Large-scale validation of methods for cytotoxic T-lymphocyte epitope prediction

Background  

Reliable predictions of Cytotoxic T lymphocyte (CTL) epitopes are essential for rational vaccine design. Most importantly, they can minimize the experimental effort needed to identify epitopes. NetCTL is a web-based tool designed for predicting human CTL epitopes in any given protein. It does so by integrating predictions of proteasomal cleavage, TAP transport efficiency, and MHC class I affinity. At least four other methods have been developed recently that likewise attempt to predict CTL epitopes: EpiJen, MAPPP, MHC-pathway, and WAPP. In order to compare the performance of prediction methods, objective benchmarks and standardized performance measures are needed. Here, we develop such large-scale benchmark and corresponding performance measures and report the performance of an updated version 1.2 of NetCTL in comparison with the four other methods.     

Construction and immunogenicity prediction of Plasmodium falciparum CTL epitope minigene vaccine

The minigenes encoding Plasmodiumfalciparum CTL epitopes restricted to human MHC class I molecular HLA-A2 and HLA-B51, which were both at high frequency among Chinese population, were constructed as mono-epitope CTL vaccines named pcDNA3.1/tr and pcDNA3.1/sh. The minigenes of the two epitopes were then tandem linked to form a dimeric CTL epitope minigene recombinant vaccine. After DNA transfection, the epitope minigenes were expressed respectively in two human cell lines, each bearing one MHC class I molecule named CIR/HLA-A2.1 and K562/HLA-B51. The intracellular expression of the CTL epitope minigenes not only enhanced the stability of HLA-A2.1 and HLA-B51 molecules but also increased the assemblage of MHC class I molecules on cell surfaces, which testified the specific process and presentation of those endogenous expressed epitopes. For the cells transfected with the dimeric minigene encoding two tandem linked epitopes, the expression and presentation of each epitope were also detected on cell membranes that bore different MHC class I molecules. It meant that the adjacency of the two CTL epitopes did not interfere with the specific process and presentation of each epitope. Compared with the ordinary CTL studies that inoculated synthesized epitope peptides with peripheral blood cells, this work aimed to process the epitopes directly inside HLA class I allele specific human cells, and thus theoretically imitated the same procedurein vivo. It was also an economical way to predict the immunogenicity of CTL epitopes at an early stage especially in laboratories with limited financial resource.     

Integrated modeling of the major events in the MHC class I antigen processing pathway

Rational design of epitope-driven vaccines is a key goal of immunoinformatics. Typically, candidate selection relies on the prediction of MHC-peptide binding only, as this is known to be the most selective step in the MHC class I antigen processing pathway. However, proteasomal cleavage and transport by the transporter associated with antigen processing (TAP) are essential steps in antigen processing as well. While prediction methods exist for the individual steps, no method has yet offered an integrated prediction of all three major processing events. Here we present WAPP, a method combining prediction of proteasomal cleavage, TAP transport, and MHC binding into a single prediction system. The proteasomal cleavage site prediction employs a new matrix-based method that is based on experimentally verified proteasomal cleavage sites. Support vector regression is used for predicting peptides transported by TAP. MHC binding is the last step in the antigen processing pathway and was predicted using a support vector machine method, SVMHC. The individual methods are combined in a filtering approach mimicking the natural processing pathway. WAPP thus predicts peptides that are cleaved by the proteasome at the C terminus, transported by TAP, and show significant affinity to MHC class I molecules. This results in a decrease in false positive rates compared to MHC binding prediction alone. Compared to prediction of MHC binding only, we report an increased overall accuracy and a lower rate of false positive predictions for the HLA-A*0201, HLA-B*2705, HLA-A*01, and HLA-A*03 alleles using WAPP. The method is available online through our prediction server at http://www-bs.informatik.uni-tuebingen.de/WAPP     

Antigen processing by proteasomes: insights into the molecular basis of crypticity

Eight to eleven amino acid residues are the sizes of predominant peptides found to be associated with MHC class I molecules. Proteasomes have been implicated in antigen processing and generation of such peptides. Advanced methodologies in peptide elution together with sequence determination have led to the characterisation of MHC class I binding motifs. More recently, screening of random peptide phage display libraries and synthetic combinatorial peptide libraries have also been successfully used. This has led to the development and use of predictive algorithms to screen antigens for potential CTL epitopes. Not all predicted epitopes will be generated in vivo and the emerging picture suggests differential presentation of predicted CTL epitopes ranging from cryptic to immunodominant. The scope of this review is to discuss antigen processing by proteasomes, and to put forward a hypothesis that the molecular basis of immunogenicity can be a function of proteasomal processing. This may explain how pathogens and tumours are able to escape immunosurveillance by altering sequences required by proteasomes for epitope generation. Abbreviations: CTL – cytotoxic T lymphocytes; DRiPs – defective ribosomal products; ER – endoplasmic reticulum; Hsps – heat shock proteins; LMP – low molecular weight peptide; MHC – major histocompatibility complex; TAP – transporter associated with antigen processing.     

Enhancement to the RANKPEP resource for the prediction of peptide binding to MHC molecules using profiles

We introduced previously an on-line resource, RANKPEP that uses position specific scoring matrices (PSSMs) or profiles for the prediction of peptide-MHC class I (MHCI) binding as a basis for CD8 T-cell epitope identification. Here, using PSSMs that are structurally consistent with the binding mode of MHC class II (MHCII) ligands, we have extended RANKPEP to prediction of peptide-MHCII binding and anticipation of CD4 T-cell epitopes. Currently, 88 and 50 different MHCI and MHCII molecules, respectively, can be targeted for peptide binding predictions in RANKPEP. Because appropriate processing of antigenic peptides must occur prior to major histocompatibility complex (MHC) binding, cleavage site prediction methods are important adjuncts for T-cell epitope discovery. Given that the C-terminus of most MHCI-restricted epitopes results from proteasomal cleavage, we have modeled the cleavage site from known MHCI-restricted epitopes using statistical language models. The RANKPEP server now determines whether the C-terminus of any predicted MHCI ligand may result from such proteasomal cleavage. Also implemented is a variability masking function. This feature focuses prediction on conserved rather than highly variable protein segments encoded by infectious genomes, thereby offering identification of invariant T-cell epitopes to thwart mutation as an immune evasion mechanism.     

Identifying MHC class I epitopes by predicting the TAP transport efficiency of epitope precursors

We are able to make reliable predictions of the efficiency with which peptides of arbitrary lengths will be transported by TAP. The pressure exerted by TAP on Ag presentation thus can be assessed by checking to what extent MHC class I (MHC-I)-presented epitopes can be discriminated from random peptides on the basis of predicted TAP transport efficiencies alone. Best discriminations were obtained when N-terminally prolonged epitope precursor peptides were included and the contribution of the N-terminal residues to the score were down-weighted in comparison with the contribution of the C terminus. We provide evidence that two factors may account for this N-terminal down-weighting: 1) the uncertainty as to which precursors are used in vivo and 2) the coevolution in the C-terminal sequence specificities of TAP and other agents in the pathway, which may vary among the various MHC-I alleles. Combining predictions of MHC-I binding affinities with predictions of TAP transport efficiency led to an improved identification of epitopes, which was not the case when predictions of MHC-I binding affinities were combined with predictions of C-terminal cleavages made by the proteasome.     

Construction and immunogenicity prediction of Plasmodiumfalciparum CTL epitope minigene vaccine

The minigenes encoding Plasmodium falciparum CTL epitopes restricted to human MHC class I molecular HLA-A2 and HLA-B51, which were both at high frequency among Chinese population, were constructed as mono-epitope CTL vaccines named pcDNA3.1/frand pcDNA3.1/ sh. The minigenes of the two epitopes were then tandem linked to form a dimeric CTL epitope minigene recombinant vaccine. After DNA transfection, the epitope minigenes were expressed respectively in two human cell lines, each bearing one MHC class I molecule named CIR/HLA-A2.1 and K562/HLA-B51. The intraceliular expression of the CTL epitope minigenes not only enhanced the stability of HLA-A2.1 and HLA-B51 molecules but also increased the assemblage of MHC class I molecules on cell surfaces, which testified the specific process and presentation of those endogenous expressed epitopes. For the cells transfected with the dimeric minigene encoding two tandem linked epitopes, the expression and presentation of each epitope were also detected on cell membran     

The specificity and polymorphism of the MHC class I prevents the global adaptation of HIV-1 to the monomorphic proteasome and TAP

The large diversity in MHC class I molecules in a population lowers the chance that a virus infects a host to which it is pre-adapted to escape the MHC binding of CTL epitopes. However, viruses can also lose CTL epitopes by escaping the monomorphic antigen processing components of the pathway (proteasome and TAP) that create the epitope precursors. If viruses were to accumulate escape mutations affecting these monomorphic components, they would become pre-adapted to all hosts regardless of the MHC polymorphism. To assess whether viruses exploit this apparent vulnerability, we study the evolution of HIV-1 with bioinformatic tools that allow us to predict CTL epitopes, and quantify the frequency and accumulation of antigen processing escapes. We found that within hosts, proteasome and TAP escape mutations occur frequently. However, on the population level these escapes do not accumulate: the total number of predicted epitopes and epitope precursors in HIV-1 clade B has remained relatively constant over the last 30 years. We argue that this lack of adaptation can be explained by the combined effect of the MHC polymorphism and the high specificity of individual MHC molecules. Because of these two properties, only a subset of the epitope precursors in a host are potential epitopes, and that subset differs between hosts. We estimate that upon transmission of a virus to a new host 39%-66% of the mutations that caused epitope precursor escapes are released from immune selection pressure.     

The varicellovirus-encoded TAP inhibitor UL49.5 regulates the presentation of CTL epitopes by Qa-1b1

Impairment of MHC class I Ag processing is a commonly observed mechanism that allows viruses and tumors to escape immune destruction by CTL. The peptide transporter TAP that is responsible for the delivery of MHC class I-binding peptides into the endoplasmic reticulum is a pivotal target of viral-immune evasion molecules, and expression of this transporter is frequently lost in advanced cancers. We recently described a novel population of CTL that intriguingly exhibits reactivity against such tumor-immune escape variants and that recognizes self-peptides emerging at the cell surface due to defects in the processing machinery. Investigations of this new type of CTL epitopes are hampered by the lack of an efficient inhibitor for peptide transport in mouse cells. In this article, we demonstrate that the varicellovirus protein UL49.5, in contrast to ICP47 and US6, strongly impairs the activity of the mouse transporter and mediates degradation of mouse TAP1 and TAP2. Inhibition of TAP was witnessed by a strong reduction of surface MHC class I display and a decrease in recognition of conventional tumor-specific CTL. Analysis of CTL reactivity through the nonclassical molecule Qa-1(b) revealed that the presentation of the predominant leader peptide was inhibited. Interestingly, expression of UL49.5 in processing competent tumor cells induced the presentation of the new category of peptides. Our data show that the varicellovirus UL49.5 protein is a universal TAP inhibitor that can be exploited for preclinical studies on CTL-based immune intervention.     

MAPPP: MHC class I antigenic peptide processing prediction

MAPPP is a bioinformatics tool for the prediction of potential antigenic epitopes presented on the cell surface by major histocompatibility complex class I (MHC I) molecules to CD8 positive T lymphocytes. It combines existing predictions for proteasomal cleavage with peptide anchoring to MHC I molecules.     

Intramembrane proteolysis of signal peptides: an essential step in the generation of HLA-E epitopes.

Signal sequences of human MHC class I molecules are a unique source of epitopes for newly synthesized nonclassical HLA-E molecules. Binding of such conserved peptides to HLA-E induces its cell surface expression and protects cells from NK cell attack. After cleavage from the pre-protein, we show that the liberated MHC class I signal peptide is further processed by signal peptide peptidase in the hydrophobic, membrane-spanning region. This cut is essential for the release of the HLA-E epitope-containing fragment from the lipid bilayer and its subsequent transport into the lumen of the endoplasmic reticulum via the TAP.     

An empirical method for the prediction of T-cell epitopes

Identification of T-cell epitopes from foreign proteins is the current focus of much research. Methods using simple two or three position motifs have proved useful in epitope prediction for major histocompatibility complex (MHC) class I, but to date not for MHC class II molecules. We utilized data from pool sequence analysis of peptides eluted from two HLA-DR13 alleles to construct a computer algorithm for predicting the probability that a given sequence will be naturally processed and presented on these alleles. We assessed the ability of this method to predict know self-peptides from these DR-13 alleles, DRB1 ^*1301 and *1302, as well as an immunodominant T-cell epitope. We also compared the predictions of this scoring procedure with the measured binding affinities of a panel of overlapping peptides from hepatitis B virus surface antigen. We concluded that this method may have wide application for the prediction of T-cell epitopes for both MHC class I and class II molecules.     

EpiJen: a server for multistep T cell epitope prediction

Background  

The main processing pathway for MHC class I ligands involves degradation of proteins by the proteasome, followed by transport of products by the transporter associated with antigen processing (TAP) to the endoplasmic reticulum (ER), where peptides are bound by MHC class I molecules, and then presented on the cell surface by MHCs. The whole process is modeled here using an integrated approach, which we call EpiJen. EpiJen is based on quantitative matrices, derived by the additive method, and applied successively to select epitopes. EpiJen is available free online.     

The varicellovirus UL49.5 protein blocks the transporter associated with antigen processing (TAP) by inhibiting essential conformational transitions in the 6+6 transmembrane TAP core complex

TAP translocates virus-derived peptides from the cytosol into the endoplasmic reticulum, where the peptides are loaded onto MHC class I molecules. This process is crucial for the detection of virus-infected cells by CTL that recognize the MHC class I-peptide complexes at the cell surface. The varicellovirus bovine herpesvirus 1 encodes a protein, UL49.5, that acts as a potent inhibitor of TAP. UL49.5 acts in two ways, as follows: 1) by blocking conformational changes of TAP required for the translocation of peptides into the endoplasmic reticulum, and 2) by targeting TAP1 and TAP2 for proteasomal degradation. At present, it is unknown whether UL49.5 interacts with TAP1, TAP2, or both. The contribution of other members of the peptide-loading complex has not been established. Using TAP-deficient cells reconstituted with wild-type and recombinant forms of TAP1 and TAP2, TAP was defined as the prime target of UL49.5 within the peptide-loading complex. The presence of TAP1 and TAP2 was required for efficient interaction with UL49.5. Using deletion mutants of TAP1 and TAP2, the 6+6 transmembrane core complex of TAP was shown to be sufficient for UL49.5 to interact with TAP and block its function. However, UL49.5-induced inhibition of peptide transport was most efficient in cells expressing full-length TAP1 and TAP2. Inhibition of TAP by UL49.5 appeared to be independent of the presence of other peptide-loading complex components, including tapasin. These results demonstrate that UL49.5 acts directly on the 6+6 transmembrane TAP core complex of TAP by blocking essential conformational transitions required for peptide transport.     

T cell recognition of QA-1b antigens on cells lacking a functional Tap-2 transporter.

MHC class Ia H chains and beta 2-microglobulin assemble with appropriate peptides to form stable cell surface molecules that serve as targets for Ag-specific CTL. The structural similarities of class Ia and the less polymorphic Q/T/M (class Ib) molecules suggest that class Ib molecules also play a role in antigen presentation, although the origin of the peptides they present remains mostly unclear. The cell line RMA-S has a defect in class I Ag presentation, presumably due to a mutation in a peptide transporter gene. This defect can be overcome by transfection of RMA-S cells with the Tap-2 gene (formerly Ham-2) that encodes an ATP-binding transporter protein. We now show that a substantial portion of alloreactive CTL specific for Qa-1 class Ib molecules recognize Qa-1b on RMA-S cells and thus differ from most class Ia specific CTL. Those anti-Qa-1b CTL that do not recognize untransfected RMA-S do lyse RMA-S transfected with Tap-2. We also examine the effects of Qdm, a gene that maps to the D region and alters recognition of Qa-1. Qdm(k) strains lack an epitope(s) recognized by some (Qdm dependent) anti-Qa-1 CTL whereas Qdm+ strains express this epitope. Thus, Qdm-dependent CTL do not recognize Qa-1 on Qdm(k) targets whereas Qdm-independent CTL recognize Qa-1 epitopes in all strains. Although Qdm-independent CTL varied as to whether they recognized RMA-S vs RMA, all nine Qdm-dependent clones only recognized Qa-1b on RMA and not RMA-S. This result is consistent with Qdm encoding a peptide dependent upon the TAP transporter for cell membrane expression.     

Prediction of proteasome cleavage motifs by neural networks

We present a predictive method that can simulate an essential step in the antigen presentation in higher vertebrates, namely the step involving the proteasomal degradation of polypeptides into fragments which have the potential to bind to MHC Class I molecules. Proteasomal cleavage prediction algorithms published so far were trained on data from in vitro digestion experiments with constitutive proteasomes. As a result, they did not take into account the characteristics of the structurally modified proteasomes--often called immunoproteasomes--found in cells stimulated by gamma-interferon under physiological conditions. Our algorithm has been trained not only on in vitro data, but also on MHC Class I ligand data, which reflect a combination of immunoproteasome and constitutive proteasome specificity. This feature, together with the use of neural networks, a non-linear classification technique, make the prediction of MHC Class I ligand boundaries more accurate: 65% of the cleavage sites and 85% of the non-cleavage sites are correctly determined. Moreover, we show that the neural networks trained on the constitutive proteasome data learns a specificity that differs from that of the networks trained on MHC Class I ligands, i.e. the specificity of the immunoproteasome is different than the constitutive proteasome. The tools developed in this study in combination with a predictor of MHC and TAP binding capacity should give a more complete prediction of the generation and presentation of peptides on MHC Class I molecules. Here we demonstrate that such an approach produces an accurate prediction of the CTL the epitopes in HIV Nef. The method is available at www.cbs.dtu.dk/services/NetChop/.     

Sequential cleavage by metallopeptidases and proteasomes is involved in processing HIV-1 ENV epitope for endogenous MHC class I antigen presentation

Antigenic peptides derived from viral proteins by multiple proteolytic cleavages are bound by MHC class I molecules and recognized by CTL. Processing predominantly takes place in the cytosol of infected cells by the action of proteasomes. To identify other proteases involved in the endogenous generation of viral epitopes, specifically those derived from proteins routed to the secretory pathway, we investigated presentation of the HIV-1 ENV 10-mer epitope 318RGPGRAFVTI327 (p18) to specific CTL in the presence of diverse protease inhibitors. Both metalloproteinase and proteasome inhibitors decreased CTL recognition of the p18 epitope expressed from either native gp160 or from a chimera based on the hepatitis B virus secretory core protein as carrier protein. Processing of this epitope from both native ENV and the hepatitis B virus secretory core chimeric protein appeared to proceed by a TAP-dependent pathway that involved sequential cleavage by proteasomes and metallo-endopeptidases; however, other protease activities could replace the function of the lactacystin-sensitive proteasomes. By contrast, in a second TAP-independent pathway we detected no contribution of metallopeptidases for processing the ENV epitope from the chimeric protein. These results show that, in the classical TAP-dependent MHC class I pathway, endogenous Ag processing of viral proteins to yield the p18 10-mer epitope requires metallo-endopeptidases in addition to proteasomes.     

Phosphorylated peptides can be transported by TAP molecules, presented by class I MHC molecules, and recognized by phosphopeptide-specific CTL.

CTL recognize short peptide fragments presented by class I MHC molecules. In this study, we examined the effect of phosphorylation on TAP transport, binding to class I MHC molecules, and recognition by CTL of peptide fragments from known phosphorylated oncogene proteins or virus phosphoproteins. We show that phosphopeptides can be efficiently transported from the cytosol to the endoplasmic reticulum by the TAP. Furthermore, we show that phosphorylation can have a neutral, negative, or even a positive effect on peptide binding to class I MHC. Finally, we have generated phosphopeptide-specific CTL that discriminate between the phosphorylated and the nonphosphorylated versions of the peptide. We conclude that phosphopeptide-specific CTL responses are likely to constitute a subset of the class I MHC-restricted CTL repertoire in vivo.