Antitumor Efficacy of Oblimersen Bcl-2 Antisense Oligonucleotide Alone and in Combination with Vinorelbine in Xenograft Models of Human Non–Small Cell Lung Cancer

7662 Vol. 10, 7662–7670, November 15, 2004 Clinical Cancer Research

YanPing Hu,1 Gwyn Bebb,2,3 Sophia Tan,2 These data suggest that the combination of oblimersen and Rebecca Ng,2 Hong Yan,2 Jason R. Sartor,2 VNB may provide enhanced antitumor activities against non– Lawrence D. Mayer,2,4,5 and Marcel B. Bally2,5,6 small-cell lung cancer. 1Genzyme Corporation, Framingham, Massachusetts; 2Department of Advanced Therapeutics and 3Systemic Therapy Program, British INTRODUCTION Columbia Cancer Agency, Vancouver, British Columbia, Canada; Lung cancer causes almost a third (ϳ30%) of all cancer- 4Celator Technologies, Vancouver, British Columbia, Canada; related deaths, a death toll that is greater than that attributed to 5 Faculty of Pharmaceutical Sciences, University of British Columbia, colorectal, breast, and prostate cancer (the second through fourth Vancouver, British Columbia, Canada; and 6Department of Pathology, University of British Columbia, Vancouver, British Columbia, Canada leading causes of cancer mortality) combined (1). Treating lung cancer is a frustrating process because improvements in overall outcome remain elusive, despite the introduction of molecularly ABSTRACT targeted therapy. A majority of patients present with advanced, Overexpression of Bcl-2 protein in cancer cells can inhibit incurable disease (stages IIIB and IV) and have a very poor out- programmed cell death and engender chemoresistance. Reduc- come. In this setting, the median survival with best supportive care ing Bcl-2 protein levels by using antisense oligonucleotides is 4 to 6 months, with Ͻ10% of patients still alive at 1 year (2). targeting the gene message can increase the sensitivity of can- Platinum-based doublet chemotherapy can extend the median sur- cer cells to cytotoxic agents. The objective of this work was to vival to 8 to 11 months and the 1-year survival to 30% (3). investigate the antitumor efficacy of the Bcl-2 antisense oligo- Changing the natural course of advanced or metastatic non–small- nucleotide oblimersen (Genasense; G3139), alone and in com- cell lung cancer (NSCLC) is thus proving very difficult. However, bination with vinorelbine (VNB), in an ectopic and orthotopic our increased understanding of the molecular anomalies associated xenograft model of NCI-H460 human non–small-cell lung can- with NSCLC has raised hopes that targeted therapies, especially if cer. In addition to assessing therapeutic effect, Bcl-2 protein used in combination with standard cytotoxic agents, may lead to expression in tumor tissue isolated from lung and heart was improved outcomes (4). Recent data, for example, have demon- measured. In the ectopic xenograft model, oblimersen at 5 and strated the efficacy of anti-epidermal growth factor receptor– 10 mg/kg significantly inhibited tumor growth compared with targeted therapies (5). In this study, we have initiated preclinical saline-treated control groups, and furthermore, the antitumor drug combination studies with another molecularly targeted thera- effect of oblimersen was associated with down-regulation of peutic agent, oblimersen, which effectively reduces expression of Bcl-2 protein in isolated tumor tissue. Moreover, the combina- the antiapoptotic molecule Bcl-2 (6). tion of oblimersen with VNB was more active in inhibiting Since confirmation of its antiapoptosis role in follicular tumor growth than either drug used alone. In the orthotopic lymphoma, the role of Bcl-2 in other malignancies has been model, oblimersen treatment (5 mg/kg) increased the median extensively explored. Initially, dysregulation and overexpres- survival time of mice to 33 days in comparison with a median sion of Bcl-2 in follicular lymphoma was found to be the result survival time of 21 days in the control animals. With this of the t(14;18)(q32;q31) chromosomal translocation character- model, the anticancer effect was demonstrated by assessing istic of the disease (7). Subsequent research has demonstrated tumor growth in lung and heart tissues by hematoxylin and Bcl-2 overexpression in other lymphoid malignancies as well as in eosin staining and Bcl-2 expression by immunohistochemistry. solid tumors that do not exhibit the translocation. It is believed that When VNB at 5 mg/kg was combined with oblimersen admin- the attenuation of apoptosis by overexpression of Bcl-2 is a con- istered at 5 mg/kg, 33% of mice survived more than 90 days. tributing factor associated with malignant transformation and resistance to chemotherapy-induced apoptosis (8–10). Although Bcl-2 overexpression is seen more commonly in small cell lung cancer, it is also seen in NSCLC tumors. Recent Received 5/28/04; revised 8/3/04; accepted 8/11/04. experience at our institution revealed that about 15% of non–small- Grant support: National Cancer Institute of Canada and the Canadian cell tumors overexpressed Bcl-2 by immunohistochemistry (11). Institutes of Health Research. Several studies have examined the role of Bcl-2 expression as a The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked predictive or prognostic marker in lung cancer, but the impact of advertisement in accordance with 18 U.S.C. Section 1734 solely to Bcl-2 expression on outcome has not been shown to be as signif- indicate this fact. icant as that of p53 expression (12). Importantly, Bcl-2 overexpres- Requests for reprints: Marcel B. Bally, Department of Advanced sion in these malignancies is not associated with a specific chro- Therapeutics, British Columbia Cancer Agency, 601 West 10th Avenue, Vancouver, British Columbia, V5Z 4E6 Canada. Phone: 604-877-6098, mosomal translocation and therefore is most likely a consequence ext. 3191; Fax: 604-877-6011; E-mail: [email protected]. of the general molecular dysregulation associated with lung carci- ©2004 American Association for Cancer Research. nogenesis, possibly caused by mechanisms such as changes in

promoter methylation status (13). It is believed that overexpression the British Columbia Cancer Agency Joint Animal Facility breed- of Bcl-2 will impart a survival advantage to cells in the face of ing colony and maintained in a pathogen-free environment. Two treatment. Furthermore, there is some evidence to suggest that NCI-H460 models were used in these studies: A solid tumor model down-regulation of Bcl-2 may increase sensitivity to cytotoxic was developed for antitumor efficacy studies by subcutaneous chemotherapy and radiation (14). Consequently, therapeutic strat- implantation of 106 NCI-H460 cells in the flank of SCID-RAG2 egies that can directly or indirectly inhibit or suppress Bcl-2 are mice; and an “orthotopic” NCI-H460 lung tumor model was es- garnering great clinical interest. tablished to assess the effects of treatment on systemic disease (106 Antisense oligonucleotides (ASOs) are one category of NCI-H460 cells were implanted through the chest wall into the left agents capable of affecting Bcl-2 expression. These short se- pleural space of SCID-RAG2 mice in a volume of 100 ␮L using a quences of nucleotide bases complementary to coding sequences of 26-gauge needle). The depth of needle penetration through the a gene of interest can be designed to bind RNA molecules in a intercostal muscles was controlled to avoid lung injury and hem- sequence-specific manner (6, 15–17). Binding can directly impair orrhage into the pleural space. All animal protocols were approved interaction with factors in the cytoplasm that are required for by the University of British Columbia Animal Care Committee, translation of the message into protein and recruit endogenous and all studies were completed in accordance with the current RNase H to cleave the RNA backbone (18). The latter results in a guidelines of the Canadian Council on Animal Care. reduction in the targeted RNA pool, which subsequently leads to In vivo Antitumor Activity. Efficacy experiments were reductions in the target protein encoded by that RNA. For ASOs to conducted in male SCID-RAG2M mice bearing established (0.05 3 be active, delivery to the inside of a target cell must be achieved cm ) NCI-H460 tumors (6 mice per group). Day 1 was defined as efficiently and, in tissue culture, for example, this requires associ- the day treatments started, typically 12 to 14 days after tumor cell ation with a delivery system such as cationic lipids (19, 20). In vivo, inoculation. Oblimersen or reverse polarity control (RPO) was however, ASOs are active when given in free form, possibly due to administered intraperitoneally daily at 5 or 10 mg/kg for 5 doses interactions with plasma lipoproteins (21). per week (Monday to Friday) over a 3-week period. Saline was Anti-Bcl-2–based treatment using oblimersen, an 18-mer ol- administered using the same schedule for control mice. VNB (5 or igonucleotide sequence targeting the first six codons of the open 10 mg/kg) was administered intravenously every 4th day ϫ 3 (days reading frame of the Bcl-2 message, is being tested in phase III 1, 5, and 9) via a lateral tail vein, either alone or in combination clinical trials as a chemosensitizing agent in melanoma and other with ASO (administered as described above). The dose schedules tumors (22, 23). It has been shown previously that Bcl-2 ASOs used were based on previous reports demonstrating that phospho- have an inhibitory effect on the growth of Bcl-2–overexpressing rothioate ASOs have elimination half-lives of 12 to 72 hours (27). lung cancer cell lines (24, 25), and clinical studies evaluating the When ASOs were administered in combination with VNB, the use of Bcl-2 ASOs in patients with lung cancer have been reported intravenous drug treatment was staggered 4 hours after intraperi- (26). Here, we present data demonstrating inhibition of tumor toneal ASO treatment. growth, anticancer effect, and prolongation of survival of tumor- Mice were observed daily, and body weight measurements bearing animals by oblimersen in two murine xenograft models of and signs of stress (e.g., lethargy, ruffled coat, and ataxia) were NSCLC: a subcutaneous model and an orthotopic model. Further- used to detect toxicities. Animals with ulcerated tumors or whose 3 more, tumor inhibition is shown to be associated with down- tumors exceeded 1 cm were euthanized. Electronic caliper (Mitu- regulation of Bcl-2 protein in tumor tissue. In addition, we dem- toyo Corp., Kawasaki, Japan) measurements of tumors were con- 3 onstrate greater antitumor effects when oblimersen was given in verted into mean tumor size (cm ) using the following formula: 2 combination with vinorelbine, a chemotherapeutic agent used in the 1⁄2[length (cm)] ϫ [width (cm)] . An average tumor size per mouse treatment of NSCLC. was used to calculate the group mean tumor size Ϯ SE (n ϭ 6–12 mice) from two independent experiments. The effects of treatment on progression of systemic disease MATERIALS AND METHODS and survival time of mice were studied using the intrapleural Tumor Cell Lines and Drugs. NCI-H460 human xenograft model. Treatments were initiated 7 days after tumor cell NSCLC cells were obtained from the National Cancer Institute injection. Histologic analysis had demonstrated consistent involve- tumor repository and maintained in RPMI 1640 supplemented ment of lung and heart by tumor by this stage. Saline, oblimersen with 10% fetal calf serum at 37°C in a humidified atmosphere (Bcl-2 ASO), or Bcl-2 RPO sequence was administered intraperi- containing 5% CO2. Cells were used in exponential growth toneally daily at 5 or 10 mg/kg for 5 doses per week over a 3-week phase, up to a maximum of 10 in vitro passages. Phosphoro- regimen. VNB (2.5 and 5 mg/kg) was administered intravenously thioate 18-mer ASO with a sequence complementary to the first on days 7, 11, and 15, after tumor implant, via a lateral tail vein, six codons of the open reading frame of Bcl-2 mRNA (5Ј-tct- either alone or in combination with ASOs, which were adminis- ccc-agc-gtg-cgc-cat-3; 18-mer) was generously provided by tered intraperitoneally. When ASOs were administered in combi- Genta Inc. (Berkeley Heights, NJ). This was used as the ASO nation with VNB, the drug treatment was staggered 4 hours after (oblimersen; G3139) to Bcl-2. A reverse polarity phosphoro- intraperitoneal ASO treatment. Each treated group consisted of 6 to thioate sequence, 5Ј-tac-cgc-gtg-cga-ccc-tct-3Ј, was used as 12 mice, and control groups consisted of 26 mice, pooled from control (RPO; Genta Inc.). Vinorelbine (VNB; GSK Inc. Mis- independent experiments. All experiments designed to assess ther- sissauga, Ontario, Canada) was purchased from the British apeutic activity included control animals (which were treated with Columbia Cancer Agency Pharmacy. saline). Because the NCI-H460 model was highly robust and Tumor Models. Male severe combined immunodeficient reproducible, these controls exhibited very comparable growth (SCID)-RAG2 mice (7–9 weeks old; 23–26 g) were obtained from curves and survival times. For this reason, all controls were aver-

aged, and experimental groups were compared with these controls. Where experimental groups were repeated, the results of these studies were also averaged. In general, studies done in this labora- tory consist of smaller n (typically 6 mice per treatment arm), but the studies are repeated to demonstrate consistency of the results. When the studies are repeated, the sample size is considered to be 12. Animal health was checked daily, and antitumor activity was evaluated as follows: treated versus control % ϭ median survival time (MST) of treated group/MST of control group ϫ 100. Histologic Study. Cohesive NCI-H460 tumors were har- vested on the last day of treatment for the subcutaneous xenograft model. For the intrapleural xenograft model, lung and heart tissue was removed at day 7 after tumor implantation and the day of last treatment. All tumors and tissues were fixed in formalin and embedded in paraffin. Four-micrometer sections were stained with hematoxylin and eosin (H&E) and assessed for immunohistochemical expression of Bcl-2 (mouse mono- clonal antihuman Bcl-2; DAKO, Glostrup, Denmark) using the ABC peroxidase labeling procedure (DAKO). Microscopic Fig. 1 Antitumor efficacy of systemically delivered oblimersen (Bcl-2 evaluation and scoring were carried out by a pathologist blinded AS) and Bcl-2 RPO (Bcl-2 RP) on the growth of subcutaneous human to the treatment group. lung cancer (NCI-H460) xenografts in male SCID-RAG2M mice. The tumor inoculation, treatment schedules, and tumor measurement were as Statistical Analyses. An unpaired Student’s t test (para- described in Material and Methods. All data are expressed as mean Ϯ metric) was used to measure statistical significance between two SE (n ϭ 6 mice per group). The effects of 5- or 10-mg/kg doses of ASOs treatment groups. Multiple comparisons were done using one- on tumor growth were statistically significant, e.g., oblimersen (5 or 10 Ͻ Ͻ way analysis of variance and post hoc test that compared dif- mg/kg) versus control, P 0.0225 or P 0.0008, respectively, on day 13. Arrows, days of ASO administration. ferent treatment groups by the Scheffe´ test criteria (Statistica release 4.5; Soft Inc., Tulsa, OK). A comparison of the survival curves among all of the treated and control groups was per- formed with Kaplan-Meier survival analysis, which takes cen- tumors from animals treated with oblimersen showed an in- sored values into account (SPSS release 12.0; SPSS Inc., Chi- crease in the number of dead cells and exhibited regions of cago, IL). Data were considered significant for P of Ͻ0.05. necrosis compared with controls. Necrosis was identified morphologically by areas of cells characterized by amorphous shape and condensed nuclear material (Fig. 2A, right panel). Bcl-2 RESULTS protein expression in these tumor sections was reduced (Fig. 2B, Antitumor Efficacy of Oblimersen in the Subcutaneous right panel), compared with control groups. Tumor Model. The antitumor effects of oblimersen in SCID- Mice bearing established NCI-H460 tumors were treated RAG2M mice bearing established subcutaneous NCI-H460 human with a combination of oblimersen and VNB. Treatment was NSCLC tumors are summarized in Fig. 1. Saline-treated control initiated when tumor size was approximately 50 mm3. Fig. 3A tumors grew reproducibly from a measurable size of 0.05 cm3 to presents the in vivo efficacy results obtained after treatment with 0.60 cm3 within approximately 13 days (ϳ 25–27 days after tumor VNB given Q4D ϫ 3 at 5 and 10 mg/kg doses. The two cell inoculation). Comparing treatment effects measured on day 13 regimens induced significant tumor growth suppression (P Ͻ of this study indicated that oblimersen significantly (P Ͻ 0.05) 0.05) in a dose-dependent manner (Fig. 3A) without showing inhibited tumor growth. The therapeutic effects were dose depend- significant signs of undesirable toxicity (i.e., body weight loss). ent; the mean tumor sizes on day 13 were 0.60, 0.27, and 0.16 cm3 On day 13, mean tumor sizes were 0.60, 0.28, and 0.10 cm3 in in saline and 5 and 10 mg/kg oblimersen treatment groups, respec- saline and 5 and 10 mg/kg VNB treatment groups, respectively tively. In contrast, administration of Bcl-2 RPO at 10 mg/kg pro- (Fig. 3A). The antitumor activities of VNB were partially asso- vided no therapeutic activity. None of the mice treated with ciated with down-regulation of Bcl-2 protein expression in these oblimersen displayed any significant signs of toxicity. tumors, as judged by immunohistochemistry (data not shown). We assessed the consequences of oblimersen treatment by When oblimersen treatment was combined with VNB (5 examining tumor tissue by histologic (H&E staining) and im- mg/kg), an even more pronounced delay of NCI-H460 tumor munohistochemical analysis of Bcl-2 expression in sections growth was observed compared with either treatment admin- derived from tumors obtained on the last day of treatment (day istered alone (Fig. 3B). On day 13, mean tumor sizes were 19). Tumors from mice treated with 10 mg/kg oblimersen, Bcl-2 0.60, 0.16, and 0.11 cm3 in groups treated with saline alone RPO, or saline were evaluated, and representative sections are or with 5 or 10 mg/kg oblimersen in combination with VNB shown in Fig. 2. Tumors from control (saline-treated) mice or (5 mg/kg), respectively. The mean tumor sizes in mice treated RPO-treated mice were composed of densely packed, viable with 5 mg/kg VNB in the presence or absence of oblimersen tumor cells with a very small number of dead cells (Fig. 2A, left (5 and 10 mg/kg) on day 17 were also compared (Fig. 3B). and middle panels). The tumors also exhibited extensive Bcl-2 The tumor sizes in the group of combined VNB and 5 or 10 protein expression (Fig. 2B, left and middle panels). In contrast, mg/kg oblimersen were 0.28 and 0.15 cm3, respectively,

Fig. 2 Histopathology of NCI-H460 tumors implanted in SCID-RAG2M mice after 10 mg/kg systemic dosing of oblimersen (Bcl-2 AS; right panels), RPO control sequence (Bcl-2 RP; middle panels), or saline control (left panels) for 19 days. A. Tumor sections were stained with H&E. Representative tumor pictographs are shown (taken with a ϫ10 objective). B. Bcl-2 protein expression in tumor sections stained using immunohistochemistry; representative tumor pictographs are shown (taken with a ϫ40 objective).

which were significantly (P Ͻ 0.05) smaller than the tumor sensitivity of the NSCLC tumor model to the antitumor size of 0.42 cm3 in the groups treated with 5 mg/kg VNB effects of oblimersen. These mice did not show any signifi- alone (Fig. 3B). Bcl-2 RPO did not enhance the activity of cant signs of overt toxicity, and weight loss was not increased VNB. These data suggest one of two possibilities: (a) by combination with oblimersen. However, in all treatment Oblimersen increased the chemosensitivity of the NSCLC groups, tumor growth persisted after an initial response, subcutaneous tumor model to VNB; or (b) VNB increased the despite concurrent oblimersen and/or VNB treatment.

Fig. 3 Treatment of NCI-H460 tumors xenografts in SCID-RAG2M mice. Methods of tumor inoculation, treatment schedules, and tumor measurement were as described in Material and Methods. A, antitumor efficacy of VNB (5 or 10 mg/kg). VNB treatment resulted in significantly lower tumor weights on day 13 compared with controls (5 mg/kg VNB, P Ͻ 0.03; 10 mg/kg VNB, P Ͻ 0.0018). B, antitumor efficacy of VNB (5 mg/kg) or oblimersen (5 or 10 mg/kg) alone and in combination. The combination of VNB (5 mg/kg) ϩ oblimersen (5 or 10 mg/kg) was significantly (P Ͻ 0.05) better than VNB alone when comparisons were made based on tumor sizes determined on day 13 and 17. All data are expressed as means Ϯ SE (n ϭ 6 mice per group). Arrows, days of VNB or oblimersen ϩ VNB administration.

Antitumor Activity of Oblimersen and Vinorelbine in heart tissues from control mice or RPO-treated controls exhibited the Orthotopic NCI-H460 Model. After inoculation of 106 obvious Bcl-2 protein expression (see Fig. 5B, left and middle NCI-H460 tumor cells into the pleural cavity, numerous tumor panels). In contrast, the lung and heart tissue from oblimersen- nodules were observed on the inner surface of ribs, diaphragm, treated mice showed a complete lack of Bcl-2 protein expression lung, and heart within 1 week. Local invasion of lung and heart (Fig. 5B, right panel), suggesting that the oblimersen antitumor tissues by tumor cells was observed in 100% of mice and effect was mediated through down-regulation of Bcl-2 protein confirmed by histologic analysis on day 7 (data not shown). expression. Positive-staining leukocytes within tissue provided When these animals were treated with oblimersen given daily at positive internal controls of Bcl-2 expression. These cells, normal- 5 and 10 mg/kg for 15 doses, a significant therapeutic response benign bronchial columnar cells and light staining striated muscle, was observed (Fig. 4). In comparison with untreated animals, were not included in the Bcl-2 score. oblimersen treatment at 5 and 10 mg/kg significantly increased To determine whether improved therapeutic effects could the MST of these NCI-H460 tumor-bearing mice from 21 days be provided when oblimersen was used in combination with to 33 and 36 days, respectively (Table 1). Nonetheless, no VNB in the NCI-H460 orthotopic model, the therapeutic effect long-term survivors were observed. In contrast to oblimersen- of oblimersen at 5 mg/kg with VNB (2.5 and 5 mg/kg) was treated animals, however, administration of Bcl-2 RPO at 10 compared with the effect of the agents individually. In the mg/kg provided no therapeutic activity (Fig. 4). Saline- and absence of oblimersen, VNB treatment at 2.5 and 5 mg/kg Bcl-2 RPO-treated control animals were terminated or died of significantly increased the MST of NCI-H460 tumor bearing progressive tumor growth by day 29 after tumor cell inoculation mice from 21 days to 24 and 35 days, respectively (Table 2). with a MST of 21 days (Table 1). The combination of oblimersen at 5 mg/kg with VNB (2.5 and An evaluation of H&E-stained sections of lung and heart of 5 mg/kg) produced an even more pronounced increase in MST these control animals demonstrated that tumor cells intensively (Table 2). In addition, more long-term survivors were observed invaded these tissues (Fig. 5A, left and middle panels). In contrast, among animals treated with the combination (Fig. 6). The com- these tissues exhibited little tumor growth and reduced invasion in bined treatment resulted in 17% and 33% long-term survivors animals given oblimersen treatment at 10 mg/kg (Fig. 5A, right (Ͼ90 days) and dose-dependent therapeutic activity when VNB panel). As noted in the methodology, all H&E-stained sections was administered at 2.5 and 5 mg/kg in combination with from groups of six control and treated mice were scored qualita- oblimersen at a dose of 5 mg/kg (every day ϫ 5 for 3 weeks). tively by a pathologist blinded to the treatment groups. Results In contrast, administration of Bcl-2 RPO at 5 mg/kg with VNB showed limited disease in the lung and no detectable disease in the at 5 mg/kg did not provide any increase in MST compared with heart. In addition, Bcl-2 protein expression was assessed by immu- VNB treatment alone (Table 2). It is worth noting that pathological nohistochemical methods. The results demonstrated that lung and evaluation at necropsy was completed in all long-term survivors, and those animals showed no evidence of tumor either on gross inspection or on histologic examination of tissues. Selected mice that survived as a consequence of treatment received reinoculations of 1 ϫ 106 NCI-H460 cells to assess whether the SCID animals had developed an immune response to the inoculated tumor cells. These animals died within 28 days.

DISCUSSION Most anticancer drugs induce activation of apoptotic pathways as part of their cytotoxic activity (28). Relevant progress has been made in understanding the events underlying drug- induced apoptosis, and the Bcl-2 family of proteins plays a crucial role in regulating this process (29, 30). Overexpression of the antiapoptotic molecule Bcl-2 can engender resistance to anticancer drugs (29, 31), and, conversely, use of oblimersen to reduce Bcl-2 levels has been proven to be directly cytotoxic and to enhance chemosensitivity. These effects have been docu- mented in vitro for morphologically distinct NSCLC cell lines (14, 25), small-cell lung cancer cell lines (32), and cell lines derived from other tumor types (6). Several studies on the in vivo biological effects of oblimersen have also been reported, and only a few of these are referenced here (22, 32–36). In this study, the therapeutic effects of oblimersen, alone and in com- Fig. 4 Survival curves of SCID mice bearing an intrapleural inocula- bination with VNB, are described using a subcutaneous and an 6 tion of 1 ϫ 10 NCI-H460 cells and treated with saline, control ASO orthotopic NSCLC tumor model. These results are correlated (Bcl-2 RP), or oblimersen (Bcl-2 AS; 5 or 10 mg/kg). When comparing MSTs of oblimersen-treated groups versus saline- or RPO-treated con- with ASO-mediated suppression of Bcl-2 protein in tumor tis- trols, the differences were statistically significant (P Ͻ 0.00001). Ar- sue. The studies presented suggest that the combination appears rows, days of ASO administration. to be more effective than either individual agent used alone.

Table 1 Effect of saline, RPO control sequence, or oblimersen (Bcl-2 ASO) treatment on SCID-RAG2M male mice bearing NCI-H460 cells inoculated intrapleurally Treatment Dose (mg/kg) Schedule n MST (days) % 90-day survival T/C % * P † None NA NA 26 21 0 100 NA RPO 10 QD 8 21.5 0 102 Ͼ0.05 ‡ ASO 5 QD 7 33 0 157 Ͻ0.00001 ASO 10 QD 9 36 0 171 Ͻ0.00001 Abbreviations: MST, median survival time; NA, not applicable; QD, daily treatment for 15 days, excluding weekends. * T/C (%) ϭ MST of treated mice/MST of untreated mice ϫ 100%. † P values were derived by using the log-rank test comparing each group versus control. ‡ Not significant when compared with control animals.

The immunohistochemical results data presented here dem- sequence control Bcl-2 RPO. Importantly, the RPO sequence onstrate that daily intraperitoneal oblimersen (QD ϫ 5 days for used did contain the immunostimulatory CpG motifs, suggest- 3 weeks) is capable of achieving qualitatively substantial down- ing that the activity seen in this NCI-H460 model was not due regulation of Bcl-2 expression in the subcutaneous solid tumor to nonspecific immune stimulation. Histologic assessment of model as well as in the orthotopic model, where cells grew in the tumors derived from mice treated with oblimersen demonstrates lung and heart (Figs. 2B and 5B). Furthermore, treatment was significant levels of tumor cell death and suggests that the also associated with significant antitumor activity. The effects decrease in tumor size and increase in survival of animals appeared to be specific for oblimersen because no therapeutic treated with oblimersen were caused by tumor cell kill associ- response was observed in animals treated with the reverse ated with down-regulation of Bcl-2 protein (Figs. 1, 2, 4, and 5).

Fig. 5 Histopathology of lung and heart tissues from SCID- RAG2M mice that received intrapleural inoculation with 1 ϫ 106 NCI-H460 cells and were treated with oblimersen (Bcl-2 ASO), RPO control sequence, or saline control. A. Lung (top panels) and heart (bottom panels) sections were stained with H&E. Representative pictographs are shown (taken with a ϫ10 objective). B. Bcl-2 protein expression in lung (top panels) and heart (bottom panels) sections stained using immunohistochemistry; representative tissue pictographs are shown (taken with a ϫ40 objective). Saline-treated control (left panels), Bcl-2 RPO sequence treatment (middle panels), and oblimersen Bcl-2 ASO treatment (right panels).

Table 2 Treatment of SCID-RAG2M male mice that received inoculations of 106 NCI-H460 cells in pleural cavity Treatment Dose (mg/kg) Schedule n MST (days) % 90-day survival T/C% * P † Saline NA NA 26 21 0 100 NA VNB 2.5 Days 7, 11, 15 6 24 0 114 0.0051 VNB 5 Days 7, 11, 15 6 35 0 167 Ͻ0.00001 RPO ϩ VNB 5 ϩ 5QDϩ days 7, 11, 15 6 35 0 167 Ͼ0.05 ‡ ASO ϩ VNB 5 ϩ 2.5 QD ϩ days 7, 11, 15 6 37 17 154 0.019 ASO ϩ VNB 5 ϩ 5QDϩ days 7, 11, 15 6 62 33 177 0.0005 NOTE. Animals were treated with saline control, VNB alone, or in combination with RPO control sequence or oblimersen (Bcl-2 ASO). VNB was injected intravenously for three doses after tumor inoculation, and ASOs were given intraperitoneally. Abbreviations: MST, median survival time; NA, not applicable; QD, daily treatment for 15 days, excluding weekends. * T/C (%) ϭ MST of treated mice/MST of untreated mice ϫ 100%. † P values were derived by using the log-rank test comparing each group versus the corresponding VNB concentration and schedule. Animals treated with VNB alone were compared with the control group. ‡ Not significant when compared with the corresponding VNB alone concentration and schedule.

As stated above, some ASOs containing unmethylated CpG Madagascan periwinkle leaves, functions by interfering with motifs within their sequence have been shown to be potent immune microtubule assembly (45). It appears to have preferential se- stimulators (37, 38). Our results suggest that the therapeutic effects lectivity for mitotic microtubules as opposed to axonal micro- obtained were specific for the oblimersen sequence, confirming the tubules (46). VNB is approved for use as a single agent and in observations by Klasa et al. (35), which likewise showed specific combination with cisplatin for patients with stage IV NSCLC, efficacy due to the G3139 molecule and not Bcl-2 RPO in perforin- where it is associated with a 15% to 25% and a 30% to 35% knockout (natural killer-deficient) SCID mice bearing systemic overall response rate, respectively (47). Moreover, its acceptable follicular lymphoma. In combination, these results argue against side effect profile as a single agent makes it suitable for patients possible immunostimulatory action, which is dependent on func- of poorer performance status (Eastern Cooperative Oncology tional natural killer activity for oblimersen used here. Furthermore, Group performance status ϭ 2; ref. 48). The combination of the others have shown that a substitution of unmethylated CpG motifs anticancer agent VNB with oblimersen presents an additional within the oblimersen sequence with 5-methylation of cytosine therapeutic strategy for the treatment of this challenging disease. (Genta G4232) to abrogate immune stimulation resulted in an antitumor activity similar to that of the oblimersen (39). It should be noted that mice treated with oblimersen had significantly increased spleen weight relative to those treated with G4232 or saline, a phenomenon that has been associated with immune stimulation (40). We have observed that when oblimersen was given to non–tumor-bearing and tumor-bearing SCID mice, spleen weight increased significantly in both groups. Spleen weight in both groups returned to normal within 1 month after the end of treatment (data not shown). It is important to note that this oblimersen- associated splenic hyperplasia may be specific to murine species because it was not observed in primates and has not been witnessed in the clinical setting in humans. It is also important to note that the RPO ASO control sequence used in these studies caused increases in spleen weight that were not significantly different than those seen in the oblimersen-treated group. Although spleen weight increases are a very crude indication of immune stimulation, by this measurement both sequences used were immunostimulatory, yet only oblimersen exhibited significant therapeutic effects. The results presented do not exclude the possibility that some of the therapeutic effects of oblimersen are due to immune stimulation. Perhaps more importantly, it is acknowledged here that the therapeutic potential of oblimersen is enhanced due to the fact that it likely mediates anticancer activity through several mechanisms. Fig. 6 Survival curves of SCID mice bearing an intrapleural inocula- As shown here and by others (34, 41–43), oblimersen effectively tion of 1 ϫ 106 NCI-H460 cells and treated with oblimersen ASO (5 inhibits expression of Bcl-2. It is accepted that the oligonucleotide mg/kg) and VNB (2.5 and 5 mg/kg), alone and in combination. When sequence used can stimulate the immune system in a manner that comparing MSTs of the combination of oblimersen (5 mg/kg) and VNB can indirectly result in antitumor effects. It is also possible that the (2.5 mg/kg) versus VNB (2.5 mg/kg), the differences were significant (P Ͻ 0.019); when comparing MSTs of the combination of oblimersen immune effects engendered by administration of oblimersen may ASO (5 mg/kg) and VNB (5 mg/kg) versus VNB (5 mg/kg), the directly influence induction of apoptosis (44). differences were significant (P Ͻ 0.0005). Arrows, days of oblimersen VNB, a Vinca alkaloid antineoplastic agent derived from ASO and VNB administration.

It is interesting to note that treatment with VNB did cause an only to minimize unwarranted side effects but also to reduce the apparent down-regulation of Bcl-2, as judged by immunohisto- costs incurred by health care systems. This study did not address chemistry. This was observed in both subcutaneous and in- the issue of correlating response to oblimersen to the level of trapleural tumor models. There is no evidence that Bcl-2 ex- Bcl-2 protein or mRNA. The NCI-H460 NSCLC cell line is pression is a predictor of response to VNB in the clinical setting known to be a express high levels of Bcl-2 and may not be (49). Despite this, there is indirect evidence associating VNB typical of all NSCLC tumors (4). However, it has been shown with altered apoptosis signal transduction pathways (50), and it that anti Bcl-2 strategies are effective in cells that express has been suggested that VNB induces apoptosis via Bcl-2 phos- relatively low amounts of Bcl-2 protein (33). With this in mind, phorylation in MCF7 cells (51). The precise means by which further work is required to assess potential predictive molecular VNB induces a decrease in Bcl-2 protein expression in this markers of response to oblimersen. model warrants further study, but the observation suggests that The data presented here are the first to address the potential oblimersen and VNB act at least in an additive manner on a role of ASO directed at the Bcl-2 gene message alone in en- Bcl-2–dependent apoptotic pathway in bringing about cell hancing the therapeutic efficacy of VNB in both models of death. To better define the type of interactions these two agents NSCLC. This study successfully used this strategy to increase exhibit, further dose titration information is required. As noted the number of long-term survivors (33%) when mice were by others, dose effects in vitro and in vivo are often nonlinear treated at a point when they had systemically distributed disease. and a dose that is synergistic (or superadditive) at one affect Moreover, the data suggest that improved clinical outcomes level may not be synergistic at another dose level. The only way could be achieved with standard or even lower doses of anti- to gain this understanding is through dose-effect curves, and it cancer drugs when combined with oblimersen, potentially im- is critical to identify combinations that appear to be additive or pacting overall clinical tolerance and costs of care. Further synergistic at maximum effect levels. investigations, possibly leading to clinical trials of these drugs in The data presented here add weight to the assertion that combination in metastatic NSCLC, are warranted. anticancer drugs are more effective in combination than as single agents. Traditional anticancer treatment protocols contain mixtures ACKNOWLEDGMENTS of several non–cross-resistant cytotoxic agents (52), but their ap- We thank Dana Masin and Natashia Harasym (BC Cancer Agency, plication is limited by broad toxicities, and their use is precluded in Vancouver, BC, Canada) for their excellent technical assistance and patients with poor performance status. The possibility of combining efforts in performing the in vivo studies. cytotoxic chemotherapy with more rational molecularly targeted treatment has excited medical oncologists and cancer biologists REFERENCES alike, but attempts at doing so in clinical trials have shown incon- 1. World Health Organization. The World Health Report Statistical sistent results. Most recently, promising results have been obtained Annex. Health systems: improving performance; Geneva, Switzerland, using a combination of oblimersen with dacarbazine in the setting World Health Organization 2000. p. 164–9. of metastatic melanoma (23), where the response rate and progres- 2. Smith W, Khuri FR. The care of the lung cancer patient in the 21st sion-free survival were significantly longer in the G3139 treatment century: a new age. Semin Oncol 2004;31:11–5. arm with an early trend to improved survival (53). The data 3. Scagliotti GV, De Marinis F, Rinaldi M, et al. Phase III randomized trial comparing three platinum-based doublets in advanced non-small- presented here suggest that such a combination can likewise be cell lung cancer. J Clin Oncol 2002;20:4285–91. effective in the setting of metastatic NSCLC, a malignancy also 4. Danesi R, de Braud F, Fogli S, et al. Pharmacogenetics of anticancer associated with a poor outcome. Currently, single-agent VNB in drug sensitivity in non-small cell lung cancer. Pharmacol Rev 2003;55: the NSCLC setting is reserved for the more elderly and frail patient 57–103. deemed not to be a candidate for “doublet” chemotherapy. The 5. Kris MG, Natale RB, Herbst RS, et al. Efficacy of gefitinib, an possibility of improving response rate and prolonging survival in inhibitor of the epidermal growth factor receptor tyrosine kinase, in symptomatic patients with non-small cell lung cancer: a randomized such patients without incurring significant additional side effects is trial. J Am Med Assoc 2003;290:2149–58. obviously attractive. 6. Klasa RJ, Gillum AM, Klem RE, Frankel SR. Oblimersen Bcl-2 To date, clinical trials involving oblimersen have used a antisense: facilitating apoptosis in anticancer treatment. Antisense Nu- schedule that predoses patients for several days with the ASO, cleic Acid Drug Dev 2002;12:193–213. occasionally using continuous infusion pumps, before adminis- 7. Bakhshi A, Jensen JP, Goldman P, et al. Cloning the chromosomal tering the cytotoxic drug. The intention is to cause sufficient breakpoint of t(14;18) human lymphomas: clustering around JH on down-regulation of Bcl-2 protein and lower the apoptotic chromosome 14 and near a transcriptional unit on 18. Cell 1985; 41(Suppl 2):899–906. threshold in the malignant cells and improve cell kill (6). In this 8. Hockenbery D, Nunez G, Milliman C, Schreiber RD, Korsmeyer SJ. study, no such pretreatment was given, and the drugs were Bcl-2 is an inner mitochondrial membrane protein that blocks pro- administered concurrently with impressive results in inhibiting grammed cell death. Nature (Lond) 1990;348:334–6. tumor growth and down-regulating Bcl-2 protein expression. 9. Hockenbery DM, Zutter M, Hickey W, Nahm M, Korsmeyer SJ. Whether this effect is specific only for this combination or BCL2 protein is topographically restricted in tissues characterized by suggests that oblimersen can lower the apoptotic threshold more apoptotic cell death. Proc Natl Acad Sci USA 1991;88:6961–5. rapidly than previously assumed is unclear. However, in design- 10. Nunez G, London L, Hockenbery D, et al. Deregulated Bcl-2 gene expression selectively prolongs survival of growth factor-deprived he- ing a schedule for clinical trials of this combination of drugs, it mopoietic cell lines. J Immunol 1990;144:3602–10. may be possible to dispense with the pretreatment protocol. 11. Bebb GAN, Cheang M, Tucker C, et al. Cyclin D1 is over- Predicting which tumors will respond to treatment has expressed in a high proportion of non small cell lung cancer. Lung always been an important goal in cancer therapeutic agents, not Cancer 2003;41:S258.

12. Laudanski J, Niklinska W, Burzykowski T, Chyczewski L, ing G3139) in PC3 prostate carcinoma cells are recapitulated at least in Niklinski J. Prognostic significance of p53 and bcl-2 abnormalities in part by treatment with interferon-beta and -gamma. Clin Cancer Res operable nonsmall cell lung cancer. Eur Respir J 2001;17:660–6. 2004;10:3678–88. 13. Nagatake M, Osada H, Kondo M, et al. Aberrant hypermethylation 35. Klasa RJ, Bally MB, Ng R, et al. Eradication of human non- at the bcl-2 locus at 18q21 in human lung cancers. Cancer Res 1996; Hodgkin’s lymphoma in SCID mice by BCL-2 antisense oligonucleo- 56:1886–91. tides combined with low-dose cyclophosphamide. Clin Cancer Res 14. Koty PP, Tyurina YY, Tyurin VA, Li SX, Kagan VE. Depletion of 2000;6:2492–500. Bcl-2 by an antisense oligonucleotide induces apoptosis accompanied 36. Raffo A, Lai JC, Stein CA. Antisense RNA down-regulation of bcl-2 by oxidation and externalization of phosphatidylserine in NCI-H226 expression in DU145 prostate cancer cells does not diminish the cytostatic lung carcinoma cells. Mol Cell Biochem 2002;234–235:125–33. effects of G3139 (oblimersen). Clin Cancer Res 2004;10:3195–206. 15. Stephenson ML, Zamecnik PC. Inhibition of Rous sarcoma viral 37. Weiner GJ, Liu HM, Wooldridge JE, Dahle CE, Krieg AM. Immu- RNA translation by a specific oligodeoxyribonucleotide. Proc Natl Acad nostimulatory oligodeoxynucleotides containing the CpG motif are ef- Sci USA 1978;75:285–8. fective as immune adjuvants in tumor antigen immunization. Proc Natl 16. Zamecnik PC, Stephenson ML. Inhibition of Rous sarcoma virus Acad Sci USA 1997;94:10833–7. replication and cell transformation by a specific oligodeoxynucleotide. 38. Ballas ZK, Rasmussen WL, Krieg AM. Induction of NK activity in Proc Natl Acad Sci USA 1978;75:280–4. murine and human cells by CpG motifs in oligodeoxynucleotides and 17. Cotter FE. Antisense therapy for lymphomas. Hematol Oncol 1997; bacterial DNA. J Immunol 1996;157:1840–5. 15:3–11. 39. Wacheck V, Krepler C, Strommer S, et al. Antitumor effect of 18. Wintersberger U. Ribonucleases H of retroviral and cellular origin. G3139 Bcl-2 antisense oligonucleotide is independent of its immune Pharmacol Ther 1990;48:259–80. stimulation by CpG motifs in SCID mice. Antisense Nucleic Acid Drug 19. Wong FMP, Macadam SA, Kim A, et al. Formation and delivery of Dev 2002;12:359–67. novel oligonucleotide/cationic lipid complexes. Proc Am Assoc Cancer 40. Wacheck V, Heere-Ress E, Halaschek-Wiener J, et al. Bcl-2 anti- Res 1998;39:646. sense oligonucleotides chemosensitize human gastric cancer in a SCID 20. Shoji Y, Nakashima H. Current status of delivery systems to improve mouse xenotransplantation model. J Mol Med 2001;79:587–93. target efficacy of oligonucleotides. Curr Pharm Des 2004;10:785–96. 41. Vollmer J, Weeratna R, Payette P, et al. Characterization of three 21. Wasan EK, Waterhouse D, Sivak O, et al. Plasma protein binding, CpG oligodeoxynucleotide classes with distinct immunostimulatory ac- lipoprotein distribution and uptake of free and lipid-associated BCL-2 tivities. Eur J Immunol 2004;34:251–62. antisense oligodeoxynucleotides (G3139) in human melanoma cells. Int 42. Cornelie S, Poulain-Godefroy O, Lund C, et al. Methylated CpG- J Pharm 2002;241:57–64. containing plasmid activates the immune system. Scand J Immunol 22. Jansen B, Schlagbauer-Wadl H, Brown BD, et al. bcl-2 antisense 2004;59:143–51. therapy chemosensitizes human melanoma in SCID mice. Nat Med 43. Lai JC, Benimetskaya L, Santella RM, et al. G3139 (oblimersen) 1998;4:232–4. may inhibit prostate cancer cell growth in a partially bis-CpG-dependent 23. Jansen B, Wacheck V, Heere-Ress E, et al. Chemosensitisation of non-antisense manner. Mol Cancer Ther 2003;2:1031–43. malignant melanoma by BCL2 antisense therapy. Lancet 2000;356: 44. Chen Q, Ray S, Hussein MA, Srkalovic G, Almasan A. Role of 1728–33. Apo2L/TRAIL and Bcl-2-family proteins in apoptosis of multiple my- 24. Koty PP, Zhang H, Levitt ML. Antisense bcl-2 treatment increases eloma. Leuk Lymphoma 2003;44:1209–14. programmed cell death in non-small cell lung cancer cell lines. Lung 45. Binet S, Fellous A, Lataste H, et al. In situ analysis of the action of Cancer 1999;23:115–27. Navelbine on various types of microtubules using immunofluorescence. 25. Ziegler A, Luedke GH, Fabbro D, et al. Induction of apoptosis in Semin Oncol 1989;16(Suppl 4):5–8. small-cell lung cancer cells by an antisense oligodeoxynucleotide tar- 46. Fellous A, Ohayon R, Vacassin T, et al. Biochemical effects of geting the Bcl-2 coding sequence. J Natl Cancer Inst (Bethesda) 1997; Navelbine on tubulin and associated proteins. Semin Oncol 1989; 89:1027–36. 16(Suppl 4):9–14. 26. Rudin CM, Otterson GA, Mauer AM, et al. A pilot trial of G3139, 47. Le Chevalier T, Brisgand D, Soria JC, et al. Long term analysis of a bcl-2 antisense oligonucleotide, and paclitaxel in patients with che- survival in the European randomized trial comparing vinorelbine/ morefractory small-cell lung cancer. Ann Oncol 2002;13:539–45. cisplatin to vindesine/cisplatin and vinorelbine alone in advanced non- 27. Agrawal S, Temsamani J, Tang JY. Pharmacokinetics, biodistribu- small cell lung cancer. Oncologist 2001;6:8–11. tion, and stability of oligodeoxynucleotide phosphorothioates in mice. 48. Rossi A, Gridelli C, Gebbia V, et al. Single agent vinorelbine as Proc Natl Acad Sci USA 1991;88:7595–9. first-line chemotherapy in elderly patients with advanced breast cancer. 28. Miyashita T, Reed JC. Bcl-2 oncoprotein blocks chemotherapy- Anticancer Res 2003;23:1657–64. induced apoptosis in a human leukemia cell line. Blood 1993;81:151–7. 49. Krug LM, Miller VA, Filippa DA, et al. Bcl-2 and bax expression 29. Reed JC. Bcl-2: prevention of apoptosis as a mechanism of drug in advanced non-small cell lung cancer: lack of correlation with chem- resistance. Hematol Oncol Clin N Am 1995;9:451–73. otherapy response or survival in patients treated with docetaxel plus 30. Kroemer G. The proto-oncogene Bcl-2 and its role in regulating vinorelbine. Lung Cancer 2003;39:139–43. apoptosis. Nat Med 1997;3:614–20. 50. Wang LG, Liu XM, Kreis W, Budman DR. The effect of antimi- 31. Reed JC, Miyashita T, Takayama S, et al. BCL-2 family proteins: crotubule agents on signal transduction pathways of apoptosis: a review. regulators of cell death involved in the pathogenesis of cancer and Cancer Chemother Pharmacol 1999;44:355–61. resistance to therapy. J Cell Biochem 1996;60:23–32. 51. Liu XM, Jiang JD, Ferrari AC, Budman DR, Wang LG. Differential 32. Zangemeister-Wittke U, Schenker T, Luedke GH, Stahel RA. Syn- effect of vinorelbine versus paclitaxel on ERK2 kinase activity during ergistic cytotoxicity of bcl-2 antisense oligodeoxynucleotides and eto- apoptosis in MCF-7 cells. Br J Cancer 2001;85:1403–11. poside, doxorubicin and cisplatin on small-cell lung cancer cell lines. 52. Goldie JH, Coldman AJ, Gudauskas, GA. Rationale for the use of Br J Cancer 1998;78:1035–42. alternating non-cross-resistant chemotherapy. Cancer Treat Rep 1982; 33. Chi KC, Wallis AE, Lee CH, et al. Effects of Bcl-2 modulation with 66:439–49. G3139 antisense oligonucleotide on human breast cancer cells are 53. Millward MJ, Bedikian AY, Conry RM, et al. Randomized multi- independent of inherent Bcl-2 protein expression. Breast Cancer Res national phase 3 trial of dacarbazine (DTIC) with or without Bcl-2 Treat 2000;63:199–212. antisense (oblimersen sodium) in patients (pts) with advanced malignant 34. Benimetskaya L, Wittenberger T, Stein CA, et al. Changes in gene melanoma (MM): analysis of long-term survival. Proc Am Soc Clin expression induced by phosphorothioate oligodeoxynucleotides (includ- Oncol 2004;22:711.

Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2004 American Association for Cancer Research. Antitumor Efficacy of Oblimersen Bcl-2 Antisense Oligonucleotide Alone and in Combination with Vinorelbine in Xenograft Models of Human Non−Small Cell Lung Cancer

YanPing Hu, Gwyn Bebb, Sophia Tan, et al.

Clin Cancer Res 2004;10:7662-7670.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/22/7662

Cited articles This article cites 50 articles, 17 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/22/7662.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/22/7662.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/10/22/7662. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.