An integrated approach to epitope analysis I: Dimensional reduction, visualization and prediction of MHC binding using amino acid principal components and regression approaches

High-throughput, targeted MHC class I immunopeptidomics using a functional genetics screening platform

Identification of CD8+ T cell epitopes is critical for the development of immunotherapeutics. Existing methods for major histocompatibility complex class I (MHC class I) ligand discovery are time intensive, specialized and unable to interrogate specific proteins on a large scale. Here, we present EpiScan, which uses surface MHC class I levels as a readout for whether a genetically encoded peptide is an MHC class I ligand. Predetermined starting pools composed of >100,000 peptides can be designed using oligonucleotide synthesis, permitting large-scale MHC class I screening. We exploit this programmability of EpiScan to uncover an unappreciated role for cysteine that increases the number of predicted ligands by 9-21%, reveal affinity hierarchies by analysis of biased anchor peptide libraries and screen viral proteomes for MHC class I ligands. Using these data, we generate and iteratively refine peptide binding predictions to create EpiScan Predictor. EpiScan Predictor performs comparably to other state-of-the-art MHC class I peptide binding prediction algorithms without suffering from underrepresentation of cysteine-containing peptides. Thus, targeted immunopeptidomics using EpiScan will accelerate CD8+ T cell epitope discovery toward the goal of individual-specific immunotherapeutics.

Determinants of tumor immune evasion: the role of T cell exposed motif frequency and mutant amino acid exposure

Few neoepitopes detected in tumor biopsies are immunogenic. Tumor-specific T cell responses require both the presentation of an epitope that differs from wildtype and the presence of T cells with neoepitope-cognate receptors. We show that mutations detected in tumor biopsies result in an increased frequency of rare amino acid combinations compared to the human proteome and gastrointestinal microorganisms. Mutations in a large data set of oncogene and tumor suppressor gene products were compared to wildtype, and to the count of corresponding amino acid motifs in the human proteome and gastrointestinal microbiome. Mutant amino acids in T cell exposed positions of potential neoepitopes consistently generated amino acid motifs that are less common in both proteome reference datasets. Approximately 10% of the mutant amino acid motifs are absent from the human proteome. Motif frequency does not change when mutants were positioned in the MHC anchor positions hidden from T cell receptors. Analysis of neoepitopes in GBM and LUSC cases showed less common T cell exposed motifs, and HLA binding preferentially placing mutant amino acids in an anchor position for both MHC I and MHC II. Cross-presentation of mutant exposed neoepitopes by MHC I and MHC II was particularly uncommon. Review of a tumor mutation dataset known to generate T cell responses showed immunogenic epitopes were those with mutant amino acids exposed to the T cell receptor and with exposed pentamer motifs present in the human and microbiome reference databases. The study illustrates a previously unrecognized mechanism of tumor immune evasion, as rare T cell exposed motifs produced by mutation are less likely to have cognate T cells in the T cell repertoire. The complex interactions of HLA genotype, binding positions, and mutation specific changes in T cell exposed motif underscore the necessity of evaluating potential neoepitopes in each individual patient.

Unraveling the structure and function of CdcPDE: A novel phosphodiesterase from Crotalus durissus collilineatus snake venom

This study reports the isolation, structural, biochemical, and functional characterization of a novel phosphodiesterase from Crotalus durissus collilineatus venom (CdcPDE). CdcPDE was successfully isolated from whole venom using three chromatographic steps and represented 0.7% of total protein content. CdcPDE was inhibited by EDTA and reducing agents, demonstrating that metal ions and disulfide bonds are necessary for its enzymatic activity. The highest enzymatic activity was observed at pH 8-8.5 and 37 °C. Kinetic parameters indicated a higher affinity for the substrate bis(p-nitrophenyl) phosphate compared to others snake venom PDEs. Its structural characterization was done by the determination of the protein primary sequence by Edman degradation and mass spectrometry, and completed by the building of molecular and docking-based models. Functional in vitro assays showed that CdcPDE is capable of inhibiting platelet aggregation induced by adenosine diphosphate in a dose-dependent manner and demonstrated that CdcPDE is cytotoxic to human keratinocytes. CdcPDE was recognized by the crotalid antivenom produced by the Instituto Butantan. These findings demonstrate that the study of snake venom toxins can reveal new molecules that may be relevant in cases of snakebite envenoming, and that can be used as molecular tools to study pathophysiological processes due to their specific biological activities.

CD4+ T Cells in the Blood of MS Patients Respond to Predicted Epitopes From B cell Receptors Found in Spinal Fluid

B cells are important pathogenic players in multiple sclerosis (MS), but their exact role is not known. We have previously demonstrated that B cells from cerebrospinal fluid (CSF) of MS patients can activate T cells that specifically recognize antigenic determinants (idiotopes) from their B cell receptors (BCRs). The aim of this study was to evaluate whether in silico prediction models could identify antigenic idiotopes of immunoglobulin heavy-chain variable (IGHV) transcriptomes in MS patients. We utilized a previously assembled dataset of CSF IGHV repertoires from MS patients. To guide selection of potential antigenic idiotopes, we used in silico predicted HLA-DR affinity, endosomal processing, as well as transcript frequency from nine MS patients. Idiotopes with predicted low affinity and low likelihood of cathepsins cleavage were inert controls. Peripheral blood mononuclear cells from these patients were stimulated with the selected idiotope peptides in presence of anti-CD40 for 12 h. T cells were then labeled for activation status with anti-CD154 antibodies and CD3+CD4+ T cells phenotyped as memory (CD45RO+) or naïve (CD45RO-), with potential for brain migration (CXCR3 and/or CCR6 expression). Anti-CD14 and -CD8 were utilized to exclude monocytes and CD8+ T cells. Unstimulated cells or insulin peptides were negative controls, and EBNA-1 peptides or CD3/CD28 beads were positive controls. The mean proportion of responding memory CD4+ T cells from all nine MS patients was significantly higher for idiotope peptides with predicted high HLA-DR affinity and high likelihood of cathepsin cleavage, than toward predicted inert peptides. Responses were mainly observed toward peptides affiliated with the CDR3 region. Activated memory CD4+ T cells expressed the chemokine receptor CCR6, affiliated with a Th17 phenotype and allowing passage into the central nervous system (CNS). This in vitro study suggests that that antigenic properties of BCR idiotopes can be identified in silico using HLA affinity and endosomal processing predictions. It further indicates that MS patients have a memory T cell repertoire capable of recognizing frequent BCR idiotopes found in endogenous CSF, and that these T cells express chemokine receptors allowing them to reach the CSF B cells expressing these idiotopes.

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.

Human Cysteine Cathepsins Degrade Immunoglobulin G In Vitro in a Predictable Manner

Cysteine cathepsins are critical components of the adaptive immune system involved in the generation of epitopes for presentation on human leukocyte antigen (HLA) molecules and have been implicated in degradation of autoantigens. Immunoglobulin variable regions with somatic mutations and random complementarity region 3 amino acid composition are inherently immunogenic. T cell reactivity towards immunoglobulin variable regions has been investigated in relation to specific diseases, as well as reactivity to therapeutic monoclonal antibodies. Yet, how the immunoglobulins, or the B cell receptors, are processed in endolysosomal compartments of professional antigen presenting cells has not been described in detail. Here we present in silico and in vitro experimental evidence suggesting that cysteine cathepsins S, L and B may have important roles in generating peptides fitting HLA class II molecules, capable of being presented to T cells, from monoclonal antibodies as well as from central nervous system proteins including a well described autoantigen. By combining neural net models with in vitro proteomics experiments, we further suggest how such degradation can be predicted, how it fits with available cellular models, and that it is immunoglobulin heavy chain variable family dependent. These findings are relevant for biotherapeutic drug design as well as to understand disease development. We also suggest how these tools can be improved, including improved machine learning methodology.

Computational B-cell epitope identification and production of neutralizing murine antibodies against Atroxlysin-I

Epitope identification is essential for developing effective antibodies that can detect and neutralize bioactive proteins. Computational prediction is a valuable and time-saving alternative for experimental identification. Current computational methods for epitope prediction are underused and undervalued due to their high false positive rate. In this work, we targeted common properties of linear B-cell epitopes identified in an individual protein class (metalloendopeptidases) and introduced an alternative method to reduce the false positive rate and increase accuracy, proposing to restrict predictive models to a single specific protein class. For this purpose, curated epitope sequences from metalloendopeptidases were transformed into frame-shifted Kmers (3 to 15 amino acid residues long). These Kmers were decomposed into a matrix of biochemical attributes and used to train a decision tree classifier. The resulting prediction model showed a lower false positive rate and greater area under the curve when compared to state-of-the-art methods. Our predictions were used for synthesizing peptides mimicking the predicted epitopes for immunization of mice. A predicted linear epitope that was previously undetected by an experimental immunoassay was able to induce neutralizing-antibody production in mice. Therefore, we present an improved prediction alternative and show that computationally identified epitopes can go undetected during experimental mapping.

A Role for Epitope Networking in Immunomodulation by Helminths

Helminth infections, by nematodes, trematodes, or cestodes, can lead to the modulation of host immune responses. This allows long-duration parasite infections and also impacts responses to co-infections. Surface, secreted, excreted, and shed proteins are thought to play a major role in modulation. A commonly reported feature of such immune modulation is the role of T regulatory (Treg) cells and IL-10. Efforts to identify helminth proteins, which cause immunomodulation, have identified candidates but not provided clarity as to a uniform mechanism driving modulation. In this study, we applied a bioinformatics systems approach, allowing us to analyze predicted T-cell epitopes of 17 helminth species and the responses to their surface proteins. In addition to major histocompatibility complex (MHC) binding, we analyzed amino acid motifs that would be recognized by T-cell receptors [T-cell-exposed motifs (TCEMs)]. All the helminth species examined have, within their surface proteins, peptides, which combine very common TCEMs with predicted high affinity binding to many human MHC alleles. This combination of features would result in large cognate T cell and a high probability of eliciting Treg responses. The TCEMs, which determine recognition by responding T-cell clones, are shared to a high degree between helminth species and with Plasmodium falciparum and Mycobacterium tuberculosis, both common co-infecting organisms. The implication of our observations is not only that Treg cells play a significant role in helminth-induced immune modulation but also that the epitope specificities of Treg responses are shared across species and genera of helminth. Hence, the immune response to a given helminth cannot be considered in isolation but rather forms part of an epitope ecosystem, or microenvironment, in which potentially immunosuppressive peptides in the helminth network via their common T-cell receptor recognition signals with T-cell epitopes in self proteins, microbiome, other helminths, and taxonomically unrelated pathogens. Such a systems approach provides a high-level view of the antigen-immune system signaling dynamics that may bias a host's immune response to helminth infections toward immune modulation. It may indicate how helminths have evolved to select for peptides that favor long-term parasite host coexistence.

In Silico Prediction Analysis of Idiotope-Driven T-B Cell Collaboration in Multiple Sclerosis

Memory B cells acting as antigen-presenting cells are believed to be important in multiple sclerosis (MS), but the antigen they present remains unknown. We hypothesized that B cells may activate CD4⁺ T cells in the central nervous system of MS patients by presenting idiotopes from their own immunoglobulin variable regions on human leukocyte antigen (HLA) class II molecules. Here, we use bioinformatics prediction analysis of B cell immunoglobulin variable regions from 11 MS patients and 6 controls with other inflammatory neurological disorders (OINDs), to assess whether the prerequisites for such idiotope-driven T–B cell collaboration are present. Our findings indicate that idiotopes from the complementarity determining region (CDR) 3 of MS patients on average have high predicted affinities for disease associated HLA-DRB1*15:01 molecules and are predicted to be endosomally processed by cathepsin S and L in positions that allows such HLA binding to occur. Additionally, complementarity determining region 3 sequences from cerebrospinal fluid (CSF) B cells from MS patients contain on average more rare T cell-exposed motifs that could potentially escape tolerance and stimulate CD4⁺ T cells than CSF B cells from OIND patients. Many of these features were associated with preferential use of the IGHV4 gene family by CSF B cells from MS patients. This is the first study to combine high-throughput sequencing of patient immune repertoires with large-scale prediction analysis and provides key indicators for future in vitro and in vivo analyses.

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.

Computational prediction of immunodominant antigenic regions & potential protective epitopes for dengue vaccination

Background & objectives:

Epitope-based vaccines (EVs) are specific, safe and easy to produce. However, vaccine failure has been frequently reported due to variation within epitopic regions. Therefore, development of vaccines based on conserved epitopes may prevent such vaccine failure. This study was undertaken to identify highly conserved antigenic regions in the four dengue serotypes to produce an epitope-based dengue vaccine.

Methods:

Polyprotein sequences of all four dengue serotypes were collected and aligned using MAFFT multiple sequence alignment plugin with Geneious Pro v6.1. Consensus sequences of the polyproteins for all four dengue serotypes were designed and screened against experimentally proven epitopes to predict potential antigenic regions that are conserved among all four dengue serotypes.

Results:

The antigenic region VDRGWGNGCGLFGKG was 100 per cent conserved in the consensus polyprotein sequences of all four dengue serotypes. Fifteen experimentally proven epitopes were identical to the immunodominant antigenic region.

Interpretation & conclusions:

Computationally predicted antigenic regions may be considered for use in the development of EVs for protection against dengue virus. Such vaccines would be expected to provide protection against dengue infections caused by all dengue serotypes because these would contain antigenic regions highly conserved across those serotypes. Therefore, the immunodominant antigenic region (VDRGWGNGCGLFGKG) and 15 potential epitopes may be considered for use in dengue vaccines.

The quantitative prediction of HLA-B*2705 peptide binding affinities using Support Vector Regression to gain insights into its role for the Spondyloarthropathies

Computational methods are increasingly utilised in many immunoinformatics problems such as the prediction of binding affinity of peptides. The peptides could provide valuable insight into the drug design and development such as vaccines. Moreover, they can be used to diagnose diseases. The presence of human class I MHC allele HLA-B*2705 is one of the strong hypothesis that would lead spondyloarthropathies. In this paper, Support Vector Regression is used in order to predict binding affinity of peptides with the aid of experimentally determined peptide-MHC binding affinities of 222 peptides to HLA-B*2705 to get more insight into this problematic disease. The results yield a high correlation coefficient as much as 0.65 and the SVR-based predictive models can be considered as a useful tool in order to predict the binding affinities for newly discovered peptides.

Classification epitopes in groups based on their protein family

Background

The humoral immune system response is based on the interaction between antibodies and antigens for the clearance of pathogens and foreign molecules. The interaction between these proteins occurs at specific positions known as antigenic determinants or B-cell epitopes. The experimental identification of epitopes is costly and time consuming. Therefore the use of in silico methods, to help discover new epitopes, is an appealing alternative due the importance of biomedical applications such as vaccine design, disease diagnostic, anti-venoms and immune-therapeutics. However, the performance of predictions is not optimal been around 70% of accuracy. Further research could increase our understanding of the biochemical and structural properties that characterize a B-cell epitope.

Results

We investigated the possibility of linear epitopes from the same protein family to share common properties. This hypothesis led us to analyze physico-chemical (PCP) and predicted secondary structure (PSS) features of a curated dataset of epitope sequences available in the literature belonging to two different groups of antigens (metalloproteinases and neurotoxins). We discovered statistically significant parameters with data mining techniques which allow us to distinguish neurotoxin from metalloproteinase and these two from random sequences. After a five cross fold validation we found that PCP based models obtained area under the curve values (AUC) and accuracy above 0.9 for regression, decision tree and support vector machine.

Conclusions

We demonstrated that antigen's family can be inferred from properties within a single group of linear epitopes (metalloproteinases or neurotoxins). Also we discovered the characteristics that represent these two epitope groups including their similarities and differences with random peptides and their respective amino acid sequence. These findings open new perspectives to improve epitope prediction by considering the specific antigen's protein family. We expect that these findings will help to improve current computational mapping methods based on physico-chemical due it's potential application during epitope discovery.

Frequency Patterns of T-Cell Exposed Amino Acid Motifs in Immunoglobulin Heavy Chain Peptides Presented by MHCs

Immunoglobulins are highly diverse protein sequences that are processed and presented to T-cells by B-cells and other antigen presenting cells. We examined a large dataset of immunoglobulin heavy chain variable regions (IGHV) to assess the diversity of T-cell exposed motifs (TCEMs). TCEM comprise those amino acids in a MHC-bound peptide, which face outwards, surrounded by the MHC histotope, and which engage the T-cell receptor. Within IGHV there is a distinct pattern of predicted MHC class II binding and a very high frequency of re-use of the TCEMs. The re-use frequency indicates that only a limited number of different cognate T-cells are required to engage many different clonal B-cells. The amino acids in each outward-facing TCEM are intercalated with the amino acids of inward-facing MHC groove-exposed motifs (GEM). Different GEM may have differing, allele-specific, MHC binding affinities. The intercalation of TCEM and GEM in a peptide allows for a vast combinatorial repertoire of epitopes, each eliciting a different response. Outcome of T-cell receptor binding is determined by overall signal strength, which is a function of the number of responding T-cells and the duration of engagement. Hence, the frequency of TCEM re-use appears to be an important determinant of whether a T-cell response is stimulatory or suppressive. The frequency distribution of TCEMs implies that somatic hypermutation is followed by T-cell clonal expansion that develops along repeated pathways. The observations of TCEM and GEM derived from immunoglobulins suggest a relatively simple, yet powerful, mechanism to correlate T-cell polyspecificity, through re-use of TCEMs, with a very high degree of specificity achieved by combination with a diversity of GEMs. The frequency profile of TCEMs also points to an economical mechanism for maintaining T-cell memory, recall, and self-discrimination based on an endogenously generated profile of motifs.

Recognition of higher order patterns in proteins: immunologic kernels

By applying analysis of the principal components of amino acid physical properties we predicted cathepsin cleavage sites, MHC binding affinity, and probability of B-cell epitope binding of peptides in tetanus toxin and in ten diverse additional proteins. Cross-correlation of these metrics, for peptides of all possible amino acid index positions, each evaluated in the context of a ±25 amino acid flanking region, indicated that there is a strongly repetitive pattern of short peptides of approximately thirty amino acids each bounded by cathepsin cleavage sites and each comprising B-cell linear epitopes, MHC–I and MHC-II binding peptides. Such “immunologic kernel” peptides comprise all signals necessary for adaptive immunologic cognition, response and recall. The patterns described indicate a higher order spatial integration that forms a symbolic logic coordinating the adaptive immune system.

Hybrid biogeography based simultaneous feature selection and MHC class I peptide binding prediction using support vector machines and random forests

Accurate detection of peptides binding to specific Major Histocompatibility Complex Class I (MHC-I) molecules is extremely important for understanding the underlying process of the immune system, as well as for effective vaccine design and developing immunotherapies. Development of learning algorithms and their application for binding predictions have thus speeded up the state-of-the-art in immunological research, in a cost-effective manner. In this work, we propose the application of a hybrid filter-wrapper algorithm employing concepts from the recently developed biogeography based optimization algorithm, in conjunction with SVM and Random Forests for identification of MHC-I binding peptides. In the process, we demonstrate the effectiveness of this evolutionary technique, coupled with weighted heuristics, for the construction of improved prediction models. The experiments have been carried out for the CoEPrA competition datasets (accessible online at: http://www.coepra.org) and the results show a marked improvement over the winner results in some situations and comparably good with regard to others .We thus hope to initiate further research on the application of this new bio-inspired methodology for immunological research.

Patterns of predicted T-cell epitopes associated with antigenic drift in influenza H3N2 hemagglutinin

Antigenic drift allowing escape from neutralizing antibodies is an important feature of transmission and survival of influenza viruses in host populations. Antigenic drift has been studied in particular detail for influenza A H3N2 and well defined antigenic clusters of this virus documented. We examine how host immunogenetics contributes to determination of the antibody spectrum, and hence the immune pressure bringing about antigenic drift. Using uTOPE™ bioinformatics analysis of predicted MHC binding, based on amino acid physical property principal components, we examined the binding affinity of all 9-mer and 15-mer peptides within the hemagglutinin 1 (HA1) of 447 H3N2 virus isolates to 35 MHC-I and 14 MHC-II alleles. We provide a comprehensive map of predicted MHC-I and MHC-II binding affinity for a broad array of HLA alleles for the H3N2 influenza HA1 protein. Each HLA allele exhibited a characteristic predicted binding pattern. Cluster analysis for each HLA allele shows that patterns based on predicted MHC binding mirror those described based on antibody binding. A single amino acid mutation or position displacement can result in a marked difference in MHC binding and hence potential T-helper function. We assessed the impact of individual amino acid changes in HA1 sequences between 10 virus isolates from 1968-2002, representative of antigenic clusters, to understand the changes in MHC binding over time. Gain and loss of predicted high affinity MHC-II binding sites with cluster transitions were documented. Predicted high affinity MHC-II binding sites were adjacent to antibody binding sites. We conclude that host MHC diversity may have a major determinant role in the antigenic drift of influenza A H3N2.

An integrated approach to epitope analysis II: A system for proteomic-scale prediction of immunological characteristics

Background

Improving our understanding of the immune response is fundamental to developing strategies to combat a wide range of diseases. We describe an integrated epitope analysis system which is based on principal component analysis of sequences of amino acids, using a multilayer perceptron neural net to conduct QSAR regression predictions for peptide binding affinities to 35 MHC-I and 14 MHC-II alleles.

Results

The approach described allows rapid processing of single proteins, entire proteomes or subsets thereof, as well as multiple strains of the same organism. It enables consideration of the interface of diversity of both microorganisms and of host immunogenetics. Patterns of binding affinity are linked to topological features, such as extracellular or intramembrane location, and integrated into a graphical display which facilitates conceptual understanding of the interplay of B-cell and T-cell mediated immunity.

Patterns which emerge from application of this approach include the correlations between peptides showing high affinity binding to MHC-I and to MHC-II, and also with predicted B-cell epitopes. These are characterized as coincident epitope groups (CEGs). Also evident are long range patterns across proteins which identify regions of high affinity binding for a permuted population of diverse and heterozygous HLA alleles, as well as subtle differences in reactions with MHCs of individual HLA alleles, which may be important in disease susceptibility, and in vaccine and clinical trial design. Comparisons are shown of predicted epitope mapping derived from application of the QSAR approach with experimentally derived epitope maps from a diverse multi-species dataset, from Staphylococcus aureus, and from vaccinia virus.

Conclusions

A desktop application with interactive graphic capability is shown to be a useful platform for development of prediction and visualization tools for epitope mapping at scales ranging from individual proteins to proteomes from multiple strains of an organism. The possible functional implications of the patterns of peptide epitopes observed are discussed, including their implications for B-cell and T-cell cooperation and cross presentation.