AMG 102, A Fully Human Anti-Hepatocyte Growth Factor/Scatter Factor Neutralizing Antibody, Enhances the Efficacy of Temozolomide or Docetaxel in U-87 MG Cells and Xenografts

Materials and Methods

Cell lines and reagents. Recombinant human d5-HGF/SF and AMG 102 expressed by Chinese hamster ovary cells were prepared as described (24). U-87 MG cells were obtained from the American Type Culture Collection. Docetaxel was obtained from Aventis Pharmaceuticals and resuspended in DMSO for in vitro cell assays. For the in vivo assays, docetaxel was resuspended in the manufacturer-provided diluent and adjusted to the final concentration used before injection with PBS. Temozolomide was purified in house from capsule form and resuspended in a 20% captisol solution (pH 7.5).

U-87 MG growth/survival. The U-87 MG cell proliferation assay was done as described (24) with modifications. Two thousand cells per well in MEM medium + 1% fetal bovine serum was used for the docetaxel studies, and 3,000 cells per well in MEM medium + 1% fetal bovine serum was used for the temozolomide studies. The absorbance at 540 nm was read on a microtiter plate reader (SpectraMax PLUS, Molecular Devices). To calculate the percentage of control, the value of treated cells was compared with cells treated with an isotype control antibody or no antibody treatment. Data were generated as either triplicate or quadruplicate points and averaged. Growth data were evaluated with repeated-measures ANOVA using StatView followed by Bonferroni/Dunn to calculate P values.

Caspase-3/7 activity. U87-MG cells were plated at 10,000 per well in Costar 96-well black assay plates into MEM + 1% fetal bovine serum and treated 24 h later. Twenty-four hours later, the Apo-ONE reagent was prepared according to the manufacturer's instructions, added to the plate, and read using the Victor2 fluorescent microtiter plate reader at excitation of 485 nm and emission at 530 nm. Data points were generated in triplicate and averaged.

Western blot analysis. Cells were treated with huIgG₂ control antibody or AMG 102 alone or in combination with docetaxel at 100 nmol/L for 24 h, spun down, and resuspended in 2× Tris-glycine SDS sample buffer (Invitrogen) with 5% of β-mercaptoethanol (Sigma). Lysates were sheared through an 18.5-gauge needle four times and stored at −80°C. For Western blots, lysate was loaded on either 4% to 20% or 10% Tris-glycine gels (Invitrogen) and transferred to nitrocellulose membranes (Schleicher and Schnell). Blots were probed for 20 h at 4°C with a 1:500 dilution of a rabbit polyclonal anti–poly(ADP)ribose polymerase (PARP) antibody (Cell Signaling). Densitometry analysis was done using Labworks software from UVP, Inc. Values of cleaved PARP were normalized to actin and expressed as fold increase compared with huIgG₂ control antibody.

Fluorescence-activated cell sorting analysis. Cells were treated for 24 h with the indicated agents and harvested in 1 mL of dissociation buffer (Invitrogen). Cells were then suspended in cytofix/cytoperm solution (BD Biosciences) for 30 min. Cells were washed once in perm/wash buffer (BD Biosciences), transferred to a 96-well U-bottomed plate, and stained with a 1:100 dilution of cleaved caspase-7 antibody or FITC-conjugated cleaved PARP polyclonal antibody (Cell Signaling). After washing once with perm/wash buffer, Alexa 568–conjugated goat anti-rabbit secondary antibody (Invitrogen) was used for the caspase-7 antibody–stained samples. All samples were washed once with perm/wash buffer before transfer to fluorescence-activated cell sorting (FACS) tubes and run on the FACScan instrument (BD Biosciences). Cells that stained positive in the FL-2 channel for cleaved caspase-7 or the FL-1 channel for cleaved PARP were identified and counted to determine the percentage of positive cells compared with total number of cells analyzed.

Tumor growth and measurements. Xenograft studies were done as described (24) with modifications. Eight-week-old female CD1 nude mice (Charles River Laboratories) were injected s.c. with 5 × 10⁶ U-87 MG cells. Tumors were allowed to grow to ∼200 mm³ and then the mice were randomized into individual treatment groups (n = 10/group). Tumor measurements and body weights were recorded twice per week. Tumor volume was calculated as length × width × height in mm³. Results are expressed as mean ± SE. All animal experiments were done in a blinded fashion where the person measuring the tumor volumes was unaware of the treatment groups. These studies were conducted in accordance with federal animal care guidelines and were preapproved by the Amgen Institutional Animal Care and Use Committee. Statistical analysis was performed using repeated-measures ANOVA using StatView followed by Scheffé's post hoc test to determine P values.

Results

Temozolomide can effectively inhibit U-87 MG growth in vitro. U-87 MG cells were tested for a single-agent dose response to temozolomide. Temozolomide undergoes spontaneous hydrolysis at physiologic pH into its active metabolite 5-(3-methyl triazen-1-yl)imidazole-4-carboxamide (27), which is then thought to mediate cytotoxicity through DNA methylation (27, 28). Cell response was analyzed using sulforhodamine B (SRB), a total protein dye stain that assesses the total number of surviving cells (29). This method has a linear correlation to cell number and provides a measure of cell growth and survival. Duplicate data points were used to generate a four-parameter curve fit to determine a single agent IC₅₀ value. After 8 days of exposure in vitro, temozolomide was determined to have an IC₅₀ of 21.2 ± 12.5 μmol/L (n = 5) for the U-87 MG cell line. Data from a representative experiment are shown in Fig. 1A .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Effects of temozolomide on U-87 MG cell growth. A, the percentage inhibition of proliferation/survival of U-87 MG cells cultured in the presence of temozolomide (TMZ) was measured over a 12-point dose curve by SRB dye staining. B, HGF/SF can protect cells from the antigrowth effects of temozolomide. Cells were preincubated with HGF/SF for 24 h before addition of 27.8 μmol/L temozolomide. C, anti-HGF/SF–neutralizing antibody enhances temozolomide-induced antigrowth effects. U-87 MG cells were treated with 83 nmol/L of IgG₂ alone, 12.5 μmol/L temozolomide + 83 nmol/L IgG₂, 83 nmol/L (12.5 μg/mL) AMG 102 alone, or the combination of AMG 102 and temozolomide. At least four independent points were used to generate the average percentage of control graphed. Human IgG₂ was used as a control treatment for the antibody. *, P ≤ 0.0001, statistically significant compared with treatment with 83 nmol/L IgG₂ alone, 83 nmol/L AMG 102 alone, or 83 nmol/L IgG₂ + 12.5 μmol/L temozolomide. Percentage inhibition, percentage of control calculations, and statistical analysis for all figures were generated as described in Materials and Methods.

Addition of HGF/SF can protect U-87 MG cells against temozolomide-induced cell death. Previous studies have shown that the HGF/SF/SF:c-Met axis drives U-87 MG proliferation (19, 24). We confirmed the presence of c-Met on the cell surface by FACS analysis using a fluorescently labeled polyclonal anti–c-Met antibody (data not shown). An HGF/SF ELISA was done on cell culture medium supernatants to quantify the levels of HGF/SF secreted in the growth medium 5 days after plating 1 × 10⁶ cells in a 10-cm dish. Approximately 340 ± 30 pg/mL of HGF/SF was detected in the medium of U-87 MG cells grown for 5 days postplating.

To simulate a paracrine model of c-Met stimulation, exogenous HGF/SF was added to the medium to supplement the HGF/SF:c-Met autocrine loop in the presence of a suboptimal concentration of temozolomide. HGF/SF was added for 24 h to the medium in low serum conditions (1% fetal bovine serum) before addition of temozolomide. Addition of 75 ng/mL of HGF/SF to the medium was able to completely protect against the effects of a 27.8 μmol/L dose of temozolomide when assayed by the SRB method (Fig. 1B). This suggests that HGF/SF signaling can provide some measure of resistance to temozolomide treatment, and despite the secreted HGF/SF from U-87 MG cells, functional c-Met receptor signaling is not saturated in these cells under in vitro culture conditions. At higher concentrations of temozolomide (>100 μmol/L), however, additional HGF/SF was not able to protect the cells (data not shown).

AMG 102 enhances the antigrowth effects of temozolomide in vitro. Because U-87 MG cells are sensitive to both temozolomide and AMG 102 as single agents, we measured the effects of the combination of temozolomide and AMG 102 to assess any increased activity. In a preliminary set of experiments, a dose range of temozolomide was combined with a dose range of AMG 102 and cell growth was measured using the SRB method (data not shown). Each single agent alone was able to inhibit U-87 MG cells effectively at high concentrations [>100 μmol/L for temozolomide, >333 nmol/L (50 μg/mL) for AMG 102]. However, optimal antigrowth activity was also observed using a combination of suboptimal doses of the two agents, 83 nmol/L (12.5 μg/mL) AMG 102 and 12.5 μmol/L temozolomide (Fig. 1C). Cell growth/survival after treatment with the combination was significantly impaired when compared with control cells or either single agent alone (P < 0.0001 for all groups).

AMG 102 enhances the antigrowth effects of temozolomide in vivo. U-87 MG tumors are sensitive to HGF/SF-neutralizing antibodies when grown s.c. in athymic nude mice, demonstrating that the growth of this model is driven by the presence of an HGF/SF/SF:c-Met autocrine loop (23, 24). AMG 102 (30 μg) dosed twice per week can completely inhibit tumor growth in this model as a single agent. This tumor model is also sensitive to temozolomide treatment, as >7 mg/kg of temozolomide dosed once per week completely inhibits tumor growth (data not shown). To assess the combination effects of AMG 102 and temozolomide, suboptimal doses of 3 μg twice per week of AMG 102 and 5 mg/kg once per week of temozolomide were selected and tumor growth was evaluated. Each drug was dosed alone and in combination in an established disease model, when tumors reached an average size of 200 mm³ (day 14; Fig. 2 ). Treatment with 3 μg of AMG 102 did not result in statistically significant tumor inhibition in this experiment (33.7% inhibition; P > 0.05). Treatment with 5 mg/kg temozolomide resulted in 76.2% inhibition of tumor growth on day 29 that was statistically significant when compared with control (P < 0.0001). However, the combination of temozolomide and AMG 102 resulted in a 93% inhibition of tumor growth that was statistically significant compared with treatment with either single agent alone (P < 0.0001). The combination of the drugs did not seem to negatively affect the efficacy of either agent, suggesting that they do not interfere with therapeutic efficacy when dosed together. No treatment effect on animal body weight was observed either as single agents or in combination throughout the study, suggesting that the combination is not overtly toxic (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Combination of AMG 102 with temozolomide results in enhanced tumor inhibition in vivo. U-87 MG cells were injected s.c. into female athymic nude mice (n = 10/group). Treatment with 3 μg AMG 102 twice per week, 5 mg/kg temozolomide once per week, or the combination was initiated after tumors reached an average size of 200 mm³. The combination of AMG 102 and temozolomide significantly increased inhibition of tumor growth compared with treatment with either single agent alone (#, P < 0.0001). *, P < 0.0001 compared with vehicle and IgG₂ control antibody. Points, mean; bars, SE.

Docetaxel can effectively inhibit U-87 MG growth in vitro as a single agent. Docetaxel is a tubulin-stabilizing agent currently used as a therapy in metastatic breast cancer. This agent is not used clinically in GBM; however, in vitro and in vivo models for testing anti-HGF/SF–neutralizing antibodies are limited due to the specificity of the antibodies for human HGF/SF. We have not yet identified a human breast cancer cell line that contains a functional HGF/SF:c-Met autocrine loop appropriate for testing anti-HGF/SF agents in mouse xenografts models. Therefore, we used U-87 MG cells to determine the dose response to docetaxel as a single agent. At high concentrations (>100 nmol/L), docetaxel could completely inhibit U-87 MG cell growth and displayed a single-agent IC₅₀ value of 5.8 ± 0.4 nmol/L (n = 4). A representative IC₅₀ curve is shown in Fig. 3A .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

A, docetaxel inhibits U-87 MG cell growth. The percentage inhibition of proliferation/survival of U-87 MG cells cultured in the presence of docetaxel was measured over a nine-point dose curve by SRB dye staining. B, AMG 102 enhances docetaxel-induced antigrowth effects. U-87 MG cells were treated with 2 nmol/L docetaxel, 83 nmol/L (12.5 μg/mL) AMG 102, or the combination of antibody with docetaxel. At least four independent points were used to generate the average percentage of control. IgG₂ was used as a control treatment for the antibody. *, P ≤ 0.0001, statistically significant compared with treatment with control IgG alone, AMG 102 alone, or control IgG + 2 nmol/L docetaxel. Percentage inhibition, percentage of control calculations, and statistical analysis for all figures were generated as described in Materials and Methods.

AMG 102 enhances docetaxel antigrowth effects in vitro. To test the combination effects of docetaxel and anti-HGF/SF–neutralizing antibodies, a dose range of docetaxel was combined with a dose range of AMG 102 in a series of preliminary experiments. Suboptimal concentrations of 83 nmol/L (12.5 μg/mL) AMG 102 and 2 nmol/L of docetaxel treatment of U-87 MG cell growth over 6 days was chosen for further study. As shown in Fig. 3B, docetaxel in combination with AMG 102 significantly inhibited cell growth/survival compared with either docetaxel in combination with IgG₂ or AMG 102 alone (P < 0.0001 for all groups).

AMG 102 increases docetaxel-induced apoptosis. To investigate whether loss of cell viability was due to increased apoptosis, cells were examined for activation of caspase-3/7 activity. Caspase-3 and caspase-7 are key executioners in the apoptotic process, cleaving downstream enzymes such as PARP, as well as other caspase enzymes. Increased caspase-3/7 activity was observed in a dose-dependent manner in response to docetaxel, with an IC₅₀ of 11.0 nmol/L (Fig. 4A ) similar to the growth/survival IC₅₀ as assessed by SRB (see Fig. 3A). The addition of 83 nmol/L (12.5 μg/mL) AMG 102 increased the extent of caspase-3/7 activity across a dose range of docetaxel (Fig. 4B), and slightly reduced the IC₅₀ of docetaxel. Similar effects were observed at 166 nmol/L (25 μg/mL) of AMG 102 across a dose range of docetaxel (data not shown). AMG 102 alone did not induce measurable caspase-3/7 activity when added to U-87 MG cells in vitro across a large dose range (up to 1.3 μmol/L; data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

HGF-neutralizing antibody increases docetaxel-induced caspase activity in U-87 MG cells in vitro. A, U-87 MG cells were tested for caspase-3/7 activity in response to a wide concentration range of docetaxel alone. Increasing caspase-3/7 activity is observed in response to increasing concentrations of docetaxel. B, U-87 MG cells were tested for caspase-3/7 activity in response to a wide concentration of docetaxel in combination with 83 nmol/L (12.5 μg/mL) AMG 102. AMG 102 is able to increase the total caspase-3/7 activity generated by docetaxel. Caspase-3/7 activity was measured using an enzymatic activity kit as described in Materials and Methods. Each curve was generated using triplicate data points. FU, fluorescence unit.

To confirm the increase in apoptotic activity, caspase-3/7 substrates were examined for increased cleavage in response to combination treatment. PARP is a key downstream substrate of caspase activation. Lysates from U-87 MG cells treated for 24 h with IgG₂, AMG 102, docetaxel alone, or the combinations were analyzed by immunoblot for PARP cleavage (Fig. 5A ). Treatment with docetaxel (in combination with IgG₂) caused a detectable increase in PARP cleavage. When 50 or 100 nmol/L docetaxel was combined with 83 nmol/L (12.5 μg/mL) AMG 102, however, PARP cleavage was enhanced compared with docetaxel alone (Fig. 5A). We have previously noted that high doses of anti-HGF antibodies can lead to a small but significant increase in tumor cell apoptosis in the U-87 MG model in vivo (24). At a higher concentration of 333 nmol/L (50 μg/mL) AMG 102 in vitro, there was a small but detectable increase in PARP cleavage, which was enhanced by 5-fold when combined with 100 nmol/L docetaxel (Fig. 5B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Neutralizing anti-HGF/SF antibody increases docetaxel-induced apoptotic markers. A and B, cleaved PARP was detected by immunoblot of cell lysates after treatment with 83 nmol/L (12.5 μg/mL) of IgG₂ control antibody or AMG 102, alone or in combination with 100 or 50 nmol/L docetaxel. Actin was probed as a loading control (bottom blot). Densitometry was used to estimate the fold increase of cleaved PARP compared with IgG₂ control antibody–treated cells (bottom). C and D, U-87 MG cells were treated with 83 nmol/L (12.5 μg/mL) IgG₂ alone (top left), 83 nmol/L (12.5 μg/mL) AMG 102 alone (top right), 100 nmol/L of docetaxel alone (bottom left), or the combination of docetaxel and AMG 102 (bottom right), stained with an antibody that specifically recognizes cleaved caspase-7 (C) or cleaved PARP (D) and analyzed by FACS analysis.

The observation of increased caspase-3/7 activity in the combination-treated samples (Fig. 4B) could be attributed to either increased caspase activity in the same subpopulation of cells or to the generation of additional cells entering apoptosis. FACS analysis was used to quantify the number of cells displaying apoptotic markers after single agent or combination treatments. Cells were stained with antibodies specific for cleaved caspase-7 (Fig. 5C) or cleaved PARP (Fig. 5D) and analyzed by FACS. Consistent with results observed in the caspase-3/7 assay, treatment with AMG 102 alone was not able to induce measurable cleavage of either caspase-7 or PARP, indicating that caspase-3/7 was not substantially activated. However, the combination of AMG 102 with docetaxel increased the percentage of cells that stained for caspase-7 cleavage from 2.74% with docetaxel treatment alone to 5.50% when treated with the combination (Fig. 5C). Similarly, the cells staining for PARP cleavage increased from 2.58% when treated with docetaxel alone to 6.13% when treated in combination (Fig. 5D). These data suggest that the increased antigrowth/survival effects observed with combination treatment of docetaxel and AMG 102 are at least in part due to an increase in the proportion of cells undergoing apoptosis.

AMG 102 enhances docetaxel antigrowth effects in vivo. Docetaxel can completely inhibit tumor growth in the U-87 MG xenograft model as a single agent when used at 20 mg/kg once per week (data not shown). To assess combination effects of AMG 102 and docetaxel, suboptimal doses of 3 μg twice per week of AMG 102 or 10 mg/kg once per week of docetaxel were selected for use in an established disease xenograft model. Each agent was dosed either alone or in combination once tumors had reached an average size of 200 mm³ (14 days after cell implantation). Treatment with docetaxel or AMG 102 as single agents significantly inhibited tumor growth compared with control (40.5% and 55.8% inhibition, respectively, P < 0.05 on day 27). The combination of AMG 102 and docetaxel had a significant additive inhibitory effect on tumor growth compared with treatment with either single agent alone as shown in Fig. 6 (84.7% inhibition of control, P < 0.015 day 30). Furthermore, the combination of drugs showed no evidence of interference in therapeutic efficacy or overt toxicity, as might be indicated by body weight loss, of any of the test groups in the study (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Increased inhibition of U-87 MG tumor growth with AMG 102 in combination with docetaxel. Cells were injected s.c. into female athymic nude mice, and treatment with 3 μg AMG 102 twice per week, 10 mg/kg docetaxel once per week, or the combination was initiated once tumors reached an average size of 200 mm³. The combination of docetaxel with AMG 102 significantly increased efficacy compared with either single agent alone (#, P < 0.015; n = 10/group). Points, mean; bars, SE. *, P < 0.05, compared with vehicle and isotype control antibody on day 27.

Discussion

In this study, we show that AMG 102, a fully human anti-HGF/SF–neutralizing antibody, can enhance the activity of either of the chemotherapeutic agents temozolomide or docetaxel in vitro and in vivo. Combined treatment at suboptimal doses led to greater efficacy compared with treatment with single agents for either cell growth inhibition in vitro or when administered to animals to inhibit xenograft tumor growth. There was no evidence of interference between anti-HGF/SF–neutralizing antibodies and either docetaxel or temozolomide. In the case of docetaxel, the combined treatment with an anti-HGF/SF–neutralizing antibody led to increased apoptosis in vitro and may represent the mechanism of increased efficacy observed in the xenograft studies.

In an attempt to confirm that apoptosis was the in vivo mechanism of action of the combination effect of docetaxel and AMG 102, a short-term study was done to test for evidence of apoptosis. Unfortunately, no consistent trend supporting our hypothesis of increased apoptosis could be observed (data not shown). The failure to translate the in vitro results to the in vivo setting could be due to both the transient characteristics of the apoptotic cells as well as the extensive manipulation required to dissociate the tumor into a single-cell suspension for analysis. Further studies, although challenging, may be warranted.

The IC₅₀ of temozolomide in our study for the U-87 MG cell line averaged 21.2 μmol/L, and we observed statistically significant combined effects with AMG 102 at a lower concentration of 12.5 μmol/L. These levels are within the observed plasma concentration range for temozolomide in patients; in a study assessing plasma concentrations with continuous dosing at 75 to 200 mg/m²/d, calculated plasma concentrations ranged from 0.10 to 13.99 μg/mL (0.52-72.1 μmol/L; ref. 30). However, the experimental concentrations of temozolomide used in this study exceed the observed clinical cerebrospinal fluid concentrations reported of 0.16 to 9.93 μg/mL (0.82-9.94 μmol/L; ref. 30). Although it is possible that longer-term exposure at lower concentrations would cause some effect in vitro, overall, we do not know how U-87 MG sensitivity to temozolomide translates to the response observed in patients.

A significant finding of this study was that exogenous HGF/SF added to the medium was able to protect cells from temozolomide-induced growth inhibition. HGF/SF, through its receptor c-Met, can activate a number of downstream signaling molecules including the Ras/Erk and phosphatidylinositol 3-kinase/Akt pathways. Previous reports using other glioma cell lines, such as U373, have also shown limited cytoprotective action of HGF/SF against agents such as cisplatin, camptothecin, Adriamycin, and Taxol as well as γ irradiation (31, 32). This protection was abrogated by phosphatidylinositol 3-kinase inhibitors but not by mitogen-activated protein kinase or protein kinase C pathway inhibitors, suggesting a phosphatidylinositol 3-kinase/Akt–dependent mechanism (33). That additional HGF/SF in our study can increase this protective effect suggests that the c-Met receptors in U-87 MG cells are not saturated by the autocrine loop expressed by these cells in vitro; indeed, increased phosphorylation of c-Met is observed with HGF/SF addition to the medium in culture (data not shown). We hypothesize that this increased signaling is responsible for the increased cytoprotective effects versus temozolomide treatment.

In contrast to temozolomide, exogenous HGF/SF was not able to protect against docetaxel in U-87 MG cells (data not shown). However, there was clearly an additive effect in growth inhibition in vitro and in vivo when combined with anti-HGF/SF–neutralizing antibodies. These data may emphasize a fundamental difference in mechanism between these two cytotoxics; docetaxel triggers apoptosis after mitotic arrest (reviewed in ref. 33), and additional HGF/SF may not be able to overcome this block before drug-induced apoptosis. However, apoptosis caused by docetaxel was enhanced with anti-HGF/SF antibody, suggesting that the autocrine HGF/SF produced by U-87 MG cells does have measurable cytoprotective effects against docetaxel. It is interesting that additive effects are observed in a growth inhibition assay, despite the action of these two drugs on distinct phases of the cell cycle. This may be due to the ability of each agent to affect a distinct population of the cells resulting in an overall additive effect upon combination. Alternatively, signaling pathways such as the phosphatidylinositol 3-kinase/Akt pathway that is activated by the active HGF/SF:c-Met autocrine loop may offer some cytoprotective effect that is blocked upon addition of AMG 102. Taken together, these data imply that there may be utility for an anti-HGF/SF–neutralizing therapeutic in addressing disease that is resistant to chemotherapy treatment alone, and these data provide evidence that the neutralization of HGF/SF using monoclonal antibodies might lead to sensitization toward chemotherapeutic drugs.

Despite the strong evidence implicating the HGF/SF:c-Met axis in GBM, the blood-brain barrier represents a significant hurdle for antibody therapies in treatment of this disease. However, it has been suggested that there is a significant difference between normal and tumor vasculature, which is known to lack tight junctions and be leaky compared with normal host vasculature, potentially permitting drug access to the tumor. For example, recent preclinical evidence suggests that systemic administration of an anti-HGF/SF antibody therapy may be a viable approach in treating tumors of the central nervous system. In a mouse intracranial U-87 MG tumor model, i.p. dosing of a murine anti-HGF/SF monoclonal antibody was able to dramatically prolong survival (25). Other antibody therapies have also shown promising results both preclinically and clinically. For example, anti–vascular endothelial growth factor antibodies can prolong animal survival when given systemically in an intracranial G55 glioma model, although the tumor showed significant adaptation in co-opting the host vasculature posttreatment (34). In patients, a phase II trial with bevicizumab, a humanized anti–vascular endothelial growth factor antibody, in combination with irinotecan led to a radiographic response rate of 63% (20 of 32) in glioma patients with one complete response (35) and resulted in improved progression-free survival compared with irinotecan alone. Other published reports suggest that radiation therapy can open the blood-tumor barrier from ∼1 week to 1 month after completion of treatment (36). As radiation is part of the current standard of care for advanced glioblastoma, this suggests that anti-HGF/SF–neutralizing antibodies may be able to show efficacy if added to a regimen that includes radiation therapy. However, the size limitation of these blood-tumor barrier openings has not been fully characterized and may not accommodate a large molecule therapeutic.

Finally, this study is restricted to the U-87 MG model of glioblastoma due to its sensitivity to anti-human HGF/SF–neutralizing antibodies. Stromal effects from the host cannot be assessed because the mouse HGF/SF does not activate the human receptor and AMG 102 does not recognize or neutralize mouse HGF. However, in human disease, HGF/SF derived from the stroma could be playing a significant role in disease progression via paracrine activation of c-Met in tumors and/or the vasculature (37). Studies have been published suggesting a paracrine role for HGF/SF in multiple cancer types, including cancers of the breast (37), lung (38), prostate (39), etc. (1, 2). In many of these cases, HGF/SF expression correlates with poor prognosis for disease outcome. In these patient populations, there may exist an opportunity for an anti-HGF/SF therapy to show efficacy. Therefore, the application of AMG 102 in the clinic may be more widely applied to cancers that express the c-Met receptor. This report suggests that in cases where docetaxel, temozolomide, or similar mechanistic agents are used, combination therapy may further enhance efficacy.