Abstract CD8⁺ T cells in solid cancers progressively lose anti-tumor activity, yet the cell-intrinsic mechanisms driving this loss of function remain incompletely defined. Here, we performed matched proteomic and transcriptomic profiling of dysfunctional and bystander CD8⁺ tumor-infiltrating T cells isolated from primary tumors of treatment-naïve non-small cell lung cancer patients. Proteomic analysis revealed widespread discordance with mRNA expression, with 8% of all quantified proteins displaying differential expression exclusively at the protein level. Genetic perturbation of such differentially expressed proteins identified the chromatin remodeler CHD4 and fatty acid synthase (FASN) as cell-intrinsic regulators of T cell function. CHD4 deletion resulted in altered gene-regulatory networks that promoted effector differentiation and enhanced cytokine production. In contrast, FASN deletion preserved mitochondrial fitness and sustained T cell functionality under chronic T cell receptor stimulation. Together, these findings demonstrate that proteomic profiling uncovers regulators of T cell functionality that are not apparent from transcriptomic analyses alone, highlighting an additional layer of regulatory control. One Sentence Summary Integrated multi-omic profiling of human tumor-infiltrating T cells reveals cell-intrinsic regulators of T cell dysfunction that are missed by transcriptomic analyses alone.

While microsatellite-instable (MSI) colorectal cancers (CRC), reflecting mismatch repair deficiency, often respond to immune checkpoint inhibitors, microsatellite-stable (MSS) tumors remain largely resistant. This disparity is typically attributed to differences in neoantigen load. However, whether antigen-independent mechanisms contribute to immune evasion in MSS-CRC remains unclear. To address this, we engineered a model in which MSI- and MSS-CRC cells express identical levels of a defined antigen recognized by TCR-engineered T cells. Despite equivalent antigen presentation, MSS tumors exhibited impaired T-cell activation, reduced cytotoxicity, and resistance to killing. We linked this immune evasion to the MSS tumor secretome, which suppressed immune responses even in immunogenic MSI cells by impairing immune synapse formation. Surfaceome profiling by mass spectrometry identified glycosylation-dependent alterations that impair immune recognition. Our findings demonstrate that MSS-CRC evades immune attack via intrinsic secretome-driven mechanisms, independent of antigenicity. Targeting glycosylation-linked suppressive pathways may restore T-cell responsiveness and improve immunotherapy efficacy in MSS-CRC.

MOTIVATION:Functional screening of patient-derived T cell receptor (TCR)-neoantigen pairs via co-culture experiments is a way to design personalised immunotherapy or to investigate its mechanism of action. Current computational toolkits can either generate and prioritise candidate epitopes from tumour variants or count barcodes in sequencing data. However, they lack modules to support experimental screening, such as sample demultiplexing, construct quality control, and downstream analysis. To bridge these gaps, we present pepitope, an R package that integrates minigene library generation, sequencing-based quality control (QC), and differential abundance analysis of co-culture screens into a single software package within the accessible R/Bioconductor ecosystem. RESULTS:pepitope workflows include the extraction of mutant and reference peptides with customisable flanking regions from tumour variant calls using Bioconductor annotation resources; demultiplexing and barcode counting for construct QC; and negative-binomial-based differential testing built on DESeq2 to identify immunogenic epitopes in TCR co-culture assays. By remaining within R, pepitope lowers the barrier for lab-based biologists familiar with R and Bioconductor to perform end-to-end co-culture screen analyses without needing dedicated computational support. AVAILABILITY:pepitope (R ≥ 4.5.0) is freely available on GitHub under the GPL-3.0 license, with detailed vignettes hosted at https://mschubert.github.io/pepitope/. Installation is facilitated via the remotes package in R.

Details about methods used for mIF, DNA- and RNA isolation and sequencing and single-cell RNA sequencing

T cells perform critical roles in orchestrating immunity in health and disease. However, decoding what individual T cells recognize has long been challenging due to the immense diversity of both T cell receptors (TCRs) and potential antigens. Recent advances in high-throughput TCR screening approaches now provide an opportunity to map the antigen specificity landscape of T cells with unprecedented depth. Here, we outline these recent developments in screening methodologies and discuss how these can help advance our fundamental understanding of T cell-based immunity.

Accurate prediction of TCR specificity forms a holy grail in immunology and large language models and computational structure predictions provide a path to achieve this. Importantly, current TCR-pMHC prediction models have been trained and evaluated using historical data of unknown quality. Here, we develop and utilize a high-throughput synthetic platform for TCR assembly and evaluation to assess a large fraction of VDJdb-deposited TCR-pMHC entries using a standardized readout of TCR function. Strikingly, this analysis demonstrates that claimed TCR reactivity is only confirmed for 50% of evaluated entries. Intriguingly, the use of TCRbridge to analyze AlphaFold3 confidence metrics reveals a substantial performance in distinguishing functionally validating and non-validating TCRs even though AlphaFold3 was not trained on this task, demonstrating the utility of the validated VDJdb (TCRvdb) database that we generated. We provide TCRvdb as a resource to the community to support training and evaluation of improved predictive TCR specificity models.

Patient fractions per molecular subtype for response groups a. Fraction of patients per molecular subtype (n=5) according to the consensus classification. b. Fraction of patients per molecular subtype (n=4) according to the TCGA classification. There were no neuro-endocrine subtypes in our patient cohort. Responders (<=ypT1N0) are depicted in blue, non-responders (ypT2-4aNx or ypTxN1-3) are depicted in orange.

Tumor mutational burden (TMB) in responders vs non-responders for ipilimumab-high and ipilimumab-low. Baseline TMB (somatic non-synonymous TMB per Mb) in the ipilimumab-high cohorts (cohort 1 and 2A; left panel) and the ipilimumab-low cohort (cohort 2B; right panel) for responders (blue; <=ypT1N0) and non-responders (orange; ypT2-4aNx or ypTxN1-3).

Consort diagram of available samples per analysis for each arm at baseline and on-treatment. One patient in arm 2A and 3 patients in arm 2B did not undergo cystectomy.

PURPOSE:In NABUCCO, the safety and efficacy of preoperative ipilimumab plus nivolumab were assessed in stage III urothelial cancer. Encouraging responses were achieved, and ipilimumab 3 mg/kg (ipilimumab-high) seemed more effective than ipilimumab 1 mg/kg (ipilimumab-low). We explored ipilimumab plus nivolumab response biomarkers and tumor microenvironment (TME) treatment dynamics. PATIENTS AND METHODS:Baseline formalin-fixed, paraffin-embedded tumor tissue was analyzed using PD-L1 IHC (n = 51) and whole-exome and transcriptome sequencing (both n = 53) and correlated with response. Baseline infiltration of CD8+ T cells (n = 51) and at cystectomy (n = 42) was examined. Single-cell RNA sequencing (scRNA-seq) of CD3+ T cells was conducted on on-treatment resection tissue of two responders to ipilimumab-high to explore the characteristics of CD8+ T cells within the TME. RESULTS:High tumor mutational burden and PD-L1 positivity were associated with response to ipilimumab plus nivolumab. Nonresponding patients exhibited increased expression of a TGFβ signature. We observed increased transcription of the g2m checkpoint and e2f target in responders to ipilimumab-high and enhanced transcription of IFN-α and IFN-γ hallmarks in responders to ipilimumab-low. CD8+TCF7+ T cells accumulated in the TME of responders to ipilimumab-high. scRNA-seq of CD8A+TCF7+ T cells demonstrated enhanced expression of IL7R, CCR7, GPR15, XCL1, SELL, and LEF1. CONCLUSIONS:Our data indicate that tumor mutational burden, PD-L1, and TGFβ are potential biomarkers for response to ipilimumab plus nivolumab in stage III urothelial cancer. An inflammatory TME might be relevant for responding to ipilimumab-low. We found that in responders to ipilimumab-high, TCF7+CD8+ T cells accumulated in the TME. scRNA-seq in two responders suggested that TCF7+CD8A+ T cells express genes associated with immunologic memory formation and T-cell homing.

Mutational analysis DDR genes. Fraction and type of alterations for various DNA damage response genes in responders (blue) and non-responders (orange). The different treatment arms are depicted in light green (arm 1), dark green (arm 2A) and lilac (arm 2B).

Representativeness of Study Participants

T cell receptor (TCR) gene therapy is a potent form of cellular immunotherapy in which patient T cells are genetically engineered to express TCRs with defined tumor reactivity. However, the isolation of therapeutic TCRs is complicated by both the general scarcity of tumor-specific T cells among patient T cell repertoires and the patient-specific nature of T cell epitopes expressed on tumors. Here we describe a high-throughput, personalized TCR discovery pipeline that enables the assembly of complex synthetic TCR libraries in a one-pot reaction, followed by pooled expression in reporter T cells and functional genetic screening against patient-derived tumor or antigen-presenting cells. We applied the method to screen thousands of tumor-infiltrating lymphocyte (TIL)-derived TCRs from multiple patients and identified dozens of CD4+ and CD8+ T-cell-derived TCRs with potent tumor reactivity, including TCRs that recognized patient-specific neoantigens. Tumor-specific T cell receptors (TCRs) are identified by functional screening of synthetic TCR libraries.

The prediction of peptide-MHC (pMHC) recognition by αβ T-cell receptors (TCRs) remains a major biomedical challenge. Here, we develop STAPLER (Shared TCR And Peptide Language bidirectional Encoder Representations from transformers), a transformer language model that uses a joint TCRαβ- peptide input to allow the learning of patterns within and between TCRαβ and peptide sequences that encode recognition. First, we demonstrate how data leakage during negative data generation can confound performance estimates of neural network-based models in predicting TCR – pMHC specificity. We then demonstrate that, because of its pre-training and fine-tuning masked language modeling tasks, STAPLER outperforms both neural network-based and distance-based ML models in predicting the recognition of known antigens in an independent dataset, in particular for antigens for which little related data is available. Based on this ability to efficiently learn from limited labeled TCR- peptide data, STAPLER is well-suited to utilize growing TCR – pMHC datasets to achieve accurate prediction of TCR – pMHC specificity.

Cancer neoantigens that arise from tumor mutations are drivers of tumor-specific T cell responses, but identification of T cell-recognized neoantigens in individual patients is challenging. Previous methods have restricted antigen discovery to selected HLA alleles, thereby limiting the breadth of neoantigen repertoires that can be uncovered. Here, we develop a genetic neoantigen screening system that allows sensitive identification of CD4+ and CD8+ T cell-recognized neoantigens across patients' complete HLA genotypes.

γδT cell receptors (γδTCRs) recognize a broad range of malignantly transformed cells in mainly a major histocompatibility complex (MHC)-independent manner, making them valuable additions to the engineered immune effector cell therapy that currently focuses primarily on αβTCRs and chimeric antigen receptors (CARs). As an exception to the rule, we have previously identified a γδTCR, which exerts antitumor reactivity against HLA-A*24:02-expressing malignant cells, however without the need for defined HLA-restricted peptides, and without exhibiting any sign of off-target toxicity in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mouse models. This particular tumor-HLA-A*24:02-specific Vγ5Vδ1TCR required CD8αα co-receptor for its tumor reactive capacity when introduced into αβT cells engineered to express a defined γδTCR (TEG), referred to as TEG011; thus, it was only active in CD8 + TEG011. We subsequently explored the concept of additional redirection of CD4 + T cells through co-expression of the human CD8α gene into CD4 + and CD8 + TEG011 cells, later referred as TEG011_CD8α. Adoptive transfer of TEG011_CD8α cells in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mice injected with tumor HLA-A*24:02 + cells showed superior tumor control in comparison to TEG011, and to mock control groups. The total percentage of mice with persisting TEG011_CD8α cells, as well as the total number of TEG011_CD8α cells per mice, was significantly improved over time, mainly due to a dominance of CD4 + CD8 + double-positive TEG011_CD8α, which resulted in higher total counts of functional T cells in spleen and bone marrow. We observed that tumor clearance in the bone marrow of TEG011_CD8α-treated mice associated with better human T cell infiltration, which was not observed in the TEG011-treated group. Overall, introduction of transgenic human CD8α receptor on TEG011 improves antitumor reactivity against HLA-A*24:02 + tumor cells and further enhances in vivo tumor control.

T cell engineering strategies offer cures to patients and have entered clinical practice with chimeric antibody-based receptors; alpha beta T cell receptor (alpha beta TCR)-based strategies are, however, lagging behind. To allow a more rapid and successful translation to successful concepts also using alpha beta TCRs for engineering, incorporating a method for the purification of genetically modified T cells, as well as engineered T cell deletion after transfer into patients, could be beneficial. This would allow increased efficacy, reduced potential side effects, and improved safety of newly to-be-tested lead structures. By characterizing the antigen-binding interface of a good manufacturing process (GMP)-grade anti-alpha beta TCR antibody, usually used for depletion of abT cells from stem cell transplantation products, we developed a strategy that allows for the purification of untouched alpha beta TCR-engineered immune cells by changing 2 amino acids only in the TCR beta chain constant domain of introduced TCR chains. Alternatively, we engineered an antibody that targets an extended mutated interface of 9 amino acids in the TCR beta chain constant domain and provides the opportunity to further develop depletion strategies of engineered immune cells.