Scholarly article on topic 'ABSTRACTS OF ORAL PRESENTATIONS'

ABSTRACTS OF ORAL PRESENTATIONS Academic research paper on "Biological sciences"

Share paper
Academic journal
Cancer Gene Therapy
OECD Field of science

Academic research paper on topic "ABSTRACTS OF ORAL PRESENTATIONS"


001 Second Pasman Keynote Lecture: Gene and oncolytic virus therapies

J. Zwiebel

Cancer Therapy Evaluation Program, National Cancer Institute, National Institutes of Health, Bethesda, Maryland.

Cancer accounts by far for most of the gene therapy trials that are being carried out both in the United States and worldwide. Seventy percent of the gene therapy protocols that have been reviewed by the National Institutes of Health Recombinant DNA Advisory Committee are for the treatment of cancer. Of these cancer gene therapy protocols, two-thirds are for cancer immunotherapy, involving transfer of cytokines, immune accessory molecules, or tumor antigens into a variety of cellular targets. The majority of the remaining protocols involve the transfer of chemoprotection (e.g., the multidrug resistance gene), prodrug activation (e.g., herpes thymidine kinase, which activates ganciclovir), or tumor suppressor genes (e.g., p53). Local or distal bystander effects of transduced cells may mediate antitumoral effects. These bystander mechanisms may help to overcome poor transduction efficiencies by currently available vectors. Replicating oncolytic viruses that have entered or are soon to enter the clinic include herpes virus, Newcastle disease virus, and reovirus. All of these viruses have been shown to replicate selectively in cancer cells, albeit by different mechanisms. Reovirus, for example, requires the presence of an activated Ras signaling pathway to replicate and destroy cells. In addition, one can confer selectivity by placing an essential viral gene under the control of a tumor-specific promoter. Because all of these viruses can replicate conditionally and spread in tumors, they may serve as more effective gene vectors for cancer gene therapy. For example, inserting a cytokine gene into these viruses may enhance antitumoral immunity. Clearly, much work needs to be done both in the laboratory and in the clinic to exploit the full potential of these novel viral therapies.

002 Tumor vaccination using genetic modification to enhance interactions of dendritic cells with the host immune system

Ronald G. Crystal

Institute of Genetic Medicine, Weill Medical College of Cornell University, New York, New York.

Dendritic cells (DC) are powerful antigen (Ag)-presenting cells (APCs) that function as the principal activators of T cells. Using adenovirus gene transfer vectors, we have assessed strategies that capitalize on genetic modification to enhance the interaction of DC with tumor cells in vivo, thus intensifying the process of presentation of tumor Ags to the host immune system. Three approaches have been used. First, based on data that the human CC chemokine, macrophage inflammatory protein-3 is chemotactic for DC in vitro, we assessed the hypothesis that adenovirus-mediated gene transfer of macrophage inflammatory protein-3 to tumors might induce a local accumulation of DC and inhibit the growth of pre-existing tumors. This strategy may be a useful approach to enlisting the help of DC in boosting antitumoral immunity in treating both local and metastatic disease, without the necessity of ex vivo isolation and manipulation of DC. Second, CD40 ligand (CD40L), a CD4+ T-cell ligand for CD40 on APC, is essential for the initiation of antigen-specific T-cell responses, an important component of the immune response to tumors. We have examined the hypothesis that in vivo genetic modification of tumor cells to express CD40L will trigger CD40 on local APCs to present tumor Ag to the cellular immune systems, thus eliciting antitumoral immunity to suppress growth of the tumor. The results support the concept that transduction of tumors with a recombinant CD40L adenovirus vector initiates antitumoral immunity sufficient to suppress the growth of pre-existing tumors. Finally, rather than genetically modifying the tumor cells with CD40L, we have modified the DC ex vivo to express the CD40L CD4+ T-cell gene and then transferred the genetically modified DC directly to the tumor. The results show a striking CD4+

T-cell-independent suppression of the growth of pre-existing primary and metastatic tumors. Taken together, these murine experimental models support the concept that gene transfer can be used to enhance the role of DC in strategies that should be applicable to the treatment of human malignancies.

003 Plasmid DNA vaccines for cancer

Albert F. LoBuglio

University of Alabama, Birmingham, Alabama.

The past decade has seen a substantial research effort in the use of plasmid DNA (polynucleotide) vaccines for the induction of immune responses to protein antigens (Ags) relevant to infectious disease and cancer. These studies have been directed at the prevention of infectious diseases, as well as therapeutic efforts directed at existing infections (HIV) or cancer. Most of the work regarding cancer therapy has used animal tumor models modified for the expression of human tumor Ags so that they can provide data on the preclinical efficacy of a vaccine to be used subsequently in human trials. Our own studies have used a syngeneic murine colon carcinoma model stably trans-fected for human carcinoembryonic Ag (CEA). We have developed a variety of plasmid constructs expressing the entire CEA protein or components of the protein. We have shown in animal models that significant B- and T-cell immune responses occur generally over a 4-to 8-week timeframe and include the capacity to reject CEA expressing tumor challenge. The immune response varies with the route of administration and can be enhanced by the incorporation of immune accessory factors and/or cytokines. Studies in primates with this plasmid DNA vector demonstrated a slower onset of immune response and a differential effect of intramuscular versus intradural routes of administration. An initial phase I trial of a plasmid DNA vaccine expressing CEA was carried out in patients with metastatic colorectal cancer, and a current trial is ongoing with a plasmid DNA vector expressing MART-1 in patients with malignant melanoma. The design of the vectors and a current evaluation of the trials will be presented. This strategy has many advantages regarding costs and standardization of the DNA vaccine product but will clearly need continued development and amplification of effects to have antitumoral efficacy.

004 Cancer immunotherapy with tumor RNA-trans-fected dendritic cells

E. Gilboa, Smita K. Nair, and David Boczkowski Department of Surgery, The Center for Cellular and Genetic Therapies, Duke University Medical Center, Durham, North Carolina. Vaccination with the mRNA content of tumor cells would extend the scope of vaccination to the majority of cancer patients who lack an identified tumor antigen and cannot provide sufficient tumor tissue for antigen preparation because RNA can be amplified from very few cancer cells. Extensive preclinical studies carried out in our laboratory have shown that RNA-transfected dendritic cells (DC) are effective inducers of cytotoxic T-lymphocyte (CTL) and tumor immunity, as effective if not more so as other strategies under consideration. Importantly, DC pulsed with whole cell tumor RNA from cancer patients were capable of stimulating polyclonal CTL responses in vitro that recognized and lysed the patient's own tumor cells. In recent studies, RNA was successfully amplified from tumor cells without loss of function. In animal studies, RNA amplified from murine tumor cells was capable of inhibiting or preventing the growth of metastases in a postsurgical metastasis model. Furthermore, RNA was successfully isolated and amplified from human colon and prostate cancer cells obtained by microdissection from a pathology slide, and DC pulsed with the amplified RNA were able to stimulate a potent CTL response in vitro. Clinical testing of vaccination with RNA-transfected DC is ongoing in patients with colon cancer, prostate cancer, and breast cancer. Similar programs focusing on kidney cancer, multiple myeloma, glioma, pancreatic cancer, and stomach cancer have been established recently.

Cancer Gene Therapy, Vol 7, No 8, 2000: pp 1203-1210

005 Antigen-presenting vesicles as novel vectors for immunotherapy of cancer

Laurence Zitvogel, Eric Angevin, Joseph Wülfers, Fabrice André, Carole Masurier, Anne Lozier, Nadine Fernandez, Clotilde Thery, Armelle Regnault, Graca Raposo, and Sebastian Amigorena Institut Gustave Roussy, Villejuif, France; and Institut National de la Santé et de la Recherche Médicale and Institut Curie, Paris, France. Dendritic cells (DC) proved to be efficient cellular adjuvants to prime naive T cells and mediate antitumoral effects in cancer patients in the pioneering clinical trials. Ideal vaccination processes should be cell free, use immunorelevant tumor antigens (Ags) regardless of patient human histocompatibility leukocyte Ag haplotype, and allow manufacturing on a large scale. We found that DC spontaneously release 50-90 nm of Ag-presenting and -containing vesicles of endosomal origin, namely "exosomes" (DEX), that elicit potent T-cell-mediated antitumoral immune responses in tumor bearers.1 The protein pattern of DC-derived exosomes differs from that of cytosolic or plasma membrane components with overexpression of hsc73, molecules involved in Ag presentation, or in targeting to T cells or Ag-presenting cells (APCs).2 These data support the implementation of DEX for cancer immunotherapy as a novel and safe cell-free vaccine. One of the other problems that needs to be solved is the efficient loading of DC with immunorelevant tumor Ags. In parallel, we found an alternative option of efficient DC loading, taking advantage of our knowledge of late endocytic compartment-derived exosomes. Indeed, not only APCs but also tumor cells secrete such exosomes that engulf, during their formation, cytosolic material and heat shock proteins (Wolfers et al, submitted for publication). Indeed, tumor-derived exo-somes (TEX) bear functional peptide/major histocompatibility complex class I complexes on their membrane but also contain heat shock proteins and cytosolic tumor Ags (i.e., p53, gfp, MART1). TEX pulsed onto DC induce potent CD8+ T-cell-restricted, tumor Ag-specific cross-presentation associated with tumor regression in mice. TEX could be a valuable and suitable option for DC loading in innovative cancer vaccines and could account for the in vivo cross-talk between tumors and APCs.


1. Zitvogel L, Regnault A, Lozier A, et al. Eradication of established murine tumors using a novel cell-free vaccine: dendritic cell-derived exosomes. Nat Med. 1998;4:594-600.

2. Theéry C, Regnault A, Garin J, et al. Molecular characterization of dendritic cell-derived exosomes: Selective accumulation of the heat shock protein hsc73. J Cell Biol. 1999;147:599-610.

006 Immunological, pathological, and long-term clinical data of vaccination with autologous granulocyte-macrophage colony-stimulating factor-transduced tumor cells in metastatic melanoma

G. C. de Gast, M. P. W. Gallee, H. Spits, W. R. Gerritsen, D. Batchelor, S. Clift, W. J. Nooijen, B. B. R. Kroon, A. M. Berns, and E. M. Rankin

Netherlands Cancer Institute, Amsterdam, The Netherlands. Immunological, pathological, and long-term clinical data of vaccination with autologous granulocyte-macrophage colony-stimulating factor (GM-CSF)-transduced tumor cells was studied in 64 patients with progressive metastatic melanoma. Transduction of tumor cells failed (no growth, insufficient tumor cells) in eight patients. Median time to start of vaccination was 66 days. A total of 28 patients did not receive three vaccinations (17 due to PD before, 10 due to PD after one or two vaccinations, 1 refusal), of whom 19 had central nervous system metastases. Patients treated with three vaccinations (n = 28-44%) received either 5 X 106 (arm I) or 5 X 107 (arm II) transduced cells intramuscularly and subcutaneously. DTH reactions became positive in all patients. Tumor-specific cytotoxic T lymphocytes were detected in 5 of 12 patients studied. Cytotoxic T-lymphocyte activity was blocked by anti-human histocompatibility leukocyte antigen class I and CD8 antibodies. Vaccination sites showed an infiltration of T cells (CD4 55> CD8), macrophages, dendritic cells, and eosinophils and no natural killer cells. Intradermal injection sites of DTH and GM-CSF alone showed dendritic cells, macrophages, and CD4+ T cells. Metastatic

sites removed after vaccination showed the presence of T cells (CD4 > CD8), macrophages, and plasma cells. All patients showed short-term eosinophilia in the blood, increasing after each vaccination. No tumor responses (CR/PR) were seen, and 9 patients had NED, 2 patients had SD, and 17 patients had PD. After a follow-up of 3-5 years, six patients are still alive: two with relapses responding to chemoimmunotherapy, three with NED since vaccination, and one with NED after resection of a relapse. There was no difference in survival between arm I and arm II. We conclude that vaccination with autologous GM-CSF-transduced tumor cells is not a good option for stage IV progressive metastatic melanoma because of the time delay, despite the induction of antitumoral immunity. GM-CSF production by transduced tumor cells causing eosinophilia was of very short duration. Sensitized CD4+ T cells probably contribute to the eosinophilia at the second and third vaccination. Long-term survival was only inducted in selected patients with low tumor load in whom all tumor sites could be removed. It still may be effective for stage III disease.

007 Ectopic expression of telomere reverse leads to immortalization of human CD8+ T-cell clones

Erik Hooijberg,1,2 J. J. Ruizendaal,1,2 P. J. F. Snijders,2 E. Kueter,1 J. J. M. Walboomers,2 and H. Spits1

1Department of Immunology, The Netherlands Cancer Institute, Amsterdam, The Netherlands; and 2Department of Pathology (PA 312), AZVU, Amsterdam, The Netherlands.

Replicative senescence of T cells is correlated with erosion of telomere ends. Telomerase plays a key role in maintaining telomere length. It is therefore thought that telomerase regulates the life span of T cells. To test this hypothesis, we have overexpressed the active component of telomerase, telomerase reverse transcriptase (TERT) in human CD8+ T cells and have analyzed the effect on the life span of these cells. A recombinant retrovirus was made that harbors hTERT-IRES-GFP; this retrovirus was used to transduce a Mart-1/HLA-A2-specific T cell with an in vitro history of 12-16 population doublings (PD), which was obtained from a melanoma patient. The transduction efficiency was low (6%), but the proportion of hTERT-IRES-GFP+ cells increased strongly on further expansion of the cells by weekly stimulations in the presence of interleukin-2. After 7 weeks, >95% of the cells were hTERT-IRES-GFP+, indicating a strong growth advantage of the transduced cells. Wild-type or control GFP-only transduced T cells could not be maintained beyond 60 days of in vitro culture. In contrast to untransduced T cells, the hTERT-IRES-GFP+ cells could be subcloned at least three times. After >80 PD, the sub-sub-clones maintain GFP expression and are still growing well, indicating the stability of the introduced gene. The phenotype of the TERT-trans-duced cells was virtually indistinguishable from that of the untrans-duced cells. Also the specificity and the cytotoxic activity of the cells as determined by MART1/HLA-A2 tetramer staining and cytotoxic assays were not altered by introduction of TERT. Most importantly, the cells remained dependent upon antigen stimulation and cytokines for their expansion and survival, indicating that the cells retained their normal physiology and were not transformed. The hTERT mRNA levels and telomere activity were both higher in hTERT-IRES-GFP+ cells compared with wild-type cells. Moreover, telomere lengths were stabilized at 7 kb in the sub-sub-clones, differing by >60-70 PD from the wild-type cells analyzed. These observations prove that TERT directly regulates the replicative life span of human CD8+ T cells. In addition, the fact that tumor-specific T cells can be immortalized without changes in specificity and physiology has implications for the use of these cells in adoptive transfer protocols to treat cancer patients.

008 Regulatory affairs and safety aspects of gene therapy

A. Meager

National Institute of Biological Standards and Control, Hertfordshire, United Kingdom. No abstract available

009 CAR-independent gene transfer to accomplish efficient and specific genetic modification of target cells via adenoviral vectors

David T. Curiel

Gene Therapy Center and Division of Human Gene Therapy, University of Alabama, Birmingham, Alabama.

Adenoviruses have been widely employed for a variety of gene therapy applications, owing largely to their unparalleled efficiency in accomplishing in vivo gene transfer. Despite this unique capacity, full utility of these vectors for gene therapy applications has been limited by their reliance on target cell entry via the native adenovirus receptor, CAR. On this basis, target cells that express low amounts of CAR are relatively resistant to adenoviral vectors (Ad). Alternatively, recognition of CAR may allow gene transfer via Ad to ectopic target cell sites with attendant toxic/morbid consequences. Thus, the ability to direct adenovirus to cell-specific receptors in a CAR-independent manner would potentially allow circumvention of these two key limitations of Ad. To this end, we have endeavored tropism modifications of Ad to allow cell-specific gene delivery. This has been achieved via heterolo-gous retargeting complexes and via genetic modification of the Ad capsid. Both of these strategies have allowed the achievement of CAR-independent gene delivery to target cells. Furthermore, such CAR-independent gene delivery has allowed the achievement of both cell-specific gene delivery as well as gene transfer efficiency augmentations. On this basis, it is clear that strategies to alter Ad tropism may allow greatly improved utilities of Ad for gene therapy applications.

010 Adenovirus vectors for human gene therapy

Peter W. Roelvink, David A. Einfeld, Gai Mi Lee, Rosanna Schroeder, Yuan Li, Imre Kovesdi, C. Richter King, and Thomas J. Wickham GenVec Inc., Gaithersburg, Maryland.

The ability of the adenovirus vector to bind to the native Coxsackie and adenovirus receptor (CAR) has been an obstacle to the development of targetable adenovirus vectors in human gene therapy and cancer treatment. Recent studies have resulted in the identification of a conserved receptor binding site on the fibers of the CAR binding Adenoviridae. We have mutated the AB loop of the Ad5 fiber and introduced a HA tag into the HI loop of the fiber. This has resulted in a vector that no longer binds to CAR and, instead, is directed to a pseudoreceptor in a cell line that was created to allow for the production of mutant Ad vectors. We demonstrate that it is feasible to redirect an Ad vector to a new receptor. Some residual transduction (<2%) of this cell line is observed with the singly ablated vector. We therefore improved this novel vector further by removal of the RGD sequence from the penton base protein. This resulted in a so-called "doubly ablated" vector that has no affinity for the fiber receptor and the a integrins. We will show the first results obtained with these vectors in vitro and in vivo and discuss configurations that are presently being considered for application in cancer gene therapy.

011 Targeting viral membrane glycoproteins for cy-toreductive gene therapy

Stephen J. Russell

Mayo Clinic, Rochester, Minnesota.

An ideal vector for cancer gene therapy should deliver a gene that encodes a cytotoxic protein that has very potent bystander killing capability and should selectively deliver and express this gene in disseminated tumor cells after systemic administration. We have developed convenient strategies for the display of foreign polypeptides on the surface of retroviral, lentiviral, and measles virus vectors as genetically encoded extensions of their envelope glycoproteins. Non-cleavable and protease cleavable linkers were used to tether oligomer-izing polypeptides and high-affinity polypeptide ligands to the surface glycoproteins of these vectors, leading in some cases to potentially useful alterations in their host range properties. In vivo targeting studies are currently underway. For the elimination of tumor cells, we have been exploring the use of genes coding for viral membrane

glycoproteins that cause extensive cell-cell fusion. Our initial studies have focused on the measles virus F and H glycoproteins and a hyperfusogenic mutant of the Gibbon ape leukemia virus envelope that lacks the R-peptide. The genes that code for these proteins give rise to a very potent bystander killing effect because nontransduced tumor cells are recruited into multinucleated syncytia by cells that express the fusogenic proteins, and the syncytia eventually die. Encouraging tumor regressions were observed when lentiviral and measles virus vectors were used for delivery and expression of fusogenic proteins in human tumor xenografts in immunocompromised mice. Also, the specificity of cell-cell fusion could be successfully regulated by displaying polypeptide growth factors and cleavable blocking polypeptides as genetically encoded extensions of the hyperfusogenic GALV envelope glycoprotein.

012 Regulation of therapeutic gene expression with rapa-mycin

Roy Pollock,1 Robbyn Issner,1 Victor Rivera,1 Sridaran Natesan,1 Karen Zoller,1 Phil Zoltick,2 Guang-ping Gao,2 James M. Wilson,2 and Tim Clackson1

1ARIAD Gene Therapeutics Inc., Cambridge, Massachusetts; and 2Uni-versity of Pennsylvania, Philadelphia, Pennsylvania. Safe and effective delivery of therapeutic proteins by gene therapy will require regulatory systems that permit the natural levels and kinetics of protein expression to be reproduced. We have developed a system that allows pharmacological control of protein production at the level of transcription. In this system, expression of a gene is made dependent upon a transcription factor whose activity is regulated by the orally bioavailable drug rapamycin, or nonimmunosuppressive derivatives. Using this system, long-term, tightly regulated expression of growth hormone and Epo has been demonstrated in mice upon intramuscular delivery using AAV vectors. Regulated expression of Epo for >1 year has also been achieved in non-human primates. These studies involved delivery of the regulated target gene and transcription factor cassettes on separate AAV vectors. More recently, we have optimized the system for retroviral delivery and have exploited the increased insert capacity of this vector to deliver both the transcription factors and target gene in a single retrovirus. In vitro studies with model reporter genes indicate that a retrovirus containing all components of the rapamycin-regulated gene expression system is able to induce transcription by at least three orders of magnitude on addition of drug. A complete absence of detectable gene expression in the absence of drug suggests that the retroviral rapamycin system may be particularly suited to the anticancer therapies involving expression of highly cytotoxic proteins. The retrovirally delivered rapamycin system was found to function effectively in a variety of mammalian cell lines including fibroblasts, myoblasts, and pre-B cells. These results indicate that the rapamycin system is adaptable to multiple delivery platforms. In addition, the ability to introduce all components in a single vector may prove important in situations in which efficient delivery remains a challenge.

013 Transductional and transcriptional targeting of tumor endothelial cells for the expression of a novel extracellular prodrug-converting enzyme

S. Brusselbach, W. Graulich, D. M. Nettelbeck, V. Jerome, T. Nahde, K. Muller, D. Weyel, J. Adamkiewicz, A. Fahr, and R. Miiller

Philipps-University, Marburg, Germany.

A major challenge in cancer gene therapy is the development of efficient strategies for the targeting of tumor or tumor endothelial cells (ECs). Targeting the tumor vasculature has several potential advantages compared with direct tumor cell targeting, such as vector accessibility, lack of treatment resistance, and tumor type independence. We have developed a new type of anionic liposomal vector based on the composition of retroviral envelopes. The packaging of plasmid DNA into these lipids in the presence of polyethylenimine yields serum-resistant liposomes of high stability. These liposomes were endowed with a specificity for av^3 integrin-expressing cells (such

as tumor ECs) by attaching a cyclic RGD-containing peptide. These targeted liposomes are nontoxic, transduce cycling ECs in cell culture with an efficiency of ~80% in the presence of sera, and show a high degree of cell-type specificity. To increase this specificity, we have combined transductional targeting with transcriptional targeting. To this end, we have established a number of novel tools: (a) We have found a preferential expression of endoglin (CD 105) in the angiogenic endothelium of human lung tumors. We have cloned the human endoglin promoter and can show a strong promoter activity selectively in proliferating ECs in the context of an adenoviral vector. (b) We have established a generally applicable strategy for enhancing the activity of weak cell-type-specific promoters without loss of specificity, which leads to a > 100-fold enhancement of transcriptional activity as exemplified by the EC-specific vWF promoter. To develop an extracellular cytotoxic effector system (GDEPT), we have constructed secreted and transmembrane forms of ^-glucuronidase. We were able to demonstrate that both forms of the protein in combination with doxorubicin glucuronide efficiently induce tumor cell killing in cell culture and in animal models, which is mediated through a pronounced bystander effect of the generated cytotoxic drug.

014 Polymer-based systems for tumor-targeted gene delivery

Ralf Kircheis,1 Lionel Wightman,1 Birgit Robitza,1 Manfred Ogris,2 Sylvia Brunner,2 and Ernst Wagner1,2

1Boehringer Ingelheim Austria, Vienna, Austria; and 2Institute of Biochemistry, Vienna University Biocenter, Vienna, Austria. DNA complexed with transferrin-polyethylenimine (Tf-PEI) conju-gates,1,2 combining the gene transfer efficiency of PEI with the specific mechanism of receptor-mediated endocytosis, was tested for in vivo application. Systemic application of optimized, surface-shielded Tf-PEI800/DNA complexes into the tail vein of Neuro2A-bearing mice results in preferential gene delivery into distantly growing subcutaneous tumors.3,4 In contrast, application of standard DNA/Tf-PEI800 complexes caused gene transfer primarily to the lung as well as toxicity. The physical and colloidal parameters of the transfection complexes, such as particle size, stability and surface charge, modification with targeting ligand, and the molecular weight of polymeric carrier (PEI800 versus PEI25 versus PEI22), were found to strongly influence in vivo DNA biodistribution, toxicity, and gene transfer efficacy. Two major mechanisms are considered to contribute to the tumor-specific targeting as found in our model: passive targeting (by shielding the surface of complexes from undesired interaction) and active targeting (by incorporation of a cell surface receptor ligand such as Tf). Passive targeting can be achieved by shielding the surface charge of complexes through covalent coating DNA/Tf-PEI800 complexes with polyethyl-eneglycol (PEG), resulting in a reduced surface charge and plasma protein and erythrocyte binding. Prolonged circulation in blood allows extravasation of DNA complexes into distant tumor tissue, which is an area of vascular leakiness. Shielding of surface charges in PEI25 or PEI22 complexes by linkage of sufficient Tf ligand also leads to preferential gene delivery into tumors, even in the absence of PEGylation.


1. Kircheis R, Kichler A, Wallner G, et al. Coupling of cell-binding ligands to polyethylenimine for targeted gene delivery. Gene Ther. 1997;4:409-418.

2. Ogris M, Steinlein P, Kursa M, et al. The size of DNA/transferrin-PEI complexes is an important factor for gene expression in cultured cells. Gene Ther. 1998;5:1425-1433.

3. Ogris M, Brunner S, Schuller S, et al. PEGylated DNA/transferrin-PEI complexes: reduced interaction with blood components, extended circulation in blood, and potential for systemic gene delivery. Gene Ther. 1999;6:595-605.

4. Kircheis R, Schuller S, Brunner, S, et al. Polycation-based DNA complexes for tumor-targeted gene delivery in vivo. J Gene Med. 1999;1:111-120.

015 Cancer therapy based on p53: The adenovirus mutant dl1520 (0NYX-015)

W. Michael Korn

Division of Gastroenterology and Comprehensive Cancer Center, University of California, San Francisco, California.

Cells infected by adenovirus resemble tumor cells in many remarkable ways, including suppression of apoptosis due to loss of p53 function and loss of cell cycle control through loss of the Rb and p53. New therapeutic strategies take advantage of this similarity by using adenovirus mutants that are defective in replication in normal cells but competent to replicate in tumor cells. The adenovirus mutant dl1520 (0NYX-015) lacks expression of the E1B55k protein, which binds and inactivates p53. Infection of normal cells harboring wild-type (wt) p53 should result in an abortive infection, because p53 counteracts the infection by inducing apoptosis or cell cycle arrest. In contrast, in tumor cells with mutant p53, the E1B55k gene is dispensable, allowing 0NYX-015 to replicate efficiently. Based on this concept, 0NYX-015 is expected to selectively destroy tumors with mutations of p53 (50-60% of all cancer types). Studies in nude mice as well as initial clinical trials demonstrated that direct injection of ONYX-

015 into solid tumors was safe and induced objective tumor remissions. However, in vitro studies showed no strong correlation between p53 status and 0NYX-015 replication. We investigated the molecular mechanisms that allow ONYX-015 to replicate in tumor cells with wt p53. The human MDM2 protein and adenovirus E1B55k show striking functional similarities, including degradation of p53, oncogenicity, and a role in RNA transport. We propose a model in which deregulated MDM2 compensates for the lack of E1B55k in 0NYX-015, enabling this virus to replicate in cells with wt p53. The deregulation of MDM2 is a consequence of loss of its major negative regulator, the tumor-suppressor protein p14ARF. This protein is lacking in most tumors cells with wt p53 and is induced after infection of normal cells with 0NYX-015. Reintroduction of p14ARF into tumor cells suppresses replication of 0NYX-015, but not wt adenovirus, in a p53-dependent fashion. These data support the therapeutic use of 0NYX-015 not only in tumors with mutant p53 but also in tumors with other lesions within the p53 pathway.

016 Cytoreductive gene therapy for human prostate cancer

Jonathan W. Simons

The Brady Urological Institute and The Johns Hopkins Oncology Center and National Institutes of Health Special Program in Research Excellence in Prostate Cancer

New therapies are urgently needed for radiation recurrent and hormone refractory metastatic prostate cancer (PCA). Cytoreductive gene therapy using tissue-specific expression of cytotoxic genes is one novel approach. We have created a number cytotoxic adenoviruses for PCA by inserting minimal enhancer/promoter elements from the PSA gene into the position of the Ad5 E1A gene promoter and expressing cytotoxic genes 3' to it. Screened head to head as potential lead compounds, a clinical candidate with impressive specificity and antineoplastic properties in clonogenic survival assays was identified. The vector, CN706 (E3del), does not carry a prodrug-activated suicide gene but rather contains the Ad5 E1A gene immediately 3' to PSA gene enhancer and promoter elements. The E1A gene is critical in triggering early events in adenoviral replication. Thus, we created a PSA expression-selective, oncolytic virus. It kills by replication, independent of cell cycle or androgen or p53 status. CN706 has promising tumoricidal properties in vivo in mice and now in phase I trials treating men with stereotactic injections who have radiation recurrent prostate cancer.

017 Modifications to improve the efficacy of replicative adenoviral vectors

R. Alemany, K. Suzuki, and D. T. Curiel

Gene Therapy Center and Division of Human Gene Therapy, University of Alabama, Birmingham, Alabama.

Conditionally replicative adenoviruses (CRAds) can be improved at the level of transduction, replication, and interaction with the immune

system. Tumor-selective replication was obtained by regulating viral genes with tumor- and tissue-specific promoters and deleting the Rb binding of E1A. For a localized viral spread and to allow insertion of longer promoters and transgenes, helper-dependent or gutless vector systems were adapted for oncolysis. Copropagation of two complementary oncolytic vectors was observed throughout injected tumors. To overcome the transduction limitations imposed by the paucity of CAR in tumors, we incorporated the integrin-binding peptide RGD into the fiber of a CRAd. Enhanced infectivity CRAds were more oncolytic both in vitro and in vivo. We also begin to address the hurdles imposed by the innate and adaptive host antiviral defense systems. In particular, clearance of CRAds from blood after systemic administration was shown to be very efficient. Possible strategies to improve tumor targeting were examined.

018 DNA replication-competent natural and recombinant parvoviruses

Jean Rommelaere

Applied Tumor Virology, Abt. F0100 and Institut National de la Santé et de la Recherche Médicale U375, Deutsches Krebsforschungszentrum, Heidelberg, Germany.

The predilection of several rodent parvoviruses for neoplastic cells has been known for a long time, as these agents were isolated from tumor implants or tumor cell filtrates under conditions in which they could not be detected in normal tissues. This oncotropism was recently exploited to isolate a new rat parvovirus from a tumor implanted in an animal that was suspected of being parvovirus-infected. Some tumor cells appear to provide these viruses with an environment beneficial to their amplification and expression. Indeed, the in vitro transformation of a number of human and rodent cells by various treatments (radiation, chemicals, oncogenes) was found to correlate with their sensitization to parvovirus-induced killing and their increased capacity for sustaining certain steps of the viral life-cycle. Transformation-enhanced events during the parvovirus growth cycle include DNA amplification and gene expression in particular, while no significant change in virus uptake has been reported thus far. Interestingly, the production and toxic activity of the nonstructural polypeptide NS1, the main viral effector of parvovirus replication and cytolysis, are both stimulated in oncogene-transformed cells. It should be stated, however, that transformed cells showing an increased permissiveness for parvovirus DNA replication, gene expression, and the resultant cyto-pathic effect are not always able to support the full infection that would lead to the release of progeny particles. The intrinsic propensity of various rodent parvoviruses to perform at least part of their lytic life-cycle in (pre)neoplastic cells makes them appealing anticancer tools. However, the mere fact that these agents were often isolated from growing tumors shows that the natural viruses are not always successful in eradicating infected tumors. However, the oncotropic and oncolytic properties of these parvoviruses may be exploited by using them as DNA replication-competent vectors to achieve a targeted coexpression of viral and heterologous genes with therapeutic potential in tumors.

019 First A. J. van der Eb Honorary Lecture: Viruses as tools for cancer gene therapy: promises and problems

A. J. van der Eb

Department of Molecular Cell Biology and Department of Radiation Genetics and Chemical Mutagenesis, Leiden University Medical Center, Leiden, The Netherlands.

The study of viruses has not only contributed to a better understanding of the mechanism by which they cause disease or how viral infection can be treated or prevented, but has also led to important new insights into cellular functions, such as the control of cell proliferation. More recently, viruses have also provided us with tools for the treatment of disease: they are sources of curative genes and have been modified to become carriers of genetic information for gene therapy. In the past decade, we have been using viruses for the latter purpose, in particular for gene therapy of cancer. I will summarize some of our results in this area of the past years and focus on some (potential) problems that we

encountered. Among the many known strategies for the genetic treatment of cancer, the combination of the herpes simplex virus thymidine kinase (HSV-tk) gene with the prodrug ganciclovir (GCV) has been the most widely used method in both preclinical and clinical trials, with some limited success for the treatment of brain cancer. We have explored the efficacy of this combination for the treatment of hepatic metastases of colorectal cancer, using adenovirus vectors expressing HSV-tk and the rat tumor model CC531 in syngeneic WAG/Rij rats. Control experiments had shown that the CC531 rat colon cancer cells were susceptible to HSV-tk/GCV treatment. The in vivo efficacy was tested by intratumoral injection or intraportal administration of the adenovirus-tk vector. Although the validity of the approach could be demonstrated, particularly after intratumoral injection, expression of HSV-tk in normal liver cells, surprisingly, was shown to provoke severe liver dysfunction after GCV treatment. This showed that, contrary to common belief, non-mitotic cells can be seriously affected by the HSV-tk/GCV treatment. Studies on the possible cause of this unexpected result revealed that mitochondrial dysfunction is most likely responsible. We showed that GCV-phos-phates are efficiently imported into mitochondria of hepatocytes that express HSV-tk, leading to abnormal mitochondria and a decreased mitochondrial membrane potential. These results demonstrate that great care should be taken if tumors outside the brain are targeted with HSV-tk/GCV. It also shows that this system does not possess the necessary tumor cell specificity required for all types of cancer gene therapy. In search of genes that have an intrinsic tumor cell specificity, we discovered the Apoptin gene, a gene derived from the chicken anemia virus. The results obtained in preclinical gene therapy studies with Apoptin will be reported in this meeting by Dr. Noteborn. However, I would like to discuss an interesting phenomenon related to Apoptin that we may have to consider when developing gene therapy strategies with this highly promising tumor cell-specific gene. The product of the gene, called Apoptin, has a unique characteristic in that it only kills tumor cells but leaves normal cells unaffected. Recently, we observed that normal fibroblasts from certain hereditary cancer-prone syndromes can become susceptible to the apoptotic activity of the protein if they are exposed to DNA-damaging agents. We also evaluated the use of adenoviruses carrying human p53 cDNA in the CC531 rat tumor model. Intratumoral administration of Ad-p53 resulted in a significant reduction in tumor size in the recipient animals and was not associated with any overt adverse side effects. High doses of Ad-p53 were well tolerated. To study the feasibility of intra-arterial administration of the vector for the treatment of hepatic micrometas-tases, we noted that although gene transfer to the liver parenchymal cells was highly efficient, virtually no cells in the tumor expressed the reporter gene. The low transduction efficiencies can most likely be attributed to the anatomical structure of the CC531 tumors, in which the tumor cells are separated from the vasculature by an extracellular-matrix barrier and layers of stromal cells. These examples show that there can be unexpected bumps on the road to gene therapy of cancer, and they also demonstrate the value of animal experiments in gene therapy studies, even if they are apparently well documented.

020 Tumor-selective transcriptional targeting for cancer gene therapy

Nick Lemoine

Imperial Cancer Research Fund Molecular Oncology Unit, ICSM at Hammersmith Hospital, London, United Kingdom. One of the most attractive ways of targeting gene therapy is by exploitation of the transcriptional regulatory elements of genes that display tissue- or tumor-selective patterns of expression. Our understanding of transcriptional control is continuing to increase and the power of expression profiling by chip technology will expand the range of candidate genes enormously. Genetic prodrug activation therapy depends upon the conditional expression of a gene encoding an enzyme capable of converting a nontoxic prodrug into an active cytotoxic agent. We have developed prototype systems based on the transcriptional regulatory elements of the human ERBB2 oncogene driving a variety of suicide genes in plasmid, retroviral, and adenoviral vectors. We have completed a phase 1 clinical trial of direct intratu-moral injection of an ERBB2-CD plasmid in patients with advanced

breast cancer, and the system is about to be applied to other tumor types. Tissue-selective targeting has been explored using the promoter and enhancer elements of the MUC1 gene, which is characteristically expressed by simple ductal epithelial cells, including those of the breast and pancreas. Combination of the MUC1 and ERBB2 elements has proved effective in mix-and-match constructs that allow dual specificity targeting. The range of suicide genes that can be delivered with such targeting devices is expanded by the use of high-capacity adenoviral vectors that allow combinations of genes to be made to increase therapeutic effect without loss of selectivity.

021 Targeted enzymes for molecular chemotherapy

H. J. Haisma, V. W. van Beusechem, D. T. Curiel, H. M. Pinedo, and W. R. Gerritsen

Department of Medical Oncology, Division of Gene Therapy, University Hospital VU, Amsterdam, The Netherlands; and Gene Therapy Center, University of Alabama, Birmingham, Alabama.

Gene therapy offers a new approach for the treatment of cancer. Recombinant adenoviral vectors are attractive in the context of cancer gene therapy because they are capable of delivering genes to a wide variety of tissues. The utility of adenoviruses is limited by their lack of specificity and by the absence of the receptor(s) for these viruses on many tumor cells. In addition, tissue penetration of the adenoviral vectors is limited to several cell layers. Therefore, adenoviral vectors have mainly been used for intratumoral injection. New strategies have been developed to overcome these problems. 0ne such approach is molecular chemotherapy. The concept underlying molecular chemotherapy is localization of chemotherapy in the region in which the transgene has been transferred. This is accomplished by injection of a vector encoding an enzyme capable of activating a nontoxic prodrug into a toxic drug. The drug is capable of diffusion into surrounding cells. However, as the enzyme is expressed intracellularly, the cells producing the enzyme are preferentially killed, which will limit the conversion of prodrug into drug and reduce the efficacy of the treatment. In addition, diffused drug may leak into the circulation, resulting in systemic toxicity. We developed vectors encoding a secreted form of the enzyme, which is also targeted to the cell surface of tumor cells by fusion to a single-chain Fv antibody. After gene transfer into tumor cells within a solid tumor, the secreted targeted enzyme, which is relatively small compared with the viral vector, will spread through the tumor tissue and bind to a great number of tumor cells. After systemic administration of a nontoxic prodrug, a drug with high cellular retention is released at the tumor site, resulting in an enhancement of the drug concentration in the tumor and, as a consequence, in a better antitumoral effect and a reduction of the systemic toxicity.

022 Herpes simplex virus thymidine kinase gene/gan-ciclovir gene therapy for primary malignant glioma: Results of a phase III study

Nikolai G. Rainov, on behalf of the multicenter North American-European study group

Department of Neurosurgery, Martin-Luther-University, Halle, Germany. The transduction of tumor cells with the herpes simplex virus thymi-dine kinase (HSV-tk) gene and their successful destruction by subsequent administration of nucleoside analogs has been demonstrated in animal models. This has been explored in uncontrolled human studies. We report here the first randomized, controlled clinical trial carried out for a malignant disease most likely to respond favorably to gene therapy. The study was conducted in adults with primary glioblastoma multiforme suitable to gross total resection. Standard therapy in all patients consisted of resection (day 0) followed by fractionated radiotherapy (56-60 Gy over 6 weeks) beginning between days 14 and 21. Gene therapy consisted of multiple injections of HSV-tk retrovirus (RV) vector-producing cells into the walls of the resection cavity at craniotomy, followed by intravenous infusion of ganciclovir (5 mg/kg twice a day on postoperative days 14-27). Tumor specimens were reviewed by a reference neuropathologist. Tumor recurrences were evaluated by enhanced magnetic resonance imaging scans at regular intervals, following a strict protocol with central neuroradiological review. Recurrence-free and overall survival times were compared in

both groups. Safety was evaluated clinically and by biological monitoring for replication-competent RV. A total of 240 patients were included, of whom 120 were randomized to receive adjuvant gene therapy. Progression-free median survival in the gene therapy group was 195 days compared with 207 days in controls; median overall survival was 368 days compared with 376 days. The 12-month survival rate was 52% compared with 44% in controls. Kaplan-Meier analysis did not indicate significant differences in progression or survival times between the groups. There was evidence of transduction in 7 of 15 samples available from recurrent tumors. Injection of vector-producing cells was well tolerated. Two treatment-related deaths were recorded: one from infection after ganciclovir-induced neutropenia and the other from cerebral hemorrhage associated with a pre-existing bleeding disorder. Apart from these events, clinical safety was comparable in both groups. There was no evidence for replication-competent RV. With no significant advantage of the gene therapy having been demonstrated, the value of this study is discussed in the context of possible reasons for failure principally related to the physical characteristics of the tumors being treated and to the gene delivery rate and the subsequent tumor-killing strategy. Other methods for enhancing the concept and improving the existing clinical protocols are also considered. Further pursuit of this gene therapy strategy must consider all of the aspects essential to adequate evaluation of the concept, such as improved delivery modes and in vivo assessment of transduction rates.

023 New suicide gene therapies

C. J. Springer,1 F. Friedlos,1 R. Spooner,2 S. Stribbling,1 J. Martin,1 L. Davies,1 I. Scanlon,1 and R. Marais2

1Cancer Research Campaign-Center for Cancer Therapeutics, Institute of Cancer Research, Sutton, Surrey, United Kingdom; and 2Cancer Research Campaign-Center for Cell and Molecular Biology, Institute of Cancer Research, London, United Kingdom.

In cancer suicide gene therapy, a vector containing the gene for a foreign enzyme is administered with the aim of restricting expression to tumor cells. Once expression of the foreign enzyme has occurred, a nontoxic prodrug is administered that is converted by the foreign enzyme at the tumor to a toxic drug. We have investigated novel suicide gene therapy regimes with carboxypeptidase G2 (CPG2), which has no mammalian homolog, in conjunction with the mustard prodrug 4-[2-chloroethyl)(2-mesyloxyethyl) amino]benzoyl-L-glutamic acid (CMDA). We found that the mutated CPG2 gene could be expressed stably in tumor cells in an enzymatically active surface-tethered form, called stCPG2(Q)3. The role of the bystander effect was examined with stCPG2(Q)3 and CMDA in the treatment of a human breast carcinoma xenograft, MDA MB 361. Cells stably expressing stCPG2(Q)3 were mixed with control ß-galactosidase (lacZ)-expressing cells to give stCPG2(Q)3:lacZ ratios of 0:100, 10:90 50:50, and 100:0 for groups 1, 2, 3, and 4, respectively. Four days after injection into nude mice, the prodrug CMDA was administered. Tumor growth delay correlated well with the levels of stCPG2(Q)3 expression: group 1, 0-day delay; group 2, 10-day delay; group 3, 16-day delay; group 4, 90-day delay. Similarly, the number of cures was strongly correlated with the level of stCPG2(Q)3 expression: group 1, 0% cured; group 2, 17% cured; group 3, 50% cured; group 4, 67% cured. There was good correlation between the CPG2 activity in the tumors and the numbers of cures. Immunohistochemical staining showed that the majority of cells from groups 2 and 3 were apoptotic, whereas those from group 1 were not, indicating a substantial bystander effect in the tumors. These results indicate that a bystander effect plays the major role in suicide gene therapy regimes with stCPG2(Q)3 and CMDA. This work was supported by the Cancer Research Campaign.

024 Bispecific single-chain diabodies secreted from mammalian producer cells for the recruitment of effector mechanisms to tumor cells

Roland Kontermann, Tina Korn, Tina Völkel, Margitta Alt, Rolf Muller, and Sabine Brusselbach

Institut fur Molekularbiologie und Tumorforschung, Universität Marburg, Marburg, Germany.

The expression of therapeutic proteins in vivo might be an alternative to the injection of the purified proteins. For example, the secretion of bispecific antibodies from producer cells in vivo could be used to target effector molecules or cells (e.g., cytotoxic T lymphocytes or prodrug-converting enzymes) to tumor cells. To explore this approach, we engineered a novel bispecific antibody format (single-chain diabody (scDb)) for secretion from mammalian producer cells. As a model, we used an scDb (scDb CEAGal) directed against carcinoembryonic antigen (CEA) and Escherichia coli ^-galactosidase (Gal) for the recruitment of an enzyme to tumor cells. Active scDb CEAGal was secreted from stably transfected human cells (HEK293) and specifically recruited ^-galactosi-dase to cocultivated CEA-expressing LoVo cells. Furthermore, specific tumor cell killing was mediated by scDb CEAGal in vitro through enzymatic conversion of a daunomycin-prodrug to the toxic drug dauno-mycin. Currently, the approach is extended by using a bispecific scDb directed against CEA and the T-cell coreceptor CD3 for the retargeting of cytotoxic T lymphocytes to tumor cells. These bispecific scDbs secreted from mammalian producer cells may be useful for gene-directed antibody therapy by combining antibody-mediated effector recruitment with a gene-therapeutic approach.

025 Imaging of herpes simplex virus thymidine kinase (HSVtk) gene expression with positron emission tomography (PET)

E. F. J. de Vries,1 P. Doze,1 A. van Waarde,1 W. Vaalburg,1 N. H. Mulder,2 and G. A. P. Hospers2

1PET Center and 2Department of Medical Oncology, Groningen University Hospital, Groningen, The Netherlands.

Objective: Although the herpes simplex virus thymidine kinase (HS-Vtk) gene has been applied in numerous gene therapy protocols, a noninvasive method to measure HSVtk expression in vivo is still unavailable. Such a method would not only be extremely helpful for optimizing gene transfer techniques, but would also allow tailoring the gene therapy protocol to the individual patient. 0ur aim is to investigate the feasibility of positron emission tomography (PET), with the ganciclovir derivative [18F]-9-[(1-fluoro-3-hydroxy-2-propoxy) methyl]guanine ([18F]FHPG) as the radio-tracer, as a noninvasive technique to image HSVtk expression.

Methods: After subcutaneous injection of C6tK+ and C6tk~ rat glioma cells, one or more HSVtk-expressing tumors and a HSVtk-negative control tumor were grown in the flanks of nude rats. After intravenous injection of [18F]FHPG, dynamic PET images were acquired. The tumors and relevant organs were dissected, uptake of radioactivity was measured, and metabolite analysis was performed. Results: Tissue distribution studies and dynamic PET imaging showed irreversible trapping of [18F]FHPG in HSVtk-expressing tumors. The tracer was rapidly cleared from control tumors and other non-target tissues into the urine. As a result, only HSVtk-positive tumors (tumorplasma ratio of 13.3 ± 3.6) and the kidneys and bladder remained visible in the PET images 2 hours after injection of the tracer. Excellent linear correlations were found between the fraction of HSVtk-expressing cells in the tumor and the [18F]FHPG tumor uptake as determined by PET (r2 = 0.998), and between the in vitro cellular [18F]FHPG accumulation and the in vivo [18F]FHPG tumor uptake (r2 = 0.997). Metabolite analysis revealed the formation of three metabolites, presumably phosphates of [18F]FHPG, in HSVtk-positive tumors. No metabolites were found in control tumors and in plasma. Conclusion: These results demonstrate that [18F]FHPG is phosphory-lated, and thus trapped, in HSVtk-expressing cells only. As a result, high target-background ratios were observed. Therefore, [18F]FHPG PET appears to be a promising technique for imaging of HSVtk gene expression in vivo.

026 Gene therapy targeting overexpression of the tumor suppressor gene p53 represents a novel approach acting at the molecular level

Concha Fernandez-Chacon

DIA/PR/Oncology, Aventis Pharma, Alcorcon, Madrid, Spain.

In three different phase II studies, RPR/INGN 201 (Ad5CMV-p53) was injected intratumorally daily in a schedule of 1, 3, or 6 biconsecu-tive days every 4 weeks (1 cycle) in patients with recurrent or refractory SCCHN. The median dose range was 2.5 X 1010 vp to 1.5 X 1012 vp. All patients had unresectable tumors. The majority of patients had disease progression after chemotherapy and/or radiation therapy. Evidence of p53 gene dysfunction was not required at study entry. The primary end-point was response rate; secondary endpoints included duration of response, overall survival, pain (VAS), and quality of life. A total of 170 (median age 62 years, male/female ratio of 4:1) had been treated, with 145 evaluable for response. A complete response was achieved in 6 lesions (with 2 being pathologically confirmed by open biopsy), a partial response was obtained in 11 lesions, and stable disease was observed for 82 lesions (for 3 to >13 months). 0verall median survival was 7.2 months. Univariate Cox analyses showed that tumor size (sum of the single longest diameter < or > than 7.5 cm) at study entry was a significant predictor of both tumor growth control and survival. In an effort to further evaluate the tolerance and safety of the product, the data from 307 patients enrolled worldwide in 47 different centers were reviewed, out of either phase I or II, by the intratumoral route of administration. Four-week cycles were repeated for as many as 18+ months, and the database reflected 702 cycles of therapy. No toxic death was reported. The most frequently related adverse events were fever/ chills/flu-like syndrome and pain at the injection site, with the majority graded mild to moderate. This significant accumulation of clinical safety and activity data justifies phase III trials (single-agent or combination trials) to evaluate the clinical benefit in this patient population with highly unmet medical needs.

027 The p53 tumor suppressor gene: From the bench top to the clinic

Stephen Chang, Ken Wills, G. William Demers, Heidrun Engler, Loretta Nielsen, Duane Johnson, JoAnn Horowitz, and Daniel Maneval

Canji, Inc., San Diego, California; and Schering-Plough Research Institute, Kenilworth, New Jersey.

Mutations in the p53 tumor suppressor gene occur in half of all cancers, and reintroduction of normal p53 function can inhibit cell proliferation and induce apoptosis in tumor cells. To evaluate the therapeutic potential of p53 gene therapy for cancer, we have developed a replication-deficient recombinant human adenovirus encoding wild-type p53 (rAd-p53). Preclinical studies with rAd-p53 using p53-altered human tumor cell lines have demonstrated inhibition of cell cycle, induction of apoptosis, and increased sensitivity to chemother-apeutic agents. Studies in animal models confirm the p53-specific antitumoral effects of rAd-p53 but highlight the challenges for efficient delivery to tumors in vivo. We have developed clinical strategies to maximize p53 gene delivery to tumors focusing on locoregional delivery to intraperitoneal ovarian cancer and liver malignancies. 0ver 150 patients have been treated in phase I clinical trials, and safe doses for future investigation have been identified. Analysis of tissue biopsies demonstrated transgene expression in both normal and tumor tissue that was not abrogated by serum anti-adenovirus antibodies. Preclin-ical data supporting these clinical targets will be summarized together with results from ongoing clinical investigation.

028 Apoptin

M. H. M. Noteborn

LEADD BV and Department of Molecular Cell Biology, Leiden University Medical Center, Leiden, The Netherlands.

Apoptin, derived from an avian virus, induces apoptosis in various human transformed and/or tumorigenic cell lines, but not in normal primary cells.1 However, loss of a tumor-suppressor gene renders normal human diploid cells susceptible to Apoptin-induced apoptosis, if they are exposed to a mutagenic radiation dose, which can be "handled" in diploid cells from healthy individuals.2 In Apoptin-induced apoptosis, upstream caspases are not involved but downstream caspases seem to be of importance. In all (human) tumor cells, analyzed so far, Apoptin is located mainly in the nucleus, whereas in

"normal" cells, it is present in the cytoplasm.3 Furthermore, Apoptin-induced apoptosis is independent of p53, is not inhibited by BAG-1 or Bcr-Abl, and is even accelerated by Bcl-2. These features indicate that Apoptin induces apoptosis via a novel tumor-specific pathway and that, more importantly, Apoptin will be able to induce apoptosis when many (chemo)therapeutic agents are known to fail (Noteborn and Van der Eb, Biogenic Amines, Volume 15, page 73). To explore the effect of Apoptin in vivo, we have constructed an adenovirus vector expressing the Apoptin gene (Ad-Apoptin). By means of terminal deoxynucleoti-dyltransferase-mediated deoxyuridine triphosphate nick end labeling assays, Ad-Apoptin was shown to induce apoptosis specifically in tumor cells. Experiments in rats show that Ad-Apoptin can be safely administered via a single as well as repeated intravenous injections, indicating that Ad-Apoptin has no severe adverse effects. A single intratumoral injection of Ad-Apoptin into a xenogeneic hepatoma in nude mice resulted in a significant reduction of tumor growth.4 Various aspects of the tumor cell specificity of Apoptin and its unique apoptotic pathway will be discussed that make Apoptin a potential antitumoral agent.


1. Danen-Van Oorschot AA, Fischer DF, Grimbergen JM, et al. Apoptin induces apoptosis in human transformed and malignant cells but not in normal cells. Proc Natl Acad Sci USA. 1997;94:5843-5847.

2. Abrahams PJ, van der Eb AJ, Noteborn MH. The viral protein Apoptin induces apoptosis in UV-C-irradiated cells from individuals with various hereditary cancer-prone syndromes. Cancer Res. 1999;59:3010-3015.

3. Danen-Van Oorschot AA, Zhang Y, Erkeland SJ, Fischer DF, van der Eb AJ, Noteborn MH. The effect of Bcl-2 on Apoptin in "normal" vs transformed human cells.Leukemia. 1999;13(Suppl 1):S75-S77.

4. Pietersen AM, van der Eb MM, Rademaker HJ, et al. Specific tumor-cell killing with adenovirus vectors containing the apoptin gene. Gene Ther. 1999;6:882-892.

029 Multimodality anticancer treatment based on ad-enovirus-mediated delivery of bax

J. Gomez-Navarro, J. Xiang, W. Arafat, D. Buchsbaum, M. Stack-house, R. D. Alvarez, and D. T. Curiel

Division of Human Gene Therapy and Radiation Biology, University of Alabama, Birmingham, Alabama.

Bax, a member of the Bcl-2 family that can act as a tumor suppressor, potently induces apoptosis. We have explored its therapeutic potential. First, we constructed a recombinant adenoviral vector encoding Bax (Ad/Bax) without compromising the survival of virus-packaging cells using the inducible Cre-loxP system. Next, we evaluated the cytotox-icity of Bax in cancer cells representing a wide variety of tissues. We infected the cells with increasing viral doses and measured viability 48 hours later by a quantitative MTS assay. Expression of Bax resulted in apoptotic cell death in most human cancer cells, although the sensitivity of each cell line varied. Frequently, 80-90% of cell killing was achieved with a multiplicity of infection of 100, although occasionally higher doses were needed or resistance was observed. Differential sensitivity did not correlate with cell infectibility by Ad. Importantly, Bax-mediated death was observed in immunopurified, primary patient-derived ovarian cancer cells from several individuals but did not occur when peritoneal mesothelial cells were infected and expressed Bax. This apparent specificity was supported by an in vivo experiment, whereby Ad/Bax and Ad/Cre were administered intravenously into severe combined immunodeficient mice at a dose of 1 X 109 plaque-forming units. After 72 hours, high levels of Bax were detected in the liver by Western blot, but no evidence of toxicity was microscopically evident. Then, to establish the dependence of Bax-mediated cytotox-icity upon other mediators of the apoptosis pathways, we determined by Western blot the endogenous levels of both Bcl-2 and Bax proteins and p53 status. We found no correlation with any of those parameters. Next, to evaluate the potential of Bax for sensitizing tumors that are refractory to conventional chemotherapy and radiotherapy, we tested

the combined treatments and determined dose-response curves and cell colony formation assays in several cell lines and primary ovarian cancer cells and found a very significant enhancement of cytotoxicity in cells particularly refractory to both chemotherapy and radiotherapy. Finally, we evaluated the capacity of the combined Bax and radiation treatment to inhibit tumor growth in a murine subcutaneous model. When mice were injected with SKOV3.ip1 cells exposed to Bax, RT, or both, tumor nodules developed in all groups except in those animals that received the combination. Thus, production and delivery of Bax via a recombinant adenovirus is feasible, induces apoptosis robustly and preferentially in human cancer cells, and augments the efficacy of both chemotherapy and radiotherapy in otherwise refractory tumor cells. The independence of these biological effects of the Bcl-2, Bax, and p53 status suggests its potential utility for overcoming tumor heterogeneity. Further vector developments to allow efficient intratu-moral Bax delivery in vivo are ongoing.

030 Inducing apoptosis in brain tumor cells with adenovirus-mediated transfer of fas ligand-green fluorescent protein fusion gene

Danher Wang,1 Semyon Rubinchik,1 Rong-xian Ding,1 Mitchel S. Berger,2 and Jian-yun Dong1

'Medical University of South Carolina, Charleston, South Carolina; and 2University of California, San Francisco, California. Brain tumors, including astrocytoma and glioblastoma, are incurable and highly malignant. Surgery and high-dose radiation therapy only provide short-term repression, and complications of these therapies are often severe. Our data have shown that gene therapy with apoptotic signaling genes may provide a promising alternative to treat brain tumors. To study the feasibility of forcing brain tumor cells into apoptosis, we constructed fusion proteins that contain an extracellular domain that binds to Fas receptor and an intracellular domain derived from the jellyfish green fluorescent protein. Delivery of this fusion protein with an adenovirus vector efficiently induced apoptosis in cells from nine clinical brain tumor isolates. In vivo studies using mice with xenografted brain tumors showed dramatic suppression of tumor growth; in some cases, mice remained tumor-free until the time of sacrifice for histological analysis. In contrast, all control mice had to be sacrificed within one month due to massive tumor. Our data support the notion that gene therapy with genes that signal apoptosis may be an applicable strategy for treating brain tumors.

031 Modification of vascular tone using inducible nitric oxide synthase under the control of a radiation-inducible promoter

J. Worthington, M. Murray, M. O'Rourke, G. Keilty, D. G. Hirst, and Tracy Robson

School of Biomedical Sciences, University of Ulster, Jordanstown, N. Ireland.

It may be therapeutically advantageous to alter tumor blood supply specifically. Nitric oxide (NO) is a potent vasodilator that is produced in many tissues by the enzyme NO synthase (NOS). We have trans-fected cDNA for the inducible isoform of this enzyme (iNOS) under the control of the radiation-inducible promoter WAF1. The activity of the promoter was initially assessed using green fluorescent protein in both endothelial cells and rat tail artery segments. Induction of protein expression by 9.5- and 4.5-fold, respectively, was observed after a radiation dose of 4 Gy. Artery sections were then transfected with the WAF1/iNOS construct; this gave a 5-fold induction of iNOS protein after a dose of 4 Gy. The transfected artery was also tested functionally for relaxation, which is indicative of NO production. One hour after exposure to 4 Gy, there was a significant (65%) relaxation of artery segments that had been preconstricted with phenylephrine. This could be partially reversed by the NOS inhibitor nitro-L-arginine. This study demonstrates that we can regulate vascular tone using an x-ray-inducible promoter.