Nab-paclitaxel Efficacy in the Orthotopic Model of Human Breast Cancer Is Significantly Enhanced By Concurrent Anti-Vascular Endothelial Growth Factor A Therapy

Abstract

Nab-paclitaxel is an albumin-bound 130-nm particle form of paclitaxel that has shown an improved efficacy in experimental tumor models and clinical studies compared with solvent-based paclitaxel. Anti-vascular endothelial growth factor A (VEGF-A) antibody bevacizumab is known to enhance antitumor activity of cytotoxic drugs. This study evaluated the effects of combined nab-paclitaxel and bevacizumab therapy on growth and metastatic spread of orthotopic breast tumors. Cytotoxic and clonogenic assays measured VEGF-A-dependent modulation of nabpaclitaxel toxicity on cultured tumor cells. Antitumor effects were assessed in mice with luciferase-tagged, wellestablished MDA-MB-231 tumors (250–310 mm³) treated with one, two, or three cycles of nab-paclitaxel (10 mg/kg, daily for five consecutive days), bevacizumab (2–8 mg/kg, twice a week), or with combination of both drugs. VEGF-A protected MDA-MB-231 cells against nab-paclitaxel cytotoxicity, whereas bevacizumab sensitized cells to the effect of the drug. Combined bevacizumab and nab-paclitaxel treatment synergistically inhibited tumor growth and metastasis resulting in up to 40% of complete regressions of well-established tumors. This therapy also decreased the incidence of lymphatic and pulmonary metastases by 60% and 100%, respectively. The significant increase in the cure of tumor-bearing mice in the nab-paclitaxel/bevacizumab combined group compared with mice treated with single drugs strongly advocates for implementing such strategy in clinics.

Introduction

Paclitaxel is a potent cytotoxic agent that is widely used against various refractory and metastatic malignancies [1,2]. Paclitaxel induces the assembly of tubulin subunits and prevents microtubule depolymerization, thereby disrupting normal microtubule reorganization, leading to G₂/M cell cycle arrest [3]. In addition, paclitaxel enhances apoptosis of cancer cells by promoting the phosphorylation and down-regulation of prosurvival protein bcl-2 [4]. Recently, the clinical use of paclitaxel has been significantly improved by formulating this drug in Cremophor-free, albumin-bound, 130-nm particles named nab-paclitaxel (paclitaxel protein-bound particles for injectable suspension or nab-paclitaxel also named Abraxane; Abraxis BioScience, Inc., Los Angeles, CA). Both clinical [2,5] and experimental studies demonstrated several advantages of nab-paclitaxel compared to conventional solvent-based paclitaxel (Taxol; Bristol-Myers Squibb Co., Princeton, NJ) [6,7] and docetaxel (Taxotere; Sanofi Aventis, Bridgewater, NJ). The main advantages include linear pharmacokinetics, high tumor retention, improved antitumor efficacy, and reduced toxicity [8] because of the elimination of Cremaphor [2]. However, the tumor response even to the improved formulation of paclitaxel is typically only 30% to 35% [7].

Several mechanisms may account for resistance to nab-paclitaxel treatment in some tumors, and targeting these mechanisms would greatly enhance antitumor efficacy. One mechanism that might increase tumor resistance to chemotherapy is increased expression of vascular endothelial growth factor A (VEGF-A) that is primarily responsible for the vascularization of solid tumors [9,10]. Although some studies have reported inhibition of VEGF-A by paclitaxel [11] and docetaxel [12,13] suggesting that these drugs are antiangiogenic in vivo [14,15], the overwhelming body of evidence indicates induction rather than suppression of VEGF-A after cytotoxic treatments. Chemodrugs that have been shown to induce VEGF-A expression include paclitaxel [16], docetaxel [16], carboplatin [17], cisplatin [18], 5-fluorouracil [19], dacarbazine [20], and anthracyclines [21]. Chemotherapy-induced VEGF-A production is possibly mediated by mitogen-activated protein kinase kinase/extracellular regulated kinase pathway [16], nuclear factor-κB [22], and phosphatidylinositol 3 kinase/AKT pathways [23,24] that are typically activated in response to stress in both tumor [18,19] and endothelial cells [17]. In vivo, VEGF-A production and subsequent angiogenesis can also be augmented by chemotherapyassociated hypoxia [18,25] that results from tumor cell death and vascular collapse in the drug-affected areas. In addition, docetaxel-induced, hypoxia-inducible factor 1 alpha-independent increase in VEGF-A has been also reported in patients with chemoresistant breast carcinoma [26].

VEGF-A induced by chemotherapy can promote survival of endothelial [12] and tumor cells [27], thus protecting both neoplastic and stroma cells from cytotoxic death while supporting angiogenesis in vivo. Thus, activation of proangiogenic programs by low concentrations of chemotherapeutic drugs might be responsible, in part, for tumor resistance to chemotherapy and tumor regrowth after initial drug-mediated suppression. This concept is supported by the recently recognized autocrine-positive VEGF-A·VEGF receptors loop detected in malignant cells of the lung [28], breast [28], colon [29], and renal [30] tumors. Coexpression of VEGF-A and its tyrosine kinase [29] or nontyrosine kinase [31] cognate receptors in neoplastic cells enhances tumor motility [32] and plays a major role in tumor cell survival under adverse conditions. Taken together, these observations lend to the idea that antitumor and antimetastatic effects of chemotherapy can be both potentiated and sustained by concurrent administration of agents that disrupt VEGF-A signaling. Not surprisingly, the combination of cytotoxic drugs including paclitaxel with VEGF-A targeting agents reportedly improved tumor response to chemotherapy [33,34] and increased patient survival [35].

An additional reason for combined therapy is the need to decrease high oncotic and intratumoral pressure that prevent tumor penetration by chemoagents, including that of nab-paclitaxel [36]. Sequestration of VEGF-A by bevacizumab, a specific VEGF-A-neutralizing monoclonal antibody [10,37] (Avastin; Genentech, Inc., San Francisco, CA), increased vessel stability and reduced both protein extravasation and intratumoral pressure, thus increasing concentration of chemoagents within the tumor mass [38]. Response to paclitaxel improved in patients treated with bevacizumab and paclitaxel combined compared with paclitaxel alone [39]. However, vessel stabilization was generally transient, lasting 5 to 8 days before the tumor reverted to abnormal vessel structure [40] with poor transport of molecules, thus requiring chronic bevacizumab treatment. This could lead to overpruning of the vasculature, resulting in reduced blood flow and intratumoral penetration. Understanding this paradox and developing combinatorial treatments that include targeting angiogenic response to chemotherapy will greatly increase the efficacy of the treatment and durability of the tumor response to chemodrugs. We therefore performed a series of studies in orthotopic breast cancer models examining the combined effects of a novel paclitaxel, nab-paclitaxel, and anti-VEGF-A antibody, bevacizumab, on tumor growth and on hematogenous and lymphatic metastases.

Materials and Methods

Materials

Dulbecco's modified Eagle's medium (DMEM), glutamine, sodium pyruvate, and nonessential amino acids were obtained from Life Technologies (Grand Island, NY). Human recombinant VEGF-A protein (VEGF-A₁₆₅ isoform) was purchased from PeproTech (Rocky Hill, NJ). Mouse anesthetic cocktail was a mixture of ketamine (Fort Dodge Animal Health, Fort Dodge, IA), xylazine (Phoenix Scientific, Inc., St. Joseph, MO), and sterile water. Endotoxin-free sterile saline was purchased from Sigma (St. Louis, MO).

Study Drugs

Paclitaxel albumin-bound particles for injectable suspension (nabpaclitaxel or Abraxane) was obtained from Abraxis BioScience, Inc., Los Angeles, CA. Drugs were reconstituted in saline, prepared fresh daily as required, and given within 1 hour of preparation. Bevacizumab (Avastin), humanized anti-VEGF-A antibody, manufactured by Genentech, Inc., was obtained from a local pharmacy.

Human MDA-MB-231 and MDA-MB-435 Carcinoma Cell Lines and Their Luciferase Derivatives

Luciferase-expressing cell line derivative of MDA-MB-231, 231-Luc+, has been extensively characterized in vitro and in vivo in prior studies and repeatedly assessed of reproducible spontaneous lymphatic and pulmonary metastases on orthotopic implantation in immunodeficient mice [41]. Luciferase-tagged MDA-MB-435 derivative subline was a generous gift from Dr. Sierra (Universitaria de Bellvitge, Barcelona, Spain). Cells were cultured at 37°C in an incubator gassed with 10% CO₂ in humidified air. The culture medium for cell lines consisted of DMEM supplemented with 5% fetal bovine serum, 2 mM glutamine, 1 mM sodium pyruvate, and nonessential amino acids. Tumor cells were harvested for passage by washing the monolayer with PBS, followed by 3 to 4 minutes of exposure to EDTA (0.5 mM) diluted in PBS. Cells were subcultured twice weekly and routinely tested for mycoplasma using an immunodetection kit from Roche Diagnostics GmbH (Penzberg, Germany).

Total RNA Extraction and Reverse Transcription-Polymerase Chain Reaction Analysis

Total RNA was extracted using an RNeasy Mini kit (Qiagen, Valencia, CA) and reverse-transcribed using RTG You-Prime Reaction beads (Amersham, Piscataway, NJ) and random hexamer primers (Invitrogen, Carlsbad, CA). Primer sequences, which were designed using GeneRunner software, are available on request. Human Universal cDNA (SuperArray, Frederick, MD) and a housekeeping gene, namely, β-actin, were used as positive controls. Standard end point reverse transcription-polymerase chain reaction (RT-PCR) was performed and analyzed using a FluroChem5500 (Alpha Innotech, San Leandro, CA) imaging system and software.

Western Blot Analysis

Antibodies for Western blot analysis were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Cells were lysed in cell lysis buffer (50 mM Tris-CHCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, PMSF, and protease inhibitors). Thirty micrograms of total protein was separated on 12% SDS-PAGE gels and transferred to nitrocellulose membranes (Pierce, Rockford, IL). Membranes were blocked for 1 hour with 5% BSA in 0.1% Tween 20 in Tris-buffered saline and were incubated with primary antibodies overnight at 4°C. Membranes were extensively washed and incubated with horseradish peroxidase-Clabeled secondary antibodies from Jackson Immunoresearch Laboratories (West Grove, PA) for 60 minutes before development using Super Signal Chemiluminescence substrate (Pierce).

Cytotoxic Assay

MDA-MB-231 cells were seeded at a density of 50,000 cell per well in 24-well plates and allowed to attach for 2 hours before treatment. Cells were treated with nab-paclitaxel alone (0.2 to 50 nM), VEGF-A₁₆₅ alone (20 ng/ml), or with combination of nab-paclitaxel and anti-VEGF-A antibody, bevacizumab (10 µg/ml). Control wells received an equal volume of growth medium. After 72 hours, cells were dislodged by EDTA-saline solution, and viable cells were enumerated using the Trypan blue exclusion method. Each experimental condition was tested in triplicate and reproduced in two separate experiments. The results are presented as percent of viable cells ± SD of the mean cell number in control wells.

Clonogenic Assay

Cells were seeded in triplicate at a density of 10,000 live cells per 100-mm dish containing 10 ml of complete medium. Two hours after seeding, cells were treated with nab-paclitaxel (10 nM), VEGF-A (20 ng/ml), bevacizumab (10 µg/ml), and combinations of nabpaclitaxel with either VEGF-A or bevacizumab diluted in 20 µl of growth medium. Control plates received medium only. After 14 days of incubation at 37°C, plates were washed with PBS and stained with 0.5% crystal violet dissolved in absolute methanol. Images were taken using an Alpha Innotech image analyzer FluroChem5500, and colonies with diameter >0.05 mm were enumerated using Alpha Ease FC software. The results are presented as the mean survival fraction derived from averaged values of three plates per experimental group ± SD.

Mouse Models of Human Breast Cancer

Orthotopic model of luciferase-tagged MDA-MB-231 cells was previously described [41]. Briefly, 4 x 10⁶ cells and 50% Matrigel were implanted into the mammary fat pad of 4- to 6-week-old female nu/nu mice (Harlan Sprague-Delaney, Indianapolis, IN). Every 2 to 3 days, perpendicular tumor diameters were measured by digital caliper and used to calculate tumor volume according to the formula: volume = Dd²π/6, where D equals larger diameter and d equals smaller diameter. The 231-Luc⁺ tumors and nontransfected parental cell line had an identical proliferation rate. Animal care was in accordance with institutional guidelines.

Treatment of Tumor-Bearing Mice with Nab-paclitaxel and Bevacizumab

Mice bearing 231-Luc⁺ tumors of 200 to 250 mm³ in volume were randomized into groups (5–10 per group) and treated with saline, nab-paclitaxel alone [10 mg/kg, intravenously (i.v.), daily for five consecutive days (qdx5)], bevacizumab alone [4 mg/kg, intraperitoneally (i.p.), twice a week], or nab-paclitaxel followed by bevacizumab. Nab-paclitaxel treatment was given once for five consecutive days (one cycle) or repeated two or three times with 1 week of rest interval (two and three cycles). The averaged tumor volume of mice treated with three cycles of combined therapy was 310 mm³. Bevacizumab treatment began 24 hours after the first injection of nabpaclitaxel, with a dose of 2, 4, or 8 mg/kg dissolved in 0.1 ml of sterile endotoxin-free saline and injected i.p. twice a week, for a total period of 5 weeks. The control group received saline, injected i.v. or i.p. (0.1 ml) on the same days as nab-paclitaxel and bevacizumab treatments, respectively. To allow maximal development of metastatic foci, mice were killed when the average tumor volume in the control group reached 2000 mm³, and tumor metastasis was determined by measuring luciferase activity in tissue extracts derived from 10 lymph nodes (LNs) and both lobes of lungs. All experiments were repeated at least twice with similar results.

Detection of Metastases By Measuring Luciferase Activity

Lymph nodes and lungs were excised, washed in PBS, and homogenized in 0.35 ml of cold cell culture lysis reagent buffer (Promega, Madison, WI) containing protease inhibitor cocktail and PMSF (Sigma, MO). Cell debris was removed by centrifugation. Protein concentrations of cleared lysates were determined by Bradford assay (Bio-Rad, Hercules, CA). Fifty microliters of Luciferase Assay Reagent (Promega) was mixed with 10 µl of lysate, and a 10-second average of luminescence was detected using a single-tube luminometer (Berthold Technologies, Bad Wildbad, Germany). Cell culture lysis reagent buffer without tissue homogenates and LNs or lungs of non-tumor-bearing mice were considered as background and subtracted from the results. The net results are expressed as relative light units normalized per milligram of total protein. To assess incidence of metastasis, extracts that had a luminescence signal of 400 light units above background (approximately 100 relative light units) were rated positive. This value was chosen because it was the minimal signal reproducibly detected above background in metastatic tissues, independently confirmed by immunohistochemical detection.

Statistical Analysis

Statistical analysis was performed using GraphPad InStat (GraphPad, Inc., San Diego, CA) or SPSS 14.0 (SPSS, Inc., Chicago, IL). Results were expressed as mean ± SE. Statistically significant differences of mean tumor volume among experimental groups were analyzed using a nonparametric Kruskal-Wallis test. Statistically significant differences of LN and lung metastases were analyzed by Fisher's exact test or Mann-Whitney U test. Metastatic burden was transformed with a log transformation and analyzed using analysis of variance with Tukey's follow-up comparisons.

Results

Low Concentrations of Nab-paclitaxel Increase VEGF-A Production in Breast Carcinoma Cells

To access the response of tumor cells toward chemotherapy, we first investigated whether VEGF-A expression in cultured MDAMB-231 cells could be up-regulated by nab-paclitaxel in vitro. As shown in Figure 1A, nab-paclitaxel applied at concentrations below 10 nM increased VEGF-A mRNA expression measured by quantitative RT-PCR (QRT-PCR) by two- to three-fold in a dose-dependent manner. Nab-paclitaxel at concentrations more than 20 nM decreased VEGF-A mRNA production compared with medium control, likely due to the cytotoxic effect of nab-paclitaxel. The changes in VEGF-A mRNA production as determined by QRT-PCR were confirmed by measuring secreted VEGF-A protein with ELISA (Figure 1B).Moreover, intracellular VEGF-A protein, as detected byWestern blot analysis, was also proportionally increased after exposure to 2.5 to 30 nM of nab-paclitaxel (Figure 1C). Similar results have been obtained with MDA-MB-435 cells. These results show that low concentrations of nab-paclitaxel could directly stimulate both VEGF-A mRNA and protein production in tumor cells in vitro.

Figure 1.

Low concentrations of nab-paclitaxel elicit VEGF-A production in cultured breast carcinoma cells and in breast tumors in vivo. (A) MDA-MB-231 cells were exposed to indicated concentrations of nab-paclitaxel for 72 hours and analyzed for mRNA expression of VEGF-A by QRT-PCR. (B) Tumors from mice treated with escalating doses of nab-paclitaxel were homogenized and analyzed for the presence of human VEGF-A protein by ELISA. (C) Lysates of cells treated by nab-paclitaxel at indicated doses were separated on SDS-PAGE, transferred to nitrocellulose membrane, and probed for expression of VEGF-A. β-Actin was also detected on the same blot and shown as a loading control. Note the increase in VEGF-A expression in lysates of nab-paclitaxel-treated cells relative to control.

Some Breast Carcinoma Lines Express VEGF-A Receptors Suggesting an Existence of Autocrine Loop

MDA-MB-231 and MDA-MB-435 have been previously reported to express functional VEGF-A receptors that have been implicated in the generation of VEGF-A-dependent autocrine loops [31,42,43]. Both tyrosine kinase VEGF-A receptors [43] and neuropilin-1 (NP-1) have been detected in MDA-MB-231 [31] and to a lesser extent, in MDA-MB-435 [31,42]. We have used luciferase-tagged derivatives of these lines that have not been analyzed previously either for the expression of VEGF-A receptors by neoplastic cells or for evidence of VEGF-A-driven autocrine regulation. In accord with published studies analyzing parental lines [43], three VEGF-A isoforms were detected in both 231-Luc⁺ and 435-Luc⁺ with predominant expression in the 231 derivative line (Figure 2A). VEGF-A receptors, VEGFR-2 and NP-1, were strongly detected in 231-Luc⁺ and weakly in 435-Luc⁺ line (Figure 2A). Expression of neuropilin-2 (NP-2) was not previously examined in either 231-Luc⁺ or 435-Luc⁺line. We detected strong NP-2 expression in 435-Luc⁺ cells only (Figure 2). This profile of mRNA expression was confirmed by immunohistochemical detection of the targets proteins (Figure 2B). As has been suggested previously [44,45], coexpression of VEGF-A and its specific receptors in either parental human breast carcinoma or metastatic derivatives suggests an autocrine role for VEGF-A in the regulation of vital tumor cell activities such as proliferation and survival. Both of these activities can contribute to the resistance of tumor cells to the action of cytotoxic drugs.

Figure 2.

Luciferase-tagged derivatives of human breast carcinoma lines express VEGF-A receptors. (A) Total RNA was isolated from 231-Luc⁺ and 435-Luc⁺ cells and used to synthesize cDNA for PCR amplification using primers for VEGF-A and its receptors. Universal human cDNA was used as a positive control. Amplification of β-actin is shown as a reference for housekeeping gene expression and for comparing a relative gene target expression in 231-Luc⁺ and 435-Luc⁺ cell lines. (B) 231-Luc⁺ and 435-Luc⁺ cells were stained with anti-Chuman VEGFR-2 and anti-CNP-1 and NP-2 antibodies followed by appropriate horseradish peroxidase-Cconjugated secondary antibodies and DAB development. Arrows indicate positively stained cells. Brown color precipitate was not developed on the application of secondary antibodies alone. Original magnifications, x400.

Anti-VEGF-A Antibody, Bevacizumab, Abolished the Protective Effect of VEGF-A Against Paclitaxel Cytotoxicity on Cultured Breast Carcinoma Cells In Vitro

Because VEGF-A protects endothelial cells against chemotherapies [12], we examined whether it also enhances survival of VEGF-A receptors positive tumor cells exposed to cytotoxic drugs. We hypothesized that low concentrations of nab-paclitaxel might induce VEGFR signaling leading to enhanced tumor cell survival and, subsequently, to decreased therapeutic efficacy. To test this hypothesis, we measured the cytotoxicity and clonogenic survival of 231-Luc⁺ cells treated with nab-paclitaxel (0.2 to 50 nM), VEGF-A (20 ng/ml), and bevacizumab (10 µg/ml) alone or in combination (Figure 3). In the cytotoxic assay, the IC₅₀ of nab-paclitaxel for cultured 231-Luc⁺ cells was determined as 9.9 nM (Figure 3A). A 50-nM dose of nabpaclitaxel reduced the cell number by 90% compared with culture medium-only control cells. Exogenous VEGF-A shifted the IC₅₀ to 14.2 nM, thus decreasing the efficacy of a nab-paclitaxel-mediated cytotoxic effect. In contrast, neutralization of VEGF-A by bevacizumab caused an opposite effect by shifting the IC₅₀ to 2.27 nM, thus strongly increasing nab-paclitaxel cytotoxicity. The most conspicuous changes in shifting cell sensitivity to nab-paclitaxel were observed in the IC₅₀ range. This suggests that very low concentrations of cytotoxic drugs may not be sufficient for eliciting VEGF-A signaling, whereas cell death at high drug doses occurs before VEGF-A can play a protective role.

Figure 3.

Nab-paclitaxel cytotoxicity is negated by VEGF-A and is enhanced by anti-VEGF-A antibody bevacizumab. (A) MDA-MB-231 cells were treated with nab-paclitaxel only (black circles), with nab-paclitaxel and VEGF-A (20 ng/ml, white circles), with a mixture of nabpaclitaxel and bevacizumab (10 µg/ml, black triangles) or left untreated. Seventy-two hours later, cell number and viability was determined using Trypan blue exclusion method. The results were averaged from triplicate per treatment and presented as percentage of viable cells ± SD from the cell number in untreated wells. (B) The survival capacity of MDA-MB-231 cells was determined by seeding cells at the low density of 10,000 per 100-mm dish followed by treatment with nab-paclitaxel (10 nM), VEGF-A (20 ng/ml), bevacizumab (10 µg/ml), a combination of nab-paclitaxel with VEGF-A, or bevacizumab given at the same concentrations as a single drug. The results are presented as the mean number of colonies per dish ± SD, detected 14 days after seeding.

We also examined single-cell survival of 231-Luc⁺ cells after treatment with nab-paclitaxel, VEGF-A, and bevacizumab alone or in combination. Exposure to 10-nM nab-paclitaxel reduced cell survival by 43%, whereas addition of VEGF-A (20 ng/ml) increased survival by 49%, compared with medium-treated control cells (Figure 3B). Bevacizumab alone had relatively modest effect by decreasing cell survival by 23% compared with control. Importantly, VEGF-A virtually negated the cytotoxic effect of the nab-paclitaxel, whereas combination of bevacizumab and nab-paclitaxel greatly potentiated the drug's efficacy resulting in a 76% decrease in the number of colonies compared with vehicle-treated controls. The differences in cell survival after treatment with nab-paclitaxel/bevacizumab combination and all other groups were statistically significant (P < .05 by Student's t-test). Similar effects were observed in experiments with MDA-MB-435 and 435-Luc⁺ cells. These findings in the models of tumor cell responses to chemotherapy in vitro suggest that nabpaclitaxel/bevacizumab combination strategy could significantly improve therapeutic outcome in vivo.

Bevacizumab Potentiates Nab-paclitaxel Inhibition of Tumor Growth In Vivo

The in vitro results suggested that enhanced VEGF-A production by paclitaxel-treated cells might contribute to tumor chemoresistance through stimulation of angiogenic and cell survival pathways. On the basis of these findings, we hypothesized that neutralizing VEGF-A concurrently with chemotherapy would enhance antitumor efficacy and sustain therapeutic gains after cessation of the treatment. To test this hypothesis, we examined the antitumor efficacy of nab-paclitaxel alone and in combination with bevacizumab in an orthotopic model of luciferase-tagged human breast carcinoma, MDA-MB-231 (231-Luc⁺). This model is extensively characterized in our laboratory and is known to reproducibly generate LN and pulmonary metastases in 100% and 70% of tumor-bearing mice, respectively [41]. Moreover, recent studies in our laboratory established that 50% of control mice (n = 15) have metastases in contralateral LN and 40% of animals have metastasis to the brain. To our knowledge, this is the first described animal model of orthotopically implanted human breast carcinoma that faithfully mimics the severity of the human metastatic disease including spread not only to the proximal LN but also to the contralateral LN, the lungs, and the brain.

The first experiment was conducted with nab-paclitaxel (10 mg/kg) administered i.v. qdx5 either alone or combined with bevacizumab (2 mg/kg, i.p., twice a week, for the duration of the experiment). Combination therapy showed an additive effect in the suppression of tumor growth as evidenced by the stronger reduction of tumor volume (87.7% vs 71.3% and 54.1% for nab-paclitaxel and bevacizumab monotherapies, respectively, P < .05; Table 1). Tumor growth delay (TGD) was also increased from 13 and 20 days for bevacizumab and nab-paclitaxel, respectively, to 33 days in mice treated with combined therapy (Figure 4; Table 1). However, despite this evidence of improved tumor responses after combination therapy, no complete responses were observed in any of the groups. Moreover, tumors regrew in 100% of the treated mice within 60 to 70 days after cessation of either single or combined therapies (Table 1; Figure 4).

Table 1.

Effect of Nab-paclitaxel, Bevacizumab, and Combined Therapies on Tumor Growth in the MDA-MB-231-Luc⁺ Tumor Model.

Treatment	Bevacizumab Dose (mg/kg)	Mean Tumor Volume (mm3)	%TGI†vs Control	TGD‡vs Control	% Complete Responses§
One cycle of treatment
Control		2079 ± 287
Bev*	2	953 ± 127	54.16	13	0
ABX*		596 ± 98	71.33	20	0
ABX + Bev*	2	255 ± 46	87.73	33	0
Two cycles of treatment
Control		2391 ± 432			0
Bev	4	2089 ± 251	12.56	7	0
ABX		117 ± 38	95.11	>65	0
ABX + Bev	2	138 ± 42	94.23	>65	20
ABX + Bev	4	60 ± 17	97.49	>65	40
ABX + Bev	8	36 ± 16	98.49	>65	40
Three cycles of treatment
Control		1829 ± 241			0
ABX + Bev	4	52 ± 17	97.16	>107	80

^*Bev, ABX, and Bev + ABX denote treatment with bevacizumab (4 mg/kg), nab-paclitaxel/Abraxane (10 mg/kg), and combination of the two drugs, respectively.

^†TGI (tumor growth inhibition) is presented as percent reduction in the mean tumor volume in experimental groups compared with that in saline-treated control groups.

^‡TGD (tumor growth delay) is defined as a number of days that delayed the mean tumor volume per group reaching 1000 mm³ as compared with the saline-treated control mice.

^§Complete response was defined as the absence of palpable tumor at the original tumor injection site for the entire length of the experiments (75–110 days).

Figure 4.

Bevacizumab significantly improves antitumor effect of nab-paclitaxel in vivo. (A, B) Mice bearing orthotopic MDA-MB-231-Luc⁺ tumors of 200 to 250 mm³ were treated with one or two cycles of saline, nab-paclitaxel alone (10 mg/kg, qdx5 cycle), bevacizumab alone (4 mg/kg), or both drugs. (C) A similar experiment was performed in mice bearing ∼310 mm³ tumors that were treated with three cycles of nab-paclitaxel and concurrent bevacizumab or with saline control. The results are presented as the mean tumor volume ± SE per group measured at indicated days after treatment.

Optimized Regimen of Bevacizumab and Nab-paclitaxel Combined Therapy Potentiates and Sustains Efficacy of Cytotoxic Therapy

To further improve tumor responses, we modified the regimen by increasing the dose of bevacizumab from 2 mg/kg to 4 and 8 mg/kg and the number of nab-paclitaxel treatments to two or three cycles given with 1 week of interval between each cycle. The change in regimen was possible because both bevacizumab and nab-paclitaxel have relatively high maximum tolerated doses. This is particularly significant for nab-paclitaxel whose maximum tolerated dose is 2.5-fold higher than that of Taxol in both rodents [6,46] and in cancer patients [47–49]. This property of nab-paclitaxel allowed us to administer repetitive chemotherapy cycles without causing unacceptable toxicity as judged by weight loss (more than 20% of the body weight) and behavioral changes. Addition of bevacizumab to nab-paclitaxel did not change the safety profile, although increased toxicity for the combined Taxol and bevacizumab therapy has been documented in human subjects.

We performed three experiments with one, two, or three cycles of nab-paclitaxel. Those results are presented in Figure 4, A, B, and C, respectively. The average tumor volume before the treatments was approximately 250 mm³ in the experiments with one and two cycles (Figure 4, A and B), and approximately 310 mm³ in the experiment with three cycles of chemotherapy (Figure 4C). The increase of bevacizumab from 2 to 4 mg/kg showed some improvement in tumor suppression but further escalation to 8 mg/kg had no additional effect (Table 1). Bevacizumab alone had no effect in TGD in any of the experiments. In contrast, significantly higher tumor inhibition was observed after two cycles of nab-paclitaxel compared with one cycle (95.11% vs 71.33%, P < .005; Table 1). Notably, two cycles of nabpaclitaxel delayed tumor regrowth for longer than 65 days after treatment compared with only 20 days after one cycle (Table 1). However, the most significant improvement in tumor response was observed on the combination of the two drugs: 40% of mice treated with nabpaclitaxel and either 4 or 8 mg/kg of anti-VEGF-A antibody (total N = 10) had complete and durable tumor regressions. Mice treated with the same regimen with 2 mg/kg of bevacizumab had 20% complete responses suggesting a trend toward dose dependency of the ancillary anti-VEGF-A therapy.

No palpable or measurable tumors were observed in the cured mice for the duration of the experiment (>65 days). Moreover, taking advantage of the great sensitivity of luciferase detection of tumor cells (one cell is equivalent to 10 light units), we measured luciferase activity in the extract of the mammary fat pad used for 231-Luc⁺ cell implantation. The extracts contained no measurable enzymatic activity indicating complete tumor eradication in 231-Luc⁺ mice. The remarkable effect on eradication of well-established tumors (>200 mm³ in volume before the treatment) was reproduced in the third experiment in which the number of nab-paclitaxel cycles was increased to three and bevacizumab's dose fixed to 4 mg/kg. Tumor regrowth was observed in one mouse in the treated group (Figure 4C; three cycles) but only after 107 days after treatment. After three cycles, tumor growth was suppressed 5.35- and 1.64-fold days longer, compared with one and two cycles, respectively. Apart from one escapee, most of the mice (80%) treated with this combinatory regimen had complete regressions (Table 1; Figure 4C). Thus, three cycles of the combined therapy doubles the number of complete responses compared with two cycles (40%), whereas a one-cycle combination treatment had no capacity to eradicate tumors (Table 1; Figure 4A). These data indicate that repetitive delivery of the well-tolerated dose of nab-paclitaxel (10 mg/kg) supplemented with anti-VEGF-A therapy can eradicate well-established tumors.

Nab-paclitaxel/Bevacizumab Combination Therapy Significantly Suppresses Both Lymphatic and Pulmonary Metastases

Metastasis is the most devastating consequence of the malignant disease. The 231-Luc⁺ model induces both intratumoral [41] and peritumoral [50] lymphatic vessels and first metastasizes to LNs with subsequent spread to visceral organs [41]. Because one tumor cell produces 10 light units on exposure to luciferin, as few as 50 cells can reliably be detected in the lysate samples. This model, therefore, allows sensitive assessment of the antimetastatic potential of nabpaclitaxel/bevacizumab combination therapy with regard to both incidence and total metastatic burden. These parameters were analyzed in all experiments by measuring luciferase activity in tissue extracts from ipsilateral LNs and lungs. Measurements of luciferase activity in contralateral LNs and brain extracts were additionally performed after three cycles of combined treatment.

No effect on metastasis was observed after one cycle of nab-paclitaxel, bevacizumab or combined therapy (data not shown). This is despite the fact that one cycle of combination therapy inhibited tumor growth by 87% and delayed regrowth of primary tumors by 33 days (Table 1). In contrast, two cycles of the combined therapy, but not the single treatments, reduced the burden and the incidence in the proximal LN by 80% and 60%, respectively (Table 2; Figure 5). Notably, two cycles of the combined therapy eliminated pulmonary metastases in 100% of the treated mice (n = 5) whereas monotherapies were ineffective (Table 2). Escalation of the bevacizumab dose (2, 4, or 8 mg/kg) had no significant impact on the overall effect (Figure 5), although the higher doses of the antibody treatment (4 or 8 mg/kg) tend to be more effective in all experiments. Bevacizumab at 4 mg/kg was used to study the effect of a three-cycle combination therapy. In this experiment, contralateral LN and brain metastases were eliminated in 100% of the treated mice (n = 5), whereas ipsilateral LN and lung metastases were eradicated in 80% of the mice (Table 2; Figure 5). All differences in the average metastatic burden between control- and the combined therapy-treated groups were statistically significant for all organs examined in this study (Table 2).

Table 2.

Effect of Nab-paclitaxel, Bevacizumab, and Combined Therapy on Metastatic Incidence and Burden.

Tissue	Treatment	Two Cycles of Nab-paclitaxel			Three Cycles of Nab-paclitaxel
		Incidence†	Burden‡	P§	Incidence	Burden	P
Proximal LN	Control	10/10 (100%)	31.1 ± 9.4		5/5 (100%)	123.7 ± 73.0
	Bev*	5/5 (100%)	31.8 ± 9.1	NS¶	ND
	ABX*	5/5 (100%)	26.0 ± 12.8	NS	ND
	Bev + ABX*	2/5 (40%)	6.0 ± 4.3	.022**	1/5 (20%)	16.7 ± 3.4	.013**
Contralateral LN	Control	ND#			3/5 (60%)	72.3 ± 47.0
	Bev + ABX	ND			0/5 (0%)	0	<.001**
Lungs	Control	7/10 (70%)	7.7 ± 2.2		5/5 (100%)	21.5 ± 9.4
	Bev	3/5 (100%)	3.2 ± 2.0	NS	ND
	ABX	4/5 (100%)	7.9 ± 2.7	NS	ND
	Bev + ABX	0/5 (0%)	0	.025**	1/5 (20%)	5.7 ± 1.4	.041**
Brain	Control	ND			2/5 (40%)	1.3 ± 0.8
	Bev + ABX	ND			0/5 (0%)	0	<.001**

^*Bev, ABX (nab-paclitaxel), and Bev + ABX denote treatment with bevacizumab (4 mg/kg), nab-paclitaxel (10 mg/kg), and combination of the two drugs, respectively.

^†Incidence is defined as number of mice with LN or lung metastasis per total number of mice with tumors in each experimental group. Percentage of mice with metastases is shown in parentheses.

^‡Burden is presented as luciferase activity that was measured in extracts from 10 tumor-proximal or contralateral LNs and both lobes of lungs of each mouse. In the experiment with three ABX cycles, the brain was analyzed as well. Samples that contained a minimum of 400 light units per 20-µl lysate were rated positive. Lysates from tissues of non.tumor-bearing mice produced a luminescence signal of 100 light units, which was considered as background and subtracted from positive values. Results are presented as the average amount of relative light units per group x 10⁴ ± SE.

^§P value was calculated by Fisher's exact test.

^¶NS, nonsignificant; the difference between experimental and control groups does not reach statistical significance when analyzed by the Fisher's exact test.

^#ND, not determined; in the experiment with three ABX cycles, the comparison was done between control and combination therapy only.

^**Denotes statistical significance when log transformed values for metastatic burden in experimental and control groups were analyzed as described under the Materials and Methods section.

Figure 5.

Effect of nab-paclitaxel and bevacizumab on lymphatic and systemic metastases. Mice were treated with either two (A) or three (B) cycles of nab-paclitaxel alone, bevacizumab, or combination. Tumor-proximate LNs (n = 10) and lungs were harvested from mice in experimental groups when the mean tumor volume in the saline-treated group reached 2 cm³. After three cycles of chemotherapy, luciferase activity was determined in proximal and contralateral LNs, the lungs, and the brain. The results are presented as the mean light units x 10⁴ per group ± SE normalized per milligram of total protein. x-axis: Bevacizumab and ABX denote bevacizumab and nabpaclitaxel, respectively. Numbers above and below the solid line separating ABX and bevacizumab indicate the drugs and a dose expressed in milligrams per kilogram used for the treatment of each of the experimental groups.

In summary, stark differences were observed between a completely ineffective one-cycle treatment, singular or combined, and the twoor three-cycle therapies that drastically reduced both incidence and the averaged metastatic burden. Bevacizumab seems to be essential for the effective suppression of metastasis by nab-paclitaxel as only the combined therapy group displayed appreciable antimetastatic effect (Table 2). These results suggest that repetitive cycles of nabpaclitaxel when administered concurrently with bevacizumab have a strong potential to eradicate breast cancer metastasis in ipsilateral LN and other metastatic sites.

Discussion

The main findings of this study are that tumor cells respond to cytotoxic stress by producing VEGF-A and that neutralization of VEGF-A during chemotherapy is essential for sustaining inhibition of both primary tumors and metastasis. Moreover, synergy developed between nab-paclitaxel and anti-VEGF-A antibody has strong potential to eradicate lymphatic- and organ-specific metastases that were not achievable by using either of these drugs alone.

Our in vitro studies demonstrated that nontyrosine kinase and tyrosine kinase VEGF-A receptors are present in both MDA-MB-231 and MDA-MB-435 cells. Coexpression of VEGF-A and its cognate receptors in malignant cells have been previously suggested as responsible for the formation of positive circuit that protects tumor cells from adverse effects of chemotherapy [24,44]. This presumably occurs because VEGF-A binding to VEGFR-2 activates the AKT pathway and increases the expression of prosurvival factors, such as bcl-2 [51] and survivin [52]. A previous study by Sweeney et al. [12] showed that VEGF-A protects endothelial cells from the cytotoxic effect of docetaxel. In accord with the published findings, we observed that VEGF-A significantly reduces nab-paclitaxel cytotoxicity whereas anti-VEGF-A antibody, bevacizumab, abrogates the VEGF-A dependent protective effect on tumor cells (Figure 3). Thus, over-expression of VEGF-A induced by cytotoxic therapy not only protects endothelial cells and increases angiogenesis, but also can prolong survival of the damaged neoplastic cells during a vulnerable period of tumor vasculature regeneration.

Because of the crucial role of VEGF-A in chemoresistance [53], VEGF-A targeting agents combined with chemotherapy show significantly improved tumor responses to treatment in a variety of preclinical [33,35,54,55] and clinical studies [37,56]. In particular, the anti-VEGF-A antibody bevacizumab, which was approved in 2004 for use in colorectal cancer [10] combined with 5-fluorouracil, leucovorin, oxaliplatin, or irinotecan, showed an additive or synergistic effect when combined with chemotherapeutic drugs [33,55]. In 2006, bevacizumab in combination with carboplatin and paclitaxel significantly improved overall survival and was approved as the first line treatment of non-small cell lung carcinoma [57]. It has recently been approved in February 2008 for metastatic breast cancer in combination with paclitaxel (Bloom; www.medpagetoday.com). Bevacizumab also showed efficacy against renal and ovarian cancers when used as a single agent, and in patients with lung and breast cancer when combined with chemotherapy [39]. Clinical studies demonstrated that bevacizumab prolonged survival in previously untreated patients with advanced non-small cell lung cancer when combined with paclitaxel/carboplatin [58]. The E2100 trial showed that bevacizumab and paclitaxel combined therapy did not affect the progression-free survival but doubled the response rate compared with the treatment with paclitaxel alone [39,59]. Bevacizumab combined with taxanes (paclitaxel or docetaxel) has also been shown to significantly improve the cancer-related symptoms in patients with refractory epithelial ovarian tumors, albeit this improvement was relatively short-lived [56].

Our study, for the first time, demonstrated the effectiveness of combination therapy with bevacizumab and nab-paclitaxel, the next generation of taxanes in an experimental orthotopic and metastatic model of human breast cancer. In vitro, sequestration of VEGF-A by bevacizumab significantly enhanced nab-paclitaxel cytotoxicity to breast carcinoma MDA-MB-231 cells leading to the restriction of their clonogenic survival. The latter is crucially important for the successful establishment of secondary tumors in distant organs because of the inherent inefficiency of the metastatic process [60]. In vivo, bevacizumab synergistically increased the antitumor efficacy of nab-paclitaxel against breast tumors. Although nab-paclitaxel alone potently inhibited tumor growth, particularly after the two-cycle therapy, tumor growth resumed in all mice regardless of the extent of tumor suppression and TGD. Additionally, no complete regressions were observed in any of the mice treated with either nab-paclitaxel or bevacizumab alone. In contrast, nab-paclitaxel/bevacizumab combination therapy resulted in a significant number of complete responses. Two qdx5 cycles of nab-paclitaxel alone and in combination with bevacizumab, showed more efficacy than a single cycle of monotherapy or combined therapy, and retained limited toxicity. More importantly, for the first time, the superior effect of the combined therapy was demonstrated here not only for inhibition of the primary breast tumors but also for suppression of both lymphatic- and organ-specific metastases.

The superior efficacy of nab-paclitaxel/bevacizumab combination therapy might be explained by several mechanisms. First, bevacizumab, although not highly effective as a monotherapy, may block the reactionary angiogenesis by neutralizing nab-paclitaxel-induced VEGF-A in tumor cells. Reactionary angiogenesis can be promoted by multiple VEGF-A dependent mechanisms including recruitment of proangiogenic macrophages and endothelial progenitor cells [61]. Second, as was recently suggested [62], anti-VEGF-A antibody therapy might promote normalization of tumor vessels leading simultaneously to both increased drug delivery and decreased metastasis [62]. This is because depletion of VEGF-A from the tumor environment eliminates the majority of structurally defective and malfunctioning tumor microvessels. The few well-constructed vessels left behind can function properly, potentially improving drug delivery to the tumor [62]. Third, bevacizumab could block enhanced VEGF-A production by tumor cells exposed to chemotherapy and a variety of stress conditions, thus depriving a vital survival factor from both endothelial and malignant cells [24]. Fourth, the combination of cytotoxic and anti-VEGF therapies has been reported to reduce recruitment of tumor stem-like cells that may contribute to recovery of tumor vasculature after partial response to chemotherapy [63]. Finally, we have recently demonstrated in the orthotopic 231-Luc⁺ breast carcinoma model that anti-VEGF-A therapy significantly inhibits tumor lymphangiogenesis [41], which would explain the inability of nab-paclitaxel alone to reduce lymphatic metastasis, and shows an additional advantage of its combination with bevacizumab.

In summary, our data support the hypothesis that VEGF-A plays a critical role in tumor recovery after nab-paclitaxel chemotherapy either by acting directly on malignant cells expressing VEGF-A receptors or indirectly, by facilitating both angiogenesis and lymphangiogenesis [64] in drug-injured regions. We also conclude that tumor regrowth and metastasis to LNs and lungs can be prevented by the combination of cytotoxic therapy with VEGF-A-neutralizing agents. Although the conclusions from experimental models must be confirmed in multiple clinical settings, these results are in line with the encouraging outcome of the recent limited clinical trials demonstrating benefits of the nab-paclitaxel/bevacizumab combination therapy in breast cancer patients [8,49]. The improved antitumor efficacy and better safety profile of nab-paclitaxel over conventional solvent-based paclitaxel [49,65] make it a promising candidate for combining with anti-VEGF-A antibody. Our study demonstrates a high efficacy of nab-paclitaxel and anti-VEGF-A combined therapy against aggressive experimental metastatic breast tumors and strongly advocates for the inclusion of bevacizumab as a standard modality administrated concurrently with nab-paclitaxel therapy in clinics.

Abbreviations

Nab-paclitaxel

paclitaxel albumin-bound particles also known as Abraxane or ABI-007

i.p.

intraperitoneally

i.v.

intravenously

LN

lymph node

NP-1 and NP-2

neuropilin-1 and -2

RT-PCR

reverse transcription-polymerase chain reaction

TGD

tumor growth delay

VEGF-A

vascular endothelial growth factor-A

VEGFR-2

vascular endothelial growth factor receptor 2