Rapid and cost-effective epitope mapping using PURE ribosome display coupled with next-generation sequencing and bioinformatics.

A novel, efficient and cost-effective approach for epitope identification of an antibody has been developed using a ribosome display platform. This platform, known as PURE ribosome display, utilizes an Escherichia coli-based reconstituted cell-free protein synthesis system (PURE system). It stabilizes the mRNA-ribosome-peptide complex via a ribosome-arrest peptide sequence. This system was complemented by next-generation sequencing (NGS) and an algorithm for analyzing binding epitopes. To showcase the effectiveness of this method, selection conditions were refined using the anti-PA tag monoclonal antibody with the PA tag peptide as a model. Subsequently, a random peptide library was constructed using 10 NNK triplet oligonucleotides via the PURE ribosome display. The resulting random peptide library-ribosome-mRNA complex was selected using a commercially available anti-HA (YPYDVPDYA) tag monoclonal antibody, followed by NGS and bioinformatic analysis. Our approach successfully identified the DVPDY sequence as an epitope within the hemagglutinin amino acid sequence, which was then experimentally validated. This platform provided a valuable tool for investigating continuous epitopes in antibodies.

De novo design and Rosetta-based assessment of high-affinity antibody variable regions (Fv) against the SARS-CoV-2 spike receptor binding domain (RBD).

The continued emergence of new SARS-CoV-2 variants has accentuated the growing need for fast and reliable methods for the design of potentially neutralizing antibodies (Abs) to counter immune evasion by the virus. Here, we report on the de novo computational design of high-affinity Ab variable regions (Fv) through the recombination of VDJ genes targeting the most solvent-exposed hACE2-binding residues of the SARS-CoV-2 spike receptor binding domain (RBD) protein using the software tool OptMAVEn-2.0. Subsequently, we carried out computational affinity maturation of the designed variable regions through amino acid substitutions for improved binding with the target epitope. Immunogenicity of designs was restricted by preferring designs that match sequences from a 9-mer library of "human Abs" based on a human string content score. We generated 106 different antibody designs and reported in detail on the top five that trade-off the greatest computational binding affinity for the RBD with human string content scores. We further describe computational evaluation of the top five designs produced by OptMAVEn-2.0 using a Rosetta-based approach. We used Rosetta SnugDock for local docking of the designs to evaluate their potential to bind the spike RBD and performed "forward folding" with DeepAb to assess their potential to fold into the designed structures. Ultimately, our results identified one designed Ab variable region, P1.D1, as a particularly promising candidate for experimental testing. This effort puts forth a computational workflow for the de novo design and evaluation of Abs that can quickly be adapted to target spike epitopes of emerging SARS-CoV-2 variants or other antigenic targets.

Computational design of a neutralizing antibody with picomolar binding affinity for all concerning SARS-CoV-2 variants.

Coronavirus disease 2019, caused by SARS-CoV-2, remains an on-going pandemic, partly due to the emergence of variant viruses that can "break-through" the protection of the current vaccines and neutralizing antibodies (nAbs), highlighting the needs for broadly nAbs and next-generation vaccines. We report an antibody that exhibits breadth and potency in binding the receptor-binding domain (RBD) of the virus spike glycoprotein across SARS coronaviruses. Initially, a lead antibody was computationally discovered and crystallographically validated that binds to a highly conserved surface of the RBD of wild-type SARS-CoV-2. Subsequently, through experimental affinity enhancement and computational affinity maturation, it was further developed to bind the RBD of all concerning SARS-CoV-2 variants, SARS-CoV-1 and pangolin coronavirus with pico-molar binding affinities, consistently exhibited strong neutralization activity against wild-type SARS-CoV-2 and the Alpha and Delta variants. These results identify a vulnerable target site on coronaviruses for development of pan-sarbecovirus nAbs and vaccines.

Identification of Electrostatic Epitopes in Flavivirus by Computer Simulations: The PROCEEDpKa Method.

Viruses are enthusiastically studied due to the great impact that these organisms can have on human health. Computational approaches can contribute offering tools that can shed light on important molecular mechanisms that help to design new diagnostic procedures. Several cellular processes between the immune-host system and the pathogenic organism are dependent on specific intermolecular interactions. In this study, we evaluated theoretical approaches to understand some properties of the antigen-antibody interactions considering the titratable properties of all ionizable residues of the nonstructural viral protein 1 (NS1) of the West Nile virus (WNV) and the Zika virus (ZIKV). Constant-pH Monte Carlo simulations were performed to estimate electrostatic properties such as the pKa shifts (ΔpKa). We proposed an alternative criterion for the discrimination of antigenic residues based on ΔpKas. Our outcomes were analyzed by an evaluation of the sensitivity and specificity through a receiver operating characteristic (ROC). As a starting point, we used the known crystallographic structure for the complex of NS1WNV(176-352) and the specific antibody 22NS1 (PDB ID 4OII ) to differentiate the residues belonging to that interface. With an optimal threshold for the absolute value of the pKa shifts, we found that is possible to predict antigenic epitopes reproducing the interfaces as defined by the X-ray structure. After this validation, we evaluated theoretical predictions based on protein-protein (PP) complexation simulations. From them, we observe amino acids with an antigenic potential and defined the optimum threshold that was applied for two strains of ZIKV (i.e., Uganda and Brazil). Several ionizable residues with antigenic capacity were identified. This is favorably related to some studies that show the high immunogenicity of secreted NS1. This approach opens up an important discussion about what are termed here "electrostatic epitopes" and how they work as an important reference in the paratope-epitope interaction for viral systems.

Antibody Epitopes of Pneumovirus Fusion Proteins.

The pneumoviruses respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) are two widespread human pathogens that can cause severe disease in the young, the elderly, and the immunocompromised. Despite the discovery of RSV over 60 years ago, and hMPV nearly 20 years ago, there are no approved vaccines for either virus. Antibody-mediated immunity is critical for protection from RSV and hMPV, and, until recently, knowledge of the antibody epitopes on the surface glycoproteins of RSV and hMPV was very limited. However, recent breakthroughs in the recombinant expression and stabilization of pneumovirus fusion proteins have facilitated in-depth characterization of antibody responses and structural epitopes, and have provided an enormous diversity of new monoclonal antibody candidates for therapeutic development. These new data have primarily focused on the RSV F protein, and have led to a wealth of new vaccine candidates in preclinical and clinical trials. In contrast, the major structural antibody epitopes remain unclear for the hMPV F protein. Overall, this review will cover recent advances in characterizing the antigenic sites on the RSV and hMPV F proteins.

In silico selection of functionally important proteins from the mialome of Ornithodoros erraticus ticks and assessment of their protective efficacy as vaccine targets.

BACKGROUND: New candidate protective antigens for tick vaccine development may be identified by selecting and testing antigen candidates that play key biological functions. After blood-feeding, tick midgut overexpresses proteins that play essential functions in tick survival and disease transmission. Herein, Ornithodoros erraticus midgut transcriptomic and proteomic data were examined in order to select functionally significant antigens upregulated after feeding to be tested as vaccine candidate antigens. METHODS: Transcripts annotated as chitinases, tetraspanins, ribosomal protein P0 and secreted proteins/peptides were mined from the recently published O. erraticus midgut transcriptome and filtered in a second selection step using criteria based on upregulation after feeding, predicted antigenicity and expression in the midgut proteome. Five theoretical candidate antigens were selected, obtained as recombinant proteins and used to immunise rabbits: one chitinase (CHI), two tetraspanins (TSPs), the ribosomal protein P0 (RPP0) and one secreted protein PK-4 (PK4). RESULTS: Rabbit vaccination with individual recombinant candidates induced strong humoral responses that mainly reduced nymph moulting and female reproduction, providing 30.2% (CHI), 56% (TSPs), 57.5% (RPP0) and 57.8% (PK4) protection to O. erraticus infestations and 19.6% (CHI), 11.1% (TSPs), 0% (RPP0) and 8.1% (PK4) cross-protection to infestations by the African tick Ornithodoros moubata. The joint vaccine efficacy of the candidates was assessed in a second vaccine trial reaching 66.3% protection to O. erraticus and 25.6% cross-protection to O. moubata. CONCLUSIONS: These results (i) indicate that argasid chitinases and RPP0 are promising protective antigens, as has already been demonstrated for ixodid chitinases and RPP0, and could be included in vaccines targeting multiple tick species; (ii) reveal novel protective antigens tetraspanins and secreted protein PK-4, never tested before as protective antigens in ticks; and (iii) demonstrate that multi-antigenic vaccines increased vaccine efficacy compared with individual antigens. Lastly, our data emphasize the value of the tick midgut as a source of protective candidate antigens in argasids for tick control.

Characterizing the Diversity of the CDR-H3 Loop Conformational Ensembles in Relationship to Antibody Binding Properties.

We present an approach to assess antibody CDR-H3 loops according to their dynamic properties using molecular dynamics simulations. We selected six antibodies in three pairs differing substantially in their individual promiscuity respectively specificity. For two pairs of antibodies crystal structures are available in different states of maturation and used as starting structures for the analyses. For a third pair we chose two antibody CDR sequences obtained from a synthetic library and predicted the respective structures. For all three pairs of antibodies we performed metadynamics simulations to overcome the limitations in conformational sampling imposed by high energy barriers. Additionally, we used classic molecular dynamics simulations to describe nano- to microsecond flexibility and to estimate up to millisecond kinetics of captured conformational transitions. The methodology represents the antibodies as conformational ensembles and allows comprehensive analysis of structural diversity, thermodynamics of conformations and kinetics of structural transitions. Referring to the concept of conformational selection we investigated the link between promiscuity and flexibility of the antibodies' binding interfaces. The obtained detailed characterization of the binding interface clearly indicates a link between structural flexibility and binding promiscuity for this set of antibodies.

Modeling the Role of Epitope Arrangement on Antibody Binding Stoichiometry in Flaviviruses.

Cryo-electron-microscopy (cryo-EM) structures of flaviviruses reveal significant variation in epitope occupancy across different monoclonal antibodies that have largely been attributed to epitope-level differences in conformation or accessibility that affect antibody binding. The consequences of these variations for macroscopic properties such as antibody binding and neutralization are the results of the law of mass action-a stochastic process of innumerable binding and unbinding events between antibodies and the multiple binding sites on the flavivirus in equilibrium-that cannot be directly imputed from structure alone. We carried out coarse-grained spatial stochastic binding simulations for nine flavivirus antibodies with epitopes defined by cryo-EM or x-ray crystallography to assess the role of epitope spatial arrangement on antibody-binding stoichiometry, occupancy, and neutralization. In our simulations, all epitopes were equally competent for binding, representing the upper limit of binding stoichiometry that results from epitope spatial arrangement alone. Surprisingly, our simulations closely reproduced the relative occupancy and binding stoichiometry observed in cryo-EM, without having to account for differences in epitope accessibility or conformation, suggesting that epitope spatial arrangement alone may be sufficient to explain differences in binding occupancy and stoichiometry between antibodies. Furthermore, we found that there was significant heterogeneity in binding configurations even at saturating antibody concentrations, and that bivalent antibody binding may be more common than previously thought. Finally, we propose a structure-based explanation for the stoichiometric threshold model of neutralization.

Hammock: a hidden Markov model-based peptide clustering algorithm to identify protein-interaction consensus motifs in large datasets.

MOTIVATION: Proteins often recognize their interaction partners on the basis of short linear motifs located in disordered regions on proteins' surface. Experimental techniques that study such motifs use short peptides to mimic the structural properties of interacting proteins. Continued development of these methods allows for large-scale screening, resulting in vast amounts of peptide sequences, potentially containing information on multiple protein-protein interactions. Processing of such datasets is a complex but essential task for large-scale studies investigating protein-protein interactions. RESULTS: The software tool presented in this article is able to rapidly identify multiple clusters of sequences carrying shared specificity motifs in massive datasets from various sources and generate multiple sequence alignments of identified clusters. The method was applied on a previously published smaller dataset containing distinct classes of ligands for SH3 domains, as well as on a new, an order of magnitude larger dataset containing epitopes for several monoclonal antibodies. The software successfully identified clusters of sequences mimicking epitopes of antibody targets, as well as secondary clusters revealing that the antibodies accept some deviations from original epitope sequences. Another test indicates that processing of even much larger datasets is computationally feasible. AVAILABILITY AND IMPLEMENTATION: Hammock is published under GNU GPL v. 3 license and is freely available as a standalone program (from http://www.recamo.cz/en/software/hammock-cluster-peptides/) or as a tool for the Galaxy toolbox (from https://toolshed.g2.bx.psu.edu/view/hammock/hammock). The source code can be downloaded from https://github.com/hammock-dev/hammock/releases. CONTACT: muller@mou.cz SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Broad spectrum assessment of the epitope fluctuation--Immunogenicity hypothesis.

Prediction of immunogenicity is a substantial barrier in vaccine design. Here, a molecular dynamics approach to assessing the immunogenicity of nanoparticles based on structure is presented. Molecular properties of epitopes on nonenveloped viral particles are quantified via a set of metrics. One such metric, epitope fluctuation (and implied flexibility), is shown to be inversely correlated with immunogenicity for each of a broad spectrum of nonenveloped viruses. The molecular metrics and experimentally determined immunogenicities for these viruses are archived in the open-source vaccine computer-aided design database. Results indicate the promise of computer-aided vaccine design to bring greater efficiency to traditional lab-based vaccine discovery approaches.

Assessment of 3D models for allergen research.

Allergenic proteins must crosslink specific IgE molecules, bound to the surface of mast cells and basophils, to stimulate an immune response. A structural understanding of the allergen-IgE interface is needed to predict cross-reactivities between allergens and to design hypoallergenic proteins. However, there are less than 90 experimentally determined structures available for the approximately 1500 sequences of allergens and isoallergens cataloged in the Structural Database of Allergenic Proteins. To provide reliable structural data for the remaining proteins, we previously produced more than 500 3D models using an automated procedure, with strict controls on template choice and model quality evaluation. Here, we assessed how well the fold and residue surface exposure of 10 of these models correlated with recently published experimental 3D structures determined by X-ray crystallography or NMR. We also discuss the impact of intrinsically disordered regions on the structural comparison and epitope prediction. Overall, for seven allergens with sequence identities to the original templates higher than 27%, the backbone root-mean square deviations were less than 2 Å between the models and the subsequently determined experimental structures for the ordered regions. Further, the surface exposure of the known IgE epitopes on the models of three major allergens, from peanut (Ara h 1), latex (Hev b 2), and soy (Gly m 4), was very similar to the experimentally determined structures. For the three remaining allergens with lower sequence identities to the modeling templates, the 3D folds were correctly identified. However, the accuracy of those models is not sufficient for a reliable epitope mapping.

Expression of recombinant Ara h 6 in Pichia pastoris but not in Escherichia coli preserves allergic effector function and allows assessment of specific mutations.

SCOPE: Ara h 6 has recently been recognized as an important peanut allergen. Recombinant allergens have been used for analysis of IgE binding, but have not been used to analyze the allergic effector activity that is more relevant to allergic reactions. METHODS AND RESULTS: Ara h 6 was expressed as a recombinant protein in both Escherichia coli and Pichia pastoris (rAra h 6-E. coli and rAra h 6-Pichia, respectively). Effector activity was assayed by measuring degranulation of RBL SX-38 cells sensitized with IgE from patients with severe peanut allergy. Compared to native Ara h 6 (nAra h 6), rAra h 6-Pichia had intact effector function whereas rAra h 6-E. coli had significantly reduced function. The lower effector activity in rAra h 6-E. coli compared to nAra h 6 and rAra h 6-Pichia did not appear to be due to differences in posttranslational modifications (analyzed by mass spectrometry and staining for carbohydrates) and may be due to subtle alteration(s) of folding seen on CD analysis and on nonreduced gels. Finally, we introduced point mutations in four important IgE-binding linear epitopes of Ara h 6 and found dramatically reduced allergic effector activity. CONCLUSION: Our studies demonstrate the utility of fully functional rAra h 6-Pichia as a starting point for analysis of specific mutations that adversely affect allergic effector function.

Comparative assessment of the recognition of domain-specific CD163 monoclonal antibodies in human monocytes explains wide discrepancy in reported levels of cellular surface CD163 expression.

BACKGROUND: CD163 is expressed exclusively on cells of the monocyte/macrophage lineage and is widely used as a marker of human macrophages. Further, it has been suggested as a diagnostic marker of monocyte/macrophage activity in inflammatory conditions and as a therapeutic target. However, studies continue to exhibit great discrepancy in the measured percentage of CD163-expressing blood monocytes in healthy individuals. In this study we sought to clarify this inconsistency in reported levels of CD163 surface expression by a detailed analysis of a panel of CD163 antibodies used in previous studies. MATERIALS AND METHODS: The cellular distribution of CD163 on human peripheral blood monocytes in freshly drawn blood and peripheral blood mononuclear cells isolated from buffy-coats was investigated by flow cytometry using CD163 monoclonal antibodies recognizing scavenger receptor cysteine-rich (SRCR) domain 1 (MAC2-158), domain 4 (R-20), domain 7 (GHI/61), and domain 9 (RM3/1). The CD163 monoclonal antibodies were characterized in binding and endocytosis experiments in human macrophages and CD163-transfected Flp-In CHO cells. Calcium-dependent ligand binding was assessed using surface plasmon resonance, and the specificity of the CD163 monoclonal antibodies was analyzed by western blotting. RESULTS AND DISCUSSION: Flow cytometric analysis revealed that the estimated proportion of CD163-expressing human peripheral blood monocytes increased when using CD163 monoclonal antibodies recognizing epitopes in the N-terminal part of CD163, remote from the membrane surface. Moreover, the proportion of CD163 positive monocytes observed was highly dependent on free calcium. GHI/61 did not exhibit CD163 binding in the presence of calcium as measured by surface plasmon resonance, which was in agreement with the concordant loss of binding in heparin-stabilized whole blood observed by flow cytometry. In contrast, RM3/1 exhibited weak binding to CD163 in the absence of calcium but high affinity binding to CD163 in the presence of calcium. R-20 and MAC2-158 were unaffected by extracellular calcium levels. The latter SRCR domain 1mAb consistently recognized more than 80% CD163-positive monocytes in human peripheral blood. CONCLUSION: Epitope accessibility and extracellular calcium dependence elucidate discrepancies in reported levels of monocytic CD163 expression. Utilizing monoclonal antibodies to the N-terminal part of CD163 more than 80% monocytes in human peripheral blood could be identified as CD163 positive, indicating that most, and conceivably all, human peripheral blood monocytes do express CD163.

Increased rheumatoid factor interference observed during immunogenicity assessment of an Fc-engineered therapeutic antibody.

Protein therapeutics may elicit an anti-therapeutic antibody (ATA) response in patients. This response depends on a number of factors including patient population, disease state, route of delivery or characteristics specific to the product. Therapeutics for immunological indications often target relatively young and healthy patients with hyperactive immune systems who have periodic flares and remissions. The hyperactive immune system of these patients can add several levels of bioanalytical complexity due to the presence of cross reactive molecules such as autoantibodies. In addition, the long-term chronic dosing regimen often necessary in this patient population can increase their risks of immunogenicity against the therapeutic and lead to safety concerns. Therefore, development of a sensitive and drug-tolerant ATA method is important. Bridging ATA assays are usually very sensitive and drug-tolerant methods for immunogenicity assessment; however these methods are particularly vulnerable to any factor that is able to bridge the conjugated therapeutics used as reagents and can generate false positive signal. Although there are many potential interfering factors in serum, rheumatoid factors (RFs), autoantibodies associated with rheumatoid arthritis (RA), are of particular concern in this type of assay. MTRX1011A is a non-depleting anti-CD4 monoclonal antibody therapeutic that was clinically tested in RA patients. This paper will discuss the bioanalytical challenges encountered during development of a clinical ATA assay for MTRX1011A. These challenges highlight interference due to patient disease state, in this case presence of RF in RA patients, as well as specific molecule-related interference caused by an engineered mutation in the Fc region of MTRX1011A designed to enhance its binding to the neonatal Fc receptor (FcRn). We will discuss the characterization work used to identify the cross-reactive epitope and our strategy to overcome this interference during development of an effective ATA assay to support clinical evaluation of MTRX1011A.

Towards universal structure-based prediction of class II MHC epitopes for diverse allotypes.

The binding of peptide fragments of antigens to class II MHC proteins is a crucial step in initiating a helper T cell immune response. The discovery of these peptide epitopes is important for understanding the normal immune response and its misregulation in autoimmunity and allergies and also for vaccine design. In spite of their biomedical importance, the high diversity of class II MHC proteins combined with the large number of possible peptide sequences make comprehensive experimental determination of epitopes for all MHC allotypes infeasible. Computational methods can address this need by predicting epitopes for a particular MHC allotype. We present a structure-based method for predicting class II epitopes that combines molecular mechanics docking of a fully flexible peptide into the MHC binding cleft followed by binding affinity prediction using a machine learning classifier trained on interaction energy components calculated from the docking solution. Although the primary advantage of structure-based prediction methods over the commonly employed sequence-based methods is their applicability to essentially any MHC allotype, this has not yet been convincingly demonstrated. In order to test the transferability of the prediction method to different MHC proteins, we trained the scoring method on binding data for DRB1*0101 and used it to make predictions for multiple MHC allotypes with distinct peptide binding specificities including representatives from the other human class II MHC loci, HLA-DP and HLA-DQ, as well as for two murine allotypes. The results showed that the prediction method was able to achieve significant discrimination between epitope and non-epitope peptides for all MHC allotypes examined, based on AUC values in the range 0.632-0.821. We also discuss how accounting for peptide binding in multiple registers to class II MHC largely explains the systematically worse performance of prediction methods for class II MHC compared with those for class I MHC based on quantitative prediction performance estimates for peptide binding to class II MHC in a fixed register.

Structural allele-specific patterns adopted by epitopes in the MHC-I cleft and reconstruction of MHC:peptide complexes to cross-reactivity assessment.

The immune system is engaged in a constant antigenic surveillance through the Major Histocompatibility Complex (MHC) class I antigen presentation pathway. This is an efficient mechanism for detection of intracellular infections, especially viral ones. In this work we describe conformational patterns shared by epitopes presented by a given MHC allele and use these features to develop a docking approach that simulates the peptide loading into the MHC cleft. Our strategy, to construct in silico MHC:peptide complexes, was successfully tested by reproducing four different crystal structures of MHC-I molecules available at the Protein Data Bank (PDB). An in silico study of cross-reactivity potential was also performed between the wild-type complex HLA-A2-NS31073 and nine MHC:peptide complexes presenting alanine exchange peptides. This indicates that structural similarities among the complexes can give us important clues about cross reactivity. The approach used in this work allows the selection of epitopes with potential to induce cross-reactive immune responses, providing useful tools for studies in autoimmunity and to the development of more comprehensive vaccines.

Epitope characterization of an aromatase monoclonal antibody suitable for the assessment of intratumoral aromatase activity.

Immunohistochemistry is one of the most suitable methods for the detection of intratumoral aromatase in order to identify patients who may respond to aromatase inhibitor therapy in hormone-dependent breast cancer. Previous studies showed statistically significant correlation between results of immnuohistochemistry and biochemical analysis in carcinoma components stained by aromatase monoclonal antibody 677. In this study, determination of the antigenic peptides recognized by aromatase antibodies through epitope mapping, combined with the new knowledge on aromatase-reductase interaction, provide insights for understanding various immunostaining patterns using different aromatase antibodies. Our studies on aromatase-reductase interaction also provided critical information on how aromatase and reductase interact with each other on the endoplasmic reticulum membrane, and identified key residues, including K108 of aromatase, that are involved in the interaction with reductase. Through epitope mapping and taking into consideration the interference with aromatase immunohistochemical staining by NADPH-cytochrome P450 reductase, we demonstrated that monoclonal antibody 677 is a suitable antibody for an assessment of intratumoral aromatase activity in breast cancer patients for making clinical management decisions. These results also provide valuable information to identify new aromatase antibodies for immunohistochemical diagnosis of hormone-dependent breast cancer in future.

A shooting algorithm for complex immunodominance control problems.

Although T cells are able to recognize a wide variety of target peptides, they are often strongly focused on a few of the peptides and leave the rest of them unattended. This phenomenon of strongly biased immune response is known as immunodominance. Mathematically, an immunodominance problem can be formulated using optimal control principles as a two-point boundary-value problem. The solution of this problem is challenging especially when the control variables are bounded. In this work, we develop a numerical algorithm based on the shooting technique for bounded optimal control problems. The algorithm is applied to a group of immunodominance problems. Numerical simulations reveal that the immune system selects either a broad or a specific strategy of immunodominance based on different optimization goals. The shooting algorithm can also be utilized to solve other complex optimal control problems.

Evaluation of MHC-II peptide binding prediction servers: applications for vaccine research.

BACKGROUND: Initiation and regulation of immune responses in humans involves recognition of peptides presented by human leukocyte antigen class II (HLA-II) molecules. These peptides (HLA-II T-cell epitopes) are increasingly important as research targets for the development of vaccines and immunotherapies. HLA-II peptide binding studies involve multiple overlapping peptides spanning individual antigens, as well as complete viral proteomes. Antigen variation in pathogens and tumor antigens, and extensive polymorphism of HLA molecules increase the number of targets for screening studies. Experimental screening methods are expensive and time consuming and reagents are not readily available for many of the HLA class II molecules. Computational prediction methods complement experimental studies, minimize the number of validation experiments, and significantly speed up the epitope mapping process. We collected test data from four independent studies that involved 721 peptide binding assays. Full overlapping studies of four antigens identified binding affinity of 103 peptides to seven common HLA-DR molecules (DRB1*0101, 0301, 0401, 0701, 1101, 1301, and 1501). We used these data to analyze performance of 21 HLA-II binding prediction servers accessible through the WWW. RESULTS: Because not all servers have predictors for all tested HLA-II molecules, we assessed a total of 113 predictors. The length of test peptides ranged from 15 to 19 amino acids. We tried three prediction strategies - the best 9-mer within the longer peptide, the average of best three 9-mer predictions, and the average of all 9-mer predictions within the longer peptide. The best strategy was the identification of a single best 9-mer within the longer peptide. Overall, measured by the receiver operating characteristic method (AROC), 17 predictors showed good (AROC > 0.8), 41 showed marginal (AROC > 0.7), and 55 showed poor performance (AROC < 0.7). Good performance predictors included HLA-DRB1*0101 (seven), 1101 (six), 0401 (three), and 0701 (one). The best individual predictor was NETMHCIIPAN, closely followed by PROPRED, IEDB (Consensus), and MULTIPRED (SVM). None of the individual predictors was shown to be suitable for prediction of promiscuous peptides. Current predictive capabilities allow prediction of only 50% of actual T-cell epitopes using practical thresholds. CONCLUSION: The available HLA-II servers do not match prediction capabilities of HLA-I predictors. Currently available HLA-II prediction servers offer only a limited prediction accuracy and the development of improved predictors is needed for large-scale studies, such as proteome-wide epitope mapping. The requirements for accuracy of HLA-II binding predictions are stringent because of the substantial effect of false positives.