Preclinical Evaluation of Sequential Combination of Oncolytic Adenovirus Delta-24-RGD and Phosphatidylserine-Targeting Antibody in Pancreatic Ductal Adenocarcinoma

This article is featured in Highlights of This Issue, p. 553

Pancreatic ductal adenocarcinoma (PDAC) remains one of the leading causes of cancer-related death (1). Although advances have been made, the overall 5-year survival rate of patients with PDAC remains less than 7% mainly due to late diagnosis and limited number of available systemic agents. New effective therapeutic agents for this cancer are still urgently needed.

Oncolytic virus are engineered to take advantage of genetic mutations of specific tumor suppressor genes or oncogenes in cancer cells to allow preferential replication within and lysis of cancer cells while sparing normal cells (2, 3). Multiple types of oncolytic viruses including Ad5/3 adenovirus, herpes simplex viruses (HSV), vaccinia virus, Newcastle disease virus, and measles virus have been evaluated clinically (3). The first FDA-approved HSV1-based oncolytic virus T-VEC carrying a granulocyte-macrophage colony-stimulating factor (GM-CSF) gene significantly improved durable response rate of patients with unresectable melanoma compared with GM-CSF treatment alone (16.3% vs. 2.1%, P < 0.0001; ref. 4). The approval of T-VEC for late-stage melanoma brings the hope in oncolytic virus–mediated immunotherapy for cancer treatment. Delta-24-RGD is an adenovirus-based oncolytic virus with a deletion of 24 basepairs in the E1A region and a modification in virus fiber with a RGD-4C motif to enhance its infection of cancer cells independent of the expression of coxsackievirus and adenovirus receptor (CAR; refs. 5, 6). Adenovirus E1A gene codes a 19-kDa protein that binds to RB protein, thus releasing E2F factor from RB/E2F complex for cell-cycle progression. The deletion of 24 basepairs in E1A region suppresses virus replication in normal cells but not in cancer cells with defect of p16/RB/E2F pathway. Delta-24-RGD has shown promising anticancer effect by stimulating anticancer immune response in brain tumor patients (7) and is currently in phase II clinical trials with combination of chemotherapy for brain tumor. Because p16/RB/E2F pathway is also frequently altered in pancreatic cancer due to the deletion, mutation, or promoter methylation of CDKN2A gene which encodes the p16 protein, we hypothesize that Delta-24-RGD could be an active agent for pancreatic cancer therapy, especially for the tumors with abnormal p16/RB/E2F pathway.

Phosphatidylserine, a membrane phospholipid, is localized in the inner leaflet of a plasma membrane in normal nontumorigenic cells but is presented on the surface of apoptotic cells and cancer cells within the tumor microenvironment (8, 9). Although a signal for cell engulfment, phosphatidylserine is known to dampen the immune response. Phosphatidylserine exposure on the outer membrane also occurs during viral cellular infection and replication. Monoclonal antibodies have been raised to target phosphatidylserine and investigated as anti-viral therapy (10). Recent data from an animal model of melanoma demonstrated that combining phosphatidylserine-targeting antibodies improved the effectiveness of immune checkpoint inhibitors, suggesting that antibodies to phosphatidylserine can reverse its immune dampening signals (11). Phosphatidylserine-targeting antibodies, bavituximab, have also been raised to target phosphatidylserine-expressing tumor cells and investigated in phase I clinical trials of several solid tumors systems including metastatic breast and lung cancers (12, 13). Together, these studies suggest that anti-phosphatidylserine antibodies could augment the anticancer effects of oncolytic virus therapy.

In this study, we evaluated the anticancer activity of Delta-24-RGD in multiple pancreatic cancer cell lines and primary pancreatic cells established from patient-derived xenograft tumors (PDX) and explored potential predictive biomarkers for sensitivity. We found that Delta-24-RGD induced dramatic cytotoxicity in a subset of pancreatic cancer cell lines with high expression of cyclin D1 and induced phosphatidylserine exposure in infected cells. In addition, combination with a phosphatidylserine-targeting antibody further enhanced the anticancer effects of Delta-24-RGD in vivo, possibly through stimulating immune responses. Our study demonstrates that Delta-24-RGD could be a promising agent, alone or in combination, for pancreatic cancer therapy.

Cell culture medium RPMI-1640 was purchased from HyClone. FBS and MTT cell viability reagents were purchased from Life Science Technologies. Penicillin/streptomycin/neomycin (PSN) Antibiotic Mixture 100× was purchased from Sigma. Antibodies for PARP, cleaved PARP, caspase-9, cleaved caspase-9, caspase-7, cleaved caspase-7, Ki-67, and β-actin were purchased from Cell Signaling Technology. E1A antibody and phosphatidylserine antibody [4b6] for in vitro immunofluorescent staining were purchased from Abcam. CD68 and NKp46 antibodies were from Biolegend. Phosphatidylserine antibody 1N11, also known as PGN635, a fully human anti-phosphatidylserine antibody was prepared by Dr. Rolf A. Brekken's laboratory (University of Texas Southwestern Medical Center, Dallas, TX) and was used for in vivo study (14). Conventional pancreatic cancer cell lines were from ATCC, and primary pancreatic cancer cells were established in our laboratory as described previously (15). All primary pancreatic cancer cells were authenticated with unique fingerprinting. NOD/SCID and nude mice (female, 6 weeks) were purchased from the National Cancer Institute (NIH, Bethesda, MD) and Jackson Laboratories.

Delta-24-RGD and Ad-GFP-RGD virus were prepared as described previously (16). For cell infection with Ad-GFP-RGD or Delta-24-RGD, 1 × 10⁴ cells were seeded in 6-well plates, and after 24 hours, cells were infected with adenovirus at different multiplicity of infections (MOI).

Cells (1 × 10⁴) were seeded in 6-well plates and infected with Delta-24-RGD or Ad-GFP-RGD control virus at different MOIs, and then 10 days postinfection, cells were fixed with ice-cold methanol for 10 minutes. Cells were washed with PBS and stained with 0.5% crystal violet solution under room temperature for 10 minutes. The plates were rinsed with water until no color came off. The plates were dried at room temperature for overnight.

Cells were seeded at the density of 1 × 10³ cells per well in a 96-well plate, and 24 hours later, the cells were infected with Dleta-24-RGD at MOIs of 0.1, 0.3, 1, 3, and 10. Cell viability was measured with MTT assay as described previously (15). The dose that causes 50% of cells death (IC₅₀) of infected cells compared with noninfected control was calculated with GraphPad Prism software (6.0).

Cells were harvested with trypsinization and washed with PBS containing 10% FBS. About 2 × 10⁶ cells in 100 μL FACS buffer were added in polystyrene round-bottom tube and incubated with another 100 μL Fc blocking buffer on ice for 20 minutes. Cells were centrifuged at 1,500 rpm for 5 minutes at 4°C and resuspended with 100 μL FACS buffer. Primary antibodies were added in the buffer with a concentration of 1 μg/mL. Cells were incubated for 60 minutes at 4°C. After incubation, cells were washed with FACS buffer 3 times by centrifugation at 1,500 rpm and resuspended in 200 μL cold FACS buffer. Diluted fluorochrome-labeled secondary antibody in FACS buffer was added to the cells and incubated them for 30 minutes at 4°C. Cells were washed 3 times with FACS buffer and resuspended in 200 μL FACS buffer. FACS analysis of fluorescence intensity was performed with FACSCalibur (BD), and the results were analyzed with FlowJo 10 software (FlowJo).

Cells (1 × 10⁵) seeded in the wells of chamber slide were treated with PBS or Delta-24-RGD virus at 1 MOI for 3 to 5 days, and 10 μL LC3B-GFP virus (BacMam 2.0, Thermo Fisher) was added into the well and incubated for 24 hours. Cells were checked under fluorescence microscope and LC3-GFP–positive autophagosome were counted. For acridine orange staining and flow cytometric analysis, cells (1 × 10⁶) in 6-well plates were infected with Delta-24-RGD virus at MOIs of 0.1, 1, and 10 for 5 days. Cells were harvested, washed with PBS, and incubated with acridine orange (1 μg/mL) for 15 minutes. Cells were washed with PBS twice and resuspended in 400 μL PBS for FACS analysis.

Cells (1 × 10⁵) were seeded in 6-well plates and infected with Delta-24-RGD at 1 MOI. Cells were harvested and counted, and cell lysates were collected at days 1, 3, and 7 postinfection. Virus DNA was isolated and purified with Adeno-X qPCR Titration Kit from Clontech Company. The copy numbers of adenovirus in the cells were quantified with the kit following manufacturer's instruction and normalized by cell number.

Cells after treatment or infection were washed with cold PBS, and tissue lysates were extracted with RIPA buffer and quantified for protein concentration with bicinchoninic acid (BCA) method. Thirty to 50 micrograms of protein was used for Western blotting. Detailed methods can be found in our previous publication (17).

Tumor tissues harvested from mice were fixed for overnight. Paraffin-embedded tumor tissues were cut into 5-μm sections. Immunohistochemical or immunofluorescent staining was performed using the methods described previously (18). Images were captured with an Olympus DP72 camera and CellSens software on an Olympus BX51 microscope.

Animal experiment protocol (00001089-RN00) was reviewed and approved by The University of Texas MD Anderson Cancer Center (Houston, TX) institutional review board and in accordance with the Guidelines for the Care and Use of Laboratory Animals published by the NIH. MDA-PATC53 cells were harvested, and 2 × 10⁶ cells in 100 μL PBS were mixed with equal volume of Matrigel (Invitrogen) and implanted into nude mice. When tumors reach the size of about 100 mm³, mice were randomly divided into 4 groups with 5 mice in each group. Mice were treated with PBS or Delta-24-RGD adenovirus 3 times weekly by intratumoral injection of 1 × 10⁸ plaque-forming units (pfu) virus. Treatment with 1N11 antibody was started after the third virus treatment by intraperitoneal injection of 100 μL antibody with the dose of 1 mg/kg. A total of 3 doses of 1N11 were administrated. The tumor volumes were calculated using the formula: length × width² × 0.52. When tumors reached the size of 12 mm in diameter, all mice were sacrificed, and tumors were harvested for histochemical analysis.

The significance of differences between different treatment groups was analyzed by two-way ANOVA or t test (two tailed), and P < 0.05 was considered as significant. Correlation of cyclin D1 expression with sensitivity of cell lines to Delta-24-RGD virus was analyzed with Pearson correlation method. All statistical analyses were done with GraphPad Prism 6.0 (GraphPad Software Inc.).

To test the cytotoxicity of Delta-24-RGD in pancreatic cancer cells, 4 cell lines, BxPC3, PANC1, MiaPaCa2, and MDA-PATC53, a primary cell line established in our laboratory, were infected with Ad-GFP-RGD control and Delta-24-RGD at different MOIs followed by crystal violet staining. Infection of cells with Delta-24-RGD virus induced dramatic cytotoxicity effects in PANC1, MiaPaCa2, and MDA-PATC53 cells but not in BxPC3 cells (Fig. 1A). We then used the cell viability assay to test cytotoxicity of Delta-24-RGD in 6 classic and 6 primary pancreatic cancer cell lines derived from PDAC PDX models (Fig. 1B and C). Six of 12 tested cell lines were sensitive to Delta-24-RGD. Notably, PANC1, MiaPaCa2, and AsPC1 have similar sensitivity as human glioma cell line (U251), which was used as a positive control as it has been previously shown to be sensitive to Delta-24-RGD (5). The IC₅₀ values of Delta-24-RGD in each of the cell lines were calculated (Supplementary Table S1). We used the IC₅₀ less than 10 MOIs as a sensitivity cutoff. On the basis of this cutoff, of the 12 lines tested, MDA-PATC53, MiaPaCa2, PANC1, MDA-PATC108, AsPC1, and MDA-PATC118 are sensitive to Delta-24-RGD. These results suggest that Delta-24-RGD induced dramatic cytotoxicity in a subset of pancreatic cancer cells.

To determine the molecular markers that correlate with sensitivity of pancreatic cancer cells to Delta-24-RGD (Supplementary Table S1), we analyzed the protein expressions of the major components of p16/RB/cyclin D1/CDK4 pathway using Western blotting. The expression of p16 was low in all of the tested cells, except MDA-PATC43 (Fig. 2A), which is consistent with high-frequency deletion and promoter methylation of p16 gene in pancreatic cancer. This result indicates that low or no expression of p16 is not a predictor of sensitivity to Delata-24-RGD. The expression of RB gene was low in all the sensitive cells, whereas 2 of 5 resistant cells also have low RB expression, suggesting that low RB expression alone is not a predictor of sensitivity either (Fig. 2A). Overall, the expression of cyclin D1 and CDK4 was higher in sensitive cells than in resistant cells (Fig. 2B). Using Pearson correlation analysis, we found that cyclin D1 expression is reversely correlated with the IC₅₀ of Dela-24-RGD (R = −0.8506; Fig. 2C). To examine for differences in virus replication between sensitive and resistant cells, virus copy number was analyzed with a PCR-based assay. As was shown in Fig. 2D, virus copy number increased dramatically over time in 2 sensitive cell lines, MiaPaCa2 and MDA-PATC53, but not in the resistant BxPC3 cells, suggesting that replication of Delta-24-RGD in infected cells is associated with its cytotoxicity effects.

Previous studies have found that oncolytic virus induces autophagy-mediated cancer cells death (19). To investigate whether the cytotoxicity induced by Delta-24-RGD in pancreatic cancer cells is also mediated by autophagy, LC3B-GFP expression reporter was introduced into the cells after 5 days of infection with Delta-24-RGD. Infection of Delta-24-RGD indeed dramatically induced autophagosome formation evidenced by LC3B-GFP punctae in Delta-24-RGD–sensitive and not in -resistant cells (Fig. 3A and B). Acridine orange staining further confirmed that acidic vesicular organelles (AVO) were significantly increased in sensitive cells after infection with Delta-24-RGD (Fig. 3C). In addition, Delta-24-RGD induced dramatic expression of LC3B-II, an autophagy marker, in sensitive cells (Fig. 3D). The result also showed that high MOI of Delta-24-RGD induced cleaved PARP, caspase-7, and caspase-9, indicating that caspase activation was induced (Fig. 3D).

Phosphatidylserine is an aminophospholipid distributed in inner leaflet of cell membrane. It is externalized to the outer leaflet when cells undergo apoptosis. Exposure of phosphatidylserine is a signaling for phagocytosis of apoptotic cells. Phosphatidylserine exposure on a virus's envelop or vesicle is a strategy called apoptotic mimicry for virus entry into host cells (20–22). Previous studies have revealed that some viruses can induce exposure of phosphatidylserine on the membrane of infected cells, which has been explored as a strategy for antivirus therapy (10). Phosphatidylserine exposure on cancer-specific endothelial cells and cancer cells was also reported and plays a role in cancer-related immunosuppression (8, 23–25). Targeting phosphatidylserine with an anti-phosphatidylserine antibody has been under clinical investigation for cancer therapy (11, 26, 27). To check whether Delta-24-RGD can also induce phosphatidylserine exposure in infected pancreatic cancer cells, immunofluorescent staining followed by flow cytometric analysis of Detla-24-RGD–infected cells was performed with a phosphatidylserine-specific antibody. As was shown in Fig. 4A, infection with Delta-24-RGD virus induced phosphatidylserine exposure in both resistant and sensitive pancreatic cancer cells. Further microscopic analysis also confirmed that Delta-24-RGD induced phosphatidylserine externalization to the outer leaflet of the cell membrane in 3 PDAC cell lines (Fig. 4B). These results suggest that Delta-24-RGD can induce phosphatidylserine exposure, which may not depend on the sensitivity of infected cells.

Previous studies have shown that phosphatidylserine exposure in tumor cells induces immunosuppression and targeting phosphatidylserine promotes anticancer immune response (11, 26). We hypothesize that the combination of Delta-24-RGD oncolytic virus with phosphatidylserine-targeting antibody may further enhance anticancer effects through both antibody-dependent cell-mediated cytotoxicity (ADCC) and enhanced anticancer immune response. To confirm this hypothesis, mice bearing tumors derived from a sensitive cell line MDA-PATC53 were first treated with 1 × 10⁸ pfu of Delta-24-RGD for 3 times by intratumoral injections. Then, mice were treated with intraperitoneal injections of phosphatidylserine-targeting antibody 1N11, a fully human phosphatidylserine antibody. The detailed treatment schedule is shown in Fig. 5A. The result showed that treatment with Delta-24-RGD alone significantly inhibited tumor growth compared with nontreated control (P < 0.0001). Phosphatidylserine-targeting antibody 1N11 had a moderate antitumor effect that was less effective than Delta-24-RGD alone (Fig. 5B and C). However, the combination of Delta-24-RGD virus and phosphatidylserine-targeting antibody was more effective than each agent alone (Delta-24-RGD plus 1N11 vs. 1N11 alone, P < 0.0001; Delta-24-RGD plus 1N11 vs. Delta-24-RGD, P < 0.01). Expression of adenovirus antigen E1A was confirmed in Delta-24-RGD or combination–treated tumors, indicating efficient virus infection and caspase-related cell death induced by Delta-24-RGD in vivo (Fig. 5C). Consistent with previous studies that phosphatidylserine-targeting antibody induces microphage activation (27), our study shows enhanced staining of CD68, a marker of macrophages, after treatment with 1N11 alone or after the combination treatment (Fig. 6). In addition, we also observed that infiltration of activated NK cells was enhanced in the tumor tissue with single or combined treatments (Fig. 6).

A recent clinical trial of Delta-24-RGD in patients with glioblastoma demonstrated favorable toxicity profile and remarkable clinical efficacy (7). This prompted us to evaluate its anticancer activity in pancreatic cancer, as 85% of PDACs have an incomplete p16/RB/E2F pathway due to promoter methylation and homozygous deletion of CDKN2A locus encoding p16 protein (28). In our study, Delta-24-RGD induced dramatic cytotoxicity in more than 50% of the tested cell lines. Specifically, 3 of 6 tested pancreatic cancer cell lines have sensitivity comparable to that of the human glioma cell line U251 (Fig. 1A), which has been shown, in previous study, to be a sensitive cell line to Delta-24-RGD in vitro and in vivo (19). One primary pancreatic cancer cell line, MDA-PATC53, which was established in our laboratory from a liver metastasis of pancreatic cancer (15), was particularly sensitive to Delta-24-RGD (Fig. 1A and C). Because most of metastatic pancreatic cancer cells are usually resistant to chemotherapies, this result suggests that Delta-24-RGD might be a promising agent for pancreatic cancer therapy. Importantly, a telomerase reverse transcriptase (TERT)-immortalized human pancreatic duct epithelia cell line, HPNE-tert, was not sensitive to Delta-24-RGD, suggesting that Delta-24-RGD may not be toxic to normal pancreatic cells.

In this study, 2 conventional cell lines, BxPC3 and Hs776T, were not sensitive to Delta-24-RGD, and these 2 cell lines, based on previous studies, have wild-type form of p16 gene (29). However, the genetic status of p16 seems to not be the principal driver of cell sensitivity to Delta-24-RGD, as p16 was very low in almost all tested cells lines (Fig. 2B). Likewise, RB expression was not found to be the determinant of sensitivity, although all the sensitive cells have low expression of RB (Fig. 2B). These results suggest that although Delta-24-RGD was originally engineered to preferably replicate in cancer cells with abnormality in p16/RB pathway, genetic or molecular changes of other factors might also affect the sensitivity of cancer cells to Delta-24-RGD. We found that among the tested components of p16/RB pathway, the expression level of cyclin D1 was most correlated with the sensitivity of PDAC cells to Delta-24-RGD. Cyclin D1 can bind to and activate CDK4/CDK6, which phosphorylates RB protein and promotes the release of E2F and cell-cycle progression. This pathway is negatively regulated by p16; however, in cells with deletion or suppressed expression of p16, cyclin D1 becomes the key driver for the activation of this pathway. This could explain why, in our study, high expression of cyclin D1 was closely linked to Delta-24-RGD sensitivity. Considering that about 40% to 80% of pancreatic cancers have cyclin D1 overexpression or gene amplification (30, 31), we speculate that Delta-24-RGD could be active in a large fraction of pancreatic cancers.

Autophagy or autophagy-mediated apoptosis are a described mechanism of oncolytic virus–induced cytotoxicity (19). In our study, we used multiple approaches, including autophagy reporter LC3-GFP, acridine orange staining, and Western blotting to confirm autophagy induction by Delta-24-RGD in pancreatic cancer cells (Fig. 3). Exposure of phosphatidylserine on enveloped viruses or on the vesicular membrane of unenveloped viruses, a strategy called apoptosis mimicry, is a described method of virus infection (32). In addition, some viruses induce phosphatidylserine exposure in infected cells, opening the possibility of using anti-phosphatidylserine antibodies as an antiviral therapy (10). However, the mechanism of virus-induced phosphatidylserine exposure in host cells has not been well studied. We found that Delta-24-RGD induces phosphatidylserine exposure in both sensitive and resistance cells (Fig. 4). We speculate that oncolytic virus–induced phosphatidylserine exposure could be caused by specific viral genes, as it reports of E1B 19-kDa protein–induced apoptotic mimicry in host cells to modulate innate immune responses (33). Alternatively, phosphatidylserine exposure could be related with metabolic changes induced by oncolytic virus (34). Further study is necessary to determine the mechanism of adenovirus-induced phosphatidylserine exposure to develop further rationale treatment combinations with virus-mediated gene therapy.

Phosphatidylserine exposure induced by oncolytic viruses can help the clearance of infected cells by immune system. This offers an opportunity for combination therapy with phosphatidylserine-targeting antibodies. Phosphatidylserine-targeting antibodies, including bavituximab and 1N11, have been evaluated in metastatic breast and lung cancers in phase I clinical trials showing promising results (12, 13). In this study, we tested the combination of Delta-24-RGD and 1N11, a fully human phosphatidylserine-targeting antibody similar to bavituximab, by using a sequential combination strategy. The major consideration for this sequential combination strategy is that phosphatidylserine-targeting antibody may suppress virus production and spreading if given concurrently. In a mouse pancreatic cancer model, sequential combination of Delta-24-RGD and 1N11 was superior to each single treatment alone (Fig. 5B). Mechanistically, combination treatment induced macrophage-mediated anticancer immune response, which was evidenced by increased staining of microphage marker, CD68, in the tumors treated with single and combination regimen. Stimulating of differentiation of M2 tumor–associated microphage (TAM) to M1 TAM by phosphatidylserine-targeting antibody has been reported previously (11, 27). In addition, modulating anticancer immune response by Delta-24-RGD has been observed in patients with glioblastoma and in immunocompetent mouse brain tumor models (7, 35–37). Notably, single treatment with Delta-24-RGD also induced strong anticancer effects, which may involve direct tumor lyses, autophagy, as well as immune responses, as the nude mice used for the tumor model keep natural killer cells and macrophages (38). Nevertheless, more pancreatic tumor animal models, including immunocompetent mouse models, might be used to evaluate the anticancer effects induced by Delta-24-RGD as a single agent and in combination with other agents, such as phosphatidylserine-targeting antibodies.

In summary, in this study, we evaluated the anticancer activity of oncolytic virus Delta-24-RGD in pancreatic cancer cells in vitro and in preclinical animal models in vivo. We demonstrate that Delta-24-RGD is active against pancreatic cancer cells, and cyclin D1 expression is associated with the vulnerability of pancreatic cancer cells to oncolytic adenovirus treatment. In addition, we demonstrate that Delta-24-RGD induced phosphatidylserine exposure, and combination of phosphatidylserine-targeting antibody and Delta-24-RGD viral therapy induced enhanced antitumor activity. Further studies are needed for full understanding of the mechanism of oncolytic virus–induced cytotoxicity, phosphatidylserine exposure, and oncolytic virus–induced immune response to develop rational combination therapies as well as predictive biomarkers for patients' selection. Nevertheless, our study demonstrates that Delta-24-RGD is a promising agent alone or in combination, for the treatment of pancreatic cancer, a devastating disease with limited therapeutic regimens available. Given that Delta-24-RGD and phosphatidylserine-targeting antibodies are being evaluated in clinical trials and have favorable toxicity profile, combination of these 2 agents could be quickly translated into clinical testing.

R.A. Brekken reports receiving a commercial research grant from and serves as a Medical and Scientific Advisor at Peregrine Pharmaceuticals. J. Fueyo-Margareto is shareholder of and is a consultant/advisory board member for DNAtrix. F.F. Lang reports receiving other commercial research support from DNAtrix, Inc. and is a patent holder for Delta-24-RGD. No potential conflicts of interest were disclosed by the other authors.

Conception and design: B. Dai, J.B. Fleming

Development of methodology: B. Dai, D. Roife, Y. Kang, J. Gumin, X. Li, F.F. Lang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B. Dai, D. Roife, M.V.R. Perez, M. Pratt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B. Dai, M.V.R. Perez

Writing, review, and/or revision of the manuscript: B. Dai, R.A. Brekken, J. Fueyo-Margareto, F.F. Lang, J.B. Fleming

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Dai, D. Roife, X. Li, R.A. Brekken, J. Fueyo-Margareto, F.F. Lang, J.B. Fleming

Study supervision: J.B. Fleming

This work was funded by The Skip Viragh Family Foundation, the Various Donors in Pancreatic Cancer Research Fund (to J.B. Fleming), the Research Animal Support Facility—Houston under NIH/NCI award P30CA016672 (to R.A. Depinho), the National Cancer InstituteP50 CA127001, the Broach Foundation for Brain Cancer Research, the Howard and Susan Elias Foundation, and the Jason & Priscilla Hiley Fund (to F.F. Lang). This work was also supported in part by the NIH grant T32CA009599 (D. Roife and M.V. Rios Perez).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.