Atypical acute myeloid leukemia-specific transcripts generate shared and immunogenic MHC class-I-associated epitopes.

Acute myeloid leukemia (AML) has not benefited from innovative immunotherapies, mainly because of the lack of actionable immune targets. Using an original proteogenomic approach, we analyzed the major histocompatibility complex class I (MHC class I)-associated immunopeptidome of 19 primary AML samples and identified 58 tumor-specific antigens (TSAs). These TSAs bore no mutations and derived mainly (86%) from supposedly non-coding genomic regions. Two AML-specific aberrations were instrumental in the biogenesis of TSAs, intron retention, and epigenetic changes. Indeed, 48% of TSAs resulted from intron retention and translation, and their RNA expression correlated with mutations of epigenetic modifiers (e.g., DNMT3A). AML TSA-coding transcripts were highly shared among patients and were expressed in both blasts and leukemic stem cells. In AML patients, the predicted number of TSAs correlated with spontaneous expansion of cognate T cell receptor clonotypes, accumulation of activated cytotoxic T cells, immunoediting, and improved survival. These TSAs represent attractive targets for AML immunotherapy.

The cryptic immunopeptidome in health and disease.

Peptides presented by MHC proteins regulate all aspects of T cell biology. These MHC-associated peptides (MAPs) form what is known as the immunopeptidome and their comprehensive analysis has catalyzed the burgeoning field of immunopeptidomics. Advances in mass spectrometry (MS) and next-generation sequencing have facilitated significant breakthroughs in this area, some of which are highlighted in this article on the cryptic immunopeptidome. Here, 'cryptic' refers to peptides and proteins encoded by noncanonical open reading frames (ORFs). Cryptic MAPs derive mainly from short unstable proteins found in normal, infected, and neoplastic cells. Cryptic MAPs show minimal overlap with cryptic proteins found in whole-cell extracts. In many cancer types, most cancer-specific MAPs are cryptic.

Regulation of MHC Class I Expression in Lung Epithelial Cells during Inflammation.

Lung infections are a perennial leading cause of death worldwide. The lung epithelium comprises three main cell types: alveolar type I (AT1), alveolar type II (AT2), and bronchiolar cells. Constitutively, these three cell types express extremely low amounts of surface MHC class I (MHC I) molecules, that is, <1% of levels found on medullary thymic epithelial cells (ECs). We report that inhalation of the TLR4 ligand LPS upregulates cell surface MHC I by ∼25-fold on the three subtypes of mouse lung ECs. This upregulation is dependent on Nlrc5, Stat1, and Stat2 and caused by a concerted production of the three IFN families. It is nevertheless hampered, particularly in AT1 cells, by the limited expression of genes instrumental in the peptide loading of MHC I molecules. Genes involved in production and response to cytokines and chemokines were selectively induced in AT1 cells. However, discrete gene subsets were selectively downregulated in AT2 or bronchiolar cells following LPS inhalation. Genes downregulated in AT2 cells were linked to cell differentiation and cell proliferation, and those repressed in bronchiolar cells were primarily involved in cilium function. Our study shows a delicate balance between the expression of transcripts maintaining lung epithelium integrity and transcripts involved in Ag presentation in primary lung ECs.

Immunopeptidomic analyses of colorectal cancers with and without microsatellite instability.

Colorectal cancer is the second leading cause of cancer death worldwide, and the incidence of this disease is expected to increase as global socioeconomic changes occur. Immune checkpoint inhibition therapy is effective in treating a minority of colorectal cancer tumors; however, microsatellite stable tumors do not respond well to this treatment. Emerging cancer immunotherapeutic strategies aim to activate a cytotoxic T cell response against tumor-specific antigens, presented exclusively at the cell surface of cancer cells. These antigens are rare and are most effectively identified with a mass spectrometry-based approach, which allows the direct sampling and sequencing of these peptides. Although the few tumor-specific antigens identified to date are derived from coding regions of the genome, recent findings indicate that a large proportion of tumor-specific antigens originate from allegedly noncoding regions. Here, we employed a novel proteogenomic approach to identify tumor antigens in a collection of colorectal cancer-derived cell lines and biopsy samples consisting of matched tumor and normal adjacent tissue. The generation of personalized cancer databases paired with mass spectrometry analyses permitted the identification of more than 30,000 unique MHC I-associated peptides. We identified 19 tumor-specific antigens in both microsatellite stable and unstable tumors, over two-thirds of which were derived from noncoding regions. Many of these peptides were derived from source genes known to be involved in colorectal cancer progression, suggesting that antigens from these genes could have therapeutic potential in a wide range of tumors. These findings could benefit the development of T cell-based vaccines, in which T cells are primed against these antigens to target and eradicate tumors. Such a vaccine could be used in tandem with existing immune checkpoint inhibition therapies, to bridge the gap in treatment efficacy across subtypes of colorectal cancer with varying prognoses. Data are available via ProteomeXchange with identifier PXD028309.

OpenCustomDB: Integration of Unannotated Open Reading Frames and Genetic Variants to Generate More Comprehensive Customized Protein Databases.

Proteomic diversity in biological samples can be characterized by mass spectrometry (MS)-based proteomics using customized protein databases generated from sets of transcripts previously detected by RNA-seq. This diversity has only been increased by the recent discovery that many translated alternative open reading frames rest unannotated at unsuspected locations of mRNAs and ncRNAs. These novel protein products, termed alternative proteins, have been left out of all previous custom database generation tools. Consequently, genetic variations that impact alternative open reading frames and variant peptides from their translated proteins are not detectable with current computational workflows. To fill this gap, we present OpenCustomDB, a bioinformatics tool that uses sample-specific RNaseq data to identify genomic variants in canonical and alternative open reading frames, allowing for more than one coding region per transcript. In a test reanalysis of a cohort of 16 patients with acute myeloid leukemia, 5666 peptides from alternative proteins were detected, including 201 variant peptides. We also observed that a significant fraction of peptide-spectrum matches previously assigned to peptides from canonical proteins got better scores when reassigned to peptides from alternative proteins. Custom protein libraries that include sample-specific sequence variations of all possible open reading frames are promising contributions to the development of proteomics and precision medicine. The raw and processed proteomics data presented in this study can be found in PRIDE repository with accession number PXD029240.

Identification of antigenic epitopes recognized by tumor infiltrating lymphocytes in high grade serous ovarian cancer by multi-omics profiling of the auto-antigen repertoire.

Immunotherapeutic strategies aimed at enhancing tumor cell killing by tumor-specific T cells hold great potential for reducing tumor burden and prolonging survival of cancer patients. Although many potential tumor antigens have been described, identifying relevant targets when designing anti-cancer vaccines or targeted cell therapies remains a challenge. To identify novel, potentially immunogenic candidate tumor antigens, we performed integrated tumor transcriptomic, seromic, and proteomic analyses of high grade serous ovarian cancer (HGSC) patient tumor samples. We identified tumor neo-antigens and over-expressed antigens using whole exome and RNA sequencing and examined these in relation to patient-matched auto-antibody repertoires. Focusing on MHC class I epitopes recognized by CD8+ T cells, HLA-binding epitopes were identified or predicted from the highly expressed, mutated, or auto-antibody target antigen, or MHC-associated peptides (MAPs). Recognition of candidate antigenic peptides was assessed within the tumor-infiltrating T lymphocyte (TIL) population expanded from each patient. Known tumor-associated antigens (TAA) and cancer/testis antigens (CTA) were commonly found in the auto-antibody and MAP repertoires and CD8+ TILs recognizing epitopes from these antigens were detected, although neither expression level nor the presence of auto-antibodies correlated with TIL recognition. Auto-antibodies against tumor-mutated antigens were found in most patients, however, no TIL recognition of the highest predicted affinity neo-epitopes was detected. Using high expression level, auto-antibody recognition, and epitope prediction algorithms, we identified epitopes in 5 novel antigens (MOB1A, SOCS3, TUBB, PRKAR1A, CCDC6) recognized by HGSC patient TILs. Furthermore, selection of epitopes from the MAP repertoire identified 5 additional targets commonly recognized by multiple patient TILs. We find that the repertoire of TIL specificities includes recognition of highly expressed and immunogenic self-antigens that are processed and presented by tumors. These results indicate an ongoing autoimmune response against a range of self-antigens targeted by HGSC TILs.

CAMAP: Artificial neural networks unveil the role of codon arrangement in modulating MHC-I peptides presentation.

MHC-I associated peptides (MAPs) play a central role in the elimination of virus-infected and neoplastic cells by CD8 T cells. However, accurately predicting the MAP repertoire remains difficult, because only a fraction of the transcriptome generates MAPs. In this study, we investigated whether codon arrangement (usage and placement) regulates MAP biogenesis. We developed an artificial neural network called Codon Arrangement MAP Predictor (CAMAP), predicting MAP presentation solely from mRNA sequences flanking the MAP-coding codons (MCCs), while excluding the MCC per se. CAMAP predictions were significantly more accurate when using original codon sequences than shuffled codon sequences which reflect amino acid usage. Furthermore, predictions were independent of mRNA expression and MAP binding affinity to MHC-I molecules and applied to several cell types and species. Combining MAP ligand scores, transcript expression level and CAMAP scores was particularly useful to increase MAP prediction accuracy. Using an in vitro assay, we showed that varying the synonymous codons in the regions flanking the MCCs (without changing the amino acid sequence) resulted in significant modulation of MAP presentation at the cell surface. Taken together, our results demonstrate the role of codon arrangement in the regulation of MAP presentation and support integration of both translational and post-translational events in predictive algorithms to ameliorate modeling of the immunopeptidome.

IFN-λ Enhances Constitutive Expression of MHC Class I Molecules on Thymic Epithelial Cells.

Regulation of MHC class I (MHC I) expression has been studied almost exclusively in hematolymphoid cells. We report that thymic epithelial cells (TECs), particularly the medullary TECs, constitutively express up to 100-fold more cell surface MHC I proteins than epithelial cells (ECs) from the skin, colon, and lung. Differential abundance of cell surface MHC I in primary ECs is regulated via transcription of MHC I and of genes implicated in the generation of MHC I-binding peptides. Superior MHC I expression in TECs is unaffected by deletion of Ifnar1 or Ifngr1, but is lessened by deletion of Aire, Ifnlr1, Stat1, or Nlrc5, and is driven mainly by type III IFN produced by medullary TECs. Ifnlr1 -/- mice show impaired negative selection of CD8 thymocytes and, at 9 mo of age, present autoimmune manifestations. Our study shows unanticipated variation in MHC I expression by ECs from various sites and provides compelling evidence that superior expression of MHC I in TECs is crucial for proper thymocyte education.

Factorized embeddings learns rich and biologically meaningful embedding spaces using factorized tensor decomposition.

MOTIVATION: The recent development of sequencing technologies revolutionized our understanding of the inner workings of the cell as well as the way disease is treated. A single RNA sequencing (RNA-Seq) experiment, however, measures tens of thousands of parameters simultaneously. While the results are information rich, data analysis provides a challenge. Dimensionality reduction methods help with this task by extracting patterns from the data by compressing it into compact vector representations. RESULTS: We present the factorized embeddings (FE) model, a self-supervised deep learning algorithm that learns simultaneously, by tensor factorization, gene and sample representation spaces. We ran the model on RNA-Seq data from two large-scale cohorts and observed that the sample representation captures information on single gene and global gene expression patterns. Moreover, we found that the gene representation space was organized such that tissue-specific genes, highly correlated genes as well as genes participating in the same GO terms were grouped. Finally, we compared the vector representation of samples learned by the FE model to other similar models on 49 regression tasks. We report that the representations trained with FE rank first or second in all of the tasks, surpassing, sometimes by a considerable margin, other representations. AVAILABILITY AND IMPLEMENTATION: A toy example in the form of a Jupyter Notebook as well as the code and trained embeddings for this project can be found at: https://github.com/TrofimovAssya/FactorizedEmbeddings. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

PSMB11 Orchestrates the Development of CD4 and CD8 Thymocytes via Regulation of Gene Expression in Cortical Thymic Epithelial Cells.

T cell development depends on sequential interactions of thymocytes with cortical thymic epithelial cells (cTECs) and medullary thymic epithelial cells. PSMB11 is a catalytic proteasomal subunit present exclusively in cTECs. Because proteasomes regulate transcriptional activity, we asked whether PSMB11 might affect gene expression in cTECs. We report that PSMB11 regulates the expression of 850 cTEC genes that modulate lymphostromal interactions primarily via the WNT signaling pathway. cTECs from Psmb11 -/- mice 1) acquire features of medullary thymic epithelial cells and 2) retain CD8 thymocytes in the thymic cortex, thereby impairing phase 2 of positive selection, 3) perturbing CD8 T cell development, and 4) causing dramatic oxidative stress leading to apoptosis of CD8 thymocytes. Deletion of Psmb11 also causes major oxidative stress in CD4 thymocytes. However, CD4 thymocytes do not undergo apoptosis because, unlike CD8 thymocytes, they upregulate expression of chaperones and inhibitors of apoptosis. We conclude that PSMB11 has pervasive effects on both CD4 and CD8 thymocytes via regulation of gene expression in cTECs.

Immunogenic stress and death of cancer cells: Contribution of antigenicity vs adjuvanticity to immunosurveillance.

Cancer cells are subjected to constant selection by the immune system, meaning that tumors that become clinically manifest have managed to subvert or hide from immunosurveillance. Immune control can be facilitated by induction of autophagy, as well as by polyploidization of cancer cells. While autophagy causes the release of ATP, a chemotactic signal for myeloid cells, polyploidization can trigger endoplasmic reticulum stress with consequent exposure of the "eat-me" signal calreticulin on the cell surface, thereby facilitating the transfer of tumor antigens into dendritic cells. Hence, both autophagy and polyploidization cause the emission of adjuvant signals that ultimately elicit immune control by CD8+ T lymphocytes. We investigated the possibility that autophagy and polyploidization might also affect the antigenicity of cancer cells by altering the immunopeptidome. Mass spectrometry led to the identification of peptides that were presented on major histocompatibility complex (MHC) class I molecules in an autophagy-dependent fashion or that were specifically exposed on the surface of polyploid cells, yet lost upon passage of such cells through immunocompetent (but not immunodeficient) mice. However, the preferential recognition of autophagy-competent and polyploid cells by the innate and cellular immune systems did not correlate with the preferential recognition of such peptides in vivo. Moreover, vaccination with such peptides was unable to elicit tumor growth-inhibitory responses in vivo. We conclude that autophagy and polyploidy increase the immunogenicity of cancer cells mostly by affecting their adjuvanticity rather than their antigenicity.

MAPDP: A Cloud-Based Computational Platform for Immunopeptidomics Analyses.

The immunopeptidome corresponds to the repertoire of peptides presented at the cell surface by the major histocompatibility complex (MHC) molecules. Cytotoxic T cells scan this repertoire to identify nonself antigens that can arise from tumors or infected cells. The identification of actionable antigenic targets is key to the development of therapeutic cancer vaccines, T-cell therapy, and other T-cell receptor-based biologics. The growing clinical interest for immunopeptidomics has accelerated the development of high throughput proteogenomic platforms that provide a system-level analysis of MHC-associated peptides. Improvement in sensitivity and throughput of mass spectrometers now allows the detection of a few thousands of peptides from less than 100 million cells. To manage the amount of data generated by these instruments, we have developed the MHC-associated peptide discovery platform (MAPDP), a novel open-source cloud-based computational platform for immunopeptidomic analyses. It provides convenient access from a web portal to immunopeptidomes stored in the database, filtering tools, various visualizations, annotations (e.g., IEDB, dbSNP, gnomAD), peptide-binding affinity prediction (mhcflurry, NetMHC), HLA genotyping, and the generation of personalized proteome databases. MAPDP functionalities are demonstrated here by the discovery of MHC peptides featuring new genetic variants identified in two previously published ovarian carcinoma data sets.

Major multilevel molecular divergence between THP-1 cells from different biorepositories.

The THP-1 cell line is broadly used as a model for acute myeloid leukemia (AML) with MLL fusion and to study monocyte differentiation and function. We studied THP-1 cells obtained from two major biorepositories. The two cell lines were closely related with a percentage match of short tandem repeat (STR) profiles ranging from 93.75% to 100%, depending on the algorithm used. Nevertheless, we found that the two cell lines presented discordant HLA type, cytogenetic aberrations and AML-related gene expression (including critical targets of MLL fusion). These discrepancies resulted mainly from loss of heterozygosity (LOH) involving five chromosomal regions. In view of their aberrant expression of key "leukemia" genes (e.g., LIN28B, MEIS1 and SPARC), we argue that one of the THP-1 cell lines may not be a reliable model for studying leukemia. Their defective expression of HLA molecules and abnormal adhesion properties is also a caveat for studies of antigen presentation. In a more general perspective, our findings show that seemingly minor discrepancies in STR profiles among cell lines may be the sign of major genetic drift, of sufficient magnitude to affect the reliability of cell line-based research.

Extending the Comprehensiveness of Immunopeptidome Analyses Using Isobaric Peptide Labeling.

Defining the repertoire of peptides presented by the major histocompatibility complex class I (MHC I) is a key step toward the identification of relevant antigens for cancer immunotherapy. However, the identification of cancer-specific antigens is a significant analytical challenge in view of their low abundance and low mutational load found in most primary cancer specimens. Here, we describe the application of isobaric peptide labeling with tandem mass tag (TMT) to improve the detection of the MHC I peptides. Isobaric peptide labeling was found to promote the formation of multiply charged ions and to enhance the formation of b-type fragment ions, thus resulting in a 50% improvement of MHC I peptide identification. The gain in sensitivity obtained using TMT labeling enabled the detection of low-abundance MHC I peptides including tumor-specific antigens (TSAs) and minor histocompatibility antigens (MiHAs). We further demonstrate the application of this approach to quantify MiHAs presented by B-cell lymphocytes and determined their expression levels by LC-MS/MS using both synchronous precursor selection (SPS) and high-field asymmetric waveform ion mobility spectrometry (FAIMS).

The SysteMHC Atlas project.

Mass spectrometry (MS)-based immunopeptidomics investigates the repertoire of peptides presented at the cell surface by major histocompatibility complex (MHC) molecules. The broad clinical relevance of MHC-associated peptides, e.g. in precision medicine, provides a strong rationale for the large-scale generation of immunopeptidomic datasets and recent developments in MS-based peptide analysis technologies now support the generation of the required data. Importantly, the availability of diverse immunopeptidomic datasets has resulted in an increasing need to standardize, store and exchange this type of data to enable better collaborations among researchers, to advance the field more efficiently and to establish quality measures required for the meaningful comparison of datasets. Here we present the SysteMHC Atlas (https://systemhcatlas.org), a public database that aims at collecting, organizing, sharing, visualizing and exploring immunopeptidomic data generated by MS. The Atlas includes raw mass spectrometer output files collected from several laboratories around the globe, a catalog of context-specific datasets of MHC class I and class II peptides, standardized MHC allele-specific peptide spectral libraries consisting of consensus spectra calculated from repeat measurements of the same peptide sequence, and links to other proteomics and immunology databases. The SysteMHC Atlas project was created and will be further expanded using a uniform and open computational pipeline that controls the quality of peptide identifications and peptide annotations. Thus, the SysteMHC Atlas disseminates quality controlled immunopeptidomic information to the public domain and serves as a community resource toward the generation of a high-quality comprehensive map of the human immunopeptidome and the support of consistent measurement of immunopeptidomic sample cohorts.

Allodepleted T-cell immunotherapy after haploidentical haematopoietic stem cell transplantation without severe acute graft-versus-host disease (GVHD) in the absence of GVHD prophylaxis.

Graft-versus-host disease (GVHD) is a major cause of transplant-related mortality (TRM) after allogeneic haematopoietic stem cell transplantation (HSCT) and presents a challenge in haploidentical HSCT. GVHD may be prevented by ex vivo graft T-cell depletion or in vivo depletion of proliferating lymphocytes. However, both approaches pose significant risks, particularly infections and relapse, compromising survival. A photodepletion strategy to eliminate alloreactive T cells from mismatched donor lymphocyte infusions (enabling administration without immunosuppression), was used to develop ATIR101, an adjunctive therapy for use after haploidentical HSCT. In this phase I dose-finding study, 19 adults (median age: 54 years) with high-risk haematological malignancies were treated with T-cell-depleted human leucocyte antigen-haploidentical myeloablative HSCT followed by ATIR101 at doses of 1 × 104 -5 × 106  CD3+  cells/kg (median 31 days post-transplant). No patient received post-transplant immunosuppression or developed grade III/IV acute GVHD, demonstrating the feasibility of ATIR101 infusion for evaluation in two subsequent phase 2 studies. Additionally, we report long-term follow -up of patients treated with ATIR101 in this study. At 1 year, all 9 patients receiving doses of 0·3-2 × 106  CD3+  cells/kg ATIR101 remained free of serious infections and after more than 8 years, TRM was 0%, relapse-related mortality was 33% and overall survival was 67% in these patients.

An Unbiased Linkage Approach Reveals That the p53 Pathway Is Coupled to NK Cell Maturation.

Natural killer cells constitute potent innate lymphoid cells that play a major role in both tumor immunosurveillance and viral clearance via their effector functions. A four-stage model of NK cell functional maturation has been established according to the expression of CD11b and CD27, separating mature NK (mNK) cells into distinct populations that exhibit specific phenotypic and functional properties. To identify genetic factors involved in the regulation of NK cell functional maturation, we performed a linkage analysis on F2 (B6.Rag1-/- × NOD.Rag1-/- intercross) mice. We identified six loci on chromosomes 2, 4, 7, 10, 11, and 18 that were linked to one or more mNK cell subsets. Subsequently, we performed an in silico analysis exploiting mNK cell subset microarray data, highlighting various genes and microRNAs as potential regulators of the functional maturation of NK cells. Together, the combination of our unbiased genetic linkage study and the in silico analysis positions genes known to affect NK cell biology along the specific stages of NK cell functional maturation. Moreover, this approach allowed us to uncover a novel candidate gene in the regulation of NK cell maturation, namely Trp53 Using mice deficient for Trp53, we confirm that this tumor suppressor regulates NK cell functional maturation. Additional candidate genes revealed in this study may eventually serve as targets for the modulation of NK cell functional maturation to potentiate both tumor immunosurveillance and viral clearance.

Exploiting non-canonical translation to identify new targets for T cell-based cancer immunotherapy.

Cryptic MHC I-associated peptides (MAPs) are produced via two mechanisms: translation of protein-coding genes in non-canonical reading frames and translation of allegedly non-coding sequences. In general, cryptic MAPs are coded by relatively short open reading frames whose translation can be regulated at the level of initiation, elongation or termination. In contrast to conventional MAPs, the processing of cryptic MAPs is frequently proteasome independent. The existence of cryptic MAPs derived from allegedly non-coding regions enlarges the scope of CD8 T cell immunosurveillance from a mere ~2% to as much as ~75% of the human genome. Considering that 99% of cancer-specific mutations are located in those allegedly non-coding regions, cryptic MAPs could furthermore represent a particularly rich source of tumor-specific antigens. However, extensive proteogenomic analyses will be required to determine the breath as well as the temporal and spatial plasticity of the cryptic MAP repertoire in normal and neoplastic cells.