Neoantigen Prioritization Research Articles

Cancer neoantigen prioritization through sensitive and reliable proteogenomics analysis

Genomics-based neoantigen discovery can be enhanced by proteomic evidence, but there remains a lack of consensus on the performance of different quality control methods for variant peptide identification in proteogenomics. We propose to use the difference between accurately predicted and observed retention times for each peptide as a metric to evaluate different quality control methods. To this end, we develop AutoRT, a deep learning algorithm with high accuracy in retention time prediction. Analysis of three cancer data sets with a total of 287 tumor samples using different quality control strategies results in substantially different numbers of identified variant peptides and putative neoantigens. Our systematic evaluation, using the proposed retention time metric, provides insights and practical guidance on the selection of quality control strategies. We implement the recommended strategy in a computational workflow named NeoFlow to support proteogenomics-based neoantigen prioritization, enabling more sensitive discovery of putative neoantigens.

PTuneos: prioritizing tumor neoantigens from next-generation sequencing data

BackgroundCancer neoantigens are expressed only in cancer cells and presented on the tumor cell surface in complex with major histocompatibility complex (MHC) class I proteins for recognition by cytotoxic T cells. Accurate and rapid identification of neoantigens play a pivotal role in cancer immunotherapy. Although several in silico tools for neoantigen prediction have been presented, limitations of these tools exist.ResultsWe developed pTuneos, a computational pipeline for prioritizing tumor neoantigens from next-generation sequencing data. We tested the performance of pTuneos on the melanoma cancer vaccine cohort data and tumor-infiltrating lymphocyte (TIL)-recognized neopeptide data. pTuneos is able to predict the MHC presentation and T cell recognition ability of the candidate neoantigens, and the actual immunogenicity of single-nucleotide variant (SNV)-based neopeptides considering their natural processing and presentation, surpassing the existing tools with a comprehensive and quantitative benchmark of their neoantigen prioritization performance and running time. pTuneos was further tested on The Cancer Genome Atlas (TCGA) cohort data as well as the melanoma and non-small cell lung cancer (NSCLC) cohort data undergoing checkpoint blockade immunotherapy. The overall neoantigen immunogenicity score proposed by pTuneos is demonstrated to be a powerful and pan-cancer marker for survival prediction compared to traditional well-established biomarkers.ConclusionsIn summary, pTuneos provides the state-of-the-art one-stop and user-friendly solution for prioritizing SNV-based candidate neoepitopes, which could help to advance research on next-generation cancer immunotherapies and personalized cancer vaccines. pTuneos is available at https://github.com/bm2-lab/pTuneos, with a Docker version for quick deployment at https://cloud.docker.com/u/bm2lab/repository/docker/bm2lab/ptuneos.

Abstract CT006: Clinical study of a personalized neoantigen cancer vaccine in treating patients with advanced malignant tumor

Abstract Despite recent advances in cancer treatment by immune checkpoint blockade therapies, the overall response rate is still low in patients with advanced malignancies. Therapeutic cancer vaccines designed to reactivate T cells against tumor cells has achieved great progress. Neoantigens arisen from tumor somatic mutations provide tumor specific targets for developing personalized cancer vaccines, which are able to elicit strong T cell mediated immune responses due to the immunogenicity and the overcoming immune tolerance. Therefore, we developed a peptide-based personalized cancer vaccines (iNeo-Vac-P01) to mobilize immunity against a spectrum of tumor neoantigens, and initiated a clinical trial (NCT03662815). This study was a single-arm, open-label, Phase I trial, carried out in Sir Run Run Shaw Hospital, Zhejiang University School of Medicine in China with approval by the institutional review board and independent ethics committee. The main objectives were to assess safety, progression-free survival (PFS) and overall survival (OS). Patients with advanced malignancies were enrolled. For each patient, Exome-seq and RNA-seq was conducted. The analyses of somatic mutations, tumor neoantigens, as well as neoantigen prioritization and long peptide design, were performed by in-house pipeline (iNeo). The clinical-grade long peptides incorporating multiple neo-epitopes of both HLA class I and II were synthesized. The iNeo-Vac-P01 consisting of 10~20 long peptides per patient (grouped into 3~4 pools with 0.1mg of each peptide) was administered subcutaneously on day 1, 4, 8, 15, and 22 (priming phase) and day 78 and 162 (booster phase) with GM-CSF as adjuvant. The Cox regression model was used to assess the risk factors for OS and PFS. Total 30 patients were enrolled with different cancer types, including non-small cell lung cancer, pancreatic cancer, colorectal cancer, melanoma, ovarian cancer and others. The iNeo-Vac-P01 vaccines were successfully manufactured and administered for all patients. None of them had a grade 3 or 4 adverse event. 18 of them completed priming phase were included in the efficacy evaluation. The disease control rate (percentage of patients whose best response was not PD) was 75%. The median PFS was 5.37 months, and the median OS was not available yet. Overall, T cell responses were detected against 80% of peptide pools by ELISpot assays using post-vaccination blood samples. No response was detected in pre-vaccination blood samples, except for 2 patients who underwent pretreatment of radiofrequency ablation (RFA), which might cause tumor antigens release to trigger immune response. The preliminary results of this study demonstrated that iNeo-Vac-P01 was feasible, safe, able to elicit T cell mediated immune response specifically targeting tumor cells, and beneficial to patients with advanced diseases. The pre-existing responses against vaccine peptides in two patients with RFA pretreatment suggested that neoantigens incorporated in iNeo-Vac-P01 were accurately predicted. Citation Format: Yong Fang, Jiawei Shou, Fan Mo, Huimin Wang, Shanshan Zhang, Hongsen Li, Xiaoyun Zhou, Hongming Pan, Shuqing Chen. Clinical study of a personalized neoantigen cancer vaccine in treating patients with advanced malignant tumor [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr CT006.

Abstract B086: Exploiting large-scale HLA peptidomics to generate novel immunotherapies: A data-driven approach to true neoantigen prioritization

Abstract Purpose of Study: In this study we demonstrate accurate prediction of the impact of somatic mutations on the HLA presentation landscape achieved by interrogating a large scale database of 1.4 million unique HLA peptide sequences that have been directly identified by mass spectrometry. Background: Peptides presented to the immune system on HLA complexes are valuable targets for immunotherapeutic treatments. Identifying the full complement of peptides derived from a particular protein that are presented on major class I HLA restrictions will provide a vital step toward increasing the speed and viability of many immunotherapeutic strategies. Advances in next-generation sequencing (NGS) and single-cell technologies have enabled the accurate capture of somatic mutations accumulated by a tumor, yet a significant hurdle remains how this information can be utilized for immunotherapeutic benefit. In particular, identifying which somatic mutations produce neoantigens (peptides that contain a somatic mutation and are presented to the immune system in complex with HLA) is crucial to linking genetic changes with immunologic impact. Materials and Methods: Our approach to understanding the targetable human HLA peptidome is based on three key principles: achieving full proteome coverage, maximising individual protein coverage, and focusing on dominant HLA restrictions. By integrating novel cell biology, mass spectrometry, and bioinformatic technologies across over 1,000 individual experiments we have dramatically increased the depth of the HLA ligandome captured and achieved near total coverage of the protein-coding genome. Over 90% of the proteome has been captured for the restriction HLA-A*02:01, dominant in Caucasian populations. Our comprehensive genome coverage has enabled us to probe both directly and indirectly for the presence of neoantigens. Known somatic mutations within immortalized lines were used to generate bespoke reference databases that has led to direct identification of many hundreds of neoantigens. Results: Proteins that were found to contain neoantigens appeared to follow the same pattern of antigen processing and presentation as their unmutated equivalents. We have therefore found our HLA peptide dataset is able to offer significant value in predicting the likelihood of a somatic mutation creating a neoantigen. To test this, somatic mutations reported in 980 cell lines were probed against the database of HLA peptides. On average we find one peptide containing the mutated amino acid for every five somatic mutations reported. By incorporating the HLA background of the cell carrying the mutation, we narrow this prediction to one high-affinity HLA peptide for every fourteen somatic mutations reported. Comparing the peptides predicted in this analysis with those directly identified by mass spectrometry, we are able to show that we can prioritize mutation data by accurately predicting the presence and relative abundance of neoantigens. Our neoantigen prediction process is fully incorporated into a large scale database system, enabling us to seamlessly integrate NGS data from individual tissue and use peptidomic data to rapidly define the targetable landscape of an individual. Conclusions: An integrative approach to HLA peptidomics has delivered a powerful reference database for developing novel immunotherapies. Citation Format: Alex S. Powlesland, Geert P.M. Mommen, Ricardo J. Carreira, Jacob Hurst, Michael J. Cundell, David Lowne, Floriana Capuano, Bent K. Jakobsen. Exploiting large-scale HLA peptidomics to generate novel immunotherapies: A data-driven approach to true neoantigen prioritization [abstract]. In: Proceedings of the Fourth CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival; Sept 30-Oct 3, 2018; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2019;7(2 Suppl):Abstract nr B086.

Abstract PL02-02: Immunopeptidomics: Accelerating the development of personalized cancer immunotherapy

Abstract Cancer immunotherapy amplifies the inherent capacity of cytotoxic T cells to eliminate tumor cells by recognizing human leukocyte antigen-binding peptides (HLAp) presented specifically on tumors but not on normal cells. Recent data show that recognition of mutated neoantigens plays a key role. Currently, mass spectrometry (MS) is the only unbiased methodology to comprehensively uncover the repertoire of HLAp presented in vivo. Recently, we have developed an in-depth MS-based immunopeptidomics approach combined with exome sequencing analysis to directly identify neoantigens from human melanoma tumors (Bassani-Sternberg et al., Nat Commun 2016). This approach is not feasible for substantial number of patients for whom not enough tissue material is available. So far, the discovery of neoantigens relies mainly on prediction-based interrogation of the "mutanome." In a proof-of-concept study we showed that incorporation of deconvoluted HLAp data in ligand prediction algorithms can improve their accuracy (Bassani-Sternberg and Gfeller, J Immunol 2016). In addition, we assembled a comprehensive HLAp database in term of number of peptides and diversity of HLA-I molecules. We showed that by taking advantage of co-occurring HLA-I alleles, we can rapidly and accurately identify HLA-I binding motifs and map them to their corresponding alleles without any a priori knowledge of HLA-I binding specificity. Our novel and scalable approach uncovers new motifs for several alleles that up to now had no known ligands. HLA-ligand predictors trained on such data substantially improve neoantigen predictions in four melanoma and two lung cancer patients, indicating that unbiased HLAp data are ideal for in silico identification of neoantigens (Bassani-Sternberg et al., PLOS Comp Biol 2017). Furthermore, our immunopeptidomics database revealed hotspots that are sub-sequences of proteins frequently presented. We introduce two immunopeptidomics MS-based features to guide prioritization of neoantigens: the number of peptides matching a protein in our database and the overlap of the predicted wild-type peptide with other peptides in our database. We show as a proof of concept that our immunopeptidomics MS-based features improved neoantigen prioritization by up to 50% (Mueller et al., submitted). Overall, our work shows that, in addition to providing huge training data to improve the HLA binding prediction, immunopeptidomics also captures other aspects of the natural in vivo presentation that significantly improve prediction of clinically relevant neoantigens. Citation Format: Michal Bassani-Sternberg, Chloe Chong, Fabio Marino, HuiSong Pak, David Gfeller, Markus Müller, George Coukos. Immunopeptidomics: Accelerating the development of personalized cancer immunotherapy [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2017 Oct 26-30; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2018;17(1 Suppl):Abstract nr PL02-02.

'Hotspots' of Antigen Presentation Revealed by Human Leukocyte Antigen Ligandomics for Neoantigen Prioritization.

The remarkable clinical efficacy of the immune checkpoint blockade therapies has motivated researchers to discover immunogenic epitopes and exploit them for personalized vaccines. Human leukocyte antigen (HLA)-binding peptides derived from processing and presentation of mutated proteins are one of the leading targets for T-cell recognition of cancer cells. Currently, most studies attempt to identify neoantigens based on predicted affinity to HLA molecules, but the performance of such prediction algorithms is rather poor for rare HLA class I alleles and for HLA class II. Direct identification of neoantigens by mass spectrometry (MS) is becoming feasible; however, it is not yet applicable to most patients and lacks sensitivity. In an attempt to capitalize on existing immunopeptidomics data and extract information that could complement HLA-binding prediction, we first compiled a large HLA class I and class II immunopeptidomics database across dozens of cell types and HLA allotypes and detected hotspots that are subsequences of proteins frequently presented. About 3% of the peptidome was detected in both class I and class II. Based on the gene ontology of their source proteins and the peptide’s length, we propose that their processing may partake by the cellular class II presentation machinery. Our database captures the global nature of the in vivo peptidome averaged over many HLA alleles, and therefore, reflects the propensity of peptides to be presented on HLA complexes, which is complementary to the existing neoantigen prediction features such as binding affinity and stability or RNA abundance. We further introduce two immunopeptidomics MS-based features to guide prioritization of neoantigens: the number of peptides matching a protein in our database and the overlap of the predicted wild-type peptide with other peptides in our database. We show as a proof of concept that our immunopeptidomics MS-based features improved neoantigen prioritization by up to 50%. Overall, our work shows that, in addition to providing huge training data to improve the HLA binding prediction, immunopeptidomics also captures other aspects of the natural in vivo presentation that significantly improve prediction of clinically relevant neoantigens.

Abstract 604: Accurate identification and prioritization of candidate neoantigens from integrated cancer exome and transcriptome sequencing of FFPE samples

Abstract Precise identification and characterization of candidate neoantigens is important for the development of effective cancer vaccines, adoptive T-cell transfer, and prediction of response to checkpoint inhibitors. The candidate tumor neoantigens are actionable only when expressed, however, current prediction methods lack the capacity to evaluate neoantigen expression. Sequencing both DNA and RNA from a patient’s tumor tissue enables identification of mutations and evaluation of their expression leading to accurate identification of putative neoantigens. The purpose of this study was to develop and validate a methodology for co-extraction and sequencing of DNA and RNA from formalin-fixed paraffin-embedded (FFPE) samples to enable a robust neoantigen prediction protocol that integrates whole exome and transcriptome data to identify and prioritize tumor neoantigens for application in immuno-oncology research and clinical trials. In order to prepare high-quality sequencing libraries from FFPE specimens, the tissue was macrodissected to enrich for tumor-specific material, and improve the overall accuracy of next-generation sequencing for detection of somatic alterations. Total DNA and RNA was co-extracted and purified. The DNA was used to prepare whole exome sequencing (WES) libraries, while the co-extracted RNA was ribosome-depleted, and reverse-transcribed to prepare RNA sequencing (RNAseq) libraries. The WES and RNAseq data was then analyzed using a multi-algorithm HLA typing and neoantigen prediction protocol (ImmunoSelect-RTM). ImmunoSelect-R evaluates somatic genomic alterations identified from WES of tumor and matched normal tissue to ensure appropriate prediction of candidate neoantigens. The process of neoantigen prediction was then refined by integration of patient tumor-matched RNAseq data, which allowed for removal of non-expressed putative neoantigens. To further validate the approach, we applied the methodology to a set of experimentally validated neoantigens. In this setting, ImmunoSelect-R correctly classified 18 out of 19 as strong neoantigen candidates, suggesting a sensitivity of greater than 90%. Moreover, in a set of 10 patients, ImmunoSelect-R consistently ranked experimentally validated neoantigens within the top 20% of all neoantigen candidates derived from whole exome sequencing. In summary, our combined tissue processing, macrodissection, co-extraction, and neoantigen prediction methodology is able to identify and prioritize candidate neoantigens. Our approach is unique in combining high-fidelity sequencing (WES) and expression (RNAseq) data to accurately inform the selection of actionable tumor neoantigens for immuno-oncology applications. Citation Format: Marián Novak, Sam Angiuoli, Luis A. Diaz, Andrew Georgiadis, Sian Jones, Peter R. Loverso, Sonya Parpart-Li, Maria Sevdali, Victor E. Velculescu, Ellen L. Verner, James White, Theresa Zhang, Mark Sausen. Accurate identification and prioritization of candidate neoantigens from integrated cancer exome and transcriptome sequencing of FFPE samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 604. doi:10.1158/1538-7445.AM2017-604

Abstract 632: Genome-scale neoantigen screening using ATLAS™ prioritizes candidate antigens for immunotherapy in a non-small cell lung cancer patient

Abstract Despite the unprecedented efficacy of checkpoint blockade (CPB) therapy in treating some cancers, the majority of patients fail to respond. Several lines of evidence support that the combination of CPB and neoantigen vaccine prolongs survival curves in cancer patients. Capitalizing on neoantigens derived from non-synonymous somatic mutations is a good strategy for therapeutic immunization. Current approaches to neoantigen prioritization involve mutanome sequencing, in silico epitope prediction algorithms, and experimental validation of cancer neoepitopes. Even the best in class in silico epitope prediction algorithms lack the accuracy necessary for efficacious personalized cancer vaccines. We sought to circumvent some of the limitations of currently available prediction algorithms by prioritizing neoantigens empirically using ATLAS™, a technology developed to screen T cell responses from any subject against their entire complement of potential neoantigens. Exome sequences were obtained from peripheral blood mononuclear cells (PBMC) and tumor biopsies from a non-small cell lung cancer patient who had been successfully treated with pembrolizumab. The tumor exome was sequenced and somatic mutations were identified. Individual DNA sequences (399 nucleotides) spanning each mutation site were built, cloned and expressed in E. coli co-expressing listeriolysin O. Polypeptide expression was validated using a surrogate T cell assay or by Western Blotting. Frozen PBMCs, collected pre- and post-therapy, were used to derive dendritic cells (MDDC). Both CD4+ and CD8+ T cells were enriched and expanded using microbeads. The E. coli clones were pulsed onto MDDC in an ordered array, then co-cultured either with CD8+ or with CD4+ T cells overnight. T cell activation was detected by analyzing cytokines in supernatants. Antigens were identified as clones that induced a cytokine response that exceeded three standard deviations of the mean of all negative control wells, then their identities compared with T cell epitopes predicted using previously described algorithms. We found biological evidence for neoantigens that were specifically responsive to peripheral CD8+ and CD4+ T cells, derived from the patient’s tumor, pre- and post-CPB intervention. Some of these neoantigens were identified as a T cell target both pre- and post-CPB therapy. We identified neoantigens for which no epitopes were predicted by in silico methods. These data represent evidence that multiple patient-specific neoantigens can be identified through functional evidence of T cell response from peripheral blood without epitope prediction. By profiling natural and CPB-enhanced immunity to neoantigens, a broad catalog of T cell targets can be identified for development of immunotherapies that engage T cells against cancer to improve outcomes for patients for whom current therapies are ineffective. Citation Format: Lila Ghamsari, Emilio Flano, Judy Jacques, Biao Liu, Zheng Yan, Aula Alami, Christine Kelley, Theresa Zhang, Jonathan Havel, Vladimir Makarov, Taha Merghoub, Jedd D. Wolchok, Matthew Hellman, Pamela Carroll, Timothy Chan, Jessica B. Flechtner. Genome-scale neoantigen screening using ATLAS™ prioritizes candidate antigens for immunotherapy in a non-small cell lung cancer patient [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 632. doi:10.1158/1538-7445.AM2017-632

Abstract A039: Accurate identification and prioritization of candidate neoantigens from cancer exome sequencing

Abstract Somatic, nonsynonymous genetic alterations present in cancer can lead to the formation of novel protein sequences and thus production of immunogenic “non-self” neoantigens. Those neoantigens with sufficient expression will be processed and presented on MHC molecules, subsequently inducing a tumor-specific T cell response. Emerging data from clinical studies suggest that neoantigen load and composition is associated with the efficacy of some immunotherapies. As neoantigens play central a role in the cancer-immunity cycle, it is critical to identify the most potent immunogenic neoantigens effectively and accurately. Here we have leveraged highly accurate cancer whole exome (WES) analyses from FFPE tumor tissue with a state-of-the-art analysis protocol to identify and prioritize candidate neoantigens most likely to promote an immune response. ImmunoSelect-R utilizes somatic variants from WES to ensure detection of true somatic peptides and minimize false positives, and provides accurate HLA typing from whole exome sequencing data with >99.9% sensitivity and specificity. When applied to a set of experimentally validated neoantigens, ImmunoSelect correctly classified 18 out of 19 as strong neoantigen candidates, suggesting a sensitivity of greater than 90%. Moreover, in a small set of 10 patients, ImmunoSelect consistently ranked experimentally validated neoantigens within top 20% of all neoantigen candidates derived from whole exome sequencing. In summary, ImmunoSelect is able to identify and prioritize candidate neoantigens from cancer exome sequencing results effectively and accurately, enabling personalized cancer vaccine development, adoptive T-cell transfer, and prediction of response to checkpoint inhibitors Citation Format: James White, John Simmons, Sam Angiuoli, Mark Sausen, Sian Jones, Lisa Kann, Manish Shukla, Maria Sevdali, Victor Velculescu, Luis Diaz, Theresa Zhang. Accurate identification and prioritization of candidate neoantigens from cancer exome sequencing [abstract]. In: Proceedings of the Second CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival; 2016 Sept 25-28; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2016;4(11 Suppl):Abstract nr A039.

Abstract SY18-02: Immunogenomics and precision cancer medicine

Abstract The expansion of precision cancer medicine through robust interrogation of cancer genomes at the point of care has enhanced patient stratification and treatment selection strategies in numerous cancer types. In parallel, the rise of multiple immunotherapies, such as immune checkpoint blockade, has significantly improved clinical outcomes for multiple patient populations. The intersection of clinical cancer genomics and immuno-oncology has created an opportunity to learn how the tumor and germline genomes impact immune checkpoint blockade response, and how the immune system can be leveraged with knowledge of the cancer genome to more effectively stratify patients and tailor immune-specific therapies. Here, we describe the landscape of genomic correlates of immune checkpoint blockade, and examine how emerging genomic studies of immunotherapy responders and non-responders may inform patient selection strategies for clinical trial design. Furthermore, we explore how knowledge of the cancer genome and prioritization of neoantigens that arise from somatic mutations may impact personalized cancer vaccine development. Finally, we examine emerging data on genomic resistance to immunotherapies and the implications for combination treatment strategies. Broadly, the convergence of comprehensive clinical genomics and immuno-oncology treatment strategies may inform precision medicine for immunotherapies across cancer types. Citation Format: Eliezer Van Allen. Immunogenomics and precision cancer medicine. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr SY18-02.