Nature Communications | Extracellular Vesicle Proteins Monitor and Interfere with CLDN18.2-Targeted CAR-T and Antibody Therapies for Gastrointestinal Cancers

This study reveals how specific proteins carried by extracellular vesicles induce immune cell exhaustion and remodel the tumor microenvironment, providing critical experimental design insights for investigating resistance mechanisms to immunotherapy and developing combination strategies for gastrointestinal cancers.

 

Literature Overview

The article "Extracellular vesicle proteins monitor and interfere with CLDN18.2-targeted CAR-T cell and antibody therapies against gastrointestinal cancers," published in Nature Communications, systematically explores the dual role of proteins carried by circulating extracellular vesicles (EVs) in CLDN18.2-targeted therapies: serving as predictive biomarkers for efficacy and elucidating the molecular mechanisms that interfere with therapeutic outcomes.

Background Knowledge

As a key target for gastrointestinal cancers such as gastric and pancreatic cancer, CLDN18.2-targeted CAR-T and monoclonal antibody therapies have demonstrated clinical potential. However, patient response rates vary significantly, and there is a lack of reliable biomarkers to predict long-term benefits. The primary bottleneck in current CLDN18.2 research is the absence of liquid biopsy indicators capable of dynamically reflecting changes in the tumor immune microenvironment (TIME), and the mechanism by which tumor-derived extracellular vesicles interfere with treatment efficacy via specific cargo proteins (e.g., PD-1, PD-L2) remains unclear. This study utilizes high-throughput EV protein array technology to analyze plasma samples from clinical trial participants, aiming to identify key EV proteins that predict efficacy and reveal resistance mechanisms, thereby proposing targeted sensitization strategies.

 

 

Research Methods and Core Experiments

The authors integrated longitudinal plasma samples from two clinical trials, CT041-CG4006 (CLDN18.2-CAR-T) and GLOW (Zolbetuximab), and utilized an antibody sandwich EV expression array (EV-array) to quantitatively detect dynamic changes in 50 key EV proteins. A composite scoring system (CPP-score) incorporating EV-CLDN18.2, EV-PD-1, and EV-PD-L2 was constructed, and its ability to predict progression-free survival (PFS) and overall survival (OS) was evaluated in both discovery and validation cohorts.

In terms of mechanistic validation, the research team generated HEK293F cells overexpressing CLDN18.2 to isolate EVs. In vitro co-culture experiments confirmed that EV-CLDN18.2 could be internalized by CAR-T cells, inducing an exhaustion phenotype. Single-cell transcriptome sequencing (scRNA-seq) was employed to analyze subpopulation changes and trajectory evolution of CAR-T cells following EV treatment. Furthermore, using TurboID proximity labeling combined with mass spectrometry, the study identified an interaction between CLDN18.2 and the cancer-associated fibroblast (CAF) surface protein S100A4. This was validated in PDX mouse models, demonstrating that EV-CLDN18.2 inhibits CAR-T efficacy in vivo by activating a CAF subpopulation with high IL-6 expression.

Key Conclusions and Perspectives

• High expression of EV-CLDN18.2 and EV-PD-1, coupled with low expression of EV-PD-L2, is significantly associated with poor prognosis in CLDN18.2-targeted therapies. The integrated CPP-score demonstrates superior predictive power compared to single markers or histological detection, serving as a liquid biopsy biomarker for clinical monitoring.
• EV-CLDN18.2 directly binds to surface receptors on CAR-T cells, inducing the upregulation of exhaustion markers such as PD-1 and LAG-3. This leads to impaired CAR-T cell proliferation and loss of cytotoxic function, revealing a novel mechanism of extracellular vesicle-mediated immune exhaustion within the tumor microenvironment.
• EV-CLDN18.2 specifically activates an IL-6-secreting subpopulation of CAFs by binding to its ligand S100A4, thereby remodeling the tumor microenvironment and suppressing CAR-T cell function. This provides a theoretical basis for the combined use of CAF inhibitors.
• EV-CLDN18.2 can specifically neutralize targeting antibodies such as Zolbetuximab, reducing their anti-tumor activity. Conversely, EV-PD-L2 enhances anti-tumor immunity by modulating the Th1/Th2 balance, indicating that different EV proteins exert opposing regulatory effects on therapeutic outcomes.
• Clearing EV-CLDN18.2 from plasma via plasma exchange (PE) or intraperitoneal irrigation (IPI), or combining treatment with PD-1 inhibitors (with caution regarding AICD risk) and CAF inhibitors, can significantly reverse EV-mediated resistance and enhance the in vivo efficacy of CAR-T and antibody therapies.

Research Significance and Prospects

This discovery offers new combination therapy strategies for drug development, particularly for patients with high CPP-scores, suggesting approaches such as plasma EV clearance or combined CAF targeting to overcome resistance. In terms of clinical monitoring, the CPP-score serves as an effective tool for dynamically assessing CAR-T cell expansion and survival in vivo, guiding adjustments in treatment timing. Additionally, the disease modeling framework established in this study (e.g., using PDX models to simulate EV-mediated immune suppression) provides a standardized evaluation system for screening novel immune-sensitizing drugs, having profound implications for optimizing immunotherapy pathways for gastrointestinal cancers.

 

 

Conclusion

This study provides an in-depth analysis of the complex role of extracellular vesicles in CLDN18.2-targeted therapies. It not only establishes the plasma EV protein-based CPP-score as a novel biomarker for predicting treatment response in gastrointestinal cancer patients but also reveals the dual resistance pathways of EV-CLDN18.2-induced CAR-T exhaustion and CAF activation at the molecular level. This finding extends traditional tissue biomarker detection into the dynamic realm of liquid biopsy, offering a new dimension for the precise clinical assessment of therapeutic efficacy. From the perspective of translational research, developing sensitization strategies to clear plasma EVs or block S100A4-CLDN18.2 interactions to counteract the EV-mediated immunosuppressive microenvironment holds promise for significantly improving the clinical benefit rates of CAR-T and antibody drugs. This lays a new foundation for the care systems of refractory diseases such as gastric cancer and drives the optimization and implementation of personalized immunotherapy protocols.

 

Changsong Qi, Mingyang Ma, Kaijie Liang, Lin Shen, and Cheng Zhang. Extracellular vesicle proteins monitor and interfere with CLDN18.2-targeted CAR-T cell and antibody therapies against gastrointestinal cancers. Nature Communications.