Precise score for the prediction of peptides cleaved by the proteasome

Counting is almost all you need

The immune memory repertoire encodes the history of present and past infections and immunological attributes of the individual. As such, multiple methods were proposed to use T-cell receptor (TCR) repertoires to detect disease history. We here show that the counting method outperforms two leading algorithms. We then show that the counting can be further improved using a novel attention model to weigh the different TCRs. The attention model is based on the projection of TCRs using a Variational AutoEncoder (VAE). Both counting and attention algorithms predict better than current leading algorithms whether the host had CMV and its HLA alleles. As an intermediate solution between the complex attention model and the very simple counting model, we propose a new Graph Convolutional Network approach that obtains the accuracy of the attention model and the simplicity of the counting model. The code for the models used in the paper is provided at: https://github.com/louzounlab/CountingIsAlmostAllYouNeed.

Naive and memory T cells TCR-HLA-binding prediction

T cells recognize antigens through the interaction of their T cell receptor (TCR) with a peptide-major histocompatibility complex (pMHC) molecule. Following thymic-positive selection, TCRs in peripheral naive T cells are expected to bind MHC alleles of the host. Peripheral clonal selection is expected to further increase the frequency of antigen-specific TCRs that bind to the host MHC alleles. To check for a systematic preference for MHC-binding T cells in TCR repertoires, we developed Natural Language Processing-based methods to predict TCR-MHC binding independently of the peptide presented for Class I MHC alleles. We trained a classifier on published TCR-pMHC binding pairs and obtained a high area under curve (AUC) of over 0.90 on the test set. However, when applied to TCR repertoires, the accuracy of the classifier dropped. We thus developed a two-stage prediction model, based on large-scale naive and memory TCR repertoires, denoted TCR HLA-binding predictor (CLAIRE). Since each host carries multiple human leukocyte antigen (HLA) alleles, we first computed whether a TCR on a CD8 T cell binds an MHC from any of the host Class-I HLA alleles. We then performed an iteration, where we predict the binding with the most probable allele from the first round. We show that this classifier is more precise for memory than for naïve cells. Moreover, it can be transferred between datasets. Finally, we developed a CD4-CD8 T cell classifier to apply CLAIRE to unsorted bulk sequencing datasets and showed a high AUC of 0.96 and 0.90 on large datasets. CLAIRE is available through a GitHub at: https://github.com/louzounlab/CLAIRE, and as a server at: https://claire.math.biu.ac.il/Home.

Cytotoxic T lymphocyte epitopes identified from a contemporary strain of porcine reproductive and respiratory syndrome virus enhance CD4+CD8+ T, CD8+ T, and γδ T cell responses

Immuno-stimulatory class I-restricted cytotoxic T lymphocytes (CTL) epitopes of porcine reproductive and respiratory syndrome virus (PRRSV) are important for vaccine development. In this study we first determined the expression frequency of swine leukocyte antigen (SLA) class I alleles in commercial pigs in the United States. The SLA genotyping result allowed us to predict potential CTL epitopes from a contemporary strain of PRRSV (RFLP 1-7-4) by using bioinformatic tools. The predicted epitopes were then evaluated in an ex vivo stimulation assay with peripheral blood mononuclear cells isolated from pigs experimentally-infected with PRRSV. Using flow-cytometry analysis, we identified a number of immuno-stimulatory CTL epitopes, including two peptides from GP3 and two from Nsp9 that significantly improved both degranulation marker CD107a and IFN-γ production in cytotoxic CD4+CD8+ T cells, CD8+ T cells, and γδ T cells, and two peptides that inhibited IFN-γ production. These CTL epitopes will aid future vaccine development against PRRSV.

Systems Immunology: Learning the Rules of the Immune System

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

Level of neo-epitope predecessor and mutation type determine T cell activation of MHC binding peptides

BACKGROUND

Targeting epitopes derived from neo-antigens (or "neo-epitopes") represents a promising immunotherapy approach with limited off-target effects. However, most peptides predicted using MHC binding prediction algorithms do not induce a CD8 + T cell response, and there is a crucial need to refine the predictions to readily identify the best antigens that could mediate T-cell responses. Such a response requires a high enough number of epitopes bound to the target MHC. This number is correlated with both the peptide-MHC binding affinity and the number of peptides reaching the ER. Beyond this, the response may be affected by the properties of the neo-epitope mutated residues.

METHODS

Herein, we analyzed several experimental datasets from cancer patients to elaborate better predictive algorithms for T-cell reactivity to neo-epitopes.

RESULTS

Indeed, potent classifiers for epitopes derived from neo-antigens in melanoma and other tumors can be developed based on biochemical properties of the mutated residue, the antigen expression level and the peptide processing stage. Among MHC binding peptides, the present classifiers can remove half of the peptides falsely predicted to activate T cells while maintaining the absolute majority of reactive peptides.

CONCLUSIONS

The classifier properties further highlight the contribution of the quantity of peptides reaching the ER and the mutation type to CD8 + T cell responses. These classifiers were then validated on neo-antigens obtained from other datasets, confirming the validity of our prediction. Algorithm Availability: http://peptibase.cs.biu.ac.il/Tcell_predictor/ or by request from the authors as a standalone code.

Computational Approaches for the Discovery of Human Proteasome Inhibitors: An Overview

Proteasome emerged as an important target in recent pharmacological research due to its pivotal role in degrading proteins in the cytoplasm and nucleus of eukaryotic cells, regulating a wide variety of cellular pathways, including cell growth and proliferation, apoptosis, DNA repair, transcription, immune response, and signaling processes. The last two decades witnessed intensive efforts to discover 20S proteasome inhibitors with significant chemical diversity and efficacy. To date, the US FDA approved to market three proteasome inhibitors: bortezomib, carfilzomib, and ixazomib. However new, safer and more efficient drugs are still required. Computer-aided drug discovery has long being used in drug discovery campaigns targeting the human proteasome. The aim of this review is to illustrate selected in silico methods like homology modeling, molecular docking, pharmacophore modeling, virtual screening, and combined methods that have been used in proteasome inhibitors discovery. Applications of these methods to proteasome inhibitors discovery will also be presented and discussed to raise improvements in this particular field.

Estimate of within population incremental selection through branch imbalance in lineage trees

Incremental selection within a population, defined as limited fitness changes following mutation, is an important aspect of many evolutionary processes. Strongly advantageous or deleterious mutations are detected using the synonymous to non-synonymous mutations ratio. However, there are currently no precise methods to estimate incremental selection. We here provide for the first time such a detailed method and show its precision in multiple cases of micro-evolution. The proposed method is a novel mixed lineage tree/sequence based method to detect within population selection as defined by the effect of mutations on the average number of offspring. Specifically, we propose to measure the log of the ratio between the number of leaves in lineage trees branches following synonymous and non-synonymous mutations. The method requires a high enough number of sequences, and a large enough number of independent mutations. It assumes that all mutations are independent events. It does not require of a baseline model and is practically not affected by sampling biases. We show the method's wide applicability by testing it on multiple cases of micro-evolution. We show that it can detect genes and inter-genic regions using the selection rate and detect selection pressures in viral proteins and in the immune response to pathogens.

Role of peptide processing predictions in T cell epitope identification: contribution of different prediction programs

Proteolysis is the general term to describe the process of protein degradation into peptides. Proteasomes are the main actors in cellular proteolysis, and their activity can be measured in in vitro digestion experiments. However, in vivo proteolysis can be different than what is measured in these experiments if other proteases participate or if proteasomal activity is different in vivo. The in vivo proteolysis can be measured only indirectly, by the analysis of peptides presented on MHC-I molecules. MHC-I presented peptides are protected from further degradation, thus enabling an indirect view on the underlying in vivo proteolysis. The ligands presented on different MHC-I molecules enable different views on this process; in combination, they might give a complete picture. Based on in vitro proteasome-only digestions and MHC-I ligand data, different proteolysis predictors have been developed. With new in vitro digestion and MHC-I ligand data sets, we benchmarked how well these predictors capture in vitro proteasome-only activity and in vivo whole-cell proteolysis, respectively. Even though the in vitro proteasome digestion patterns were best captured by methods trained on such data (ProteaSMM and NetChop 20S), the in vivo whole-cell proteolysis was best predicted by a method trained on MHC-I ligand data (NetChop Cterm). Follow-up analysis showed that the likely source of this difference is the activity from proteases other than the proteasome, such as TPPII. This non-proteasomal in vivo activity is captured by NetChop Cterm and should be taken into account in MHC-I ligand predictions.

Viral CD8 T cell epitope nucleotide composition shows evidence of short- and long-term evolutionary strategies

Viral epitopes have a distinct codon usage that reflects their dual role in infection and immunity. On the one hand, epitopes are part of proteins important to viral function; on the other hand, they are targets of the immune response. Studies of selection are most commonly based on changes of amino acid and seen through the accumulation of non-synonymous mutations. An independent measure of selection is the codon usage and underlying changeability of the nucleotide sequences. We here use multiple tools and a large-scale analysis of viral genomes to demonstrate that viral epitopes have a distinct codon usage and that this codon usage reflects distinct short- and long-term types of selection during viral evolution. We show that CD8(+) T cell epitopes are encoded by codons more distant from stop codons and more changeable than codons outside epitopes. This biased codon usage reflects the viral population toggling back and forth from a wild-type sequence to an escape mode, which enable them to avoid immune detection when needed, and go back to the functionally favorable form when the threat is removed (i.e., in a new host).

Computational prediction of cleavage using proteasomal in vitro digestion and MHC I ligand data

Proteasomes are responsible for the production of the majority of cytotoxic T lymphocyte (CTL) epitopes. Hence, it is important to identify correctly which peptides will be generated by proteasomes from an unknown protein. However, the pool of proteasome cleavage data used in the prediction algorithms, whether from major histocompatibility complex (MHC) I ligand or in vitro digestion data, is not identical to in vivo proteasomal digestion products. Therefore, the accuracy and reliability of these models still need to be improved. In this paper, three types of proteasomal cleavage data, constitutive proteasome (cCP), immunoproteasome (iCP) in vitro cleavage, and MHC I ligand data, were used for training cleave-site predictive methods based on the kernel-function stabilized matrix method (KSMM). The predictive accuracies of the KSMM+pair coefficients were 75.0%, 72.3%, and 83.1% for cCP, iCP, and MHC I ligand data, respectively, which were comparable to the results from support vector machine (SVM). The three proteasomal cleavage methods were combined in turn with MHC I-peptide binding predictions to model MHC I-peptide processing and the presentation pathway. These integrations markedly improved MHC I peptide identification, increasing area under the receiver operator characteristics (ROC) curve (AUC) values from 0.82 to 0.91. The results suggested that both MHC I ligand and proteasomal in vitro degradation data can give an exact simulation of in vivo processed digestion. The information extracted from cCP and iCP in vitro cleavage data demonstrated that both cCP and iCP are selective in their usage of peptide bonds for cleavage.

Viral proteome size and CD8+ T cell epitope density are correlated: the effect of complexity on selection

•

We analyze the relation between viral complexity and their adaptation to the host immune system.

•

Viruses with few proteins and low number of nucleotides remove more CD8+ T cell epitopes.

•

Within a virus, short proteins (with fewer amino acids) adapt better than long ones.

•

The relation between total size and adaptation is host specific.

•

Complexity limits genetic adaptation in the high-mutation rate strong selection regime.

The relation between the complexity of organisms and proteins and their evolution rates has been discussed in the context of multiple generic models. The main robust claim from most such models is the negative relation between complexity and the accumulation rate of mutations.

Viruses accumulate escape mutations in their epitopes to avoid detection and destruction of their host cell by CD8+ T cells. The extreme regime of immune escape, namely, strong selection and high mutation rate, provide an opportunity to extend and validate the existing models of relation between complexity and evolution rate as proposed by Fisher and Kimura.

Using epitope prediction algorithms to compute the epitopes presented on the most frequent human HLA alleles in over 100 fully sequenced human viruses, and over 900 non-human viruses, we here study the correlation between viruses/proteins complexity (as measured by the number of proteins in the virus and the length of each protein, respectively) and the rate of accumulation of escape mutation. The latter is evaluated by measuring the normalized epitope density of viral proteins.

If the virus/protein complexity prevents the accumulation of escape mutations, the epitope density is expected to be positively correlated with both the number of proteins in the virus and the length of proteins. We show that such correlations are indeed observed for most human viruses. For non-human viruses the correlations were much less significant, indicating that the correlation is indeed induced by human HLA molecules.

Establishing a relationship between bacteria in the human gut and complex regional pain syndrome

Complex Regional Pain Syndrome (CRPS) is a serious and painful condition involving the peripheral and central nervous systems. Full comprehension of the disorder's pathophysiology remains incomplete, but research implicates the immune system as a contributor to chronic pain. Because of the impact gastrointestinal bacteria have in the development and behavior of the immune system, this study compares the GI microbial communities of 16 participants with CRPS (5 of whom have intestinal discomforts) and 16 healthy controls using 454 sequencing technology. CRPS subjects were found to have significantly less diversity than their healthy counterparts. Statistical analysis of the phylogenetic classifications revealed significantly increased levels of Proteobacteria and decreased levels of Firmicutes in CRPS subjects. Clustering analysis showed significant separation between healthy controls and CRPS subjects. These results support the hypothesis that the GI microbial communities of CRPS participants differ from those of their healthy counterparts. These variations may hold the key to understanding how CRPS develops and provide information that could yield a potential treatment.

Predictor for the effect of amino acid composition on CD4+ T cell epitopes preprocessing

Predictive tools for all levels of CD8+ T cell epitopes processing have reached a maturation level. Good prediction algorithms have been developed for proteasomal cleavage, TAP and MHC class I peptide binding. The same cannot be said of CD4+ T cell epitopes. While multiple algorithms of varying accuracy have been proposed for MHC class II peptide binding, the preprocessing of CD4+ T cell epitopes is still lacking a good prediction algorithm. CD4+ T cell epitopes generation includes several stages, not all which are well-defined. We here group these stages to produce a generic preprocessing stage predictor for the cleavage processes preceding the presentation of epitopes to CD4+ T cell. The predictor is learnt using a combination of in vitro cleavage experiments and observed naturally processed MHC class II binding peptides. The properties of the predictor highlight the effect of different factors on CD4+ T cell epitopes preprocessing. The most important factor emerging from the predictor is the secondary structure of the cleaved region in the protein. The effect of the secondary structure is expected since CD4+ T cell epitopes are not denatured before cleavage. A website developed based on this predictor is available at: http://peptibase.cs.biu.ac.il/PepCleave_cd4/.

Evolution of viral life-cycle in response to cytotoxic T lymphocyte-mediated immunity

Viruses in mammals are constantly faced with the problem of elimination by the host immunity. Cytotoxic T lymphocyte (CTL) responses are thought to play a major role in the control and clearance of several viral infections in mice and humans. It is therefore expected that over evolutionary time, viruses would be forced to evolve to avoid recognition by CTLs. Indeed, a number of studies have documented the accumulation of viral variants with escape mutations. These mutations allow viruses to hide from CTL responses common in the host population. CTLs recognize viruses by short protein sequences, named epitopes, derived from viral proteins. The efficiency of viral recognition by epitope-specific CTL responses depends on the expression pattern of the proteins carrying these epitopes, and the total amount of that protein (and thus epitopes) in the cell. When a virus replicates in a cell, some viral genes are expressed early in the life cycle of the virus, while other proteins are expressed late. For example, HIV infected cells first express Rev and Tat proteins, and the Gag proteins are expressed late. Here we propose a dynamical model of the viral life cycle to study how expression level of early vs. late genes may affect viral dynamics within the host and virus transmission over the course of infection. We find that for acute and chronic viral infections lower expression of early genes than that of the late genes is expected to give selective advantage and higher transmission to viruses.

Structure of the proteasome

The ubiquitin-proteasomal system is an essential element of the protein quality control machinery in cells. The central part of this system is the 20S proteasome. The proteasome is a barrel-shaped multienzyme complex, containing several active centers hidden at the inner surface of the hollow cylinder. So, the regulation of the substrate entry toward the inner proteasomal surface is a key control mechanism of the activity of this protease. This chapter outlines the knowledge on the structure of the subunits of the 20S proteasome, the binding and structure of some proteasomal regulators and inducible proteasomal subunits. Therefore, this chapter imparts the knowledge on proteasomal structure which is required for the understanding of the following chapters.

Bacteria modulate the CD8+ T cell epitope repertoire of host cytosol-exposed proteins to manipulate the host immune response

The main adaptive immune response to bacteria is mediated by B cells and CD4+ T-cells. However, some bacterial proteins reach the cytosol of host cells and are exposed to the host CD8+ T-cells response. Both gram-negative and gram-positive bacteria can translocate proteins to the cytosol through type III and IV secretion and ESX-1 systems, respectively. The translocated proteins are often essential for the bacterium survival. Once injected, these proteins can be degraded and presented on MHC-I molecules to CD8+ T-cells. The CD8+ T-cells, in turn, can induce cell death and destroy the bacteria's habitat. In viruses, escape mutations arise to avoid this detection. The accumulation of escape mutations in bacteria has never been systematically studied. We show for the first time that such mutations are systematically present in most bacteria tested. We combine multiple bioinformatic algorithms to compute CD8+ T-cell epitope libraries of bacteria with secretion systems that translocate proteins to the host cytosol. In all bacteria tested, proteins not translocated to the cytosol show no escape mutations in their CD8+ T-cell epitopes. However, proteins translocated to the cytosol show clear escape mutations and have low epitope densities for most tested HLA alleles. The low epitope densities suggest that bacteria, like viruses, are evolutionarily selected to ensure their survival in the presence of CD8+ T-cells. In contrast with most other translocated proteins examined, Pseudomonas aeruginosa's ExoU, which ultimately induces host cell death, was found to have high epitope density. This finding suggests a novel mechanism for the manipulation of CD8+ T-cells by pathogens. The ExoU effector may have evolved to maintain high epitope density enabling it to efficiently induce CD8+ T-cell mediated cell death. These results were tested using multiple epitope prediction algorithms, and were found to be consistent for most proteins tested.

Optimal viral immune surveillance evasion strategies

Following cell entry, viruses can be detected by cytotoxic T lymphocytes. These cytotoxic T lymphocytes can induce host cell apoptosis and prevent the propagation of the virus. Viruses with fewer epitopes have a higher survival probability, and are selected through evolution. However, mutations have a fitness cost and on evolutionary periods viruses maintain some epitopes. The number of epitopes in each viral protein is a balance between the selective advantage of having fewer epitopes and the reduced fitness following the epitope removing mutations. We discuss a bioinformatic analysis of the number of epitopes in various viral proteins and propose an optimization framework to explain these numbers. We show, using a genomic analysis and a theoretical optimization framework, that a critical factor affecting the number of presented epitopes is the expression stage in the viral life cycle of the gene coding for the protein. The early expression of epitopes can lead to the destruction of the host cell before budding can take place. We show that a lower number of epitopes is expected in early proteins even if late proteins have a much higher copy number.

Universal peptide vaccines - optimal peptide vaccine design based on viral sequence conservation

Rapidly mutating viruses such as the hepatitis C virus (HCV), the human immunodeficiency virus (HIV), or influenza viruses (Flu) call for highly effective universal peptide vaccines, i.e. vaccines that do not only yield broad population coverage but also broad coverage of various viral strains. The efficacy of such vaccines is determined by multiple properties of the epitopes they comprise. Beyond the specific properties of each epitope, properties of the corresponding source antigens are of great importance. If a response is mounted against viral proteins with a low copy number within the cell or against proteins expressed very late, this response may fail to induce lysis of the infected cells before budding can take place. We here propose a novel methodology to optimize the epitope composition and assembly in order to induce maximum protection. In order for a peptide vaccine to yield the best possible universal protection, several conditions should be met: (a) an optimal choice of target antigens, (b) an optimal choice of highly conserved epitopes, (c) maximum coverage of the target population, and (d) the proper ordering of the epitopes in the final vaccine to ensure favorable cleavage. We propose a mathematical formalism for epitope selection and ordering that balances the constraints imposed by these different conditions. Focusing on HCV, HIV, and Flu, we show that not all of the conditions can be satisfied for all viruses. Depending on the virus, different constraints are harder to fulfill: for Flu, the conservation constraint is violated first, while for HIV, it is difficult to focus the response at the optimal target antigens. The proposed methodology can be applied to any virus to assess the feasibility of optimally combining the above-mentioned constraints.

Whole-genome immunoinformatic analysis of F. tularensis: predicted CTL epitopes clustered in hotspots are prone to elicit a T-cell response

The cellular arm of the immune response plays a central role in the defense against intracellular pathogens, such as F. tularensis. To date, whole genome immunoinformatic analyses were limited either to relatively small genomes (e.g. viral) or to preselected subsets of proteins in complex pathogens. Here we present, for the first time, an unbiased bacterial global immunoinformatic screen of the 1740 proteins of F. tularensis subs. holarctica (LVS), aiming at identification of immunogenic peptides eliciting a CTL response. The very large number of predicted MHC class I binders (about 100,000, IC₅₀ of 1000 nM or less) required the design of a strategy for further down selection of CTL candidates. The approach developed focused on mapping clusters rich in overlapping predicted epitopes, and ranking these “hotspot” regions according to the density of putative binding epitopes. Limited by the experimental load, we selected to screen a library of 1240 putative MHC binders derived from 104 top-ranking highly dense clusters. Peptides were tested for their ability to stimulate IFNγ secretion from splenocytes isolated from LVS vaccinated C57BL/6 mice. The majority of the clusters contained one or more CTL responder peptides and altogether 127 novel epitopes were identified, of which 82 are non-redundant. Accordingly, the level of success in identification of positive CTL responders was 17–25 fold higher than that found for a randomly selected library of 500 predicted MHC binders (IC₅₀ of 500 nM or less). Most proteins (ca. 2/3) harboring the highly dense hotspots are membrane-associated. The approach for enrichment of true positive CTL epitopes described in this study, which allowed for over 50% increase in the dataset of known T-cell epitopes of F. tularensis, could be applied in immunoinformatic analyses of many other complex pathogen genomes.

Signal peptides and trans-membrane regions are broadly immunogenic and have high CD8+ T cell epitope densities: Implications for vaccine development

Cell mediated immune response has a major role in controlling the elimination of infectious agents. The rational design of sub-unit peptide vaccines against intracellular pathogens or cancer requires the use of antigenic sequence/s that can induce highly potent, long lasting and antigen-specific responses in the majority of the population. A promising peptide selection strategy is the detection of multi-epitope peptide sequences with an ability to bind multiple MHC alleles. While past research sought the best epitopes based on their specific antigenicity, we ask whether specific defined domains have high epitope densities. Signal peptides and trans-membrane domains were found to have exceptionally high epitope densities. The improved MHC binding of these domains relies on their hydrophobic nature and, in signal peptides, also on their specific sequence. The high epitope density of SP was computed using in-silico methods and corroborated by the high percentage of identified SP epitope in the IEDB (immune epitope database). The enhanced immunogenicity of SP was then experimentally confirmed using a panel of nine peptides derived from Mycobacterium tuberculosis (MTb) proteins used in human PBMC proliferation assays and T cell lines functional assays. Our results show the exceptionally high antigen specific response rates and population coverage to SP sequences compared with non-SP peptide antigens derived from the same proteins. The results suggest a novel scheme for the rational design of T cell vaccines using a domain based rather than an epitope based approach.

Immune-induced evolutionary selection focused on a single reading frame in overlapping hepatitis B virus proteins

Viruses employ various means to evade immune detection. Reduction of CD8(+) T cell epitopes is one of the common strategies used for this purpose. Hepatitis B virus (HBV), a member of the Hepadnaviridae family, has four open reading frames, with about 50% overlap between the genes they encode. We computed the CD8(+) T cell epitope density within HBV proteins and the mutations within the epitopes. Our results suggest that HBV accumulates escape mutations that reduce the number of epitopes. These mutations are not equally distributed among genes and reading frames. While the highly expressed core and X proteins are selected to have low epitope density, polymerase, which is expressed at low levels, does not undergo the same selection. In overlapping regions, mutations in one protein-coding sequence also affect the other protein-coding sequence. We show that mutations lead to the removal of epitopes in X and surface proteins even at the expense of the addition of epitopes in polymerase. The total escape mutation rate for overlapping regions is lower than that for nonoverlapping regions. The lower epitope replacement rate for overlapping regions slows the evolutionary escape rate of these regions but leads to the accumulation of mutations more robust in the transfer between hosts, such as mutations preventing proteasomal cleavage into epitopes.

State of the art and challenges in sequence based T-cell epitope prediction

Sequence based T-cell epitope predictions have improved immensely in the last decade. From predictions of peptide binding to major histocompatibility complex molecules with moderate accuracy, limited allele coverage, and no good estimates of the other events in the antigen-processing pathway, the field has evolved significantly. Methods have now been developed that produce highly accurate binding predictions for many alleles and integrate both proteasomal cleavage and transport events. Moreover have so-called pan-specific methods been developed, which allow for prediction of peptide binding to MHC alleles characterized by limited or no peptide binding data. Most of the developed methods are publicly available, and have proven to be very useful as a shortcut in epitope discovery. Here, we will go through some of the history of sequence-based predictions of helper as well as cytotoxic T cell epitopes. We will focus on some of the most accurate methods and their basic background.

Computational analysis and modeling of cleavage by the immunoproteasome and the constitutive proteasome

Background

Proteasomes play a central role in the major histocompatibility class I (MHCI) antigen processing pathway. They conduct the proteolytic degradation of proteins in the cytosol, generating the C-terminus of CD8 T cell epitopes and MHCI-peptide ligands (P1 residue of cleavage site). There are two types of proteasomes, the constitutive form, expressed in most cell types, and the immunoproteasome, which is constitutively expressed in mature dendritic cells. Protective CD8 T cell epitopes are likely generated by the immunoproteasome and the constitutive proteasome, and here we have modeled and analyzed the cleavage by these two proteases.

Results

We have modeled the immunoproteasome and proteasome cleavage sites upon two non-overlapping sets of peptides consisting of 553 CD8 T cell epitopes, naturally processed and restricted by human MHCI molecules, and 382 peptides eluted from human MHCI molecules, respectively, using N-grams. Cleavage models were generated considering different epitope and MHCI-eluted fragment lengths and the same number of C-terminal flanking residues. Models were evaluated in 5-fold cross-validation. Judging by the Mathew's Correlation Coefficient (MCC), optimal cleavage models for the proteasome (MCC = 0.43 ± 0.07) and the immunoproteasome (MCC = 0.36 ± 0.06) were obtained from 12-residue peptide fragments. Using an independent dataset consisting of 137 HIV1-specific CD8 T cell epitopes, the immunoproteasome and proteasome cleavage models achieved MCC values of 0.30 and 0.18, respectively, comparatively better than those achieved by related methods. Using ROC analyses, we have also shown that, combined with MHCI-peptide binding predictions, cleavage predictions by the immunoproteasome and proteasome models significantly increase the discovery rate of CD8 T cell epitopes restricted by different MHCI molecules, including A*0201, A*0301, A*2402, B*0702, B*2705.

Conclusions

We have developed models that are specific to predict cleavage by the proteasome and the immunoproteasome. These models ought to be instrumental to identify protective CD8 T cell epitopes and are readily available for free public use at http://imed.med.ucm.es/Tools/PCPS/.

Towards in silico design of epitope-based vaccines

BACKGROUND

Epitope-based vaccines (EVs) make use of immunogenic peptides (epitopes) to trigger an immune response. Due to their manifold advantages, EVs have recently been attracting growing interest. The success of an EV is determined by the choice of epitopes used as a basis. However, the experimental discovery of candidate epitopes is expensive in terms of time and money. Furthermore, for the final choice of epitopes various immunological requirements have to be considered.

METHODS

Numerous in silico approaches exist that can guide the design of EVs. In particular, computational methods for MHC binding prediction have already become standard tools in immunology. Apart from binding prediction and prediction of antigen processing, methods for epitope design and selection have been suggested. We review these in silico approaches for epitope discovery and selection along with their strengths and weaknesses. Finally, we discuss some of the obvious problems in the design of EVs.

CONCLUSION

State-of-the-art in silico approaches to MHC binding prediction yield high accuracies. However, a more thorough understanding of the underlying biological processes and significant amounts of experimental data will be required for the validation and improvement of in silico approaches to the remaining aspects of EV design.

Viruses selectively mutate their CD8+ T-cell epitopes--a large-scale immunomic analysis

Motivation: Viruses employ various means to evade immune detection. One common evasion strategy is the removal of CD8+cytotoxic T-lymphocyte epitopes. We here use a combination of multiple bioinformatic tools and large amount of genomic data to compute the epitope repertoire presented by over 1300 viruses in many HLA alleles. We define the ‘Size of Immune Repertoire score’, which represents the ratio between the epitope density within a protein and the expected density. This score is used to study viral immune evasion.

Results: We show that viral proteins in general have a higher epitope density than human proteins. This difference is due to a good fit of the human MHC molecules to the typical amino-acid usage of viruses. Among different viruses, viruses infecting humans present less epitopes than non-human viruses. This selection is not at the amino-acid usage level, but through the removal of specific epitopes. Within a single virus, not all proteins express the same epitopes density. Proteins expressed early in the viral life cycle have a lower epitope density than late proteins. Such a difference is not observed in non-human viruses. The removal of early epitopes and the targeting of the cellular immune response to late viral proteins, allow the virus a time interval to propagate before its host cells are destroyed by T cells.

Contact:louzouy@math.biu.ac.il

The HIV hide and seek game: an immunogenomic analysis of the HIV epitope repertoire

BACKGROUND

Viruses employ various means to evade immune detection. One common evasion strategy is the removal of CD8 cytotoxic T-lymphocyte (CTL) epitopes.

METHOD

Here, we use bioinformatic tools to compute the HIV CTL epitope repertoire presented by over 8000 HIV sequences in multiple Human Leukocyte Antigen alleles. We define the 'Size of Immune Repertoire' (SIR) score, which represents the ratio between the number of the predicted epitopes within a protein and their expected number within a scrambled version of the same protein.

RESULTS

We show that HIV proteins present less epitopes than expected and that the number of epitopes gradually decreases from SIV to recent HIV sequences. The decrease of the SIR score of HIV is accompanied by a high frequency of replacement mutations within epitopes. The SIR score of the different HIV proteins is not uniform. The regulatory proteins, Tat and Rev, expressed early during cellular infection have a low SIR score, whereas virion-associated genes that are expressed later, such as Env, Pol and Gag, have a higher SIR score. Actually, the SIR score of Gag keeps increasing over time.

CONCLUSION

We hypothesize that our results reflect an HIV immune evasion strategy. This involves the targeting of the CTL immune response to viral structural and enzyme proteins, allowing the virus a time interval to propagate before its host cells are destroyed by CTLs. An efficient anti-HIV CTL response against HIV should thus also target the regulatory genes that HIV seeks to hide from the immune system.

Human self-protein CD8+ T-cell epitopes are both positively and negatively selected

The cellular immune system recognizes self-epitopes in the context of MHC-I molecules. The immunological general view presumes that these self-epitopes are just a background, both positively and negatively selecting T cells. We here estimate the number of epitopes in each human protein for many frequent HLA alleles, and a score representing over or under presentation of epitopes on these proteins. We further show that there is a clear selection for the presentation of specific self-protein types. Proteins presenting many epitopes include, for example, autoimmune regulator (AIRE) upregulated tissue-specific antigens, immune system receptors and proteins with a high expression level. On the other hand, proteins that may be considered less "useful" for the immune system, such as low expression level proteins, are under-presented. We combine our epitope estimate with single nucleotide polymorphism (SNP) measures to show that this selection can be directly observed through the fraction of non-synonymous SNP (replacement fraction), which is significantly higher inside epitopes than outside.