Journal Club: Decoding T Cell Exhaustion Across Cancers with Pan-Cancer Multi-Omics

· 10 min read · Paper Feed

Journal Club: Decoding T Cell Exhaustion Across Cancers with Pan-Cancer Multi-Omics

Conceptual illustration: Decoding T Cell Exhaustion Across Cancers with Pan-Cancer Multi-Omics

Introduction

T cell exhaustion has become one of the most consequential—and frustrating—features of the tumor microenvironment (TME). When T cells encounter persistent antigen stimulation, they progressively lose cytotoxic function and upregulate a characteristic repertoire of inhibitory receptors, locking them into a hyporesponsive state that undermines anti-tumor immunity. Despite the clinical success of checkpoint inhibitors targeting PD-1, LAG-3, and TIM-3, responses remain inconsistent across cancer types, suggesting that the molecular texture of T cell exhaustion is far more heterogeneous than once appreciated.

This journal club synthesizes three recent papers that tackle T cell exhaustion from complementary angles. Juthi et al. (2025) provide a focused, pan-cancer characterization of exhausted T cell (Tex) marker signatures using TCGA data, Cho et al. (2025) introduce an integrative omics platform (OncoDB 2.0) that enables the kind of cross-cancer analyses needed to study immune dysfunction at scale, and Tan et al. (2026) demonstrate how a standard systemic therapy—androgen deprivation therapy (ADT)—actively remodels the prostate cancer immune microenvironment toward an immunosuppressive, Tex-like state. Together, these works chart a trajectory from gene-level signatures, to platform-level infrastructure, to therapy-induced remodeling.

Read on to see how these threads fit together into a coherent research agenda for the field.

Paper 1: A Six-Gene Pan-Cancer Signature of T Cell Exhaustion

Juthi et al., International Journal of Molecular Sciences, 2025

Key Findings

  • A curated six-gene Tex signature (PDCD1, LAG3, TIGIT, HAVCR2, CXCL13, LAYN) shows significant differential expression in KIRC (clear cell renal cell carcinoma) and is associated with poor prognosis across multiple cancers, positioning these markers as both prognostic indicators and potential therapeutic targets.
  • The signature plays dual pathway roles: activating apoptosis, EMT, and hormone ER pathways while inhibiting DNA damage response and RTK oncogenic pathways—a context-dependent functional fingerprint.
  • Tex marker expression correlates with immune cell infiltration patterns in most cancer types, supporting the emerging model that exhausted T cells arise in inflamed but dysfunctional microenvironments.
  • Drug sensitivity analyses linked Tex marker expression to response profiles of multiple anti-cancer drugs, hinting at potential combinatorial strategies with checkpoint blockade.

Methodology & Study Design

The team integrated data from The Cancer Genome Atlas (TCGA), the Gene Expression Omnibus (GEO), and The Cancer Proteome Atlas (TCPA) to systematically profile Tex markers at the mRNA level, with downstream analyses of mutation profiles, pathway enrichment, immune infiltration (using deconvolution approaches), and drug sensitivity correlations. The study is observational and bioinformatics-driven, relying on the assumption that mRNA expression of these six genes reasonably approximates the abundance and dysfunction state of exhausted T cells across tumors.

Significance

This work fills a translational gap by moving from a vague concept of "exhausted T cells" to a measurable, six-gene signature applicable across cancer types. It validates Tex markers as clinically meaningful biomarkers and provides a reusable framework that can plug into larger omics platforms like OncoDB 2.0 (discussed below). Importantly, the authors emphasize that "reviving exhausted T cells" via these insights may unlock synergistic immunotherapy responses in Tex-high subtypes.

Paper 2: OncoDB 2.0 — Infrastructure for Integrated Pan-Cancer Omics

Cho et al., Nucleic Acids Research, 2025

Key Findings

  • OncoDB 2.0 (freely available at oncodb.org) expands the original platform to integrate RNA expression, DNA methylation, somatic mutations, clinical annotations, oncoviral signatures, proteomic data from CPTAC, and chromatin accessibility data from TCGA in a single, queryable interface.
  • The platform now includes a somatic mutation atlas built from combined DNA and RNA sequencing, enabling cross-tumor mutation pattern discovery linked to clinical features.
  • New multi-omics analysis modules allow users to jointly examine RNA expression, DNA methylation, and somatic mutations—precisely the cross-omic exploration needed for studies like Juthi et al.'s Tex signature work or Tan et al.'s ADT-induced remodeling.

Methodology & Study Design

Rather than a hypothesis-driven study, OncoDB 2.0 is a computational infrastructure project. It uses TCGA and Genotype-Tissue Expression (GTEx) data as primary backbones, augments with CPTAC proteomics and oncoviral integration, and exposes precomputed analyses through a web interface. The authors prioritized usability, reproducibility, and extensibility—features that lower the barrier for labs without heavy bioinformatics infrastructure to perform pan-cancer analyses.

Significance

OncoDB 2.0 represents a methodological democratization of pan-cancer omics. For labs studying immune signatures like Tex, the platform eliminates redundant data wrangling and provides a standardized reference for cross-cohort validation. It also enables researchers to test whether observations from one study (e.g., Tex marker expression in KIRC) hold across related omics layers—methylation, proteomics, or chromatin accessibility—facilitating the kind of vertical validation increasingly demanded in translational research.

Paper 3: ADT Weakens T Cell Killing and Enriches Immunosuppression in Prostate Cancer

Tan et al., Journal of Peking University. Health Sciences, 2026

Key Findings

  • In 12 prostate cancer patients receiving ADT at Peking University People's Hospital, serum PSA and testosterone dropped significantly, but peripheral blood immune cell proportions and most cytokines showed no significant change—suggesting systemic immune effects of ADT are subtler than local tumor effects.
  • PDX model transcriptomic analysis (GSE41193, n=5) revealed that castration-induced differentially expressed genes enriched in neural-related and immune regulation pathways, with trending changes in immune cell infiltration.
  • Single-cell RNA-seq on 4 prostate tumor tissues (2 neoadjuvant ADT, 2 untreated) showed a decreased proportion of intratumoral effector CD8⁺ T cells with downregulated cytotoxicity genes (GZMA, GZMB, GNLY, NKG7) and relative enrichment of regulatory T cells (Tregs).
  • Hallmark pathway analysis revealed suppression of interferon signaling after ADT, reinforcing the immunosuppressive TIME interpretation.

Methodology & Study Design

This is a multi-level (multi-scale) study combining three complementary data types: (1) prospective patient serum monitoring for systemic immunity, (2) public PDX transcriptomic data for in vivo modeling, and (3) in-house scRNA-seq for high-resolution tumor profiling. The integration of systemic, animal-model, and single-cell human data is the study's distinguishing strength.

Significance

Prostate cancer is historically resistant to immune checkpoint inhibitors, and Tan et al. provide a mechanistic explanation: ADT, the backbone of prostate cancer treatment, may actively drive the TIME toward an immunosuppressive, Tex-like state. This reframes ADT as not merely a hormonal therapy but as an immune-modulating intervention with complex consequences. The findings provide a theoretical basis for combining ADT with immune-modulating strategies (e.g., checkpoint blockade, myeloid-targeted therapy) and underscore the importance of timing and sequence in such combinations.

Synthesis & Discussion

How These Papers Complement Each Other

These three papers form a logical progression:

  1. Juthi et al. (1) define what Tex looks like at the gene-expression level across cancers.
  2. Cho et al. (2) provide the infrastructure to query, validate, and extend such signatures across multi-omic layers.
  3. Tan et al. (3) demonstrate why Tex-like states matter clinically—showing that standard therapies can drive tumors into immunosuppressive, exhausted-like phenotypes.

Agreements and Nuances

All three support the principle that specific molecular signatures capture clinically relevant immune dysfunction. Juthi et al. and Tan et al. independently link T cell functional impairment (Tex signatures; reduced GZMA/GZMB/GNLY/NKG7 expression with Treg enrichment) with worse outcomes or immunosuppressive TMEs. Cho et al. implicitly support this by making such signatures discoverable and analyzable at scale.

However, a critical nuance emerges: density ≠ quality of immune response. Juthi et al. show that Tex-high tumors often have high immune infiltration but dysfunctional T cells—a "hot but suppressed" phenotype. Tan et al. show that ADT-treated tumors can have decreased effector CD8⁺ T cells with Treg enrichment—a "cold and suppressed" phenotype. Both are immunosuppressive but mechanistically and therapeutically distinct, suggesting that stratifying tumors by both immune density and exhaustion quality may be necessary for predicting checkpoint inhibitor response.

Emerging Patterns

  • From descriptive to integrative: Tex signature work, ADT-induced immune profiling, and omics platforms are converging toward standardized cross-cancer immune signatures supported by multi-omics infrastructure.
  • Therapy-induced immune remodeling: ADT exemplifies how standard systemic therapies reshape the TME toward immunosuppression, paralleling concerns about treatment-induced Tex and immune escape in other cancers.
  • From static to dynamic: The field is moving from baseline biomarkers toward dynamic profiling of the TIME during treatment, as Tan et al. demonstrate with longitudinal ADT monitoring.

Unanswered Questions

  • Do ADT-induced changes in prostate cancer partly operate through canonical Tex programs, or via distinct androgen-receptor-driven mechanisms?
  • Can pan-cancer Tex signatures (like Juthi et al.'s six-gene set) predict functional exhaustion under treatment pressure, or do they need augmentation with treatment-specific markers?
  • How can platforms like OncoDB 2.0 incorporate single-cell, spatial, and longitudinal data to capture dynamic Tex and ADT-induced changes at higher resolution?

For Your Lab Meeting

  1. On signature robustness: Juthi et al. use six Tex markers. Given known heterogeneity of exhausted T cell states (progenitor, intermediate, terminal), is a six-gene signature sufficient to capture clinically relevant subtypes, or would a larger gene panel (e.g., the 40+ genes referenced in recent work) yield better stratification?

  2. On cross-platform validation: If you replicated Juthi et al.'s KIRC findings in OncoDB 2.0 using methylation or proteomic layers, would you expect concordance, or would each omics layer tell a different story about Tex biology?

  3. On experimental design: The Tan et al. scRNA-seq cohort (n=4) is small. What statistical and biological considerations should guide interpretation, and what would be the minimum viable cohort size for definitive claims about ADT-induced immune remodeling?

  4. On therapeutic implications: If ADT actively suppresses T cell killing, should we reconsider the timing of checkpoint inhibitor administration relative to ADT in prostate cancer trials? What preclinical models would best test this?

  5. On platform integration: How might single-cell and spatial omics data be best integrated into platforms like OncoDB 2.0 to support dynamic, treatment-aware immune profiling rather than static baseline analyses?

Key Terms

  • T Cell Exhaustion (Tex): A hyporesponsive state of T cells characterized by sustained inhibitory receptor expression (PD-1, LAG-3, TIM-3, TIGIT), loss of effector function (reduced cytokine production, cytotoxicity), and a distinct transcriptional program that emerges under chronic antigen stimulation.

  • Tumor Immune Microenvironment (TIME): The cellular and molecular ecosystem surrounding a tumor, including immune cells (T cells, B cells, macrophages, myeloid-derived suppressor cells), stromal cells, cytokines, and extracellular matrix components, which collectively shape anti-tumor immunity.

  • Androgen Deprivation Therapy (ADT): A systemic treatment for prostate cancer that reduces circulating androgens (testosterone, DHT), depriving androgen receptor–driven tumors of proliferative signals. Tan et al. show it has complex, partially suppressive effects on tumor immunity.

  • Pan-Cancer Analysis: A research approach that examines molecular features across many cancer types simultaneously, often using TCGA or similar large-cohort datasets, to identify both cancer-specific and shared biological patterns.

  • Hallmark Pathway Analysis: A gene set enrichment method that tests whether predefined biological pathways (e.g., interferon signaling, apoptosis, EMT) are coordinately up- or downregulated in a given gene list, commonly used to interpret transcriptomic and proteomic data.

Bottom line: These three papers illustrate how gene-level signatures, omics infrastructure, and therapy-induced remodeling research are converging into a coherent framework for understanding—and ultimately reversing—immune dysfunction in cancer. For labs working in this space, the path forward involves integrating dynamic single-cell and spatial data into platforms like OncoDB 2.0, validating pan-cancer Tex signatures across treatment contexts, and designing rational combination strategies that account for therapy-induced immunosuppression.

References

  1. Rifat Tasnim Juthi, Saiful Arefeen Sazed, M. Mareboina, A. Zaravinos, & I. Georgakopoulos-Soares (2025). Characterization of Exhausted T Cell Signatures in Pan-Cancer Settings. International Journal of Molecular Sciences. https://doi.org/10.3390/ijms26052311

  2. Minsu Cho, Gongyu Tang, Charles S. Rogers, Makayla Dove, Xinyi Liu, Yuanxiang Li, et al. (2025). OncoDB 2.0: a comprehensive platform for integrated pan-cancer omics analysis. Nucleic Acids Research. https://doi.org/10.1093/nar/gkaf952

  3. Tan Z, DU Y, Qin C, & Xu T (2026). [A multi-level study of androgen deprivation therapy on the immune microenvironment in prostate cancer].. Beijing da xue xue bao. Yi xue ban = Journal of Peking University. Health sciences. https://europepmc.org/article/MED/42493444

Cancer Genomics Oncology Tumor