blocked HGF/SF-induced tubulogenesis in a matrigel assay,7 the MAPK inhibitor failed to induce a similar effect (see Data Supplement Figure II). This was consistent with a recent report in which U0106, a MAPK inhibitor, had no effect on HUVEC tubulogenesis.25
Mutated c-met receptors that hyperactivated Ras interact more efficiently with PI3K,26 and activated mutants of PI3K can activate Akt.27 HGF/SF was found to phosphorylate Akt in vivo, and treatment with the PI3K inhibitors completely abolished the angiogenic effects of HGF/SF and the phosphorylation of Akt. Furthermore, PI3K inhibitors resulted in the loss of cells below basal level, suggesting that PI3K could be mediating an additional survival signal, which can be studied further.
An interesting observation in this study, implications of which on angiogenic outcome is yet unclear, was that pretreatment with LY294002 inhibited the late-phase HGF/SF-induced phosphorylation of ERK1/2, suggesting a contribution of the PI3K pathway into the prolonged activation of the MAPK pathways, though the fast activation of the latter less susceptible to such modulation. Additionally, the failure of PTK787 to block the late-phase phosphorylation of ERKs indicates that this component is independent of the paracrine involvement of VEGF. It has been demonstrated that signals emerging from PI3Kγ could phosphorylate MAPK,28 and this sequence of activation was crucial for the HGF/SF-induced scattering of human hepatoma cells.29 The current observation could explain PI3K as a key and sufficient signal for mediating the HGF/SF-induced angiogenesis.
Sebolt-Leopold22 demonstrated the antitumorigenic potential of targeting the MAPK cascade. Furthermore, Solorzano et al30 suggested that 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. The present study is the first to evaluate the possible targeting of these two signaling pathways for the inhibition of HGF/SF-induced angiogenesis. Furthermore, both these pathways were implicated in mediating the HGF/SF-induced release of proangiogenic cytokines,9 and obviously the inhibitors would disrupt the carcinoma-EC crosstalk. These pathways are additionally upregulated by other angiogenic factors, such as VEGF, FGF. And so forth, and therefore a downstream blockade could implement a global blockade of pathological angiogenesis. Although the present study did not raise any obvious toxicity profile, there are still insufficient safety profiles on the inhibitors of the MAPK or PI3K module; however, this may emerge as a putative strategy to manage conditions characterized by overt activation of the signals.
Additional information, including 2 figures, is available in the online-only Data Supplement at http://www.circulationaha.org.
Gherardi E, Gray J, Stoker M, et al. Purification of scatter factor, a fibroblast-derived basic protein that modulates epithelial interactions and movement. Proc Natl Acad Sci U S A. 1989; 86: 5844–5848.
Bussolino F, Di Renzo MF, Ziche M, et al. Hepatocyte growth factor is a potent angiogenic factor which stimulates endothelial cell motility and growth. J Cell Biol. 1992; 119: 629–641.
Wojta J, Kaun C, Breuss JM, et al. Hepatocyte growth factor increases expression of vascular endothelial growth factor and plasminogen activator inhibitor-1 in human keratinocytes and the vascular endothelial growth factor receptor flk-1 in human endothelial cells. Lab Invest. 1999; 79: 427–438.
Sengupta S, Gherardi E, Sellers LA, et al. Hepatocyte growth factor/scatter factor can induce angiogenesis independently of vascular endothelial growth factor. Arterioscler Thromb Vasc Biol. 2003; 23: 69–75.
Furge KA, Kiewlich D, Le P, et al. Suppression of Ras-mediated tumorigenicity and metastasis through inhibition of the Met receptor tyrosine kinase. Proc Natl Acad Sci U S A. 2001; 98: 10722–10727.
Dong G, Chen Z, Li Z, et al. Hepatocyte growth factor/scatter factor-induced activation of MEK and PI3K signal pathways contributes to expression of pro-angiogenic cytokines interleukin-8 and vascular endothelial growth factor in head and neck squamous cell carcinoma. Cancer Res. 2001; 61: 5911–5918.
Grant DS, Kleinman HK, Goldberg ID, et al. Scatter factor induces blood vessel formation in vivo. Proc Natl Acad Sci U S A. 1993; 90: 1937–1941.
Nagashima M, Hasegawa J, Kato K, et al. Hepatocyte growth factor (HGF), HGF activator, and c-Met in synovial tissues in rheumatoid arthritis and osteoarthritis. J Rheumatol. 2001; 28: 1772–1778.
Paumelle R, Tulasne D, Leroy C, et al. Sequential activation of ERK and repression of JNK by scatter factor/hepatocyte growth factor in Madin-Darby Canine Kidney epithelial cells. Mol Biol Cell. 2000; 11: 3751–3763.
Graziani A, Gramaglia D, Dalla Zonca P, et al. Hepatocyte growth factor/scatter factor stimulates the Ras-guanine nucleotide exchanger. J Biol Chem. 1993; 268: 9165–9168.
Milella M, Kornblau SM, Estrov Z, et al. Therapeutic targeting of the MEK/MAPK signal transduction module in acute myeloid leukemia. J Cell Invest. 2001; 108: 851–859.
Trusolino L, Cavassa S, Angelini P, et al. HGF/Scatter factor selectively promotes cell invasion by increasing integrin avidity. FASEB J. 2000; 14: 1629–1640.
Grove AD, Prabhu VV, Young BL, et al. Both Protein activation and gene expression are involved in early vascular tube formation in vitro. Clin Cancer Res. 2002; 9: 3019–3026.
Giordano S, Maffe A, Williams TA, et al. Different point mutations in the met oncogene elicit distinct biological properties. FASEB J. 2000; 14: 399–406.
Franke TF, Kaplan DR, Cantley LC, et al. Direct regulation of the Akt proto-oncogene product by phosphatidylinositol-3, 4-biphosphate. Science. 1997; 275: 665–668.
Bondeva T, Pirola L, Bulgarelli-Leva G, et al. Bifurcation of lipid and protein kinase signals of PI3Kg to protein kinases PKB and MAPK. Science. 1998; 282: 293–296.
Solorzano CC, Jung YD, Bucana CD, et al. In vivo intracellular signaling as a marker for antiangiogenic activity. Cancer Res. 2001; 61: 7048–7051.