Immunodominance in major histocompatibility complex class I-restricted T lymphocyte responses

Establishment of a novel tumor neoantigen prediction tool for personalized vaccine design

The personalized neoantigen nanovaccine (PNVAC) platform for patients with gastric cancer we established previously exhibited promising anti-tumor immunoreaction. However, limited by the ability of traditional neoantigen prediction tools, a portion of epitopes failed to induce specific immune response. In order to filter out more neoantigens to optimize our PNVAC platform, we develop a novel neoantigen prediction model, NUCC. This prediction tool trained through a deep learning approach exhibits better neoantigen prediction performance than other prediction tools, not only in two independent epitope datasets, but also in a totally new epitope dataset we construct from scratch, including 25 patients with advance gastric cancer and 150 candidate mutant peptides, 13 of which prove to be neoantigen by immunogenicity test in vitro. Our work lay the foundation for the improvement of our PNVAC platform for gastric cancer in the future.

A therapeutic HBV vaccine containing a checkpoint modifier enhances CD8+ T cell and antiviral responses

In patients who progress from acute hepatitis B virus (HBV) infection to a chronic HBV (CHB) infection, CD8+ T cells fail to eliminate the virus and become impaired. A functional cure of CHB likely requires CD8+ T cell responses different from those induced by the infection. Here we report preclinical immunogenicity and efficacy of an HBV therapeutic vaccine that includes herpes simplex virus (HSV) glycoprotein D (gD), a checkpoint modifier of early T cell activation, that augments CD8+ T cell responses. The vaccine is based on a chimpanzee adenovirus serotype 6 (AdC6) vector, called AdC6-gDHBV2, which targets conserved and highly immunogenic regions of the viral polymerase and core antigens fused to HSV gD. The vaccine was tested with and without gD in mice for immunogenicity, and in an AAV8-1.3HBV vector model of antiviral efficacy. The vaccine encoding the HBV antigens within gD stimulates potent and broad CD8+ T cell responses. In a surrogate model of HBV infection, a single intramuscular injection achieved pronounced and sustained declines of circulating HBV DNA copies and HBV surface antigen; both inversely correlated with HBV-specific CD8+ T cell frequencies in spleen and liver.

Probiotic neoantigen delivery vectors for precision cancer immunotherapy

Microbial systems have been synthetically engineered to deploy therapeutic payloads in vivo1,2. With emerging evidence that bacteria naturally home in on tumours3,4 and modulate antitumour immunity5,6, one promising application is the development of bacterial vectors as precision cancer vaccines2,7. Here we engineered probiotic Escherichia coli Nissle 1917 as an antitumour vaccination platform optimized for enhanced production and cytosolic delivery of neoepitope-containing peptide arrays, with increased susceptibility to blood clearance and phagocytosis. These features enhance both safety and immunogenicity, achieving a system that drives potent and specific T cell-mediated anticancer immunity that effectively controls or eliminates tumour growth and extends survival in advanced murine primary and metastatic solid tumours. We demonstrate that the elicited antitumour immune response involves recruitment and activation of dendritic cells, extensive priming and activation of neoantigen-specific CD4+ and CD8+ T cells, broader activation of both T and natural killer cells, and a reduction of tumour-infiltrating immunosuppressive myeloid and regulatory T and B cell populations. Taken together, this work leverages the advantages of living medicines to deliver arrays of tumour-specific neoantigen-derived epitopes within the optimal context to induce specific, effective and durable systemic antitumour immunity.

Advancing one health vaccination: In silico design and evaluation of a multi-epitope subunit vaccine against Nipah virus for cross-species immunization using immunoinformatics and molecular modeling

The resurgence of the Nipah virus (NiV) in 2023 has raised concerns for another potentially severe pandemic, given its history of high mortality from previous outbreaks. Unfortunately, no therapeutics and vaccines have been available for the virus. This study used immunoinformatics and molecular modeling to design and evaluate a multi-epitope subunit vaccine targeting NiV. The designed vaccine construct aims to stimulate immune responses in humans and two other intermediate animal hosts of the virus-swine and equine. Using several epitope prediction tools, ten peptides that induced B-lymphocyte responses, 17 peptides that induced cytotoxic T-lymphocyte (CTL) responses, and 12 peptides that induced helper T-lymphocyte (HTL) responses were mapped from nine NiV protein sequences. However, the CTL and HTL-inducing peptides were reduced to ten and eight, respectively, following molecular docking and dynamics. These screened peptides exhibited stability with 30 common major histocompatibility complex (MHC) receptors found in humans, swine, and equine. All peptides were linked using peptide linkers to form the multi-epitope construct and various adjuvants were tested to enhance its immunogenicity. The vaccine construct with resuscitation-promoting factor E (RpfE) adjuvant was selected as the final design based on its favorable physicochemical properties and superior immune response profile. Molecular docking was used to visualize the interaction of the vaccine to toll-like receptor 4 (TLR4), while molecular dynamics confirmed the structural stability of this interaction. Physicochemical property evaluation and computational simulations showed that the designed vaccine construct exhibited favorable properties and elicited higher antibody titers than the six multi-epitope NiV vaccine designs available in the literature. Further in vivo and in vitro experiments are necessary to validate the immunogenicity conferred by the designed vaccine construct and its epitope components. This study demonstrates the capability of computational methodologies in rational vaccine design and highlights the potential of cross-species vaccination strategies for mitigating potential NiV threats.

IMGT/RobustpMHC: robust training for class-I MHC peptide binding prediction

The accurate prediction of peptide-major histocompatibility complex (MHC) class I binding probabilities is a critical endeavor in immunoinformatics, with broad implications for vaccine development and immunotherapies. While recent deep neural network based approaches have showcased promise in peptide-MHC (pMHC) prediction, they have two shortcomings: (i) they rely on hand-crafted pseudo-sequence extraction, (ii) they do not generalize well to different datasets, which limits the practicality of these approaches. While existing methods rely on a 34 amino acid pseudo-sequence, our findings uncover the involvement of 147 positions in direct interactions between MHC and peptide. We further show that neural architectures can learn the intricacies of pMHC binding using even full sequences. To this end, we present PerceiverpMHC that is able to learn accurate representations on full-sequences by leveraging efficient transformer based architectures. Additionally, we propose IMGT/RobustpMHC that harnesses the potential of unlabeled data in improving the robustness of pMHC binding predictions through a self-supervised learning strategy. We extensively evaluate RobustpMHC on eight different datasets and showcase an overall improvement of over 6% in binding prediction accuracy compared to state-of-the-art approaches. We compile CrystalIMGT, a crystallography-verified dataset presenting a challenge to existing approaches due to significantly different pMHC distributions. Finally, to mitigate this distribution gap, we further develop a transfer learning pipeline.

Energy landscapes of peptide-MHC binding

Molecules of the Major Histocompatibility Complex (MHC) present short protein fragments on the cell surface, an important step in T cell immune recognition. MHC-I molecules process peptides from intracellular proteins; MHC-II molecules act in antigen-presenting cells and present peptides derived from extracellular proteins. Here we show that the sequence-dependent energy landscapes of MHC-peptide binding encode class-specific nonlinearities (epistasis). MHC-I has a smooth landscape with global epistasis; the binding energy is a simple deformation of an underlying linear trait. This form of epistasis enhances the discrimination between strong-binding peptides. In contrast, MHC-II has a rugged landscape with idiosyncratic epistasis: binding depends on detailed amino acid combinations at multiple positions of the peptide sequence. The form of epistasis affects the learning of energy landscapes from training data. For MHC-I, a low-complexity problem, we derive a simple matrix model of binding energies that outperforms current models trained by machine learning. For MHC-II, higher complexity prevents learning by simple regression methods. Epistasis also affects the energy and fitness effects of mutations in antigen-derived peptides (epitopes). In MHC-I, large-effect mutations occur predominantly in anchor positions of strong-binding epitopes. In MHC-II, large effects depend on the background epitope sequence but are broadly distributed over the epitope, generating a bigger target for escape mutations due to loss of presentation. Together, our analysis shows how an energy landscape of protein-protein binding constrains the target of escape mutations from T cell immunity, linking the complexity of the molecular interactions to the dynamics of adaptive immune response.

APEX-pHLA: A novel method for accurate prediction of the binding between exogenous short peptides and HLA class I molecules

Human leukocyte antigen (HLA) molecules play critically significant role within the realm of immunotherapy due to their capacities to recognize and bind exogenous antigens such as peptides, subsequently delivering them to immune cells. Predicting the binding between peptides and HLA molecules (pHLA) can expedite the screening of immunogenic peptides and facilitate vaccine design. However, traditional experimental methods are time-consuming and inefficient. In this study, an efficient method based on deep learning was developed for predicting peptide-HLA binding, which treated peptide sequences as linguistic entities. It combined the architectures of textCNN and BiLSTM to create a deep neural network model called APEX-pHLA. This model operated without limitations related to HLA class I allele variants and peptide segment lengths, enabling efficient encoding of sequence features for both HLA and peptide segments. On the independent test set, the model achieved Accuracy, ROC_AUC, F1, and MCC is 0.9449, 0.9850, 0.9453, and 0.8899, respectively. Similarly, on an external test set, the results were 0.9803, 0.9574, 0.8835, and 0.7863, respectively. These findings outperformed fifteen methods previously reported in the literature. The accurate prediction capability of the APEX-pHLA model in peptide-HLA binding might provide valuable insights for future HLA vaccine design.

Characterisation of novel influenza-derived HLA-B*18:01-restricted epitopes

Objectives

Seasonal influenza viruses cause roughly 650 000 deaths annually despite available vaccines. CD8+ T cells typically recognise influenza-derived peptides from internal structural and non-structural influenza proteins and are an attractive avenue for future vaccine design as they could reduce the severity of disease following infection with diverse influenza strains. CD8+ T cells recognise peptides presented by the highly polymorphic Human Leukocyte Antigens class I molecules (HLA-I). Each HLA-I variant has distinct peptide binding preferences, representing a significant obstacle for designing vaccines that elicit CD8+ T cell responses across broad populations. Consequently, the rational design of a CD8+ T cell-mediated vaccine would require the identification of highly immunogenic peptides restricted to a range of different HLA molecules.

Methods

Here, we assessed the immunogenicity of six recently published novel influenza-derived peptides identified by mass-spectrometry and predicted to bind to the prevalent HLA-B*18:01 molecule.

Results

Using CD8+ T cell activation assays and protein biochemistry, we showed that 3/6 of the novel peptides were immunogenic in several HLA-B*18:01+ individuals and confirmed their HLA-B*18:01 restriction. We subsequently compared CD8+ T cell responses towards the previously identified highly immunogenic HLA-B*18:01-restricted NP219 peptide. Using X-ray crystallography, we solved the first crystal structures of HLA-B*18:01 presenting immunogenic influenza-derived peptides. Finally, we dissected the first TCR repertoires specific for HLA-B*18:01 restricted pathogen-derived peptides, identifying private and restricted repertoires against each of the four peptides.

Conclusion

Overall the characterisation of these novel immunogenic peptides provides additional HLA-B*18:01-restricted vaccine targets derived from the Matrix protein 1 and potentially the non-structural protein and the RNA polymerase catalytic subunit of influenza viruses.

Tumor-Derived Antigenic Peptides as Potential Cancer Vaccines

Peptide antigens derived from tumors have been observed to elicit protective immune responses, categorized as either tumor-associated antigens (TAAs) or tumor-specific antigens (TSAs). Subunit cancer vaccines incorporating these antigens have shown promise in inducing protective immune responses, leading to cancer prevention or eradication. Over recent years, peptide-based cancer vaccines have gained popularity as a treatment modality and are often combined with other forms of cancer therapy. Several clinical trials have explored the safety and efficacy of peptide-based cancer vaccines, with promising outcomes. Advancements in techniques such as whole-exome sequencing, next-generation sequencing, and in silico methods have facilitated the identification of antigens, making it increasingly feasible. Furthermore, the development of novel delivery methods and a deeper understanding of tumor immune evasion mechanisms have heightened the interest in these vaccines among researchers. This article provides an overview of novel insights regarding advancements in the field of peptide-based vaccines as a promising therapeutic avenue for cancer treatment. It summarizes existing computational methods for tumor neoantigen prediction, ongoing clinical trials involving peptide-based cancer vaccines, and recent studies on human vaccination experiments.

Informing immunotherapy with multi-omics driven machine learning

Progress in sequencing technologies and clinical experiments has revolutionized immunotherapy on solid and hematologic malignancies. However, the benefits of immunotherapy are limited to specific patient subsets, posing challenges for broader application. To improve its effectiveness, identifying biomarkers that can predict patient response is crucial. Machine learning (ML) play a pivotal role in harnessing multi-omic cancer datasets and unlocking new insights into immunotherapy. This review provides an overview of cutting-edge ML models applied in omics data for immunotherapy analysis, including immunotherapy response prediction and immunotherapy-relevant tumor microenvironment identification. We elucidate how ML leverages diverse data types to identify significant biomarkers, enhance our understanding of immunotherapy mechanisms, and optimize decision-making process. Additionally, we discuss current limitations and challenges of ML in this rapidly evolving field. Finally, we outline future directions aimed at overcoming these barriers and improving the efficiency of ML in immunotherapy research.

Polytopic fractional delivery of an HIV vaccine alters cellular responses and results in increased epitope breadth in a phase 1 randomized trial

BACKGROUND

Elicitation of broad immune responses is understood to be required for an efficacious preventative HIV vaccine. This Phase 1 randomized controlled trial evaluated whether administration of vaccine antigens separated at multiple injection sites vs combined, fractional delivery at multiple sites affected T-cell breadth compared to standard, single site vaccination.

METHODS

We randomized 90 participants to receive recombinant adenovirus 5 (rAd5) vector with HIV inserts gag, pol and env via three different strategies. The Standard group received vaccine at a single anatomic site (n = 30) compared to two polytopic (multisite) vaccination groups: Separated (n = 30), where antigens were separately administered to four anatomical sites, and Fractioned (n = 30), where fractions of each vaccine component were combined and administered at four sites. All groups received the same total dose of vaccine.

FINDINGS

CD8 T-cell response rates and magnitudes were significantly higher in the Fractioned group than Standard for several antigen pools tested. CD4 T-cell response magnitudes to Pol were higher in the Separated than Standard group. T-cell epitope mapping demonstrated greatest breadth in the Fractioned group (median 8.0 vs 2.5 for Standard, Wilcoxon p = 0.03; not significant after multiplicity adjustment for co-primary endpoints). IgG binding antibody response rates to Env were higher in the Standard and Fractioned groups vs Separated group.

INTERPRETATION

This study shows that the number of anatomic sites for which a vaccine is delivered and distribution of its antigenic components influences immune responses in humans.

FUNDING

National Institute of Allergy and Infectious Diseases, NIH.

Virus-like particle-mediated delivery of structure-selected neoantigens demonstrates immunogenicity and antitumoral activity in mice

BACKGROUND

Neoantigens are patient- and tumor-specific peptides that arise from somatic mutations. They stand as promising targets for personalized therapeutic cancer vaccines. The identification process for neoantigens has evolved with the use of next-generation sequencing technologies and bioinformatic tools in tumor genomics. However, in-silico strategies for selecting immunogenic neoantigens still have very low accuracy rates, since they mainly focus on predicting peptide binding to Major Histocompatibility Complex (MHC) molecules, which is key but not the sole determinant for immunogenicity. Moreover, the therapeutic potential of neoantigen-based vaccines may be enhanced using an optimal delivery platform that elicits robust de novo immune responses.

METHODS

We developed a novel neoantigen selection pipeline based on existing software combined with a novel prediction method, the Neoantigen Optimization Algorithm (NOAH), which takes into account structural features of the peptide/MHC-I interaction, as well as the interaction between the peptide/MHC-I complex and the TCR, in its prediction strategy. Moreover, to maximize neoantigens' therapeutic potential, neoantigen-based vaccines should be manufactured in an optimal delivery platform that elicits robust de novo immune responses and bypasses central and peripheral tolerance.

RESULTS

We generated a highly immunogenic vaccine platform based on engineered HIV-1 Gag-based Virus-Like Particles (VLPs) expressing a high copy number of each in silico selected neoantigen. We tested different neoantigen-loaded VLPs (neoVLPs) in a B16-F10 melanoma mouse model to evaluate their capability to generate new immunogenic specificities. NeoVLPs were used in in vivo immunogenicity and tumor challenge experiments.

CONCLUSIONS

Our results indicate the relevance of incorporating other immunogenic determinants beyond the binding of neoantigens to MHC-I. Thus, neoVLPs loaded with neoantigens enhancing the interaction with the TCR can promote the generation of de novo antitumor-specific immune responses, resulting in a delay in tumor growth. Vaccination with the neoVLP platform is a robust alternative to current therapeutic vaccine approaches and a promising candidate for future personalized immunotherapy.

Antibody-epitope conjugates deliver immunogenic T-cell epitopes more efficiently when close to cell surfaces

Antibody-mediated delivery of immunogenic viral CD8+ T-cell epitopes to redirect virus-specific T cells toward cancer cells is a promising new therapeutic avenue to increase the immunogenicity of tumors. Multiple strategies for viral epitope delivery have been shown to be effective. So far, most of these have relied on a free C-terminus of the immunogenic epitope for extracellular delivery. Here, we demonstrate that antibody-epitope conjugates (AECs) with genetically fused epitopes to the N-terminus of the antibody can also sensitize tumors for attack by virus-specific CD8+ T cells. AECs carrying epitopes genetically fused at the N-terminus of the light chains of cetuximab and trastuzumab demonstrate an even more efficient delivery of the T-cell epitopes compared to AECs with the epitope fused to the C-terminus of the heavy chain. We demonstrate that this increased efficiency is not caused by the shift in location of the cleavage site from the N- to the C-terminus, but by its increased proximity to the cell surface. We hypothesize that this facilitates more efficient epitope delivery. These findings not only provide additional insights into the mechanism of action of AECs but also broaden the possibilities for genetically fused AECs as an avenue for the redirection of multiple virus-specific T cells toward tumors.

Antibody-mediated delivery of viral epitopes to redirect EBV-specific CD8+ T-cell immunity towards cancer cells

Antibody-mediated delivery of immunogenic epitopes to redirect virus-specific CD8+ T-cells towards cancer cells is an emerging and promising new therapeutic strategy. These so-called antibody-epitope conjugates (AECs) rely on the proteolytic release of the epitopes close to the tumor surface for presentation by HLA class I molecules to eventually redirect and activate virus-specific CD8+ T-cells towards tumor cells. We fused the immunogenic EBV-BRLF1 epitope preceded by a protease cleavage site to the C-terminus of the heavy and/or light chains of cetuximab and trastuzumab. We evaluated these AECs and found that, even though all AECs were able to redirect the EBV-specific T-cells, AECs with an epitope fused to the C-terminus of the heavy chain resulted in higher levels of T-cell activation compared to AECs with the same epitope fused to the light chain of an antibody. We observed that all AECs were depending on the presence of the antibody target, that the level of T-cell activation correlated with expression levels of the antibody target, and that our AECs could efficiently deliver the BRLF1 epitope to cancer cell lines from different origins (breast, ovarian, lung, and cervical cancer and a multiple myeloma). Moreover, in vivo, the AECs efficiently reduced tumor burden and increased the overall survival, which was prolonged even further in combination with immune checkpoint blockade. We demonstrate the potential of these genetically fused AECs to redirect the potent EBV-specific T-cells towards cancer in vitro and in vivo.

ACE configurator for ELISpot: optimizing combinatorial design of pooled ELISpot assays with an epitope similarity model

The enzyme-linked immunosorbent spot (ELISpot) assay is a powerful in vitro immunoassay that enables cost-effective quantification of antigen-specific T-cell reactivity. It is used widely in the context of cancer and infectious diseases to validate the immunogenicity of predicted epitopes. While technological advances have kept pace with the demand for increased throughput, efforts to increase scale are bottlenecked by current assay design and deconvolution methods, which have remained largely unchanged. Current methods for designing pooled ELISpot experiments offer limited flexibility of assay parameters, lack support for high-throughput scenarios and do not consider peptide identity during pool assignment. We introduce the ACE Configurator for ELISpot (ACE) to address these gaps. ACE generates optimized peptide-pool assignments from highly customizable user inputs and handles the deconvolution of positive peptides using assay readouts. In this study, we present a novel sequence-aware pooling strategy, powered by a fine-tuned ESM-2 model that groups immunologically similar peptides, reducing the number of false positives and subsequent confirmatory assays compared to existing combinatorial approaches. To validate ACE's performance on real-world datasets, we conducted a comprehensive benchmark study, contextualizing design choices with their impact on prediction quality. Our results demonstrate ACE's capacity to further increase precision of identified immunogenic peptides, directly optimizing experimental efficiency. ACE is freely available as an executable with a graphical user interface and command-line interfaces at https://github.com/pirl-unc/ace.

Correlative CD4 and CD8 T-cell immunodominance in humans and mice: Implications for preclinical testing

Antigen-specific T-cell recognition is restricted by Major Histocompatibility Complex (MHC) molecules, and differences between CD4 and CD8 immunogenicity in humans and animal species used in preclinical vaccine testing are yet to be fully understood. In this study, we addressed this matter by analyzing experimentally identified epitopes based on published data curated in the Immune Epitopes DataBase (IEDB) database. We first analyzed SARS-CoV-2 spike (S) and nucleoprotein (N), which are two common targets of the immune response and well studied in both human and mouse systems. We observed a weak but statistically significant correlation between human and H-2b mouse T-cell responses (CD8 S specific (r = 0.206, p = 1.37 × 10-13); CD4 S specific (r = 0.118, p = 2.63 × 10-5) and N specific (r = 0.179, p = 2.55 × 10-4)). Due to intrinsic differences in MHC molecules across species, we also investigated the association between the immunodominance of common Human Leukocyte Antigen (HLA) alleles for which HLA transgenic mice are available, namely, A*02:01, B*07:02, DRB1*01:01, and DRB1*04:01, and found higher significant correlations for both CD8 and CD4 (maximum r = 0.702, p = 1.36 × 10-31 and r = 0.594, p = 3.04-122, respectively). Our results further indicated that some regions are commonly immunogenic between humans and mice (either H-2b or HLA transgenic) but that others are human specific. Finally, we noted a significant correlation between CD8 and CD4 S- (r = 0.258, p = 7.33 × 1021) and N-specific (r = 0.369, p = 2.43 × 1014) responses, suggesting that discrete protein subregions can be simultaneously recognized by T cells. These findings were confirmed in other viral systems, providing general guidance for the use of murine models to test T-cell immunogenicity of viral antigens destined for human use.

A cell-free antigen processing system informs HIV-1 epitope selection and vaccine design

Distinct CD4+ T cell epitopes have been associated with spontaneous control of HIV-1 replication, but analysis of antigen-dependent factors that influence epitope selection is lacking. To examine these factors, we used a cell-free antigen processing system that incorporates soluble HLA-DR (DR1), HLA-DM (DM), cathepsins, and full-length protein antigens for epitope identification by LC-MS/MS. HIV-1 Gag, Pol, Env, Vif, Tat, Rev, and Nef were examined using this system. We identified 35 novel epitopes, including glycopeptides. Epitopes from smaller HIV-1 proteins mapped to regions of low protein stability and higher solvent accessibility. HIV-1 antigens associated with limited CD4+ T cell responses were processed efficiently, while some protective epitopes were inefficiently processed. 55% of epitopes obtained from cell-free processing induced memory CD4+ T cell responses in HIV-1+ donors, including eight of 19 novel epitopes tested. Thus, an in vitro processing system utilizing the components of Class II processing reveals factors influencing epitope selection of HIV-1 and represents an approach to understanding epitope selection from non-HIV-1 antigens.

Comparison of methods generating antibody-epitope conjugates for targeting cancer with virus-specific T cells

Therapeutic antibody-epitope conjugates (AECs) are promising new modalities to deliver immunogenic epitopes and redirect virus-specific T-cell activity to cancer cells. Nevertheless, many aspects of these antibody conjugates require optimization to increase their efficacy. Here we evaluated different strategies to conjugate an EBV epitope (YVL/A2) preceded by a protease cleavage site to the antibodies cetuximab and trastuzumab. Three approaches were taken: chemical conjugation (i.e. a thiol-maleimide reaction) to reduced cysteine side chains, heavy chain C-terminal enzymatic conjugation using sortase A, and genetic fusions, to the heavy chain (HC) C-terminus. All three conjugates were capable of T-cell activation and target-cell killing via proteolytic release of the EBV epitope and expression of the antibody target was a requirement for T-cell activation. Moreover, AECs generated with a second immunogenic epitope derived from CMV (NLV/A2) were able to deliver and redirect CMV specific T-cells, in which the amino sequence of the attached peptide appeared to influence the efficiency of epitope delivery. Therefore, screening of multiple protease cleavage sites and epitopes attached to the antibody is necessary. Taken together, our data demonstrated that multiple AECs could sensitize cancer cells to virus-specific T cells.

Survivin (BIRC5) Peptide Vaccine in the 4T1 Murine Mammary Tumor Model: A Potential Neoadjuvant T Cell Immunotherapy for Triple Negative Breast Cancer: A Preliminary Study

A triple negative breast cancer model using the murine 4T1 tumor cell line was used to explore the efficacy of an adjuvanted survivin peptide microparticle vaccine using tumor growth as the outcome metric. We first performed tumor cell dose titration studies to determine a tumor cell dose that resulted in sufficient tumor takes but allowed multiple serial measurements of tumor volumes, yet with minimal morbidity/mortality within the study period. Later, in a second cohort of mice, the survivin peptide microparticle vaccine was administered via intraperitoneal injection at the study start with a second dose given 14 days later. An orthotopic injection of 4T1 cells into the mammary tissue was performed on the same day as the administration of the second vaccine dose. The mice were followed for up to 41 days with subcutaneous measurements of tumor volume made every 3–4 days. Vaccination with survivin peptides was associated with a peptide antigen-specific gamma interferon enzyme-linked immunosorbent spot response in the murine splenocyte population but was absent from the control microparticle group. At the end of the study, we found that vaccination with adjuvanted survivin peptide microparticles resulted in statistically significant slower primary tumor growth rates in BALB/c mice challenged with 4T1 cells relative to the control peptideless vaccination group. These studies suggest that T cell immunotherapy specifically targeting survivin might be an applicable neoadjuvant immunotherapy therapy for triple negative breast cancer. More preclinical studies and clinical trials are needed to explore this concept further.

SARS-CoV-2 epitope-specific T cells: Immunity response feature, TCR repertoire characteristics and cross-reactivity

The devastating COVID-19 pandemic caused by SARS-CoV-2 and multiple variants or subvariants remains an ongoing global challenge. SARS-CoV-2-specific T cell responses play a critical role in early virus clearance, disease severity control, limiting the viral transmission and underpinning COVID-19 vaccine efficacy. Studies estimated broad and robust T cell responses in each individual recognized at least 30 to 40 SARS-CoV-2 antigen epitopes and associated with COVID-19 clinical outcome. Several key immunodominant viral proteome epitopes, including S protein- and non-S protein-derived epitopes, may primarily induce potent and long-lasting antiviral protective effects. In this review, we summarized the immune response features of immunodominant epitope-specific T cells targeting different SRAS-CoV-2 proteome structures after infection and vaccination, including abundance, magnitude, frequency, phenotypic features and response kinetics. Further, we analyzed the epitopes immunodominance hierarchy in combination with multiple epitope-specific T cell attributes and TCR repertoires characteristics, and discussed the significant implications of cross-reactive T cells toward HCoVs, SRAS-CoV-2 and variants of concern, especially Omicron. This review may be essential for mapping the landscape of T cell responses toward SARS-CoV-2 and optimizing the current vaccine strategy.

Decoupling peptide binding from T cell receptor recognition with engineered chimeric MHC-I molecules

Major Histocompatibility Complex class I (MHC-I) molecules display self, viral or aberrant epitopic peptides to T cell receptors (TCRs), which employ interactions between complementarity-determining regions with both peptide and MHC-I heavy chain 'framework' residues to recognize specific Human Leucocyte Antigens (HLAs). The highly polymorphic nature of the HLA peptide-binding groove suggests a malleability of interactions within a common structural scaffold. Here, using structural data from peptide:MHC-I and pMHC:TCR structures, we first identify residues important for peptide and/or TCR binding. We then outline a fixed-backbone computational design approach for engineering synthetic molecules that combine peptide binding and TCR recognition surfaces from existing HLA allotypes. X-ray crystallography demonstrates that chimeric molecules bridging divergent HLA alleles can bind selected peptide antigens in a specified backbone conformation. Finally, in vitro tetramer staining and biophysical binding experiments using chimeric pMHC-I molecules presenting established antigens further demonstrate the requirement of TCR recognition on interactions with HLA framework residues, as opposed to interactions with peptide-centric Chimeric Antigen Receptors (CARs). Our results underscore a novel, structure-guided platform for developing synthetic HLA molecules with desired properties as screening probes for peptide-centric interactions with TCRs and other therapeutic modalities.

Subcellular location of source proteins improves prediction of neoantigens for immunotherapy

Antigen presentation via the major histocompatibility complex (MHC) is essential for anti-tumor immunity. However, the rules that determine which tumor-derived peptides will be immunogenic are still incompletely understood. Here, we investigated whether constraints on peptide accessibility to the MHC due to protein subcellular location are associated with peptide immunogenicity potential. Analyzing over 380,000 peptides from studies of MHC presentation and peptide immunogenicity, we find clear spatial biases in both eluted and immunogenic peptides. We find that including parent protein location improves the prediction of peptide immunogenicity in multiple datasets. In human immunotherapy cohorts, the location was associated with a neoantigen vaccination response, and immune checkpoint blockade responders generally had a higher burden of neopeptides from accessible locations. We conclude that protein subcellular location adds important information for optimizing cancer immunotherapies.

Understanding the Role of HLA Class I Molecules in the Immune Response to Influenza Infection and Rational Design of a Peptide-Based Vaccine

Influenza A virus is a respiratory pathogen that is responsible for regular epidemics and occasional pandemics that result in substantial damage to life and the economy. The yearly reformulation of trivalent or quadrivalent flu vaccines encompassing surface glycoproteins derived from the current circulating strains of the virus does not provide sufficient cross-protection against mismatched strains. Unlike the current vaccines that elicit a predominant humoral response, vaccines that induce CD8+ T cells have demonstrated a capacity to provide cross-protection against different influenza strains, including novel influenza viruses. Immunopeptidomics, the mass spectrometric identification of human-leukocyte-antigen (HLA)-bound peptides isolated from infected cells, has recently provided key insights into viral peptides that can serve as potential T cell epitopes. The critical elements required for a strong and long-living CD8+ T cell response are related to both HLA restriction and the immunogenicity of the viral peptide. This review examines the importance of HLA and the viral immunopeptidome for the design of a universal influenza T-cell-based vaccine.

Durable CD8 T Cell Memory against SARS-CoV-2 by Prime/Boost and Multi-Dose Vaccination: Considerations on Inter-Dose Time Intervals

Facing the COVID-19 pandemic, anti-SARS-CoV-2 vaccines were developed at unprecedented pace, productively exploiting contemporary fundamental research and prior art. Large-scale use of anti-SARS-CoV-2 vaccines has greatly limited severe morbidity and mortality. Protection has been correlated with high serum titres of neutralizing antibodies capable of blocking the interaction between the viral surface protein spike and the host SARS-CoV-2 receptor, ACE-2. Yet, vaccine-induced protection subsides over time, and breakthrough infections are commonly observed, mostly reflecting the decay of neutralizing antibodies and the emergence of variant viruses with mutant spike proteins. Memory CD8 T cells are a potent weapon against viruses, as they are against tumour cells. Anti-SARS-CoV-2 memory CD8 T cells are induced by either natural infection or vaccination and can be potentially exploited against spike-mutated viruses. We offer here an overview of current research about the induction of anti-SARS-CoV-2 memory CD8 T cells by vaccination, in the context of prior knowledge on vaccines and on fundamental mechanisms of immunological memory. We focus particularly on how vaccination by two doses (prime/boost) or more (boosters) promotes differentiation of memory CD8 T cells, and on how the time-length of inter-dose intervals may influence the magnitude and persistence of CD8 T cell memory.

Immunopeptidomic Analysis of BoLA-I and BoLA-DR Presented Peptides from Theileria parva Infected Cells

The apicomplexan parasite Theileria parva is the causative agent of East Coast fever, usually a fatal disease for cattle, which is prevalent in large areas of eastern, central, and southern Africa. Protective immunity against T. parva is mediated by CD8+ T cells, with CD4+ T-cells thought to be important in facilitating the full maturation and development of the CD8+ T-cell response. T. parva has a large proteome, with >4000 protein-coding genes, making T-cell antigen identification using conventional screening approaches laborious and expensive. To date, only a limited number of T-cell antigens have been described. Novel approaches for identifying candidate antigens for T. parva are required to replace and/or complement those currently employed. In this study, we report on the use of immunopeptidomics to study the repertoire of T. parva peptides presented by both BoLA-I and BoLA-DR molecules on infected cells. The study reports on peptides identified from the analysis of 13 BoLA-I and 6 BoLA-DR datasets covering a range of different BoLA genotypes. This represents the most comprehensive immunopeptidomic dataset available for any eukaryotic pathogen to date. Examination of the immunopeptidome data suggested the presence of a large number of coprecipitated and non-MHC-binding peptides. As part of the work, a pipeline to curate the datasets to remove these peptides was developed and used to generate a final list of 74 BoLA-I and 15 BoLA-DR-presented peptides. Together, the data demonstrated the utility of immunopeptidomics as a method to identify novel T-cell antigens for T. parva and the importance of careful curation and the application of high-quality immunoinformatics to parse the data generated.

Multiple instance neural networks based on sparse attention for cancer detection using T-cell receptor sequences

Early detection of cancers has been much explored due to its paramount importance in biomedical fields. Among different types of data used to answer this biological question, studies based on T cell receptors (TCRs) are under recent spotlight due to the growing appreciation of the roles of the host immunity system in tumor biology. However, the one-to-many correspondence between a patient and multiple TCR sequences hinders researchers from simply adopting classical statistical/machine learning methods. There were recent attempts to model this type of data in the context of multiple instance learning (MIL). Despite the novel application of MIL to cancer detection using TCR sequences and the demonstrated adequate performance in several tumor types, there is still room for improvement, especially for certain cancer types. Furthermore, explainable neural network models are not fully investigated for this application. In this article, we propose multiple instance neural networks based on sparse attention (MINN-SA) to enhance the performance in cancer detection and explainability. The sparse attention structure drops out uninformative instances in each bag, achieving both interpretability and better predictive performance in combination with the skip connection. Our experiments show that MINN-SA yields the highest area under the ROC curve scores on average measured across 10 different types of cancers, compared to existing MIL approaches. Moreover, we observe from the estimated attentions that MINN-SA can identify the TCRs that are specific for tumor antigens in the same T cell repertoire.

Amino acids at position 5 in the peptide/MHC binding region of a public virus-specific TCR are completely inter-changeable without loss of function

Anti-viral T-cell responses are usually directed against a limited set of antigens, but often contain many T cells expressing different T-cell receptors (TCRs). Identical TCRs found within virus-specific T-cell populations in different individuals are known as public TCRs, but also TCRs highly-similar to these public TCRs, with only minor variations in amino acids on specific positions in the Complementary Determining Regions (CDRs), are frequently found. However, the degree of freedom at these positions was not clear. In this study, we used the HLA-A*02:01-restricted EBV-LMP2^FLY -specific public TCR as model and modified the highly-variable position 5 of the CDR3β sequence with all 20 amino acids. Our results demonstrate that amino acids at this particular position in the CDR3β region of this TCR are completely inter-changeable, without loss of TCR function. We show that the inability to find certain variants in individuals is explained by their lower recombination probability rather than by steric hindrance.

The role of antigen expression in shaping the repertoire of HLA presented ligands

Human leukocyte antigen (HLA) presentation of peptides is a prerequisite of T cell immune activation. The understanding of the rules defining this event has large implications for our knowledge of basic immunology and for the rational design of immuno-therapeutics and vaccines. Historically, most of the available prediction methods have been solely focused on the information related to antigen processing and presentation. Recent work has, however, demonstrated that method performance can be boosted by integrating information related to antigen abundance. Here we expand on these later findings and develop an extended version of NetMHCpan, called NetMHCpanExp, integrating information on antigen abundance from RNA-Seq experiments. In line with earlier works, the model demonstrates improved performance for both HLA ligand and cancer neoantigen epitope prediction. Optimal results are obtained by use of sample-specific abundance information but also reference datasets can be applied with a limited performance drop. The developed tool is available at https://services.healthtech.dtu.dk/service.php?NetMHCpanExp-1.0.

Modeling ligand-macromolecular interactions as eigenvalue-based transition-state dissociation constants may offer insights into biochemical function of the resulting complexes

A ligand when bound to a macromolecule (protein, DNA, RNA) will influence the biochemical function of that macromolecule. This observation is empirical and attributable to the association of the ligand with the amino acids/nucleotides that comprise the macromolecule. The binding affinity is a measure of the strength-of-association of a macromolecule for its ligand and is numerically characterized by the association/dissociation constant. However, despite being widely used, a mathematically rigorous explanation by which the association/dissociation constant can influence the biochemistry and molecular biology of the resulting complex is not available. Here, the ligand-macromolecular complex is modeled as a homo- or hetero-dimer with a finite and equal number of atoms/residues per monomer. The pairwise interactions are numeric, empirically motivated and are randomly chosen from a standard uniform distribution. The transition-state dissociation constants are the strictly positive real part of all complex eigenvalues of this interaction matrix, belong to the open interval (0,1), and form a sequence whose terms are finite, monotonic, non-increasing and convergent. The theoretical results are rigorous, presented as theorems, lemmas and corollaries and are complemented by numerical studies. An inferential analysis of the clinical outcomes of amino acid substitutions of selected enzyme homodimers is also presented. These findings are extendible to higher-order complexes such as those likely to occur in vivo. The study also presents a schema by which a ligand can be annotated and partitioned into high- and low-affinity variants. The influence of the transition-state dissociation constants on the biochemistry and molecular biology of non-haem iron (Ⅱ)- and 2-oxoglutarate-dependent dioxygenases (catalysis) and major histocompatibility complex (Ⅰ) mediated export of high-affinity peptides (non-enzymatic association/dissociation) are examined as special cases.

SILVI, an open-source pipeline for T-cell epitope selection

High-throughput screening of available genomic data and identification of potential antigenic candidates have promoted the development of epitope-based vaccines and therapeutics. Several immunoinformatic tools are available to predict potential epitopes and other immunogenicity-related features, yet it is still challenging and time-consuming to compare and integrate results from different algorithms. We developed the R script SILVI (short for: from in silico to in vivo), to assist in the selection of the potentially most immunogenic T-cell epitopes from Human Leukocyte Antigen (HLA)-binding prediction data. SILVI merges and compares data from available HLA-binding prediction servers, and integrates additional relevant information of predicted epitopes, namely BLASTp alignments with host proteins and physical-chemical properties. The two default criteria applied by SILVI and additional filtering allow the fast selection of the most conserved, promiscuous, strong binding T-cell epitopes. Users may adapt the script at their discretion as it is written in open-source R language. To demonstrate the workflow and present selection options, SILVI was used to integrate HLA-binding prediction results of three example proteins, from viral, bacterial and parasitic microorganisms, containing validated epitopes included in the Immune Epitope Database (IEDB), plus the Human Papillomavirus (HPV) proteome. Applying different filters on predicted IC50, hydrophobicity and mismatches with host proteins allows to significantly reduce the epitope lists with favourable sensitivity and specificity to select immunogenic epitopes. We contemplate SILVI will assist T-cell epitope selections and can be continuously refined in a community-driven manner, helping the improvement and design of peptide-based vaccines or immunotherapies. SILVI development version is available at: github.com/JoanaPissarra/SILVI2020 and https://doi.org/10.5281/zenodo.6865909.

Comprehensive analysis of mycobacterium tuberculosis antigen-specific CD4+ T cell responses restricted by single HLA class II allotype in an individual

Mycobacterium tuberculosis infection is generally asymptomatic as latent tuberculosis, but it is still known as the world’s leading bacterial cause of death. The diagnosis of latent tuberculosis infection relies on the evidence of cellular immunity to mycobacterial antigens. Since the association between HLA class II and tuberculosis infection has been reported in several population groups, a detailed study on the CD4+ T cell response to major tuberculosis antigens is needed. To elucidate which HLA class II allotypes in an individual are preferentially used in tuberculosis, CD4+ T cells specific to TB10.4, Ag85b, ESAT-6, and CFP-10 of Mycobacterium tuberculosis antigens were analyzed comprehensively. A total of 33 healthy donors were analyzed by ex vivo and cultured ELISPOT using panels of artificial antigen-presenting cells expressing a single HLA class II allotype. The CD4+ T cell responses were increased by an average of 39-fold in cultured ELISPOT compared with ex vivo ELISPOT. In ex vivo and cultured ELISPOT, CD4+ T cell responses showed significantly higher by HLA-DR than those of HLA-DQ and HLA-DP locus. In cultured ELISPOT, 9 HLA-DR allotypes, 4 HLA-DQ allotypes, and 3 HLA-DP allotypes showed positive CD4+ T cell responses. Among ten donors with positive CD4+ T cell responses when tested for mixed Mycobacterium tuberculosis antigens, seven donors were positive for only a single allotype, and three were positive for two allotypes in an individual. However, only one allotype was used for a single antigen-specific response when a single tuberculosis antigen was used individually. These results on the distribution of HLA class II allotypes showing high CD4+ T-cell responses to Mycobacterium tuberculosis antigens and the intra-individual allotype dominance will provide valuable information for understanding the immunobiology and immunogenetics of tuberculosis, which can contribute to the development of more effective vaccines.

A comprehensive analysis of the IEDB MHC class-I automated benchmark

In 2014, the Immune Epitope Database automated benchmark was created to compare the performance of the MHC class I binding predictors. However, this is not a straightforward process due to the different and non-standardized outputs of the methods. Additionally, some methods are more restrictive regarding the HLA alleles and epitope sizes for which they predict binding affinities, while others are more comprehensive. To address how these problems impacted the ranking of the predictors, we developed an approach to assess the reliability of different metrics. We found that using percentile-ranked results improved the stability of the ranks and allowed the predictors to be reliably ranked despite not being evaluated on the same data. We also found that given the rate new data are incorporated into the benchmark, a new method must wait for at least 4 years to be ranked against the pre-existing methods. The best-performing tools with statistically indistinguishable scores in this benchmark were NetMHCcons, NetMHCpan4.0, ANN3.4, NetMHCpan3.0 and NetMHCpan2.8. The results of this study will be used to improve the evaluation and display of benchmark performance. We highly encourage anyone working on MHC binding predictions to participate in this benchmark to get an unbiased evaluation of their predictors.

Understanding the constitutive presentation of MHC class I immunopeptidomes in primary tissues

Understanding the molecular principles that govern the composition of the MHC-I immunopeptidome across different primary tissues is fundamentally important to predict how T cells respond in different contexts in vivo. Here, we performed a global analysis of the MHC-I immunopeptidome from 29 to 19 primary human and mouse tissues, respectively. First, we observed that different HLA-A, HLA-B, and HLA-C allotypes do not contribute evenly to the global composition of the MHC-I immunopeptidome across multiple human tissues. Second, we found that tissue-specific and housekeeping MHC-I peptides share very distinct properties. Third, we discovered that proteins that are evolutionarily hyperconserved represent the primary source of the MHC-I immunopeptidome at the organism-wide scale. Fourth, we uncovered new components of the antigen processing and presentation network, including the carboxypeptidases CPE, CNDP1/2, and CPVL. Together, this study opens up new avenues toward a system-wide understanding of antigen presentation in vivo across mammalian species.

•

Tissue-specific and housekeeping MHC class I peptides share distinct properties

•

HLA-A, HLA-B, and HLA-C allotypes contribute very unevenly to the pool of class I peptides

•

MHC-I immunopeptidomes are represented by evolutionarily conserved proteins

•

An extended antigen processing and presentation pathway is uncovered

Mechanistic diversity in MHC class I antigen recognition

Throughout its evolution, the human immune system has developed a plethora of strategies to diversify the antigenic peptide sequences that can be targeted by the CD8+ T cell response against pathogens and aberrations of self. Here we provide a general overview of the mechanisms that lead to the diversity of antigens presented by MHC class I complexes and their recognition by CD8+ T cells, together with a more detailed analysis of recent progress in two important areas that are highly controversial: the prevalence and immunological relevance of unconventional antigen peptides; and cross-recognition of antigenic peptides by the T cell receptors of CD8+ T cells.

T Cell Epitope Prediction and Its Application to Immunotherapy

T cells play a crucial role in controlling and driving the immune response with their ability to discriminate peptides derived from healthy as well as pathogenic proteins. In this review, we focus on the currently available computational tools for epitope prediction, with a particular focus on tools aimed at identifying neoepitopes, i.e. cancer-specific peptides and their potential for use in immunotherapy for cancer treatment. This review will cover how these tools work, what kind of data they use, as well as pros and cons in their respective applications.

Immunodominance complexity: lessons yet to be learned from dominant T cell responses to SARS-COV-2

Immunodominance is a complex and highly debated topic of T cell biology. The current SARS-CoV-2 pandemic has provided the opportunity to profile adaptive immune responses and determine molecular factors contributing to emerging responses towards immunodominant viral epitopes. Here, we discuss parameters that alter the dynamics of CD8 viral epitope processing, generation and T-cell responses, and how immunodominance counteracts viral immune escape mechanisms that develop in the context of emerging SARS-CoV-2 variants.

Borrelia Infection in Bank Voles Myodes glareolus Is Associated With Specific DQB Haplotypes Which Affect Allelic Divergence Within Individuals

The high polymorphism of Major Histocompatibility Complex (MHC) genes is generally considered to be a result of pathogen-mediated balancing selection. Such selection may operate in the form of heterozygote advantage, and/or through specific MHC allele–pathogen interactions. Specific MHC allele–pathogen interactions may promote polymorphism via negative frequency-dependent selection (NFDS), or selection that varies in time and/or space because of variability in the composition of the pathogen community (fluctuating selection; FS). In addition, divergent allele advantage (DAA) may act on top of these forms of balancing selection, explaining the high sequence divergence between MHC alleles. DAA has primarily been thought of as an extension of heterozygote advantage. However, DAA could also work in concert with NFDS though this is yet to be tested explicitly. To evaluate the importance of DAA in pathogen-mediated balancing selection, we surveyed allelic polymorphism of MHC class II DQB genes in wild bank voles (Myodes glareolus) and tested for associations between DQB haplotypes and infection by Borrelia afzelii, a tick-transmitted bacterium causing Lyme disease in humans. We found two significant associations between DQB haplotypes and infection status: one haplotype was associated with lower risk of infection (resistance), while another was associated with higher risk of infection (susceptibility). Interestingly, allelic divergence within individuals was higher for voles with the resistance haplotype compared to other voles. In contrast, allelic divergence was lower for voles with the susceptibility haplotype than other voles. The pattern of higher allelic divergence in individuals with the resistance haplotype is consistent with NFDS favouring divergent alleles in a natural population, hence selection where DAA works in concert with NFDS.

SARS-CoV-2 human T cell epitopes: Adaptive immune response against COVID-19

Over the past year, numerous studies in the peer reviewed and preprint literature have reported on the virological, epidemiological and clinical characteristics of the coronavirus, SARS-CoV-2. To date, 25 studies have investigated and identified SARS-CoV-2-derived T cell epitopes in humans. Here, we review these recent studies, how they were performed, and their findings. We review how epitopes identified throughout the SARS-CoV2 proteome reveal significant correlation between number of epitopes defined and size of the antigen provenance. We also report additional analysis of SARS-CoV-2 human CD4 and CD8 T cell epitope data compiled from these studies, identifying 1,400 different reported SARS-CoV-2 epitopes and revealing discrete immunodominant regions of the virus and epitopes that are more prevalently recognized. This remarkable breadth of epitope repertoire has implications for vaccine design, cross-reactivity, and immune escape by SARS-CoV-2 variants.

Mathematical modeling and stochastic simulations suggest that low-affinity peptides can bisect MHC1-mediated export of high-affinity peptides into "early"- and "late"-phases

The peptide loading complex (PLC) is a multi-protein complex of the endoplasmic reticulum (ER) which optimizes major histocompatibility I (MHC1)-mediated export of intracellular high-affinity peptides. Whilst, the molecular biology of MHC1-mediated export is well supported by empirical data, the stoichiometry, kinetics and spatio-temporal profile of the participating molecular entities are a matter of considerable debate. Here, a low-affinity peptide-driven (LAPD)-model of MHC1-mediated high-affinity peptide export is formulated, implemented, analyzed and simulated. The model is parameterized in terms of the contribution of the shunt reaction to the concentration of exportable MHC1. Theoretical analyses and simulation studies of the model suggest that low-affinity peptides can bisect MHC1-mediated export of high-affinity peptides into time-dependent distinct “early”- and “late”-phases. The net exportable MHC1 (eM1β(t)) is a function of the retrograde (rM1β(t))- and anterograde (aM1β(t))-derived fractions. The “early”-phase is dominated by the contribution of the retrograde/recyclable (rM1β≈61%,aM1β≈39%) pathway to exportable MHC1, is characterized by Tapasin-mediated peptide-editing and is ATP-independent. The “late”-phase on the other hand, is characterized by de novo PLC-assembly, rapid disassembly and a significant contribution of the anterograde pathway to exportable MHC1 (rM1β≈21%,aM1β≈79%). The shunt reaction is rate limiting and may integrate peptide translocation with PLC-assembly/disassembly thereby, regulating peptide export under physiological and pathological (viral infections, dysplastic alterations) conditions.

MHCVision: estimation of global and local false discovery rate for MHC class I peptide binding prediction

Motivation

MHC-peptide binding prediction has been widely used for understanding the immune response of individuals or populations, each carrying different MHC molecules as well as for the development of immunotherapeutics. The results from MHC-peptide binding prediction tools are mostly reported as a predicted binding affinity (IC₅₀) and the percentile rank score, and global thresholds e.g. IC₅₀ value < 500 nM or percentile rank < 2% are generally recommended for distinguishing binding peptides from non-binding peptides. However, it is difficult to evaluate statistically the probability of an individual peptide binding prediction to be true or false solely considering predicted scores. Therefore, statistics describing the overall global false discovery rate (FDR) and local FDR, also called posterior error probability (PEP) are required to give statistical context to the natively produced scores.

Result

We have developed an algorithm and code implementation, called MHCVision, for estimation of FDR and PEP values for the predicted results of MHC-peptide binding prediction from the NetMHCpan tool. MHCVision performs parameter estimation using a modified expectation maximization framework for a two-component beta mixture model, representing the distribution of true and false scores of the predicted dataset. We can then estimate the PEP of an individual peptide’s predicted score, and conversely the probability that it is true. We demonstrate that the use of global FDR and PEP estimation can provide a better trade-off between sensitivity and precision over using currently recommended thresholds from tools.

Availability and implementation

https://github.com/PGB-LIV/MHCVision.

Supplementary information

Supplementary data are available at Bioinformatics online.

Effect of epitope variant co-delivery on the depth of CD8 T cell responses induced by HIV-1 conserved mosaic vaccines

To stop the HIV-1 pandemic, vaccines must induce responses capable of controlling vast HIV-1 variants circulating in the population as well as those evolved in each individual following transmission. Numerous strategies have been proposed, of which the most promising include focusing responses on the vulnerable sites of HIV-1 displaying the least entropy among global isolates and using algorithms that maximize vaccine match to circulating HIV-1 variants by vaccine cocktails of optimized complementing sequences. In this study, we investigated CD8 T cell responses induced by a bi-valent mosaic of highly conserved HIVconsvX regions delivered by a combination of simian adenovirus ChAdOx1 and poxvirus MVA. We compared partially and fully mono- and bi-valent prime-boost regimens and their ability to elicit T cells recognizing natural epitope variants using an interferon-γ enzyme-linked immunospot (ELISPOT) assay. We used 11 well-defined CD8 T cell epitopes in two mouse haplotypes and, for each epitope, assessed recognition of the two vaccine forms together with the other most frequent epitope variants in the HIV-1 database. We conclude that for the magnitude and depth of epitope recognition, CD8 T cell responses benefitted in most comparisons from the combined bi-valent mosaic and envisage the main advantage of the bi-valent vaccine during its deployment to diverse populations.

Challenges targeting cancer neoantigens in 2021: a systematic literature review

Introduction: Cancer neoantigens represent important targets of cancer immunotherapy. The goal of cancer neoantigen vaccines is to induce neoantigen-specific immune responses and antitumor immunity while minimizing the potential for autoimmune toxicity. Advances in sequencing technologies, neoantigen prediction algorithms, and other technologies have dramatically improved the ability to identify and prioritize cancer neoantigens. Unfortunately, results from preclinical studies and early phase clinical trials highlight important challenges to the successful clinical translation of neoantigen cancer vaccines.Areas covered: In this review, we provide an overview of current strategies for the identification and prioritization of cancer neoantigens with a particular emphasis on the two most common strategies used for neoantigen identification: (1) direct identification of peptide ligands eluted from peptide-MHC complexes, and (2) next-generation sequencing combined with neoantigen prediction algorithms. We highlight the limitations of current neoantigen prediction pipelines, and discuss broader challenges associated with cancer neoantigen vaccines including tumor purity/heterogeneity and the immunosuppressive tumor microenvironment.Expert opinion: Despite current limitations, neoantigen prediction is likely to improve rapidly based on advances in sequencing, machine learning, and information sharing. The successful development of robust cancer neoantigen prediction strategies is likely to have a significant impact, with the potential to facilitate cancer neoantigen vaccine design.

Broad-Based Influenza-Specific CD8+ T Cell Response without the Typical Immunodominance Hierarchy and Its Potential Implication

Syngeneic murine systems have pre-fixed MHC, making them an imperfect model for investigating the impact of MHC polymorphism on immunodominance in influenza A virus (IAV) infections. To date, there are few studies focusing on MHC allelic differences and its impact on immunodominance even though it is well documented that an individual’s HLA plays a significant role in determining immunodominance hierarchy. Here, we describe a broad-based CD8⁺ T cell response in a healthy individual to IAV infection rather than a typical immunodominance hierarchy. We used a systematic antigen screen approach combined with epitope prediction to study such a broad CD8⁺ T cell response to IAV infection. We show CD8⁺ T cell responses to nine IAV proteins and identify their minimal epitope sequences. These epitopes are restricted to HLA-B*44:03, HLA-A*24:02 and HLA-A*33:03 and seven out of the nine epitopes are novel (NP_319–330^# (known and demonstrated minimal epitope positions are subscripted; otherwise, amino acid positions are shown as normal text (for example NP 319–330 or NP 313–330)), M1_124–134, M2_7–15, NA_337–346, PB2_39–49, HA_445–453 and NS1_195–203). Additionally, most of these novel epitopes are highly conserved among H1N1 and H3N2 strains that circulated in Australia and other parts of the world.

DeepNetBim: deep learning model for predicting HLA-epitope interactions based on network analysis by harnessing binding and immunogenicity information

BACKGROUND

Epitope prediction is a useful approach in cancer immunology and immunotherapy. Many computational methods, including machine learning and network analysis, have been developed quickly for such purposes. However, regarding clinical applications, the existing tools are insufficient because few of the predicted binding molecules are immunogenic. Hence, to develop more potent and effective vaccines, it is important to understand binding and immunogenic potential. Here, we observed that the interactive association constituted by human leukocyte antigen (HLA)-peptide pairs can be regarded as a network in which each HLA and peptide is taken as a node. We speculated whether this network could detect the essential interactive propensities embedded in HLA-peptide pairs. Thus, we developed a network-based deep learning method called DeepNetBim by harnessing binding and immunogenic information to predict HLA-peptide interactions.

RESULTS

Quantitative class I HLA-peptide binding data and qualitative immunogenic data (including data generated from T cell activation assays, major histocompatibility complex (MHC) binding assays and MHC ligand elution assays) were retrieved from the Immune Epitope Database database. The weighted HLA-peptide binding network and immunogenic network were integrated into a network-based deep learning algorithm constituted by a convolutional neural network and an attention mechanism. The results showed that the integration of network centrality metrics increased the power of both binding and immunogenicity predictions, while the new model significantly outperformed those that did not include network features and those with shuffled networks. Applied on benchmark and independent datasets, DeepNetBim achieved an AUC score of 93.74% in HLA-peptide binding prediction, outperforming 11 state-of-the-art relevant models. Furthermore, the performance enhancement of the combined model, which filtered out negative immunogenic predictions, was confirmed on neoantigen identification by an increase in both positive predictive value (PPV) and the proportion of neoantigen recognition.

CONCLUSIONS

We developed a network-based deep learning method called DeepNetBim as a pan-specific epitope prediction tool. It extracted the attributes of the network as new features from HLA-peptide binding and immunogenic models. We observed that not only did DeepNetBim binding model outperform other updated methods but the combination of our two models showed better performance. This indicates further applications in clinical practice.

Predicting TCR-Epitope Binding Specificity Using Deep Metric Learning and Multimodal Learning

Understanding the recognition of specific epitopes by cytotoxic T cells is a central problem in immunology. Although predicting binding between peptides and the class I Major Histocompatibility Complex (MHC) has had success, predicting interactions between T cell receptors (TCRs) and MHC class I-peptide complexes (pMHC) remains elusive. This paper utilizes a convolutional neural network model employing deep metric learning and multimodal learning to perform two critical tasks in TCR-epitope binding prediction: identifying the TCRs that bind a given epitope from a TCR repertoire, and identifying the binding epitope of a given TCR from a list of candidate epitopes. Our model can perform both tasks simultaneously and reveals that inconsistent preprocessing of TCR sequences can confound binding prediction. Applying a neural network interpretation method identifies key amino acid sequence patterns and positions within the TCR, important for binding specificity. Contrary to common assumption, known crystal structures of TCR-pMHC complexes show that the predicted salient amino acid positions are not necessarily the closest to the epitopes, implying that physical proximity may not be a good proxy for importance in determining TCR-epitope specificity. Our work thus provides an insight into the learned predictive features of TCR-epitope binding specificity and advances the associated classification tasks.

Applications of Machine and Deep Learning in Adaptive Immunity

Adaptive immunity is mediated by lymphocyte B and T cells, which respectively express a vast and diverse repertoire of B cell and T cell receptors and, in conjunction with peptide antigen presentation through major histocompatibility complexes (MHCs), can recognize and respond to pathogens and diseased cells. In recent years, advances in deep sequencing have led to a massive increase in the amount of adaptive immune receptor repertoire data; additionally, proteomics techniques have led to a wealth of data on peptide-MHC presentation. These large-scale data sets are now making it possible to train machine and deep learning models, which can be used to identify complex and high-dimensional patterns in immune repertoires. This article introduces adaptive immune repertoires and machine and deep learning related to biological sequence data and then summarizes the many applications in this field, which span from predicting the immunological status of a host to the antigen specificity of individual receptors and the engineering of immunotherapeutics.

Deciphering and predicting CD4+ T cell immunodominance of influenza virus hemagglutinin

The importance of CD4⁺ T helper (Th) cells is well appreciated in view of their essential role in the elicitation of antibody and cytotoxic T cell responses. However, the mechanisms that determine the selection of immunodominant epitopes within complex protein antigens remain elusive. Here, we used ex vivo stimulation of memory T cells and screening of naive and memory T cell libraries, combined with T cell cloning and TCR sequencing, to dissect the human naive and memory CD4⁺ T cell repertoire against the influenza pandemic H1 hemagglutinin (H1-HA). We found that naive CD4⁺ T cells have a broad repertoire, being able to recognize naturally processed as well as cryptic peptides spanning the whole H1-HA sequence. In contrast, memory Th cells were primarily directed against just a few immunodominant peptides that were readily detected by mass spectrometry–based MHC-II peptidomics and predicted by structural accessibility analysis. Collectively, these findings reveal the presence of a broad repertoire of naive T cells specific for cryptic H1-HA peptides and demonstrate that antigen processing represents a major constraint determining immunodominance.

Deep learning pan-specific model for interpretable MHC-I peptide binding prediction with improved attention mechanism

Accurate prediction of peptide binding affinity to the major histocompatibility complex (MHC) proteins has the potential to design better therapeutic vaccines. Previous work has shown that pan-specific prediction algorithms can achieve better prediction performance than other approaches. However, most of the top algorithms are neural networks based black box models. Here, we propose DeepAttentionPan, an improved pan-specific model, based on convolutional neural networks and attention mechanisms for more flexible, stable and interpretable MHC-I binding prediction. With the attention mechanism, our ensemble model consisting of 20 trained networks achieves high and more stabilized prediction performance. Extensive tests on IEDB's weekly benchmark dataset show that our method achieves state-of-the-art prediction performance on 21 test allele datasets. Analysis of the peptide positional attention weights learned by our model demonstrates its capability to capture critical binding positions of the peptides, which leads to mechanistic understanding of MHC-peptide binding with high alignment with experimentally verified results. Furthermore, we show that with transfer learning, our pan model can be fine-tuned for alleles with few samples to achieve additional performance improvement. DeepAttentionPan is freely available as an open-source software at https://github.com/jjin49/DeepAttentionPan.