1174 Comparison of E1a Modifications To Achieve Tumor Cell Selective Viral Replication and Toxicity Molecular Therapy �������� ��� ���� ���������������� �������� ��� ������®������������ �!����� ����"�[.]
CANCER TARGETED GENE THERAPY I 1172 Minimally Invasive In Vivo Imaging of an Oncolytic Herpes Simplex Virus Mutant Expressing Green Fluorescent Protein 1173 Selection and Application of Chronic Lymphocytic Leukemia Binding Ligands from Peptide-Presenting Phage Libraries Brendon M Stiles,1 Prasad Adusumilli,1 Amit Bhargava,1 Stanziale F Stephen,1 Richard J Wong,1 Valerie W Rusch,1 Fong Yuman.1 Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, United States Satoshi Takahashi,1 Hoyin Mok,2 Frank C Marini III,3 Michael Andreef,3 Malcolm K Brenner,1 Michael A Barry.1,2 Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX; 2Department of Bioengineering, Rice University, Houston, TX; 3Department of Molecular Hematology and Therapy, M.D Anderson Cancer Center, Houston, TX Introduction: Herpes simplex virus-one (HSV-1) oncolytic therapy and viral-based gene therapy are promising treatment modalities against cancer A replication-competent HSV-1 mutant, NV1066, carries a transgene for enhanced green fluorescent protein (EGFP), which is commonly used in vitro as a reporter/marker gene While techniques for assessing EGFP expression in cell culture or animal models are readily available, methods for detecting regional distribution and activity in man are needed The purpose of this study was to determine whether: 1) NV1066 is cytotoxic to a lung cancer cell line in vitro and in vivo, 2) EGFP is a detectable marker of viral infection, and 3) EGFP expression and viral distribution could be localized in an animal model of pleural cancer using a minimally invasive system Methods: A549 human lung cancer cells were infected in vitro at multiplicities of infection (MOI: number of viral particles per tumor cell) of 0.1 and 1.0 Viral replication was determined by plaque assay, cell kill by LDH release assay, and EGFP expression by flow cytometry To mimic pleural spread of lung cancer, x 106 A549 cells were injected percutaneously into the pleural cavity of athymic mice Mice were treated day following tumor implantation with intrapleural injection of either NV1066 (1 x 107 plaque forming units) or PBS, then sacrificed at weeks to determine treatment effect For EGFP localization studies, mice were treated with NV1066 day 14 after tumor implantation Mice were examined 48 hours later for EGFP expression in tumor deposits, using an Olympus thoracoscopic/laparoscopic system equipped with the appropriate fluorescent filters Results: NV1066 progressively replicated in and killed tumor cells Peak viral titers demonstrated a 54-fold increase over the initial infecting dose By day 9, cell kill at MOIs of 0.1 and 1.0 was 83% and 96% respectively (p < 01, t-test) EGFP expression in infected cells was initially dose-dependent, but increased over time with viral replication By day 5, over 95% of all remaining live cells expressed EGFP at both MOIs (p < 01, t-test) Treatment with NV1066 decreased pleural disease burden after four weeks Average chest wall nodule count was 71 in the PBS group compared to in the NV1066 group (p < 01, t-test) Following intrapleural administration of NV1066, EGFP was easily visualized in tumor deposits in the pleural cavity using a thoracoscope with a fluorescent filter system Expression could be used to identify tumor deposits as small as mm in diameter EGFP expression and viral localization to tumor were confirmed by histopathology Conclusions: NV1066 has significant oncolytic activity against A549 lung cancer cells in vivo Expression of EGFP in infected cells can be used to localize the virus and helps to identify tumor foci By incorporating fluorescent filters into thoracoscopes, laparoscopes, or other endoscopic systems, a minimally invasive means of detection and localization of viral therapy may be developed, using the EGFP marker gene The technique reported here can be applied clinically in many oncolytic viral or gene therapy trials to monitor viral distribution and activity in vivo Molecular Therapy Vol 7, No 5, May 2003, Part of Parts Copyright ® The American Society of Gene Therapy Previous work with random phage libraries demonstrated that cell binding peptide ligands can be selected from phage libraries by binding directly to cells in culture (Barry et al Nat Med 2(3):299305 1996) or by binding to vascular cells in vivo (Pasqualini et al Nature 380: 364-366) One goal is to identify ligands that bind directly to human cancer cells to use these with gene therapy vectors to mediate gain of function transduction in vitro or for the direct targeting in vivo To test this application, we have selected peptidepresenting phage libraries against human prostate and breast cancer cell lines and against primary patient chronic lymphocytic leukemia (CLL) cells In this work, libraries with large peptides have always out-competed libraries with smaller peptides and the build up of consensus peptide motifs is unusual To provide cell-binding ligands for ex vivo gene therapy and CLL-targeting ligands for in vivo drug and gene therapy, we have selected 44 20-mer peptides from peptidepresenting phage libraries by panning against patient primary CLL cancer cells 29 of the selected peptides were assayed for cell binding Fourteen of the selected peptides bound CLL, B, T, and monocyte cells, six bound only CLL and B cells, and one peptide bound only B cells However, eight of the selected peptides were CLL specific The degree if specificity of the peptides was highly dependent on the target cell used for selection, with some patient’s cells producing highly specific peptides while other’s generating largely promiscuous peptides When peptides were tested out of the context of phage, synthetic peptide 1-5 was able to functionally re-target adenoviral vectors for increased ex vivo gene delivery to primary CLL cells These data demonstrate the ability to identify cancer-targeting peptides by selection of phage libraries against primary human cancers This work also emphasizes the importance of patient to patient heterogeneity in cancer cells for the identification of targeting peptides and their functional application for gene therapy vector targeting 1174 Comparison of E1a Modifications To Achieve Tumor Cell Selective Viral Replication and Toxicity Harald Sauthoff,1 Teona Pipiya,1 Sheila Heitner,1 William Chang,1 Shu Chen,1 William N Rom,1 John G Hay.1 Medicine, NYU School of Medicine, New York, NY, United States Replicating adenoviral vectors have the capability of multiplying up to a thousand fold in the target cell and may have a tremendous potential for cancer therapy However, restricting viral replication to cancer cells is of great importance to limit toxicity The adenoviral E1a protein binds to and modifies Rb and p300 functions, forcing resting cells into S-phase and creating an environment suitable for viral replication It has been proposed that modifications of the E1a protein that impair Rb- or/and p300-binding prevent S-phase induction in normal cells resulting in selective viral replication in tumor cells However, it remains unclear which of several possible E1a modifications would confer the most protection to normal cells without compromising the oncolytic effect of the vector The purpose of this work was to determine the best E1a modification to target a replicating adenovirus to cancer cells The evaluated vectors contained the following E1a modifications Deletion of the E1a domain that binds to Rb (dl121-128) Deletion of the Rb binding domain plus S453 CANCER TARGETED GENE THERAPY I deletion of the N-terminal region that binds to p300 (dl4–25/121128) Deletion of the E1a domain that binds to the transcriptional adapter motif (TRAM) in p300 (dl65-68) Deletion of the Rb binding domain plus deletion of the domain that binds to TRAM in p300 (dl65-68/121-128) Deletion of the E1a core binding domain to Rb plus deletion of an adjacent region which functions independently of Rb binding in the prevention of cell-cycle arrest (dl121-138) Results: Deletion of the domain that binds to Rb plus deletion of the region that binds to p300 (dl4–25/121-128) were required to effectively block S-phase induction in growth arrested normal cells In three growth arrested normal cell types, a small decrease in new virus production was seen for all E1a-modified viruses (except dl121138) without correlation to S-phase induction All viruses were equally cytotoxic to cancer cells, as compared to wild-type virus Only dl65-68 and dl4–25/121-128 showed evidence of decreased toxicity in at least of normal cell types Conclusion: Deletion of only the Rb-binding domain of E1a in a replication-competent adenovirus does not effectively block S-phase induction or reduce toxicity to normal cells Deletion of the Rb binding domain plus deletion of the region that binds to p300 (dl4– 25/121-128) or deletion of the E1a domain that binds to the transcriptional adapter motif (TRAM) in p300 (dl65-68) is associated with reduced toxicity to some normal cell types, probably by a mechanism that is independent of the ability of the virus to induce S-phase 1175 Oncolytic Herpes Simplex Virus-1 G207 Induces Cyclooxygenase-2 Gene Expression by Activation of the Mitogen Activated Protein Kinase Pathway Teresa H Kim,1 Amit Bhargava,1 Brendon M Stiles,1 Richard J Wong,1 Yuman Fong, Andrew J Dannenberg.2 Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY; 2Department of Medicine, New York Presbyterian Hospital-Cornell, New York, NY Introduction: Oncolytic viral therapy stimulates various mediators of inflammation Increased cyclooxygenase-2 (COX-2) levels are found in inflammatory as well as neoplastic states Recent investigations have found enhanced antitumor effects combining COX-2 inhibitor with various other cancer treatment modalities The purpose of this study was to investigate the effects of G207, an oncolytic herpes simplex virus type-1 (HSV-1) on COX-2 levels and to investigate the effects of combination therapy Methods: RAW 264.7, mouse macrophage cells, were treated with G207 (multiplicity of infection 1.0 ) and harvested at 6, 12, and 24 hours post infection Total cellular protein and RNA were isolated for Western and Northern blotting respectively PGE2 levels were measured using enzyme immunoassay PD98059, extracellularregulated protein kinase and (ERK1/2) inhibitor and SB202190, p38 mitogen-activated protein kinase (MAPK) inhibitor, were added to G207 treated RAW cells and cellular protein was harvested for COX-2 Western blotting analysis 24 hours post infection HTB147, human pancreatic adenocarcinoma cells, were implanted in the flanks of nude mice and treated with intratumoral G207 injection (5 x 106 plaque forming units) alone or in combination with celecoxib feed (150 parts per million powder chow) Results: There was an increase in COX-2 protein levels with G207 treatment seen by Western blot analysis (Figure 1) Measurement of PGE2 levels showed a 10-fold increase in production by cells treated with G207 Northern blot showed an increase in COX-2 mRNA levels in G207 treated RAW cells PD98059 and SB202190 both blocked the G207 induced increase in COX-2 protein level (Figure 2) Pancreatic flank tumor model showed smaller average S454 tumor volume in the combination group as compared to the each of the single treatment groups 28 days post treatment (combination vs G207, p