Intratumoral 5-Fluorouracil Produced by Cytosine Deaminase/5-Fluorocytosine Gene Therapy Is Effective for Experimental Human Glioblastomas1

[CANCER RESEARCH 62, 773–780, February 1, 2002]

Intratumoral 5-Fluorouracil Produced by Cytosine Deaminase/5-Fluorocytosine Gene Therapy Is Effective for Experimental Human Glioblastomas1 C. Ryan Miller, Christopher R. Williams, Donald J. Buchsbaum, and G. Yancey Gillespie2 Department of Radiation Oncology, Division of Radiation Biology [C. R. M., D. J. B.], and Department of Surgery, Divisions of Neurosurgery [C. R. M., G. Y. G.] and Urology [C. R. W.], University of Alabama at Birmingham, Birmingham, Alabama 35294

ABSTRACT 5-Fluorouracil (5-FU) is a potent antimetabolite used for chemotherapy of gastrointestinal (GI), breast, and head and neck malignancies. Although clinical trials have been conducted, the poor therapeutic index of 5-FU has precluded its clinical use for a number of other tumor types. It is unclear whether this lack of utility is due to problems with drug delivery or inherent insensitivity. Adenovirus (Ad) vector-mediated cytosine deaminase (CD)/5fluorocytosine (5-FC) gene therapy has the potential to overcome pharmacokinetic issues associated with systemic 5-FU and is particularly well suited to use with tumors in which local control is paramount, such as recurrent, localized prostate cancer and malignant gliomas. In this study, the in vitro response by a panel of human tumor cell lines derived from both GI (colon, pancreas) and non-GI (prostate, glioma) tumors to 5-FU and to AdCMVCD (an Ad encoding Escherichia coli CD)/5-FC was examined. Whereas the sensitivity (IC50) of individual cell lines to these agents varied, no significant difference in median IC50 for either 5-FU or AdCMVCD/5-FC was evident for the four tumor types tested (P > 0.1). The relevant contributions of Ad gene transfer efficiency and inherent 5-FU sensitivity in determining response to AdCMVCD/5-FC were then assessed. Multiple linear regression analysis revealed that whereas both factors significantly contribute to the response, inherent 5-FU sensitivity was substantially more important (␤ ⴝ 0.78 versus 0.48; P < 0.001). Finally, the therapeutic efficacy of a single intratumoral injection of AdCMVCD followed by systemic 5-FC was assessed in three intracranial C.B17 severe combined immunodeficient mouse models of human glioma. AdCMVCD/5-FC efficacy was specific, virus dose-dependent, and closely paralleled in vitro 5-FU and CD/5-FC sensitivity in two of three models tested. These results reveal that glioma cells are as sensitive as GI tumor cells to the antineoplastic effects of 5-FU, identify inherent 5-FU sensitivity as an important factor in determining CD/5-FC efficacy, and confirm previous findings in rat models that demonstrate the potential clinical utility of AdCMVCD/5-FC gene therapy for gliomas.

INTRODUCTION 5-FU,3 first described in 1957 (1), remains an essential component of chemotherapy for a number of solid tumors, particularly GI, breast, and head and neck malignancies. Overall, the single-agent response rate of 5-FU in GI tumors is 10 –20% (2). 5-FU has also been actively investigated during the last 40 years for many non-GI tumors. However, the role of systemic 5-FU in therapy of these tumors has been Received 4/13/01; accepted 11/29/01. 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. 1 Supported by a fellowship from the American Brain Tumor Association (to C. R. M. and G. Y. G.), a grant from the Pediatric Brain Tumor Foundation of the United States (to C. R. M. and G. Y. G.), and NIH Grants CA73636 (to D. J. B.) and CA71933 (to G. Y. G.). 2 To whom requests for reprints should be addressed, at Department of Surgery, Division of Neurosurgery, 1032 Tinsley Harrison Tower, 1530 Third Avenue South, Birmingham, AL 35294-0006. Phone: (205) 934-7227; Fax: (205) 934-7676; E-mail: [email protected]. 3 The abbreviations used are: 5-FU, 5-fluorouracil; Ad, adenovirus; CAR, Coxsackie adenovirus receptor; CD, cytosine deaminase; CMV, cytomegalovirus; 5-FC, 5-fluorocytosine; GDEPT, gene-directed enzyme-prodrug therapy; GFP, green fluorescent protein; GI, gastrointestinal; ILS, increased length of survival; MOI, multiplicity of infection; pfu, plaque-forming unit(s); SCCHN, squamous cell carcinoma of the head and neck; SCID, severe combined immunodeficient; UPRT, uracil phosphoribosyltransferase; VDEPT, virus-directed enzyme-prodrug therapy; XRT, radiation therapy; GFP, green fluorescent protein; b.i.d., bis in die (twice a day); CNS, central nervous system; BCNU, carmustine; FDA, Food and Drug Administration.

limited by the fact that dose-limiting toxicities (myelosuppression and stomatitis) are usually reached before evidence of antitumor response. Potential explanations for this low therapeutic index include pharmacokinetic factors limiting accessibility of systemic 5-FU to tumor cells and inherent insensitivity of tumor cells to 5-FU. Intratumoral delivery of genes encoding prodrug-activating enzymes, an approach termed GDEPT (3) or molecular chemotherapy (4), has the potential to become a powerful alternative method of drug delivery. In particular, the GDEPT combination of CD (EC 3.5.4.1) and 5-FC may be used to circumvent the pharmacokinetic limitations of systemic 5-FU. In this system, the nontoxic antifungal metabolite 5-FC is deaminated to the potent antitumor agent 5-FU within tumor cells expressing CD. Murine studies with CD-conjugated monoclonal antibodies (5), encapsulated CD protein (6), or tumor cells engineered to express the CD gene (7) have shown that high intratumoral concentrations of 5-FU, with minimal diffusion back into circulation, may be attained on conversion of systemically administered 5-FC. The CD/5-FC GDEPT system likely to be used clinically involves intratumoral delivery of CD via viral vectors (VDEPT). We have demonstrated previously that intratumoral injection of an Ad encoding Escherichia coli CD (AdCMVCD) into mice bearing s.c. colon (8) and bile duct (4, 8) tumors followed by systemic 5-FC administration results in significant suppression of tumor growth. CD/5-FC VDEPT is logical because systemic 5-FU chemotherapy is standard of care for these tumors. However, GI tumors, particularly colorectal carcinomas, are frequently metastatic on initial presentation (9). With current vector technology, CD/5-FC VDEPT is best suited to local control of cancers that are easily accessible to intratumoral vector injection. Examples of such non-GI tumors include cutaneous SCCHN (10), recurrent and/or hormone-refractory localized prostate cancer (11), and gliomas (12). With the exception of SCCHN, 5-FU is not routinely used in the clinical management of these tumors. It is not clear whether this is due to an inherent lack of sensitivity or, alternatively, to poor drug delivery. In this report, we address both of these important issues in determining the potential clinical utility of interstitial 5-FU in such tumors. To address the former, the in vitro sensitivities of a panel of 14 human cell lines derived from both GI (colon, pancreas) and non-GI (prostate, glioma) tumors to 5-FU and to AdCMVCD/5-FC were quantified. To address the latter, the relative importance of Ad gene transfer efficiency versus inherent 5-FU sensitivity in determining AdCMVCD/5-FC toxicity in vitro was estimated using multiple linear regression analysis. Finally, we chose gliomas as a representative non-GI tumor system for which (a) direct intratumoral vector injection is likely to be used clinically and (b) 5-FU is not routinely used. The CD/5-FC system has been previously investigated in preclinical models of glioma (6, 13–19). All of these studies used either rat C6 glioma cells injected intracranially in Wistar rats, an allogeneic model (20), or rat 9L gliosarcoma cells injected intracranially in syngeneic Fisher 344 rats or s.c. in xenogeneic nude mice. Whereas most of these studies used cell lines stably transfected to express CD, two investigated intratumoral injection of Ad vectors encoding CD (15, 18), and two used a fusion gene encoding CD and herpes simplex virus thymidine kinase (16, 18). Although comparisons of results with such diverse models

773

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

with different transgenes and prodrug treatment regimens are difficult to make, CD/5-FC therapy, in general, produced significant tumor volume reduction (s.c.) or ILS (intracranial) in each model. Both C6 and 9L rat glioma models have been frequently used for experimental therapeutic studies. However, these models suffer from a number of problems that limit extrapolation to the clinical situation, most notably the lack of a syngeneic host (C6) and an intense antitumor immune response even in the absence of therapy [C6 and 9L (20)]. Orthotopic xenograft SCID mouse models of human glioma do not suffer from these limitations, permitting evaluation of experimental therapeutics in the absence of tumor immunogenicity. Thus, this is the first report to assess the therapeutic efficacy of CD/5-FC VDEPT in intracranial C.B17 SCID mouse models of human glioma using three human glioma cell lines displaying different genetic, morphological, and phenotypic profiles. Data presented provide the rationale for further evaluation of CD/5-FC-based VDEPT for human gliomas in clinical trials. MATERIALS AND METHODS

Corning, NY) overnight at 37°C. Cells were infected with varying amounts of AdCMVCD (0 –300 pfu/cell) in 2 ml of OptiMEM (Life Technologies, Inc., Gaithersburg, MD) for 1 h at 37°C with continuous rocking. Twenty-four h after infection, cells were harvested with 0.25% trypsin/EDTA, plated into 96-well tissue culture plates at 5000 cells/well in 100 ␮l of complete media, and allowed to adhere overnight at 37°C. One hundred ␮l of media supplemented with serial dilutions of 5-FC were added, and 5-FC toxicity (IC50) was determined at 5 days as described above. Quantification of Ad Transducibility. Ad gene transfer efficiency was quantified by monitoring expression of GFP in AdCMVGFP-infected cells as described previously (22). Briefly, cells (0.5–1 ⫻ 106) were plated in 6-well plates (Corning), allowed to adhere overnight, and subsequently infected with AdCMVGFP at various MOIs (0 –500 pfu/cell) in 0.6 ml of OptiMEM for 1 h at 37°C with continuous rocking. Twenty-four h after infection, cells were harvested with 0.25% trypsin/EDTA (Mediatech), washed with buffer (PBS, 0.1% sodium azide, and 0.1% BSA), and resuspended at 1–2 ⫻ 106 cells/ml. Cells (104 cells/sample) were analyzed by flow cytometry at the University of Alabama at Birmingham Department of Rheumatology FACS Core Facility. MOI50 values, defined as the AdCMVGFP MOI required to produce detectable GFP in 50% of cells, were determined for each cell line by a piecewise linear regression of data plotted as the logarithm of AdCMVGFP MOI versus the percentage of GFP-positive cells. AdCMVCD/5-FC Treatment of C.B17 SCID Mice Bearing Intracranial Human Gliomas. U87MG, D54MG, or U251MG tumors were established intracranially in C.B17 SCID mice as described previously (23–25). Briefly, mice were anesthetized with ketamine, and 0.5–1 ⫻ 106 cells (1 ⫻ 107 ml⫺1) were inoculated into the right frontal cerebral hemisphere via stereotactic injection. Tumors were allowed to grow for 5 days before stereotactic injection of 10 ␮l of saline, AdCMVCD, or AdCMVhSSTr2 (109 pfu). Two days after infection, mice were treated with either saline or 500 mg/kg 5-FC b.i.d. i.p. for 7 days and monitored daily for survival. When tumor-bearing mice displayed overt signs of neurological dysfunction, manifested primarily as a hunched appearance, lack of grooming, and lack of avoidance behavior when handled, they were sacrificed by lethal CO2 inhalation, and their brains were harvested for histopathological examination, confirming the presence of tumor in all sacrificed mice. Statistics. Pairwise comparisons of mean 5-FU or 5-FC IC50 values calculated as described above were made using one-way ANOVA with Fisher’s modification using MINITAB v13.3 for Windows (Minitab, State College, PA). Median 5-FU and 5-FC IC50 values for each tumor type were calculated by grouping cell lines according to tissue of origin and compared by KruskalWallis analysis using MINITAB. Pairwise comparisons of median IC50 values were performed with Dunn’s post-test using GraphPad InStat v3.05 (GraphPad, San Diego, CA). Multiple linear regression analysis on log-converted 5-FU IC50, 5-FC IC50, and AdCMVGFP MOI50 values was performed using GraphPad InStat v3.05. Assessment of the relative importance of each variable was made using the ␤ standardized regression coefficient (26). Kaplan-Meier survival curves were analyzed by the log-rank test using GB-STAT v6.5 for Macintosh (Dynamic Microsystems, Silver Springs, MD), and specific pairwise comparisons were made. All comparisons were made using the 0.05 level of significance, unless otherwise stated.

Tumor Cells, Animals, and Chemicals. Human glioma cell lines D54MG, U87MG, U251MG (obtained from Darell Bigner; Duke University Medical Center, Durham, NC) and U118MG (American Type Culture Collection, Manassas, VA) were cultured in DMEM/F12 (Mediatech, Herndon, VA) containing 7% fetal bovine serum (Summit Biotechnology, Fort Collins, CO) and 2 mM glutamine. Human colon (LS174T and WiDr), pancreatic (AsPC-1, BxPC-3, and MIA PaCa-2), and prostate (DU145, LNCaP, and PC-3) carcinoma cell lines (American Type Culture Collection) were maintained in RPMI 1640 (Mediatech) containing 10% fetal bovine serum and 2 mM glutamine. All cells were cultured at 37°C in a 5% CO2 atmosphere without antibiotics and passaged ⬍12 times during the course of these experiments. C.B17 SCID mice were purchased from Frederick Cancer Research Facility (Bethesda, MD) and housed under aseptic conditions in microisolator cages with sterile food and water ad libitum. 5-FC and 5-FU were purchased from Sigma Chemical Co. (St. Louis, MO) and SP Pharmaceuticals (Albuquerque, NM), respectively. Viruses and Monoclonal Antibodies. AdCMVGFP, a first-generation E1-, E3-deleted vector expressing GFP from the CMV immediate early promoter, was obtained from Corey Goldman (Cleveland Clinic, Cleveland, OH). The construction and characterization of AdCMVCD (4) and AdCMVhSSTr2 (21), an Ad encoding human somatostatin receptor subtype 2, have been described previously. All viruses were constructed by two-plasmid rescue in 293 cells using pACCMVpLpA shuttle vector and pJM17 rescue vector (4). Viruses were propagated and plaque titered on permissive 293 cells and purified twice by centrifugation on CsCl gradients. All virus aliquots were maintained at ⫺80°C until use. Single lots of AdCMVCD, AdCMVGFP, and AdCMVhSSTr2 were used throughout these experiments. The AdCMVCD lot proved free of contaminating wild-type Ad after infection of nonpermissive cells (A549 and HeLa; data not shown) and neurotoxicity after injection of up to 109 pfu into the brains of non-tumor-bearing C.B17 SCID mice as described below. 5-FU and AdCMVCD/5-FC Toxicity. Confluent cell monolayers were harvested with 0.25% trypsin/EDTA, plated (5000 cells/well in 100 ␮l of RESULTS complete media) in 96-well tissue culture plates, and allowed to adhere 5-FU Sensitivity. Sensitivity to continuous exposure of 5-FU was overnight at 37°C. Ten serial half-log dilutions of 5-FU (six replicates/dilution) were made, and 100 ␮l of 5-FU-containing media were added directly to cells assessed for 14 cell lines derived from malignant gliomas and colon, to achieve final concentrations of 0 –200 ␮g/ml. Cells were incubated at 37°C pancreatic, and prostate carcinomas and was found to be variable, for 5 days, and cellular respiration was determined using 3-(4,5-dimethylthia- within a 2-log range (Fig. 1; IC50 range, 0.07–1.44 ␮g/ml). Cells zol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfonyl)-2H-tetrazolium (Pro- could be grouped into three categories based on statistical analysis of mega, Madison, WI) as per the manufacturer’s protocol. Fractional cell sur- the individual mean 5-FU IC for each cell line. Group I, consisting 50 vival at each drug concentration was calculated as the ratio of absorbance at only of BxPC-3 pancreatic carcinoma cells, was most sensitive 490 nm of cells incubated in the presence versus absence of drug, corrected for (IC50 ⫽ 0.07 ⫾ 0.03 ␮g/ml). Group III, consisting of MIA PaCa-2 background absorbance of media alone. Fractional cell survival data were pancreatic carcinoma cells and U87MG and D54MG glioma cells, plotted against the logarithm of drug concentration and IC50 values extrapolated by piecewise linear regression as the concentration of drug producing a was the least sensitive (IC50 ⫽ 1.24 –1.44 ␮g/ml). The largest group of cells, Group II, displayed intermediate 5-FU sensitivity. Only the 50% reduction in corrected absorbance. For AdCMVCD/5-FC toxicity experiments, cells were harvested from con- difference in mean IC50 values for group I versus group III cells was fluent monolayers and plated at 1.5–3 ⫻ 106 cells/well in T25 flasks (Corning, statistically significant (P ⬍ 0.05). A notable exception was LS174T 774

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

(a) Ad gene transfer efficiency; and (b) inherent drug sensitivity. These variables, represented as MOI50 and IC50, respectively, were used to calculate 5-FC sensitivity (IC50) after infection of 11 cell lines with AdCMVCD at 100 pfu/cell. Multiple regression analysis was performed on log-converted 5-FU IC50, AdCMVGFP MOI50, and 5-FC IC50 values, and coefficients were determined for the equation shown below.

Log共5-FCIC50兲 ⫽ b0 ⫹ b1 ⫻ Log共5-FUIC50兲 ⫹ b2 ⫻ Log共AdCMVGFPMOI50兲 (1)

Fig. 1. 5-FU sensitivity. Mean 5-FU sensitivity (IC50; ␮g/ml) of cell lines derived from human colon, pancreatic, and prostatic carcinomas and from gliomas determined from two to nine separate experiments/cell line. Difference between individual mean IC50 values of group I cells (⽧) versus group III cells (f) was statistically significant (P ⬍ 0.05). Differences in IC50 values between groups I and II (F) and groups II and III were not statistically significant (P ⬎ 0.05). Differences in median IC50 (bars) for each of the four tumor types were not significant (P ⫽ 0.10).

colon carcinoma cells, which were significantly more sensitive to 5-FU than were U87MG and D54MG glioma cells (P ⬍ 0.05). AdCMVCD/5-FC Sensitivity. 5-FC sensitivity (IC50 in ␮g/ml) was assessed with a subpanel of 11 cell lines after infection with AdCMVCD at 1, 10, 30, 100, and 300 MOI (pfu/cell). Dose-response curves from representative pancreatic, glioma, and colon cancer cell lines (AsPC-1, D54MG, and WiDR, respectively) are shown in Fig. 2. A significant linear relation (P ⬍ 0.01) between AdCMVCD MOI and 5-FC sensitivity (log-log scale) for each of the three cell lines was observed. Similar results were obtained with BxPC-3 and LS174T cells as described previously (27). The dose response with PC-3 and U251MG cells was not significantly linear (P ⬎ 0.05) despite sufficient data for analysis (three and four MOIs, respectively). Only two MOIs (100 and 300 MOI) were tested for four cell lines (DU145, LNCaP, U118MG, and U87MG), precluding accurate analysis of the linear relation with these cell lines (data not shown). Like results with 5-FU, the 5-FC responses in the subpanel of cell lines infected with AdCMVCD at 100 pfu/cell were variable within a 2-log range (Fig. 3; IC50 range, 1.0 –196 ␮g/ml). Statistical analysis of mean 5-FC IC50 for each individual cell line revealed two groups with similar responses. Group II was comprised of three cell lines that were significantly less sensitive than group I cells (P ⬍ 0.05): U118MG and U87MG glioma and AsPC-1 pancreatic cells (IC50 ⫽ 83, 96, and 196 ␮g/ml, respectively). The remaining eight cell lines comprised group I (IC50 range, 1.0 –18.7 ␮g/ml). Tumor-type Response to 5-FU and AdCMVCD/5-FC. Median 5-FU toxicity (IC50) varied according to tumor type: 0.30, 0.37, 0.32, and 1.42 ␮g/ml for colon, pancreatic, prostate, and glioma cell lines, respectively (Fig. 1). Similarly, response to 5-FC based on infection with AdCMVCD (100 pfu/cell) varied because the median 5-FC IC50 under these infection conditions was 4.1, 99.8, 4.2, and 53.2 ␮g/ml, respectively (Fig. 3). Kruskal-Wallis analysis revealed no significant difference in either the median 5-FU IC50 (P ⫽ 0.10) or the 5-FC IC50 (P ⫽ 0.24) for the four cell types tested. These results were also independent of infection conditions because no difference in tumor-type response was evident at 10, 30, and 300 pfu/cell AdCMVCD (P ⬍ 0.21, 0.21, and 0.65, respectively; data not shown). Contribution of Inherent 5-FU Sensitivity and Ad Transducibility in Determining AdCMVCD/5-FC Response. For AdCMVCD/5-FC GDEPT, two factors may influence the cellular response:

This model described the data well (Table 1; P ⬍ 0.001) and revealed that both AdCMVGFP gene transfer efficiency and inherent 5-FU sensitivity significantly contributed to overall 5-FC response after AdCMVCD infection. However, inherent 5-FU sensitivity was the more important factor of the two, as demonstrated by a b1 ␤ weight of 0.78 compared with b2 ␤ weight of 0.48 for Ad transducibility. Finally, as shown in Fig. 4, 5-FC IC50 predicted by the above equation correlated strongly with measured 5-FC IC50 (P ⫽ 0.0001). Intratumoral 5-FU Chemotherapy of Established Intracranial Gliomas in C.B17 SCID Mice. The dose response of intratumoral 5-FU chemotherapy via AdCMVCD/5-FC was assessed in an estab-

Fig. 2. AdCMVCD/5-FC dose response. 5-FC sensitivity (IC50; ␮g/ml) was quantified using cells infected with increasing MOIs (pfu/cell) of AdCMVCD, and the dose response was determined. Data from three representative cell lines are shown (R2 ⬎ 0.94; P ⬍ 0.01): AsPC-1, ⽧; D54MG, F; and WiDR, Œ.

Fig. 3. AdCMVCD/5-FC sensitivity. Mean 5-FC sensitivity (IC50; ␮g/ml) of cell lines derived from human colon, pancreatic, and prostatic carcinomas and from gliomas after infection with AdCMVCD at 100 MOI (pfu/cell). Difference between individual mean 5-FC IC50 values of group I cells (F) and group II cells (f) were statistically significant (P ⬍ 0.05). Differences in median 5-FC IC50 (bars) for each of the four tumor types were not significant (P ⫽ 0.24).

775

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

Table 1 Contributions of inherent 5-FU sensitivity and Ad transducibility in determining AdCMVCD/5-FC sensitivity to human tumor cells in vitroa Log(5-FCIC50) ⫽ b0 ⫹ b1 ⫻ Log(5-FUIC50) ⫹ b2 ⫻ Log(AdCMVGFPMOI50) P ⬍ 0.001

Coefficient

Value

Standardized coefficient (␤)

b0 b1 b2

0.81 1.21 0.56

0.78 0.48

Significance (P) 0.022 0.001 0.013

a Multiple regression analysis performed on log-converted 5-FU IC50, AdCMVGFP MOI50, and 5-FC IC50 (AdCMVCD, MOI 100) values.

two additional intracranial C.B17 SCID mouse models of human glioma (24). AdCMVCD/5-FC therapy significantly prolonged survival of mice bearing either intracranial D54MG tumors [Fig. 6B; median survival, 43 days versus 27 days (control); P ⫽ 0.002] or U251MG tumors [Fig. 6C; median survival, 35 days versus 22 days (control); P ⫽ 0.001]. This effect was dependent on AdCMVCD/5-FC because no prolongation of survival with control virus (AdCMVhSSTr2 ⫾ 5-FC) and control drug (AdCMVCD ⫹ saline) regimens was seen with either model (P ⬎ 0.05). Importantly, the magnitude of survival advantage in AdCMVCD/5-FC-treated animals differed among the three models. As shown in Table 2, the ILS for U87MG, D54MG, and U251MG tumors was 1.67, 2.25, and 1.59, respectively. With the exception of results with U251MG, the treatment response correlated well with the in vitro sensitivity of these cells to both 5-FU and AdCMVCD/5-FC (Figs. 1 and 3, U251MG ⬎ D54MG ⬎ U87MG). DISCUSSION This study addresses three important issues regarding CD/5-FC VDEPT. First, data presented in Figs. 1 and 3 demonstrate that cell lines derived from both GI and non-GI tumors display similar in vitro sensitivity to both 5-FU and AdCMVCD/5-FC on continuous 5-day drug exposure using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfonyl)-2H-tetrazolium assay of cellular respiration. Nonparametric analysis of median 5-FU (Fig. 1) and AdC-

Fig. 4. Prediction of AdCMVCD/5-FC response from inherent 5-FU sensitivity and AdCMVGFP transducibility. Multiple regression analysis on log-converted 5-FU IC50, AdCMVGFP MOI50, and 5-FC IC50 (AdCMVCD, MOI 100 pfu/cell) values was performed. Experimental data were entered into the equation in Table 1, and 5-FC IC50 values predicted from this model (Y axis) were compared with actual 5-FC IC50 values (X axis).

lished model of intracranial U87MG gliomas in C.B17 SCID mice (23). As shown in Fig. 5, a single intratumoral injection of 108 pfu AdCMVCD followed by a twice daily, 7-day course of i.p. 5-FC at 500 mg/kg significantly prolonged median survival of mice bearing U87MG tumors (42 days) as compared with that of mice treated with intratumoral saline plus 5-FC (31 days; P ⬍ 0.05). Mice receiving 109 pfu AdCMVCD with systemic 5-FC lived significantly longer (61 days) than both control, saline/5-FC-treated mice (P ⬍ 0.01) and mice receiving 108 pfu AdCMVCD/5-FC (P ⬍ 0.05). The specificity of the AdCMVCD/5-FC therapeutic effect was determined with mice bearing intracranial U87MG tumors and treated with either saline, AdCMVCD, or the control virus AdCMVhSSTr2 (109 pfu/mouse), followed by systemic saline or 5-FC as described above. As shown in Fig. 6A, a single intratumoral injection of 109 pfu AdCMVCD significantly prolonged the survival of C.B17 SCID mice bearing intracranial U87MG tumors compared with control, mock-infected animals treated with 5-FC (median survival, 67 days versus 49 days, respectively; P ⬍ 0.05). This effect was not due to AdCMVCD infection alone because survival of mice receiving 109 pfu AdCMVCD but no 5-FC was not significantly prolonged compared with that of control mice (median survival, 49.5 days; P ⬎ 0.22). The prolongation of survival was specific for AdCMVCD/5-FC because infection with AdCMVhSSTr2, which encodes the human SSTr2 (21) and does not induce CD activity in infected cells (data not shown), did not significantly prolong survival of mice in either the presence (53.5 days) or absence (46 days; data not shown) of 5-FC (P ⬎ 0.4). The potential clinical efficacy of this approach was assessed in

Fig. 5. AdCMVCD/5-FC dose response in an intracranial C.B17 SCID mouse model of human U87MG glioma. U87MG gliomas (0.5 ⫻ 106 cells/mouse) were established in the right frontal cortex of C.B17 SCID mice (8 –10 mice/group). Five days after tumor implantation, saline (Œ) or 108 (circles) or 109 (squares) pfu/mouse AdCMVCD were injected intratumorally. Mice then received i.p. saline (open symbols) or 500 mg/kg 5-FC (closed symbols) b.i.d. and were subsequently monitored for survival. Differences in median survival of mice receiving 108 pfu versus saline (ⴱ), 109 pfu versus saline (ⴱⴱ), and 108 versus 109 pfu (ⴱⴱⴱ) AdCMVCD plus systemic 5-FC were significant at Ps of 0.05, 0.01, and 0.05, respectively.

Table 2 Efficacy of AdCMVCD/5-FC therapy in three intracranial C.B17 SCID mouse models of human gliomaa ILS Tumor

Expt 1

U87MG D54MG U251MG

1.37 1.59 1.59

b

Expt 2

Mean

1.97 2.90c ND

1.67 2.25 1.59

a ILS of animals receiving 109 pfu of AdCMVCD with systemic 5-FC (500 mg/kg b.i.d. i.p. for 7 days) versus animals receiving no AdCMVCD plus 5-FC. 5-FC regimen started 2 days after virus infection. ILS was determined from one to two separate experiments per tumor with 7–11 animals/group. b Expt, experiment. c 5-FC regimen started 3 days after virus infection.

776

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

MVCD/5-FC (Fig. 3) sensitivity (IC50) of 14 tumor cell lines (2 colon, 5 pancreatic, 4 glioma, and 3 prostatic cell lines) revealed no significant differences among the four tumor cell types tested (P ⫽ 0.1 and 0.24, respectively). A similar analysis was performed on publicly available 5-FU toxicity data from the National Cancer Institute Developmental Therapeutics Program Disease-oriented Anticancer Drug Screen (28). The National Cancer Institute screens approximately 10,000 compounds/year using a sulforhodamine B protein biomass assay to assess acute drug toxicity (2 days) with a panel of ⬎60 human tumor cell lines from 9 different tissues (breast, CNS, colon, lung, leukemia, melanoma, ovarian, prostate, and renal). Analysis of data in the August 2000 database4 revealed no significant difference in median 5-FU GI50 (50% growth inhibition) values (P ⫽ 0.17; data not shown) for the nine different tissues. This analysis confirms that our results with 5-FU were not an artifact of the cell lines included in our panel, the toxicity assay used, or the time point analyzed. These results are consistent with the broad spectrum of activity demonstrated clinically with systemic 5-FU and suggest that intratumoral 5-FU may be active, with optimal drug delivery via CD/5-FC GDEPT, for both GI and non-GI tumors. CD/5-FC VDEPT is a two-component system designed to produce high intratumoral 5-FU concentrations from systemically administered 5-FC. Both CD expression level within infected tumor cells and inherent 5-FU sensitivity of the tumor cells may contribute to its overall efficacy. Thus, the second focus of this study was to determine the relative importance of these two factors in vitro. We have previously shown that AdCMVCD MOI correlates directly with CD mRNA and protein expression, CD activity, and 5-FC toxicity in vitro using LS174T colon and BxPC-3 pancreatic cells (27). We have extended this analysis to include five additional cell lines, three of which demonstrated a significant dose-response relationship between 5-FC toxicity (IC50) and increasing AdCMVCD MOI (Fig. 2). To our knowledge, these data are the first to demonstrate a clear doseresponse relationship between AdCMVCD dose and 5-FC toxicity. Multiple linear regression analysis confirmed that both Ad gene transfer efficiency and inherent 5-FU sensitivity contributed significantly to AdCMVCD/5-FC toxicity in vitro (Table 1) and that quantitation of these two factors could accurately predict 5-FC toxicity (Fig. 4). We have previously reported a facile method of quantifying Ad transduction efficiency by flow cytometry using AdCMVGFP that permits the statistical comparison of transduction efficiencies across different cell types, represented as a MOI50 value for each cell sample (22). We chose the method using AdCMVGFP for these studies, rather than AdCMVCD, due to its simplicity (direct autofluorescence) and its amenability to subsequent statistical analysis. However, differences in transcription and translation efficiencies of different transgenes within the same viral vector and driven from the same promoter may produce differing results. Preliminary experiments involving quantitation of AdCMVCD transduction efficiency using flow cytometric detection of CD by indirect immunofluorescence on a subset of the cell lines used in this study have demonstrated the concordance of these two methods (27, 29), suggesting that GFP is a suitable surrogate reporter gene for these studies. These data also illustrate the importance of Ad gene transfer efficiency in determining CD/5-FC VDEPT efficacy. Ad gene transfer has been shown to depend on expression of the cellular receptors necessary for Ad entry, CAR (30, 31) and ␣v integrins (32), on the surface of target cells. Lack of expression of CAR or ␣v integrins on primary tumor cells from different tissues significantly limits Ad gene transfer (33–35). Moreover, in animal models, intratumoral Ad gene 4

Fig. 6. Efficacy of AdCMVCD/5-FC in three intracranial C.B17 SCID mouse models of human glioma. U87MG (A), D54MG (B), or U251MG (C) human glioma cells (0.5–1 ⫻ 106 cells/mouse) were injected into the right frontal cortex of C.B17 SCID mice (7–11 mice/group). Five days after tumor implantation, a single dose of saline (triangle) or 109 pfu of AdCMVhSSTr2 (circles) or AdCMVCD (squares) was injected intratumorally. Mice then received i.p. saline (open symbols) or 500 mg/kg 5-FC (closed symbols) b.i.d. and were subsequently monitored for survival. A representative of one to two experiments for each tumor is shown. Differences in median survival were significant at 0.05 (ⴱ) and 0.01 (ⴱⴱ), respectively.

http://www.dtp.nci.nih.gov.

transfer is limited to the area around the needle track, presumably due to high interstitial pressure within the tumor that limits vector diffusion (36). Therefore, to maximize intratumoral CD expression and hence CD/5-FC efficacy, targeted Ad vectors capable of CAR-inde-

777

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

pendent gene transfer (37), replicative Ad vectors capable of lateral infection throughout the tumor mass (38), or a combination of the two (39) may be used in future Ad-based CD vectors to overcome these problems. Whereas both Ad gene transfer efficiency and inherent 5-FU sensitivity contributed significantly to CD/5-FC VDEPT efficacy in vitro, inherent 5-FU sensitivity was the more important factor (␤ ⫽ 0.48 versus 0.78, respectively; Table 1). There is, therefore, no irony in the fact that CD/5-FC GDEPT has primarily been investigated in preclinical models of both GI [bile duct (4), colon (8), and pancreas (40)] and non-GI [breast (41) and SCCHN (42)] malignancies for which systemic 5-FU has a clearly defined role in clinical management (2). However, these tumors, particularly colon and breast cancers, are frequently metastatic on initial presentation (9, 43). Whereas systemic CD/5-FC has been investigated for treatment of metastatic disease (44), current Ad-based VDEPT technology is most likely to be efficacious on direct injection into a tumor mass (45) or surgically voided tumor cavity (46). Thus, it is better suited to use in low-stage disease, where local control is possible, such as cutaneous SCCHN (10) or recurrent and/or hormone-refractory localized prostate cancer (11), or in high-grade tumors that display locally invasive growth and rarely metastasize, such as gliomas (12). Based on the demonstrated efficacy of CD/5-FC VDEPT in animal models of GI tumors and the equivalent 5-FU in vitro sensitivity profile of both GI and non-GI tumor cells, the above-mentioned studies provided the rationale for further evaluation of CD/5-FC VDEPT in tumors not traditionally treated with systemic 5-FU. CD/ 5-FC VDEPT is particularly well suited to the treatment of gliomas because 90% of these tumors recur within 2 cm of the resection margin and rarely metastasize outside the CNS (47). Gliomas remain one of the more therapeutically intractable types of cancer; the median survival is ⬍12 months even with surgery, XRT, and chemotherapy (48). Whereas the roles of cytoreductive surgery and external beam XRT are established in the clinical management of newly diagnosed and recurrent gliomas, the value of adjunctive systemic chemotherapy is less well defined. BCNU, a lipophilic, non-phase-specific alkylating agent, is the most commonly used systemic chemotherapeutic and, until 1999, was the only FDA-approved drug for malignant gliomas (49). Prolongation of patient survival by nitrosourea-based mono- and poly-chemotherapy has been modest, primarily due to the emergence of drug-resistant cells. FDA approval of temozolomide in 1999 for recurrent anaplastic astrocytoma marked the biggest change in neurooncology in 20 years (50), but its role in the management of malignant gliomas is far from clear (51). Combination systemic chemotherapy and chemoradiotherapy with BCNU and non-cross-resistant, phasespecific drugs such as 5-FU, 5-bromodeoxyuridine, 5-iododeoxyuridine, hydroxyurea, and other antimetabolites has been investigated in various clinical trials for gliomas over the past 40 years (see review, Refs. 52 and 53). Whereas no clear role for these drugs has been established, antimetabolites remain attractive due to their capacities to potentiate the effects of XRT (54) and to overcome or circumvent BCNU resistance (49). Narrow therapeutic indices, the intact bloodbrain barrier at the actively proliferating tumor margin, and the relatively low growth fraction of gliomas have limited the utility of antimetabolites for gliomas and precluded clinical evaluation of their direct antineoplastic effect (49). Alternative drug delivery approaches may overcome the obstacles associated with systemic 5-FU chemotherapy. Interstitial chemotherapy has been previously investigated in gliomas, and sustained release, biodegradable BCNU polymers (Gliadel) have recently been approved by the FDA for use in recurrent gliomas (55). This approach has also been attempted for antimetabolites such as 5-FU. Menei et al. (56) recently reported results of a Phase I trial investigating the

combination of interstitial 5-FU with concurrent external beam XRT. High levels of 5-FU in the CSF, but not blood, were seen up to 30 days after surgical implantation of biodegradable microspheres containing 5-FU, with minimal neurological side effects. Although only eight patients were studied, there was a distinct trend toward improved local control and prolonged survival in patients receiving high-dose interstitial 5-FU with concurrent XRT. CD/5-FC GDEPT is an attractive alternative to polymer-based interstitial 5-FU. Unlike 5-FU, 5-FC is amenable to oral administration due to its efficient GI absorption, readily penetrates the bloodbrain barrier after systemic administration, and is FDA-approved for CNS fungal chemotherapy (57). Thus, the third aim of this study was to investigate the in vivo efficacy of intratumoral 5-FU CD/5-FC VDEPT in experimental human gliomas. In all three intracranial C.B17 SCID mouse models, combined intratumoral injection of AdCMVCD with systemic 5-FC significantly prolonged survival (P ⬍ 0.05). This effect was virus dose dependent (Fig. 5) and specific for the CD/5-FC combination (Fig. 6) because treatment with neither AdCMVhSSTr2 ⫾ 5-FC nor AdCMVCD without 5-FC significantly prolonged survival (P ⬎ 0.05). Interestingly, the mean ILS in two of three models tested (U87MG and D54MG; Table 2) closely mirrored the in vitro response of these cells to both 5-FU (Fig. 1) and AdCMVCD/5-FC (Fig. 3). Results with U251MG cells were the exception, displaying an intermediate ILS but superior sensitivity to both 5-FU and AdCMVCD/5-FC. Perhaps additional in vivo experiments with this model or modification of the treatment regimen would yield results that correlate more closely with in vitro sensitivity profiles. Alternatively, there may be undetected differences in the growth profiles of U251MG cells in vitro and in vivo that may explain the differences in sensitivity. Future testing of AdCMVCD/5-FC therapy in multiple different tumor models, such as pancreatic and prostate cancer, should permit a more definitive evaluation of this correlation. A recent study by Lambin et al. (58) quantified the therapeutic gain that may be obtained with CD/5-FC GDEPT and concurrent fractionated XRT. Based on published pharmacokinetic data for 5-FC, these authors suggested that, assuming only 1–3% local CD conversion efficiency, maintenance of serum 5-FC at 25–100 ␮g/ml (standard antifungal treatment values) may permit intratumoral 5-FU production of 0.6 – 0.9 ␮g/ml, a level previously shown to achieve radiosensitization values of 1.1–1.2 in vitro (59). This level of sensitization, when combined with 60 Gy of fractionated XRT (2 Gy/fraction), might increase local tumor control 20 – 40%. This level of increased local control may be detected in a randomized, two-arm clinical trial involving 260 or 60 total patients, respectively (60). Data in Fig. 1 demonstrate that the mean 5-FU IC50 values for three of four human glioma cell lines tested were in this range (1.17–1.44 ␮g/ml), and the mean 5-FU IC50 for the other cell line, U251MG, was even lower (0.35 ␮g/ml). 19F magnetic resonance spectroscopy studies are currently under way to assess intratumoral CD conversion efficiency on intratumoral AdCMVCD infection. In addition, concurrent CD/5-FC VDEPT and fractionated XRT studies with these cell lines and tumor xenografts are currently ongoing to directly assess the potential gain of this approach for experimental malignant gliomas. GDEPT is also potentially more flexible than polymer delivery systems because additional genes may be included to potentiate the GDEPT effect. For example, we and others (61– 63) have conducted preclinical GDEPT studies to investigate the combination of CD and UPRT, an enzyme that catalyzes production of (5fluoro-)UMP in the first step of the (fluoro)uracil salvage pathway and is down-regulated in most 5-FC-resistant fungi (57). Coexpression of CD and UPRT has been shown to increase the toxic effect of 5-FC by 1–3 orders of magnitude over treatment with CD/5-FC alone (61– 63). Preclinical studies with CD/UPRT-based

778

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

VDEPT and combination chemotherapy (BCNU, hydroxyurea, irinotecan, and leucovorin) and chemoradiotherapy are currently ongoing to maximize the potential of CD/5-FC GDEPT for malignant gliomas. Such manipulation of downstream 5-FU metabolism and modulation of its cytotoxic effects via adjunctive therapies should improve efficacy because inherent 5-FU sensitivity was shown to be the more important factor in Ad-mediated CD/5-FC VDEPT (Table 1). This study represents the first study to use CD/5-FC VDEPT in experimental human gliomas and demonstrates the efficacy of intratumoral 5-FU produced via a single intratumoral injection of AdCMVCD and a 7-day b.i.d. regimen of systemic 5-FC (500 mg/kg). These efficacy results compare favorably with other reports of CD/5-FC GDEPT and suggest that future protocol optimization of viral injection (fractionated dosing), 5-FC regimen (prolonged administration), and concurrent fractionated external beam XRT may be even more efficacious, providing the rationale for further investigation of this approach via clinical trials. This study is also the first to systematically evaluate VDEPT in multiple murine models of the same tumor type using cell lines displaying distinct genetic, morphological, and phenotypic properties. Results demonstrate that response to CD/5-FC VDEPT is variable and that both Ad gene transfer efficiency and inherent 5-FU sensitivity are critical factors in determining tumor response. Most importantly, these findings are likely to reflect the variable responses to CD/5-FC VDEPT that may be encountered clinically in human tumors that are even more heterogeneous. ACKNOWLEDGMENTS We thank Robert Fulgham, Barbara Krum, Suzanne Randall, and Jennifer Trevor for technical assistance and Martin Johnson, Joe Kelly, and Burt Nabors for fruitful discussions and critical reading of the manuscript.

REFERENCES 1. Heidelberger, C., Chaudhuri, N. K., Danenberg, P., Mooren, D., Griesbach, L., Duschinsky, R., Schnitzer, R. J., Pleven, E., and Scheiner, J. Fluorinated pyrimidines, a new class of tumor inhibitory compounds. Nature (Lond.), 179: 663– 666, 1957. 2. Sotos, G. A., Grogan, L., and Allegra, C. J. Preclinical and clinical aspects of biomodulation of 5-fluorouracil. Cancer Treat. Rev., 20: 11– 49, 1994. 3. Springer, C. J., and Niculescu-Duvaz, I. Prodrug-activating systems in suicide gene therapy. J. Clin. Investig., 105: 1161–1167, 2000. 4. Pederson, L. C., Buchsbaum, D. J., Vickers, S. M., Kancharla, S. R., Mayo, M. S., Curiel, D. T., and Stackhouse, M. A. Molecular chemotherapy combined with radiation therapy enhances killing of cholangiocarcinoma cells in vitro and in vivo. Cancer Res., 57: 4325– 4332, 1997. 5. Wallace, P. M., MacMaster, J. F., Smith, V. F., Kerr, D. E., Senter, P. D., and Cosand, W. L. Intratumoral generation of 5-fluorouracil mediated by an antibody-cytosine deaminase conjugate in combination with 5-fluorocytosine. Cancer Res., 54: 2719 – 2723, 1994. 6. Nishiyama, T., Kawamura, Y., Kawamoto, K., Matsumura, H., Yamamoto, N., Ito, T., Ohyama, A., Katsuragi, T., and Sakai, T. Antineoplastic effects in rats of 5-fluorocytosine in combination with cytosine deaminase capsules. Cancer Res., 45: 1753– 1761, 1985. 7. Huber, B. E., Austin, E. A., Richards, C. A., Davis, S. T., and Good, S. S. Metabolism of 5-fluorocytosine to 5-fluorouracil in human colorectal tumor cells transduced with the cytosine deaminase gene: significant antitumor effects when only a small percentage of tumor cells express cytosine deaminase. Proc. Natl. Acad. Sci. USA, 91: 8302– 8306, 1994. 8. Stackhouse, M. A., Pederson, L. C., Grizzle, W. E., Curiel, D. T., Gebert, J., Haack, K., Vickers, S. M., Mayo, M. S., and Buchsbaum, D. J. Fractionated radiation therapy in combination with adenoviral delivery of the cytosine deaminase gene and 5-fluorocytosine enhances cytotoxic and antitumor effects in human colorectal and cholangiocarcinoma models. Gene Ther., 7: 1019 –1026, 2000. 9. Skibber, J. M., Minsky, B. D., and Hoff, P. M. Cancer of the colon. In: V. T. DeVita, S. Hellman, and S. A. Rosenberg (eds.), Cancer: Principles and Practice of Oncology, 6th ed., pp. 1216 –1270. Philadelphia: Lippincott Williams & Wilkins, 2001. 10. Khuri, F. R., Shin, D. M., Glisson, B. S., Lippman, S. M., and Hong, W. K. Treatment of patients with recurrent or metastatic squamous cell carcinoma of the head and neck: current status and future directions. Semin. Oncol., 27: 25–33, 2000. 11. Steiner, M. S., and Gingrich, J. R. Gene therapy for prostate cancer: where are we now? J. Urol., 164: 1121–1136, 2000. 12. Alavi, J. B., and Eck, S. L. Gene therapy for malignant gliomas. Hematol. Oncol. Clin. N. Am., 12: 617– 629, 1998.

13. Aboody, K. S., Brown, A., Rainov, N. G., Bower, K. A., Liu, S. X., Yang, W., Small, J. E., Herrlinger, U., Ourednik, V., Black, P. M., Breakefield, X. O., and Snyder, E. Y. Neural stem cells display extensive tropism for pathology in adult brain: evidence from intracranial gliomas. Proc. Natl. Acad. Sci. USA, 97: 12846 –12851, 2000. 14. Aghi, M., Kramm, C. M., Chou, T. C., Breakefield, X. O., and Chiocca, E. A. Synergistic anticancer effects of ganciclovir/thymidine kinase and 5-fluorocytosine/ cytosine deaminase gene therapies. J. Natl. Cancer Inst. (Bethesda), 90: 370 –380, 1998. 15. Dong, Y., Wen, P., Manome, Y., Parr, M., Hirshowitz, A., Chen, L., Hirschowitz, E. A., Crystal, R., Weichselbaum, R., Kufe, D. W., and Fine, H. A. In vivo replicationdeficient adenovirus vector-mediated transduction of the cytosine deaminase gene sensitizes glioma cells to 5-fluorocytosine. Hum. Gene Ther., 7: 713–720, 1996. 16. Ge, K., Xu, L., Zheng, Z., Xu, D., Sun, L., and Liu, X. Transduction of cytosine deaminase gene makes rat glioma cells highly sensitive to 5-fluorocytosine. Int. J. Cancer, 71: 675– 679, 1997. 17. Rogulski, K. R., Zhang, K., Kolozsvary, A., Kim, J. H., and Freytag, S. O. Pronounced antitumor effects and tumor radiosensitization of double suicide gene therapy. Clin. Cancer Res., 3: 2081–2088, 1997. 18. Chang, J. W., Lee, H., Kim, E., Lee, Y., Chung, S. S., and Kim, J. H. Combined antitumor effects of an adenoviral cytosine deaminase/thymidine kinase fusion gene in rat C6 glioma. Neurosurgery, 47: 931–938, 2000. 19. Stegman, L. D., Rehemtulla, A., Hamstra, D. A., Rice, D. J., Jonas, S. J., Stout, K. L., Chenevert, T. L., and Ross, B. D. Diffusion MRI detects early events in the response of a glioma model to the yeast cytosine deaminase gene therapy strategy. Gene Ther., 7: 1005–1010, 2000. 20. Barth, R. F. Rat brain tumor models in experimental neuro-oncology: the 9L, C6, T9, F98, RG2 (D74), RT-2 and CNS-1 gliomas. J. Neuro-Oncol., 36: 91–102, 1998. 21. Rogers, B. E., McLean, S. F., Kirkman, R. L., Della Manna, D., Bright, S. J., Olsen, C. C., Myracle, A. D., Mayo, M. S., Curiel, D. T., and Buchsbaum, D. J. In vivo localization of [111In]-DTPA-D-Phe1-octreotide to human ovarian tumor xenografts induced to express the somatostatin receptor subtype 2 using an adenoviral vector. Clin. Cancer Res., 5: 383–393, 1999. 22. Miller, C. R., Kelly, F. J., Williams, C. W., Curiel, D. T., Buchsbaum, D. J., and Gillespie, G. Y. A quantitative method for determining adenovirus gene transfer efficiency with established and primary human tumor cell cultures. Mol. Ther., 3: S165, 2001. 23. Bradley, J. D., Kataoka, Y., Advani, S., Chung, S. M., Arani, R. B., Gillespie, G. Y., Whitley, R. J., Markert, J. M., Roizman, B., and Weichselbaum, R. R. Ionizing radiation improves survival in mice bearing intracranial high-grade gliomas injected with genetically modified herpes simplex virus. Clin. Cancer Res., 5: 1517–1522, 1999. 24. Andreansky, S., Soroceanu, L., Flotte, E. R., Chou, J., Markert, J. M., Gillespie, G. Y., Roizman, B., and Whitley, R. J. Evaluation of genetically engineered herpes simplex viruses as oncolytic agents for human malignant brain tumors. Cancer Res., 57: 1502–1509, 1997. 25. Andreansky, S. S., He, B., Gillespie, G. Y., Soroceanu, L., Markert, J., Chou, J., Roizman, B., and Whitley, R. J. The application of genetically engineered herpes simplex viruses to the treatment of experimental brain tumors. Proc. Natl. Acad. Sci. USA, 93: 11313–11318, 1996. 26. Statsoft Electronic Statistics Textbook. Tulsa, OK: StatSoft, 2001. 27. Miller, C. R., Gustin, A. N., Buchsbaum, D. J., Vickers, S. M., Manne, U., Grizzle, W. E., Cloud, G. A., Diasio, R. B., and Johnson, M. R. Quantitation of cytosine deaminase mRNA by real time RT-PCR: a sensitive method for assessing 5-fluorocytosine toxicity in vitro. Anal. Biochem., in press, 2002. 28. Weinstein, J. N., Myers, T. G., O’Connor, P. M., Friend, S. H., Fornace, A. J., Jr., Kohn, K. W., Fojo, T., Bates, S. E., Rubinstein, L. V., Anderson, N. L., Buolamwini, J. K., van Osdol, W. W., Monks, A. P., Scudiero, D. A., Sausville, E. A., Zaharevitz, D. W., Bunow, B., Viswanadhan, V. N., Johnson, G. S., Wittes, R. E., and Paull, K. D. An information-intensive approach to the molecular pharmacology of cancer. Science (Wash. DC), 275: 343–349, 1997. 29. Miller, C. R., Buchsbaum, D. J., Williams, C. W., Kiss, P., Elgavish, G., Grizzle, W. E., Diasio, R. B., Johnson, M. R., and Gillespie, G. Y. Cytosine deaminase/5fluorocytosine therapy for gastrointestinal (GI) and non-GI malignancies. Proc. Am. Assoc. Cancer Res., 42: 28, 2001. 30. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg, R. W. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science (Wash. DC), 275: 1320 –1323, 1997. 31. Tomko, R. P., Xu, R., and Philipson, L. HCAR and MCAR: the human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA, 94: 3352–3356, 1997. 32. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. Integrins ␣v␤3 and ␣v␤5 promote adenovirus internalization but not virus attachment. Cell, 73: 309 –319, 1993. 33. Bergelson, J. M. Receptors mediating adenovirus attachment and internalization. Biochem. Pharmacol., 57: 975–979, 1999. 34. Nemerow, G. R. Cell receptors involved in adenovirus entry. Virology, 274: 1– 4, 2000. 35. Nemerow, G. R., and Stewart, P. L. Role of ␣v integrins in adenovirus cell entry and gene delivery. Microbiol. Mol. Biol. Rev., 63: 725–734, 1999. 36. Stohrer, M., Boucher, Y., Stangassinger, M., and Jain, R. K. Oncotic pressure in solid tumors is elevated. Cancer Res., 60: 4251– 4255, 2000. 37. Krasnykh, V. N., Douglas, J. T., and van Beusechem, V. W. Genetic targeting of adenoviral vectors. Mol. Ther., 1: 391– 405, 2000. 38. Heise, C., and Kirn, D. H. Replication-selective adenoviruses as oncolytic agents. J. Clin. Investig., 105: 847– 851, 2000.

779

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

CD/5-FC GENE THERAPY FOR HUMAN GLIOMAS

39. Rogulski, K. R., Wing, M. S., Paielli, D. L., Gilbert, J. D., Kim, J. H., and Freytag, S. O. Double suicide gene therapy augments the antitumor activity of a replicationcompetent lytic adenovirus through enhanced cytotoxicity and radiosensitization. Hum. Gene Ther., 11: 67–76, 2000. 40. Evoy, D., Hirschowitz, E. A., Naama, H. A., Li, X. K., Crystal, R. G., Daly, J. M., and Lieberman, M. D. In vivo adenoviral-mediated gene transfer in the treatment of pancreatic cancer. J. Surg. Res., 69: 226 –231, 1997. 41. Li, Z., Shanmugam, N., Katayose, D., Huber, B., Srivastava, S., Cowan, K., and Seth, P. Enzyme/prodrug gene therapy approach for breast cancer using a recombinant adenovirus expressing Escherichia coli cytosine deaminase. Cancer Gene Ther., 4: 113–117, 1997. 42. Hamstra, D. A., Rice, D. J., Fahmy, S., Ross, B. D., and Rehemtulla, A. Enzyme/ prodrug therapy for head and neck cancer using a catalytically superior cytosine deaminase. Hum. Gene Ther., 10: 1993–2003, 1999. 43. Winer, E. P., Morrow, M., Osborne, C. K., and Harris, J. R. Malignant tumors of the breast. In: V. T. DeVita, S. Hellman, and S. A. Rosenberg (eds.), Cancer: Principles and Practice of Oncology, 6th ed., pp. 1651–1716. Philadelphia: Lippincott Williams & Wilkins, 2001. 44. Topf, N., Worgall, S., Hackett, N. R., and Crystal, R. G. Regional ‘pro-drug’ gene therapy: intravenous administration of an adenoviral vector expressing the E. coli cytosine deaminase gene and systemic administration of 5-fluorocytosine suppresses growth of hepatic metastasis of colon carcinoma. Gene Ther., 5: 507–513, 1998. 45. Khuri, F. R., Nemunaitis, J., Ganly, I., Arseneau, J., Tannock, I. F., Romel, L., Gore, M., Ironside, J., MacDougall, R. H., Heise, C., Randlev, B., Gillenwater, A. M., Bruso, P., Kaye, S. B., Hong, W. K., and Kirn, D. H. A controlled trial of intratumoral ONYX-015, a selectively-replicating adenovirus, in combination with cisplatin and 5-fluorouracil in patients with recurrent head and neck cancer. Nat. Med., 6: 879 – 885, 2000. 46. Eck, S. L., Alavi, J. B., Alavi, A., Davis, A., Hackney, D., Judy, K., Mollman, J., Phillips, P. C., Wheeldon, E. B., and Wilson, J. M. Treatment of advanced CNS malignancies with the recombinant adenovirus H5.010RSVTK: a Phase I trial. Hum. Gene Ther., 7: 1465–1482, 1996. 47. Hochberg, F. H., and Pruitt, A. Assumptions in the radiotherapy of glioblastoma. Neurology, 30: 907–911, 1980. 48. Levin, V. A., Leibel, S. A., and Gutin, P. H. Neoplasms of the central nervous system. In: V. T. DeVita, S. Hellman, and S. A. Rosenberg (eds.), Cancer: Principles and Practice of Oncology, 6th ed., pp. 2100 –2160. Philadelphia: Lippincott Williams & Wilkins, 2001. 49. Patchell, R. A. Chemotherapy of primary brain tumors. In: M. C. Perry (ed.), The Chemotherapy Source Book, 2nd ed., pp. 1071–1081. Baltimore: Williams & Wilkins, 1996. 50. Yung, W. K., Prados, M. D., Yaya-Tur, R., Rosenfeld, S. S., Brada, M., Friedman, H. S., Albright, R., Olson, J., Chang, S. M., O’Neill, A. M., Friedman, A. H., Bruner, J., Yue, N., Dugan, M., Zaknoen, S., and Levin, V. A. Multicenter Phase II trial of

51. 52.

53.

54. 55.

56.

57.

58.

59.

60.

61.

62.

63.

temozolomide in patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. Temodal Brain Tumor Group. J. Clin. Oncol., 17: 2762–2771, 1999. Batchelor, T. Temozolomide for malignant brain tumours. Lancet, 355: 1115–1116, 2000. Mahaley, M. S., Jr. Neuro-oncology index and review (adult primary brain tumors). Radiotherapy, chemotherapy, immunotherapy, photodynamic therapy. J. NeuroOncol., 11: 85–147, 1991. Roth, W., and Weller, M. Chemotherapy and immunotherapy of malignant glioma: molecular mechanisms and clinical perspectives. Cell. Mol. Life Sci., 56: 481–506, 1999. Pu, A. T., Robertson, J. M., and Lawrence, T. S. Current status of radiation sensitization by fluoropyrimidines. Oncology (Huntingt.), 9: 707–714, 1995. Brem, H., Piantadosi, S., Burger, P. C., Walker, M., Selker, R., Vick, N. A., Black, K., Sisti, M., Brem, S., and Mohr, G. Placebo-controlled trial of safety and efficacy of intraoperative controlled delivery by biodegradable polymers of chemotherapy for recurrent gliomas. The Polymer-brain Tumor Treatment Group. Lancet, 345: 1008 – 1012, 1995. Menei, P., Venier, M. C., Gamelin, E., Saint-Andre, J. P., Hayek, G., Jadaud, E., Fournier, D., Mercier, P., Guy, G., and Benoit, J. P. Local and sustained delivery of 5-fluorouracil from biodegradable microspheres for the radiosensitization of glioblastoma: a pilot study. Cancer (Phila.), 86: 325–330, 1999. Vermes, A., Guchelaar, H. J., and Dankert, J. Flucytosine: a review of its pharmacology, clinical indications, pharmacokinetics, toxicity and drug interactions. J. Antimicrob. Chemother., 46: 171–179, 2000. Lambin, P., Nuyts, S., Landuyt, W., Theys, J., De Bruijn, E., Anne, J., Van Mellaert, L., and Fowler, J. The potential therapeutic gain of radiation-associated gene therapy with the suicide gene cytosine deaminase. Int. J. Radiat. Biol., 76: 285–293, 2000. Byfield, J. E., Calabro-Jones, P., Klisak, I., and Kulhanian, F. Pharmacologic requirements for obtaining sensitization of human tumor cells in vitro to combined 5fluorouracil or ftorafur and X rays. Int. J. Radiat. Oncol. Biol. Phys., 8: 1923–1933, 1982. Bentzen, S. M. Towards evidence based radiation oncology: improving the design, analysis, and reporting of clinical outcome studies in radiotherapy. Radiother. Oncol., 46: 5–18, 1998. Adachi, Y., Tamiya, T., Ichikawa, T., Terada, K., Ono, Y., Matsumoto, K., Furuta, T., Hamada, H., and Ohmoto, T. Experimental gene therapy for brain tumors using adenovirus-mediated transfer of cytosine deaminase gene and uracil phosphoribosyltransferase gene with 5-fluorocytosine. Hum. Gene Ther., 11: 77– 89, 2000. Erbs, P., Regulier, E., Kintz, J., Leroy, P., Poitevin, Y., Exinger, F., Jund, R., and Mehtali, M. In vivo cancer gene therapy by adenovirus-mediated transfer of a bifunctional yeast cytosine deaminase/uracil phosphoribosyltransferase fusion gene. Cancer Res., 60: 3813–3822, 2000. Williams, C. W., Buchsbaum, D. J., and Miller, C. R. Uracil phosphoribosyltransferase potentiates 5-fluorouracil and cytosine deaminase/5-fluorocytosine cytotoxicity in prostate cancer. Proc. Am. Assoc. Cancer Res., 42: 455, 2001.

780

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.

Intratumoral 5-Fluorouracil Produced by Cytosine Deaminase/5-Fluorocytosine Gene Therapy Is Effective for Experimental Human Glioblastomas C. Ryan Miller, Christopher R. Williams, Donald J. Buchsbaum, et al. Cancer Res 2002;62:773-780.

Updated version

Cited articles Citing articles

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/62/3/773

This article cites 57 articles, 20 of which you can access for free at: http://cancerres.aacrjournals.org/content/62/3/773.full.html#ref-list-1 This article has been cited by 15 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/62/3/773.full.html#related-urls

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cancerres.aacrjournals.org on August 12, 2015. © 2002 American Association for Cancer Research.