Rapid TCR:Epitope Ranker (RAPTER): a primary human T cell reactivity screening assay pairing epitope and TCR at single cell resolution.

Identifying epitopes that T cells respond to is critical for understanding T cell-mediated immunity. Traditional multimer and other single cell assays often require large blood volumes and/or expensive HLA-specific reagents and provide limited phenotypic and functional information. Here, we present the Rapid TCR:Epitope Ranker (RAPTER) assay, a single cell RNA sequencing (scRNA-SEQ) method that uses primary human T cells and antigen presenting cells (APCs) to assess functional T cell reactivity. Using hash-tag oligonucleotide (HTO) coding and T cell activation-induced markers (AIM), RAPTER defines paired epitope specificity and TCR sequence and can include RNA- and protein-level T cell phenotype information. We demonstrate that RAPTER identified specific reactivities to viral and tumor antigens at sensitivities as low as 0.15% of total CD8+ T cells, and deconvoluted low-frequency circulating HPV16-specific T cell clones from a cervical cancer patient. The specificities of TCRs identified by RAPTER for MART1, EBV, and influenza epitopes were functionally confirmed in vitro. In summary, RAPTER identifies low-frequency T cell reactivities using primary cells from low blood volumes, and the resulting paired TCR:ligand information can directly enable immunogenic antigen selection from limited patient samples for vaccine epitope inclusion, antigen-specific TCR tracking, and TCR cloning for further therapeutic development.

Spontaneous tumor regression mediated by human T cells in a humanized immune system mouse model.

Immunodeficient mice reconstituted with a human immune system (HIS mice) give rise to human T cells, which make them an attractive system to study human immune responses to tumors. However, such HIS mice typically exhibit sub-optimal responses to immune challenges as well as fail to develop antigen-specific B or T cell memory. Here we report HIS mice mediate spontaneous regression of human B cell lymphoma Raji. Tumor regression was dependent on CD4+ and CD8+ T cell responses and resulted in T cell memory. The T cell memory elicited was mainly Raji-specific, however some level of cross-protection was also elicited to a related B cell lymphoma cell line Ramos. Single-cell RNAseq analysis indicated activation of CD8+ T cells in regressing Raji tumors as well as clonal expansion of specific T cell receptors (TCRs). Cloning of TCRs from Raji-infiltrating T cells into a Jurkat reporter cell line showed reactivity specific for Raji tumor cells. Overall, we report a platform for studying in vivo human T cell tumor immunity by highlighting spontaneous Raji tumor regression, clonal TCR expansion, and T cell memory in HIS mice.

A single-cell map of dynamic chromatin landscapes of immune cells in renal cell carcinoma.

A complete chart of the chromatin regulatory elements of immune cells in patients with cancer and their dynamic behavior is necessary to understand the developmental fates and guide therapeutic strategies. Here, we map the single-cell chromatin landscape of immune cells from blood, normal tumor-adjacent kidney tissue and malignant tissue from patients with early-stage clear cell renal cell carcinoma (ccRCC). We catalog the T cell states dictated by tissue-specific and developmental-stage-specific chromatin accessibility patterns, infer key chromatin regulators and observe rewiring of regulatory networks in the progression to dysfunction in CD8+ T cells. Unexpectedly, among the transcription factors orchestrating the path to dysfunction, NF-κB is associated with a pro-apoptotic program in late stages of dysfunction in tumor-infiltrating CD8+ T cells. Importantly, this epigenomic profiling stratified ccRCC patients based on a NF-κB-driven pro-apoptotic signature. This study provides a rich resource for understanding the functional states and regulatory dynamics of immune cells in ccRCC.

Humanization of T cell-mediated immunity in mice.

Despite the enormous promise of T cell therapies, the isolation and study of human T cell receptors (TCRs) of dedicated specificity remains a major challenge. To overcome this limitation, we generated mice with a genetically humanized system of T cell immunity. We used VelociGene technology to replace the murine TCRαß variable regions, along with regions encoding the extracellular domains of co-receptors CD4 and CD8, and major histocompatibility complex (MHC) class I and II, with corresponding human sequences. The resulting "VelociT" mice have normal myeloid and lymphoid immune cell populations, including thymic and peripheral αß T cell subsets comparable with wild-type mice. VelociT mice expressed a diverse TCR repertoire, mounted functional T cell responses to lymphocytic choriomeningitis virus infection, and could develop experimental autoimmune encephalomyelitis. Immunization of VelociT mice with human tumor-associated peptide antigens generated robust, antigen-specific responses and led to identification of a TCR against tumor antigen New York esophageal squamous cell carcinoma-1 with potent antitumor activity. These studies demonstrate that VelociT mice mount clinically relevant T cell responses to both MHC-I­ and MHC-II­restricted antigens, providing a powerful new model for analyzing T cell function in human disease. Moreover, VelociT mice are a new platform for de novo discovery of therapeutic human TCRs.

A framework for highly multiplexed dextramer mapping and prediction of T cell receptor sequences to antigen specificity.

T cell receptor (TCR) antigen-specific recognition is essential for the adaptive immune system. However, building a TCR-antigen interaction map has been challenging due to the staggering diversity of TCRs and antigens. Accordingly, highly multiplexed dextramer-TCR binding assays have been recently developed, but the utility of the ensuing large datasets is limited by the lack of robust computational methods for normalization and interpretation. Here, we present a computational framework comprising a novel method, ICON (Integrative COntext-specific Normalization), for identifying reliable TCR-pMHC (peptide-major histocompatibility complex) interactions and a neural network-based classifier TCRAI that outperforms other state-of-the-art methods for TCR-antigen specificity prediction. We further demonstrated that by combining ICON and TCRAI, we are able to discover novel subgroups of TCRs that bind to a given pMHC via different mechanisms. Our framework facilitates the identification and understanding of TCR-antigen-specific interactions for basic immunological research and clinical immune monitoring.

Endogenous retroviral proteins provide an immunodominant but not requisite antigen in a murine immunotherapy tumor model.

Clinical observations suggest that responses to cancer immunotherapy are correlated with intra-tumoral T cell receptor (TCR) clonality, tumor mutation burden (TMB) and host HLA genotype, highlighting the importance of host T cell recognition of tumor antigens. However, the dynamic interplay between T cell activation state and changes in TCR repertoire in driving the identification of potential immunodominant antigen(s) remains largely unexplored. Here, we performed single-cell RNA-sequencing on CD8+ tumor-infiltrating T cells (TILs) using the murine colorectal tumor model MC38 to identify unique TCR sequences and validate their tumor reactivity. We found that the majority of clonally expanded TILs are tumor-reactive and their TCR repertoire is unique amongst individual MC38 tumor-bearing mice. Our query identified that multiple expanded TCR clones recognized the retroviral epitope p15E as an immunodominant antigen. In addition, we found that the endogenous retroviral genome encoding for p15E is highly expressed in MC38 tumors, but not in normal tissues, due to epigenetic derepression. Further, we demonstrated that the p15E-specific TILs exhibit an activated phenotype and an increase in frequency upon treatment with anti-41BB and anti-PD-1 combination immunotherapy. Importantly, we showed that although p15E-specific TILs are not required to mount a primary anti-tumor response, they contributed to the development of strong immune memory. Overall our results revealed that endogenous retroviral antigens expressed by tumor cells may represent an important and underappreciated category of tumor antigens that could be readily targeted in the clinic.

In-silico discovery of cancer-specific peptide-HLA complexes for targeted therapy.

BACKGROUND: Major Histocompatibility Complex (MHC) or Human Leukocyte Antigen (HLA) Class I molecules bind to peptide fragments of proteins degraded inside the cell and display them on the cell surface. We are interested in peptide-HLA complexes involving peptides that are derived from proteins specifically expressed in cancer cells. Such complexes have been shown to provide an effective means of precisely targeting cancer cells by engineered T-cells and antibodies, which would be an improvement over current chemotherapeutic agents that indiscriminately kill proliferating cells. An important concern with the targeting of peptide-HLA complexes is off-target toxicity that could occur due to the presence of complexes similar to the target complex in cells from essential, normal tissues. RESULTS: We developed a novel computational strategy for identifying potential peptide-HLA cancer targets and evaluating the likelihood of off-target toxicity associated with these targets. Our strategy combines sequence-based and structure-based approaches in a unique way to predict potential off-targets. The focus of our work is on the complexes involving the most frequent HLA class I allele HLA-A*02:01. Using our strategy, we predicted the off-target toxicity observed in past clinical trials. We employed it to perform a first-ever comprehensive exploration of the human peptidome to identify cancer-specific targets utilizing gene expression data from TCGA (The Cancer Genome Atlas) and GTEx (Gene Tissue Expression), and structural data from PDB (Protein Data Bank). We have thus identified a list of 627 peptide-HLA complexes across various TCGA cancer types. CONCLUSION: Peptide-HLA complexes identified using our novel strategy could enable discovery of cancer-specific targets for engineered T-cells or antibody based therapy with minimal off-target toxicity.

Targeting the Src Homology 2 (SH2) Domain of Signal Transducer and Activator of Transcription 6 (STAT6) with Cell-Permeable, Phosphatase-Stable Phosphopeptide Mimics Potently Inhibits Tyr641 Phosphorylation and Transcriptional Activity.

Signal transducer and activator of transcription 6 (STAT6) transmits signals from cytokines IL-4 and IL-13 and is activated in allergic airway disease. We are developing phosphopeptide mimetics targeting the SH2 domain of STAT6 to block recruitment to phosphotyrosine residues on IL-4 or IL-13 receptors and subsequent Tyr641 phosphorylation to inhibit the expression of genes contributing to asthma. Structure-affinity relationship studies showed that phosphopeptides based on Tyr631 from IL-4Rα bind with weak affinity to STAT6, whereas replacing the pY+3 residue with simple aryl and alkyl amides resulted in affinities in the mid to low nM range. A set of phosphatase-stable, cell-permeable prodrug analogues inhibited cytokine-stimulated STAT6 phosphorylation in both Beas-2B human airway cells and primary mouse T-lymphocytes at concentrations as low as 100 nM. IL-13-stimulated expression of CCL26 (eotaxin-3) was inhibited in a dose-dependent manner, demonstrating that targeting the SH2 domain blocks both phosphorylation and transcriptional activity of STAT6.

DINC: a new AutoDock-based protocol for docking large ligands.

BACKGROUND: Using the popular program AutoDock, computer-aided docking of small ligands with 6 or fewer rotatable bonds, is reasonably fast and accurate. However, docking large ligands using AutoDock's recommended standard docking protocol is less accurate and computationally slow. RESULTS: In our earlier work, we presented a novel AutoDock-based incremental protocol (DINC) that addresses the limitations of AutoDock's standard protocol by enabling improved docking of large ligands. Instead of docking a large ligand to a target protein in one single step as done in the standard protocol, our protocol docks the large ligand in increments. In this paper, we present three detailed examples of docking using DINC and compare the docking results with those obtained using AutoDock's standard protocol. We summarize the docking results from an extended docking study that was done on 73 protein-ligand complexes comprised of large ligands. We demonstrate not only that DINC is up to 2 orders of magnitude faster than AutoDock's standard protocol, but that it also achieves the speed-up without sacrificing docking accuracy. We also show that positional restraints can be applied to the large ligand using DINC: this is useful when computing a docked conformation of the ligand. Finally, we introduce a webserver for docking large ligands using DINC. CONCLUSIONS: Docking large ligands using DINC is significantly faster than AutoDock's standard protocol without any loss of accuracy. Therefore, DINC could be used as an alternative protocol for docking large ligands. DINC has been implemented as a webserver and is available at http://dinc.kavrakilab.org. Applications such as therapeutic drug design, rational vaccine design, and others involving large ligands could benefit from DINC and its webserver implementation.

Binding modes of peptidomimetics designed to inhibit STAT3.

STAT3 is a transcription factor that has been found to be constitutively activated in a number of human cancers. Dimerization of STAT3 via its SH2 domain and the subsequent translocation of the dimer to the nucleus leads to transcription of anti-apoptotic genes. Prevention of the dimerization is thus an attractive strategy for inhibiting the activity of STAT3. Phosphotyrosine-based peptidomimetic inhibitors, which mimic pTyr-Xaa-Yaa-Gln motif and have strong to weak binding affinities, have been previously investigated. It is well-known that structures of protein-inhibitor complexes are important for understanding the binding interactions and designing stronger inhibitors. Experimental structures of inhibitors bound to the SH2 domain of STAT3 are, however, unavailable. In this paper we describe a computational study that combined molecular docking and molecular dynamics to model structures of 12 peptidomimetic inhibitors bound to the SH2 domain of STAT3. A detailed analysis of the modeled structures was performed to evaluate the characteristics of the binding interactions. We also estimated the binding affinities of the inhibitors by combining MMPB/GBSA-based energies and entropic cost of binding. The estimated affinities correlate strongly with the experimentally obtained affinities. Modeling results show binding modes that are consistent with limited previous modeling studies on binding interactions involving the SH2 domain and phosphotyrosine(pTyr)-based inhibitors. We also discovered a stable novel binding mode that involves deformation of two loops of the SH2 domain that subsequently bury the C-terminal end of one of the stronger inhibitors. The novel binding mode could prove useful for developing more potent inhibitors aimed at preventing dimerization of cancer target protein STAT3.

On modeling peptidomimetics in complex with the SH2 domain of Stat3.

Signal transducer and activator of transcription 3 (Stat3) plays a role in human cancers. One of the main approaches towards inhibiting its activity is the development of phosphopetides or peptidomimetics that competitively bind to the SH2 domain of Stat3. This work reports, to the best of our knowledge, the first computational molecular docking study to model all of the 142 peptidomimetics that mimic the Stat3 inhibitory pTyr-X-X-Glu motif. We used the docking programs AUTODOCK and VINA to model SH2 domain-peptidomimetic complexes and estimate their binding affinities. We obtained better screening accuracy using AUTODOCK which ranked the most potent inhibitor as second highest. Experimental binding energy values and scores from docking programs correlated poorly, confirming the limitations of many current docking programs when dealing with ligands that have a large number of rotatable bonds. Nevertheless, for close to 65% of peptidomimetics, the structures of complexes computed by AUTODOCK are in agreement with current understanding of the structures. Modeling of the SH2 domain-peptidomimetic complexes is essential to better understand and design drug compounds for curing cancer. Our study is an important first step forward towards that goal.

Modeling discrete heterogeneity in X-ray diffraction data by fitting multi-conformers.

The native state of a protein is regarded to be an ensemble of conformers, which allows association with binding partners. While some of this structural heterogeneity is retained upon crystallization, reliably extracting heterogeneous features from diffraction data has remained a challenge. In this study, a new algorithm for the automatic modelling of discrete heterogeneity is presented. At high resolution, the authors' single multi-conformer model, with correlated structural features to represent heterogeneity, shows improved agreement with the diffraction data compared with a single-conformer model. The model appears to be representative of the set of structures present in the crystal. In contrast, below 2 A resolution representing ambiguous electron density by correlated multi-conformers in a single model does not yield better agreement with the experimental data. Consistent with previous studies, this suggests that variability in multi-conformer models at lower resolution levels reflects uncertainty more than coordinated motion.

Efficient algorithms to explore conformation spaces of flexible protein loops.

Several applications in biology - e.g., incorporation of protein flexibility in ligand docking algorithms, interpretation of fuzzy X-ray crystallographic data, and homology modeling - require computing the internal parameters of a flexible fragment (usually, a loop) of a protein in order to connect its termini to the rest of the protein without causing any steric clash. One must often sample many such conformations in order to explore and adequately represent the conformational range of the studied loop. While sampling must be fast, it is made difficult by the fact that two conflicting constraints - kinematic closure and clash avoidance - must be satisfied concurrently. This paper describes two efficient and complementary sampling algorithms to explore the space of closed clash-free conformations of a flexible protein loop. The "seed sampling" algorithm samples broadly from this space, while the "deformation sampling" algorithm uses seed conformations as starting points to explore the conformation space around them at a finer grain. Computational results are presented for various loops ranging from 5 to 25 residues. More specific results also show that the combination of the sampling algorithms with a functional site prediction software (FEATURE) makes it possible to compute and recognize calcium-binding loop conformations. The sampling algorithms are implemented in a toolkit (LoopTK), which is available at https://simtk.org/home/looptk.