Introduction
Despite advances in primary surgical therapy and first line chemotherapy, the majority of patients suffering from advanced surgical cancer ultimately relapse. Salvage therapy for platinum refractory disease or heavily pre-treated patients has proven unsuccessful with regard to survival. Therefore, therapy in these patients is primarily aimed at improving quality of life. Symptomatic malignant ascites remains a severe therapeutic challenge in patients suffering from recurrent ovarian cancer or other cancer types such as colorectal, stomach, pancreatic or lung cancer. Malignant ascites and pleural effusion are significant causes of morbidity and mortality, and frequently impair the quality of life in patients who are already distressed by the underlying disease. It is known that the median survival of patients suffering from malignant ascites is significantly reduced to 20 weeks following diagnosis if all cancer types are included (1,2). However, effective treatment of malignant ascites in palliative patients is a therapeutic challenge and often frustrating. The established method of treating malignant ascites is paracentesis and the intraperitoneal (i.p.) instillation of chemotherapeutic agents such as cisplatin, 5-fluorouracil or doxorubicin (3,4). However, the efficacy of i.p. chemotherapy in a palliative setting has only been investigated in small cohorts of highly selected patients. Although studies have shown partial responses, the results have been inconsistent and some studies have reported significant side-effects (5).
A better understanding of the underlying molecular mechanisms that regulate the formation of malignant effusions have led to the concept that the angiogenic biomolecule vascular epithelial growth factor (VEGF) is important in the pathobiology of a variety of tumours and the pathogenesis of malignant and non-malignant effusions (6). It promotes tumour angiogenesis, formation of ascites and pleural effusions, and metastasis in ovarian cancer (7). Inhibitors of VEGF, such as the antibody bevacizumab, suppress tumour growth in ovarian cancer models, possibly by inhibiting angiogenesis (7). It is known that in patients with malignant effusions, high levels of VEGF are found in serum/plasma as well as in the malignant effusion (8). Previous studies showed a positive therapeutic effect of bevacizumab in the systemic treatment of patients with ovarian carcinoma and formation of ascites (9). Numnum et al demonstrated that patients undergoing treatment with bevacizumab showed a symptomatic relief of ascites and required no therapeutic paracentesis after treatment with the antibody (9). The most common serious bevacizumab-related toxicities are hypertension, fatigue, proteinuria, bleeding and pain (10). Some cases of gastrointestinal perforation or fistula, brain ischemia, pulmonary hypertension, gastrointestinal bleeding and wound healing complications have been reported (11).
Intraperitoneal administration of bevacizumab has recently been demonstrated to be efficacious in locally pretreated cancer patients suffering from malignant ascites (12). Since malignant pleural effusion contains high numbers of active tumour cells, the local intrapleural application of bevacizumab is a rational alternative to intravenous (i.v.) administration. However, to date only a few patients with malignant ascites have been treated with intraperitoneally delivered bevacizumab, and therefore this route of administration needs to be further explored. In this study, we have compared i.v. and i.p. administration of bevacizumab in an ovarian cancer mouse model with intraperitoneal metastasis. We have examined whether a single i.p. administration of bevacizumab is capable of reducing ascites-related body surface and prolonging survival.
Materials and methods
Cells and cell lines
SKOV-3.ip1 ovarian adenocarcinoma cell lines were kind gifts from Drs Judy Wolf and Janet Price (University of Texas MD Anderson Cancer Center, Houston, TX, USA), respectively. Cell lines were maintained in recommended conditions. Cells were grown at 37˚C in a humidified atmosphere of 5% CO2.
To generate resistant ovarian cancer cells, SKOV-3.ip1 cells were cultured in growth medium (GM) containing cisplatin. Initially, 10 ng/ml cisplatin was added to the GM and the dose was increased by 10 ng/ml per week to a final concentration of 60 ng/ml.
Anticancer drugs
Paclitaxel and novantrone were obtained from the pharmacy of the University of Düsseldorf Medical Centre and handled according to the manufacturer’s instructions. Paclitaxel (6 mg/ml) was diluted at a ratio of 1:15 with saline. Novantrone (2 mg/ml) was diluted at a ratio of 1:50 with saline. Bevacizumab was purchased from Roche (Basel, Switzerland). Each mouse assigned to a bevacizumab treatment received a single application. Bevacizumab was diluted in 200 μl of phosphate-buffered saline (PBS) immediately prior to administration.
Therapeutic ovarian cancer model
Female CB17 SCID mice were obtained at 3-4 weeks of age and quarantined at least 1 week prior to the study. Mice were kept under pathogen-free conditions according to the American Association for Accreditation of Laboratory Animal Care guidelines. Animal protocols were reviewed and approved by the state authorities according to the German animal protection law. Mice were injected i.p. with 1×107 drug resistant SKOV-3.ip1 cells on day 0. Mice were assigned into six treatment groups, each consisting of 11 mice: Group 1 served as the control (injection of PBS); Group 2 and 3 received bevacizumab i.p. or i.v., respectively; Group 4 was treated with paclitaxel; the therapy of Group 5 and 6 consisted of paclitaxel plus bevacizumab i.p. or i.v., respectively. Each mouse assigned to a chemotherapy treatment received 2 applications of paclitaxel (10 mg per kg body weight) on days 11 and 13. Bevacizumab was administered at a total dose of 10 mg per kg body weight on day 12. Drug injections were performed either intravenously through the lateral tail vein (i.v. group) or via the peritoneum (i.p. group).
Abdominal circumference and body weight were measured weekly for calculation of the body surface until 35 days following initial treatment. Variations in body surface are dependent on ascites volume plus tumour burden and therefore the body surface is considered a valuable parameter for disease progression.
A total of 21 days following the initial treatment, a representative mouse from each treatment group was sacrificed and the expression of VEGF was examined in the macroscopic disease and in the ascites. In the remaining mice, survival was followed daily until mortality.
RNA preparation and quantitative RT-PCR
The total cellular RNA of tumour cells was extracted from 2×105 cells using the RNeasy mini prep kit (Qiagen, Santa Clarita, CA, USA) and additionally treated with DNase I (Life Technologies Inc; Rockville, MD, USA) for 30 min. PCR products from the VEGF gene were used for the creation of the standard curve. A GeneAmp RNA PCR core kit (Applied Biosystems, Foster City, CA, USA) was used for cDNA synthesis and PCR amplification of cDNA products. TaqMan primers and probes were designed by the Primer Express 1.0 software and synthesized by Applied Biosystems. Oligonucleotide sequences for the amplification of the VEGF gene were: forward primer CAT GCA GAT TAT GCG GAT CAA; reverse primer TTT GTT GTG CTG TAG GAA GCT CA; and probe 6FAM-CCT CAC CAA GGC CAG CAC ATA GGA GA-TAMRA. The human housekeeping gene glyceraldehydes-3-phosphate dehydrogenase (GAPDH) served as an internal control.
Each single real-time PCR contained 1X TaqMan® EZ RT-PCT kit (Applied Biosystems), 100 nM forward primer, 100 nM reverse primer, 100 nM probe and 0.025% bovine serum albumin in a final volume of 9 μl per reaction. The PCR contained 1 μl of template or water as the control.
PCR was carried out using a LightCyclerTMSystem (Roche Molecular Biochemicals, Indianapolis, IN, USA) according to the manufacturer’s instructions. Thermal cycling conditions were: 2 min at 50˚C, 30 min at 60˚C, 5 min at 95˚C and 40 cycles of 20 sec at 94˚C and 1 min at 60˚C. Data were analyzed with LightCycler software. Known amounts of human total RNA (200, 20, 2 and 0.2 ng/μl) were amplified to generate a standard curve for the determination of the concentration of the unknown samples.
Statistical analysis
Data were presented as the mean values ± standard deviation. The body surface among groups was assessed with a two-tailed Student’s t-test. P<0.05 was considered to indicate a statistically significant difference. Survival data were plotted as a Kaplan-Meier curve. Statistical analyses were performed using Prism 4 software (GraphPad Software, San Diego, CA, USA).
Results
The orthotopic murine model of peritoneal disseminated ovarian cancer
One representative mouse per treatment group was investigated to confirm the tumour model. The dissemination of ovarian cancer was confirmed surgically. Samples of each tumour burden were successfully checked for VEGF overexpression. Fig. 1 shows the data of all six sacrificed mice. VEGF was expressed 2841-fold (p<0.001) higher in the tumour tissues than in the peritoneum, which was defined as the healthy control tissue.
Overall survival
Advanced carcinomatosis was allowed to develop for 12 days and mice were treated with either a single dose of PBS (negative control), bevacizumab i.v., bevacizumab i.p., paclitaxel i.p. or a combination of bevacizumab and paclitaxel. Fig. 2 shows the results in a plot according to Kaplan-Meier calculations. In pairwise comparisons, survival of mice in all treatment groups was significantly enhanced when compared to the non-treatment control group (p<0.001 for bevacizumab + paclitaxel and p<0.05 for single agent bevacizumab i.p., bevacizumab i.v. and paclitaxel i.p.). Survival of mice treated with a combination regimen of bevacizumab and paclitaxel was significantly prolonged compared to mice treated with a single agent (p<0.001). However, differences between i.p. and i.v. administration of bevacizumab were not statistically significant in the single agent regimen or combination regimen with paclitaxel.
Therapeutic success
Success of the treatment was assessed via reduction of the body surface, which is proportional to the intracorporal ascites volume.
As shown in Fig. 3, the mean abdominal surface (AS) 24 days after initiation of treatment was 376, 266, 298, 272, 231 and 234 mm2 for mice treated with PBS, bevacizumab i.p., bevacizumab i.v., paclitaxel i.p, bevacizumab i.p. + paclitaxel i.p. or bevacizumab i.v. + paclitaxel i.p., respectively. The mean AS of all mice prior to treatment was 227±22 mm2. The results of the statistical analyses among the treatment arms are shown in Table I. Each treatment resulted in a significant decrease of body surface with the exception of bevacizumab i.v. In that group, two mice were non-responders to the therapy and showed a body surface comparable to the animals without treatment. It is noteworthy that the combination of paclitaxel plus bevacizumab significantly improved body surface as well as overall survival in comparison to a treatment with one of the drugs.
Discussion
Ovarian carcinoma is the leading cause of mortality from gynaecological cancer (7). In general, the tumour progression is already advanced at the time of diagnosis. Despite advances in primary surgical therapy and first line chemotherapy, the majority of patients suffering from advanced surgical cancer ultimately relapse. Salvage therapy for platinum refractory disease has shown limited efficacy with regard to survival. Therapy in these patients is therefore primarily aimed at improving quality of life. Secondary symptoms such as symptomatic malignant ascites are a therapeutic challenge in the palliative setting. In addition, it is well known that the prognosis of patients with malignant pleural effusion is worse compared to those without secondary symptoms (1). The effect of paracentesis is usually of short duration and the effect of i.p. administration of chemotherapeutic agents is limited.
The humanized monoclonal antibody bevacizumab that targets angiogenesis by binding to VEGF has shown significant single agent activity in patients suffering from recurrent ovarian cancer. Studies have shown response rates between 16 and 18% in platinum-sensitive and platinum-refractory disease (10,13). Despite the improvement of progression-free survival in these trials, bevacizumab appears to improve the control of ascites. These data are in line with the observation of elevated levels of VEGF in malignant ascites (6,14).
Evidence suggests that angiogenesis is important in the prognosis and progression of epithelial ovarian cancer. Several studies have demonstrated a direct association of microvessel density and VEGF in primary tumours with extent of disease and time to progression (15). In addition, VEGF has relatively low levels of expression in normal tissue (16) and is therefore considered attractive as a potential target for an anti-VEGF agent such as bevacizumab. Preclinical models have demonstrated inhibition of tumour growth in various tumour cell types (17). Several alterations in vascular function, such as decreases in vessel density, diameter and permeability, result in a decreased production of malignant ascites. Consequently, various approaches have been examined to target VEGF and VEGF receptors (18), with several studies having demonstrated a reduction of ascites and tumour growth inhibition following the administration of bevacizumab in various ovarian cancer animal models (19).
Findings of a recently published study have shown prolonged survival following maintenance therapy in an intraperitoneal xenograft mouse model of previously untreated ovarian cancer (20). This study showed a significant inhibition of tumour growth following administration of 5 mg/kg bevacizumab as a single agent. In addition, bevacizumab significantly enhanced sensitivity to cisplatin. It may be speculated that enhancement of drug-induced tumour cell apoptosis by VEGF immunoneutralisation is the molecular basis which has been shown recently for cisplatin-treated ovarian cancer (21).
Another study has recently demonstrated rapid absorption of bevacizumab following i.p. administration, with a bioavailability of 92.8% (23). The authors reported a rapid and near complete absorption of bevacizumab following i.p. dosing, as well as increased animal survival following combined therapy with bevacizumab and i.p. bevacizumab and i.p. paclitaxel. As yet, little is known about the effect of bevacizumab on malignant ascites in pre-treated ovarian cancer patients. To the best of our knowledge, the present study is the first that evaluated i.v. as well as i.p. administration of bevacizumab in a xenograft model of cisplatin pre-treated ovarian cancer. Of note, we only administered a single dose of bevacizumab, paclitaxel or the combination therapy. We demonstrated significantly prolonged survival and reduction of ascites. Moreover, the addition of bevacizumab to a common cytostatic therapy such as paclitaxel resulted in further improvement in ascites reduction and survival. These findings are in agreement with preclinical and clinical studies which demonstrated that the efficacy of i.p. chemotherapy is limited by poor penetration of drugs into peritoneal tumours (23). It has been hypothesized, that adjunct therapy with anti-angiogenic agents may lead to decreased drug removal from peritoneal tumours, increased drug concentration in tumours and increased efficacy of i.p. chemotherapy (21). In a recently published report, four heavily pre-treated ovarian cancer patients received bevacizumab intravenously at a dose of 15 mg/kg every 3 weeks (9,17). Ascites were significantly reduced in all patients and no patient required a paracentesis following the initiation of therapy. Only a few patients with malignant ascites have received i.p. bevacizumab (4). Nine patients with malignant ascites from colon, breast, uterine and ovarian cancer, for whom repeated paracentesis prior to treatment was required, were administered intraperitoneal bevacizumab at 5 mg/kg monthly. Eight out of nine patients (89%) experienced prolongation of the interval between paracentesis (23). However, the i.p. route of administration in a palliative patient population needs to be examined stringently regarding the safety of the treatment. Intraperitoneal application may be the route of choice in this particular setting as it allows for the accumulation of the study drug within the body compartment where malignant ascites is promoted by VEGF secretion.
The effects of bevacizumab, either alone or in combination with i.p. paclitaxel, were not found to be significantly dependent on the route of administration. We conclude from this study that treatment of malignant ascites with i.p. bevacizumab is an effective and powerful alternative to i.v. treatment.