D2001 Nature Publishing Group 0929-1903/01/$17.00/+0 www.nature.com/cgt

Induction of therapeutic antitumor antiangiogenesis by intratumoral injection of genetically engineered - producing Semliki Forest virus Ryuya Yamanaka,1,2 Susan A. Zullo,1 Jay Ramsey,1 Masafumi Onodera,1 Ryuichi Tanaka,2 Michael Blaese,1 and Kleanthis G. Xanthopoulos1 1Clinical Gene Therapy Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD 20892; and 2Department of Neurosurgery, Brain Research Institute, Niigata University, Niigata, Japan.

Antiangiogenic therapy using Semliki Forest virus (SFV) carrying Endostatin gene for malignant brain tumor was investigated to improve the therapeutic efficacy. The efficiency of SFV-mediated gene delivery was first evaluated for B 16 cells and compared with the efficiency in cells of endothelial origin (HMVECs). HMVECs are more susceptible to SFV infection than B 16 cells. For the in vivo treatment model, phosphate-buffered saline, SFV-LacZ, retrovirus vector GCsap-Endostatin, and SFV-Endostatin were injected to mice bearing B 16 brain tumors. A very significant inhibition of tumor growth was observed in the group that had been treated with SFV-Endostatin. A marked reduction of intratumoral vascularization was seen in the tumor sections from the SFV-Endostatin group compared with tumor sections from the SFV-LacZ or GCsap-Endostatin groups. Moreover, at day 7 after intravenous administration of SFV-Endostatin, the serum level of endostatin was augmented more than 3-fold compared to that after intravenous administration of GCsap-Endostatin. The results indicated that treatment with SFV-Endostatin inhibited the with established tumors. Gene therapy with Endostatin delivered via SFV may be a candidate for the development of new therapy for brain tumors. Cancer Gene Therapy (2001) 8, 796–802

Key words: Endostatin; Semliki Forest virus; antiangiogenic therapy; malignant brain tumor.

he prognosis of patients with malignant brain tumors has not achieved satisfactory results, we investigated the ability Tnot improved much in recent years. Current treatments, of SFV vector to elevate circulating endostatin levels in mice including surgery, radiation, and chemotherapy, do not as an approach to treatment of brain tumors. change the prognosis of malignant brain tumors. Therefore, Viral vectors commonly used for therapeutic DNA the development of novel therapies for malignant brain delivery include replication-deficient forms of adenovirus tumors is essential. and retroviruses. Although retroviruses have the advantage Tumors are dependent on angiogenesis for growth,1 and of mediating stable gene transfer with a low potential for therefore the treatment of brain tumor patients with inhibitors immunogenicity, retrovirus vector delivery systems have of angiogenesis is an attractive novel strategy. Many of these some problems with respect to therapeutic use, including biological agents are not stable in vitro and are difficult to difficulties in producing high titers of retroviruses, the fact produce in large quantities.2 For these reasons, antiangio- that only actively dividing cells are capable of being infected genic gene therapy has been proposed as an alternative to and the possibility of insertional mutagenesis.5 The adeno- treatment with recombinant .3,4 Gene therapy of virus vector system, although capable of delivering genes cancer has traditionally attempted to target tumor cells with with high efficiency to a wide spectrum of nondividing cells toxic or immunogenic transgenes. The apparent lack of in vivo,6 unfortunately produces only transient expression of toxicity of antiangiogenic agents suggests that maximizing various gene products. This transient expression may result circulating levels of these endogenous agents may be a more from a strong immune response of host cells against the appropriate gene therapy strategy.3 The goal of this therapy adenovirus.7 Semliki Forest virus (SFV), a member of the would be to create a situation where the host becomes an Alphaviruses, has received considerable attention for use as a endogenous factory producing high circulating levels of the virus-based expression vector. The SFV expression system gene product. Because current gene therapy is mainly differs from currently available viral delivery systems in that conducted using adenovirus or retrovirus vectors and has SFV is an RNA virus and is known to generate high levels of protein expression in vitro. SFV is also less pathogenic 8 to Received June 28, 2001. humans than some other viral delivery systems. SFV is a Address correspondence and reprint requests to Dr. Ryuya Yamanaka, self-amplifying expression vector, with the advantages of Department of Neurosurgery, Brain Research Institute, Niigata University, ease of production of high-titered stocks of infectious Asahimachi-dori 1-757, Niigata City, Japan 951-8122. E-mail address: particles, ability to infect nondividing cells, and high-level 9 [email protected] expression of transgenes. We have reported the usefulness

796 Cancer Gene Therapy, Vol 8, No 10, 2001: pp 796–802 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR 797 of SFV-mediated immunogene therapy for malignant then washed with phosphate-buffered saline (PBS) and glioma.10,11 incubated in 10 mL of Opti-MEM medium (Gibco-BRL) The studies presented here were performed to evaluate the without FCS. After 24 hours, culture supernatants containing efficacy of genetically modified SFV for producing endo- recombinant SFV particles were harvested and aliquots were statin. The strategies were studied in a model of active gene stored at À808C. Retroviral particles were also generated as therapy for brain tumors. described elsewhere.12 Briefly, after cloning mEndostatin cDNA into the GCsap vector, the amphotropic packaging cell line PA317 and the gibbon ape leukemia virus packaging MATERIALS AND METHODS cell line PG13 were used to generate the retrovirus particles. Cell lines and animal models The infectivity of the recombinant viruses in BHK and B 16 cells was determined by transfer of the SFV and retroviral The B 16 cell line (derived from a C57BL/6 mouse) and particles that could transduce the -galactosidase gene. human microcapillary endothelial cells (HMVECs) were Various dilutions of virus-containing culture supernatants kindly provided by Riken Cell Bank (Tsukuba, Japan). Baby were added to adherent cells (2Â105 ) in six-well plates. hamster kidney (BHK) cells were obtained from Invitrogen After a 1-hour incubation at 378C, cells were washed with (San Diego, CA). The amphotropic packaging cell line PBS and incubated in growth medium at 378C for 24 hours. PA317 and the gibbon ape leukemia virus packaging cell line Viral infection was evaluated by X-gal staining. Briefly, PG13 were kindly provided by Dr. M. Onodera.12 The cell cells were fixed in PBS containing 0.5% glutaraldehyde for lines were grown in MEM medium (Gibco-BRL, Gaithers- 15 minutes and then washed with PBS three times. Cells burg, MD) containing 10% FCS. All cell lines were shown to were then stained with PBS containing 1 mg/mL X-gal, be free from mycoplasma contamination. All experiments 5 mM potassium ferricyanide, 5 mM potassium ferrocya- used 6- to 12-week-old female C57BL/6 mice (The Jackson nide, and 1 mM MgSO at 378C for 2 hours. Laboratory, Bar Harbor, ME), which were maintained in a 4 virus-free environment and treated in accordance with the NIH Laboratory Animal Resources Commission standards. Functional assay of virally generated endostatin The B 16 cell line was infected as above with SFV- Cloning of the murine endostatin gene mEndostatin or SFV-LacZ at various multiplicities of Murine cDNA was obtained by isolating RNA (Rneasy infection (MOIs). Supernatants were harvested 24 hours Mini Kit; Qiagen, Valencia, CA) from C57BL/6 mouse later and centrifuged as above. Supernatants were analyzed for their ability to inhibit endothelial cell proliferation, as liver and reverse transcribing it with Moloney murine 13 leukemia virus reverse transcriptase (Life Technologies, described previously. Briefly, 1000 human capillary Gaithersburg, MD). The murine endostatin cDNA was endothelial cells (HMVECs) were plated in complete cloned into the pSecTag 2 vector (Invitrogen, San Diego, medium in each well of collagen I–coated 96-well plates CA) after amplification by PCR using the following primers: sense (GATCTCTAGACCACCATGCATACT- CATCAGGACTT) and antisense (ACTGGAGAAA- GAGGTTTATCTAGCTACTAG). The plasmid DNA was amplified in DH5 cells and the signal sequence (IgK Leader) and murine endostatin sequence were confirmed (ABI Prism 310 autosequencer; PE Applied Biosystems, Foster City, CA).

Generation of SFV-Endostatin particles The plasmids pSFV3 and pSFV3-LacZ (Life Technologies) each contain an SP6 promoter, a 7-kb fragment encoding the SFV RNA replicase, and a subgenomic promoter that is bound by the SFV RNA replicase and directs synthesis of a large quantity of subgenomic RNA. The helper plasmid pSFV-Helper 2, which contains the genes for the structural (capsid, E3, E2, 6K, and E1) required for packaging of the viral genome, was used for construction of the recombinant envelope gene. Retrovirus vector GCsap was kindly provided by Dr. M. Onodera.12 The murine endostatin gene was cloned into pSFV3 from pSecTag 2-mEndostatin. SFV-mEndostatin RNA was transcribed in vitro and capped using SP6 RNA polymerase and capping analog from Life Technologies. For cotransfections of helper and expression Figure 1. Efficiency of SFV-mediated gene delivery into HMVECs RNA into BHK cells, electroporation was performed. After and B 16 cells. Both cell lines were infected with increasing MOIs of electroporation, cells were transferred to 10 mL of MEM SFV-LacZ virus, and cells expressing -galactosidase were containing 5% FCS and incubated for 12 hours. Cells were assessed 24 hours postinfection.

Cancer Gene Therapy, Vol 8, No 10, 2001 798 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR

Figure 2. Inhibition of endothelial cell proliferation in vitro. HMVECs (left) and B 16 cells (right) were infected with SFV-LacZ, SFV-mEndostatin or GCsap-mEdostatin. The cells that survived posttransfection were quantified by an XTT assay.

(Biocoat, Becton Dickinson, Bedford, MA). After overnight according to the manufacturer’s instructions. Inhibition of incubation at 378C, the medium was aspirated and replaced proliferation in each sample was calculated according to the with 20 L of the supernatant sample to be tested. Six formula: samples of each supernatant were tested. After a 20-minute incubation at 378C, 80 L of modified complete medium Inhibition ð%Þ¼ðmean A450 ðcontrolÞÀA450 containing 5% FCS and 1 ng/mL basic fibroblast growth ÂðsampleÞ=mean A450 ðcontrolÞÞ Â 100 factor (R&D Systems, Minneapolis, MN) was added. After a 72-hour incubation at 378C, proliferation was analyzed by where A450 is the absorbance at 450 nm measured in a the XTT assay (Boehringer Mannheim, Tokyo, Japan), Multiscan MCC/340 plate reader (Titertek, Huntsville, AL)

Figure 3. A: C57BL/6 mice were treated with three intratumoral injections of PBS, SFV-LacZ, SFV-mEndostatin or GCsap-mEndostatin. Treatment was initiated on day 10 postgrafting. The subcutaneous tumor volume was monitored until day 30 postgrafting. B: Survival of treated mice after intracranial tumor grafting. Mice were treated with three intratumoral injections of PBS, SFV-LacZ, SFV-mEndostatin, or GCsap- mEndostatin. Treatment was initiated on day 7 postgrafting.

Cancer Gene Therapy, Vol 8, No 10, 2001 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR 799 and control represents the value for the supernatant of sions using calipers. Tumor volume was calculated accord- uninfected cells. ing to the formula:

In vivo production of endostatin volume ¼ width2  length  0:52: Fifteen 8-week-old female mice were injected in the tail Survival of the mice was monitored for the evaluation of 9 vein with 10 pfu SFV-mEndostatin or no virus in 100 L the intracranially transplanted group. of PBS to assess the dose–response and toxicity. These mice were monitored for 14 days, euthanized, and Statistical analysis autopsied. On days 1, 3, 7, 10, and 14, blood samples were Survival curves and median survivals were determined using obtained by cardiac puncture (n=3 animals/treatment group the method of Kaplan and Meier. Survival data were per time point). Plasma endostatin levels were determined compared using Wilcoxon’s test. Student’s t test was used by ELISA. for calculating the significance of other data. Statistical significance was determined at the <.05 level. In vivo activity of endostatin In a treatment experiment, mice were injected intracranially RESULTS or subcutaneously with 105 B 16 cells in 2 L of PBS as previously described.14 Two, 7, and 14 days later, mice SFV-mediated endostatin administration in murine brain received injections of 108 pfu SFV-mEndostatin, SFV- tumor model LacZ, GCsap-mEndostatin or no virus, as described as Recombinant SFV virus particles carrying a mouse endo- above. Subcutaneous tumors were measured in two dimen- statin gene were injected into mice that had received B 16

Figure 4. Inhibition of intratumoral vascularization. A: The intratumoral vascularization was quantitatively assessed on day 30 postgrafting. Lanes: 1, PBS-injected group; 2, SFV-LacZ–injected group; 3, SFV-mEndostatin–injected group; 4, GCsap-mEndostatin–injected group. B: Immunostaining of representative tumors from the SFV-mEndostatin–injected group (a), GCsap-mEndostatin–injected group (b) and SFV- LacZ injected group (c) are shown at the same magnification.

Cancer Gene Therapy, Vol 8, No 10, 2001 800 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR

Figure 5. PBS, SFV-LacZ, SFV-mEndostatin or GCsap-mEndostatin was injected into the tail vein. Serum was harvested on day 1, 3, 7, 10 or 14 and endostatin levels were determined by ELISA. The serum level of endostatin was augmentated more than 3-fold on day 7 after intravenous administration of SFV-mEndostatin compared to the level after injection of GCsap-mEndostatin. cell transplants, while other groups of mice were inoculated infection with SFV was quantified after 4 days. In this assay, with SFV-LacZ or GCsap-mEndostatin. These three groups the proliferation of HMVECs was found to be sensitive, in a of mice were treated every week with injections of these dose-dependent manner, to infection with SFV-mEndosta- recombinant virus particle, and the development of tumors tin, with an ED50 (defined as the MOI required for a 50% was compared with that in uninfected controls. All of the inhibitory effect) in the range of 400 pfu/cell. In contrast, mice were sacrificed and complete necropsy was performed. infection of HMVECs with the control virus had only a marginal effect, even at the highest dose tested (Fig 2). B 16 Antitumor effect of SFV-mEndostatin in vitro cells were found to be less susceptible than HMVEC-1 cells SFV-LacZ and GCsap-mEndostatin were used as control to growth inhibition by SFV-mEndostatin in this assay at all viruses throughout this study. The efficiency of SFV- MOIs tested. mediated gene delivery was first evaluated for B 16 cells and compared with the efficiency in HMVECs. For this Antitumor effects of SFV-Endostatin in vivo purpose, we analyzed the proportion of cells that had been In a first experiment, mice were randomized 10 days after infected with the SFV-LacZ reporter virus 48 hours after tumor grafting, when the tumors had reached a mean addition of the virus to the cells. As shown in Figure 1, a volume of 150 mm3. Treatment was initiated at this time dose-dependent increase of -galactosidase–positive cells by intratumoral injection of recombinant virus (5Â109 pfu was observed for both cell lines, with an ID40 (the MOI per injection). Tumor growth was then monitored on day required to infect 40% of the cell population) of 10 and 50 30 postgrafting. As shown in Figure 3, a very significant infectious particles per cell (pfu/cells) for HMVECs and B inhibition of tumor growth was observed in the group that 16 cells, respectively. The conclusion that HMVECs are had been treated with SFV-Endostatin. At completion of more susceptible to SFV infection than B 16 cells was also the experiment, SFV-Endostatin–injected tumors were supported by specific anti-endostatin ELISA analysis of the much smaller than tumors solely injected with the SFV- cell supernatants after infection with SFV-mEndostatin. LacZ virus or GCsap-mEndostatin. Our finding that The ability of SFV-mediated delivery of endostatin to endostatin gene delivery by itself had no significant effect specifically block proliferation of endothelial cells in vitro on tumor growth indicates the importance of the viral was then analyzed. The number of cells that survived vector for endostatin-based antiangiogenic approaches to

Cancer Gene Therapy, Vol 8, No 10, 2001 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR 801 the treatment of the experimental B 16 brain tumor against tumors, suggesting that angiogenesis and growth of model. individual lesions are dependent on the local balance of When treatment was initiated at a later stage (i.e., 14 days angiogenic and antiangiogenic factors.3,26 postgrafting, when the tumor volume had reached a mean We have constructed a recombinant SFV expressing value of 180 mm3, SFV-Endostatin treatment caused a less biologically active murine endostatin, as shown in both in significant inhibition of tumor growth as assessed 30 days vitro and in vivo models of angiogenesis. Treatment of postgrafting. tumor-bearing mice with SFV-mEndostatin led to a tumor- Treatment with SFV-LacZ had very little or no effect on size reduction in subcutaneous and brain tumor models. the survival of mice after intracranial tumor grafting. SFV- Histological findings demonstrated decreased tumor vascu- Endostatin gene delivery extended survival in this experi- larization in the endostatin-treated animals. Difficulties in ment, as shown in Figure 3B, although only a 35% survival the large-scale production of recombinant endostatin have rate was observed in the SFV-Endostatin–treated group. hampered the clinical utilization of this molecule. In this report, we demonstrated that a single injection of SFV- SFV-Endostatin inhibits tumor angiogenesis Endostatin resulted in high endostatin serum levels that were Intratumoral vascularization was assessed immunohistolog- more than three times higher than those achieved with a ically by using anti–mouse VEGF (Santa Cruz retroviral gene delivery system. This system results in much Biotechnology, Santa Cruz, CA) and quantified by assessing higher endostatin levels than those reported for other delivery systems such as nonviral gene transfer meth- the area per microscopic field that scored positive in this 13,20,27 –29 assay (Fig 4A). As shown in Figure 4B, a marked reduction ods. Endostatin dose and inhibition of endothelial cell proliferation have been shown to be positively of intratumoral vascularization was seen in the tumor 20,30 27 sections from the SFV-Endostatin group as compared with correlated. Dhanabal et al reported that low levels of endostatin did not induce endothelial cell in vitro. tumor sections from the SFV-LacZ or GCsap-mEndostatin 31 groups. Sauter et al reported that the effective dose of recombinant endostatin protein for tumor growth inhibition was more than Serum level of endostatin was augmented after intravenous 10–20 mg/kg per day. This dose prevented metastasis administration of SFV-Endostatin completely, whereas only partial prevention of metastasis was reported when a low-expression gene delivery system PBS, SFV-LacZ, SFV-Endostatin, or GCsap-mEndostatin was used.13 was injected into the tail vein and serum was harvested SFV is being developed as a vector for expression of thereafter at days 1, 3, 7, 10, and 14. Endostatin levels in the heterologous genes and has many advantages as an serum were determined by ELISA. As shown in Figure 5, at expression vector system. In SFV systems, because helper day 7 after intravenous administration of SFV-Endostatin, RNA does not contain a packaging signal, it dose not form a the serum level of endostatin was augmented more than defective interfering particle or become packaged together 3-fold compared to that after intravenous administration of with recombinant RNA. Furthermore, replication occurs GCsap-mEndostatin. There was no toxic effect on the mice entirely in the cytoplasm of the infected cells via RNA in this experiment (data not shown). molecules, without a DNA intermediate.9 This is in contrast with retroviruses, which must enter the nucleus and integrate DISCUSSION into the host genome for initiation of vector activity. Thus, retrovirus vectors have applications for long-term expres- Surgical removal of human tumors often leads to rapid sion of foreign genes, whereas SFV is useful primarily for growth of distant metastases.15,16 These clinical observations transient high level expression. Furthermore, although have been investigated in animal tumor models where adenovirus vectors can express high levels of foreign genes, growth of small lesions is suppressed in the presence of a these systems are more complex than SFV and express many large primary tumor, and the phenomenon is termed highly antigenic virus-specific gene products, including concomitant tumor resistance.17,18 O’Reilly et al 19 have structural proteins.32 In contrast, the current SFV vector proposed that a primary tumor, although capable of expresses only the four viral replicase proteins (ns P 1–4) stimulating angiogenesis in its own vascular bed, simulta- required for RNA amplification in the infected cells. This neously inhibits angiogenesis in vascular beds of secondary self-replicating RNAwas shown to produce more than 200% metastatic lesions. Soluble factors produced by the primary the amount of antigen produced using a conventional DNA tumor are responsible for suppression of distant tumor immunogen.33 This system has been found to express growth,19 –21 and two such factors, identified using murine significant quantities of heterologous proteins in vitro 34 models, are angiostatin 19 and endostatin.20 and in vivo.8 Endostatin is a C-terminal fragment of collagen 18 that In summary, we have shown that high circulating levels was first purified from conditioned medium obtained from of biologically active endostatin could be achieved through the EOMA murine hemangioendothelioma cell line.20 This SFV-mediated gene therapy. In the present form, endo- factor inhibits the growth of endothelial cells in vitro, and statin itself does not provide a cure for brain tumors; systemic administration of purified recombinant endostatin however, it may be useful in combination with other has been shown to inhibit the growth of established tumors treatment modalities. Thus, the self-replicating SFV and metastases in vivo.22 –25 Those experiments demonstra- system could serve as a powerful tool for treating ted that endostatin has intrinsic antiangiogenic activity malignant brain tumors.

Cancer Gene Therapy, Vol 8, No 10, 2001 802 YAMANAKA, ZULLO, RAMSEY, ET AL: ANTIANGIOGENIC THERAPY FOR MALIGNANT BRAIN TUMOR

CONCLUSIONS Research and Therapy. Raven Press, New York, 1997, pp. 185–197. A treatment protocol using SFV-endostatin showed anti- 16. Clark WH Jr, Elder DE, Guerry D IV, et al. Model predicting tumor effects in vitro and in vivo. SFV-mEndostatin–treated survival in stage I melanoma based on tumor progression. J mice also showed decreased tumor angiogenesis. The serum Natl Cancer Inst. 1989;81:1893–1904. 17. Prehn RT. The inhibition of tumor growth by tumor mass. level of endostatin was more than three times higher after Cancer Res. 1991;51:2–4. intravenous administration of SFV-mEndostatin than after 18. Prehn RT. Two competing influences that may explain similar administration of endostatin-expressing retrovirus concomitant tumor resistance. Cancer Res. 1993;53:3266– vector GCsap-mEndostatin. Thus, the self-replicating SFV 3269. system may provide a novel approach for the treatment of 19. O’Reilly MS, Holmgren L, Shing Y, et al. Angiostatin: a malignant brain tumors. novel that mediates the suppression of metastases by a Lewis lung carcinoma. Cell. 1994;79: 315–328. 20. O’Reilly MS, Boehm T, Shing Y, et al. Endostatin: an REFERENCES endogenous inhibitor of angiogenesis and tumor growth. Cell 1997;88:277–285. 1. Folkman J. Angiogenesis in cancer, vascular, rheumatoid and 21. Volpert OV, Lawler J, Bouck NP. A human fibrosarcoma other disease. Nat Med. 1995;1:27–31. inhibits systemic angiogenesis and the growth of experimental 2. Crystal RG. The body as a manufacturer of endostatin. Nat metastases via -1. Proc Natl Acad Sci USA. Biotechnol. 1999;17:336–337. 1998;95:6343–6348. 3. Folkman J. Antiangiogenic gene therapy. Proc Natl Acad Sci 22. Holmgren L, O’Reilly MS, Folkman J. Dormancy of USA. 1998;95:9064–9066. micrometastases: balanced proliferation and apoptosis in 4. Kong HL, Crystal RG. Gene therapy strategies for tumor the presence of angiogenesis suppression. Nat Med. 1995; antiangiogenesis. J Natl Cancer Inst. 1998;90:273–286. 1:149–153. 5. Mulligan RC. The basic science of gene therapy. Science. 23. O’Reilly MS, Holmgren L, Chen C, Folkman J. Angiostatin 1993;260:926–932. induces and sustains dormancy of human primary tumors in 6. Engelhardt JF, Simon RH, Yang Y, et al. Adenovirus- mice. Nat Med. 1996;2:689–692. mediated transfer of the CFTR gene to lung of nonhuman 24. Sim BK, O’Reilly MS, Liang H, et al. A recombinant human primates: biological efficacy study. Hum Gene Ther. 1993; angiostatin protein inhibits experimental primary and meta- 4:759–769. static cancer. Cancer Res. 1997;57:1329–1334. 7. Engelhardt JF, Yang Y, Stratford-Perricaudet LD, et al. Direct 25. Boehm T, Folkman J, Browder T, O’Reilly MS. Antiangio- gene transfer of human CFTR into human bronchial epithelia genic therapy of experimental cancer does not induce acquired of xenografts with E-deleted adenoviruses. Nat Genet. drug resistance. Nature. 1997;390:404–407. 1993;4:27–34. 26. Hanahan D, Folkman J. Patterns and emerging mechanisms of 8. Piper RC, Slot JW, Li G, Stahl PD, James DE. Recombinant the angiogenic switch during tumorigenesis. Cell. Sindbis virus as an expression system for cell biology. Methods 1996;86:353–364. Cell Biol. 1994;43:55–78. 27. Dhanabal M, Ramchandran R, Volk R, et al. Endostatin: yeast 9. Strauss JH, Strauss EG. The alphaviruses: gene expression, production, mutants, and antitumor effect in renal cell replication, and evolution. Microbiol Rev. 1994;58:491–562. carcinoma. Cancer Res. 1999;59:189–197. 10. Yamanaka R, Zullo SA, Tanaka R, Ramsey J, Blaese RM, 28. Yamaguchi N, Anand-Apte B, Lee M, et al. Endostatin inhibits Xanthopoulos KG. Induction of a therapeutic antitumor VEGF-induced endothelial cell migration and tumor growth immunologic response by intratumoral injection of genetically independently of zinc binding. EMBO J. 1999;18:4414–4423. engineered Semliki Forest Virus to produce IL-12. Neurosurg 29. Chen QR, Kumar D, Stass SA, Mixson AJ. Liposomes Focus. 2000;9:Article 7. complexed to plasmids encoding angiostatin and endostatin 11. Yamanaka R, Zullo SA, Tanaka R, Blaese RM, Xanthopoulos inhibit breast cancer in nude mice. Cancer Res. 1999; KG. Enhancement of antitumor immune response in glioma 59:3308–3312. models in mice by genetically modified dendritic cells pulsed 30. Dhanabal M, Volk R, Ramchandran R, Simons M, with Semliki Forest virus-mediated by complementary DNA. J Sukhatme VP. Cloning, expression, and in vitro activity of Neurosurg. 2001;94:474–481. human endostatin. Biochem Biophys Res Commun. 1999; 12. Onodera M, Yachie A, Nelson DM, Welchlin H, Morgan RA, 258:345–352. Blaese RM. A simple and reliable method for screening 31. Sauter BV, Martinet O, Zhang WJ, Mandeli J, Woo SLC. retroviral producer clones without selectable markers. Hum Adenovirus-mediated gene transfer of endostatin in vivo Gene Ther. 1997;8:1189–1194. results in high level of transgene expression and inhibition of 13. Blezinger P, Wang J, Gondo M, et al. Systemic inhibition of tumor growth and metastases. Proc Natl Acad Sci USA. tumor growth and tumor metastases by intramuscular 2000;97:4802–4807. administration of the endostatin gene. Nat Biotechnol. 1999; 32. Rosenfeld MA, Siegfried W, Yoshimura K, et al. Adenovirus- 17:343–348. mediated transfer of a recombinant alpha 1-antitrypsin gene to 14. Yamanaka R, Tanaka R, Yoshida S, Saitoh T, Fujita K, the lung epithelium in vivo. Science. 1991;252:431–434. Naganuma H. Suppression of TGF-beta1 in human gliomas by 33. Ying H, Zaks TZ, Wang RF, et al. Cancer therapy using a self- retroviral gene transfection enhances susceptibility to LAK replicating RNA vaccine. Nat Med. 1999;5:823–827. cells. J Neurooncol. 1999;43:27–34. 34. Levis R, Schlesinger S, Huang HV. Promoter for Sindbis virus 15. Warren BA, Chauvin WJ, Philips J. In: Myers WPL, Day SB, RNA-dependent subgenomic RNA transcription. JVirol. Stansly P, Garattini S, Lewis MG (Eds), Progress in Cancer 1990;64:1726–1733.

Cancer Gene Therapy, Vol 8, No 10, 2001