Oral Abstracts

Tumor Vaccination with Interleukin-12 and Granulocyte Macrophage-Colony Stimulating Factor-Encapsulated Microspheres: Co-induction of Innate and Adaptive Immunity Promotes Cure of Disseminated Disease
Hank Hill1, Thomas F. Conway, jr.2, Michael S. Sabel1, Yong S. Jong4, Edith Mathiowitz4, Richard B. Bankert2, 3 and Nejat K. Egilmez2, 3
Departments of Surgery1 and Immunology2, Roswell Park Cancer Institute, Buffalo NY; 3Department of Microbiology, State University of New York at Buffalo, Buffalo, NY; and 4Department of Molecular Pharmacology and Biotechnology, Brown University, Providence, RI

The efficacy of neoadjuvant tumor vaccination with combined Interleukin-12 (IL-12) and Granulocyte Macrophage-Colony Stimulating Factor (GM-CSF)-encapsulated microspheres was evaluated in a murine surgical metastasis model. Primary subcutaneous Line-1 tumors (a lung alveolar carcinoma) were induced in BALB/c mice and were allowed to metastasize to the lungs spontaneously. Mice with established metastases were then vaccinated in situ with a single injection of control or cytokine-loaded microspheres into the primary tumor. Primary tumors were surgically excised one week after treatment and the mice were monitored for development of lung tumors. Analysis of the lungs 5 weeks after surgery demonstrated that vaccination with IL-12 + GM-CSF microspheres was highly effective in suppressing metastatic disease (0.4 tumor nodules/lung). Mice that received IL-12 microspheres alone were also protected but had a significantly higher tumor burden (2.4 nodules/lung). The mice that were treated with control or GM-CSF microspheres were not protected (7.4 nodules/lung and 8.4 nodules/lung, respectively) as compared to mice that were treated with surgery alone (8.2 nodules/lung). Survival analysis demonstrated that combined IL-12 and GM-CSF vaccination resulted in the complete cure of metastatic disease in the majority of the animals (60% survival versus 6.7% in the control group). In vivo lymphocyte subset depletions indicated that while both T and NK cell subsets were important in the suppression of primary tumors, the anti-metastatic effect was mediated primarily by the NK/NK-T subsets. The suppression of metastatic disease, however, also involved a cognitive immune component as long-term, tumor-specific T-cell activity was demonstrated by immunohistochemical analysis of metastatic lesions, ELISPOT assays and tumor challenge studies. These results establish that vaccination strategies targeting both innate and adaptive immune mechanisms in the minimal residual disease setting may provide a highly effective approach in the clinic.

A Randomized Phase III Trial of High-Dose Interleukin-2 (HD IL2) versus Subcutaneous (SC) IL2/Interferon (IFN) in Patients with Metastatic Renal Cell Carcinoma (RCC)
D McDermott, L Flaherty, J Clark, G Weiss, T Logan, M Gordon, J Sosman, M Ernstoff, W Urba K Margolin, J Dutcher, M Atkins
Cytokine Working Group (CWG) Beth Israel Deaconess Medical Center, Boston, MA

HD IL-2 produces durable responses (> 3 years) in approximately 10% of (pts) with RCC but is associated with significant toxicity and cost and is not universally offered and administered. A low dose outpt IL-2/IFN produced a similar response rate (RR) and median survival in a phase II trial (Dutcher, Cancer J Sci Am 1997, pS73). We conducted a Phase III trial to determine the value of outpt sc IL-2/IFN relative to HD IL-2. Pts were stratified for bone and liver metastases, primary in place and PS 0/1. Pts were randomized to receive either IL-2 (5 MIU/m2 sc, q8 hours [h] x 3 doses on day [d] 1 then daily 5 d/week [wk] x 4 wks) and IFN-a 2B (5 MIU/m2 sc, TIW x 4 wks) every 6 wks or HD IL-2 (600,000 IU/kg/dose IV q8 h, d 1-5 and 15-19 [max 28 doses]) every 12 wks. Tumor responses were assessed at wks 6 and 12, then q12 weeks. Responding pts on IL2/IFN received up to 6 cycles at 6 week intervals; on HD IL2 up to 3 cycles at 12 wk intervals. Results: 193 pts (age: 21-75, (median 54), M/F: 128/56, IL2/IFN-94, HD IL-2-99) were enrolled between 4/97 and 7/00. Treatment arms were balanced for the presence of bone or liver mets (41/40), PS 0 (61/60) and primary in place (16/17). Toxicities seen were typical for these regimens, including 1 treatment related death from progressive disease and ARDS on IL-2/IFN and one death from capillary leak on HD IL2. The RR for HD IL-2 vs IL-2/IFN was 26% (25/97) vs 11% (10/90) (p=0.01) and CR rate was 8% vs 2%. The difference in RR between the arms was 14.6%, with 95% confidence intervals of 3.9 – 25.3%. The median response duration was 16 vs 13 months (p=0.14), and median survival 15 vs 12 months (p=0.08). Ten responses on HD IL-2 vs 4 on IL-2/IFN are ongoing at ranges of 16-38 and 11-37 months. Conclusions: HD IL-2 produces significantly more responses of apparent better quality than IL-2/IFN. The primary endpoint of the study, 3-year progression free survival, has yet to be reached and overall survival results remain preliminary. At this early analysis, we conclude that HD IL-2 should remain the preferred therapy for appropriately selected patients with access to such treatment.

Title: Phenotypic Diversity of Endothelial Cells is Defined by the Organ Microenvironment Cheryl Baker, PhD

We determined whether the interaction of mouse and human endothelial cells with specific growth factors leads to upregulation of their receptor. Sparse cultures of endothelial cells of the dermis incubated in medium containing EGF or TGF-a (but not other proangiogenic molecules) expressed EGF-R and phosphorylated EGF-R demonstrable by immunohistochemical staining. Gelfoam sponges containing 0.4% agarose with different proangiogenic molecules were implanted into the subcutis of C3H/HeN mice. Microvessel density was determined histologically subsequent to immunostaining with anti-CD31 antibodies. All proangiogenic factors induced angiogenesis, but expression of EGF-R and activated EGF-R was found only on endothelial cells induced by EGF or TGF-a. PKI166, a selective inhibitor of the EGF-R protein tyrosine kinase, produced significant apoptosis only in endothelial cells expressing activated EGF-R under both in vivo and in vitro conditions. Collectively, the data show that the expression of EGF-R and activated EGF-R by endothelial cells is conditioned by the microenvironment.

The Imaging of Angiogenesis
Peter L. Choyke, M.D.

Angiogenesis is an important component of tumor growth and as volume increases the tumor becomes increasingly dependent on angiogenesis for further growth. This process is heterogeneous within tumors. Traditional imaging techniques such as CT, MRI and ultrasound produce excellent morphologic images which are capable of resolving very small lesions. However, these anatomic images do not provide insight into the physiology or “function” of angiogenesis within the tumors. The current challenge is to combine this readily available morphologic information with functional information to provide a more comprehensive evaluation of tumor response. Dynamic enhanced MRI is a technique that could be widely available with little extra cost. It involves the acquisition of serial MR images through a tumor before and after the injection of an MR contrast agent. The time course of the contrast agent within the tumor varies widely from highly vascular and permeable tumors that wash in and wash out quickly to tumors that demonstrate slow steady enhancement. Such patterns are often observed within the same tumor. These time activity curves can be fit to a simple two compartment pharmokinetic model, from which descriptive parameters, reflecting the curve fitting process, can be derived. These parameters can, in turn, be used to color encode the image. These areas can be compared with molecular studies. They can also be monitored during anti-angiogenic therapies to detect tumor response even before changes in volume can be detected. Positron Emission Tomograpy (PET) utilizing fluorodeoxy glucose (FDG) is another method of assessing tumor function. FDG PET directly measures glucose entrapment by tumors and only indirectly measures angiogenesis. Interestingly, FDG uptake on PET scanning often correlates with angiogenesis since angiogenesis and metabolism are linked. Other PET techniques include O15 water studies directly measure tumor flow. However, the resolution of this technique limits regional assessment of tumor blood flow. Both dynamic enhanced MRI and FDG PET scans often qualitatively match each other and rarely display contradictory information raising the question whether both are necessary to assess tumor function. The future use of MRI may be influenced by new types of contrast agent which may be more selective for permeable tumor vessels. As more molecular biologic correlation is obtained, the relative roles of MRI and PET will be better understood.

Molecular Determinants of the Angiogenic Switch and Tumor Progression
Marsha A. Moses, Ph.D.
Surgical Research Laboratories, Children’s Hospital and Harvard Medical School, Boston, MA 02115

Our laboratory has had a long standing interest in identifying the molecules, and understanding the mechanisms, that control angiogenesis during tumor development and progression. Recently, we have focused our interest on a critical checkpoint in tumor development known as the “switch to the angiogenic phenotype”. In order to study this rate-limiting step in successful tumor growth, we have developed an in vivo tumor model that reproducibly recapitulates the angiogenic switch. This model permits the harvest and study of tumor nodules that can be distinguished from each other on the basis of their degree of vascularization, ie. avascular and vascular tumor lesions. Using experimental strategies that include transcriptional profiling as well as molecular and biochemical approaches, we have begun to identify the molecules that trigger this angiogenic switch. Utilizing in vivo and in vitro angiogenesis assays, substrate gel electrophoresis and antisense technology, we have now demonstrated that MMP-2 (Matrix Metalloproteinase-2; gelatinase A) triggers the switch to the angiogenic phenotype in our experimental model. In a second series of studies, we have also determined that HIF-1a-mediated upregulation of VEGF (Vascular Endothelial Growth Factor), independent of bFGF (basic Fibroblast Growth Factor), is necessary for successful transition through this critical checkpoint. This finding was supported in part by the fact that the nuclear translocation of HIF-1a was detected exclusively in avascular tumor nodules when compared to their vascular counterparts. Finally, using microarray technology, differential gene expression between avascular and vascular tumor nodules was analyzed. Candidate genes were identified that represent potential positive or negative regulators of the angiogenic switch. These candidate genes include, but are not limited to, transcription factors, extracellular matrix proteins and metalloproteinases. Continued identification and characterization of the molecular mechanisms underlying the angiogenic switch during early tumorigenesis has the potential to provide insights into developing strategies for early cancer treatment, diagnosis and perhaps prevention.

Endothelial Cell Gene Expression Profiling
Steven K. Libutti, MD

Tumor angiogenesis is a complex process involving the interaction of tumor-derived factors and endothelial cells resulting in the growth of new blood vessels into the tumor. This process is required for tumors to grow larger than a few millimeters and to gain access to the blood supply in order to invade and metastasize. The complex changes in the regulation of endothelial cell proliferation, maturation and vessel formation are only now being understood in the context of the host response to the tumor. A better understanding of the pathways involved in this process will lead to the discovery of novel targets for anti-angiogenic cancer therapy. Our laboratory is studying the changes in gene expression of endothelial cells growing in a tumor microenvironment using a variety of techniques. We have been able to successfully model tumor angiogenesis using ex-vivo and in vivo models such as the rat aortic (and human saphanous vein) ring assay, as well as matrigel and tumor implants in animals. Using techniques of laser capture microdissection, RNA extraction and amplification as well as cDNA gene expression profiling, we are beginning to gain a better understand of angiogenesis and the gene regulation level. Our strategies and early results will be presented.