Metronomic Dosing of BH3 Mimetic Small Molecule Yields Robust Antiangiogenic and Antitumor Effects

Induction chemotherapy with taxanes, platinum-based compounds, and 5-fluorouracil is beneficial for head and neck cancer patients (1, 2), but the prolonged use of chemotherapeutic drugs is limited by their toxicity and by the development of resistance. The combined use of molecularly targeted agents with conventional therapies has been proposed more recently for management of patients with locally advanced head and neck squamous cell carcinomas (HNSCC; ref. 3). These types of combination therapies have shown promising results, but the survival of head and neck cancer patients has not changed dramatically (4). Improvements in the survival of these patients require mechanism-based therapeutic strategies that maximize the antitumor effect of drugs while limiting their toxicities.

Metronomic chemotherapy has been proposed as a mean to potentiate the antitumor effect of chemotherapeutic drugs and to overcome drug resistance (5–7). Several independent groups have shown benefits of the metronomic regimen in preclinical and clinical studies (8–10). Although much has been learned about the use of conventional chemotherapeutic drugs in metronomic regimen over the last 10 years, little is known about this type of regimen with molecularly targeted drugs.

A significant proportion of head and neck tumors express high levels of antiapoptotic Bcl-2 proteins (11, 12). Indeed, high Bcl-2 expression correlates directly with resistance to therapy with cisplatin (13), and is associated with poor clinical outcomes for head and neck cancer patients (14, 15). We have shown that Bcl-2 gene expression is significantly higher (approximately 60,000-fold) in the tumor-associated endothelial cells of patients with HNSCC, as compared with endothelial cells from the normal oral mucosa (16). Collectively, these studies provide strong rationale for the investigation of the antitumor and antiangiogenic effects of drugs targeted to the Bcl-2 pathway.

BH3-mimetic compounds derived from (-)-gossypol, a natural product from cotton plant (17), inhibit the survival function of Bcl-2 family members by stimulating Noxa and Puma (18). These compounds have shown antitumor effect as a single agent or in combination with standard chemoradiotherapy in various tumor models (19–21), and seem to help to overcome resistance to therapy (22–24). We have shown the potent antiangiogenic effect (25) and antitumor effect of a small molecule inhibitor of Bcl-2 in head and neck cancer models (26). However, we do not understand the impact of the regimen of administration of a BH3-mimetic drug (AT101) on tumor growth and angiogenesis. Here, we observed that a metronomic regimen (i.e., daily administration of low-dose AT101) has more potent antitumor and antiangiogenic effects than weekly administration without compromising the low systemic toxicity profile of the drug.

Primary human dermal microvascular endothelial cells (HDMEC; Lonza) were cultured in endothelial cell growth medium (EGM2-MV; Lonza). HNSCC cell lines UM-SCC-17B and UM-SCC-74A (gift from T. Carey, University of Michigan) were cultured in DMEM (Invitrogen) supplemented with 10% FBS, 200 mmol/L L-glutamine, penicillin, and streptomycin at 37°C with 5% CO₂. The identity of all tumor cell lines was confirmed by genotyping at the University of Michigan DNA sequencing core facility in May 2010. For in vitro studies, the small molecule inhibitor of Bcl-2 (AT101; ref. 27) and taxotere (TXT; LC laboratories) were dissolved in DMSO. For in vivo studies, AT101 was resuspended in carboxymethyl cellulose and sonicated for 30 minutes, whereas TXT was dissolved in 5% ethanol.

Sulforhodamine B (SRB) cytotoxicity assays were done as described (25). Briefly, cells were seeded at 2 × 10³ cells per well of 96-well plates, allowed to attach overnight, and treated with AT101 and/or taxotere for 72 to 96 hours. Cells were fixed with 10% trichloroacetic acid, stained with 0.4% SRB (Sigma-Aldrich) in 1% acetic acid, and plates were read in a microplate reader at 560 nm (GENios; Tecan). Test results were normalized against initial plating density and drug-free controls. Data were obtained from triplicate wells per condition and is representative of 3 independent experiments.

Droplets were formed by mixing HDMEC and 10% sucrose containing 0.05% puramatrix (BD Biosciences). Droplet was transferred to type I collagen (Angiotech BioMaterials) coated plates and covered with collagen containing 50 ng/mL rhVEGF₁₆₅ (R&D Systems). Treatment started as soon as sprout outgrowth was visible. Drug containing medium was changed every day in the metronomic group whereas in the bolus treatment group the complete dose was delivered in the first day. Data were obtained from triplicate wells per condition and is representative of 3 independent experiments.

Xenograft human tumors vascularized with human blood vessels were generated as described (28). Briefly, highly porous poly-(L-lactic) acid scaffolds seeded with 9 × 10⁵ HDMEC and 1 × 10⁵ tumor cells (UM-SCC-17B) were incubated for 30 minutes at 37°C. Male severe combined immunodeficient (SCID) mice (Charles River Laboratory) were anesthetized with ketamine and xylazine, and scaffolds were implanted subcutaneously in the dorsal region of each mouse. When tumor size reached 200 mm³, mice were randomized into 4 groups (n = 8): vehicle, taxotere, weekly AT101+taxotere (bolus), and daily AT101+taxotere (metronomic). At termination of the experiment, mice were euthanized, tumors were retrieved, fixed overnight in 10% buffered formalin at 4°C, and processed for histology. The histopathology was assessed by a trained pathologist unaware of treatment conditions. The mitotic index (29) was determined by counting mitotic figures in 5 high-power fields per tumor section. Mice were euthanized whenever the tumor volume reached 2,000 mm³ following the University of Michigan Guidelines for Use and Care of Animals. Tumor size was calculated using the formula: volume (mm³) = L × W²/2 (L, length; W, width).

Paraffin-embedded tissue sections were incubated in antigen retrieval solution (Dako) for 20 minutes at 90°C to 95°C. Tissues were exposed to 1:500 dilution of the anti-human factor VIII antibody (NeoMarkers) overnight at 4°C, as described (28). Blood vessels were counted in 6 random fields per tumor, from 8 tumors per group.

Time-to-failure data was analyzed using the Kaplan–Meier method and the log-rank test. Where multiple fields were measured, the average was taken for each mouse. Continuous measurements, such as tumor volume, average mitosis per mm² and average microvessel density were analyzed using nonparametric rank-based Wilcoxon tests and Kruskal–Wallis tests. The growth curve of tumor volume (mm³) was transformed by log₂ to improve linearity assumptions of residuals for longitudinal modeling. We refer to this transformation as log₂ tumor volume and can be interpreted as tumor doubling. A longitudinal mixed model was used to test the difference in growth patterns based on treatment group. Kaplan–Meier analysis, longitudinal analysis, and Wilcoxon tests were carried out using SAS v9.2 software. In vitro statistical analysis was carried out using ANOVA followed by Tukey's test (SigmaStat 2.0 software; SPSS). The synergism or additivity was calculated using the combinatorial index (CI), a mathematical and quantitative evaluation of a 2-drug pharmacologic interaction (30). Using CalcuSyn ver. 2.0 software (Biosoft), CI values were generated over a range of fractional cell kill levels (Fa) from 0.05 to 0.95 (5%–95% cell kill). Accordingly, a synergistic effect is represented by CI < 0.9; an additive effect by 0.9 ≤ CI ≤ 1.1; and absence of combinatorial effect CI > 1.1. Statistical significance was defined as P ≤ 0.05, unless otherwise specified.

To investigate the effect of metronomic (low dose, high frequency) versus bolus (high dose, low frequency) administration of a BH3-mimetic drug on tumor growth, we generated xenografts with humanized vasculature (31, 32). All experimental conditions included low-dose taxotere (5 mg/kg taxotere, once a week) to mimic the most probable clinical scenario for the use of the targeted drug in combination with a conventional chemotherapeutic drug. The maximum-tolerated dose (MTD) for BH3-mimetic was estimated to be around 40 mg/kg for 3 days (20) and for taxotere around 50 mg/kg weekly (33). In pilot experiments, we observed that oral administration of 105 mg/kg AT101 plus 5 mg/kg taxotere weekly caused 15% to 20% weight loss, and was considered to be the MTD for the combination therapies (data not shown). These studies also showed that weekly 70 mg/kg AT101 plus 5 mg/kg taxotere was well tolerated by mice (<15% weight loss).

The criterion to determine failure for the Kaplan–Meier analysis was a 5-fold increase in tumor volume as compared with baseline (i.e., immediately prior to start of treatment). These experiments were carried out with an aggressive squamous cell carcinoma (UM-SCC-17B), and therefore the time-to-failure was as low as 28 days (vehicle control) after transplantation of 100,000 tumor cells/mouse. The average time-to-failure was 30.6 days (vehicle), 33.3 days (taxotere), 37.3 days (weekly AT101+taxotere), and 48.5 days (daily AT101+taxotere). The distribution of time-to-failure was significantly different across the 4 experimental groups (P < 0.0001; Fig. 1A). Daily AT101+taxotere mediated a significant delay in time-to-failure as compared with all other experimental conditions (Fig. 1A). In pairwise comparisons, survival was prolonged by daily AT101+taxotere treatment as compared with weekly AT101+taxotere (P = 0.0290), to taxotere (P < 0.0001), or to vehicle control (P < 0.0001). Weekly AT101+taxotere did not improve survival when compared with vehicle control (P = 0.1542) or with taxotere alone (P = 0.6535). We observed an overall delay in tumor growth when taxotere and AT101 were used in combination, particularly with the daily AT101 regimen (Fig. 1B). Figure 1C depicts a detailed analysis of tumor volume after 30 days of treatment, the day when we had to begin to euthanize animals from the untreated control group. Tumor volume was reduced in the daily AT101+taxotere as compared with the weekly AT101+taxotere group at 30 days, but did not reach statistical significance (P = 0.0660). However, starting at day 31 of treatment and continuing thereafter, the tumor volume in mice that received daily AT101+taxotere was consistently lower (P ≤ 0.050) than the tumor volume in mice treated with weekly AT101+taxotere (Fig. 1B). Notably, we did not observe deaths attributed to drug toxicity with the regimens used here, and the average weight loss was consistently less than 15% (Fig. 1D). Moreover, no signs of toxicity were observed in vital organs including liver and lung (Supplementary Fig. S1).

In general, xenografts showed tumor cells with basophile cytoplasm, increased nuclear-to-cytoplasm diameter ratio, and marked cellular pleomorphism (Fig. 2A). In contrast, the metronomic AT101+taxotere group presented a distinct pattern, with a large proportion of clear cells (Fig. 2A). The tumor cells presented clear cytoplasm and smaller nucleus with higher condensation of chromatin. In addition, larger necrotic areas were observed (Fig. 2A). We observed a significant overall difference in mitotic index (Fig. 2B, P < 0.0001) and tumor microvessel density (Fig. 2C, P = 0.0003). Pairwise analysis revealed that the mitotic index of daily AT101+taxotere was lower than the mitotic index of each one of the other experimental conditions, as follows (P = 0.0047, compared with vehicle; P = 0.0044, compared with taxotere; and P = 0.0046, compared with weekly AT101+taxotere; Fig. 2B). We also observed that the microvessel density in the daily AT101+taxotere group was lower than in the vehicle control (P = 0.0047), taxotere only (P = 0.0046), and weekly AT101+taxotere (P = 0.0046; Fig. 2C). Moreover, qualitative assessment revealed that the daily AT101+taxotere regimen induced more tumor cell apoptosis, as compared with the other experimental conditions (Supplementary Fig. S2).

The IC₅₀ for AT101 was 0.80 μmol/L (HDMEC) and ranged from 2.13 to 2.54 μmol/L for head neck tumor cells (UM-SCC-17B, UM-SCC-74A; Fig. 3). The IC₅₀ for taxotere was 0.37 nmol/L (HDMEC), and ranged form 0.21 to 0.53 nmol/L for the tumor cells evaluated (Fig. 3). The IC₂₅ and IC₇₅ of these 2 drugs were also calculated in a separate set of experiments with narrower range of drug concentrations (data not shown).

To investigate whether AT101 sensitizes HDMEC and tumor cells to chemotherapy, we examined the cytotoxic effects of the IC₂₅, IC₅₀, IC₇₅ concentrations for each drug by itself or in combination. Additive effects were observed primarily when higher concentrations of AT101 were combined with taxotere for all cell types examined here (Fig. 4). Interestingly, the overall trends for the CI analysis were similar for endothelial cells and head and neck tumor cells (Fig. 4).

Next, we investigated the effect of treatment sequence on the cytotoxicity of these drugs in HDMEC and tumor cells. For each cell type, we either pretreated cells with one drug for 24 hours, or started treatment together with the 2 drugs. The drug concentration was set at the IC₅₀, and 2 time periods were evaluated to factor-in the overall time of treatment on the effects observed. We observed that starting the 2 drugs at the same time provided the highest cytotoxic effect for all cell types evaluated here (P < 0.01; Fig. 5). Propidium iodide staining followed by flow cytometry was carried out to observe the effects of different treatment sequence on apoptosis and cell cycle of UM-SCC-74A. The use of both drugs at the same time induced more apoptosis than any other experimental condition tested here (P < 0.05; Supplementary Fig. S3A). Cell-cycle analysis showed that drug sequence had no detectable effect on the proportion of UM-SCC-74A cells in S-phase (Supplementary Fig. S3B and S3C).

AT101 [(-) gossypol] has been developed and characterized as a functional inhibitor of Bcl-2 via binding to BH3-binding domain and interference with the interaction of Bcl-2 with other BH3-containing proapoptotic proteins (34, 35). As expected, Bcl-2 expression was not affected by AT101 treatment (Supplementary Fig. S4), confirming previous observations (23). To evaluate the antiangiogenic effect of AT101 in vitro, we created three dimensional (3-D) droplets of endothelial cells and plated them in collagen matrices. Endothelial cells were induced to sprout with VEGF, and 0.1 × IC₅₀ (0.08 μmol/L), IC₅₀ (0.8 μmol/L), 10 × IC₅₀ (8 μmol/L) AT101 were used. We either administered these dosages at once (bolus), or mimicked the metronomic regimen by dividing them in 4 daily administrations. The highest concentration of AT101 eliminated capillary sprouts by 4 days in both regimens (Fig. 6). For both 0.1 × IC₅₀ and IC₅₀ conditions, the daily administration of the drug led to more potent reduction in capillary sprouting than the administration of the same overall dosage at once (0.08 μmol/L daily vs. 0.08 μmol/L bolus; P < 0.001, 0.8 μmol/L daily vs. 0.8 μmol/L bolus; P = 0.023; Fig. 6B). Interestingly, the lowest concentration of AT101 used here (0.08 μmol/L AT101) induced a potent antiangiogenic effect (P < 0.001), despite the fact that this concentration is subapoptotic (Fig. 6B). These data corroborated our previous reports that Bcl-2 functions as a proangiogenic signaling molecule (36), in addition to its well-known effects on the regulation of endothelial cell survival (28).

Improvements in the survival of head and neck cancer patients will most likely come from the development and characterization of mechanism-based therapies that combine the potency of cytotoxic drugs with the specificity of molecularly targeted agents. Evidence clearly shows the role of Bcl-2 in the poor outcomes of patients with head and neck cancer (37). We have also shown that Bcl-2 plays a dominant role in head and neck tumor angiogenesis and tumor growth (16, 25, 31, 38). Therefore, targeting the Bcl-2 system is appealing because it has 2 complimentary effects mediated by the direct induction of tumor cell apoptosis and the selective disruption of tumor blood vessels (35).

The combination of the BH3-mimetic drug AT101 with taxotere, a microtube-binding drug, was selected for these experiments. Although taxotere is widely used in many tumor types including head and neck cancer, strong side effects such as myelosuppression have suggested that combination therapies that allow for a reduction of its dosage are highly desirable (1, 39, 40). Likewise, although single-drug therapy with BH3 mimetics partially inhibits tumor growth in some cancer models (41, 42), combination therapies tend to enhance the efficacy of the BH3 mimetics (43, 44). Notably, it has been recently reported that the combination of BH3 mimetic with taxotere inhibits tumor cell growth and affects expression of several genes involved in drug resistance, drug metabolism, DNA repair, cell cycle and oncogenes in a prostate tumor cell line (PC3; ref. 22). Here, we observed that the combination of AT101 with taxotere augments the antitumor cell and antiangiogenic effects of each one of these drugs used individually. Notably, best results were achieved with higher doses of AT101, a drug with minimal systemic cytotoxicities, combined with lower doses of taxotere. Collectively, these data unveil a drug combination that is highly effective in vivo while showing modest systemic cytotoxic effects.

Metronomic therapy of DNA-damaging agents or microtubule inhibitors has shown more efficiency than MTD therapy in randomized phase III clinical trials (45, 46). Metronomic therapy of certain drugs induces sustained suppression of circulating endothelial progenitor cells, whereas MTD schedules seem to allow for rebound vessel repair responses mediated primarily by bone marrow cell recruitment (43). Here, we examined the approach of using the metronomic regimen for the molecularly targeted drug in combination with conventional regimen for chemotherapy drug. Remarkably, combination of metronomic AT101 with taxotere showed potent inhibition of tumor growth, and prolongation of the survival of mice with established tumors (as compared with bolus treatment of AT101 and taxotere) despite the fact that the total amount of drug administered in both conditions was exactly the same. We believe that this difference in response was mediated by the potentiation of both the antiangiogenic and the tumor cell effect of AT101 when used in a metronomic regimen.

It is well known that vessel rebound rapidly occurs during the breaks in treatment (47). We postulate that by keeping constant the antiangiogenic pressure, achieved even when very low concentrations of BH3-mimetic are used (25, 36), one does not allow for the vascular rebound and therefore the overall antitumor growth effect is potentiated. This hypothesis is currently under investigation in our laboratory. In addition, based on our results we speculate that low-dose BH3-mimetic prevents the development of resistance of tumor cells to therapy, by interfering with the prosurvival effect of Bcl-2 family proteins. A closer look at the histopathology of tumors treated with metronomic AT101 combined with taxotere supports this hypothesis. We observed that this treatment regimen results in lower mitotic index as compared with all other treatment groups of this study. Notably, recent studies showed that the mitotic index is a potent indicator of patient survival with melanoma or breast carcinoma (48, 49). Moreover, in the metronomic AT101 group the tumor cells displayed a clear cytoplasm, suggesting glycogen accumulation compatible with chronic hypoxia and glucose deprivation. This might be related to the fact that this treatment regimen resulted in lower microvessel density.

In summary, we showed that metronomic administration of the BH3-mimetic drug AT101 in combination with low-dose taxotere improves the survival of mice-bearing HNSCC xenografts. These results were correlated with lower mitotic index and tumor microvessel density. Notably, we observed that subapoptotic concentrations of AT101 potently inhibited the angiogenic potential of endothelial cells. The fact that AT101 can be administered orally raises the possibility that patients could take low does of this drug in both therapeutic and chemopreventive regimen. Collectively, this work unveils the antitumor potential of metronomic administration of a molecularly targeted drug.

S. Wang owns stocks, stock options, and serves as a consultant for Ascenta Therapeutics that has licensed technologies related to AT101 from the University of Michigan.

The authors thank Chris Strayhorn for assistance with the histology and Naoki Ashimori for valuable suggestions.

The work was supported by the grant P50-CA97248 (University of Michigan Head & Neck SPORE) from the NIH/NCI (S. Wang and J.E. Nör) and grant U19-CA113317 (S. Wang) from the NIH/NCI; and grants R01-DE14601, R01-DE15948, R01-DE16586, and R21-DE19279 from the NIH/NIDCR (J.E. Nör).

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.