Search

Direct analysis of HLA-presented antigens by mass spectrometry provides a comprehensive view on the antigenic landscape of different tissues/malignancies and enables the identification of novel, pathophysiologically relevant T-cell epitopes. Here, we present a systematic and comparative study of the HLA class I and II presented, nonmutant antigenome of multiple myeloma (MM). Quantification of HLA surface expression revealed elevated HLA molecule counts on malignant plasma cells compared with normal B cells, excluding relevant HLA downregulation in MM. Analyzing the presentation of established myeloma-associated T-cell antigens on the HLA ligandome level, we found a substantial proportion of antigens to be only infrequently presented on primary myelomas or to display suboptimal degrees of myeloma specificity. However, unsupervised analysis of our extensive HLA ligand data set delineated a panel of 58 highly specific myeloma-associated antigens (including multiple myeloma SET domain containing protein) which are characterized by frequent and exclusive presentation on myeloma samples. Functional characterization of these target antigens revealed peptide-specific, preexisting CD8(+) T-cell responses exclusively in myeloma patients, which is indicative of pathophysiological relevance. Furthermore, in vitro priming experiments revealed that peptide-specific T-cell responses can be induced in response-naive myeloma patients. Together, our results serve to guide antigen selection for T-cell-based immunotherapy of MM.

The breakthrough development of clinically effective immune checkpoint inhibitors illustrates the potential of T-cell–based immunotherapy to effectively treat malignancies. A remaining challenge is to increase and guide the specificities of anticancer immune responses, e.g., by therapeutic vaccination or by adoptive T-cell transfer. By analyzing the landscape of naturally presented HLA class I and II ligands of primary chronic lymphocytic leukemia (CLL), we delineated a novel category of tumor-associated T-cell antigens based on their exclusive and frequent representation in the HLA ligandome of leukemic cells. These antigens were validated across different stages and mutational subtypes of CLL and found to be robustly represented in HLA ligandomes of patients undergoing standard chemo-/immunotherapy. We demonstrate specific immune recognition of these antigens exclusively in CLL patients, with the frequencies of representation in CLL ligandomes correlating with the frequencies of immune recognition by patient T cells. Moreover, retrospective survival analysis revealed survival benefits for patients displaying immune responses to these antigens. These results directly imply these nonmutant self-peptides as pathophysiologically relevant tumor antigens and encourages their implementation for cancer immunotherapy.

Identification of physiologically relevant peptide vaccine targets calls for the direct analysis of the entirety of naturally presented human leukocyte antigen (HLA) ligands, termed the HLA ligandome. In this study, we implemented this direct approach using immunoprecipitation and mass spectrometry to define acute myeloid leukemia (AML)-associated peptide vaccine targets. Mapping the HLA class I ligandomes of 15 AML patients and 35 healthy controls, more than 25 000 different naturally presented HLA ligands were identified. Target prioritization based on AML exclusivity and high presentation frequency in the AML cohort identified a panel of 132 LiTAAs (ligandome-derived tumor-associated antigens), and 341 corresponding HLA ligands (LiTAPs (ligandome-derived tumor-associated peptides)) represented subset independently in20% of AML patients. Functional characterization of LiTAPs by interferon-γ ELISPOT (Enzyme-Linked ImmunoSpot) and intracellular cytokine staining confirmed AML-specific CD8(+) T-cell recognition. Of note, our platform identified HLA ligands representing several established AML-associated antigens (e.g. NPM1, MAGED1, PRTN3, MPO, WT1), but found 80% of them to be also represented in healthy control samples. Mapping of HLA class II ligandomes provided additional CD4(+) T-cell epitopes and potentially synergistic embedded HLA ligands, allowing for complementation of a multipeptide vaccine for the immunotherapy of AML.

Background Immunotherapy for CLL with new antibodies or T-cells with modified TCR relies on attractive targets. ROR1 is such a promising target since it is highly overexpressed in CLL. Chimeric antigen receptor engineered T cells and antibodies directed against the extracellular part of ROR1 have already been developed and tested in vitro or in animal models, but still there is no MHC-class I presented peptide serving as target structure for CD8+ T cells (with or without a genetically modified T cell receptor) available. Aim The aim of this study was (1) to identify an immunogenic MHC-class I presented ROR1 peptide, (2) to generate respective ROR1 peptide specific CD8+ T cell clones, and (3) to analyze the nucleotide sequence of the CDR3 region of the expressed alpha and beta T cell receptor chain. Results In mass spectrometric-based analyses of the HLA-ligandome a HLA-B*07 presented ROR1 peptide was identified in primary CLL cells of two patients. Six T cell clones specific for this particular ROR1-peptide were generated from single CD8+ T cells from 2 healthy individuals with 3 T cell clones generated from each donor. Functionality and specificity of those T cell clones were tested in cytotoxicity assays. All 6 dextramer+ CD8+ T cell clones lysed peptide loaded and HLA-B*07+ transduced K562 cells (kindly provided by Lorenz Jahn, [Jahn et al., Blood, 2015 Feb 5;125(6):949-58]). Two selected clones (XD8 and XB6) were tested for their cytotoxic potential against 2 ROR1+ HLA-B*07+ tumor cell lines (with the ROR1 peptide identified by mass spectrometry for both of them) and against 2 primary CLL cell samples. Tested clones showed a significant lysis of the respective target cells. CDR3 regions of the alpha and beta T cell receptor chain were sequenced on a single cell level. The CDR3 alpha region from each of the 3 ROR1 specific T cell clones from donor A showed some similarities to T cell clones derived from donor B (Table 1). Conclusion For the first time a MHC-class I presented ROR1 peptide antigen is reported. ROR1 positive CLL cells can be targeted by specific HLA-B*07 restricted CTLs. Respective CD8+ T cell clones with anti-leukemic activity from 2 donors share some amino acid motifs of the CDR3 alpha and beta regions. In conclusion, this information provides the possibility of generating ROR1 specific CD8+ T cells with genetically modified T cell receptors for immunotherapy and for tracking those cells after administration with next generation sequencing in peripheral blood samples of patients. Furthermore, data suggest the existence of public TCR motifs in leukemia antigen specific CTLs, which needs to be proven in follow-up experiments with larger cohorts of donors and patients. Finally, the presented strategy to identify leukemia specific peptide antigens for CD8+ T cells might be an attractive method for similar projects. Table 1 Amino acid sequences of CDR3 alpha and beta regions of the TCR of ROR1 specific CD8+ T cell clones. When comparing two clones, matching amino acids are depicted in red. The aromatic amino acids phenylalanine (F) and tyrosine (Y) are shown in blue when situated at the same position. Gaps inserted during the sequence alignment process are indicated by a hyphen '-'. Table 1. Amino acid sequences of CDR3 alpha and beta regions of the TCR of ROR1 specific CD8+ T cell clones. When comparing two clones, matching amino acids are depicted in red. The aromatic amino acids phenylalanine (F) and tyrosine (Y) are shown in blue when situated at the same position. Gaps inserted during the sequence alignment process are indicated by a hyphen '-'. Disclosures Middeke: Sanofi: Honoraria. Schetelig:Sanofi: Honoraria.

Several studies have proven the positive immunomodulatory effect of lenalidomide, a second-generation derivative of thalidomide, on T-cell responses in CLL, in particular regarding the repair of the immune synapse between CLL and T cells (e.g., Ramsay et al. J Clin Invest 2008), as well as the upregulation of costimulatory molecules on CLL cells (e.g., Aue et al. Haematologica 2009). Therefore, lenalidomide seems to be a promising combination partner for T-cell based immunotherapy approaches in CLL patients. We recently conducted a study which directly characterized the antigenic landscape of CLL by mass spectrometric analysis of naturally presented HLA ligands and identified a panel of CLL-specific CD4+ as well as CD8+ T-cell epitopes as suitable targets for T-cell based immunotherapy (Kowalewski et al. PNAS2015). As anti-cancer drugs can have marked effects on the HLA ligandome of tumor cells, it is of great importance to thoroughly characterize and take into account the effects of the immune modulator lenalidomide not only on the effector cells, but also on the antigenic landscape of the target cells. Here we present a mass spectrometry-based study which longitudinally maps the HLA-presented immunopeptidome and in particular the CLL-associated antigens of primary CLL cells under in vitro lenalidomide treatment. We quantified HLA surface expression on primary CLL cells and autologous B cells (n=4) at t0, t24h and t48h after incubation with 0.5 µM lenalidomide. With regard to HLA class I expression on CD19+CD5+ CLL cells, no significant impact of lenalidomide was observed (fold-change 0.92-1.02, t48h), whereas a slight but also not significant increase of HLA class II molecules after treatment was detectable (fold-change 1.25-1.43, t48h) with absolute molecule counts ranging from 40,000-125,000 class I and 30,000-200,000 class II molecules/cell. Furthermore, we could show that CD19+CD5+ CLL cells (n=4) showed a significantly increased expression of HLA class I molecules compared to normal CD19+CD5-B cells. Implementing label-free quantification, we then assessed HLA class I ligand presentation during in vitro lenalidomide incubation. We observed a higher plasticity of the HLA ligandome over time compared to treatment with lenalidomide. 2.5±3.0% of HLA class I ligands showed significant modulation (fold change ≥4, p≤0.01) after 24h of mock treatment, whereas only 0.9±1.2% of the ligands were modulated upon lenalidomide treatment. At t48hsimilar proportions of modulation were observed with 4.0±1.7% of HLA ligands significantly altered in their abundance over time (mock control), while lenalidomide treatment only resulted in 0.9±1.2% modulated ligands. Out of the 6,991 different HLA class I ligands representing 3,983 source proteins identified on primary CLL cells (n=3) by mass spectrometry, we were able to detect 36 (30.5%) of the HLA-matched CLL-associated epitopes described in previous studies on these three samples. Importantly, these CLL-associated antigens showed robust presentation on CLL cells under in vitro lenalidomide treatment. In order to investigate the impact of lenalidomide-treatment on HLA class I allotype distribution within the ligandome, we analyzed the overall frequencies of HLA allotype restrictions of ligands on treated versus untreated primary CLL cells. Thereby, no changes in the HLA allotype distribution were observed. Analysis of the HLA class II ligandome of primary CLL cells (n=3) revealed a total of 12,026 unique peptides representing 2,333 source proteins. We were able to detect 133 (28.5%) of CLL-associated class II epitopes described in previous studies. Notably, most of the CLL-associated antigens showed robust presentation under lenalidomide therapy. Exemplarily for one patient, we identified 7/66 (10.6%) CLL-associated antigens modulated after 24 h of in vitro lenalidomide treatment, but 10/66 (15.2%) antigens altered over time (mock control). Taken together our study provides direct insights into the impact of lenalidomide on the HLA ligandome of primary CLL cells. We were able to show that in vitro lenalidomide has no relevant influence on, and rather seems to stabilize the HLA-presented immunopeptidome of primary CLL cells. Importantly, CLL-associated epitopes are stably presented under lenalidomide treatment. Therefore, lenalidomide appears to be an ideal combination partner for T-cell based immunotherapy in CLL patients. Disclosures Kowalewski: Immatics Biotechnologies GmbH: Employment. Schuster:Immatics Biotechnologies GmbH: Employment.

Several studies demonstrated that peptide-based cancer immunotherapy can induce specific immune responses and affect clinical outcome in a variety of different cancer entities. We recently conducted a study, which directly characterized the antigenic landscape of acute myeloid leukemia (AML) by mass spectrometric analysis of naturally presented HLA ligands and identified a panel of AML-specific CD4+ as well as CD8+ T-cell epitopes as suitable targets for T-cell based immunotherapy (Berlin et al. Leukemia 2015). One main reason for the high relapse rates in AML patients after standard polychemotherapy and allogeneic stem cell transplantation is the presence of minimal residual disease (MRD), which is associated with the persistence of leukemic stem cells (LSCs) in the bone marrow of patients. For clinically effective immunotherapy it is therefore indispensable to target the highly chemotherapy resistant LSCs. Here we present a mass spectrometry-based study, which for the first time analyzes the naturally presented HLA ligandome of stem cell enriched (LSCenr) fractions of primary AML samples to identify novel LSC-associated antigens using the approach of direct peptide isolation and identification. The enrichment of LSCs was performed using fluorescence-activated cell sorting of the originally described phenotype of lineage-negative CD34+CD38- cells of PBMCs from eight AML patients. The original stem cell containing population of 1-3% within the PBMCs of most patients was enriched to >90% purity with cell counts of 20-200x106 for the LSCenr fraction per sample. Consistent with our own previous results, all samples showed comparable expression levels of HLA class I molecules on primary leukemia blasts as well as for the LSCenr fractions, with HLA class I molecule counts ranging from 145,000 to 175,000 molecules/cell for the LSCenr fractions. To specifically identify leukemia-associated antigens on LSCenr cells, the HLA ligandome results obtained from the sorted LSCenrfractions were combined with data acquired from AML blasts of 20 AML patients (HLA class I n=19, HLA class II n=20) in previous studies as well as our normal tissue database that comprises 153 HLA class I and 82 HLA class II ligandomes of various healthy tissues (e.g. blood, bone marrow, spleen, kidney, liver, brain, skin, ovary, bowl). We identified more than 14,600 different naturally presented HLA class I ligands representing ̴6,500 source proteins on LSCenr fractions of primary AML samples (n=8) and their autologous blast cells by mass spectrometry. Overlap analysis of the HLA class I ligandomes of LSCenr fractions and autologous AML blasts with the benign peptidome revealed 45.4% (3,132/6,896) and 40.2% (4,922/12,244) of the LSCenr fraction and the autologous AML blast ligandomes to be represented in the benign-associated HLA ligandome, respectively. 79.1% (5,458/6,896) of the mapped LSCenr fraction ligandome was also presented on autologous AML blasts. 1,029 (14.9%) of these identified HLA class I ligands were presented exclusively on LSCenr fractions and not found on autologous AML blasts, previously analyzed AML blasts or any benign tissue. Furthermore, we were able to identify more than 8,000 different naturally presented HLA class II ligands representing ̴1,700 source proteins. Overlap of the HLA class II ligandomes revealed 45.0% (2,800/4,624) and 39.9% (2,706/6,790) of the LSCenr fraction and autologous AML blast ligandomes to be represented in the benign-associated HLA ligands, respectively. The HLA ligandomes of the LSCenr fraction and the autologous AML blasts showed an overlap of 69.7% (3,224/4,624). 941 (11.5%) HLA class II ligands showed exclusive representation in the LSCenr fraction ligandomes and were never identified on AML blast or benign tissue. These LSC-associated peptides represent highly interesting targets for immunotherapeutic approaches in AML patients and will be further evaluated for their potential to elicit a specific T-cell response. Taken together these preliminary results prove the feasibility of our approach to enrich leukemic progenitor cells of primary AML samples for the successful isolation and identification of HLA presented peptides associated with enriched leukemic progenitor cells. Disclosures Schuster: Immatics Biotechnologies GmbH: Employment. Kowalewski:Immatics Biotechnologies GmbH: Employment.

Background We have recently identified a skewed distribution of class I human leukocyte antigens (HLA) among patients with AML and mutated nucleophosmin gene (NPMc+ AML). A lower frequency of several HLA allelic groups in the NPMc+ AML patient cohort was in good correlation with the results of theoretical predictions for high-affinity immunopeptides derived from NPM1, suggesting that an anti-NPM1 immune response could prevent leukemia development in patients with a suitable HLA class I type. Aim We here present more detailed analysis of HLA-specific features, in a much larger cohort (N = 357) of AML patients with NPM1 mutation (type A/D) from several centers in the Czech Republic, Germany and Poland. Results Most importantly, we confirmed the lower incidence of B´40 and C´07 allelic groups in NPMc+ AML compared to the normal values. The moderate decrease of A´02 allelic group frequency became statistically significant in this larger patient cohort (see Table). On the other hand, no difference in HLA class II frequencies was found compared to the normal distribution. Furthemore, patients with B´07 allelic group had a significantly better prognosis, but only in the absence of Flt-3-ITD mutations (p = 0.049, see Figure). HLA typing was performed by molecular methods in a subgroup of patients (N = 73), allowing for discrimination between C´07 alleles. In this subgroup, C*07:01, but not C*07:02/04 expression was associated with better overall survival (p = 0.036), in agreement with theoretical predictions. Conclusion Our results support the hypothesis that anti-NPM1 immune response reduces AML development and contributes to a better outcome of NPMc+ AML patients with suitable HLA class I type (including at least A´02, B´07, B´40 and C*07:01). Furthermore, NPMc+ immunogenicity is caused rather by its aberrant localization than by the generation of a unique aminoacid sequence. Figure Figure. Disclosures Schetelig: Sanofi: Honoraria. Thiede:AgenDix: Employment, Other: Ownership.

While the discovery of BCR-ABL and the respective tyrosine kinase inhibitors (TKI) resulted in a significant prolongation of patient survival rates, there still is no curative treatment for chronic myeloid leukemia (CML) except for allogeneic stem cell transplantation. The concept of T cell-based immunotherapy is a promising opportunity to eliminate residual leukemic cells, which might promote disease relapse after TKI discontinuation. As effective antigen-specific immunotherapy requires exact knowledge of tumor-associated epitopes that can act as rejection antigens, we have developed a mass spectrometry-based approach, which allows for the direct identification of naturally presented tumor-associated HLA ligands in hematological malignancies. In this study we used this approach to identify HLA class I and II CML-associated peptides as targets for T cell-based immunotherapy. Analysis of HLA class I ligandomes of primary CML cells (n=16) identified 8,291 HLA ligands representing 4,337 source proteins. Comparative ligandome profiling using a benign HLA class I database, which includes various healthy tissues (n=188, 65,949 HLA ligands, 14,030 source proteins) originating from peripheral blood, bone marrow, kidney, lung, liver, colon, spleen and others, revealed 38 CML-exclusive HLA class I ligands with frequencies ≥ 25% of CML patients. Because of the important indirect and direct roles of CD4+ T cells in anti-cancer immune responses, an optimal immunotherapy approach requires the inclusion of HLA class II epitopes. Hence we also analyzed the HLA class II ligandomes of primary CML cells (n=15, 2,822 HLA ligands, 794 source proteins). Comparative ligandome analysis (benign tissue, n=114, 54,149 HLA ligands, 8,584 source proteins) identified 44 CML-associated HLA class II ligands showing CML-exclusive representation in > 25% of the analyzed CML samples. To validate the immunogenicity of our HLA class I and II CML-associated peptides, we performed IFNγ- ELISPOT assays after 12-days of in vitro peptide stimulation. For HLA class II antigens, a panel of 4 peptides was implemented for stimulation of PBMCs obtained from CML patients and healthy volunteers (HV). The ELISPOT assay revealed peptide-specific immune recognition of 4/4 (100%) CML-exclusive peptides in CML patients. The frequencies of the detected immune responses ranged from 17% (4/23 patients) to 4% (1/23 patients) within the tested CML samples. These immune responses were mediated by functional CML patient-derived CD4+ T cells and strictly CML-directed, as no immune response against CML-associated peptides could be detected in HV (0/8). For HLA class I antigens, ELISPOT assays were performed using a panel of 8 peptides. Immune responses were only detected for 1/8 (13%) peptides with a low frequency of 6% (1/18 patients) of tested CML patient samples. A possible explanation for the observed weak immune response to our HLA class I CML-associated peptides compared to the immune responses shown for HLA class II peptides and for HLA class I peptides in other hematological malignancies (e.g. CLL (Kowalewski et. al. PNAS 2015)) might be an inhibition of CD8+ T cell-responses, that reportedly occurs upon TKI treatment of CML patients. To prove this hypothesis in our CML patient cohort (all patients included were under TKI treatment at the time of sample collection), we compared the ELISPOT positive controls (stimulated with a set of 5 Epstein-Barr viral peptides) of all analyzed CML samples with positive controls derived from HV and CLL samples. We could show a highly significant mean spot count reduction (per 100,000 cells) in CML samples (mean 74±16 spots, n=19) compared to HV (mean 241±24 spots, n=42, p To prove the immunogenicity of our HLA class I CML-associated peptides, we performed in vitro artificial antigen-presenting cell (aAPC)-based priming experiments of HV CD8+ cells. For one HLA-A*03-restricted peptide we observed tetramer-positive CD8+ populations with frequencies ranging from 0.12% to 1.41% of viable cells in 2/2 HVs so far. Taken together, these results are a first step towards a successful validation of these newly defined HLA class I and II CML-associated antigens as prime targets for further T cell-based immunotherapy approaches in CML patients. Disclosures Kowalewski: Immatics Biotechnologies GmbH: Employment. Brümmendorf:Ariad: Consultancy, Honoraria; Bristol-Myers Squibb: Consultancy, Honoraria; Pfizer: Consultancy, Honoraria; Novartis: Consultancy, Honoraria, Research Funding; Patent on the use of imatinib and hypusination inhibitors: Patents & Royalties. Niederwieser:Amgen: Speakers Bureau; Novartis Oncology Europe: Research Funding, Speakers Bureau.

Recent studies suggest that multiple myeloma is an immunogenic disease, which might be effectively targeted by antigen-specific T-cell immunotherapy. As standard of care in myeloma includes proteasome inhibitor therapy, it is of great importance to characterize the effects of this treatment on HLA-restricted antigen presentation and implement only robustly presented targets for immunotherapeutic intervention. Here, we present a study that longitudinally and semi-quantitatively maps the effects of the proteasome inhibitor carfilzomib on HLA-restricted antigen presentation. The relative presentation levels of 4780 different HLA ligands were quantified in an in vitro model employing carfilzomib treatment of MM.1S and U266 myeloma cells, which revealed significant modulation of a substantial fraction of the HLA-presented peptidome. Strikingly, we detected selective down-modulation of HLA ligands with aromatic C-terminal anchor amino acids. This particularly manifested as a marked reduction in the presentation of HLA ligands through the HLA allotypes A*23:01 and A*24:02 on MM.1S cells. These findings implicate that carfilzomib mediates a direct, peptide motif-specific inhibitory effect on HLA ligand processing and presentation. As a substantial proportion of HLA allotypes present peptides with aromatic C-termini, our results may have broad implications for the implementation of antigen-specific treatment approaches in patients undergoing carfilzomib treatment.

In recent years, several approaches have been taken in the peptide-based immunotherapy of metastatic renal cell carcinoma (RCC), although little is known about HLA presentation on metastases compared to primary tumor and normal tissue of RCC. In this study we compared primary tumor, normal tissue and metastases with the aim of identifying similarities and differences between these tissues. We performed this comparison for two RCC patients on the level of the HLA ligandome using mass spectrometry and for three patients on the level of the transcriptome using oligonucleotide microarrays. The quantitative results show that primary tumor is more similar to metastasis than to normal tissue, both on the level of HLA ligand presentation and mRNA. We were able to characterize a total of 142 peptides in the qualitative analysis of HLA-presented peptides. Six of them were significantly overpresented on metastasis, among them a peptide derived from CD151; fourteen were overpresented on both primary tumor and metastasis compared to normal tissue, among them an HLA ligand derived from tumor protein p53. Thus, we could demonstrate that peptide-based immunotherapy might affect tumor as well as metastasis of RCC, but not healthy kidney tissue. Furthermore we were able to identify several peptides derived from tumor-associated antigens that are suitable for vaccination of metastatic RCC.

Recent studies underscore that multiple myeloma is an immunogenic disease and suggest that it can be effectively treated by T cell based immunotherapy via immunomodulation. Unspecific immune checkpoint inhibition might be synergistically complemented by therapeutic vaccination, which may help induce and guide specific anti-cancer T cell responses. We have recently conducted a study which directly characterized the antigenic landscape of myeloma by mass spectrometric analysis of naturally presented HLA ligands and identified a panel of T cell epitopes characterized by exquisite myeloma-specificity. As novel immunotherapeutic interventions will have to be implemented as 2nd line treatments, it is of great importance to thoroughly characterize and take into account the effects of previous therapies on the antigenic landscape of target cells. Here we present a longitudinal study of the HLA-presented immunopeptidome of myeloma cells under treatment with the 2nd generation proteasome inhibitor carfilzomib. We characterize and quantify the plasticity of the antigenic landscape and identify targets characterized by robust presentation. We quantified HLA surface expression on 4 human myeloma cell lines (MM.1s, U266, RPMI8226 & JJN3) at t0, t24h and t48h after incubation with carfilzomib. Strikingly, we detected elevated levels of HLA class I post treatment for 3/4 cell lines, with absolute molecule counts ranging from 50,000-400,000 molecules/cell. Out of the >5,000 different HLA ligands we could identify on MM.1s and U266 by mass spectrometry, we were able to detect and longitudinally quantify 38 and 54 myeloma-associated epitopes described in previous studies, respectively. Importantly, the vast majority of these antigens (88.2%) showed robust presentation on myeloma cells under therapy. Around 12% of HLA ligands showed significant changes in abundance after carfilzomib treatment. Strikingly, we observed a significant reduction in aromatic residues at the C-terminal anchor position, which directly reflects the mechanism of action of carfilzomib on the HLA ligand level. Our study provides direct insights into the plasticity of T cell antigen presentation on myeloma cells in a model of proteasome inhibitor therapy. We were able to characterize general changes in the immunopeptidome composition and identify robustly presented myeloma-associated epitopes. Our findings may help guide future clinical trials for 2nd line T cell immunotherapy of multiple myeloma. Citation Format: Daniel J. Kowalewski, Simon Walz, Linus Backert, Heiko Schuster, Oliver Kohlbacher, Lothar Kanz, Helmut R. Salih, Hans-Georg Rammensee, Stefan Stevanovic, Juliane S. Stickel. Mapping the impact of proteasome inhibitor therapy on the antigenic landscape of multiple myeloma: Identifying robust targets for T cell immunotherapy. [abstract]. In: Proceedings of the CRI-CIMT-EATI-AACR Inaugural International Cancer Immunotherapy Conference: Translating Science into Survival; September 16-19, 2015; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2016;4(1 Suppl):Abstract nr A112.

Chronic lymphocytic leukemia (CLL) is the most common leukemia in adults. The course of disease predisposes patients to the development of secondary malignancies, which might be due to a generalized immunosuppression induced by the disease and/or the therapies used for treatment. Skin cancer, including melanoma, is the most common secondary malignancy reported in these patients. In a recent study we reported on spontaneous memory T cell responses against CLL-associated antigens defined by HLA ligandome analysis in CLL patients. Moreover, retrospective analysis suggests a significant survival benefit for CLL patients showing immune reactions to more than one of the tested antigens (Kowalewski et. al., PNAS 2015). Here we analyzed our CLL patient cohort for the incidence of secondary malignancies, and in particular for skin cancers. A total of 15 CLL-associated HLA-peptides (6 A*02, 3 A*03, 6 B*07) were implemented for stimulation of HLA-matched PBMC obtained from CLL patients. 45 CLL patients (Binet A n=15, Binet B, n=15, Binet C n=15, median age, 33.3% (15/45) with prior CLL specific therapy) were analyzed for antigen-specific memory T-cell responses in IFNγ-ELISPOT assays. 13 patients (28.9%, response cohort) showed specific T cell response against > 1 antigen, while 32 patients (71.1%, negative cohort) showed none or only one immune response. The median age was 71 in the response cohort and 73 in the negative cohort. 40.0% (6/15) of CLL patients with Binet stage A and 33.3% (5/15) with stage B were classified as responders whereas only 13.3% (2/15) of patients with stage C showed >1 immune responses. Of the included CLL patients with prior therapy only one patient (8%) showed > 1 immune responses. In the analyzed CLL patient cohort 7 skin cancers (melanoma: n=2, squamous cell carcinoma: n=2, basal cell carcinoma: n=3), 3 precancerous skin lesions and 9 other secondary malignancies including renal cell carcinoma (n=2), colon (n=3), breast (n=2) and prostate cancer (n=2) were observed. In the negative cohort 34.4% (11/32) of the patients developed a secondary malignancy, while only 15.4% (2/13) in the immune response cohort presented with secondary neoplasms. Strikingly all skin cancers and precancerous skin lesions were observed in the non-immune responder cohort. Furthermore multiple secondary malignancies (e.g. squamous cell carcinoma and colon cancer) were only observed in the cohort that showed no immune response. Taken together our data suggest that a favorable immune signature in CLL patients, i.e. the ability to mount T-cell responses against CLL-associated antigens not only is associated with improved overall survival but also might protect these patients from the development of secondary skin cancers. Disclosures No relevant conflicts of interest to declare.

Recent studies underscore that multiple myeloma is an immunogenic disease and suggest that it can be effectively treated by T cell based immunotherapy via immunomodulation. This strategy might be synergistically complemented by therapeutic vaccination, which may help induce and guide specific anti-cancer T cell responses. We have recently conducted a study which directly characterized the antigenic landscape of myeloma by mass spectrometric analysis of naturally presented HLA ligands and identified a panel of T cell epitopes characterized by exquisite myeloma-association (Walz, Stickel et. al., Blood 2015). As standard of care in myeloma includes proteasome inhibitor therapy and the proteasome plays a central role in the generation of MHC-presented peptides, it is of great importance to thoroughly characterize and take into account the effects of this treatment on the antigenic landscape of myeloma cells and implement only robustly presented targets for peptide vaccine design. This is even more important Here we present a mass spectrometry-based study, which longitudinally and semi-quantitatively maps the effects of treatment with the 2nd generation proteasome inhibitor carfilzomib in an in vitro model of multiple myeloma. We observed considerable plasticity of the HLA class I ligandome of MM.1S cells after treatment with carfilzomib with 17.9±1.1.% (mean of 3 biological replicates ± SD) of HLA ligands showing significant modulation (fold change ≥ 4, P ≤ 0.01) at t24h compared to mock-treated controls (down-modulated: 11.5±1.1%, up-modulated: 6.3±0.6%). We were able to longitudinally tracke the abundance of 28 previously defined myeloma antigens, confirming robust (16/28, 57.1%) or even increased presentation (8/28, 28.6%) under treatment for the majority of these peptides. However, - importantly - we observed highly distortive effects of carfilzomib treatment on the HLA allotype distribution of target cells, which manifested as a marked reduction of HLA ligands restricted by HLA-A*23:01 and A*24:02 (-62.5±1.8% and -57.0±0.6%, respectively, at t=24h after treatment). These findings indicate strong allotype-specific effects of carfilzomib on the antigenic landscape of myeloma cells, which we interpret to be a direct reflection of the mechanism of action of this drug. As a significant proportion of the U.S. population are carriers of the affected alleles (A*23:01: 8.2%; A*24:02: 22.6%), these findings could have broad implications for the design or implementation of antigen-specific therapies in patients under proteasome inhibitor treatment. Furthermore, these findings might indicate the possibility of altered cancer immunosurveillance as a consequence of proteasome inhibitor therapy. Disclosures Weisel: Amgen: Consultancy, Honoraria, Other: Travel Support; Onyx: Consultancy, Honoraria; Novartis: Other: Travel Support; Janssen Pharmaceuticals: Consultancy, Honoraria, Other: Travel Support, Research Funding; Noxxon: Consultancy; Celgene: Consultancy, Honoraria, Other: Travel Support, Research Funding; BMS: Consultancy, Honoraria, Other: Travel Support.

Non-Hodgkin lymphomas (NHL) are frequent malignancies with considerable mortality. A recurrent somatic and oncogenic driver mutation in the Toll-like receptor adaptor gene MYD88, Leu265Pro (L265P) has been identified in up to 90% of certain NHL subtypes. Genetic alterations affecting a protein-coding region have the potential to generate mutation-derived peptides that are presented by HLA class I proteins and might be recognized by cytotoxic T cells. Because MYD88L265P is a widely occurring and tumor-specific mutation, we investigated the potential of MYD88L265P -containing peptides for CD8+ T cell mediated immunotherapy as a new therapeutic approach for MYD88L265P+ NHL. Based on in silico prediction we identified potential HLA ligands encompassing the MYD88L265P mutation for several HLA class I allotypes. Functional characterization of the candidate HLA class I MYD88L265P-derived HLA class I ligands with regard to induction of T cell responses identified a set of immunogenic peptides for HLA-B*07 and -B*15. In one MYD88L265P-mutated NHL patient, memory T cell responses targeting three different MYD88L265P-derived HLA class I ligands were detected by IFN-γ ELISPOT. Efficient T cell priming was demonstrated in vitro using naïve T cells of healthy volunteers (HVs). In detail, three HLA-B*07 peptides (P1-3B*07) and one HLA-B*15-restricted peptide (P4B*15) were analyzed using artificial antigen-presenting cell-based (aAPC) in vitro priming experiments in three to six HVs, respectively. For all tested peptides proliferation of peptide-specific CD8+ T cells could be detected after in vitro priming. For the HLA-B*07-restricted ligands, peptide-specific CD8+ T cells could be induced in 6/6 (P1B*15), 1/3 (P2B*07) and 3/4 (P3B*07) HVs, respectively, with a maximum frequency of 14.1% peptide-specific CD8+ T cells. For the HLA-B*15-restricted ligand (P4B*15), peptide-specific CD8+ T cells could be induced in 2/3 HVs with a maximum frequency of 9.5% tetramer-positive CD8+ T cells. The functionality and specificity of peptide-specific CD8+ T cells after aAPC-based in vitro priming was validated by intracellular cytokine staining for IFN-γ and TNF-α as well as for the expression of the degranulation marker CD107a. In 3/3 HVs primed with P1B*07 (RPIPIKYKAM) as well as in 1/2 HVs primed with P4B*15 (HQKRPIPIKY), we detected specific and functional CD8+ T cell populations after stimulation with the mutated peptides, but not after stimulation with the corresponding wild type peptides (P1WT: RLIPIKYKAM, P4WT: HQKRLIPIKY). Furthermore, the peptide-specific cytotoxic activity of specific CD8+ T cells was demonstrated in a VITAL assay. The polyclonal P1B*07- and P4B*15-specific CD8+ T cells (0.12% and 0.76% peptide-specific T cells, respectively) lysed autologous peripheral blood mononuclear cells loaded with the mutated peptides, but not cells presenting the wild type peptides. P4B*15-specific CD8+ T cells showed 17.9% (±1.2%) MYD88L265P-peptide-specific cell killing at an E/T ratio of 1:1 compared to 2.6% (±1.2%) of non-specific cell lysis of unspecific effector cells against the same targets in three independent replicates, respectively. The specific lysis showed an E/T ratio-dependent manner as the specific lysis decreases with reducing E/T ratios. P1B*07-specific CD8+ T cells specifically killed 11.4% (±1.7%) of MYD88L265P loaded targets at an E/T ratio of 0.7:1 in comparison to 2.1% unspecific lysis of unspecific effector cells. In this study, we identified and characterized MYD88L265P mutation-derived HLA class I ligands for T cell mediated immunotherapy. The strong immunogenicity of the HLA-B*07 and HLA-B*15-restricted mutation-derived peptides as well as the functionality and specificity of peptide-specific CD8+ T cells, demonstrated by cytotoxicity assays, underline the potential of the MYD88L265P mutation as tumor-specific target. These data highlight the potential of MYD88L265P mutation-specific immunotherapy as a novel broadly applicable and tumor-specific treatment approach for patients with MYD88L265P+ NHL. Disclosures Langerak: InVivoScribe: Patents & Royalties: Licensing of IP and Patent on BIOMED-2-based methods for PCR-based Clonality Diagnostics.; DAKO: Patents & Royalties: Licensing of IP and Patent on Split-Signal FISH. Royalties for Dept. of Immunology, Erasmus MC, Rotterdam, NL; Roche: Other: Lab services in the field of MRD diagnostics provided by Dept of Immunology, Erasmus MC (Rotterdam).

Effective antigen-specific cancer immunotherapy requires exact knowledge of tumor-associated epitopes that can act as rejection antigens. Although the current paradigm views mutation-derived neoantigens as the most promising targets, we have recently demonstrated that leukemia-specific T-cell responses associated with survival benefit in CLL patients target a panel of non-mutated tumor-associated antigens.

[1] Lakka TA, Venalainen JM, Rauramaa R, Salonen R, Tuomilehto J, Salonen JT. Relation of leisure-time physical activity and cardiorespiratory fitness to the risk of acute myocardial infarction. N Engl J Med 1994;330:1549–54. [2] Haapanen-Niemi N, Miilunpalo S, Pasanen M, Vuori I, Oja P, Malmberg J. Body mass index, physical inactivity and low level of physical fitness as determinants of allcause and cardiovascular disease mortality—16 y follow-up of middle-aged and elderly men and women. Int J Obes Relat Metab Disord 2000;24:1465–74. [3] Farrell SW, Braun L, Barlow CE, Cheng YJ, Blair SN. The relation of body mass index, cardiorespiratory fitness, and all-cause mortality in women. Obes Res 2002;10:417–23. [4] Zittermann A. Vitamin D, and disease prevention with special reference to cardiovascular disease. Prog Biophys Mol Biol 2006;92:39–48. [5] Ardestani A, Parker B,Mathur S, ClarksonP, Pescatello LS, HoffmanHJ, et al. Relation of vitamin D level to maximal oxygen uptake in adults. Am J Cardiol 2011;107:1246–9. [6] Farrell SW, Cleaver JP, Willis BL. Cardiorespiratory fitness, adiposity, and serum 25dihydroxyvitamin d levels in men. Med Sci Sports Exerc 2011;43:266–71. [7] Farrell SW, Willis BL. Cardiorespiratory fitness, adiposity, and serum 25dihydroxyvitamin D levels in women: the Cooper Center Longitudinal Study. J Womens Health 2012;21:80–6. [8] Jackson AS, Blair SN, Mahar MT, Wier LT, Ross RM, Stuteville JE. Prediction of functional aerobic capacity without exercise testing. Med Sci Sports Exerc 1990;22:863–70. [9] MacLaughlin J, Holick MF. Aging decreases the capacity of human skin to produce vitamin D3. J Clin Invest 1985;76:1536–8. [10] Halloran BP, Portale AA. Vitamin D metabolism: the effects of aging. In: Feldman D, Glorieux FH, Pike JW, editors. Vitamin D. San Diego, Calif: Academic Press; 1997. p. 541–55.

Polycythemia vera (PV) is a myeloproliferative disorder mainly driven by the JAK2V617F mutant kinase. Except allogeneic stem cell transplantation, the current standard therapies, including novel therapies aiming for specific inhibition of JAK2 that are presently under investigation, do not offer a chance for cure. Of note, clear clinical benefits of unspecific immunotherapy with pegylated IFN-α suggest that PV might also be targeted by specific peptide based vaccination. Here we aimed to identify tumor associated HLA ligands of PV with special focus on peptides derived from JAK2. Therefore we analyzed HLA class I and II ligands of primary PV samples using the approach of direct isolation and identification of naturally presented HLA peptides by affinity chromatography and mass spectrometry. We analyzed the HLA class I ligandomes of 7 PV patients with the JAK2V617F mutation as well as of 4 patients with secondary (JAK2V617F negative) polycythemia. As controls, we profiled the ligandomes of 40 healthy volunteers (30 blood and 10 bone marrow samples). We identified more than 7,000 unique HLA class I ligands representing 4,435 proteins on primary PV samples. Moreover, 3 different HLA ligands derived from JAK2 on PV (KYLINLETL, HLA-A*24; KIGDFGLTK, HLA-A*03; GQGTFTKIFK, HLA-A*11) were identified. None of these ligands was found to be PV-exclusive, with JAK2 being represented on 3/4 (75%) of PV samples (JAK2V617F positive) and 18/30 (60%) healthy controls. No JAK2-derived peptides were identified on JAK2V617F negative patients, and no peptides covering the JAK2V617F mutation were identified. Since our previous work strongly suggests that tumor exclusive representation of an antigen is a prerequisite for the immunogenicity of its corresponding T cell epitopes, JAK2-derived wild-type peptides are unlikely to constitute optimal vaccines in PV. On the other hand, comparative profiling of the HLA class I ligandomes identified 935 antigens to be exclusively presented in JAK2V617F positive PV with 12 ligandome-derived tumor-associated antigens (LiTAAs) representing 26 different HLA ligands with exclusive representation in ≥30% of PV patients. Overlap analysis of PV patients with and without JAK2V617F mutation revealed 92% of the exclusive antigens to be represented on both groups, while only 8% and 0% showed exclusive representation in JAK2V617F positive and negative PV patients, respectively. As CD4+ T cells play important indirect and direct roles in anti-tumor immunity, we further applied our approach to HLA class II ligandomes of 8 PV patients (6 with, 2 without JAK2V617F mutation) and 18 HV (13 blood and 5 bone marrow samples) Using this approach, we identified more than 7,700 different naturally presented HLA class II ligands (1,700 source proteins). Notably, no JAK2-derived HLA class II peptides were identified in PV patients and HV ligandomes. Applying the same antigen-ranking strategy as described for HLA class I, we identified 5 additional HLA class II LiTAAs (exclusive in >30% of PV ligandomes) represented by 19 corresponding LiTAPs (ligandome-derived tumor-associated peptides). Overlap analysis of PV-exclusive source proteins revealed 2 proteins (POLA1, polymerase (DNA directed), alpha 1 and RHOQ, ras homolog family member Q) to be exclusively represented both in HLA class I and II ligandomes of PV patients. The novel HLA class I and II ligandome-derived tumor associated antigens identified in this study may constitute prime candidates for peptide-based immunotherapy of PV. Disclosures No relevant conflicts of interest to declare.

Despite the success of targeted therapy with tyrosine kinase inhibitors (TKIs), chronic myeloid leukemia (CML) remains largely incurable. Immunotherapy, and in particular multi-peptide vaccination, may be a promising approach to eliminate residual CML cells. As of now, a multitude of potential vaccine targets have been proposed based on reverse immunology and functional genomic approaches focusing either on BCR-ABL junction peptides, which represent CML-specific neo-antigens, or on aberrantly expressed self-proteins such as WT-1, PR and hTERT. However, the results of clinical studies employing such antigens have so far not been encouraging. This might in part be due to the inherent limitations of the above mentioned approaches: evidence of natural presentation of the predicted epitopes is lacking and the correlation of transcript abundance and HLA restricted presentation of the corresponding gene product has been shown to be skewed. Modern mass spectrometry, on the other hand, enables the comprehensive analysis of the entirety of naturally presented HLA ligands on tissues of interest, termed the HLA ligandome. Here we implemented this direct approach and comparatively mapped the HLA ligandome landscape of 16 primary CML samples and 40 healthy volunteer (HV) controls (30 blood and 10 bone marrow samples). We identified more than 30,000 different naturally presented HLA class I ligands representing ~10,000 source proteins. Regression analysis suggests source protein identifications on CML (4,337 different proteins) to attain >95% of maximum achievable coverage with the implemented analytical setup. Based on this extensive dataset, we investigated the HLA restricted presentation of established CML-associated/specific antigens and applied a novel approach defining tumor-associated antigens strictly based on exclusive and frequent representation in CML ligandomes. Strikingly, we found the vast majority of previously described antigens including wild-type BCR protein (6% CML, 5% HV), Myeloperoxidase (56% CML, 15% HV) and Proteinase 3 (38% CML, 11% HV) to be (also) represented on normal PBMC or BMNC. No evidence of naturally presented BCR-ABL junction peptides was found. However, we identified a panel of 7 LiTAAs (ligandome-derived tumor-associated antigens) represented by 16 different HLA ligands, showing CML-exclusive representation in ≥25% of CML patient ligandomes. As CD4+ T cells mediate important indirect and direct effects in anti-tumor immunity, we further applied our approach to HLA class II ligandomes of 15 CML patients and 18 HV (13 blood and 5 bone marrow samples), identifying more than 9,000 different naturally presented HLA class II ligands (1,900 source proteins). Applying the same antigen-ranking strategy as described for HLA class I, we identified 7 additional HLA class II LiTAAs represented by 50 corresponding LiTAPs (ligandome-derived tumor-associated peptides). Overlap analysis of CML-exclusive source proteins revealed 6 proteins to be represented both in HLA class I and II ligandomes. Notably, for Galectin-1, which shows CML-exclusive representation in 19% of HLA class I and 13% of HLA class II ligandomes, one of the HLA class II ligands was found to contain a complete, embedded HLA class I peptide. Such naturally presented embedded HLA ligands might present optimal vaccine candidates that are recognized by both, CD4+ and CD8+ T cells. Functional analysis of the here defined HLA class I and II LiTAPs with regard to induction of T cell responses is presently ongoing and serves to validate them as prime targets for the development of an off-the-shelf peptide vaccination in CML patients. Disclosures No relevant conflicts of interest to declare.

Results of allogeneic transplantation and reports that multiple myeloma (MM) cells can act as antigen presenting cells, but also the favorable immune effector-to-target cell ratio in the minimal residual disease (MRD) setting suggest that MM, which previously was considered to be weakly immunogenic, may be targeted effectively by T cell based immunotherapy. Following conventional therapy, peptide-based immunotherapy might offer a low side-effect opportunity to extend time to disease progression and possibly eradicate MRD. Therefore we aim to develop a multi-peptide vaccine for the immunotherapy of MM patients based on naturally processed and presented HLA ligands. In a first step we analyzed HLA surface expression on MM cells, which constitutes a prerequisite for effective T cell recognition. In MM patients (n=20), HLA class I expression was found to be heterogeneous with a mean of 415,000 ± 55,000 molecules on malignant cells. This expression was significantly (p Using the approach of direct isolation and identification of naturally presented HLA peptides by affinity chromatography and mass spectrometry, we identified more than 15,000 different HLA class I ligands from 8 primary MM samples, 2 plasma cell leukemia samples and 5 myeloma cell lines (MCLs). Comparative mapping of MM HLA ligandomes and the ligandomes of 40 HVs (30 blood and 10 bone marrow samples) revealed 15 ligandome-derived tumor-associated antigens (LiTAAs) represented by 36 corresponding HLA peptides (LiTAPs), showing exclusive representation in >30% of myeloma samples. Among them we identified 3 novel HLA peptides derived from Multiple Myeloma SET domain(MMSET), which is an established driver of myeloma cell proliferation, stimulating the expression of c-myc. We further studied the effect of the second-generation proteasome inhibitor carfilzomib on HLA surface expression and HLA ligandome composition. Notably, a slight albeit not significant increase in mean HLA class I expression was observed after 24h (t24) and 48h (t48) of carfilzomib treatment in MCL (n=5, fold change t24=1.4, pt24=0.37, fold change t48=1.6, pt48=0.25) as well as in primary myeloma cells in vitro (n=7, fold change t24=1.2, pt24=0.21, fold change t48=1.1, pt48=0.46). No significant loss or down regulation of HLA class I expression on MM cells was observed in ex vivo analysis of primary samples obtained from 2 myeloma patients before therapy and after 4 weeks of carfilzomib treatment. In order to analyze the impact of proteasome inhibition on the composition of the HLA class I ligandome we incubated MCLs (U266, MM.1S) with carfilzomib and performed quantitative mass spectrometry at baseline as well as at t24 and t48. On U266, 4.6% (4.0%) and 4.4% (8.6%) of total identified peptides (n=2444) were significantly up- or down regulated after 24h/48h of treatment with carfilzomib, respectively, whereas on MM.1S (n=1693) 1.6% (12.4%) and 10.8% (18.0%) were significantly regulated after 24h/48h. Focusing on the LiTAPs presented on these MCLs, 76.5% (88.2%) on U266 and 83.3% (66.7%) on MM.1S remained stably expressed after incubation with carfilzomib. Further analyses evaluating overall changes in peptide motifs and the occurrence of novel, cryptic peptides after carfilzomib treatment are ongoing. Taken together our data demonstrate that myeloma cells are valid targets for T cell based immunotherapy and provide a panel of novel, naturally presented myeloma-associated antigens, which exhibit stable presentation under proteasome inhibitor therapy. Disclosures No relevant conflicts of interest to declare.

The breakthrough clinical success of immune checkpoint inhibition with PD-1 and CTLA-4 antibodies illustrates the potential of T cell based immunotherapy to effectively treat malignancies. However, a remaining major challenge is to increase the specificity and frequency of anti-cancer immune responses. A rational approach to achieve this goal is therapeutic vaccination, as highlighted by a recent phase II vaccination study using naturally presented HLA ligands, which induced vaccine-specific immune responses that were associated with improved clinical outcome (Walter et al., Nat. Med. 2012). This underscores the potential of identifying patho-physiologically relevant targets by direct differential analysis of the entire landscape of HLA-presented peptides, termed the HLA ligandome. Here we applied this approach to chronic lymphocytic leukemia (CLL) with the aim of developing a CLL-specific multi-peptide vaccine. Using cohorts of 30 patients and 30 healthy volunteers, we comparatively and extensively mapped the HLA ligandome landscape of CLL and identified more than 20,000 different HLA class I and II ligands representing 7,377 source proteins, attaining >95% of the maximum attainable coverage (as extrapolated from saturation analysis). A set of 49 HLA ligand source proteins (225 different HLA ligands) showed CLL-exclusive representation in >20% of CLL patients, thereby constituting the novel category of ligandome-derived tumor associated antigens (LiTAAs). These LiTAAs were validated to be broadly and frequently represented across different stages and mutational subtypes of CLL and found to be robustly represented in HLA ligandomes of patients undergoing standard chemo-/immunotherapy. Functional characterization by IFNγ-ELISPOT revealed peptide-specific immune recognition of 14/15 (93.3%) corresponding HLA ligands (LiTAPs) exclusively in CLL patients. These immune responses were strictly CLL (LiTAP-) directed and mediated by functional patient-derived CD8+ T cells. Notably, for immunoreactive LITAPs, a linear correlation between frequencies of representation in the CLL ligandomes and immune recognition by CLL patient PBMC was observed (R²=0.59), suggesting that LiTAP presentation on cancer cells is a direct prerequisite for the presence of anti-LiTAP immune responses. Moreover, we investigated the prognostic relevance of LiTAP-specific immune responses: Retrospective survival analysis of 45 CLL patients analyzed by IFNγ-ELISPOT assays comparing cases with T cell responses specific for 0-1 LiTAPs (n=32) versus >1 LiTAPs (n=13) suggests a prolonged overall survival in the multiple-response cohort as compared to the single/non-responders (P=0.0695, log-rank test, Fig. 1). Taken together these data provide clear evidence for tumor-dependent priming of LiTAP-specific T cells in CLL patients and indicate their involvement in tumor control in vivo. This directly implies these non-mutant self-peptides as pathophysiologically relevant tumor antigens in CLL thereby validating our target identification approach and encouraging future evaluation in clinical vaccination trials. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Accelerated clonal evolution and inhibition of immune effector functions are fundamental drawbacks of chemotherapeutic treatment of chronic lymphocytic leukemia (CLL) which contribute to increased clinical aggressiveness of relapsed disease. Anticancer immune responses such as graft versus leukemia effects and remissions after donor lymphocyte infusions, on the other hand, have been correlated to long-term CLL-free survival. Clonotype analysis in these cases suggested clonally expanded CD8+ T cells recognizing tumor associated antigens (TAAs) presented by HLA as mediators of the observed effects, thus making CLL an attractive target for peptide vaccine-based immunotherapy. We here report on our approach of direct isolation and identification of naturally processed and presented HLA ligands from tissues of interest by affinity chromatography and mass spectrometry. Comparative and semi-quantitative analysis of the HLA ligandomes of malignant and benign samples provided the rationale for the identification of ligandome derived TAAs (LiTAAs) and informed selection of peptide vaccine candidates. HLA class I ligands were isolated from MACS-sorted CLL cells as well as from normal B cells or PBMC of healthy volunteers using a standard immunoaffinity purification protocol. Liquid chromatography coupled mass spectrometry (LC-MS/MS) peptide analysis was performed on a LTQ Orbitrap hybrid mass spectrometer followed by database assisted processing of fragment spectra. Semi-quantitative data analysis provided information regarding the abundance of HLA ligands in the respective ligandomes. In addition, HLA surface expression on CLL cells and autologous normal B cells was quantitatively determined using a flow cytometric assay. Selected peptides were characterized functionally in IFN-γ ELISPOT assays using PBMC of healthy volunteers and CLL patients. No significant difference in HLA class I surface expression between CLL cells and autologous normal B cells was observed. So far, we were able to map the HLA class I peptidomes of 25 CLL patients and 35 healthy controls. In total, we were able to identify more than 25,000 different HLA ligands representing >8,500 different source proteins. More than 15,000 different ligands were derived from CLL cells representing a total of 6,500 source proteins. A twofold data mining approach was used to identify both, broadly presented LiTAAs suited for off-the-shelf vaccine development, and LiTAAs showing patterns of patient-specific overrepresentation allowing for actively personalized target identification. The former strategy enabled us to pinpoint the most frequently and abundantly represented targets from the bulk of over 2,000 source proteins, which were exclusively represented in the ligandomes of CLL cells. Several published CLL-associated antigens/epitopes were found to be presented (e.g. Pim-1 Oncogene, SET nuclear oncogene, Mucin-1), which served to validate our methodological approach as proof of principle. Beyond that we identified a vast array of novel proteins that are broadly and exclusively represented in the HLA peptidome of CLL cells. Based on these findings we selected HLA-A*02, A*03 and B*07 restricted ligands derived from top ranking LiTAAs (e.g.TP53I11, PARP3, CDCA7L) for immunological characterization. Using patient PBMC, we observed frequent, reproducible and specific immune recognition of the selected peptides by CD8+ T cells in recall ELISPOT assays. The observed reactivity to CLL-associated self-peptides indicates their potential as therapeutic vaccines while underlining the validity of our target identification and selection strategy. Currently we are expanding our analyses to cover a comprehensive spectrum of HLA types with the goal to develop a clinically applicable CLL-specific multi-peptide vaccine. Disclosures: No relevant conflicts of interest to declare.

Despite the success of targeted therapy using tyrosine kinase inhibitors (TKIs), chronic myeloid leukemia (CML) remains largely incurable, most likely due to the treatment resistance of leukemic stem cells. Specific T-cell based immunotherapies for patients with CML might be able to eliminate these residual CML cells. For this goal the identification of tumor associated HLA-presented peptides, which are able to induce a tumor-specific T cell response, is indispensable. However, only few tumor associated antigens for CML are described till date. Thus the aim of this study was to identify for the first time naturally processed and presented HLA ligands from the cell surface of primary CML cells. HLA class I ligands from primary CML cells as well as from bone marrow mononuclear cell (BMNCs) and peripheral blood PBMCs of healthy donors were analyzed using the approach of direct isolation and identification of naturally presented HLA peptides by affinity chromatography and mass spectrometry (LC-MS/MS). LC-MS/MS peptide analysis provided qualitative and semi-quantitative information regarding the composition of the respective ligandomes. Comparative analysis of malignant and benign samples served to identify ligandome-derived tumor associated antigens (LiTAAs) and to select peptide vaccine candidates. So far 10 CML patients (8 chronic phase, 2 accelerated phase) and 30 healthy donors were analyzed within this study. We were able to identify a total of more than 8200 CML derived HLA ligands representing >4500 different source proteins, of which 734 were exclusively represented in CML, but not in healthy PBMCs/BMNCs. 55 of these CML exclusive antigens are presented on 3 or more of all examined CML patients, representing, as broadly represented LiTAAs, promising targets for peptide vaccination. For the first time, previously predicted CML tumor associated antigens for example Myeloperoxidase (10 peptide sequences on 7 CML patients) and Proteinase 3 (5 peptide sequences on 4 CML patients) were here confirmed by direct elution from primary CML cells, which also served to validate our methodological approach. Notably, beyond that we also identified a vast array of novel proteins (e.g. Carcinoembryonic antigen-related cell adhesion molecule 8, CEACAM8; Matrix metallopeptidase 8, MMP8; intracellular adhesion molecule 3, ICAM3) that are broadly and exclusively represented in the ligandome of CML cells. By providing for the first time HLA class I tumor associated antigens, directly obtained from the HLA ligandomes of CML patients, this study may pave the way for the future development of a multipeptide-based immunotherapy approach to eradicate residual CML cells after conventional TKI therapy. Disclosures: No relevant conflicts of interest to declare.

CD4+ T cells are crucial for the induction and maintenance of cytotoxic T cell responses, but can also mediate direct tumor rejection. The therapeutic efficacy of peptide-based cancer vaccines may thus be improved by including HLA class II epitopes to stimulate T helper cells. In contrast to HLA class I ligands, only a small number of class II ligands of TAA has been described so far. We recently reported on the overexpression of HLA class II in AML cells as compared to autologous monocytes and granulocytes as well as on the first HLA class I leukemia associated antigens identified directly on the cell surface of primary AML cells (Stickel et. al. abstract in Blood 2012). In this study we characterized the HLA class II ligandome in AML to identify additional ligands for a peptide-based immunotherapy approach. HLA class II ligands from primary AML cells as well as bone marrow and peripheral blood mononuclear cell (BMNCs/PBMCs) of healthy donors were analyzed using the approach of direct isolation and identification of naturally presented HLA peptides by affinity chromatography and mass spectrometry (LC-MS/MS). LC-MS/MS peptide analysis provided qualitative and semi-quantitative information regarding the composition of the respective ligandomes. Comparative analysis of malignant and benign samples served to identify ligandome-derived tumor associated antigens (LiTAAs) and to select peptide vaccine candidates. Most abundantly detected peptides were functionally characterized with regard to their ability to induce a specific CD4+ T-cell response in healthy donors and in tumor patients using ELISpot. Samples from 10 AML patients (5 FLT3-ITD mutated) and 18 healthy donors were analyzed. We identified more than 2,100 AML-derived HLA class II ligands representing >1,000 different source proteins, of which 315 were exclusively represented in AML, but not in healthy PBMC/BMNC. Data mining for broadly represented LiTAAs pinpointed 26 HLA class II ligands from 8 source proteins that were presented exclusively on more than 40% of all analyzed AML samples as most promising targets. Amongst them were already described TAAs (e.g., RAB5A) as well as several so far understated proteins (e.g. calsyntenin 1, glycophorin A, mannose-binding lectin 2). Subset analysis revealed 58 LiTAAs presented exclusively on FLT3-ITD mutated AML cells. Additional screening for HLA class II ligands from described leukemia associated antigens showed positive results for NPM1 (1 peptide sequence) and MPO (13 peptide sequences). Peptides from calsyntenin 1 and RAB5A were able to elicit CD4+-T-cell response in 25% of tested AML patients (n=16). Thus, our study identified, for the first time, HLA class II tumor associated antigens directly obtained from the HLA ligandomes of AML patients and thereby represents a further step to our goal of developing a multipeptide vaccine for immunotherapy of AML. Disclosures: No relevant conflicts of interest to declare.

7019 Background: CLL is a malignancy of mature B cells and constiututes the most common leukemia in adults. It is characterized by a progressive accumulation of clonal B cells, which coexpress CD19, CD23 and CD5. NFAT is a family of highly phosphorylated transcription factors residing in the cytoplasm of resting cells. Upon dephosphorylation by calcineurin, NFAT proteins translocate to the nucleus where they orchestrate developmental and activation programs in diverse cell types. In this study, we investigated the significance of NFAT signaling in B-CLL. Methods: NFAT2 expression and aberrant nuclear translocation in CLL cells (n=30) was assessed by Western Blotting and immunofluorescence. In addition, NFAT2 mRNA levels were measured by qRT-PCR and its DNA binding capacity was assessed using an electrophoretic mobility shift assay. Transcriptional activity of NFAT2 proteins in CLL cells was further analyzed by determining the expression of several well characterized NFAT target genes. Results: We detected a profound overexpression of NFAT2 mRNA and protein in all CLL samples. Using qRT-PCR we found that CD19+CD5+ CLL cells exhibited a significant overexpression of NFAT2 as compared to CD19+ B cells isolated from healthy donors (8-200fold). This overexpression of NFAT2 in CLL cells could also be confirmed on the protein level. We could further demonstrate that even under resting conditions significant amounts of NFAT2 protein had translocated to the nucleus in CLL cells, whereas virtually all NFAT2 was in the cytoplasm in non-malignant B cells. Nuclear NFAT2 in CLL cells was able to bind DNA but its transcriptional activity with respect to several apoptosis-regulating genes (i.e. Spp1, Pdcd1) was severely compromised. Conclusions: These results provide strong evidence that the Ca2+/NFAT signalling axis is constitutively activated in CD5+CD19+ CLL cells. Reduced expression of several apoptosis regulators which are known target genes of NFAT2 links deregulation of this signaling cascade to CLL progression. Further investigation is warranted to investigate the therapeutic potential of modulating Ca2+/Calcineurin/ NFAT signaling in CLL.

Abstract 3909 Chronic Lymphocytic Leukemia (CLL) is a hematological malignancy of mature B cells and constitutes the most common leukemia in adults. It is characterized by a progressive accumulation of clonal B cells, which coexpress CD19, CD23 and CD5. The clinical course of CLL can be predicted by serveral prognostic markers like CD38, ZAP70 and cytogenetic abnormalities. While the treatment of CLL has significantly improved during recent years, it remains an essentially uncurable disease and the molecular events that lead to its development are still largely ellusive. NFAT is a family of highly phosphorylated transcription factors residing in the cytoplasm of resting cells. Upon dephosphorylation NFAT proteins translocate to the nucleus where they orchestrate developmental and activation programs in diverse cell types. NFAT is inactivated by a network of several kinases. Several recent studies have demonstrated that Ca2+/NFAT signaling is involved in the pathogenesis of a wide array of different tumor types including pancreatic adenocarcinoma, breast cancer and Non Hodgkin′s lymphoma. In this study we investigated the significance of the Ca2+/NFAT signaling pathway in B-CLL. For this purpose, we analyzed CLL cell lines (MEC-1, JVM-3) as well as primary blood samples from patients with CLL (n=30). The analyzed patient population exhibited a representative distribution of age, sex, Binet stage, WBC count, cytogenetics and IGVH mutational status. PBMC were obtained by ficoll density gradient centrifugation and B cells were subsequently isolated using magnetic bead technology. NFAT2 expression and aberrant nuclear translocation was then assessed by Western Blotting and Immunofluorescence staining. In addition, NFAT2 mRNA levels were measured by qRT-PCR and its DNA binding capacity was assessed using an electrophoretic mobility shift assay (EMSA). Transcriptional activity of NFAT2 proteins in CLL cells was further analyzed by determining the expression of several well characterized NFAT target genes by qRT-PCR and Western Blotting. In our analysis, we detected a profound overexpression of NFAT2 mRNA as well as NFAT2 protein in all CLL samples. Using qRT-PCR we found that CD19+CD5+ CLL cells exhibited an at least three fold overexpression of NFAT2 as compared to CD19+ B cells isolated from healthy donors. In one case, NFAT2 expression in CLL cells was 200 times higher than in the corresponding control cell population. This profound overexpression of NFAT2 in CLL cells could be confirmed on the protein level using Western Blotting and Immunocytochemistry. We could further demonstrate that even under resting conditions significant amounts of NFAT2 protein had translocated to the nucleus in CLL cells, whereas virtually all NFAT2 was in the cytoplasm in healthy B cells. NFAT2 nuclear translocation could be inhibited using pretreatment with Cyclosporin A demonstrating that this process was still calcineurin-dependent in CLL cells. We could further show that nuclear NFAT2 in CLL cells was able to bind DNA but that its transcriptional activity with respect to several apoptosis and cell cycle regulating genes was severely compromised when compared with healthy CD19+CD5+ B cells. In summary, these results provide strong evidence that the Ca2+/NFAT signaling axis is con-stitutively activated in CD19+CD5+ CLL cells. Reduced expression of several apoptosis and cell cycle regulating proteins which are known target genes of NFAT2 potentially links deregulation of this signaling cascade to CLL progression. This is in line with our observation that genetic loss of NFAT2 can induce acceleration of CLL in TCL1 mouse model. Further investigation is warranted to decipher the therapeutic potential of modulating Ca2+/Calcineurin/NFAT signaling in CLL. Disclosures: No relevant conflicts of interest to declare.

Abstract 3012 Introduction: Eradication of minimal residual disease (MRD) by antigen-specific immunotherapy after conventional chemotherapy and also after allogeneic stem cell transplantation may improve overall survival in acute myeloid leukemia (AML). To achieve this goal there is a need for novel leukemia-associated HLA-presented peptides, which are able to induce a tumor-specific T cell response. Accordingly we here aimed to identify novel leukemia associated antigens employing, for the first time, the approach of direct elution and analysis of naturally processed and presented HLA ligands from the surface of primary AML cells. In particular we focused on peptides derived from AML specific mutated proteins, which are produced only by leukemia cells, only. Methods: HLA class I ligands were isolated from MACS-purified PBMCs of AML patients using immunoprecipitation. Liquid chromatography tandem mass spectrometry (LC-MS/MS) based peptide sequencing was performed to identify HLA presented peptides. The obtained data were mined for leukemia associated peptides by comparison of HLA ligandomes of AML cells with that of PMBCs and granulocytes from healthy donors, investigation of gene expression databases and literature research. All acquired peptide sequences were blasted against the COSMIC database (http://www.sanger.ac.uk/genetics/CGP/cosmic/), to identify sequences from AML specific mutated proteins. Identified mutations were confirmed by PCR. In addition, HLA quantification experiments on the cell surface of AML cells and autologous healthy monocytes and granulocytes were performed using a flow cytometric indirect immunofluorescence assay. Results: Quantitative preliminary results showed similar amounts of HLA class I molecules on AML cells as compared to autologous healthy monocytes and granulocytes lymphocytes (p=0.56, unpaired t-test). For HLA class II we could show a clear upregulation on AML cells compared to autologous healthy leukocytes, not reaching level of significance due to the small cohort examined till now (p= 0.12, unpaired t-test). We were able to identify a total of more than 10.000 peptides from 6 AML patients, including 2 NPM1 mutated AMLs, 1 FLT3-ITD mutated AML, 1 JAK2 mutated AML from osteomyelofibrosis and two AMLs from myelodysplastic syndrome without any cytogenetical aberrations. Using a new approach of spectral counting the obtained sequences could be compared semi-quantitatively to the same amount of peptides obtained from HLA matched healthy volunteers. Thus several new naturally processed and presented leukemia-associated peptides (LEUMAPs) were identified. This comprised, amongst others, ligands of p53, c-myc oncogene, PRAME, RHAMM and NPM1. The sequences of identified LEUMAPs were all verified by LC-MS/MS-based peptide sequencing of their synthetic counterparts. Testing for immunogenicity by ELISPOTs-assay and an in vitro T cell priming approach is presently ongoing. Conclusions: Our analysis revealed several new leukemia associated antigens, which were, for the first time, directly obtained from the HLA ligandomes of AML patients. Quantitative HLA analysis showed no loss or downregulation of HLA molecules on the surface of AML cells, thereby excluding a potential major obstacle for successful immunotherapy. Our results may pave the way for future peptide based immunotherapy approaches to eradicate MRD after obtaining a remission by presently available conventional therapies in AML patients. Disclosures: No relevant conflicts of interest to declare.

Abstract 4119 Recurrent disease, selection for chemo resistant clones and inhibition of immune effector functions are limitations of chemotherapeutic treatment of cancer including CLL. This is even more since graft-versus-leukemia effects and remissions after donor lymphocyte infusions have been correlated to long-term CLL free survival. Clonotype analysis of such cases suggested clonally expanded CD8+ T cells that recognize tumor associated antigens (TAAs) presented on HLA (human leukocyte antigen) as mediators of the observed effects, thus making CLL an attractive target for peptide vaccine based immunotherapy. For this goal we established the approach of direct isolation and identification of naturally processed and presented HLA ligands from tissues of interest by affinity chromatography and mass spectrometry. Comparative, semi-quantitative analysis of the HLA ligandomes of malignant and benign samples provided the rationale for selection of potential targets. 42 CLL patients (ages 48–85; median 70 years) of different HLA types and disease stages (Binet A, 24 patients; Binet B, 11 patients; Binet C, 7 patients) were enrolled in this study. Furthermore we collected blood samples from 50 healthy volunteers. HLA class I ligands were isolated from CLL cells as well as benign B cells from healthy donors and unsorted healthy PBMC using a standard affinity chromatography protocol implementing the pan-HLA class I specific antibody W6/32. Liquid chromatography coupled mass spectrometry (LC-MS/MS) based peptide analysis was performed on an LTQ Orbitrap hybrid mass spectrometer followed by annotation of fragment spectra using the MASCOT search algorithm against the human proteome comprised in the SwissProt database. Spectral counting based analysis provided semi-quantitative information regarding the abundance of HLA ligands and their source proteins in the respective ligandomes. In addition, HLA quantification experiments on the cell surface of CLL cells and autologous healthy B cells were performed using a flow cytometric indirect immunofluorescence assay. For this interim analysis we completely analyzed the HLA ligandomes of 7 CLL patients and 10 healthy controls. In total, we identified more than 15.000 different HLA ligands representing more than 7300 different proteins. This comprised more than 6,500 different ligands from CLL cells representing a total of 4,149 source proteins. Comparative analysis of representation in the HLA ligandomes of CLL cells, healthy B cells and healthy PBMC identified 1741 different source proteins as being exclusively expressed in CLL. Semi-quantitative evaluation revealed 138 of these proteins as being highly expressed on CLL (e.g. SET proto-oncogene, Pim-1 Oncogene, Mucin 1). Flow cytomerty based quantification of surface HLA expression revealed similar amounts of HLA class I (p=0.23, unpaired t-test) and II (p=0.33, unpaired t-test) molecules on CLL cells and autologous benign B lymphocytes, with a trend to higher HLA expression on CLL cells. Taken together, the presented strategy enabled the identification of a vast array of both known and novel TAAs and their corresponding naturally processed and presented HLA ligands in CLL. It pinpointed highly overrepresented TAAs for further analysis and immunogenicity testing. Expansion of the dataset after analysis of all enrolled patients will provide an unprecedented in-depth characterization of the HLA ligandome of CLL for future immunotherapeutic approaches. Disclosures: No relevant conflicts of interest to declare.