Analysis of Changes in Gene Expression Induced by Fusion of Dendritic Cells and Melanoma Cells.

Heterokaryons formed through electrofusion of Dendritic cells (DC) with tumor cells are highly effective vaccines for cancer immunotherapy. We have demonstrated in transplantable murine tumor models that active immunotherapy using a combination of DC-tumor fusion vaccine, local tumor irradiation, and adjuvant anti-CD134 mAb can mediate regression of advanced intracranial tumors. Due to their potency as tumor vaccines, and their unique physical attributes, consisting of multinucleated cells arising from partners with disparate genetic programs, we compared the gene expression profile of DC-melanoma fusion cells with DC-DC fusion cells and tumor-tumor fusion cells. Human DC were prepared by culturing PBMC obtained by leukapheresis for 7 days in medium supplemented with GM-CSF and IL-4 with addition of PGE-2 and TNF-α for the final 48 hrs. The human melanoma cell line 888 was irradiated (50Gy) and labeled with CFSE prior to use. DC and tumor cells were mixed at a 1:1 ratio at 2 × 10⁷ cells/ml in electrofusion buffer [5% dextrose, 0.5 mM Mg(CH₃COO)₂, 0.1 mM Ca(CH₃COO)₂ pH 7.2]. Cells were subjected to dielectrophoresis using an ac current of 140V/cm × 10s to align the cells in a chain-like configuration between the electrodes, then pulsed with a 1500 V/cm × 25 μs to induce transient membrane breakdown. Heterokarons were purified through a combination of steps involving differential adherence and magnetic bead separation based on CD80 and CD86 expression resulting in 95–97% purity. Total RNA was isolated from DC-tumor fusion heterokaryons, or DC-DC fusions, or tumor-tumor fusion for 3 independent experiments and was analyzed by Affymetrix gene array. Bioinformatics analysis methods, including a relational database, were developed to identify and functionally classify genes with a robust change in expression. The criteria applied were:

1. all 9 pairwise comparisons had to exhibit a Log ratio >1 increase or decrease,

2. statistical significance had to be at least p<0.01, and

3. signals had to surpass a threshold expression of 500 units.

Hierarchical cluster analysis using the entire set of transcripts (n=22,277) separated the samples into 3 distinct clusters with the DC-tumor cluster being closed to the tumor-tumor cells than the DC-DC cells. The DC-tumor fusion cells also displayed a higher Spearman rank correlation coefficient in pair-wise comparisons with the tumor-tumor samples. A relatively small number of genes (n=137) were expressed in the DC-tumor fusion cells but not in the DC-DC control. Among these were several well-characterized melanoma tumor antigens. Similarly, a small number of genes (n=116), primarily involved in antigen processing and immune response, were contributed to the DC-tumor fusion cell by the DC partner. These data support the hypothesis that ongoing expression of tumor antigens and immune function genes continue to be expressed after heterokaryons formation. More interestingly, we identified 83 DC genes that were specifically downregulated in DC-tumor fusion cells. Moreover, 28 genes were uniquely expressed at a significant level in the DC-tumor fusion cells relative to either the DC or tumor whereas only 1 gene was significantly repressed. Uniquely expressed genes are involved in transcriptional regulation, development, growth factor receptors, signal transduction, cell adhesion, and cytoskeletal organization and in physiologic membrane fusion processes. This suggests that transcription in fusion heterokaryons can be cross-regulated by transcription factors uniquely expressed in each of the component cells.