To study mechanisms of T cell-mediated rejection of B cell lymphomas, we developed a murine lymphoma model wherein three potential rejection antigens, human c-MYC, chicken ovalbumin (OVA), and GFP are expressed. After transfer into wild-type mice 60–70% of systemically growing lymphomas expressing all three antigens were rejected; lymphomas expressing only human c-MYC protein were not rejected. OVA expressing lymphomas were infiltrated by T cells, showed MHC class I and II upregulation, and lost antigen expression, indicating immune escape. In contrast to wild-type recipients, 80–100% of STAT1-, IFN-γ-, or IFN-γ receptor-deficient recipients died of lymphoma, indicating that host IFN-γ signaling is critical for rejection. Lymphomas arising in IFN-γ- and IFN-γ-receptor-deficient mice had invariably lost antigen expression, suggesting that poor overall survival of these recipients was due to inefficient elimination of antigen-negative lymphoma variants. Antigen-dependent eradication of lymphoma cells in wild-type animals was dependent on cross-presentation of antigen by cells of the tumor stroma. These findings provide first evidence for an important role of the tumor stroma in T cell-mediated control of hematologic neoplasias and highlight the importance of incorporating stroma-targeting strategies into future immunotherapeutic approaches.

Over-expression of the proto-oncogene c-MYC is frequently observed in a variety of tumors and is a hallmark of Burkitt´s lymphoma. The fact that many tumors are oncogene-addicted to c-MYC, renders c-MYC a powerful target for anti-tumor therapy. Using a xenogenic vaccination strategy by immunizing C57BL/6 mice with human c-MYC protein or non-homologous peptides, we show that the human c-MYC protein, despite its high homology between mouse and man, contains several immunogenic epitopes presented in the context of murine H2b haplotype. We identified an MHC class II-restricted CD4+ T-cell epitope and therein an MHC class I-restricted CD8+ T-cell epitope (SSPQGSPEPL) that, after prime/boost immunization, protected up to 25% of mice against a lethal lymphoma challenge. Lymphoma-rejecting animals contained MHC multimer-binding CD8+ cell within the peripheral blood and displayed in vivo cytolytic activity with specificity for SSPQGSPEPL. Taken together these data suggest that oncogenic c-MYC can be targeted with specific T-cells.

BackgroundA given tumor is usually dependent on the oncogene that is activated in the respective tumor entity. This phenomenon called oncogene addiction provides the rationale for attempts to target oncogene products in a therapeutic manner, be it by small molecules, by small interfering RNAs (siRNA) or by antigen-specific T cells. As the proto-oncogene product is required also for the function of normal cells, this raises the question whether there is a therapeutic window between the adverse effects of specific inhibitors or T cells to normal tissue that may limit their application, and their beneficial tumor-specific therapeutic action. To address this crucial question, suitable mouse strains need to be developed, that enable expression of the human proto-oncogene not only in tumor but also in normal cells. The aim of this work is to provide such a mouse strain for the human proto-oncogene product c-MYC.Principal FindingsWe generated C57BL/6-derived embryonic stem cells that are transgenic for a humanized c-Myc gene and established a mouse strain (hc-Myc) that expresses human c-MYC instead of the murine ortholog. These transgenic animals harbor the humanized c-Myc gene integrated into the endogenous murine c-Myc locus. Despite the lack of the endogenous murine c-Myc gene, homozygous mice show a normal phenotype indicating that human c-MYC can replace its murine ortholog.ConclusionsThe newly established hc-Myc mouse strain provides a model system to study in detail the adverse effects of therapies that target the human c-MYC protein. To mimic the clinical situation, hc-Myc mice may be cross-bred to mice that develop tumors due to overexpression of human c-MYC. With these double transgenic mice it will be possible to study simultaneously the therapeutic efficiency and adverse side effects of MYC-specific therapies in the same mouse.

Adoptive T cell therapy is an important additional treatment option for malignant diseases resistant to chemotherapy. Using a murine high-grade B cell lymphoma model, we have addressed the question whether the B cell differentiation antigen CD19 can act as rejection antigen. CD19 2/2 mice inoculated with CD19 1 B cell lymphoma cells showed higher survival rates than WT mice and were protected against additional tumor challenge. T cell depletion prior to tumor transfer completely abolished the protective response. By heterotypic vaccination of CD19 2/2 mice against murine CD19, survival after tumor challenge was significantly increased. To define protective epitopes within the CD19 molecule, T cells collected from mice that had survived the tumor transfer were analyzed for IFNc secretion in response to CD19-derived peptides. The majority of mice exhibited a CD4 1 T cell response to CD19 peptide 27, which was the most dominant epitope after CD19 vaccination. A peptide 27-specific CD4 1 T cell line protected CD19 2/2 mice against challenge with CD19 1 lymphoma and also cured a significant proportion of WT mice from recurrent disease in a model of minimal residual disease after chemotherapy. In conclusion, our data highlight CD19-specific CD4 1 T cells for adoptive T cell therapy of B cell lymphomas.

4091 Poster Board III-1026 The rejection of Non Hodgkin Lymphomas expressing foreign, for example viral, antigens is compromised despite the presence of specific T-cells. The underlying immunosuppressive mechanisms are poorly understood. Using a transgenic mouse lymphoma model, where the proto-oncogene c-myc is driven by parts of the immunoglobulin lambda locus representing a t(8;22) translocation as found in Burkitt's lymphoma, we investigated the anti-lymphoma activity of specific T-cells. By retroviral transduction of a specific foreign antigen (chicken Ovalbumin-IRES-GFP, OVA) into cell lines from primary c-myc transgenic lymphomas we established a syngeneic model that would allow us to address the role of interferon gamma signalling in rejection of high grade lymphomas. All primary lymphoma cells displayed normal MHC class I and II levels on the surface when compared to wildtype splenic B-cells. This expression could be enhanced by treatment with interferon gamma (IFNg,100U/ml) up to 10 fold. When retrovirally OVA-transduced lymphoma cells were injected into either wildtype or GFP transgenic recipients, animals displayed a significant delay in lymphoma growth compared to non transduced or IRES-GFP transduced control cell lines. 80% of the recipient mice rejected OVA expressing lymphomas. By contrast, we observed 100% mortality when GFP expressing control lymphomas were injected in GFP transgenic recipients, which are tolerant for GFP. Developing OVA expressing lymphomas (20%) displayed a loss of GFP expression indicating a selection for antigen negative cells (p=0.001). In spleens from mice rejecting OVA-expressing lymphomas we found up to 1.8% SIINFEKL specific T-cells. To gain more mechanistic insights, we transferred OVA expressing lymphoma cells into IFNg, Stat1-/- or IFNg-/-receptor deficient recipients. Lack of STAT1-/- or IFNg-receptor on the recipient side or inability to secrete IFNg was associated with fast lymphoma progression and growth was not different when compared to non transduced, antigen negative cell lines. When IFNg-receptor or STAT1 deficient OVA expressing cell lines were transferred into wildytpe mice, rejection was not influenced. Outgrowing OVA expressing lymphomas in wildtype mice displayed a high MHC class I and II expression compared to the cell line prior to injection. MHC induction was absent in lymphomas transferred to Stat1 or IFNg deficient recipients. Depletion of NK cells by anti AsialoGM antibody in wildtype recipients resulted in a significant reduction of disease free survival (80% vs. 50%, p=0.002) and animals developed larger tumors which were eventually rejected resulting in a comparable overall survival. In peripheral blood of NK depleted mice significantly more OVA specific T-cells were detectable through pentamer staining. When lymphoma cell lines were injected into Rag1-/- mice, NK cell mediated rejection was also significantly impaired upon depletion. Our results suggest that T-cell mediated rejection of high grade B-cell lymphomas is strongly dependent on host IFNg secretion and that NK cells substantially contribute to T-cell mediated rejection. Disclosures: No relevant conflicts of interest to declare.

3927 The incidence of high-grade B cell lymphomas has been increasing over the last decades in western countries for unclear reasons. Relapse after conventional chemotherapy especially in high-grade B cell lymphomas remains a very difficult clinical issue. Contrary to CML and AML, the benefit of allogeneic SCT for treatment of high-grade lymphomas is not well established. Several studies suggested a potential graft versus lymphoma (GvL) effect for acute lymphoblastic leukemia (ALL) and several types of non-Hodgkin lymphomas. To study mechanisms involved in T cell-mediated rejection of B cell lymphomas, we have developed a murine lymphoma model in which three antigens, human c-MYC protein, chicken ovalbumin (OVA) and GFP, serve as foreign antigens for rejection. Lymphomas expressing all three antigens were rejected in 60 to 70% of animals after transfer into wild type mice, whereas lymphomas expressing only human c-MYC protein were not rejected. Outgrowing OVA-expressing lymphomas were infiltrated by T cells, showed MHC class I and II upregulation and loss of antigen expression, indicating immune escape. In contrast to wild type recipients of OVA-expressing lymphomas, 80 to 100% of recipient STAT1-, IFN-γ-, or IFN-γ receptor-deficient mice died due to lymphoma growth. Remarkably, lymphomas arising in IFN-γ- and IFN-γ-receptor-deficient mice also invariably showed lost antigen expression. Thus, poor overall survival of IFN-γ- and IFN-γ-receptor-deficient recipient mice is not due to a lack of antigen-specific T cell killing but due to inefficient eradication of antigen-negative variants of the lymphoma. In order to address the role of the stroma in eradication of lymphoma cells we made use of B6bm1 animals that do not present the immunodominant OVA derived peptide SIINFEKL in the context of MHC class I. Since the wildtype MHC represents an allo-antigen in B6bm1 mice, B6bm1 and B6 wildtype control recipients were T-cell depleted by 30H12 anti CD90.2 antibody prior to transfer of lymphoma cells. Anti OVA immunity was restored by adoptive transfer of 1 Mio. primed CD90.1+ OT-I-T-cells one day after lymphoma transfer. T-cell depletion was continued for 28 days biweekly. Lymphoma growth was faster in bm1 recipients and disease free survival significantly reduced (A). In addition, T-cell expansion was significantly reduced (B) in bm1 recipients as analyzed by pentamer staining of OT-I-T-cells in peripheral blood (day 21 0.84%±0.2 vs. 3.53%±0.2 of lymphocytes, p=0.001) indicating an important role of stromal crosspresentation for the rejection of lymphoma cells. Our data show that mechanisms established for solid tumors hold true also for hematologic neoplasias such as B cell lymphomas. Antigen-dependent eradication of tumor antigen-loss variants makes antigen-specific T cell therapy particularly attractive as a novel therapeutic treatment option. Disclosures: No relevant conflicts of interest to declare.

5120 Overexpression of the proto-oncogene c-myc due to chromosomal translocation is the hallmark of Burkitt-lymphoma. The evolving high grade lymphoma is dependent on the overexpression of c-myc, which provides the necessary signal to drive uncontrolled proliferation. Therefore loss of function or recognition of c-myc overexpressing cells by c-MYC specific T-cells should result in killing of the target and a halt to lymphoma progression. C-myc is also expressed in a variety of other human malignancies. Peptide prediction reveals several potential foreign epitopes in the context of murine H2b due to 87% homology between human and mouse c-MYC. In this study we explored whether the human c-myc gene product can be a target for T-cell therapy. Wildtype C57BL/6 mice were immunized with recombinant human c-MYC protein in combination with incomplete Freund′s adjuvans and CpG, and were boosted at various time points thereafter using either c-MYC protein or 40mer peptides encompassing the non homologous regions. Control animals were vaccinated with recombinant GFP or OVA protein. C-MYC vaccinated animals displayed a higher IFNg release upon re-stimulation with c-MYC pulsed dendritic cells compared to control vaccinated animals. In ELISPOT assays we observed a higher number of IFNg positive cells (299±17 vs. 122±8.5 (GFP vaccinated) vs. 66±8.5 (OVA vaccinated)). Vaccination using single peptides revealed that peptides spanning the region from amino acid 87-123, 216-255 and 334-376 produced similar results. In addition, using a human c-MYC specific ELISA we were able to detect c-MYC specific antibodies in serum from immunized mice in a concentration up to 40mg/l. Using established cell lines from l-hu-c-myc transgenic mice, where the human c-myc gene is overexpressed due to the juxtaposition of elements of the immunoglobuline lambda locus as found in t(8;22) of Burkitt's lymphoma, we investigated whether vaccination with human c-MYC protein would influence lymphoma growth in a lymphoma transfer model. Animals were s.c. challenged with 0.1 Mio 291cells overexpressing human c-MYC and were monitored for lymphoma growth. C-MYC vaccinated animals (n=15) displayed a delay in tumor onset and a significantly better disease free survival (28 vs. 22 days, p=0.012) compared to control (OVA) vaccinated animals (n=10). This delayed growth was associated with an increased number of infiltrating CD3+/Perforin+ cells. However, all mice eventually succumbed to lymphoma growth, indicating that the T-cell response was not sufficient to control lymphoma growth in the long term. From these data we conclude that the human c-MYC is a possible target antigen for T-cells, but responses are weak and presumably low in frequency. Disclosures No relevant conflicts of interest to declare.