NGS based identification of GD2-positive tumor-specific phenotype for cancer diagnostics and therapy

The tumor-associated glycosphingolipid ganglioside GD2 presents an attractive target for cancer immunotherapy. This molecule is abundant on different types of cancer cells and is characterized by restricted expression on healthy cells. Tumors exhibit heterogeneity in the expression level of GD2 and, therefore, development of methods for determination of the GD2-positive tumor phenotype for the pertinent application of targeted therapies is required. In this work, we have developed a gene expression-based classifier for the prediction of the GD2-positive tumor phenotype. We analyze RNA-seq data from GD2-positive and GD2-negative cell lines of different tumor types as well as neuroblastoma biopsy material for gene expression levels of enzymes that participate in ganglioside biosynthesis and determine the role of gene expression of these enzymes in the formation of the GD2-positive tumor phenotype. We also apply gene expression patterns known from literature to our data and use large public RNA-seq datasets to identify novel gene expression patterns associated with GD2 expression to validate the findings in our data. The results of the study may be used for the development of a gene signature which separates GD2-positive from GD2-negative tumors and for prediction of the GD2 phenotype in clinical specimens from cancer patients, and thus be used as a companion diagnostic for anti-GD2 therapy. 7 human tumor cell lines were tested for the expression of the glycosphingolipid ganglioside GD2. These cell lines were stored in RNAlater prior to RNA extraction. 3 tumor biopsies were stored as Formalin-Fixed Paraffin-Embedded (FFPE) blocks prior to RNA extraction. Following ribosomal RNA depletion and RNA libraries construction transcription profiles of the samples were obtained using Illumina HiSeq 3000.