GCV-Dependent Cytotoxic Self-Elimination of Dendritic Cells

CANCER - IMMUNOTHERAPY I minicircle DNA vectors devoid of bacterial DNA backbone result in persistent and high-level transgene expression in vivo. The higher level of target gene expression may contribute to the better immune response. However, to date, minicircle vectors have never been applied in vaccine. Safety and effective delivery systems are needed to enhance the potency of DNA vaccines. Encapsulating DNA within degradable delivery vehicles provides an effective way to protect the DNA from the surrounding environment prior to delivery. Importantly, the nanoparticles can directly come to liver and spleen by intravenous injection, where the immune cells can be activated. This procedure can avoid from muscle administration with immune clearance of antigen. Also it is absolutely safe without toxicity with intravenous injection compared with regular plasmids. We have developed the recombinant minicircle vectors carrying gag, pol and env genes of HIV-1 respectively. To improve the expression level , we optimized the codon usage of gag gene. Expressions of constructed vectors were confirmed by Western blot in transiently transfected mammalian cells. Preparation and Characterization of the biodegradable nanoparticles self-assembled from PEGylated poly (lactic-co-glycolic acid) (PEGPLGA) for the delivery of minicircle vector are in progress. The in vivo humoral immunity and cellular immunity induced by the nanoparticles will be detected in the near future.

221. Development of Baculovirus as a New Avian Influenza Vaccine Platform

Chi-Yuan Chen,1 Guan-Yu Chen,1 Hsiao-Chiao Shiah,1 Chen-Yu Chung,1 Wen-Hsin Lo,1 Yu-Chen Hu.1 1 Department of Chemical Engineering, National Tsing Hua University, Hsinchu, Taiwan, Province of China. Avian influenza virus (AIV) imposes a tremendous threat to the public health worldwide. We have previously displayed the AIV hemaglutinin protein (HA) on the baculovirus envelope and found that this pseudotyped baculovrus (Bac-HA64) induced considerable HI (HA inhibition) titers in mouse models, suggesting the potential of baculovirus as a vaccine vehicle to prevent avian influenza. Besides Bac-HA64, hereby we constructed two more baculoviruses Bac-CHA and Bac-CHA/HA64. Bac-CHA expressed the HA protein under the CMV promoter after transducing the mammalian cells while BacCHA/HA64 not only expressed HA but also possessed HA displayed on the envelope. We first confirmed the HA expression in mammalian cells after Bac-CHA and Bac-CHA/HA64 transduction by Western blot analysis. Then we compared the vaccine potential of these 3 baculovirus vectors in mouse models via different administration routes at different dosages. A virus dosage of 1010 PFU/mouse elicited significantly higher HA-specific antibody and HI titers than 108 and 109 PFU/mouse. For all 3 virus vectors, intramuscular injection induced higher HA-specific IgG titers and HI titers than subcutaneous and intranasal injections in mice. Among the 3 baculovurius vectors, both Bac-CHA and Bac-CHA/HA64 induced responses that were more biased towards Th1 immunity than Bac-HA64, but Bac-CHA and Bac-CHA/HA64 elicited similar titers of HA-specific IgG and HI. We also found baculovirus vectors with HA displayed on the envelope (Bac-HA64 and Bac-CHA/HA64) induced higher IgA titers than Bac-CHA in the bronchoalveolar lavage fluid after intranasal injection, suggesting the induction of higher mucosal immunity. These data altogether demonstrate the potential of pseudotyped baculovirus as an avian influenza vaccine candidate. Key words: baculovirus, avian influenza, vaccine, pseudotype, gene delivery.

Cancer - Immunotherapy I 222. Genome-Wide Mapping of piggyBac Transposon Integrations in Primary Human T Cells

Daniel L. Galvan,2 Aparna Kaja,2 Claudia Kettlun,2 Yozo Nakazawa,3 Cliona M. Rooney,3 Matthew H. Wilson.1,2,3 1 Michael E. DeBakey VA Medical Center, Houston; 2Department of Medicine, Baylor College of Medicine, Houston; 3Center for Cell and Gene Therapy, Baylor College of Medicine, Houston. The piggyBac transposon system represents a promising non-viral tool for gene delivery, gene discovery, and possibly clinical gene therapy applications. Recently, we have found piggyBac to be highly efficient in transduction (∼40%) of primary human T cells for potential adoptive immunotherapy applications. In this study, we evaluated 220 integration sites for piggyBac in primary human T cells. Human peripheral blood mononuclear cells (PBMC) were nucleofected with transposase and a transposon plasmid containing a bacterial origin of replication and kanamycin resistance. On day 1, PBMCs were stimulated with OKT3 and anti-CD28 antibodies. After 7-10 days of growth and culture, we used a modified plasmid rescue procedure for recovery and mapping of transposon integration sites in the human T cell genome. Our results revealed distinct differences for piggyBac integration in human T cells when compared to other derived human cell lines. PiggyBac continued to show a non-random integration profile and preferences for transcriptional units, transcriptional start sites, and long terminal repeats which differentiate it from other transposon systems such as sleeping beauty. Using public microarray database information, we evaluated the expression level of mapped genes into which piggyBac integrated. Random integrations generated in silico were compared to true integrations mapped in human T cells and two other human cell lines. PiggyBac appeared to integrate into more active genes in primary human T cells compared to random integrations and other human cell types, possibly due to concomitant T cell activation with OKT3 and anti-CD28 antibodies during transposition and transduction. It is known that piggyBac requires a TTAA nucleotide element for genomic integration. We analyzed the TTAA content and AT content contained in a 500bp window surrounding piggyBac integrations in human T cells and randomly generated integrations. Our results revealed piggyBac to apparently prefer TTAA but not necessarily AT rich regions of the human genome. Mapping of piggyBac transposon integrations in the human genome gives us insight into the biology of piggyBac, how our repetitive element rich genomes have changed over time, and how one might better consider piggyBac for gene transfer or genetic modification studies. The results of this study have important implications for consideration of piggyBac as a non-viral gene delivery system for human cell therapies and other applications.

223. SMART-DCs: Non-Integrating Multicistronic Lentiviral Vectors Result in Self-Differentiation and HSV-TK/GCV-Dependent Cytotoxic Self-Elimination of Dendritic Cells

Renata Stripecke,1 Mudita Pincha,1 Bala Sai Sundarasetty,1 Gustavo Salguero,1 Adan Jirmo,1 Axel Schambach,2 Christopher Baum,2 Arnold Ganser.1 1 Hematology, Hannover Medical School, Hannover, Germany; 2 Exerimental Hematology, Hannover Medical School, Hannover, Germany.

Background: Dendritic cells (DCs) are potent immune adjuvants capable of boosting antigenic responses against tumors and their use in tumor immunology has been extensive. Nevertheless, largescale ex vivo DC production in the laboratory remains highly costly and impractical. A goal of our laboratory is to create more effective and less costly DC vaccines by gene reprogramming. We have Molecular Therapy Volume 17, Supplement 1, May 2009 Copyright © The American Society of Gene Therapy

S87

CANCER - IMMUNOTHERAPY I demonstrated in human and mouse systems that ex vivo manipulation of DC precursors with lentiviral vectors (LV) for constitutive production of Granulocyte Macrophage Colony Stimulating Factor (GM-CSF), Interleukin 4 (IL-4) and tumor antigens was capable of inducing self-differentiation of potent therapeutic DC vaccines against melanoma (Koya et al. 2007). These cells were called “SMART”-DCs (Self-differentiated Myeloid-derived Antigenpresenting-cells Reactive against Tumors). Results: We produced LV modalities with enhanced biosafety features: integrating or nonintegrating lentiviral vectors (NILV) containing the ubiquitous CMV promoter or the antigen-presenting cell restricted MHCII promoter. Interspacing 2A elements allowed efficient co-expression of three functional proteins: HSV-TK, GM-CSF and IL-4. Mouse nonadherent bone marrow cells were transduced once with high titer LVs (2 mcg equivalent p24/ml) and maintained in culture in the absence of exogenously added cytokines. Kinetics of DC differentiation was followed by expression of immunophenotypic markers (CD11c, CD11b, MHCII, MHCI) demonstrating that all vector modalities (LVCMV, LV-MHCII, NILV-CMV, NILV-MHCII) yielded SMART-DCs. Expression of GM-CSF and IL-4 transgenes persisted stably for 2-3 weeks in all groups, correlating with results using GFP as marking gene in DCs. Analyses of the CD11c/MHCII DC differentiation markers showed approximately 43-62% DC purity at day 7 posttransduction, increasing subsequently to 72-94% with time (days 14, 21). SMART-DCs were treated with ganciclovir (GCV) on days 7, 14 and 21. Cells were analyzed by FACS for the expression of DC surface marker CD11c and propidium iodide staining was used for determine cell viability. All groups treated with GCV showed TKdependent toxicity (40-80%). Vaccination/ boost of C57BL/6 mice was performed with 10e5 freshly transduced SMART-DCs (LV-CMV transduced) co-expressing the TRP2 melanoma antigen. One week after the boost, splenocytes were harvested, re-primed with a TRP2 epitope or a control OVA peptide, and stained for CD3, CD8 and IFN-gamma. We observed high frequencies (8-10%) of TRP2-specific CD8+ effector cells. Control groups (SMART-DC vaccines not coexpressing TRP2 or splenocytes reprimed with OVA) showed no reactivity. The data demonstrates the feasibility of programming self-differentiation and self-elimination of DCs effectivelly with safety enhanced lentiviral vectors. We are currently performing vaccinations with all groups of SMART-DCs to evaluate the added-on TK-dependent cytotoxic effects.

224. HER2-Specific T Cells Recognize CD133+ Glioma Derived Stem Cells and Induce Regression of Experimsnetal Glioma in Animal Models

Nabil Ahmed,1 Vita S. Salsman,1 Yvonne Kew,2 Susanne Powell,2 Eastwood Leung,1 Winfried S. Wels,3 Robert Grossman,2 Helen E. Heslop,1 Stephen Gottschalk.1 1 Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX; 2Department of Neurosurgery, The Methodist Hospital, Houston, TX; 3Chemotherapeutisches Forschungsinstitut Georg-Speyer-Haus, Frankfurt, Germany. BACKGROUND: Glioblastoma multiforme (GBM) is a common aggressive adult brain tumor and the majority of patients die within 2 years of diagnosis. Recent studies have shown that the CD133+ stem cell populations in GBMs are relatively radioresistant, most likely explaining the poor outcome of patients undergoing standard radiation therapy. Thus, therapies that target the CD133+ glioma cell population might have the potential to improve outcomes for GBM patients. The human epidermal growth factor receptor 2 (HER2) is expressed in up to 80% of GBMs at low levels rendering monoclonal antibodies ineffective. Here we show that CD133+ glioma derived stem cells express sufficient levels of HER2 to be recognized and killed by T cells expressing HER2-specific chimeric antigen receptors (CAR; HER2-specific T cells) ex vivo. In addition, S88

adoptive transfer of HER2-specific T cells had potent anti-tumor activity in animal models. MATERIAL AND METHODS: HER2specific T cells from GBM patients were generated by transduction with a retrovirus encoding a HER2-specific CAR with a CD28.zeta signaling domain. CD133+ glioma derived stem cells were isolated from primary patients’ GBM cell lines using high-speed cell sorting. Primary GBM cells or CD133+ and CD133- subpopulation were used to test the function of the patients’ HER2-specific T cells. The ex vivo efficacy was determined by the ability of HER2-specific T cells 1) to kill HER2-positive target cells in cytoxicity assays and 2) to proliferate and secrete cytokines (IFN-γ and IL-2) in response to stimulation with HER2-positive GBM cells. The in vivo efficacy of patients’ HER2-specific T cells was tested for the ability to induce regression of autologous GBMs in an orthotopic murine xenograft model. RESULTS: The CD133+ GBM stem cell population expressed the same level of HER2 as the CD133- cell population. Patients’ HER2-specific T cells killed CD133+ and CD133- primary GBM cells in cytotoxicity assays, whereas HER2-negative tumor cells were not killed. Stimulation of HER2-specific T cells with HER2-positive primary GBM cells resulted in T-cell proliferation and secretion of IFN-γ and IL-2 in a HER2-dependent manner. Intra-tumoral injection of HER2-specific T cells induced sustained remission of established autologous GBMs in an orthotopic murine xenograft model. In contrast, delivery of non-transduced T cells did not change the tumor growth pattern. CONCLUSION: We demonstrate that T-cells expressing HER2-specific CARs can recognize and kill HER2+/CD133+ glioma derived stem cells. Their activation results in proliferation and secretion of cytokines that are important for T-cell expansion and persistence. HER2-specific T cells can effectively eradicate established GBM in an orthotopic xenograft murine model. These results indicate that HER2-specific T cells could represent a promising immunotherapeutic approach for GBM.

225. Updated Results of a Phase II Clinical Trial with a Second Generation, Gene-Modified, Immune-Enhanced, Oncolytic Herpes Virus in Unresectable Metastatic Melanoma

N. N. Senzer,1,2,3 H. Kaufman,4 T. Amatruda,5 M. Nemunaitis,1 T. Reid,6 J. Glaspy,7 H. Goldsweig,8 R. S. Coffin,9 J. Nemunaitis,1,2,3 The OncoVEX Melanoma Phase II Investigator Group.8 1 Mary Crowley Cancer Research Centers, Dallas, TX; 2Texas Oncology, P.A., Dallas, TX; 3Baylor Sammons Cancer Center, Dallas, TX; 4Columbia University, New York, NY; 5Hubert H. Humphrey Cancer Center, Minneapolis, MN; 6University of California, San Diego, San Diego, CA; 7University of California, Los Angeles, Los Angeles, CA; 8BioVex Inc., Woburn, MA.

Background: OncoVEXGM-CSF is an oncolytic HSV encoding GMCSF which replaces the deleted ICP34.5; ICP47 is likewise deleted with consequent increase in US11. We recently completed a phase II trial involving 50 advanced melanoma patients (Stages IIIc and IV) with at least one injection accessible lesion, including by ultrasound. Methods: Patients received a single IT injection of 106 pfu/ml apportioned between ≤ 10 injectable tumors, followed 3 wks later by ≤ 24 sequential injections of 108 pfu/ml every 2 wks until clinically significant disease progression or overall or injectable lesion complete response. Response (RECIST modified to allow progression before response and biopsy of residual masses) and survival were monitored. Results: All 50 pts have been enrolled and are evaluable (Stage IIIc, n=10; IV M1a, n=16; IV M1b, n=4; IV M1c, n=20). A median of 6 injections were administered. Adverse effects were limited and generally involved transient flu-like symptoms. Both injected and uninjected regional and distant disease demonstrated response including clearly documented responses at uninjected visceral sites. The overall response rate was 26% (8 CR, 5 PR); 10 responses have been maintained for >6 months and 2 are ongoing at <6months, the Molecular Therapy Volume 17, Supplement 1, May 2009 Copyright © The American Society of Gene Therapy