The tumor-associated antigen disialoganglioside-GD2 is expressed on neuroblastoma and melanoma and is an established target for passive immunotherapy. The aim of this study was to develop an active immunization strategy leading to the induction of a humoral anti-GD2 immune response. However, carbohydrates and glycolipids are T cell-independent antigens (TI) and usually evoke a poor immune response in tumor-bearing hosts.Here, we describe the identification, characterization and in vivo efficacy of cyclic peptides mimicking the structure of glycolipid GD2, i.e. GD2 mimotopes, in order to overcome T-cell independency. First, GD2 peptide mimotopes were identified by biopanning experiments of a phage-display library displaying circular decapeptides against the human/mouse chimeric anti-GD2 antibody (Ab) ch14.18. Thirteen independent phage clones were isolated which bind to ch14.18 with high specificity. Competitive binding of phages expressing GD2 peptide mimotopes to ch14.18 antibody revealed two superior peptide candidates, mimotope A (MA) and and mimotope D (MD), which were subjected to further evaluation. Second, two plasmid DNA minigene vaccines were generated by overlapping PCR encoding for MA and MD, respectively. The plasmids were based on pSecTag2-A also including a kappa leader sequence, a T-cell helper epitope from HIV-1 gp 120 (T1) and a myc-tag. Minigene expression was demonstrated following transfection of COS-7 cells in western-blots and GD2 mimikry was determined in solid phase ELISA experiments. Third, the efficacy of these mimotope DNA vaccines to induce a tumor protective anti-GD2 immune response was tested in the syngeneic NXS2 model of neuroblastoma expressing ganglioside GD2. The DNA vaccination was accomplished with attenuated Salmonella typhimurium (SL 7207) used as an oral vaccine carrier. Only mice receiving the mimotope DNA vaccines revealed a decrease in primary tumor growth by 50% and a dramatic reduction of spontaneous liver metastases with a mean liver weight of 1g in both groups (MA and MD) in contrast to negative controls (3g). Interestingly, mice immunized with KLH conjugated peptide mimotopes A and D revealed an increased rate of s.c. tumor growth and spontaneous liver metastasis with average liver weights of 5 (MA) and 7 (MD), respecively, suggesting the induction of tolerance using this peptide vaccine approach. Finally the highest anti-GD2 humoral immune response was observed in sera of mice from both GD2 mimotope DNA vaccine groups, consistent with the anti-tumor reponse observed in vivo. Based on these data, we belive that GD2 mimotope DNA vaccines may provide a useful strategy for active immunization against neuroblastoma.

Effective chemotherapy in neuroblastoma is limited by poor anti-tumor efficacy, systemic toxicity and the induction of drug resistance. Here, we provide further evidence that a hydrolytic activated prodrug design may overcome these problems. For this purpose, VP-16 was functionally blocked by a carbonate linker to generate two novel chemically stable prodrugs of VP-16, ProVP-16 I and II. We demonstrate profoundly different biological effects in vitro and in vivo of the prodrugs compared to parental VP-16. First, we established an up to >2 log higher in vitro toxicity of the two prodrugs compared to VP-16 on a panel of neuroblastoma cell lines. The highest increase of prodrug mediated cytotoxicity was observed in multi drug resistant cell lines. Second, in vivo studies showed a maximum tolerated dose (MTD) of ProVP-16 II (60 mg/kg), which was at least threefold higher than that of VP-16 (20 mg/kg). Tests of ProVP-16 II in a syngeneic NXS2 neuroblastoma model indicated that mice treated with this prodrug at 1/3 of the MTD was as effective as VP-16 parental compound used at the MTD in suppression of tumor growth. In summary, the etoposide prodrugs proved effective and less toxic and are therefore highly promising new anti-neuroblastoma compounds.

The induction of T-cell mediated immunity against neuroblastoma is a challenge due to poor immunogenicity of this malignancy. Here, we demonstrate the induction of protective immunity in a syngeneic murine neuroblastoma model following vaccination with minigenes comprising of three novel natural MHC class I ligands. First, after immunoprecipitation of MHC class I from NXS2 cells, peptides were eluted and examined in tandem-MS analysis which lead to the identification of three novel natural MHC class I peptide ligands, TEALPVKLI from ribonucleotide reductase M2, NEYIMSLI from Ser/Thr protein phosphatase 2A and FEMVSTLI with unknown origin. Second, we constructed two different minigenes, one encoding for the three novel epitopes and the second for three known mTH derived epitopes with high predicted binding affinity to MHC class I by cloning them into the mammalian expression vector pCMV-3FUB. This lead to constructs with an ubiquitin-tag upstream the inserted epitopes in order to facilitate proteasomal degradation. Furthermore the epitopes were separated by a spacer peptide (AAY), which proved to be a preferential proteasome cleavage site. Third, we demonstrate the induction of protective immunity against neuroblastoma using an attenuated strain of Salmonella typhimurium as a carrier harboring pCMV 3FUb vectors encoding for the two minigenes. These findings establish proof of concept that disruption of self tolerance against neuroblastoma associated epitopes may be an effective adjuvant therapeutic strategy.