Honokiol, a small molecular weight natural product, inhibits angiogenesis in vitro and tumor growth in vivo

Natural products comprise a major source of small molecular weight angiogenesis inhibitors. We have used the transformed endothelial cell line SVR as an effective screen of natural product extracts to isolate anti-angiogenesis and anti-tumor compounds. Aqueous extracts of Magnolia grandiflora exhibit potent activity in our SVR proliferation assays. We found that the small molecular weight compound honokiol is the active principle of magnolia extract. Honokiol exhibited potent anti-proliferative activity against SVR cells in vitro. In addition, honokiol demonstrated preferential inhibition of primary human endothelial cells compared with fibroblasts and this inhibition was antagonized by antibodies against TNF alpha-related apoptosis-inducing ligand. In vivo, honokiol was highly effective against angiosarcoma in nude mice. Our preclinical data suggests that honokiol is a systemically available and non-toxic inhibitor of angiogenesis and should be further evaluated as a potential chemotherapeutic agent.


Introduction
Angiogenesis inhibitors have been derived from a number of sources, including cleaved proteins, monoclonal antibodies, and natural products. Natural products contain a variety of chemopreventive compounds that have been shown to prevent the development of malignancies 1;2 . We and others have discovered that some of these chemopreventive agents have antiangiogenic activities which may account in part for their chemopreventive effects. These compounds include curcumin, from Curcuma longa, epicatechin gallate from tea, genistein from soybeans, and resveratrol from grapes and red wine 3 4-6 .
These compounds exert antiangiogenic and chemopreventive properties through a variety of mechanisms. Curcumin inhibits angiogenesis by both direct effects on endothelium as well as by inhibiting the COP9 signalosome-associated kinase activity, which regulates the degradation of the c-jun oncogene, with consequent downstream effects on synthesis of the potent angiogenic factor vascular endothelial growth factor (VEGF) 7;8 . Epicatechin gallate works in part through inhibiting activity of the 26S proteasome, which may also regulate synthesis of VEGF 4;9 . Genistein and resveratrol are broad spectrum protein kinase inhibitors which inhibit tumor promotion 1;2;10;11 . However, few of these compounds actually exhibit activity against established tumors in vivo.
We have developed a simple bioassay amenable to large-scale screening and fractionation of natural products, namely inhibition of proliferation of the transformed endothelial cell line SVR 12 . Using this bioassay on extracts of the seed cone of Magnolia grandiflora, we have shown that one of the active components of this extract is the small molecule honokiol. We demonstrate that honokiol inhibits angiogenesis by interfering with phosphorylation of VEGFR2 in human endothelial cells. In addition, honokiol inhibits the growth of transformed epithelial cells in vitro, thus demonstrating that it has both antiangiogenic and antitumor activity. Honokiol is well tolerated and effective against sarcomas in mice, making it an attractive candidate for clinical trials.

Extraction of Magnolia grandiflora seed cones
Magnolia grandiflora seed cones were collected and ground. The powdered Magnolia cones (100g) were extracted with 500 ml boiling water for 30 minutes, then allowed to cool to room temperature. The crude aqueous extract was clarified using a 0.45 µ filter, followed by ultrafiltration with 3000 NMWL. The ultrafiltrate was lyophilized, and then reconstituted in distilled water to give a final concentration of 500 mg/ml. The material was then fractionated by HPLC, and fractions were lyophilized and reconstituted as 10 mg/ml solutions. These fractions were tested on proliferation assays on SVR cells as described below. Honokiol and magnolol were obtained from Wako Chemical Company (Tokyo, Japan), and unsubstiuted biphenyls were obtained from Aldrich Chemical Company (St Louis, MO).

In vitro proliferation assays
10,000 SVR cells were plated in 24 well dishes. The next day, the media was replaced with fresh media containing the inhibitors or vehicle controls. Cells were incubated at 37 o C for 72 hours 12;13 , and cell number was determined in triplicate using a Coulter Counter (Hialeah, FL). Immortalized and K-ras transformed rat epithelial cells (RIEpZip and RIE pZipKRas12V) and fibroblasts (NIH3T3 pZip and NIH3T3 pZipK-Ras12V) were maintained at 37° C, 10% CO 2 , in Dulbecco's modified Eagle's medium supplemented with 5% FCS (RIE) or 10% calf serum (NIH3T3) 14;15 . Cells were plated at 10 5 per well in six well plates. Vector and ras transformed NIH3T3 and RIE cells were treated with either vehicle (20 µl DMSO) or increasing concentrations (5,10,20,40 µg/ml) of honokiol (from a 2 mg/ml DMSO stock), and observed for morphology changes after 24 h.

Apoptosis Assays
SVR cells were plated at 125,000 cells per 100 mm plate in 5% FBS/DME. After 24 hours, cells were treated with 10µg/ml magnolol or honokiol, or left untreated as control. At 18 hours, and at 48 hours of treatment, two plates per condition were analyzed. Adherent cells were washed with PBS, and the cells suspended with trypsin/EDTA treatment.
Floating cells were also collected by centrifugation of the conditioned medium, and the total cell population was analyzed. Cell surface Annexin V was measured by flow cytometry using the ApoAlert Annexin V kit (Clontech, Palo Alto, CA) as described by the manufacturer. The cells were washed in 1X Binding Buffer by centrifugation and then resuspended in 200µl of 1X Binding Buffer containing Annexin V (0.1µg/ml) and propidium iodide (PI, 0.5µg/ml). After incubation at room temperature for 15 min., the cells were analyzed by flow cytometry for the presence of Annexin V and propidium iodide.

Analysis of PI 3 kinase and MAP kinase signaling
SVR angiosarcoma cells were cultured in low glucose DMEM containing 10% fetal bovine serum. For experimental cultures, honokiol was added from a 10 mg/ml stock solution made in DMSO and used at final concentrations of 20-45 µg/ml as indicated.

Lipid kinase assays
Five µg each of HA-p85 and Myc-p110 were cotransfected into HEK 293 cells according to previously published methods 16 17 . After 24 hr, cells were treated with 10 µM honokiol or same volume of DMSO as control for another 24 hr. After removal of the culture medium, cells were washed with 5 ml ice-cold PBS twice, lysed in 0.5 ml lysis buffer A (50 mM Tris, pH 7.4, 40 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 1.5 mM Na 3 VO 4 , 50 mM NaF, 10 mM sodium pyrophosphate, 10 mM sodium βglycerophosphate, 1 mM phenylmethylsulfonyl flouride (PMSF), 5 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin A), and was centrifuged for 10 min at 14,000 x g at After 10 min incubation at room temperature, the mixture was solubilized in 8 µl of 37% HCl was added, and vortexed for a few seconds. 150 µl of 1:1 CHCl 3 : CH 3 OH was introduced and mixed, then centrifuged for 5 min. The bottom organic layer was removed into a fresh tube and air dried overnight. The next morning 10 µl of methanol was added to dissolve the lipid, then it was spotted onto a TLC plate and the lipids were separated by 65:35 (V/V) 2-propanol: 2M acetic acid. After the TLC plate was dried, it was exposed to a film.

VEGFR2 Phosphorylation Analysis
Human recombinant VEGF 165 was purchased from R&D Systems (Minneapolis, MN, USA). Anti-KDR antibody, anti-phosphotyrosine (PY99) antibody and protein A-G agarose were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). HUVECs were obtained from Emory Skin Diseases Research Center. Cells were grown on plates coated with 0.1% gelatin in EGM-MV BulletKit (Clonetics, San Diego, CA, USA; 10% fetal bovine serum in endothelial basic medium (EBM) with 12 µg/ml bovine brain extract, 1 µg/ml hydrocortisone, 1 µl/ml GA-1000, and hEGF). Experiments were performed using cells between passages 2 and 5. Growth-arrested HUVECs were stimulated with agonists at 37°C, and cells were lysed with 500 µl of ice-cold lysis buffer, pH 7.4 ((in mM) 50 HEPES, 5 EDTA, 50 NaCl), 1% Triton X-100, protease inhibitors (10 µg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin) and phosphatase inhibitors ((in mM) 50 sodium fluoride, 1 sodium orthovanadate, 10 sodium pyrophosphate). For immunoprecipitation cell lysates (600 µg) were precipitated with antibody overnight at 4°C and then incubated with 25 µl of protein A-G-agarose beads for 1.5 h at 4°C. Cell immunoprecipitates (500 µg) were separated using SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, blocked overnight in PBS containing 6% nonfat dry milk and 0.1% Tween 20, and incubated for 1 h with primary antibodies as described previously (20). After incubation with secondary antibodies, proteins were detected by ECL chemiluminescence. The amount of KDR in each cell extract was assessed by immunoblotting with anti-KDR antibody. Results were expressed as mean±S.E. Statistical significance was assessed by Student's paired two-tailed t-test or analysis of variance on untransformed data, followed by comparison of group averages by contrast analysis, using the SuperANOVA statistical program (Abacus Concepts, Berkeley, CA). A p value of <0.05 was considered to be statistically significant.
Rac Activation Assay-HUVECs were grown to confluence and made quiescent in 0.5% fetal bovine serum for 12 hours before stimulation with VEGF (20 ng/ml). Cells were lysed with ice-cold lysis buffer (pH 7.5) containing 25 mmol/L HEPES, 150 mmol/L NaCl, 1% IGEPAL CA-630, 0.25% sodium deoxycholate, 10 mmol/L MgCl 2 , 10% glycerol, 25 mmol/L NaF, 1 mmol/L EDTA, 1 mmol/L sodium orthovanadate, 10 µg/mL leupeptin, 10 µg/mL aprotinin, and 1 mmol/L phenylmethylsulfonyl fluoride. Activated (GTP-bound) Rac was affinity-precipitated with p21-activated kinase-1 (PAK-1) protein binding domain peptide, which binds only to Rac-GTP and not Rac-GDP. PAK-1 protein binding domain agarose (PAK-1 PBD, 7.5 µg/mg cell lysate) was added, and the reaction mixture was gently rocked at 4°C for 60 minutes. The agarose beads were collected by pulsing for 5 seconds in a microcentrifuge at 14 000g, and the beads were washed three times with 0.5 mL lysis buffer. The agarose beads were resuspended in 40 µL of 1x SDS sample buffer and boiled for 5 minutes. The supernatant was separated by SDS-PAGE on a 12% gel, and the proteins were transferred to nitrocellulose. After blocking for 1 hour in PBS containing 5% nonfat dry milk and 0.1% Tween 20, the membrane was incubated with anti-Rac antibody (1:1000 dilution) overnight. After incubation with the secondary antibody, Rac was detected by enhanced chemiluminescence.
Statistical Analysis-Results are expressed as mean±S.E. Statistical significance was assessed by Student's paired two-tailed t-test or analysis of variance on untransformed data, followed by comparison of group averages by contrast analysis, using the SuperANOVA statistical program (Abacus Concepts, Berkeley, CA). A p value of <0.05 was considered to be statistically significant. phosphorylation and rac activation, HUVECs were treated with TRAIL antibody or mouse IgG (30 microgram/ml) for 15 hours before addition of honokiol (10 microgram/ml) for 1 hour in 0.5% FBS containing cultured medium. Cells were then stimulated with VEGF (20 ng/ml) for 3 min. Cell lysates were assessed for measurement of tyrosine phosphorylation of VEGFR2 or Rac activity, as described above.

COP9 Signalosome-associated kinase Assays:
Kinase reaction was carried out in a final volume of 20 µl in the presence of 1 µg of recombinant c-Jun, 32 P-γ-ATP and isolated COP9 signalosome from human erythrocytes.
The reaction mix was incubated for 60 min at 37 o C. Then the complete reaction mix was separated by SDS-PAGE. The gel was dried and autoradiographed. % activity was determined by densitometry. As negative controls, assays were performed in the absence of compounds, which represent 100% activity 7;8 . Compounds were tested in two concentrations: 10 and 50 µM.

Kinase inhibition assays
Honokiol and magnolol were tested in vitro for inhibitory activity against the following

Fractionation of magnolia extracts
Aqueous magnolia extract displayed potent inhibitory effects on SVR cells (data not shown). HPLC fractions of magnolia extracts corresponded to fractions known to contain magnolol and honokiol [21][22][23] .

Effect of purified magnolia compounds on SVR proliferation
Given the potential importance of natural products as antitumor and anti-angiogenesis agents, honokiol and magnolol were tested for their effects on the survival and proliferation of SVR cells, and a steep decline in cell number was seen between 4 and 8 µg/ml of honokiol ( Figure 1A). A dose dependent decrease in cell number was seen at higher concentrations of magnolol, but given the higher potency of honokiol in our proliferation assay, we chose to focus on honokiol. Both honokiol and magnolol are substituted hydroxybiphenyls, and thus we tested the effect of nonsubstituted hydroxybiphenyls ( Figure 1B). The unsubstituted biphenyls are essentially inactive in the SVR bioassay, suggesting that the substitution is essential for bioactivity.

Effects of magnolia compounds on apoptosis
SVR cells were treated with magnolol and honokiol at 10 µg/ml. As noted above, the cellular growth rates were reduced by both agents. At 18 hours of Honokiol treatment (10 µg/ml), there was a two-fold increase in the early apoptotic cells, as measured by Annexin V positive, P.I. negative. This further increased to a 7.7 fold increase in early apoptotic cells by 48 hours of treatment, to 10.8 % of total cells (Figure 2). In contrast, magnolol, at comparable concentrations did not induce apoptosis, as assessed by annexin V positivity. These data indicate that honokiol exerts much of its suppressive effect on SVR cells by the induction of apoptosis.
The oncoprotein src can activate the phosphoinositol-3 kinase and MAPK pathways. To determine whether inhibition of Akt and MAPK phosphorylation by honokiol was due to upstream inhibition of src, SVR cells were incubated with honokiol and then examined for changes in src phosphorylation. Honokiol at high concentrations inhibited phosphorylation of c-src in SVR cells (Fig. 3A). Treatment with lower concentrations of honokiol for extended times did not cause inhibition of src phosphorylation (Fig.3A,B).
The effect of honokiol treatment on SVR cells suggested a preferential effect on PI3 kinase signaling compared with MAP kinase signaling, as phosphorylation of akt was inhibited by lower doses than that of MAP kinase. To determine whether honokiol had a direct effect on production of inositol phosphates, we examined the effect of honokiol on synthesis of these phosphates. Honokiol treatment led to approximately 50% inhibition of levels of phosphorylated inositol (data not shown). In order to determine whether honokiol directly antagonized ras, the ability of honokiol to inhibit the growth of immortalized and ras transformed epithelial and mesenchymal cells was tested. While honokiol exhibited dose-dependent inhibition of cell growth, there was no significant difference in inhibition between immortalized and ras transformed cells (data not shown).
In addition, morphologic reversion that occurs in these cells when ras is specifically inhibited was not observed as a result of honokiol treatment. These findings suggest that honokiol has activity against both preneoplastic (immortalized) and neoplastic (rastransformed) tumor cells, but does not specifically inhibit ras signaling. These findings, along with the preferential activity against PI3 kinase over MAP kinase signaling, made us consider TRAIL as a potential intermediary of honokiol activity.

Honokiol preferentially inhibits growth of primary human endothelial cells over fibroblasts
In order for a molecule to be considered an angiogenesis inhibitor, it must have preferential inhibitory activity against endothelial cells versus non-endothelial primary cells. In order to test whether this is the case for honokiol, we tested the ability of honokiol to inhibit the growth of primary fibroblasts and dermal endothelial cells.
Honokiol exhibited preferential inhibition of endothelial cells over fibroblasts, in a dose dependent fashion (Figure 4).

Honokiol mediated inhibition of endothelial cell growth is mediated by TRAIL
The lack of specific effects of honokiol on ras antagonism led us to explore alternative mechanisms of honokiol. In addition, the known antagonism that PI3 kinase shows against TRAIL activity, along with the known induction of MAP kinase activation by TRAIL made TRAIL a candidate for honokiol's effect on cell growth. We examined the effect of antibodies against TRAIL on the effect of honokiol on primary human endothelial cells ( Figure 5). Treatment with antibodies against TRAIL inhibited honokiol's activity against endothelium, while isotype control antibodies had no effect.
Thus, honokiol's activity is mediated in part by TRAIL.

Honokiol inhibits VEGF-induced Rac1 activation in human endothelial cells-We
have previously demonstrated that Rac1 activation is required for VEGF-induced production of ROS derived from NAD(P)H oxidase and subsequent KDR autophoshorylation in HUVEC 26 . Since honokiol inhibited VEGF-induced KDR autophosphorylation, we next examined whether this effect is mediated through the inhibition of Rac1. As shown in Figure 6B

Honokiol exhibits antitumor activity in mice
In order to determine whether honokiol exhibited antitumor activity in vivo, mice were inoculated with 1 x 10 6 SVR angiosarcoma cells subcutaneously and treated with honokiol or vehicle when tumors became clinically evident. Honokiol treatment led to approximately 50% inhibition of tumor growth compared with vehicle control p<0.05 ( Figure 7).

Discussion
We describe the isolation of a systemically active inhibitor of tumor growth through a rapid bioassay of proliferation of SVR transformed endothelium. We have previously shown that SVR cells accurately predict in vivo responses of two known angiogenesis inhibitors, TNP 1470 and 2-methoxyestradiol 12;29 . In addition, curcumin, a natural product not previously known to be antiangiogenic, was demonstrated to have antiangiogenic activity on both immortalized and primary endothelial cultures 3;30 .
A number of chemopreventive agents have been isolated, and characterized primarily through activities against skin and colon cancer promotion protocols. These include tea polyphenols, curcumin, caffeic acid phenethyl ester (CAPE) 6;31-33 . While many of these agents have potent chemopreventive activities, their activity against established tumors are not potent. One possible explanation is that high concentrations of drug are achievable at the skin and colon, while sustained systemic concentrations cannot be achieved. This is clearly the case with curcumin, in which previous studies have shown that it is poorly absorbed from the gastrointestinal tract and even when systemically administered, is rapidly cleared by hepatic metabolism 1 . We have found that honokiol, unlike curcumin, can inhibit growth of an established tumor, when administered systemically.
Our fractionation process, using inhibition of transformed SVR endothelial cell proliferation as a bioassay, yielded a fraction that contains magnolol and honokiol, substituted biphenols. Magnolol and honokiol are closely related, but honokiol appeared to have enhanced activity in the SVR inhibition assay. Substitution of the hydroxylated biphenyl is required for activity, as the 2,2' and 4,4' dihydroxybiphenyl had no effect on proliferation of SVR cells in vitro. Consistent with increased activity of honokiol against SVR cells, honokiol is more effective in induction of apoptosis than magnolol.
In order to determine the mechanisms of activity of honokiol, we examined its effect on expression and phosphorylation of key signal transduction pathways. Activation of both MAP kinase and akt was inhibited, indicating that a mechanism of activity is likely upstream of these pathways and possibly at the level of src. Inhibition of Akt phosphorylation occurred at an earlier time point and at lower doses than inhibition of MAP kinase phosphorylation, indicating a preferential inhibition of PI-3 kinase signaling. Given that polyphenols, including curcumin and epicatechin gallate, both target proteasomes, we examined the effect of honokiol on the COP9 signalosome. Honokiol is an effective inhibitor of COP9 signalosome kinase activity.
In conclusion, we demonstrate the utility of the SVR bioassay in isolating an active principle in a natural product and characterize its mechanism of action. The active principle, honokiol has been previously described as a component of a Japanese herbal medicine, saiboku-to, and of the Chinese medicine houpo and has been shown to have anxiolytic properties in mice 21-23;41 42 . We have demonstrated for the first time that honokiol has potent antiangiogenic and antitumor properties in vitro, and systemically active against aggressive angiosarcoma in vivo. In addition, honokiol is well tolerated by the host animal in therapeutically beneficial doses, making it an attractive candidtate for further preclinical testing as an antineoplastic agent.