Oral Abstracts

Differential impact of local production interleukin-2 and interleukin-12 on initial appearance and long-term progression of B16F10 melanoma in mice

Morihiro Watanabe and Robert H. Wiltrout
Laboratory of Experimental Immunology, NCI Center for Cancer Research

Background: Administration of Interleukin-2 (IL-2) and interleukin-12 (IL-12) proteins have been shown to induce potent antitumor activities in vivo by activating T-lymphocytes and natural killer (NK) cells. IL-12 also induces Th1 type responses and inhibits angiogenesis and seems to be a more potent anti-tumor agent than IL-2 in a variety of tumor models. However, the relative abilities of these two cytokines to induce the initial phases of tumor response vs. establish immunity to rechallenge are not clear.

Objectives: The purpose of this study is to contrast long-term vs short-term survival of mice challenged with B16 melanoma cells genetically engineered to secrete IL-12 or IL-2.

Methods: The IL-12 and IL-2 cDNAs were subcloned into the pMX retrovirus vector and polytropic retrovirus was generated using these constructs and vesicular stomatitis virus G envelope protein. Green fluorescent protein (GFP) was used as a control gene. IL-12 and IL-2 cDNAs were transduced to B16F10 melanoma. Transfected cells (2×105) were injected in the flank of C57BL/6 mice.

Results: Tumors appeared in 7 of 8 mice within 7 days after the injection of GFP-transduced B16F10 cells and in 7 of 8 mice within 11 days after injection with parental cells. Within three weeks, all the mice in these two groups were moribund and were euthanized. In contrast, IL-12 transduced B16F10 cells did not grow a palpable tumor in the first three weeks, but by day 37, 7 of 8 mice did show tumor at the injection site. Five months after tumor implantation, only 4 mice injected with the IL-12 transfectants were alive with 2 of 4 showing tumors. All mice challenged with the IL-2 secreting tumor cells initially developed tumors smaller than 64mm3, and all of these tumors disappeared by 17 days. Eighty-five percent of these mice remained tumor free 5 months later. Discussion: These findings suggest that although IL-12 inhibits tumor establishment quite effectively, the mechanisms by which it does so may fail to induce memory against the B16F10 cells in most cases. Conversely, although local production of IL-2 in the tumor microenvironment may not be potent enough to prevent initial tumor growth, the mechanism by which tumor responses are induced more effectively engages host adaptive immune responses. These findings illustrate the distinct mechanisms by which these two cytokines mediate tumor response and provide additional insight into recent observations showing therapeutic synergy between IL-2 and IL-12.

A DNA Vaccine Encoding Genetic Fusions of CEA and GMCSF

Jose Lima, Connie Jenkins, Mary Hamilton, Pierre Triozzi, Denise Shaw, and Theresa Strong
Department of Medicine and the Comprehensive Cancer Center, The University of Alabama at Birmingham, Birmingham, AL

Background: Plasmid DNA vaccines encoding tumor antigens have shown promise in animal models, but limited efficacy in the clinical setting. Carcinoembryonic antigen (CEA) is an intercellular adhesion molecule expressed in several cancers and represents a promising tumor-associated antigen for a vaccine. GM-CSF is well established as a potent immune adjuvant, in part due to its ability to recruit and activate dendritic cells.

Objective: Investigate the use of plasmid DNA encoding CEA and GM-CSF fusion proteins to enhance immune response.

Methods: We constructed two plasmids encoding fusions between CEA and murine GM-CSF (mGM-CSF). CEA was fused with GM-CSF in the carboxy or amino terminal, with a short, flexible linker joining the two moieties. Plasmids were injected i.m., and immune response was assessed by measuring anti-CEA T cell and antibody and by tumor challenge studies.

Results: In vitro studies validated that both of the fusion proteins were produced and secreted with both constructs. GM-CSF activity was confirmed with a GM-CSF-dependent mouse cell line. Immunization of C57/BL6 mice with plasmids encoding CEA-GM-CSF fusions led to T cell and antibody responses against CEA. These responses were comparable to immunization with a plasmid encoding full length CEA alone. Tumor challenge with CEA-expressing syngeneic mouse adenocarcinoma cells (MC38-CEA) resulted in the development of large tumors in control groups by day 25. In contrast, no tumors were noted at this time in mice immunized with the CEA-GM-CSF fusion plasmids or with the CEA plasmid. Subsequently, tumors developed at day 35 in mice immunized with a higher dose of the CEA-GM-CSF fusions. Immunization with lower doses of the CEA-GM-CSF fusion constructed with GM-CSF at the amino terminal, however, provided better tumor protection than the CEA plasmid. Mice injected with higher doses of both CEA-GM-CSF fusions developed IgG autoantibodies to mGM-CSF that neutralized mGM-CSF activity in vitro. Mice injected with a plasmid encoding GM-CSF alone did not produce such antibodies.

Discussion: GM-CSF-CEA fusion DNA vaccines elicited anti-CEA immune responses. Anti-mGM-CSF antibody was induced in a dose-dependent manner. A single, low dose immunization resulted in lower titers of anti-mGM-CSF antibodies and better tumor protection than a plasmid encoding CEA alone.

Systemic p53 Gene Therapy of Cancer with Immunolipoplexes Targeted by Anti-Transferrin Receptor scFv

Xu, Liang; Pirollo, Kathleen F.; Chang, Esther H.
Department of Oncology, Lombardi Cancer Center, Georgetown University Washington, DC

A long-standing goal in genetic therapy for cancer is a systemic gene delivery system that selectively targets tumor cells, including metastases. A cationic immunolipoplex system containing an anti-transferrin receptor single-chain antibody (TfRscFv) as the targeting ligand has been developed in our laboratory to target tumor cells. For characterization of the immunolipoplexes, the sizes and zeta potentials were measured by photon correlation spectroscopy (PCS). TfRscFv-liposomes have a size of 25.2 + 0.6 nm in diameter while the DNA-complexed TfRscFv-immunolipoplexes have a size of 92.9 + 8.9 nm with a zeta potential of +52.7 + 1.8 mV (n=3). More importantly, the immunolipoplexes in 5% dextrose are quite stable, with a size of 81.8 + 2.0 nm 24 hr after preparation when stored at 4 0C. Compared with untargeted lipoplexes, the TfRscFv-targeted immunolipoplexes showed enhanced tumor cell binding, improved gene delivery and transfection efficiencies both in vitro and in vivo. The tumor suppressor gene p53 has been shown to be involved in the regulation of DNA damage-induced apoptosis. Loss or malfunction of this p53-mediated apoptotic pathway has been proposed as one mechanism by which tumors become resistant to chemotherapy or radiation. In in vitro and in vivo experiments, we have stepwise proved in principle that the restoration of normal p53 functions renders the tumor cells more sensitive to the induction of apoptosis by conventional chemo/radiotherapy in various tumor models. Using the systemically administered TfRscFv-immunolipoplexes we were able to deliver the p53 gene to xenografts of human prostate, breast and head and neck cancer, and show that the therapeutic gene was functionally expressed. Moreover, in a human breast cancer metastasis mouse model, combination of systemically delivered TfRscFv-liposome-p53 and docetaxel showed improved efficacy with prolonged mouse life-span. The TfRscFv-immunolipoplex is a promising method for systemic gene therapy of cancer. The sensitization of tumors to chemotherapy and radiation could lead to lowering the effective dose of these anticancer treatments, correspondingly lessening the severe side effects, while decreasing the possibility of recurrence. More significantly, this approach can be applied to both primary and recurrent tumors as well as metastatic disease. This novel strategy combining systemically delivered molecular medicine with current chemo/radiotherapy has the potential to critically impact the clinical management of cancer.

Thalidomide and its Successor Molecules as Agents for Biological Therapy

Jerome B. Zeldis, M.D., Ph.D., Chief Medical Officer, Celgene Corporation, Warren, NJ

In every cell evaluated, thalidomide affects the cells’ ability to respond to a variety of biological stimuli. Specifically, thalidomide inhibits the production of both TNFa and COX-2 via destabilizing their mRNA. Thalidomide also affects IL-1b, IL-6, IL-10, IL-12, GM-CSF, and adhesion molecule expression. The drug is also anti-angiogenic both as a consequence of the above and by direct effects on bFGF and VEGF expression. Some of thalidomide’s effects on cytokines are paradoxical, either stimulating or inhibiting their expression depending on the type of co-stimulation that is also occurring. These biological properties partially explain its potential efficacy in treating both inflammatory and neoplastic conditions. Thalidomide has remarkable activity in multiple myeloma, myelodysplastic syndrome, colon cancer, renal cell carcinoma, prostate cancer, congestive heart failure, Crohn’s Disease, and ulcerative colitis.

Celgene had developed successor compounds to thalidomide that maintain most of their biological activities, yet are not teratogenic in the New Zealand Rabbit teratogenicity model and are not sedating or constipating in humans. The lead compound, REVIMIDÔ is currently in Phase I/II trials evaluating its safety and efficacy in solid tumors, multiple myeloma, congestive heart failure and inflammatory bowel disease.

From Coley’s Toxins to Immunotherapy with CpG DNA Motifs from Bacterial DNA

Arthur M. Krieg, Coley Pharmaceutical Group, Wellesley, MA, and University of Iowa Cancer Center, Iowa City, Iowa.

More than a century has passed since William B Coley began treating patients with advanced cancer with bacterial extracts, demonstrating a surprisingly high response rate. The active ingredient of “Coley’s toxins” was not identified during his lifetime, but current studies suggest that the bacterial DNA may have been an important component of his preparations. Bacterial and vertebrate DNAs differ by the absence of CpG suppression or methylation in bacteria. The immune system has evolved to detect the presence of unmethylated CpG dinucleotides in particular sequence contexts (“CpG motifs”) as a signal of infection.The immune stimulatory effects of bacterial DNA can be mimicked by CpG motifs contained in synthetic oligodeoxynucleotides (ODN), which are produced with a nuclease resistant phosphorothioate backbone for improved stability. B cells are induced by CpG DNA to proliferate and secrete immunoglobulin; dendritic cells secrete a wide array of cytokines, interferons, and chemokines and express increased costimulatory molecules. The cytokine response is dominated by IL-12 and type I interferons, creating an environment that promotes Th1-like immune responses and NK cell activation. CpG DNA costimulates B cell activation through cell membrane Ig thereby promoting the development of antigen-specific responses. Together, these factors activate a coordinated set of immune responses with both innate immunity (macrophages, monocytes, dendritic cells, and natural killer cells), and adaptive immune responses, including humoral and cellular immunity. When combined with protein antigens, CpG ODNs are also extremely effective vaccine adjuvants, inducing Th1 responses with both specific antibody and CTL. In mouse cancer models, immune stimulation by CpG DNA has therapeutic activity as a single agent through induction of nonspecific innate immune effector mechanisms, for vaccination against tumor antigens, and also to enhance the efficacy of antitumor antibodies via ADCC. Preliminary results from phase I human clinical trials in normal volunteers indicate that CpG DNA is relatively well tolerated. As an adjuvant for a hepatitis B vaccine, CpG DNA appears to induce earlier seroconversion with the production of increased levels of antibody. Phase II human clinical trials of CpG DNA in cancer patients are underway.


B. Jansen, V. Wacheck, E. Heere-Ress, C. Hoeller, C. Krepler, T. Lucas, K. Wolff, H. Pehamberger. Depts. of Dermatology and Clinical Pharmacology, University of Vienna, Austria

Treatment resistance has been linked to over-expression of the proto-oncogene BCL-2 in a number of human cancers, including melanoma. Advanced malignant melanoma is a paradigm of a treatment-resistant neoplasm. Genasense™ (formerly known as G3139, Genta Inc., Berkeley Heights, NJ) is an antisense phosphorothioate oligonucleotide that targets the first 6 codons of the BCL-2 mRNA. Preclinical studies have shown that Genasense™ decreased BCL-2 protein and enhanced tumor cell apoptosis both in vitro and in vivo. Genasense treatment led to major tumor responses in a SCID mouse xenotransplantation model for human melanoma when given as a single agent; treatment effects were even more pronounced when combined with systemic administration of dacarbazine (DTIC) rendering 10 of 13 animals with xenotransplanted human melanomas without detectable tumors. We recently showed that Genasense could be safely administered by continuous IV (CIV) infusion in combination with full-dose DTIC in a phase I/II trial in patients with advanced melanoma. Furthermore, this trial demonstrated that Genasense treatment down-regulated BCL-2 protein assessed by Western blot analysis in serial tumor biopsies from patients with melanoma, and that this biologic activity was associated with major clinical responses. Overall survival of advanced-stage, mainly second line patients has exceeded 1 year. Since maximal down-regulation of BCL-2 protein appeared to occur by day 4 of the Genasense infusion, an international randomized trial has been initiated in patients with advanced melanoma using a 5-day pre-treatment regimen of Genasense administered by CIV at a dose of 7 mg/kg/day, followed by DTIC (1000 mg/m2) administered immediately thereafter, with repeat courses cycled every 3 weeks. In clinical studies conducted to date in melanoma and other diseases, fever and fatigue have been the most prominent reactions up to and including Genasense doses of 9 mg/kg/day. Thrombocytopenia was dose-limiting when Genasense at 12 mg/kg/day was combined with DTIC at 1000 mg/m2. Additional controlled trials have been initiated in multiple myeloma, leukemia, and other indications in combination with other chemotherapeutic agents for these diverse diseases that are known to express BCL-2 in order to enhance the therapeutic efficacy of current anticancer treatment regimens.