Tumor vaccines and gene therapy have received significant attention as means of increasing cellular and humoral immune responses to cancer. We conducted a pilot study of seven research dogs to determine whether intradermal injection of canine tumor cells transfected via the Accell particle-mediated gene transfer device with the cDNA for human granulocyte-macrophage colony-stimulating factor (hGM-CSF) would generate biologically relevant levels of protein and result in demonstrable histological changes at sites of vaccination. Tumor cell vaccines of 10 7 irradiated canine melanoma cells were nontoxic, safe, and well tolerated. No significant alterations in blood chemistry values or hematological profiles were detected. A histological review of control vaccine sites revealed inflammatory responses predominated by eosinophils, whereas vaccine sites with hGM-CSF-transfected tumor cells had an influx of neutrophils and macrophages. Enzyme-linked immunosorbent assays of skin biopsies from vaccine sites had local hGM-CSF production (8.68 -16.82 ng/site of injection) at 24 hours after injection and detectable levels (0.014 -0.081 ng/site) for Յ2 weeks following vaccination. Flow cytometric analysis of hGM-CSF-transfected cells demonstrated Յ25% transfection efficiency, and hGM-CSF levels obtained during time-course assays demonstrated biologically relevant levels for both irradiated and nonirradiated samples. These data demonstrate the in vivo biological activity of irradiated hGM-CSF-transfected canine tumor cells and help provide evidence for a valid translational research model of spontaneous tumors.

Immunocompetent cells in bone marrow allografts have been associated with a graft-versus-leukemia (GVL) effect. To further characterize effector mechanisms that may be involved in this GVL phenomenon, we have previously established an in vitro model to identify allogeneic T- cell clones that selectively mediate cytotoxicity against a patient's leukemic cells, but not against nonleukemic lymphocytes from the same patient. We have modified this in vitro model to test whether the Ph1 chromosome and the P210 fusion protein it controls have a detectable role in leukemia-specific recognition by allogeneic T-cell clones. In this report, T-cell lines reactive with allogeneic Ph1 chromosome- bearing (Ph1+) chronic myeloid leukemia (CML) cell lines were derived and selected to be minimally reactive with Ph1 negative (Ph1-) lymphoid lines from the same patient. However, after prolonged culture, these same T-cell lines also mediated significant destruction of the Ph1- target cells from the same patients. These T-cell lines specifically recognized cells from the allogeneic CML patient to which they were sensitized, and were not contaminated by an outgrowth of natural killer cells. Furthermore, subclones could be derived from these T-cell lines, and some of these subclones again showed selective killing of the allogeneic Ph1+ leukemia cell lines, and not of the Ph1- cell line from the same patient. Analyses of T-cell receptor (TCR) genes showed the alloreactive T-cell lines and the Ph1+ selective subclones derived from them to be of the same clonal origin. This suggests that the same T cells reacting with antigens expressed on the nonleukemic Ph1- targets can at times selectively and preferentially kill the allogeneic Ph1+ cells. As the same TCR that recognizes Ph1+ cells also can recognize the Ph1- targets, it appears that the Ph1+ chromosome does not play a detectable role in recognition by these allogeneic T-cell clones. This in vitro observation may provide a model for evaluating the relationship between GVL and graft-versus-host disease effects.