Human lung cancer harbors spatially organized stem-immunity hubs associated with response to immunotherapy

The organization of immune cells in human tumors is not well understood. Immunogenic tumors harbor spatially localized multicellular 'immunity hubs' defined by expression of the T cell-attracting chemokines CXCL10/CXCL11 and abundant T cells. Here, we examined immunity hubs in human pre-immunotherapy lung cancer specimens and found an association with beneficial response to PD-1 blockade. Critically, we discovered the stem-immunity hub, a subtype of immunity hub strongly associated with favorable PD-1-blockade outcome. This hub is distinct from mature tertiary lymphoid structures and is enriched for stem-like TCF7+PD-1+CD8+ T cells, activated CCR7+LAMP3+ dendritic cells and CCL19+ fibroblasts as well as chemokines that organize these cells. Within the stem-immunity hub, we find preferential interactions between CXCL10+ macrophages and TCF7-CD8+ T cells as well as between mature regulatory dendritic cells and TCF7+CD4+ and regulatory T cells. These results provide a picture of the spatial organization of the human intratumoral immune response and its relevance to patient immunotherapy outcomes.

High-throughput RNA isoform sequencing using programmed cDNA concatenation

Full-length RNA-sequencing methods using long-read technologies can capture complete transcript isoforms, but their throughput is limited. We introduce multiplexed arrays isoform sequencing (MAS-ISO-seq), a technique for programmably concatenating complementary DNAs (cDNAs) into molecules optimal for long-read sequencing, increasing the throughput >15-fold to nearly 40 million cDNA reads per run on the Sequel IIe sequencer. When applied to single-cell RNA sequencing of tumor-infiltrating T cells, MAS-ISO-seq demonstrated a 12- to 32-fold increase in the discovery of differentially spliced genes.

The HLA-II immunopeptidome of SARS-CoV-2

Targeted synthetic vaccines have the potential to transform our response to viral outbreaks, yet the design of these vaccines requires a comprehensive knowledge of viral immunogens. Here, we report severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) peptides that are naturally processed and loaded onto human leukocyte antigen-II (HLA-II) complexes in infected cells. We identify over 500 unique viral peptides from canonical proteins as well as from overlapping internal open reading frames. Most HLA-II peptides colocalize with known CD4+ T cell epitopes in coronavirus disease 2019 patients, including 2 reported immunodominant regions in the SARS-CoV-2 membrane protein. Overall, our analyses show that HLA-I and HLA-II pathways target distinct viral proteins, with the structural proteins accounting for most of the HLA-II peptidome and nonstructural and noncanonical proteins accounting for the majority of the HLA-I peptidome. These findings highlight the need for a vaccine design that incorporates multiple viral elements harboring CD4+ and CD8+ T cell epitopes to maximize vaccine effectiveness.

Unsupervised Protein-Ligand Binding Energy Prediction via Neural Euler's Rotation Equation

Protein-ligand binding prediction is a fundamental problem in AI-driven drug discovery. Prior work focused on supervised learning methods using a large set of binding affinity data for small molecules, but it is hard to apply the same strategy to other drug classes like antibodies as labelled data is limited. In this paper, we explore unsupervised approaches and reformulate binding energy prediction as a generative modeling task. Specifically, we train an energy-based model on a set of unlabelled protein-ligand complexes using SE(3) denoising score matching and interpret its log-likelihood as binding affinity. Our key contribution is a new equivariant rotation prediction network called Neural Euler's Rotation Equations (NERE) for SE(3) score matching. It predicts a rotation by modeling the force and torque between protein and ligand atoms, where the force is defined as the gradient of an energy function with respect to atom coordinates. We evaluate NERE on protein-ligand and antibody-antigen binding affinity prediction benchmarks. Our model outperforms all unsupervised baselines (physics-based and statistical potentials) and matches supervised learning methods in the antibody case.

The HLA-II immunopeptidome of SARS-CoV-2

Targeted synthetic vaccines have the potential to transform our response to viral outbreaks; yet the design of these vaccines requires a comprehensive knowledge of viral immunogens, including T-cell epitopes. Having previously mapped the SARS-CoV-2 HLA-I landscape, here we report viral peptides that are naturally processed and loaded onto HLA-II complexes in infected cells. We identified over 500 unique viral peptides from canonical proteins, as well as from overlapping internal open reading frames (ORFs), revealing, for the first time, the contribution of internal ORFs to the HLA-II peptide repertoire. Most HLA-II peptides co-localized with the known CD4+ T cell epitopes in COVID-19 patients. We also observed that two reported immunodominant regions in the SARS-CoV-2 membrane protein are formed at the level of HLA-II presentation. Overall, our analyses show that HLA-I and HLA-II pathways target distinct viral proteins, with the structural proteins accounting for most of the HLA-II peptidome and non-structural and non-canonical proteins accounting for the majority of the HLA-I peptidome. These findings highlight the need for a vaccine design that incorporates multiple viral elements harboring CD4+ and CD8+ T cell epitopes to maximize the vaccine effectiveness.

Spatial analysis of human lung cancer reveals organized immune hubs enriched for stem-like CD8 T cells and associated with immunotherapy response

The organization of immune cells in human tumors is not well understood. Immunogenic tumors harbor spatially-localized multicellular 'immunity hubs' defined by expression of the T cell-attracting chemokines CXCL10/CXCL11 and abundant T cells. Here, we examined immunity hubs in human pre-immunotherapy lung cancer specimens, and found that they were associated with beneficial responses to PD-1-blockade. Immunity hubs were enriched for many interferon-stimulated genes, T cells in multiple differentiation states, and CXCL9/10/11 + macrophages that preferentially interact with CD8 T cells. Critically, we discovered the stem-immunity hub, a subtype of immunity hub strongly associated with favorable PD-1-blockade outcomes, distinct from mature tertiary lymphoid structures, and enriched for stem-like TCF7+PD-1+ CD8 T cells and activated CCR7 + LAMP3 + dendritic cells, as well as chemokines that organize these cells. These results elucidate the spatial organization of the human intratumoral immune response and its relevance to patient immunotherapy outcomes.

Reversal of viral and epigenetic HLA class I repression in Merkel cell carcinoma

Cancers avoid immune surveillance through an array of mechanisms, including perturbation of HLA class I antigen presentation. Merkel cell carcinoma (MCC) is an aggressive, HLA-I-low, neuroendocrine carcinoma of the skin often caused by the Merkel cell polyomavirus (MCPyV). Through the characterization of 11 newly generated MCC patient-derived cell lines, we identified transcriptional suppression of several class I antigen presentation genes. To systematically identify regulators of HLA-I loss in MCC, we performed parallel, genome-scale, gain- and loss-of-function screens in a patient-derived MCPyV-positive cell line and identified MYCL and the non-canonical Polycomb repressive complex 1.1 (PRC1.1) as HLA-I repressors. We observed physical interaction of MYCL with the MCPyV small T viral antigen, supporting a mechanism of virally mediated HLA-I suppression. We further identify the PRC1.1 component USP7 as a pharmacologic target to restore HLA-I expression in MCC.

Unannotated proteins expand the MHC-I-restricted immunopeptidome in cancer

Tumor-associated epitopes presented on MHC-I that can activate the immune system against cancer cells are typically identified from annotated protein-coding regions of the genome, but whether peptides originating from novel or unannotated open reading frames (nuORFs) can contribute to antitumor immune responses remains unclear. Here we show that peptides originating from nuORFs detected by ribosome profiling of malignant and healthy samples can be displayed on MHC-I of cancer cells, acting as additional sources of cancer antigens. We constructed a high-confidence database of translated nuORFs across tissues (nuORFdb) and used it to detect 3,555 translated nuORFs from MHC-I immunopeptidome mass spectrometry analysis, including peptides that result from somatic mutations in nuORFs of cancer samples as well as tumor-specific nuORFs translated in melanoma, chronic lymphocytic leukemia and glioblastoma. NuORFs are an unexplored pool of MHC-I-presented, tumor-specific peptides with potential as immunotherapy targets.

Spatially organized multicellular immune hubs in human colorectal cancer

Immune responses to cancer are highly variable, with mismatch repair-deficient (MMRd) tumors exhibiting more anti-tumor immunity than mismatch repair-proficient (MMRp) tumors. To understand the rules governing these varied responses, we transcriptionally profiled 371,223 cells from colorectal tumors and adjacent normal tissues of 28 MMRp and 34 MMRd individuals. Analysis of 88 cell subsets and their 204 associated gene expression programs revealed extensive transcriptional and spatial remodeling across tumors. To discover hubs of interacting malignant and immune cells, we identified expression programs in different cell types that co-varied across tumors from affected individuals and used spatial profiling to localize coordinated programs. We discovered a myeloid cell-attracting hub at the tumor-luminal interface associated with tissue damage and an MMRd-enriched immune hub within the tumor, with activated T cells together with malignant and myeloid cells expressing T cell-attracting chemokines. By identifying interacting cellular programs, we reveal the logic underlying spatially organized immune-malignant cell networks.

Optimized liquid and gas phase fractionation increases HLA-peptidome coverage for primary cell and tissue samples

Mass spectrometry is the most effective method to directly identify peptides presented on HLA molecules. However, current standard approaches often use 500 million or more cells as input to achieve high coverage of the immunopeptidome and therefore these methods are not compatible with the often limited amounts of tissue available from clinical tumor samples. Here, we evaluated microscaled basic reversed-phase fractionation to separate HLA peptide samples off-line followed by ion mobility coupled to LC-MS/MS for analysis. The combination of these two separation methods enabled identification of 20% to 50% more peptides compared to samples analyzed without either prior fractionation or use of ion mobility alone. We demonstrate coverage of HLA immunopeptidomes with up to 8,107 distinct peptides starting with as few as 100 million cells. The increased sensitivity obtained using our methods can provide data useful to improve HLA binding prediction algorithms as well as to enable detection of clinically relevant epitopes such as neoantigens.

Profiling SARS-CoV-2 HLA-I peptidome reveals T cell epitopes from out-of-frame ORFs

T cell-mediated immunity plays an important role in controlling SARS-CoV-2 infection, but the repertoire of naturally processed and presented viral epitopes on class I human leukocyte antigen (HLA-I) remains uncharacterized. Here, we report the first HLA-I immunopeptidome of SARS-CoV-2 in two cell lines at different times post infection using mass spectrometry. We found HLA-I peptides derived not only from canonical open reading frames (ORFs) but also from internal out-of-frame ORFs in spike and nucleocapsid not captured by current vaccines. Some peptides from out-of-frame ORFs elicited T cell responses in a humanized mouse model and individuals with COVID-19 that exceeded responses to canonical peptides, including some of the strongest epitopes reported to date. Whole-proteome analysis of infected cells revealed that early expressed viral proteins contribute more to HLA-I presentation and immunogenicity. These biological insights, as well as the discovery of out-of-frame ORF epitopes, will facilitate selection of peptides for immune monitoring and vaccine development.

MS-Based HLA-II Peptidomics Combined With Multiomics Will Aid the Development of Future Immunotherapies

Immunotherapies have emerged to treat diseases by selectively modulating a patient's immune response. Although the roles of T and B cells in adaptive immunity have been well studied, it remains difficult to select targets for immunotherapeutic strategies. Because human leukocyte antigen class II (HLA-II) peptides activate CD4+ T cells and regulate B cell activation, proliferation, and differentiation, these peptide antigens represent a class of potential immunotherapy targets and biomarkers. To better understand the molecular basis of how HLA-II antigen presentation is involved in disease progression and treatment, systematic HLA-II peptidomics combined with multiomic analyses of diverse cell types in healthy and diseased states is required. For this reason, MS-based innovations that facilitate investigations into the interplay between disease pathologies and the presentation of HLA-II peptides to CD4+ T cells will aid in the development of patient-focused immunotherapies.

ERAP2 Increases the Abundance of a Peptide Submotif Highly Selective for the Birdshot Uveitis-Associated HLA-A29

Birdshot Uveitis (BU) is a blinding inflammatory eye condition that only affects HLA-A29-positive individuals. Genetic association studies linked ERAP2 with BU, an aminopeptidase which trims peptides before their presentation by HLA class I at the cell surface, which suggests that ERAP2-dependent peptide presentation by HLA-A29 drives the pathogenesis of BU. However, it remains poorly understood whether the effects of ERAP2 on the HLA-A29 peptidome are distinct from its effect on other HLA allotypes. To address this, we focused on the effects of ERAP2 on the immunopeptidome in patient-derived antigen presenting cells. Using complementary HLA-A29-based and pan-class I immunopurifications, isotope-labeled naturally processed and presented HLA-bound peptides were sequenced by mass spectrometry. We show that the effects of ERAP2 on the N-terminus of ligands of HLA-A29 are shared across endogenous HLA allotypes, but discover and replicate that one peptide motif generated in the presence of ERAP2 is specifically bound by HLA-A29. This motif can be found in the amino acid sequence of putative autoantigens. We further show evidence for internal sequence specificity for ERAP2 imprinted in the immunopeptidome. These results reveal that ERAP2 can generate an HLA-A29-specific antigen repertoire, which supports that antigen presentation is a key disease pathway in BU.

SARS-CoV-2 infected cells present HLA-I peptides from canonical and out-of-frame ORFs

T cell-mediated immunity may play a critical role in controlling and establishing protective immunity against SARS-CoV-2 infection; yet the repertoire of viral epitopes responsible for T cell response activation remains mostly unknown. Identification of viral peptides presented on class I human leukocyte antigen (HLA-I) can reveal epitopes for recognition by cytotoxic T cells and potential incorporation into vaccines. Here, we report the first HLA-I immunopeptidome of SARS-CoV-2 in two human cell lines at different times post-infection using mass spectrometry. We found HLA-I peptides derived not only from canonical ORFs, but also from internal out-of-frame ORFs in Spike and Nucleoprotein not captured by current vaccines. Proteomics analyses of infected cells revealed that SARS-CoV-2 may interfere with antigen processing and immune signaling pathways. Based on the endogenously processed and presented viral peptides that we identified, we estimate that a pool of 24 peptides would provide one or more peptides for presentation by at least one HLA allele in 99% of the human population. These biological insights and the list of naturally presented SARS-CoV-2 peptides will facilitate data-driven selection of peptides for immune monitoring and vaccine development.

Cumulus provides cloud-based data analysis for large-scale single-cell and single-nucleus RNA-seq

Massively parallel single-cell and single-nucleus RNA sequencing has opened the way to systematic tissue atlases in health and disease, but as the scale of data generation is growing, so is the need for computational pipelines for scaled analysis. Here we developed Cumulus-a cloud-based framework for analyzing large-scale single-cell and single-nucleus RNA sequencing datasets. Cumulus combines the power of cloud computing with improvements in algorithm and implementation to achieve high scalability, low cost, user-friendliness and integrated support for a comprehensive set of features. We benchmark Cumulus on the Human Cell Atlas Census of Immune Cells dataset of bone marrow cells and show that it substantially improves efficiency over conventional frameworks, while maintaining or improving the quality of results, enabling large-scale studies.

Disabled Homolog 2 Controls Prometastatic Activity of Tumor-Associated Macrophages

Tumor-associated macrophages (TAM) are regulators of extracellular matrix (ECM) remodeling and metastatic progression, the main cause of cancer-associated death. We found that disabled homolog 2 mitogen-responsive phosphoprotein (DAB2) is highly expressed in tumor-infiltrating TAMs and that its genetic ablation significantly impairs lung metastasis formation. DAB2-expressing TAMs, mainly localized along the tumor-invasive front, participate in integrin recycling, ECM remodeling, and directional migration in a tridimensional matrix. DAB2⁺ macrophages escort the invasive dissemination of cancer cells by a mechanosensing pathway requiring the transcription factor YAP. In human lobular breast and gastric carcinomas, DAB2⁺ TAMs correlated with a poor clinical outcome, identifying DAB2 as potential prognostic biomarker for stratification of patients with cancer. DAB2 is therefore central for the prometastatic activity of TAMs. SIGNIFICANCE: DAB2 expression in macrophages is essential for metastasis formation but not primary tumor growth. Mechanosensing cues, activating the complex YAP-TAZ, regulate DAB2 in macrophages, which in turn controls integrin recycling and ECM remodeling in 3-D tissue matrix. The presence of DAB2⁺ TAMs in patients with cancer correlates with worse prognosis.This article is highlighted in the In This Issue feature, p. 1611.

A large peptidome dataset improves HLA class I epitope prediction across most of the human population

Prediction of HLA epitopes is important for the development of cancer immunotherapies and vaccines. However, current prediction algorithms have limited predictive power, in part because they were not trained on high-quality epitope datasets covering a broad range of HLA alleles. To enable prediction of endogenous HLA class I-associated peptides across a large fraction of the human population, we used mass spectrometry to profile >185,000 peptides eluted from 95 HLA-A, -B, -C and -G mono-allelic cell lines. We identified canonical peptide motifs per HLA allele, unique and shared binding submotifs across alleles and distinct motifs associated with different peptide lengths. By integrating these data with transcript abundance and peptide processing, we developed HLAthena, providing allele-and-length-specific and pan-allele-pan-length prediction models for endogenous peptide presentation. These models predicted endogenous HLA class I-associated ligands with 1.5-fold improvement in positive predictive value compared with existing tools and correctly identified >75% of HLA-bound peptides that were observed experimentally in 11 patient-derived tumor cell lines.

Systems Immunology: Learning the Rules of the Immune System

Given the many cell types and molecular components of the human immune system, along with vast variations across individuals, how should we go about developing causal and predictive explanations of immunity? A central strategy in human studies is to leverage natural variation to find relationships among variables, including DNA variants, epigenetic states, immune phenotypes, clinical descriptors, and others. Here, we focus on how natural variation is used to find patterns, infer principles, and develop predictive models for two areas: (a) immune cell activation-how single-cell profiling boosts our ability to discover immune cell types and states-and (b) antigen presentation and recognition-how models can be generated to predict presentation of antigens on MHC molecules and their detection by T cell receptors. These are two examples of a shift in how we find the drivers and targets of immunity, especially in the human system in the context of health and disease.

Abstract 566: Neoantigens from translated unannotated open reading frames in cancer

Somatic mutations in cancer cells can generate neoantigens which can be recognized by immune cells and trigger an immune response. Patients vaccinated with neoantigen-based peptides display expanded neoantigen-specific T cells, suggesting that this could be a promising avenue for cancer treatment. Currently, neoantigen predictions are based on mutations detected by whole exome sequencing, which covers a pre-determined set of annotated exons, and often falls short for cancers with few somatic mutations.

Ribosome profiling (Ribo-seq), which allows to monitor mRNA translation, has predicted a plethora of translated novel unannotated ORFs (nuORFs). Here we hypothesized that nuORFs can provide another source of neoantigens in cancer cells. In particular, we focused on nuORFs in the following categories: 1) nuORFs expressed in healthy and cancer cells, that have acquired tumor-specific somatic mutations; 2) nuORFs upregulated in or specific to cancer cells.

To explore this hypothesis, we performed Ribo-seq on primary healthy and cancer cells and cell lines from melanoma, glioblastoma, colon carcinoma and chronic lymphocytic leukemia. Using this extensive dataset, we performed hierarchical ORF prediction analysis to build a database of highest confidence predicted translated nuORFs across healthy and cancer cell types.

To determine if peptides from nuORFs can be a source of antigens, we searched our collection of mono-allelic MHC class I immunopeptidome mass spectrometry (MS) spectra from 94 common HLA alleles against our pan-tissue nuORF database. Additionally, we performed MHC class I immunoprecipitation followed by MS on the same cells used for Ribo-seq. We found HLA-presented unmutated peptides derived from thousands of nuORFs, found within, but out-of-frame with annotated protein-coding ORFs, within 5’ and 3’ untranslated regions of annotated protein-coding transcripts, long non-coding RNAs (lncRNAs), pseudogenes, and other RNA species. The HLA-binding motifs of peptides from nuORFs correspond to the expected motifs for given HLA types, indicating that 1) nuORFs are translated and 2) nuORF-derived peptides are presented on MHC I.

To identify tumor-specific somatic mutations in nuORFs, we performed whole genome sequencing on patient-matched healthy and cancer cells and mapped somatic mutations to annotated ORFs and nuORFs. Finally, to identify nuORFs upregulated in or specific to cancer cells, we compared translation levels of nuORFs between healthy and cancer cells of the same origin. We found translated nuORFs with cancer-specific somatic mutations and nuORFs highly upregulated in and specific to cancer cells, suggesting that they can give rise to neoantigens.

In conclusion, nuORFs are translated, contribute peptides to MHC I presentation, acquire somatic mutations, are expressed in tissue- and cancer-dependent manner and should be considered in the search for neoantigens in cancer.

Citation Format: Tamara Ouspenskaia, Travis E. Law, Karl R. Clauser, Susan Klaeger, Derin B. Keskin, Bo Li, Elena Christian, Yuen Ting Chow, Phuong M. Le, Joshua Gould, Zhe Ji, Wandi Zhang, Pavan Bachireddy, Siranush Sarkizova, Nir Hacohen, Steven A. Carr, Catherine J. Wu, Aviv Regev. Neoantigens from translated unannotated open reading frames in cancer [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 566.

Two distinct colonic CD14+ subsets characterized by single-cell RNA profiling in Crohn's disease

Inflammatory bowel diseases are associated with dysregulated immune responses in the intestinal tissue. Four molecularly identified macrophage subsets control immune homeostasis in healthy gut. However, the specific roles and transcriptomic profiles of the phenotypically heterogeneous CD14⁺ macrophage-like population in inflamed gut remain to be investigated in Crohn's disease (CD). Here we identified two phenotypically, morphologically and functionally distinct colonic HLADR⁺SIRPα⁺CD14⁺ subpopulations that were further characterized using single-cell RNA-sequencing (scRNAseq) in CD. Frequencies of CD64^hiCD163^-/dim cells selectively augmented in inflamed colon and correlated with endoscopic score of disease severity. IL-1β and IL-23-producing CD64^hiCD163^-/dim cells predominated over TNF-α-producing CD64^hiCD163^hi cells in lesions. Purified "inflammatory monocyte-like" CD163^-, but not "macrophage-like" CD163^hi cells, through IL-1β, promoted Th17/Th1 but not Th1 responses in tissue memory CD4⁺T cells. Unsupervised scRNAseq analysis that captures the entire HLADR⁺SIRPα⁺ population revealed six clusters, two of which were enriched in either CD163^- or CD163^hi cells, and best defined by TREM1/FCAR/FCN1/IL1RN or CD209/MERTK/MRCI/CD163L1 genes, respectively. Selected newly identified discriminating markers were used beyond CD163 to isolate cells that shared pro-Th17/Th1 function with CD163^- cells. In conclusion, a molecularly distinct pro-inflammatory CD14⁺ subpopulation accumulates in inflamed colon, drives intestinal inflammatory T-cell responses, and thus, might contribute to CD disease severity.

Abstract A010: Personalized neoantigen-targeting vaccines for high-risk melanoma generate epitope spreading

Cancer vaccines have been envisioned as a key tool for generating effective cancer therapy. Tumor neoantigens are ideal targets because of their exquisite tumor-specific expression (arising from somatic mutations of the tumor) and high level of immunogenicity (lacking of central tolerance against them). Recently, we and others have demonstrated that personalized neoantigen-targeting vaccines are safe, feasible and highly immunogenic in phase I trials of stage III/IV resected high-risk melanoma (Ott & Hu, Nature 2017; Sahin, Nature 2017). Our neoantigen vaccine (NeoVax), consisting of up to 20 long peptides and poly-ICLC, induced strong polyfunctional neoantigen-specific T-cells that recognized patient tumor in vitro. In addition, 2 patients who were vaccinated and received anti-PD1 checkpoint blockade (CPB) therapy upon relapse had durable complete responses (CRs). Thus far, these vaccine studies have been performed in the adjuvant setting, preventing direct assessment of on-target tumor killing in vivo due to the lack of evaluable tumor. On the other hand, the detection of epitope spreading (the broadening of the immune response from the initially targeted epitope to others) would indirectly suggest therapy-induced tumor lysis, whereby the release of additional tumor antigens leads to further tumor-specific T-cell activation. To explore the hypothesis that NeoVax+/- CPB generates epitope spreading, we evaluated the T-cell responses against neoantigens and tumor associated antigens (TAAs) that were not included in the original vaccine in 3 patients. We performed experiments for a patient with stage III melanoma who has remained disease-free (Pt.3) after vaccination and 2 patients with resected stage IV disease who recurred but achieved CR after CPB (Pts. 2&6). For the assessment of CD8+ T-cell responses, we designed 9-10 aa epitope length peptides (predicted by NetMHCpan and/or a mass spectrometry [MS]-based prediction algorithm (Abelin, Immunity 2017) or detected physically on the tumor’s surface class I complexes by MS) arising from 3 categories of antigens: (i) neoantigen peptides; (ii) TAA peptides based on high tumor gene expression; (iii) TAA peptides, detected on the tumor by MS (available for 2 of the 3 patients). For testing of CD4+ T-cell responses, we designed 15-16 aa peptides that spanned predicted neoepitopes from category i. Per patient, we designed peptides against up to 70 genes (~20 for each category). PBMCs from pre- , week 16 post-vaccination and post-CPB were stimulated with peptide pools (~10 peptides/pool) for 2 weeks, followed by restimulation with individual peptides in IFN-γ ELISPOT assays to deconvolute the peptides. Thus far, we have tested CD8+ T-cells against 71 neoantigens (category i) and 22 TAAs (ii) from Pts. 2 and 6, and CD4+ T-cells against 30 neoantigens from all 3 patients. We identified CD4+ T-cells specific for 3 peptides (mut-AGAP3 [Pt.2], -EYA3 and -P2RY4 [Pt.3]) in the week 16 samples that were not included in the original respective vaccines; these populations were expanded only post, but not pre-vaccination. For Pt.2, an additional CD4+ T-cell response against a different neoantigen peptide derived from mut-AGAP3 was detected only after CPB therapy. Lastly, all four lines of CD4+ T-cells reactive against these identified neoantigens were able to discriminate between the mutated and wild-type forms of the peptides, suggesting tumor specificity and lack of cross reactivity with normal tissues. Therefore, our results demonstrate that epitope spreading occurred in 2 patients after vaccination, and further spreading was detected in one of the two following CPB therapy. Ongoing studies are focused on screening additional peptides and investigating the association of epitope spreading and any residual tumor burden. The newly activated antigen-specific T-cells can target additional tumor antigens provided by epitope spreading, thus potentially enhancing therapeutic efficacy.

Citation Format: Zhuting Hu, Donna Leet, Siranush Sarkizova, Rebecca Holden, Jing Sun, Susan Klaeger, Karl R. Clauser, Sachet A. Shukla, Wandi Zhang, Steven A. Carr, Edward F. Fritsch, Bradley L. Pentelute, Nir Hacohen, Derin B. Keskin, Patrick A. Ott, Catherine J. Wu. Personalized neoantigen-targeting vaccines for high-risk melanoma generate epitope spreading [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 A010.

Abstract B042: Broad analysis and more accurate predictions of HLA class I epitope binding in 92 common HLA alleles profiled by mono-allelic mass spectrometry

Introduction: Cancer vaccine therapies rely on accurate personalized selection of immunizing peptides in order to potentiate tumor-specific immune responses against neoepitopes derived from somatic mutations. Given the unique accumulation of mutations in each tumor as well as the patient’s particular complement of HLA class I alleles, the ability to accurately predict which epitopes will be presented by tumor cells is a fundamental prerequisite for successful vaccine design. By utilizing a mono-allelic mass spectrometry (MS) strategy for profiling the endogenous HLA class I peptidome, we recently showed that prediction of endogenous presentation can be drastically improved when model training integrates peptide sequence along with intracellular signals such as likelihood of proteasomal processing and peptide abundance. Yet the limited set of mono-allelic data did not allow for deep comparative analysis across HLA- A, B, and C alleles, which can better inform pan-allele predictor design. Moreover, the significant variability in per-allele model performance remains unexplained. Methods: We recently developed a scalable mono-allelic MS technique to profile naturally presented peptides on HLA molecules, whereby the HLA class I deficient B721.221 cell line is transfected with HLA expression vectors coding for a single allele of interest and eluted HLA peptides are analyzed by LC-MS/MS. In addition, endogenously presented antigens on primary tumor-derived cell lines from 4 melanoma patients were also identified via MS. To extract knowledge from this unique dataset, we implemented computational tools to summarize, visualize, and compare the characteristics of HLA- A, B, C, and G alleles and developed a novel approach to define allele similarity that takes into account the collection of sub-motifs per allele. We trained neural network prediction models, validated their performance on internal and external datasets, and analyzed the variability in performance across alleles. Results: To date, we have generated binding data for 92 HLA- A, B, C and G alleles, identifying more than 190,000 peptides and covering the most frequent alleles in the population. Extensive mono-allelic profiling revealed that some alleles present non-9-mer peptides with high frequency. The availability of large number of non-9-mer peptides allowed us to build length-specific models that often performed better than the corresponding non-length-specific models currently used. We observe that HLA- A and B alleles present more peptides of length 10 and 11 than C alleles, while C alleles have a higher propensity for 8-mers. Correlation-based analysis of binding motifs revealed that HLA-A and B motifs are more specific whereas C motifs are less stringent and thus share more overlapping binders. Since binding data are available only for a fraction of all known alleles, pan-allele models implicitly embed allele similarity to predict for uncharacterized alleles based on the sequence of the binding pocket. By clustering allele-specific peptides into sub-motifs, we propose a novel explicit approach to delineate allele similarity at finer granularity that can improve pan-allele model design. We show that our allele-specific models are better at discriminating tumor-presented epitopes than state of the art predictors and investigate the relationship between false discovery rate and natural abundance of anchor residues to better understand differences in model accuracy amongst alleles. Finally, deconvolution of tumor-presented peptides demonstrated that ~10% of peptides are presented on HLA-C, which has been historically understudied. Conclusions: We have vastly expanded the collection of endogenous HLA-specific peptides deriving biologic insights into the principles of epitope presentations and valuable considerations for prediction model design and epitope selection for tumor vaccines.

Citation Format: Siranush Sarkizova, Susan Klaeger, Derin B. Keskin, Karl Clauser, Hasmik Keshishian, Christina R. Hartigan, Nir Hacohen, Steven A. Carr, Catherine J. Wu. Broad analysis and more accurate predictions of HLA class I epitope binding in 92 common HLA alleles profiled by mono-allelic mass spectrometry [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 B042.

Systematic profiling of HLA-C epitopes

Accurate predictive algorithms have required elucidation of HLA-peptide binding motifs from extensive HLA binding peptide ligand databases. Historically, however, little attention has been paid to HLA-C because of its perceived lower surface expression, even as newer studies have identified important HLA-C restricted viral epitopes. Robust predictive algorithms for HLA-C are thus lacking and yet, if improved, could markedly expand the spectrum of immune targets contributing to protective immune responses. We generated 21 distinct single HLA-C allele-expressing cell lines and subjected them to immunoaffinity purification and MS-based sequencing. From 33851 peptides identified from the 21 HLA-C expressing cell lines, we discovered novel anchor motifs. These data were used to train machine-learning algorithms to generate an accurate and novel HLA-C epitope prediction tool. Finally, we used our novel algorithm to predict HLA-C restricted neo-epitopes out of tumor biopsies.

Aryl Hydrocarbon Receptor Controls Monocyte Differentiation into Dendritic Cells versus Macrophages

After entering tissues, monocytes differentiate into cells that share functional features with either macrophages or dendritic cells (DCs). How monocyte fate is directed toward monocyte-derived macrophages (mo-Macs) or monocyte-derived DCs (mo-DCs) and which transcription factors control these differentiation pathways remains unknown. Using an in vitro culture model yielding human mo-DCs and mo-Macs closely resembling those found in vivo in ascites, we show that IRF4 and MAFB were critical regulators of monocyte differentiation into mo-DCs and mo-Macs, respectively. Activation of the aryl hydrocarbon receptor (AHR) promoted mo-DC differentiation through the induction of BLIMP-1, while impairing differentiation into mo-Macs. AhR deficiency also impaired the in vivo differentiation of mouse mo-DCs. Finally, AHR activation correlated with mo-DC infiltration in leprosy lesions. These results establish that mo-DCs and mo-Macs are controlled by distinct transcription factors and show that AHR acts as a molecular switch for monocyte fate specification in response to micro-environmental factors.

Mass Spectrometry Profiling of HLA-Associated Peptidomes in Mono-allelic Cells Enables More Accurate Epitope Prediction

Identification of human leukocyte antigen (HLA)-bound peptides by liquid chromatography-tandem mass spectrometry (LC-MS/MS) is poised to provide a deep understanding of rules underlying antigen presentation. However, a key obstacle is the ambiguity that arises from the co-expression of multiple HLA alleles. Here, we have implemented a scalable mono-allelic strategy for profiling the HLA peptidome. By using cell lines expressing a single HLA allele, optimizing immunopurifications, and developing an application-specific spectral search algorithm, we identified thousands of peptides bound to 16 different HLA class I alleles. These data enabled the discovery of subdominant binding motifs and an integrative analysis quantifying the contribution of factors critical to epitope presentation, such as protein cleavage and gene expression. We trained neural-network prediction algorithms with our large dataset (>24,000 peptides) and outperformed algorithms trained on datasets of peptides with measured affinities. We thus demonstrate a strategy for systematically learning the rules of endogenous antigen presentation.

Abstract LB-179: Next-generation epitope prediction using mass spectrometry and integrative genomics

Personalized neoantigen therapies for cancer require accurate epitope selection. However, most Class I prediction algorithms in common use today are based on biochemical binding assays that are difficult to scale and do not address processing steps upstream of peptide-MHC binding. Here we present an alternative approach based on the LC-MS/MS identification of MHC Class I-bound peptides. While MS-based profiling is not new, we optimized the system for rule learning by focusing on cell lines expressing only a single HLA-A or HLA-B allele and by collecting parallel transcriptomic and proteomic measurements. Identifying over 24,000 peptides across 16 individual alleles, we were able to discover novel binding motifs, which were validated biochemically, and develop novel neural network algorithms. Furthermore, we systematically interrogated processing rules - discovering a novel motif conserved across multiple cell types - and developed a principled framework for integrating epitope cleavability, expression, and MHC binding potential into an overall ranking. Validating on external datasets, we saw a doubling in positive predictive value with respect to standard approaches. We thus demonstrate a scalable strategy for systematically learning the rules of endogenous antigen presentation that can be deployed for the optimal selection of patient-specific cancer neoantigens.

Citation Format: Michael S. Rooney, Jenn Abelin, Siranush Sarkizova, Derin Keskin, Christine Hartigan, Wandi Zhang, John Sidney, William Lane, Jonathan Stevens, Guang L. Zhang, Karl Clauser, Nir Hacohen, Steve Carr, Cathy Wu. Next-generation epitope prediction using mass spectrometry and integrative genomics [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 LB-179. doi:10.1158/1538-7445.AM2017-LB-179

Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors

Dendritic cells (DCs) and monocytes play a central role in pathogen sensing, phagocytosis, and antigen presentation and consist of multiple specialized subtypes. However, their identities and interrelationships are not fully understood. Using unbiased single-cell RNA sequencing (RNA-seq) of ~2400 cells, we identified six human DCs and four monocyte subtypes in human blood. Our study reveals a new DC subset that shares properties with plasmacytoid DCs (pDCs) but potently activates T cells, thus redefining pDCs; a new subdivision within the CD1C+ subset of DCs; the relationship between blastic plasmacytoid DC neoplasia cells and healthy DCs; and circulating progenitor of conventional DCs (cDCs). Our revised taxonomy will enable more accurate functional and developmental analyses as well as immune monitoring in health and disease.

Abstract B089: High-throughput profiling of HLA allele-specific peptides by MS for improved epitope prediction

Cancer mutations yield neo-antigens, which are instrumental to immune-mediated recognition and control of cancer. Vaccine-based therapies targeting neo-antigens will require accurate prediction of which mutations yield peptides presented on polymorphic HLA class I. While in vitro methods have produced increasingly accurate predictors of peptide:MHC binding, there remains a need to define rules for endogenous antigen presentation. Here, we use rapid, high-resolution liquid chromatography mass spectrometry (LC-MS/MS) to identify >24,000 peptides associated with 16 HLA alleles in B cell lines that each express a single HLA allele. The elution of peptides from single HLA alleles allowed us to develop improved rules for endogenous peptide presentation based on the physicochemical properties of binding peptides, patterns of peptide cleavage and abundance of cognate transcripts. Finally, we trained models that integrated MS-derived peptide data and gene expression and demonstrate improved prediction of endogenous peptide presentation in independent datasets. Our strategy thus improves the performance of current predictive algorithms and provides a rapid and scalable method to generate rules for the massive and diverse set of human HLA alleles.

Citation Format: Michael S. Rooney, Jennifer G. Abelin, Derin B. Keskin, Siranush Sarkizova, Christina Hartigan, Wandi Zhang, John Sidney, Jonathan Stevens, William J. Lane, Guang L. Zhang, Karl R. Clauser, Nir Hacohen, Steven A. Carr, Catherine J. Wu. High-throughput profiling of HLA allele-specific peptides by MS for improved epitope prediction [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 B089.