Connecting MHC-I-binding motifs with HLA alleles via deep learning

Diverse approaches to isolate HLA class I molecules from bacterial inclusion bodies, forming heterotrimeric complexes

Production of recombinant human leukocyte antigen class I (HLA-I) proteins in vitro is fundamental for molecular immunology. However, HLA-I protein refolding has remained inefficient due to challenges in the assembling of the trimolecular complex. Here, we compare various in vitro refolding methods that address the challenges of intrachain disulfide bond formation and assembly of the complex between the light and heavy chains in the presence of the target peptide. We developed methods that uncouple the oxidation of disulfide bond formation of both subunits of HLA-I, followed by renaturation to promote complex formation. CuSO4-catalyzed air oxidation enhances correct disulfide bond formation when the protein is solubilized with N-lauryl-sarcosine (sarkosyl); however, careful removal of sarkosyl did not prevent heavy chain aggregation. We modified the classical method of HLA-I refolding by pre-oxidizing the β2m light chain before adding the HLA-I heavy chain and peptide. This method yielded successful complex refolding for HLA-A∗02:01/GILGFVFTL at 24.2 % efficiency, and HLA-C∗12:03/KAYNVTQAF at 14.5 % efficiency. Our results suggest that pre-folded β2m improves refolding efficiency of HLA-I molecules. This work presents novel approaches to HLA-I refolding that may be applied to other difficult-to-fold protein complexes.

PepCARES: A Comprehensive Advanced Refinement and Evaluation System for Peptide Design and Affinity Screening

Peptides are crucial in vaccine research, and their remarkable specificity and efficacy make them a promising potential drug class. However, designing and screening these peptides computationally is challenging. Here, we present the comprehensive advanced refinement and evaluation system (PepCARES), a program utilizing our novel model called PeptideMPNN and score evaluation for peptide design and affinity screening. PeptideMPNN, built on ProteinMPNN with transfer learning, significantly enhances sequence recovery (by 26.26%) and reduces perplexity (by 0.536) in a sequence generation task. We designed peptides targeting two HLA alleles and, using MHCfovea and PDBePISA, identified candidates with high potential. From 20 designed peptides, 14 and 7 peptides were selected, respectively. Our research provides a method for designing and screening peptides, making an important step toward the development of peptide-based vaccines.

Building trust in deep learning-based immune response predictors with interpretable explanations

The ability to predict whether a peptide will get presented on Major Histocompatibility Complex (MHC) class I molecules has profound implications in designing vaccines. Numerous deep learning-based predictors for peptide presentation on MHC class I molecules exist with high levels of accuracy. However, these MHC class I predictors are treated as black-box functions, providing little insight into their decision making. To build turst in these predictors, it is crucial to understand the rationale behind their decisions with human-interpretable explanations. We present MHCXAI, eXplainable AI (XAI) techniques to help interpret the outputs from MHC class I predictors in terms of input peptide features. In our experiments, we explain the outputs of four state-of-the-art MHC class I predictors over a large dataset of peptides and MHC alleles. Additionally, we evaluate the reliability of the explanations by comparing against ground truth and checking their robustness. MHCXAI seeks to increase understanding of deep learning-based predictors in the immune response domain and build trust with validated explanations.

ERAP1 and ERAP2 Haplotypes Influence Suboptimal HLA-B*27:05-Restricted Anti-Viral CD8+ T Cell Responses Cross-Reactive to Self-Epitopes

The human leukocyte antigen (HLA)-B*27 family of alleles is strongly associated with ankylosing spondylitis (AS), a chronic inflammatory disorder affecting the axial and peripheral joints, yet some HLA-B*27 variants not associated with AS have been shown. Since no major differences in the ligandome of associated compared to not-associated alleles have emerged, a plausible hypothesis is that the quantity rather than the quality of the presented epitopes makes the difference. In addition, the Endoplasmic Reticulum AminoPeptidases (ERAPs) 1 and 2, playing a crucial role in shaping the HLA class I epitopes, act as strong AS susceptibility factors, suggesting that an altered peptidome might be responsible for the activation of pathogenic CD8+ T cells. In this context, we have previously singled out a B*27:05-restricted CD8+ T cell response against pEBNA3A (RPPIFIRRL), an EBV peptide lacking the B*27 classic binding motif. Here, we show that a specific ERAP1/2 haplotype negatively correlates with such response in B*27:05 subjects. Moreover, we prove that the B*27:05 allele successfully presents peptides with the same suboptimal N-terminal RP motif, including the self-peptide, pDYNEIN (RPPIFGDFL). Overall, this study underscores the cooperation between the HLA-B*27 and ERAP1/2 allelic variants in defining CD8+ T cell reactivity to suboptimal viral and self-B*27 peptides and prompts further investigation of the B*27:05 peptidome composition.

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.

HLA Class I Supertype Classification Based on Structural Similarity

HLA class I proteins, a critical component in adaptive immunity, bind and present intracellular Ags to CD8+ T cells. The extreme polymorphism of HLA genes and associated peptide binding specificities leads to challenges in various endeavors, including neoantigen vaccine development, disease association studies, and HLA typing. Supertype classification, defined by clustering functionally similar HLA alleles, has proven helpful in reducing the complexity of distinguishing alleles. However, determining supertypes via experiments is impractical, and current in silico classification methods exhibit limitations in stability and functional relevance. In this study, by incorporating three-dimensional structures we present a method for classifying HLA class I molecules with improved breadth, accuracy, stability, and flexibility. Critical for these advances is our finding that structural similarity highly correlates with peptide binding specificity. The new classification should be broadly useful in peptide-based vaccine development and HLA-disease association studies.

Identifying Drug Targets of Oral Squamous Cell Carcinoma through a Systems Biology Method and Genome-Wide Microarray Data for Drug Discovery by Deep Learning and Drug Design Specifications

In this study, we provide a systems biology method to investigate the carcinogenic mechanism of oral squamous cell carcinoma (OSCC) in order to identify some important biomarkers as drug targets. Further, a systematic drug discovery method with a deep neural network (DNN)-based drug–target interaction (DTI) model and drug design specifications is proposed to design a potential multiple-molecule drug for the medical treatment of OSCC before clinical trials. First, we use big database mining to construct the candidate genome-wide genetic and epigenetic network (GWGEN) including a protein–protein interaction network (PPIN) and a gene regulatory network (GRN) for OSCC and non-OSCC. In the next step, real GWGENs are identified for OSCC and non-OSCC by system identification and system order detection methods based on the OSCC and non-OSCC microarray data, respectively. Then, the principal network projection (PNP) method was used to extract core GWGENs of OSCC and non-OSCC from real GWGENs of OSCC and non-OSCC, respectively. Afterward, core signaling pathways were constructed through the annotation of KEGG pathways, and then the carcinogenic mechanism of OSCC was investigated by comparing the core signal pathways and their downstream abnormal cellular functions of OSCC and non-OSCC. Consequently, HES1, TCF, NF-κB and SP1 are identified as significant biomarkers of OSCC. In order to discover multiple molecular drugs for these significant biomarkers (drug targets) of the carcinogenic mechanism of OSCC, we trained a DNN-based drug–target interaction (DTI) model by DTI databases to predict candidate drugs for these significant biomarkers. Finally, drug design specifications such as adequate drug regulation ability, low toxicity and high sensitivity are employed to filter out the appropriate molecular drugs metformin, gefitinib and gallic-acid to combine as a potential multiple-molecule drug for the therapeutic treatment of OSCC.

Individual HLA heterogeneity and its implications for cellular immune evasion in cancer and beyond

Structural and functional variability of human leukocyte antigen (HLA) is the foundation for competent adaptive immune responses against pathogen and tumor antigens as it assures the breadth of the presented immune-peptidome, theoretically sustaining an efficient and diverse T cell response. This variability is presumably the result of the continuous selection by pathogens, which over the course of evolution shaped the adaptive immune system favoring the assortment of a hyper-polymorphic HLA system able to elaborate efficient immune responses. Any genetic alteration affecting this diversity may lead to pathological processes, perturbing antigen presentation capabilities, T-cell reactivity and, to some extent, natural killer cell functionality. A highly variable germline HLA genotype can convey immunogenetic protection against infections, be associated with tumor surveillance or influence response to anti-neoplastic treatments. In contrast, somatic aberrations of HLA loci, rearranging the original germline configuration, theoretically decreasing its variability, can facilitate mechanisms of immune escape that promote tumor growth and immune resistance. The purpose of the present review is to provide a unified and up-to-date overview of the pathophysiological consequences related to the perturbations of the genomic heterogeneity of HLA complexes and their impact on human diseases, with a special focus on cancer.