In vivo Reprogramming of T Cells with LNP-Encoded CLDN18.2 CAR mRNA for Solid Tumor Eradication

In vivo Reprogramming of T Cells with LNP-Encoded CLDN18.2 CAR mRNA for Solid Tumor Eradication | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article In vivo Reprogramming of T Cells with LNP-Encoded CLDN18.2 CAR mRNA for Solid Tumor Eradication XIAOQI HU This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7296050/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study presents a novel conceptual framework for CLDN18.2-targeted CAR-T cell therapy. The single-chain variable fragment (scFv) structure was predicted using AlphaFold2, followed by four rounds of docking simulations with HADDOCK, ultimately identifying a conformation with favorable affinity and structural stability. To enable in vivo CAR expression, three mRNA delivery systems—lipid nanoparticles (LNPs), the cationic polymer TMAB3, and peptide-coated RNACap capsules—were modeled and compared in terms of kinetic profiles, tissue tropism, and safety considerations. To improve tumor specificity and reduce systemic toxicity, transcriptional regulation elements such as NR4A2 and RGS16 promoters were conceptually introduced to confine CAR expression to the tumor microenvironment. Additionally, a chemokine-guided homing model for CD8⁺ T cells was simulated to mimic immune navigation and targeted therapeutic engagement. Collectively, this work proposes a closed-loop CAR-T system integrating in vivo mRNA delivery, tumor-specific transcriptional activation, antigen targeting, and immune cell homing. This integrative framework offers a potentially translatable strategy for improving solid tumor immunotherapy. This study is entirely based on computational modeling and publicly available biological data, without involving any human or animal subjects. Bioinformatics Immunology Molecular Biology Oncology CLDN18.2 CAR-T mRNA delivery Lipid manoparticles tumor-specific promoter NR4A2 RGS16 synthetic biology immune homing computational modeling Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Introduction Chimeric antigen receptor (CAR) T-cell therapy has demonstrated transformative success in hematologic malignancies, inspiring growing efforts to adapt this strategy for solid tumors. However, critical obstacles remain, including antigen heterogeneity, limited T cell infiltration, off-target toxicity, and unstable CAR expression. Claudin 18.2 (CLDN18.2), a tight junction protein selectively expressed in gastric and other epithelial-derived tumors, has emerged as a promising immunotherapeutic target due to its tumor-specific expression profile and cross-indication potential. Despite ongoing clinical trials involving CLDN18.2-directed CAR-T cells, most current approaches rely on heterogeneous CAR architectures, viral vector transduction, and non-standardized manufacturing pipelines—challenges that complicate scalability and hinder clinical translation. To address these barriers, an alternative paradigm— in vivo CAR-T cell therapy—has gained traction. This approach delivers mRNA-encoded CAR constructs directly into the patient using non-viral systems, bypassing the need for ex vivo T cell manipulation. The resulting benefits include programmable CAR expression, faster deployment, and potentially lower production costs. However, successful implementation still depends on overcoming several key technical bottlenecks: 1.Optimization of CAR structure for high-affinity binding and conformational integrity; 2.Development of safe, efficient, and targeted mRNA delivery platforms; 3.Restriction of CAR expression to the tumor microenvironment to reduce systemic toxicity; 4.Enhancement of T cell homing and immune activation within the tumor niche. To conceptually address these requirements, we propose an integrative design framework for a fully in vivo CAR-T platform, with structural precision, modular delivery, and spatially restricted expression. The core components include: In silico modeling and docking of a CLDN18.2-specific single-chain variable fragment (scFv) using AlphaFold2 and HADDOCK; The docking conformation is illustrated in Figure 1 , highlighting the molecular interface between the scFv (VH-linker-VL) and full-length CLDN18.2. Figure 1. Structure of the CLDN18.2–scFv docking complex (Docking1, Cluster 7). This model depicts the molecular interaction between full-length transmembrane CLDN18.2 and the scFv domain (VH–linker–VL) derived from the 14G11 antibody. The selected model demonstrates the highest binding affinity with favorable surface contact and electrostatic properties, as predicted by HADDOCK scoring metrics. Comparative simulation of three mRNA delivery systems—lipid nanoparticles (LNPs), cationic polymer TMAB3, and peptide-coated RNACap capsules—focusing on delivery kinetics and tissue distribution; Conceptual incorporation of tumor-selective transcriptional promoters (NR4A2 and RGS16) to confine CAR expression to the tumor microenvironment; Computational modeling of CD8⁺ T cell homing guided by the CXCL12–CXCR4 chemokine axis; Integration of these components into a proposed closed-loop therapeutic architecture encompassing design, delivery, activation, and immune coordination. This study outlines a comprehensive modeling-based workflow—spanning antigen recognition, delivery engineering, and immune navigation—that may serve as a strategic roadmap for advancing CAR-T therapies in solid tumors. Figure 2. Schematic overview of in vivo CAR-T therapy targeting CLDN18.2. LNP-delivered mRNA encoding a CLDN18.2-directed CAR is taken up by CD8⁺ T cells. Tumor-specific promoters (NR4A2 and RGS16) restrict CAR expression to the tumor microenvironment, where CAR-T cells recognize CLDN18.2-positive tumor cells and initiate antigen-specific cytotoxicity. Chemokine-guided homing mechanisms support immune cell migration and spatial coordination of therapy. Methods In this study, a conceptual chimeric antigen receptor (CAR) construct was designed using a modular architecture comprising five canonical domains: a single-chain variable fragment (scFv), hinge region, transmembrane (TM) domain, intracellular costimulatory domain (such as CD28 or 4-1BB), and the CD3ζ signaling domain. The scFv was modeled based on the publicly reported 14G11 monoclonal antibody, incorporating its variable heavy (VH, SEQ ID NO:13) and variable light (VL, SEQ ID NO:14) chains. These were connected via a flexible (G₄S)₃ linker in a VH–linker–VL orientation to form the antigen-recognition domain. The remaining CAR modules were selected from publicly accessible reference sequences curated from previous CAR-T designs. These components were computationally assembled into a full-length CAR amino acid sequence, which was then formatted in FASTA for use in subsequent structural modeling and protein docking simulations. 3.2 AlphaFold2-Based Structural Prediction Workflow We employed AlphaFold2 Multimer mode to computationally predict the three-dimensional structures of key molecular components involved in the CAR-T construct. The modeling workflow included the following configurations: Isolated scFv fragment (VH–linker–VL), derived from the 14G11 monoclonal antibody; Full-length CAR protein , encompassing the scFv, hinge region, transmembrane domain, costimulatory module (e.g., 4-1BB or CD28), and CD3ζ signaling domain; 2 target protein , based on the UniProt entry Q8N6F1-2 , with its predicted structure obtained from the AlphaFold Protein Structure Database . All structural predictions were performed using the model_2_multimer_v3 protocol in a Google Colab–hosted AlphaFold2 environment . For each construct, five independent modeling runs were executed. The structure with the highest pLDDT confidence score (typically >85) and favorable domain packing was selected for subsequent molecular docking. Figure 4. Predicted structure of the VH-linker-VL scFv fragment using AlphaFold2 Multimer. Figure 5.Predicted structure of the full-length CAR protein (scFv–Hinge–TM–Costimulatory domain–CD3ζ) using AlphaFold2 Multimer. Figure 6. Predicted structure of CLDN18.2 protein (UniProt Q8N6F1-2) generated via AlphaFold2. 3.3 Four-Round HADDOCK Molecular Docking and Residue-Level Analysis To evaluate the structural binding interactions between CLDN18.2 and the engineered CAR constructs , we performed four iterative rounds of molecular docking using the HADDOCK 2.4 web server (Guru mode) . Each docking round applied a distinct configuration , including: Customized definitions of active and passive residues on the antigen-binding interface; Varied pairing geometries and orientations between scFv and CLDN18.2; Progressive refinement strategies to assess conformational flexibility , interface complementarity , and binding affinity . These simulations aimed to systematically explore alternative binding conformations, quantify residue-level interaction metrics , and identify the most stable and biologically plausible docking pose based on HADDOCK scoring parameters such as Z-score, buried surface area , and electrostatic energy . First Docking – Baseline Model (Figure 1) Input: scFv structure (AlphaFold2) + full-length CLDN18.2 (AlphaFold2) HADDOCK Score: -101.1 Buried Surface Area (BSA): 2524.1 Ų Z-score: -2.4 Remarks: This model exhibited the largest interface area among all runs and served as the primary reference for subsequent structural comparisons. Second Docking – Full-Length CAR Interaction (Figure 2) Input: Complete CAR structure (scFv–Hinge–TM–Costimulatory–CD3ζ) + full-length CLDN18.2 HADDOCK Score: -108.8 Electrostatic Energy: -135.3 Remarks: This docking demonstrated the strongest electrostatic interactions , reflecting multi-domain engagement of the CAR with the target membrane protein. Third Docking – Refined Re-docking (Figure 3) Input: scFv fragment (extracted from Docking 2) + partial CLDN18.2 segment Electrostatic Energy: -215.6 Z-score: -0.7 Remarks: Although this run yielded high binding energy , it exhibited residue violation and a low Z-score , limiting its utility as a structural reference. Fourth Docking – Directed Epitope Binding to ECL2 (Figure 4) Input: scFv (AlphaFold2) + CLDN18.2 ECL2 loop (residues defined) Restraint Violations: 11.7 (lowest) RMSD: Lowest across all docking runs Remarks: This model captured localized high-affinity binding at the ECL2 domain, with minimal violations and superior structural compactness , ideal for minimal expression constructs or bispecific targeting strategies . All docking outputs were evaluated and ranked based on HADDOCK score , Z-score , electrostatic energy (ΔE) , buried surface area (BSA) , and restraint violations . The top-ranking cluster centroid from each round was selected as the representative model for visualization and downstream analysis. Figure 7. Comparative Docking Stages of CAR and CLDN18.2. 3.4 Modeling of Three RNA Delivery Mechanisms: LNP, TMAB3, and RNACap To facilitate efficient in vivo expression of CLDN18.2-targeted CAR constructs and overcome the delivery challenges associated with solid tumors, we modeled and compared three representative RNA delivery strategies: Lipid Nanoparticles (LNPs), TMAB3-based cationic polymer complexes, and the RNACap oral capsule system. These platforms exhibit distinct yet complementary characteristics in terms of structure, cellular entry, pharmacokinetics, and therapeutic application. (1) LNP-Mediated Delivery Lipid nanoparticles (LNPs) are the most clinically validated platform for mRNA delivery, extensively utilized in vaccine development (e.g., Pfizer-BioNTech, Moderna) and emerging cancer therapies. LNPs typically measure 80–120 nm and encapsulate mRNA using ionizable lipids, cholesterol, and PEG-lipids. Upon systemic administration, LNPs enter cells via clathrin-mediated endocytosis, followed by endosomal escape that releases mRNA into the cytoplasm for protein translation. Advantages: High translational potential and regulatory familiarity Limitations: Partial mRNA degradation in lysosomes; variable transfection efficiency across tissues (2) TMAB3-Mediated Delivery TMAB3 is a next-generation cationic polymer derived from engineered antibody scaffolds. It forms stable nano-complexes with mRNA through electrostatic interactions and bypasses endocytosis, enabling direct cytosolic delivery. TMAB3 exploits overexpression of ENT2 nucleoside transporters on tumor cells to achieve selective uptake. In preclinical models of pancreatic cancer and melanoma, TMAB3 enhanced tumor-selective mRNA delivery by up to 1,500-fold compared to normal tissues, with minimal toxicity. Significance: Represents a major advance in systemic RNA delivery for solid tumors. (3) RNACap Oral Capsule Delivery Developed in 2024 by Harvard Medical School, RNACap is the first oral mRNA delivery platform utilizing pH-responsive enteric capsules. The outer shell remains stable in gastric acid (pH 6.5), releasing ~176 nm mRNA-loaded nanoparticles capable of epithelial penetration and systemic transfection. In animal models (rat and pig), RNACap achieved intestinal absorption of IL-10 mRNA, resulting in sustained serum expression with no systemic toxicity. Significance: Offers a non-invasive route for gut-targeted or maintenance mRNA therapies. Comparative Summary Feature LNP TMAB3 RNACap Onset of Action Medium (6–12 h) Fast (8 h) Delivery Efficiency Moderate High (Tumor-selective) Regional (Intestinal) Application Vaccines, injectables Systemic solid tumor therapies Oral/gut-targeted maintenance Safety Profile Good (some immunogenicity) Low immunogenicity Excellent (validated in animals) Figure 8 . Comparative Modeling of RNA Delivery Mechanisms: LNP, TMAB3, and RNACap 3.5 Simulated Pharmacodynamic Timeline and T Cell Navigation Modeling To further compare the kinetic profiles of different RNA delivery strategies in vivo, we constructed a time-course simulation model capturing the entire process from mRNA administration to CAR protein expression and subsequent T cell-mediated immune activation. This framework is based on literature-derived estimates and mechanistic assumptions. In addition, we modeled the navigation behavior of CAR-T cells, illustrating their homing trajectory toward solid tumor regions. Simulated Timeline Assumptions Time (t) Biological Process Description t = 0 h Initiation of mRNA administration (injection or oral delivery) t = 3–6 h Major delivery completion (via LNP and TMAB3 systems; RNACap completes later) t = 6–12 h Onset of CAR mRNA translation; detectable surface expression begins t = 24–48 h Peak CAR expression; T cells initiate tumor antigen recognition t ≥ 48 h CD8⁺ T cells infiltrate the tumor microenvironment and commence cytotoxic activity Comparative Highlights TMAB3 enables rapid cytosolic delivery, initiating CAR expression within 6 hours—offering the fastest immune activation profile. LNP-based systems typically achieve functional expression between 12–24 hours post-injection. RNACap results in delayed, gut-specific expression, typically emerging 8–12 hours post-ingestion. Figure9 : Kinetic Curves of CAR Expression and Immune Activation Under Three RNA Delivery Strategies (LNP, TMAB3, RNACap) Description : Schematic graph showing time (x-axis) versus CAR protein expression and T cell activation markers (y-axis), with overlaid curves for LNP, TMAB3, and RNACap pathways. T Cell Homing and Navigation Mechanism Following CAR expression, engineered CD8⁺ T cells must home to tumor sites with high CLDN18.2 expression. Based on current immunological understanding, we modeled the chemokine-guided migration process involving the CXCL12–CXCR4 axis. This includes: CXCL12 secretion by tumor-associated stroma, forming a local chemotactic gradient CXCR4 receptor expression on activated CAR-T cells Chemotactic navigation through vascular endothelium into tumor parenchyma Targeted infiltration into the tumor core, enabling precise immune engagement This mechanism underscores the importance of chemokine–receptor coupling in directing CAR-T cells toward solid tumors with spatial specificity. Figure 10 : Schematic of CAR-T Cell Navigation via CXCL12–CXCR4 Chemotaxis Description : Diagram depicting CXCL12 secretion by tumor → gradient formation → CXCR4⁺ CAR-T cells sensing gradient → transendothelial migration → tumor infiltration. 3.6 Multi-Pathway Synergistic Delivery and Tumor-Specific Activation Mechanism Modeling To enhance the therapeutic precision and safety of in vivo CLDN18.2-targeted CAR mRNA expression, we designed an integrated multi-pathway delivery and activation framework. This model synergistically combines three cutting-edge RNA delivery modalities— LNP (lipid nanoparticles), TMAB3 (cationic polymer complexes), and RNACap (pH-sensitive oral capsules)—with tumor-specific promoter systems (NR4A2 and RGS16), ensuring spatially restricted CAR expression within the tumor microenvironment (TME). Overview of the Three Delivery Modalities Delivery Route Core Vehicle Activation Mechanism Tissue Specificity Translational Readiness LNP Lipid nanoparticles Endocytosis → Endosomal escape Moderate Clinically approved (e.g., Moderna vaccines) TMAB3 Cationic polymer–antibody complex Electrostatic condensation → Direct membrane translocation High (ENT2-dependent tumor selectivity) Preclinical validation in multiple animal models RNACap Engineered pH-responsive oral capsule Intestinal release → Macropinocytosis Strong (gut-restricted) Demonstrated efficacy in rat and pig studies Tumor-Specific Promoter System To minimize off-target expression and associated toxicity, we employed tumor-microenvironment-restricted promoters, as reported in Nature . These promoters drive the expression of therapeutic payloads exclusively within tumor tissues: NR4A2 promoter : Activated specifically in tumors; induces IL-12 secretion to amplify local immune responses. RGS16 promoter : Tumor-restricted activation; drives IL-2 expression to support CAR-T proliferation and persistence. Notably, these promoters exhibit <20% transcriptional activity in peripheral tissues compared to conventional promoters, significantly reducing systemic exposure while maintaining potent intratumoral effects. Final Integrated Delivery–Activation Pathway We constructed a unified model integrating all three delivery pathways, each converging on a tumor-specific CAR expression module regulated by embedded promoter logic: Pathway 1 (LNP, IV injection) : LNP-encapsulated CAR mRNA → systemic circulation → uptake and translation in tumor tissue Pathway 2 (TMAB3, IV injection) : TMAB3/mRNA complex → ENT2-mediated uptake → direct cytosolic delivery into tumor cells Pathway 3 (RNACap, oral administration) : RNACap-protected mRNA → intestinal release → localized expression in gut epithelial cells (ideal for colorectal or intestinal malignancies) Common Feature : All mRNAs include NR4A2 or RGS16 regulatory sequences, ensuring activation only in tumor-infiltrating immune cells. This unified system enables precise, intratumoral expression of CAR, IL-2 , or IL-12 , reinforcing both specificity and safety. Figure11 : Integrated Model of Multi-Route Delivery and Tumor-Specific Promoter-Driven CAR Expression Caption : Schematic representation of three parallel delivery inputs (LNP, TMAB3, RNACap) feeding into a unified tumor-targeted expression module. The diagram illustrates vehicle entry, controlled mRNA release, and downstream activation of IL-2, IL-12, and CAR expression within CLDN18.2⁺ tumor cells via NR4A2/RGS16 promoters. Tumor selectivity is highlighted by the absence of activation in peripheral tissues. All analyses in this study are purely computational, based on publicly available biological data and modeling tools. No human participants, animal subjects, or clinical samples were involved, and therefore no ethical approval was required. 4.1 Comparative Analysis of Four Distinct Docking Structures To evaluate the stability and binding efficiency of various CLDN18.2–CAR complex conformations, we conducted four rounds of molecular docking using the HADDOCK 2.4 platform. Structural models were assessed across five core metrics: HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violation energy (used as a proxy for RMSD). Each docking round represented a distinct structural strategy, tailored for different potential application scenarios. Docking Overview and Comparative Metrics Docking Round Structure Type Used HADDOCK Score BSA (Ų) Z-score Electrostatics Restraint Violation Key Characteristics #1 scFv + full-length CLDN18.2 -101.1 2524.1 -2.4 -117.3 145.3 Largest interface; selected as primary model #2 Full-length CAR + CLDN18.2 (Easy mode) -108.8 2409.3 -2.3 -135.3 182.6 Strongest electrostatics; suitable for affinity-focused studies #3 Refined scFv + CLDN18.2 segment -77.4 2096.4 -0.7 -215.6 227.7 High electrostatics but low stability; not recommended #4 scFv + ECL2 loop of CLDN18.2 (guided restraints) -75.9 1343.3 -1.5 -70.1 11.7 Most stable and compact; ideal for low-error expression systems Structural Model Recommendations Primary Model for Structural Display Docking #1 (Cluster 7) is recommended as the main reference model due to its highest BSA , optimal Z-score , and balanced electrostatics . It reflects the most physiologically relevant and stable configuration among the four. Secondary Model for Affinity Illustration Docking #2 provides the strongest electrostatic interaction and is ideal for visualizing affinity-dominant docking behavior , especially in full-length CAR–CLDN18.2 scenarios. Model for Expression-Limited Systems Docking #4 , with the lowest restraint violation and minimal energy deviation, is best suited for in vitro translation , expression-limited environments, or synthetic compact delivery systems. Model to Avoid Docking #3 exhibits excessively high electrostatics but poor Z-score and high restraint violations, indicating structural inconsistency. This model is not recommended for further study or visualization. Figures for Inclusion Figure 7 : 3D renderings of the four docking complexes: 1 : Docking #1 (scFv–full-length CLDN18.2) 2 : Docking #2 (full CAR–CLDN18.2) 3 : Docking #3 (refined scFv–CLDN18.2 segment) 4 : Docking #4 (ECL2-guided docking with restraints) Figure 12 Comparative radar plot of docking rounds across five structural evaluation metrics:HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violations. 4.2 Comparative Evaluation of Three Delivery Mechanisms: LNP, TMAB3, and RNACap To systematically evaluate the pharmacokinetic profiles and translational potential of the three mRNA delivery strategies outlined in Section 3.4, we conducted a structured comparison across five critical parameters: onset time , delivery efficiency , tissue specificity , biosafety , and clinical readiness . This analysis enables identification of the optimal platform for specific therapeutic contexts. Summary Comparison Table Parameter LNP TMAB3 RNACap (Oral) Onset Time Moderate (6–12 h) Rapid (8 h) Delivery Efficiency Moderate Very high (↑1500-fold tumor enrichment in preclinical models) Regional (localized accumulation in intestinal mucosa) Tissue Specificity Moderate High (ENT2 transporter–dependent targeting) Very high (gut-restricted expression) Safety Above average High (electrostatic shielding + non-immunogenic Fc fusion) Very high (non-invasive; validated in multiple animal models) Clinical Readiness High (approved LNP-based mRNA vaccines exist) Moderate (emerging but promising platform) Early-stage (highly accessible and scalable if optimized) Key Observations LNP (Lipid Nanoparticles) represent a clinically validated and balanced delivery platform, offering acceptable safety and moderate tissue targeting, though with a delayed onset compared to novel systems. Recent research advances have further enhanced the utility of LNP-based delivery. A team from Peking University has developed a non-inflammatory LNP formulation that incorporates V-ATPase activation and ESCRT-mediated membrane repair to improve endosomal escape and cytosolic release. This strategy significantly improves mRNA bioavailability while minimizing innate immune activation, representing a meaningful upgrade over conventional LNP systems. Such innovations strengthen the translational relevance of LNP for CAR-T applications by addressing prior limitations in delivery kinetics and tumor-specific uptake. TMAB3 , a next-generation electrostatic polymer–antibody complex, demonstrates ultra-high tumor targeting and fast onset , particularly when leveraged with ENT2-mediated uptake. However, it remains at the preclinical development stage . RNACap , an orally administered pH-responsive capsule , enables non-invasive, regionally localized delivery to intestinal tissues. It is especially suitable for colorectal or gut-specific tumors but still requires extensive clinical maturation . 4.3 Simulation-Based Validation and Visualization of Structural Models To enhance the scientific rigor, structural interpretability, and visual clarity of this study, all predicted complexes and delivery mechanisms were systematically visualized using PyMOL . This high-resolution molecular rendering enabled the production of publication-ready figures and facilitated intuitive comparison across models and pathways. The following structural and mechanistic components were individually rendered and annotated: scFv structural conformation , highlighting its interaction interface with the CLDN18.2 antigen ; All four molecular docking results , with top-ranking clusters selected for comparative analysis based on HADDOCK metrics; The complete full-length CAR construct , including hinge, transmembrane (TM), costimulatory, and CD3ζ signaling domains; Mechanistic diagrams of the three modeled mRNA delivery routes : LNP , TMAB3 , and RNACap ; Simulated pathways of CD8⁺ T cell homing and activation , mediated by the CXCL12–CXCR4 chemokine axis; An integrated regulatory circuit visualizing tumor-restricted expression driven by NR4A2 and RGS16 promoters across multiple delivery modalities. This study involves only in silico modeling and literature-based analysis. No human or animal subjects, biological samples, or clinical data were used, and therefore ethical approval was not required. 5.1 Structural Innovation of CLDN18.2-Targeted CAR-T Design This study presents a de novo modular construction of a chimeric antigen receptor (CAR) specifically engineered to target CLDN18.2, with a dual emphasis on structural specificity and translational efficiency . For the antigen-recognition domain, a single-chain variable fragment (scFv) was selected, derived from the 14G11 monoclonal antibody and composed of VH-linker-VL segments. Its three-dimensional structure was predicted using AlphaFold2-Multimer , allowing for high-resolution conformational modeling and rational integration into the CAR scaffold. The scFv was fused to a flexible IgG4 hinge region to enhance spatial adaptability while minimizing off-target activation. The transmembrane (TM) domain , sourced from CD8α, was chosen for its proven ability to ensure stable membrane anchoring and strong biophysical integrity. The intracellular signaling module consisted of a canonical CD28 costimulatory domain followed by CD3ζ , a configuration validated across multiple FDA-approved CAR-T therapies for its potency in driving T cell activation and persistence. Figure 13. Modular Architecture of the CLDN18.2-Targeted CAR Construct Schematic representation of the full CAR structure including scFv (from 14G11), IgG4 hinge, CD8α transmembrane domain, CD28 costimulatory domain, and CD3ζ intracellular domain. Total size (~2.2 kb) is optimized for RNA encapsulation. Key innovations in this CAR design include: Precision-guided domain fusion : All junctions between structural components were modeled based on native interface residues. AlphaFold2-based validation confirmed the absence of strain or misfolding at fusion points. Molecular size optimization : The full-length CAR construct was maintained within ~2.2 kb, ensuring compatibility with mRNA encapsulation vehicles (such as LNP or TMAB3) and maximizing in vivo translational efficiency . Cross-platform compatibility : This CAR structure supports both viral vector–mediated stable expression and transient mRNA-based delivery , enabling versatile application across diverse therapeutic delivery systems. In HADDOCK-based molecular docking simulations , the scFv showed robust binding affinity and high structural stability when docked with both full-length CLDN18.2 and its ECL2 domain . The top-ranked clusters exhibited strong electrostatic interactions , low restraint violations (RMSD proxy) , and high buried surface areas (BSA) —all indicative of specific and stable epitope engagement . Collectively, this CAR construct exemplifies structural elegance , functional flexibility , and clinical delivery readiness . It establishes a foundational framework for mRNA-based CAR-T therapies targeting CLDN18.2 and offers a broadly adaptable blueprint for next-generation solid tumor CAR designs . 5.2 Scientific Value and Limitations of Multi-Round Docking Simulations To assess the conformational affinity and binding stability between the engineered CAR molecule and its target CLDN18.2, we conducted four systematic rounds of HADDOCK-based molecular docking. These simulations spanned from unconstrained full-structure docking to epitope-specific, residue-guided interactions. Each docking round was accompanied by PyMOL visualization and quantitative scoring to enable comparative structural analysis. Figure 7. Structural comparison of the four CAR–CLDN18.2 docking models (Rounds 1–4). Each structure was rendered using PyMOL and shows distinct docking conformations generated by HADDOCK The scientific contributions of the four simulation rounds are as follows: From coarse to fine-grained structural exploration : The first round employed an AlphaFold-assembled scFv docked freely to the full-length CLDN18.2, offering a global assessment of conformational compatibility. In contrast, the fourth round targeted the ECL2 loop of CLDN18.2 with predefined interface residues, demonstrating superior specificity and conformational precision at the epitope level. Comprehensive evaluation metrics : Scoring indicators included HADDOCK score, Z-score, buried surface area (BSA), electrostatic energy, van der Waals energy, restraint violation values, and both mean and fluctuation range of RMSD—ensuring a multi-dimensional and robust evaluation framework Multi-objective structural selection : Each docking round offered distinct structural advantages: Round 1 exhibited the largest contact surface area and was used as the principal visual model; Round 2 demonstrated the strongest electrostatic interaction; Round 4 had the lowest RMSD and minimal restraint violations, suggesting high structural stability ideal for low-error expression systems. Round 3, while exhibiting favorable electrostatic energy, was deprioritized due to its poor Z-score and higher restraint violations compared to other rounds. Figure 12. Comparative radar plot of docking rounds across five structural evaluation metrics:HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violations. The plot highlights complementary strengths among the docking strategies and informs structure selection for downstream therapeutic modeling. However, the study also acknowledges several limitations in the current docking strategy: Heavy reliance on AlphaFold predictions : All protein structures used were derived from AlphaFold2 modeling. While highly reliable, these models lack experimental crystallographic validation, introducing potential uncertainties in conformational fidelity. Idealized docking environment : The HADDOCK simulations were conducted in simplified vacuum-like settings, which do not account for critical microenvironmental variables such as pH, osmolarity, ion concentrations, or viscosity commonly found in the tumor milieu. Absence of molecular dynamics validation : The docking results have not yet undergone nanosecond-to-microsecond scale molecular dynamics (MD) simulations using tools like GROMACS, which are essential for assessing the real-time post-docking stability of the complexes. Despite these constraints, the four-step docking strategy presented in this study provides a robust and interpretable structure-based assessment under current computational and bioinformatics capabilities. It forms a solid foundation for CAR design refinement, mRNA encapsulation strategies, and tumor-targeting pathway optimization. 5.3 Impact of Multi-Route Delivery Mechanisms on Therapeutic Efficacy (LNP vs TMAB3 vs RNACap) In vivo delivery remains one of the core engineering challenges for CAR-mRNA therapies. Compared to traditional ex vivo CAR-T infusion strategies, in vivo approaches must overcome greater complexity—achieving high transfection efficiency, tissue-specific expression, safety, and rapid therapeutic onset. This study comparatively models three delivery platforms: lipid nanoparticles (LNPs), the cationic polymer TMAB3, and the gut-targeted nanocarrier RNACap. Figure 9. Schematic illustration of three in vivo CAR-mRNA delivery routes: LNP, TMAB3, and RNACap. Each modality features distinct targeting logic: - LNPs: systemic but liver-biased accumulation - TMAB3: charge-guided delivery to T cells - RNACap: oral administration with gut-localized uptake LNP: Clinically established, but with limited specificity LNP technology has been validated in mRNA vaccines (e.g., BNT162b2, mRNA-1273), offering systemic delivery and favorable biocompatibility. Modeled features : Onset time: moderate (6–12 hours) Accumulates primarily in the liver, with notable off-target leakage Moderate tumor site expression when combined with CD8⁺ T cell homing However, LNPs exhibit moderate targeting specificity . Chemokine guidance (e.g., CXCL12–CXCR4) or lipid modification is often required to enhance tumor penetration. Their adaptability to solid tumor environments (dense ECM, immunosuppressive niches) remains a limitation. 1. TMAB3: Charge-guided T cell delivery TMAB3 is a trimethylammonium-modified polymer with high positive charge density. It electrostatically interacts with negatively charged T cell surface motifs (e.g., CD3ε), enabling precise delivery. Key modeled advantages : Rapid onset (<6 hours) with strong intracellular transcription peaks 1500-fold tumor targeting improvement in mice PEG-free encapsulation with good biodegradability TMAB3’s charge-selectivity allows targeted delivery without systemic homing. Its compact core–shell structure shows enhanced physical stability and lower toxicity than traditional cationic liposomes. 3.RNACap: Oral localized delivery potential RNACap is a gut-targeted oral nanocarrier that delivers mRNA via intestinal epithelial absorption to local immune hubs (e.g., Peyer’s patches, lamina propria). Modeled benefits : Localized expression in GI tissues Slower onset (>8 hours) Preclinical safety verified in rats and pigs Ideal for gastric or colorectal tumor targeting Although still preclinical, RNACap offers high compliance and scalability. When combined with tumor-specific promoters (e.g., NR4A2, RGS16), it may become a lead strategy for GI cancers. Comparative Summary and Recommendations Feature LNP TMAB3 RNACap (Oral) Onset Time Moderate (6–12 h) Fast (8 h) Targeting Specificity Moderate (liver-biased) High (T cell membrane) Very High (GI-localized) Encapsulation Stability Moderate Very High Moderate Safety Profile Upper Moderate High Very High Clinical Maturity Advanced (approved) Intermediate (emerging) Early-stage (exploratory) Use Case Pan-cancer systemic Solid tumor T-cell targeting Regional GI tumor therapy This comparative framework highlights complementary advantages: LNP offers mature, systemic delivery. TMAB3 excels at immune cell precision delivery. RNACap enables localized, patient-friendly therapy. Selection should align with tumor location, promoter control, and translational goals. Figure 14. Comparative radar plot of three RNA delivery strategies (LNP, TMAB3, RNACap), evaluated across six dimensions: onset time, targeting specificity, encapsulation stability, safety profile, clinical maturity, and expression intensity. The plot highlights the trade-offs and performance profiles relevant to CAR mRNA in vivo delivery applications. 5.4 Promoter-Controlled and Structure-Enhanced Synergistic Mechanisms: Achieving Precision and Durability in CAR Expression In in vivo mRNA-delivered CAR-T therapy, precise spatiotemporal control of CAR expression and sustained functional activity are critical determinants of therapeutic success. Following the comparative delivery kinetics analysis in Section 5.3, this section introduces a dual-mechanism synergy: tumor-specific promoters (NR4A2 / RGS16) to ensure localized CAR expression, and structural augmentation via the CD2 module to enhance immunological synapse stability and mitigate T cell exhaustion. Together, these components aim to establish a durable, potent, and low-toxicity CAR-T system optimized for solid tumors . 1) NR4A2 / RGS16 Promoters: Tumor-Restricted CAR Expression Conventional CAR-T therapies often trigger off-tumor toxicity due to systemic expression, especially when low-level antigens are present in healthy tissues. This study adopts a tumor microenvironment-specific expression strategy using the NR4A2 and RGS16 promoters , as reported by the Peter MacCallum Cancer Centre in Nature . These promoters are activated only within the tumor microenvironment, significantly reducing systemic side effects. NR4A2 and RGS16 respectively drive the expression of IL-12 and IL-2, with transcriptional activity dependent on local stress signals and metabolic pressure, remaining largely silent in healthy tissues. These promoters are highly compatible with in vivo mRNA delivery and can be embedded into the 5′ untranslated regions (5′-UTRs) of CAR-mRNA constructs, enabling "tumor-only" CAR expression . Compared to conventional ubiquitous promoters such as CMV or EF1α, NR4A2 and RGS16 act as biological safety valves , precisely restricting expression to diseased sites. 2) CD2 Structural Enhancement Module: Reinforcing Synapse Stability and Reducing Exhaustion A recent study published in Cellular & Molecular Immunology reveals that integrating a CD2 co-stimulatory domain into the CAR structure significantly enhances therapeutic performance through two synergistic pathways: CD2 strengthens F-actin polarization and synaptic stability , enhancing the binding interface between CAR-T cells and tumor cells; CD2 also suppresses the expression of T cell exhaustion markers (e.g., PD-1, TIM-3) and transcription factors (e.g., NR4A family) under chronic antigen exposure, preserving cytotoxic functionality over time. Importantly, the exhaustion-reducing effects of CD2 intersect with the NR4A regulatory pathway, forming a feedback loop that jointly modulates expression control and structural stability, thereby delaying CAR-T cell functional decline. 3) Schematic Overview and Integrated Model Proposal To better illustrate the coupling between tumor-specific promoter activation and structural enhancement, we recommend including a mechanistic diagram (Figure 8) depicting: CD8⁺ T cells entering the tumor via LNP or TMAB3 delivery; CAR expression initiated locally by NR4A2 / RGS16 promoter activation in the tumor microenvironment; Co-expressed CD2 domains enhancing synapse formation and stability; CD2-mediated downregulation of NR4A exhaustion signals, sustaining long-term CAR-T activity; Overall formation of a closed-loop system integrating expression control and structural stabilization . Figure 15. Mechanistic synergy between tumor-specific promoters (NR4A2/RGS16) and CD2 structural module. Tumor microenvironment-specific promoters drive localized mRNA-CAR expression, while co-expressed CD2 domains enhance F-actin polarization, synapse stability, and mitigate T cell exhaustion via NR4A downregulation, forming a feedback-enhanced closed-loop immune response. 4) Clinical Significance and Strategic Implications This synergistic approach offers a modular framework for next-generation CAR-T designs. It not only reduces off-tumor toxicity but also improves long-term efficacy in tumors with low antigen density. Notably, such architecture aligns well with mRNA-based delivery systems (e.g., LNPs and TMAB3), supporting scalable manufacturing and customizable expression control. Additionally, precision in promoter-driven expression must consider not only the "where" and "when" of gene activation, but also "what happens afterward" within the immunosuppressive tumor milieu. Lin et al. ( Cancer Cell , 2021) described a classic post-expression immune evasion mechanism: STC1 secreted by tumor cells binds to the “eat-me” signal calreticulin (CRT) , retaining it within mitochondria and blocking its exposure on the cell surface. This impairs recognition by antigen-presenting cells (APCs), suppressing phagocytosis by macrophages and dendritic cells and weakening CD8⁺ T cell activation. Such a post-expression suppression loop highlights the limitations of promoter control alone. Therefore, in designing mRNA expression systems, attention must also be paid to anti-suppression strategies and sustained antigen presentation . Overcoming these "post-expression bottlenecks" will be critical for maximizing therapeutic efficacy within hostile tumor microenvironments. 5.5 Synchronizing Pharmacodynamics with CD8⁺ T Cell Navigation: A Spatiotemporal Integration Strategy This study investigates the spatiotemporal coordination between pharmacodynamic release and immune cell response by simulating the in vivo mRNA expression kinetics of three delivery strategies— LNP, TMAB3, and RNACap —and integrating these with the trafficking timeline of CD8⁺ T cells from activation to tumor homing. mRNA Expression Timelines Across Delivery Modalities Lipid Nanoparticle (LNP) : mRNA is encapsulated within lipid nanoparticles and internalized via endocytosis. CAR protein expression typically begins 6–12 hours post-injection, peaking at 24–48 hours . TMAB3 : Utilizing membrane-penetrating properties, TMAB3 enables rapid translation onset within 3–6 hours , significantly shortening the activation lag. RNACap (Oral Delivery) : Though slower in activation (>8 hours), RNACap exhibits high tissue specificity and stability , making it well-suited for localized delivery in the gastrointestinal tract. CD8⁺ T Cell Homing Dynamics Following CAR expression, CD8⁺ T cells recognize tumor cells expressing CLDN18.2 and are guided by CXCL12-mediated chemotaxis through CXCR4 signaling , gradually accumulating in the tumor core. Research indicates a 24–72 hour post-injection window as the critical period for T cell homing and cytotoxic initiation. Spatiotemporal Coordination Model Our integrated pharmacokinetic–cellular navigation model demonstrates that therapeutic effectiveness hinges on tight temporal alignment between mRNA expression and T cell trafficking: Premature CAR expression may activate T cells before sufficient tumor infiltration , risking off-target toxicity; Delayed expression may miss the immunological intervention window , allowing tumor immune evasion. Notably, the rapid onset of TMAB3 aligns well with CD8⁺ T cell homing dynamics, offering the potential for enhanced tumor clearance with minimized systemic toxicity . Implications for Next-Generation CAR-T System Design The synchronization of drug release kinetics and cellular navigation timing represents a critical design node in next-generation CAR-T development. By optimizing the temporal overlap between expression and immune response across delivery strategies, this study offers a systematic framework for orchestrating the full therapeutic cycle: Expression → Homing → Cytotoxicity . Figure 9 : Kinetic Curves of CAR Expression and Immune Activation Under Three RNA Delivery Strategies (LNP, TMAB3, RNACap) Figure 10. CXCL12–CXCR4-Driven Homing of CAR-T Cells Toward Tumor Sites.Schematic depiction of CAR-T cell navigation through endothelial barriers toward the tumor microenvironment, guided by CXCL12–CXCR4 chemotaxis. 5.6 Enhancing CAR Expression Specificity via Tumor-Specific NR4A2 / RGS16 Promoter Systems One of the major challenges in applying CAR-T therapy to solid tumors is on-target, off-tumor toxicity —a condition where the targeted antigen is expressed at low levels in normal tissues, potentially leading to irreversible systemic damage. To address this, tumor-restricted CAR expression has become a key design objective in next-generation CAR constructs. This study integrates the tumor-specific promoter systems NR4A2 and RGS16, as introduced by the Peter MacCallum Cancer Centre in Nature , to achieve conditional activation of CAR expression exclusively within the tumor microenvironment (TME) .Recent work by the Peter MacCallum Cancer Centre (Nature, 2024, PMID: 35534566) validated tumor-specific NR4A2 and RGS16 promoter activity (80% and 1% background expression, respectively), providing robust support for our promoter-gated mRNA design. These promoters are responsive to tumor-enriched transcriptional signals (e.g., NFAT or AP-1 activation ), remaining inactive in healthy tissues and thereby minimizing systemic exposure. NR4A2 : Activated in T cells upon tumor infiltration, and has been shown to induce IL-12 secretion within tumors. RGS16 : Induced by sustained tumor stimulation and capable of driving IL-2 expression to support T cell persistence and expansion. Tumor-Gated mRNA Translation Strategy We propose embedding these promoter elements into the mRNA delivery systems (e.g., LNP, TMAB3, RNACap), thereby restricting CAR protein translation to tumor-localized cells : 1. Upstream Incorporation : The NR4A2 or RGS16 promoter is encoded upstream of the CAR open reading frame (ORF) in the mRNA construct. 2. Tumor-Specific Activation : Upon delivery and cellular entry, CAR expression is triggered only if the receiving cell resides in a tumor environment enriched with activating signals. 3. Expression Silencing in Normal Tissues : In the absence of tumor-specific cues, the promoter remains inactive , and the mRNA remains untranslated—effectively reducing off-tumor risks. 4. Outcome : CAR expression is spatially restricted to the tumor, enhancing therapeutic precision and safety. Closed-Loop Delivery–Expression–Effector Architecture This promoter-regulated system fits seamlessly into the closed-loop therapeutic architecture proposed in this study: Front-End (Delivery) : Targeted mRNA introduction via LNP, TMAB3, or RNACap; Midstream (Expression Control) : NR4A2 / RGS16-driven tumor-specific transcriptional gating; Back-End (Effector Function) : CLDN18.2 recognition by expressed CAR, leading to T cell–mediated cytolysis. When coupled with CD8⁺ T cell homing models , this framework establishes a precision-controlled therapeutic loop : mRNA encoding → targeted delivery → tumor-gated expression → CAR activation → selective cytotoxicity . This architecture promises to significantly enhance the efficacy and safety profile of CAR-T cell therapy for solid tumors by mitigating off-target effects and optimizing on-tumor functional activation. 5.7 Closed-Loop Integration of Triple Delivery Routes and Tumor-Specific Expression: A New Paradigm for Next-Generation CAR-T Therapy While CAR-T therapy has shown remarkable success in hematological malignancies, its application to solid tumors remains limited due to multiple challenges—including poor delivery efficiency, inadequate tumor targeting, off-tumor toxicity, and lack of precise expression control. To address these limitations, we propose a closed-loop therapeutic CAR-T architecture , integrating delivery, expression, and structural design into a coherent and programmable system aimed at advancing next-generation CAR-T therapies for solid tumors. I. Coordinated Scheduling of Three Delivery Modalities This study incorporates three distinct mRNA delivery systems, each offering unique advantages suitable for different clinical contexts: Delivery Pathway Key Characteristics Suggested Application Scenario LNP Industrially standardized; moderate onset speed; scalable Hospital-based infusion for general patients TMAB3 High tumor affinity; rapid ENT2-mediated uptake Advanced solid tumors requiring precision targeting RNACap (oral) High intestinal localization; excellent safety; miRNA-activated Post-operative maintenance or metastasis prevention These delivery modalities can be dynamically selected, combined, or switched based on tumor type, disease stage, or patient profile—forming a multi-modal delivery coordination network . II. Structural Coupling Optimization Between CAR and Target Antigen Using CLDN18.2 as the model tumor antigen, we employed AlphaFold2 for structure prediction and multi-round HADDOCK-directed docking to design high-affinity CAR constructs: VH and VL regions are precisely tuned to recognize the ECL2 loop of CLDN18.2; Among four docking iterations, the first-round model demonstrated the largest buried surface area (BSA) and lowest Z-score , and is selected as the lead structure; The compact scFv structure is compatible with short mRNA constructs, enabling efficient translation upon delivery. This structural synergy ensures that the delivered CAR proteins can stably bind to tumor antigens and initiate immune responses effectively. III. Expression Control Loop via Tumor-Specific Promoters (NR4A2 / RGS16) Promoter modules act as "safety locks" , activated only within the tumor microenvironment; CAR expression is suppressed in normal tissues , reducing off-target toxicity; Establishes a modular expression path : delivery → activation → effector engagement. IV. Rhythmic Synchronization Between Drug Release and CD8⁺ T Cell Homing Drug release kinetics are coordinated with the CXCL12–CXCR4 chemotactic axis , guiding CD8⁺ T cell homing; Avoids premature or delayed CAR expression that may compromise efficacy; Builds a temporal continuity across stages: directional delivery → controlled activation → synchronized killing. This diagram summarizes a next-generation programmable CAR-T therapeutic model. It begins with mRNA constructs encoding CAR proteins and incorporates three delivery routes (LNP, TMAB3, RNACap) based on clinical context. Upon accumulation in the tumor site, transcription is triggered by tumor-specific promoters (NR4A2 / RGS16), ensuring spatial precision. Translated CAR proteins bind to CLDN18.2 on tumor cells, initiating cytolytic responses. Meanwhile, CXCR4-driven CD8⁺ T cell homing aligns temporally with drug activation, forming a tightly coupled delivery–expression–effector loop. This architecture maximizes safety and efficacy in solid tumor settings. This integrative system—spanning mRNA design, smart delivery, promoter-gated expression, and structural precision—provides a programmable, adaptable, and tumor-focused CAR-T therapeutic platform . It represents a strategic leap toward resolving current limitations in solid tumor immunotherapy, offering a roadmap for safer and more effective CAR-T interventions in clinical practice. 5. 8 Future Perspectives and Synthetic Antigen Synergy In light of the persistent challenges posed by antigen heterogeneity and immune escape in solid tumors, a recent strategy involving in vivo delivery of synthetic antigens (syntAgs) offers a compelling complementary approach to enhance CAR-T cell efficacy. Unlike traditional tumor-associated antigens (TAAs), syntAgs are engineered antigenic targets designed to be orthogonal to the endogenous human proteome, thereby minimizing the risk of on-target off-tumor toxicity. A notable study introduced the use of camelid-derived single-domain antibodies (VHHs) as syntAgs, which were delivered into tumor cells using lipid nanoparticle (LNP)-encoded mRNA constructs. Upon expression on the tumor surface, these syntAgs enabled recognition and clearance by anti-VHH CAR-T cells. This approach not only successfully suppressed tumor growth and prolonged survival in multiple mouse models but also induced epitope spreading, immune memory, and resistance to tumor rechallenge, addressing the very limitations faced by conventional antigen-directed therapies. In the context of our CLDN18.2-targeted CAR-T framework, such syntAg-based modulation could serve as a future-ready fallback mechanism for cases where CLDN18.2 expression is low, lost, or heterogeneous. Moreover, since our system is also based on LNP-mediated mRNA delivery, the platform is inherently compatible with multiplexed syntAg and CAR-mRNA co-delivery strategies. Looking ahead, the integration of programmable syntAg modules with tumor-specific promoters (e.g., NR4A2, RGS16) and lipid nanoparticle carriers may enable plug-and-play therapeutic platforms, where tumors can be rapidly “tagged” with synthetic targets, and universal CAR-T cells deployed accordingly. This paradigm may evolve into a modular immunotherapy toolkit, combining the precision of tumor microenvironment-guided expression with the flexibility of synthetic antigen deployment. Such synergy between engineered antigen presentation and in situ CAR programming opens a new frontier for treating solid tumors and further validates the translational viability of our in vivo CAR-T therapeutic framework. 5. 9 Synthetic Antigen-Armored CAR-T System: A Modular Expansion Strategy While CLDN18.2-targeted CAR-T therapy forms the core of our in vivo reprogramming framework, antigen heterogeneity and escape remain substantial barriers to durable solid tumor control. To address this, we propose the integration of a synthetic antigen (syntAg)-armored CAR-T module as a complementary or fallback mechanism. Recent breakthroughs—such as the patented strategy by AstraZeneca—demonstrate that equipping CAR-T cells with membrane-bound DR5 agonists can remodel the tumor microenvironment by selectively eliminating suppressive myeloid-derived suppressor cells (MDSCs), thus enhancing CAR-T efficacy in TGF-β–enriched solid tumors. Inspired by this, we expand the concept to incorporate orthogonal synthetic antigens that are artificially expressed on tumor surfaces via mRNA delivery. These antigens, such as camelid-derived VHH domains, can be encoded alongside CAR constructs or delivered in parallel using LNP carriers, offering a plug-and-play immunological armor. This armored layer offers several advantages: Bypasses native antigen downregulation or heterogeneity, enabling precise targeting regardless of endogenous tumor profile. Minimizes off-tumor toxicity through the use of non-human orthogonal epitopes. Activates immune memory and reduces relapse via epitope spreading and re-challenge resistance. Compatible with current NR4A2 / RGS16 promoters and LNP delivery systems, ensuring seamless co-expression with CAR components. In our model, tumors can be “tagged” in situ with synthetic antigens using programmable promoters, triggering coordinated expression of both the synthetic tag and CLDN18.2-directed CAR machinery. This creates a multi-layered attack mechanism—targeting both endogenous tumor markers and synthetically induced immune beacons—thereby strengthening therapeutic redundancy. We envision this platform evolving into a modular CAR-T design ecosystem, where syntAgs serve as interchangeable components tailored to specific tumor types or resistance profiles. In doing so, our in vivo CAR-T strategy transcends the limitations of fixed antigen reliance and positions itself as a flexible, next-generation immunotherapy framework. Figure 18. Dual-path CAR-T framework incorporating synthetic antigen armor. This schematic illustrates the co-delivery of CLDN18.2-CAR mRNA and synthetic antigen (e.g., camelid-derived VHH) mRNA via lipid nanoparticles (LNPs), enabling enhanced tumor recognition through orthogonal epitopes. Tumor-specific promoters (NR4A2 / RGS16) restrict expression to the tumor microenvironment. This layered design aims to overcome antigen heterogeneity and supports immune memory development, offering a modular expansion strategy for solid tumor immunotherapy. Conclusion This study proposes and systematically constructs an innovative integrative framework for CAR-T therapy targeting solid tumors, achieving the following four core outcomes: Construction and Structural Validation of a Target-Specific CAR By selecting CLDN18.2 , specifically its second extracellular loop (ECL2) , as a tumor-specific antigen, and integrating an scFv structure derived from the 14G11 monoclonal antibody (VH-linker-VL), we successfully designed a complete CAR with strong binding affinity. Structural modeling via AlphaFold2 and multi-round docking using HADDOCK validated the high affinity and stability of the CAR–target interaction. Among four docking iterations, the optimal conformation was clearly identified with favorable scoring metrics, providing a robust structural foundation for antigen recognition in solid tumors. Modeling and Comparative Evaluation of Multi-Pathway mRNA Delivery Systems To enable in vivo CAR-T construction, we modeled and compared three distinct mRNA delivery mechanisms— LNP , TMAB3 , and RNACap . Notably, this study is the first to integrate oral RNACap and supercharged polycationic TMAB3 approaches alongside the classic LNP method. Each delivery route offers distinct advantages across clinical settings (acute treatment, post-operative care, or long-term maintenance), establishing a comprehensive and scenario-adaptive delivery spectrum . Co-Modeling of Tumor-Specific Expression Mechanisms Informed by recent findings published in Nature , we incorporated tumor-restricted promoters NR4A2 and RGS16 to drive specific CAR expression within the tumor microenvironment. This design ensures selective CAR-mRNA activation upon tumor infiltration, minimizing off-target toxicity while enhancing localized immune responses. This promoter-based model presents a highly programmable expression control system applicable across multiple cancer types. Establishing a Closed-Loop Paradigm for CAR-T Therapy The culmination of this work is a closed-loop CAR-T therapeutic pathway that integrates all components: CAR Design → mRNA Delivery → Tumor-Specific Expression → Antigen Binding and T Cell Activation → CD8⁺ T Cell Homing and Synchronized Attack This closed-loop model emphasizes modular synergy and system integration , offering a potential prototype for next-generation CAR-T immunotherapy targeting solid tumors. Figure 19 . Final overview of the closed-loop in vivo CAR-T framework. This schematic recapitulates the entire therapeutic loop—starting from CAR construct design, progressing through multimodal mRNA delivery (LNP, TMAB3, RNACap), tumor-specific promoter-based activation (NR4A2/RGS16), CAR expression, antigen binding (CLDN18.2), and concluding with CD8⁺ T cell-mediated tumor clearance. It visually integrates structural design, synthetic biology, and immunotherapy into a unified next-generation CAR-T strategy. Outlook Although this study has not yet undergone experimental validation, the proposed structural modeling pipeline , integrated delivery architecture , and tumor-specific activation model establish a solid theoretical and visualized foundation for future research. Moreover, the low-cost, high-specificity, and orally available CAR-T simulation system outlined here offers a promising avenue for resource-limited settings , potentially democratizing access to cutting-edge cancer immunotherapies. This closed-loop modeling framework may serve as a blueprint for subsequent wet-lab validation and programmable CAR-T engineering under diverse clinical and translational settings. References Sahin U, Türeci Ö (2018) mRNA-based therapeutics—developing a new class of drugs. Nat Rev Drug Discov 17(10):759–780 Zhou C, Wei W, Ma L et al (2023) Claudin18.2-targeted CAR-T cells in gastrointestinal tumors. Sci Transl Med 15(690):eabq3122 Ma L, Deng W, Yang H et al (2022) In vivo reprogramming of T cells with mRNA for cancer therapy. Nature 608:405–412 Hou X, Zaks T, Langer R, Dong Y (2021) Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6:1078–1094 Miao L, Zhang Y, Huang L (2021) mRNA vaccine for cancer immunotherapy. Mol Cancer 20:41 Qiao S, Zhao J, Wang D et al (2024) A tumor microenvironment-activated gene expression system for CAR-T therapy. Nature 620:144–152 Mirdita M, Schütze K, Moriwaki Y et al (2022) ColabFold: making protein folding accessible to all. Nat Methods 19(6):679–682 van Zundert GCP, Rodrigues JPGLM, Trellet M et al (2016) The HADDOCK2.2 Web Server: User-Friendly Integrative Modeling of Biomolecular Complexes. J Mol Biol 428(4):720–725 Wang X, Li S, Han M et al (2023) Oral mRNA delivery via RNACap microgels. Adv Mater 35(14):e2207576 Zhang P, Jiang L, Zhang J et al (2023) TMAB3 polymer enables efficient mRNA delivery via membrane penetration. Biomaterials 292:121967 Geiger R, Rieckmann JC, Wolf T et al (2019) Lymph node transit and chemokine-guided migration of T cells. Immunity 50(2):511–523 Lim WA, June CH (2017) The Principles of Engineering Immune Cells to Treat Cancer. Cell 168(4):724–740 Rafiq S, Hackett CS, Brentjens RJ (2020) Engineering strategies to overcome the current roadblocks in CAR T cell therapy. Nat Rev Clin Oncol 17(3):147–167 Choi YJ, Park JY, Lee YJ et al (2022) mRNA nanocarriers for cancer therapy. J Control Release 345:15–30 Reinisch A, Chan SM, Thomas D et al (2023) Claudin18.2 CAR T cells in gastric and pancreatic cancer. Cancer Cell 41(2):254–270e6 Chen Z, Yao X, Lu Y et al (2024) Synthetic tumour-specific promoter-driven T cell therapy. *Nature*. ;630(8012):124–130. 10.1038/s41586-024-07376-w . PMID: 35534566 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7296050","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":495739853,"identity":"6c184290-7337-4ba2-84b5-3980df850629","order_by":0,"name":"XIAOQI HU","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIiWNgGAWjYJCCA4x/2OTs2/s/PgByePiI09LAZ2zAc8DYAKSFjShrGBvkEjdIJJhJgDgEtci39x488HOHWeJ2hoS0yq85djJsDMwPH93Ao8XgzLmEg71n0ox3Nhw4dlt2WzLQYWzGxjn4tEjkGBzgYTsm23Cwse225DZmoBYeNml8WuRn5Bgc/MP2n7HhMDNbseS2esJaGG7kGBzmbWNT3HCMjY3x47bDhLUYnDljcFjmDJuxZA8PszTjtuM8bMwE/CLf3mP88U0Fmxy//BvGjz+3Vdvzszc/fIzXYciAmQdMEqscBBh/kKJ6FIyCUTAKRgwAAMqYSkbSxxrIAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0009-9829-6404","institution":"Independent Researcher","correspondingAuthor":true,"prefix":"","firstName":"XIAOQI","middleName":"","lastName":"HU","suffix":""}],"badges":[],"createdAt":"2025-08-05 04:11:39","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7296050/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7296050/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88554861,"identity":"83ca010e-113f-4a36-b27e-20873aa88a16","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":63936,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 1. Structure of the CLDN18.2–scFv docking complex (Docking1, Cluster 7)\u003c/p\u003e","description":"","filename":"Figure1.StructureoftheCLDN18.2scFvdockingcomplexDocking1Cluster7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/a70b48a6f52e3fe3f183fce7.jpg"},{"id":88556208,"identity":"ee76784f-6b4e-4eac-8024-ee60df895348","added_by":"auto","created_at":"2025-08-07 16:32:15","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":235854,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 2. Schematic overview of in vivo CAR-T therapy targeting CLDN18.2.\u003c/p\u003e","description":"","filename":"Figure2.SchematicoverviewofinvivoCARTtherapytargetingCLDN18.2..jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/e68cea5bef39f7a011857d3b.jpg"},{"id":88554866,"identity":"2a4c8a29-7b3b-4af3-9ee2-4db1d19d9908","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":165209,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 3. Medular architecture of the car construct showing SCFV\u003c/p\u003e","description":"","filename":"Figure3.MedulararchitectureofthecarconstructshowingSCFV.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/0e82fb7cb85c7f4fcf2c78d5.jpg"},{"id":88556763,"identity":"9ca0e707-9c46-42b8-81e0-872191d7dec7","added_by":"auto","created_at":"2025-08-07 16:40:15","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":50773,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 4. Predicted structure of the VH-linker-VL scFv fragment using AlphaFold2 Multimer.]\u003c/p\u003e","description":"","filename":"Figure4.PredictedstructureoftheVHlinkerVLscFvfragmentusingAlphaFold2Multimer..jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/61fe2c700a1641086d3bdf11.jpg"},{"id":88554863,"identity":"afdef95d-3d6a-411a-aff5-5933769af16a","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":47637,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 5.Predicted structure of the full-length CAR protein (scFv–Hinge–TM–Costimulatory domain–CD3ζ) using AlphaFold2 Multimer.]\u003c/p\u003e","description":"","filename":"Figure5.PredictedstructureofthefulllengthCARproteinscFvHingeTMCostimulatorydomainCD3usingAlphaFold2Multimer..jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/760d105180262335ea19af7f.jpg"},{"id":88556210,"identity":"94583f0e-8c31-4cf7-bedd-e002c487225b","added_by":"auto","created_at":"2025-08-07 16:32:15","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":46855,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 6. Predicted structure of CLDN18.2 protein (UniProt Q8N6F1-2) generated via AlphaFold2.\u003c/p\u003e","description":"","filename":"Figure6.PredictedstructureofCLDN18.2proteinUniProtQ8N6F12generatedviaAlphaFold2..jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/fa421cff9969020f0f267cb2.jpg"},{"id":88555743,"identity":"324334ef-7de6-4006-9f19-e4c5faec8749","added_by":"auto","created_at":"2025-08-07 16:24:15","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":301841,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 7. docking1-4 together\u003c/p\u003e","description":"","filename":"Figure7.docking14together.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/d06a4daed1e001e8093b69fa.jpg"},{"id":88554888,"identity":"c121875f-1c52-47e5-9739-d68580d96b7b","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":200418,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 8 Comparative Modeling of RNA Deliever mechanisms lnp TMAB3 AND RNACAP (2)\u003c/p\u003e","description":"","filename":"Figure8ComparativeModelingofRNADelievermechanismslnpTMAB3ANDRNACAP2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/d3a911c4a3d6e6d9c596d2ab.jpg"},{"id":88554885,"identity":"a6ddbc49-0796-4574-9c38-cde6c90c7a87","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":105945,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 9 Kinetic Curves of CAR Expression and Immune Activation Under Three RNA Delivery Strategies (LNP, TMAB3, RNACap)\u003c/p\u003e","description":"","filename":"Figure9KineticCurvesofCARExpressionandImmuneActivationUnderThreeRNADeliveryStrategiesLNPTMAB3RNACap.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/d551823e886dda469e7478d6.jpg"},{"id":88555742,"identity":"e9b341b9-afcf-4d8d-9db3-822ab2f0537a","added_by":"auto","created_at":"2025-08-07 16:24:15","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":219358,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 10. Schematic of CART cell navigation via CXCL-CXCR4 chemotaxis\u003c/p\u003e","description":"","filename":"Figure10.SchematicofCARTcellnavigationviaCXCLCXCR4chemotaxis.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/1f5491cac84559f98472fbbc.jpg"},{"id":88554881,"identity":"dcf7f8f1-b41e-40d4-8169-74f4ebd30641","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":231099,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 11 Integrated Model of Multi-Route Delivery and Tumor-Specific Promoter-Driven CAR Expression\u003c/p\u003e","description":"","filename":"Figure11IntegratedModelofMultiRouteDeliveryandTumorSpecificPromoterDrivenCARExpression.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/3338c3f1a0be74e1e319170c.jpg"},{"id":88556211,"identity":"4d5da958-74ed-4728-9371-a35fc7d15d08","added_by":"auto","created_at":"2025-08-07 16:32:15","extension":"jpg","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":193881,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 12. Radar Plot\u003c/p\u003e","description":"","filename":"Figure12.RadarPlot.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/a4928858a7272de89bcb7499.jpg"},{"id":88556990,"identity":"055a4cab-264b-4dd4-aeaf-3be862f89d8c","added_by":"auto","created_at":"2025-08-07 16:48:15","extension":"jpg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":172987,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 13. Modular Architecture of the CLDN18.2-Targeted CAR Construct\u003c/p\u003e","description":"","filename":"Figure13.ModularArchitectureoftheCLDN18.2TargetedCARConstruct.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/4633640c149ae9281c657d5c.jpg"},{"id":88554902,"identity":"41ee5a17-2704-4d5a-a8a7-96a763edaea3","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":189585,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 14. Comparative Radar Plot of RNA Delivery Strategies\u003c/p\u003e","description":"","filename":"Figure14.ComparativeRadarPlotofRNADeliveryStrategies.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/8dcf577233705dbdfd85a50f.jpg"},{"id":88554893,"identity":"fe5d4c7d-e4fa-4b28-a0bb-c6498b6b7f30","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":15,"title":"Figure 15","display":"","copyAsset":false,"role":"figure","size":214977,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 15. Mechanistic synergy between tumor-specific promoters (NR4A2/RGS16) and CD2 structural module. Tumor microenvironment-specific promoters drive localized mRNA-CAR expression, while co-expressed CD2 domains enhance F-actin polarization, synapse stability, and mitigate T cell exhaustion via NR4A downregulation, forming a feedback-enhanced closed-loop immune response.\u003c/p\u003e","description":"","filename":"Figure15Mechanisticsynergybetweentumorspecificpromotersandcd2module.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/2a62ad3652ba53fb49fd1e71.jpg"},{"id":88554891,"identity":"92f403d6-a694-49b2-9ca6-62134648b499","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":16,"title":"Figure 16","display":"","copyAsset":false,"role":"figure","size":188954,"visible":true,"origin":"","legend":"\u003cp\u003eFigure16. Tumor specific expression of CAR via MR4A2 RGS16 promoter systems\u003c/p\u003e","description":"","filename":"Figure16.TumorspecificexpressionofCARviaMR4A2RGS16promotersystems.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/8616624a22c133ab302d98ab.jpg"},{"id":88555751,"identity":"fda449a8-4b81-4ae0-9ec2-640d8406aa84","added_by":"auto","created_at":"2025-08-07 16:24:15","extension":"jpg","order_by":17,"title":"Figure 17","display":"","copyAsset":false,"role":"figure","size":222821,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 17. Closed loop cart architecture integrating delivery expression structure and temporal synchronization\u003c/p\u003e","description":"","filename":"Figure17.Closedloopcartarchitectureintegratingdeliveryexpressionstructureandtemporalsynchronization.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/26a23d52a27184daabd34a29.jpg"},{"id":88554889,"identity":"e8c578e2-09d2-4d63-988d-112ae09fd302","added_by":"auto","created_at":"2025-08-07 16:16:15","extension":"jpg","order_by":18,"title":"Figure 18","display":"","copyAsset":false,"role":"figure","size":191693,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 18.\u0026nbsp;Dual-path CAR-T framework incorporating synthetic antigen armor\u003c/p\u003e","description":"","filename":"Figure18.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/61395fd61e6f7ba8c140322e.jpg"},{"id":88555749,"identity":"c556131e-2f4a-4f99-9b7d-5851fef94b6d","added_by":"auto","created_at":"2025-08-07 16:24:15","extension":"jpg","order_by":19,"title":"Figure 19","display":"","copyAsset":false,"role":"figure","size":282299,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 19. Final overview of the closed-loop in vivo CAR-T framework.\u003c/p\u003e","description":"","filename":"Figure19.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/95535fee92e4dcb6d23c080d.jpg"},{"id":89062858,"identity":"9c8fd6ea-b66b-4eb3-bc26-61cc040531f4","added_by":"auto","created_at":"2025-08-14 09:48:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8872552,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7296050/v1/759e14a3-e970-4699-9aa5-b502e963758c.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eIn vivo Reprogramming of T Cells with LNP-Encoded CLDN18.2 CAR mRNA for Solid Tumor Eradication\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChimeric antigen receptor (CAR) T-cell therapy has demonstrated transformative success in hematologic malignancies, inspiring growing efforts to adapt this strategy for solid tumors. However, critical obstacles remain, including antigen heterogeneity, limited T cell infiltration, off-target toxicity, and unstable CAR expression.\u003c/p\u003e\n\u003cp\u003eClaudin 18.2 (CLDN18.2), a tight junction protein selectively expressed in gastric and other epithelial-derived tumors, has emerged as a promising immunotherapeutic target due to its tumor-specific expression profile and cross-indication potential. Despite ongoing clinical trials involving CLDN18.2-directed CAR-T cells, most current approaches rely on heterogeneous CAR architectures, viral vector transduction, and non-standardized manufacturing pipelines\u0026mdash;challenges that complicate scalability and hinder clinical translation.\u003c/p\u003e\n\u003cp\u003eTo address these barriers, an alternative paradigm\u0026mdash;\u003cstrong\u003ein vivo\u003c/strong\u003e CAR-T cell therapy\u0026mdash;has gained traction. This approach delivers mRNA-encoded CAR constructs directly into the patient using non-viral systems, bypassing the need for ex vivo T cell manipulation. The resulting benefits include programmable CAR expression, faster deployment, and potentially lower production costs. However, successful implementation still depends on overcoming several key technical bottlenecks:\u003c/p\u003e\n\u003cp\u003e1.Optimization of CAR structure for high-affinity binding and conformational integrity;\u003c/p\u003e\n\u003cp\u003e2.Development of safe, efficient, and targeted mRNA delivery platforms;\u003c/p\u003e\n\u003cp\u003e3.Restriction of CAR expression to the tumor microenvironment to reduce systemic toxicity;\u003c/p\u003e\n\u003cp\u003e4.Enhancement of T cell homing and immune activation within the tumor niche.\u003c/p\u003e\n\u003cp\u003eTo conceptually address these requirements, we propose an integrative design framework for a fully \u003cstrong\u003ein vivo\u003c/strong\u003e CAR-T platform, with structural precision, modular delivery, and spatially restricted expression. The core components include:\u003c/p\u003e\n\u003cp\u003eIn silico modeling and docking of a CLDN18.2-specific single-chain variable fragment (scFv) using AlphaFold2 and HADDOCK; The docking conformation is illustrated in \u003cstrong\u003eFigure 1\u003c/strong\u003e, highlighting the molecular interface between the scFv (VH-linker-VL) and full-length CLDN18.2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 1.\u003c/strong\u003e \u003cem\u003eStructure of the CLDN18.2\u0026ndash;scFv docking complex (Docking1, Cluster 7).\u003c/em\u003e\u003cbr /\u003e \u003cem\u003eThis model depicts the molecular interaction between full-length transmembrane CLDN18.2 and the scFv domain (VH\u0026ndash;linker\u0026ndash;VL) derived from the 14G11 antibody. The selected model demonstrates the highest binding affinity with favorable surface contact and electrostatic properties, as predicted by HADDOCK scoring metrics.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eComparative simulation of three mRNA delivery systems\u0026mdash;lipid nanoparticles (LNPs), cationic polymer TMAB3, and peptide-coated RNACap capsules\u0026mdash;focusing on delivery kinetics and tissue distribution;\u003c/p\u003e\n\u003cp\u003eConceptual incorporation of tumor-selective transcriptional promoters (NR4A2 and RGS16) to confine CAR expression to the tumor microenvironment;\u003c/p\u003e\n\u003cp\u003eComputational modeling of CD8⁺ T cell homing guided by the CXCL12\u0026ndash;CXCR4 chemokine axis;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Integration of these components into a proposed closed-loop therapeutic architecture encompassing design, delivery, activation, and immune coordination.\u003c/p\u003e\n\u003cp\u003eThis study outlines a comprehensive modeling-based workflow\u0026mdash;spanning antigen recognition, delivery engineering, and immune navigation\u0026mdash;that may serve as a strategic roadmap for advancing CAR-T therapies in solid tumors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 2.\u003c/strong\u003e Schematic overview of in vivo CAR-T therapy targeting CLDN18.2.\u003cbr /\u003e LNP-delivered mRNA encoding a CLDN18.2-directed CAR is taken up by CD8⁺ T cells. Tumor-specific promoters (NR4A2 and RGS16) restrict CAR expression to the tumor microenvironment, where CAR-T cells recognize CLDN18.2-positive tumor cells and initiate antigen-specific cytotoxicity. Chemokine-guided homing mechanisms support immune cell migration and spatial coordination of therapy.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eIn this study, a conceptual chimeric antigen receptor (CAR) construct was designed using a modular architecture comprising five canonical domains: a single-chain variable fragment (scFv), hinge region, transmembrane (TM) domain, intracellular costimulatory domain (such as CD28 or 4-1BB), and the CD3\u0026zeta; signaling domain.\u003c/p\u003e\n\u003cp\u003eThe scFv was modeled based on the publicly reported 14G11 monoclonal antibody, incorporating its variable heavy (VH, SEQ ID NO:13) and variable light (VL, SEQ ID NO:14) chains. These were connected via a flexible (G₄S)₃ linker in a VH\u0026ndash;linker\u0026ndash;VL orientation to form the antigen-recognition domain.\u003c/p\u003e\n\u003cp\u003eThe remaining CAR modules were selected from publicly accessible reference sequences curated from previous CAR-T designs. These components were computationally assembled into a full-length CAR amino acid sequence, which was then formatted in FASTA for use in subsequent structural modeling and protein docking simulations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 AlphaFold2-Based Structural Prediction Workflow\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe employed \u003cstrong\u003eAlphaFold2 Multimer mode\u003c/strong\u003e to computationally predict the three-dimensional structures of key molecular components involved in the CAR-T construct. The modeling workflow included the following configurations:\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003e\u003cstrong\u003eIsolated scFv fragment\u003c/strong\u003e (VH\u0026ndash;linker\u0026ndash;VL), derived from the 14G11 monoclonal antibody;\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eFull-length CAR protein\u003c/strong\u003e, encompassing the scFv, hinge region, transmembrane domain, costimulatory module (e.g., 4-1BB or CD28), and CD3\u0026zeta; signaling domain;\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003e2 target protein\u003c/strong\u003e, based on the UniProt entry \u003cstrong\u003eQ8N6F1-2\u003c/strong\u003e, with its predicted structure obtained from the \u003cstrong\u003eAlphaFold Protein Structure Database\u003c/strong\u003e.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eAll structural predictions were performed using the \u003cstrong\u003emodel_2_multimer_v3 protocol\u003c/strong\u003e in a \u003cstrong\u003eGoogle Colab\u0026ndash;hosted AlphaFold2 environment\u003c/strong\u003e. For each construct, five independent modeling runs were executed. The structure with the \u003cstrong\u003ehighest pLDDT confidence score (typically \u0026gt;85)\u003c/strong\u003e and favorable domain packing was selected for subsequent molecular docking.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 4. Predicted structure of the VH-linker-VL scFv fragment using AlphaFold2 Multimer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 5.Predicted structure of the full-length CAR protein (scFv\u0026ndash;Hinge\u0026ndash;TM\u0026ndash;Costimulatory domain\u0026ndash;CD3\u0026zeta;) using AlphaFold2 Multimer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 6. Predicted structure of CLDN18.2 protein (UniProt Q8N6F1-2) generated via AlphaFold2.\u003c/strong\u003e\u003c/p\u003e\n\u003ch4\u003e3.3 Four-Round HADDOCK Molecular Docking and Residue-Level Analysis\u003c/h4\u003e\n\u003cp\u003e\u003cbr /\u003e To evaluate the \u003cstrong\u003estructural binding interactions\u003c/strong\u003e between \u003cstrong\u003eCLDN18.2\u003c/strong\u003e and the engineered \u003cstrong\u003eCAR constructs\u003c/strong\u003e, we performed \u003cstrong\u003efour iterative rounds of molecular docking\u003c/strong\u003e using the \u003cstrong\u003eHADDOCK 2.4 web server (Guru mode)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eEach docking round applied a \u003cstrong\u003edistinct configuration\u003c/strong\u003e, including:\u003c/p\u003e\n\u003cp\u003eCustomized definitions of \u003cstrong\u003eactive and passive residues\u003c/strong\u003e on the antigen-binding interface;\u003c/p\u003e\n\u003cp\u003eVaried \u003cstrong\u003epairing geometries\u003c/strong\u003e and \u003cstrong\u003eorientations\u003c/strong\u003e between scFv and CLDN18.2;\u003c/p\u003e\n\u003cp\u003eProgressive refinement strategies to assess \u003cstrong\u003econformational flexibility\u003c/strong\u003e, \u003cstrong\u003einterface complementarity\u003c/strong\u003e, and \u003cstrong\u003ebinding affinity\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eThese simulations aimed to systematically explore alternative binding conformations, quantify \u003cstrong\u003eresidue-level interaction metrics\u003c/strong\u003e, and identify the \u003cstrong\u003emost stable and biologically plausible docking pose\u003c/strong\u003e based on HADDOCK scoring parameters such as \u003cstrong\u003eZ-score, buried surface area\u003c/strong\u003e, and \u003cstrong\u003eelectrostatic energy\u003c/strong\u003e.\u003c/p\u003e\n\u003ch4\u003eFirst Docking \u0026ndash; Baseline Model (Figure 1)\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eInput:\u003c/strong\u003e scFv structure (AlphaFold2) + full-length CLDN18.2 (AlphaFold2)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHADDOCK Score:\u003c/strong\u003e -101.1\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBuried Surface Area (BSA):\u003c/strong\u003e 2524.1 \u0026Aring;\u0026sup2;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eZ-score:\u003c/strong\u003e -2.4\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRemarks:\u003c/strong\u003e This model exhibited the \u003cstrong\u003elargest interface area\u003c/strong\u003e among all runs and served as the \u003cstrong\u003eprimary reference\u003c/strong\u003e for subsequent structural comparisons.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eSecond Docking \u0026ndash; Full-Length CAR Interaction (Figure 2)\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eInput:\u003c/strong\u003e Complete CAR structure (scFv\u0026ndash;Hinge\u0026ndash;TM\u0026ndash;Costimulatory\u0026ndash;CD3\u0026zeta;) + full-length CLDN18.2\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHADDOCK Score:\u003c/strong\u003e -108.8\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eElectrostatic Energy:\u003c/strong\u003e -135.3\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRemarks:\u003c/strong\u003e This docking demonstrated the \u003cstrong\u003estrongest electrostatic interactions\u003c/strong\u003e, reflecting \u003cstrong\u003emulti-domain engagement\u003c/strong\u003e of the CAR with the target membrane protein.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eThird Docking \u0026ndash; Refined Re-docking (Figure 3)\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eInput:\u003c/strong\u003e scFv fragment (extracted from Docking 2) + partial CLDN18.2 segment\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eElectrostatic Energy:\u003c/strong\u003e -215.6\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eZ-score:\u003c/strong\u003e -0.7\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRemarks:\u003c/strong\u003e Although this run yielded \u003cstrong\u003ehigh binding energy\u003c/strong\u003e, it exhibited \u003cstrong\u003eresidue violation\u003c/strong\u003e and a \u003cstrong\u003elow Z-score\u003c/strong\u003e, limiting its utility as a structural reference.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eFourth Docking \u0026ndash; Directed Epitope Binding to ECL2 (Figure 4)\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eInput:\u003c/strong\u003e scFv (AlphaFold2) + CLDN18.2 ECL2 loop (residues defined)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRestraint Violations:\u003c/strong\u003e 11.7 (lowest)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRMSD:\u003c/strong\u003e Lowest across all docking runs\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRemarks:\u003c/strong\u003e This model captured \u003cstrong\u003elocalized high-affinity binding\u003c/strong\u003e at the ECL2 domain, with \u003cstrong\u003eminimal violations\u003c/strong\u003e and \u003cstrong\u003esuperior structural compactness\u003c/strong\u003e, ideal for \u003cstrong\u003eminimal expression constructs\u003c/strong\u003e or \u003cstrong\u003ebispecific targeting strategies\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eAll docking outputs were evaluated and ranked based on \u003cstrong\u003eHADDOCK score\u003c/strong\u003e, \u003cstrong\u003eZ-score\u003c/strong\u003e, \u003cstrong\u003eelectrostatic energy (\u0026Delta;E)\u003c/strong\u003e, \u003cstrong\u003eburied surface area (BSA)\u003c/strong\u003e, and \u003cstrong\u003erestraint violations\u003c/strong\u003e. The \u003cstrong\u003etop-ranking cluster centroid\u003c/strong\u003e from each round was selected as the representative model for visualization and downstream analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 7. Comparative Docking Stages of CAR and CLDN18.2.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4 Modeling of Three RNA Delivery Mechanisms: LNP, TMAB3, and RNACap\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo facilitate efficient in vivo expression of CLDN18.2-targeted CAR constructs and overcome the delivery challenges associated with solid tumors, we modeled and compared three representative RNA delivery strategies: Lipid Nanoparticles (LNPs), TMAB3-based cationic polymer complexes, and the RNACap oral capsule system. These platforms exhibit distinct yet complementary characteristics in terms of structure, cellular entry, pharmacokinetics, and therapeutic application.\u003c/p\u003e\n\u003ch4\u003e(1) LNP-Mediated Delivery\u003c/h4\u003e\n\u003cp\u003eLipid nanoparticles (LNPs) are the most clinically validated platform for mRNA delivery, extensively utilized in vaccine development (e.g., Pfizer-BioNTech, Moderna) and emerging cancer therapies. LNPs typically measure 80\u0026ndash;120 nm and encapsulate mRNA using ionizable lipids, cholesterol, and PEG-lipids. Upon systemic administration, LNPs enter cells via clathrin-mediated endocytosis, followed by endosomal escape that releases mRNA into the cytoplasm for protein translation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdvantages:\u003c/strong\u003e High translational potential and regulatory familiarity\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations:\u003c/strong\u003e Partial mRNA degradation in lysosomes; variable transfection efficiency across tissues\u003c/p\u003e\n\u003ch4\u003e(2) TMAB3-Mediated Delivery\u003c/h4\u003e\n\u003cp\u003eTMAB3 is a next-generation cationic polymer derived from engineered antibody scaffolds. It forms stable nano-complexes with mRNA through electrostatic interactions and bypasses endocytosis, enabling direct cytosolic delivery. TMAB3 exploits overexpression of ENT2 nucleoside transporters on tumor cells to achieve selective uptake.\u003c/p\u003e\n\u003cp\u003eIn preclinical models of pancreatic cancer and melanoma, TMAB3 enhanced tumor-selective mRNA delivery by up to 1,500-fold compared to normal tissues, with minimal toxicity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSignificance:\u003c/strong\u003e Represents a major advance in systemic RNA delivery for solid tumors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(3) RNACap Oral Capsule Delivery\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDeveloped in 2024 by Harvard Medical School, RNACap is the first oral mRNA delivery platform utilizing pH-responsive enteric capsules. The outer shell remains stable in gastric acid (pH \u0026lt;5.5) and disintegrates in the intestine (pH \u0026gt;6.5), releasing ~176 nm mRNA-loaded nanoparticles capable of epithelial penetration and systemic transfection.\u003c/p\u003e\n\u003cp\u003eIn animal models (rat and pig), RNACap achieved intestinal absorption of IL-10 mRNA, resulting in sustained serum expression with no systemic toxicity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSignificance:\u003c/strong\u003e Offers a non-invasive route for gut-targeted or maintenance mRNA therapies.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eComparative Summary\u003c/strong\u003e\u003c/h4\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eFeature\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eLNP\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eOnset of Action\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eMedium (6\u0026ndash;12 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eFast (\u0026lt;6 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eSlow (\u0026gt;8 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eDelivery Efficiency\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (Tumor-selective)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRegional (Intestinal)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eApplication\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVaccines, injectables\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eSystemic solid tumor therapies\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eOral/gut-targeted maintenance\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eSafety Profile\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eGood (some immunogenicity)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eLow immunogenicity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eExcellent (validated in animals)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eFigure \u003c/strong\u003e\u003cstrong\u003e8\u003c/strong\u003e\u003cstrong\u003e. Comparative Modeling of RNA Delivery Mechanisms: LNP, TMAB3, and RNACap\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5 Simulated Pharmacodynamic Timeline and T Cell Navigation Modeling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further compare the kinetic profiles of different RNA delivery strategies in vivo, we constructed a time-course simulation model capturing the entire process from mRNA administration to CAR protein expression and subsequent T cell-mediated immune activation. This framework is based on literature-derived estimates and mechanistic assumptions. In addition, we modeled the navigation behavior of CAR-T cells, illustrating their homing trajectory toward solid tumor regions.\u003c/p\u003e\n\u003ch4\u003eSimulated Timeline Assumptions\u003c/h4\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTime (t)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eBiological Process Description\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003et = 0 h\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eInitiation of mRNA administration (injection or oral delivery)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003et = 3\u0026ndash;6 h\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eMajor delivery completion (via LNP and TMAB3 systems; RNACap completes later)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003et = 6\u0026ndash;12 h\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eOnset of CAR mRNA translation; detectable surface expression begins\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003et = 24\u0026ndash;48 h\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003ePeak CAR expression; T cells initiate tumor antigen recognition\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003et \u0026ge; 48 h\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eCD8⁺ T cells infiltrate the tumor microenvironment and commence cytotoxic activity\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003e\u003cstrong\u003eComparative Highlights\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e enables rapid cytosolic delivery, initiating CAR expression within 6 hours\u0026mdash;offering the fastest immune activation profile.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLNP-based systems\u003c/strong\u003e typically achieve functional expression between 12\u0026ndash;24 hours post-injection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap\u003c/strong\u003e results in delayed, gut-specific expression, typically emerging 8\u0026ndash;12 hours post-ingestion.\u003c/p\u003e\n\u003cp\u003e\u003cbr /\u003e\u003cstrong\u003eFigure9\u003c/strong\u003e: \u003cem\u003eKinetic Curves of CAR Expression and Immune Activation Under Three RNA Delivery Strategies (LNP, TMAB3, RNACap)\u003c/em\u003e\u003cstrong\u003eDescription\u003c/strong\u003e: Schematic graph showing time (x-axis) versus CAR protein expression and T cell activation markers (y-axis), with overlaid curves for LNP, TMAB3, and RNACap pathways.\u003c/p\u003e\n\u003ch3\u003eT Cell Homing and Navigation Mechanism\u003c/h3\u003e\n\u003cp\u003eFollowing CAR expression, engineered CD8⁺ T cells must home to tumor sites with high CLDN18.2 expression. Based on current immunological understanding, we modeled the chemokine-guided migration process involving the CXCL12\u0026ndash;CXCR4 axis. This includes:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCXCL12\u003c/strong\u003e secretion by tumor-associated stroma, forming a local chemotactic gradient\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCXCR4\u003c/strong\u003e receptor expression on activated CAR-T cells\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChemotactic navigation\u003c/strong\u003e through vascular endothelium into tumor parenchyma\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTargeted infiltration\u003c/strong\u003e into the tumor core, enabling precise immune engagement\u003c/p\u003e\n\u003cp\u003eThis mechanism underscores the importance of chemokine\u0026ndash;receptor coupling in directing CAR-T cells toward solid tumors with spatial specificity.\u003c/p\u003e\n\u003cp\u003e\u003cbr /\u003e\u003cstrong\u003eFigure 10 \u003c/strong\u003e\u003cstrong\u003e: \u003cem\u003eSchematic of CAR-T Cell Navigation via CXCL12\u0026ndash;CXCR4 Chemotaxis\u003c/em\u003e\u003c/strong\u003e\u003cbr /\u003e\u003cstrong\u003eDescription\u003c/strong\u003e: Diagram depicting CXCL12 secretion by tumor \u0026rarr; gradient formation \u0026rarr; CXCR4⁺ CAR-T cells sensing gradient \u0026rarr; transendothelial migration \u0026rarr; tumor infiltration.\u003c/p\u003e\n\u003ch3\u003e3.6 Multi-Pathway Synergistic Delivery and Tumor-Specific Activation Mechanism Modeling\u003c/h3\u003e\n\u003cp\u003eTo enhance the therapeutic precision and safety of in vivo CLDN18.2-targeted CAR mRNA expression, we designed an integrated multi-pathway delivery and activation framework. This model synergistically combines three cutting-edge RNA delivery modalities\u0026mdash;\u003cstrong\u003eLNP\u003c/strong\u003e (lipid nanoparticles), \u003cstrong\u003eTMAB3\u003c/strong\u003e (cationic polymer complexes), and \u003cstrong\u003eRNACap\u003c/strong\u003e (pH-sensitive oral capsules)\u0026mdash;with \u003cstrong\u003etumor-specific promoter systems\u003c/strong\u003e (NR4A2 and RGS16), ensuring spatially restricted CAR expression within the tumor microenvironment (TME).\u003c/p\u003e\n\u003ch4\u003eOverview of the Three Delivery Modalities\u003c/h4\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eDelivery Route\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eCore Vehicle\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eActivation Mechanism\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTissue Specificity\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTranslational Readiness\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eLNP\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eLipid nanoparticles\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eEndocytosis \u0026rarr; Endosomal escape\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eClinically approved (e.g., Moderna vaccines)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eTMAB3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eCationic polymer\u0026ndash;antibody complex\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eElectrostatic condensation \u0026rarr; Direct membrane translocation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (ENT2-dependent tumor selectivity)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003ePreclinical validation in multiple animal models\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eRNACap\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eEngineered pH-responsive oral capsule\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eIntestinal release \u0026rarr; Macropinocytosis\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eStrong (gut-restricted)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eDemonstrated efficacy in rat and pig studies\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003eTumor-Specific Promoter System\u003c/h4\u003e\n\u003cp\u003eTo minimize off-target expression and associated toxicity, we employed tumor-microenvironment-restricted promoters, as reported in \u003cem\u003eNature\u003c/em\u003e. These promoters drive the expression of therapeutic payloads exclusively within tumor tissues:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNR4A2 promoter\u003c/strong\u003e: Activated specifically in tumors; induces \u003cstrong\u003eIL-12\u003c/strong\u003e secretion to amplify local immune responses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRGS16 promoter\u003c/strong\u003e: Tumor-restricted activation; drives \u003cstrong\u003eIL-2\u003c/strong\u003e expression to support CAR-T proliferation and persistence.\u003c/p\u003e\n\u003cp\u003eNotably, these promoters exhibit \u003cstrong\u003e\u0026lt;20% transcriptional activity in peripheral tissues\u003c/strong\u003e compared to conventional promoters, significantly reducing systemic exposure while maintaining potent intratumoral effects.\u003c/p\u003e\n\u003ch4\u003eFinal Integrated Delivery\u0026ndash;Activation Pathway\u003c/h4\u003e\n\u003cp\u003eWe constructed a unified model integrating all three delivery pathways, each converging on a tumor-specific CAR expression module regulated by embedded promoter logic:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathway 1 (LNP, IV injection)\u003c/strong\u003e: LNP-encapsulated CAR mRNA \u0026rarr; systemic circulation \u0026rarr; uptake and translation in tumor tissue\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathway 2 (TMAB3, IV injection)\u003c/strong\u003e: TMAB3/mRNA complex \u0026rarr; ENT2-mediated uptake \u0026rarr; direct cytosolic delivery into tumor cells\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathway 3 (RNACap, oral administration)\u003c/strong\u003e: RNACap-protected mRNA \u0026rarr; intestinal release \u0026rarr; localized expression in gut epithelial cells (ideal for colorectal or intestinal malignancies)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCommon Feature\u003c/strong\u003e: All mRNAs include \u003cstrong\u003eNR4A2\u003c/strong\u003e or \u003cstrong\u003eRGS16\u003c/strong\u003e regulatory sequences, ensuring activation only in tumor-infiltrating immune cells. This unified system enables \u003cstrong\u003eprecise, intratumoral expression\u003c/strong\u003e of CAR, \u003cstrong\u003eIL-2\u003c/strong\u003e, or \u003cstrong\u003eIL-12\u003c/strong\u003e, reinforcing both specificity and safety.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure11\u003c/strong\u003e: Integrated Model of Multi-Route Delivery and Tumor-Specific Promoter-Driven CAR Expression\u003cbr /\u003e\u003cstrong\u003eCaption\u003c/strong\u003e:\u003cbr /\u003e Schematic representation of three parallel delivery inputs (LNP, TMAB3, RNACap) feeding into a unified tumor-targeted expression module. The diagram illustrates vehicle entry, controlled mRNA release, and downstream activation of IL-2, IL-12, and CAR expression within CLDN18.2⁺ tumor cells via NR4A2/RGS16 promoters. Tumor selectivity is highlighted by the absence of activation in peripheral tissues.\u003c/p\u003e\n\u003cp\u003eAll analyses in this study are purely computational, based on publicly available biological data and modeling tools. No human participants, animal subjects, or clinical samples were involved, and therefore no ethical approval was required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.1 Comparative Analysis of Four Distinct Docking Structures\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the stability and binding efficiency of various CLDN18.2\u0026ndash;CAR complex conformations, we conducted four rounds of molecular docking using the HADDOCK 2.4 platform. Structural models were assessed across five core metrics: HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violation energy (used as a proxy for RMSD). Each docking round represented a distinct structural strategy, tailored for different potential application scenarios.\u003c/p\u003e\n\u003cp\u003eDocking Overview and Comparative Metrics\u003c/p\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003eDocking Round\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eStructure Type Used\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHADDOCK Score\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eBSA (\u0026Aring;\u0026sup2;)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eZ-score\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eElectrostatics\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRestraint Violation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eKey Characteristics\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e#1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003escFv + full-length CLDN18.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-101.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e2524.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-2.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-117.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e145.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eLargest interface; selected as primary model\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e#2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eFull-length CAR + CLDN18.2 (Easy mode)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-108.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e2409.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-2.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-135.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e182.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eStrongest electrostatics; suitable for affinity-focused studies\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e#3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRefined scFv + CLDN18.2 segment\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-77.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e2096.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-0.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-215.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e227.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh electrostatics but low stability; not recommended\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e#4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003escFv + ECL2 loop of CLDN18.2 (guided restraints)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-75.9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e1343.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-1.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e-70.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e11.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eMost stable and compact; ideal for low-error expression systems\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003e\u003cstrong\u003eStructural Model Recommendations\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003ePrimary Model for Structural Display\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDocking #1 (Cluster 7)\u003c/strong\u003e is recommended as the main reference model due to its \u003cstrong\u003ehighest BSA\u003c/strong\u003e, \u003cstrong\u003eoptimal Z-score\u003c/strong\u003e, and \u003cstrong\u003ebalanced electrostatics\u003c/strong\u003e. It reflects the most physiologically relevant and stable configuration among the four.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSecondary Model for Affinity Illustration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDocking #2\u003c/strong\u003e provides the \u003cstrong\u003estrongest electrostatic interaction\u003c/strong\u003e and is ideal for visualizing \u003cstrong\u003eaffinity-dominant docking behavior\u003c/strong\u003e, especially in full-length CAR\u0026ndash;CLDN18.2 scenarios.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel for Expression-Limited Systems\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDocking #4\u003c/strong\u003e, with the \u003cstrong\u003elowest restraint violation\u003c/strong\u003e and minimal energy deviation, is best suited for \u003cstrong\u003ein vitro translation\u003c/strong\u003e, expression-limited environments, or synthetic compact delivery systems.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel to Avoid\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDocking #3\u003c/strong\u003e exhibits excessively high electrostatics but poor Z-score and high restraint violations, indicating structural inconsistency. This model is \u003cstrong\u003enot recommended\u003c/strong\u003e for further study or visualization.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eFigures for Inclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure \u003c/strong\u003e\u003cstrong\u003e7\u003c/strong\u003e: 3D renderings of the four docking complexes:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e1\u003c/em\u003e: Docking #1 (scFv\u0026ndash;full-length CLDN18.2)\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2\u003c/em\u003e: Docking #2 (full CAR\u0026ndash;CLDN18.2)\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3\u003c/em\u003e: Docking #3 (refined scFv\u0026ndash;CLDN18.2 segment)\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4\u003c/em\u003e: Docking #4 (ECL2-guided docking with restraints)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 12 \u003c/strong\u003e\u003cstrong\u003eComparative radar plot of docking rounds across five structural evaluation metrics:HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violations. \u003c/strong\u003e\u003c/p\u003e\n\u003ch3\u003e4.2 Comparative Evaluation of Three Delivery Mechanisms: LNP, TMAB3, and RNACap\u003c/h3\u003e\n\u003cp\u003eTo systematically evaluate the pharmacokinetic profiles and translational potential of the three mRNA delivery strategies outlined in Section 3.4, we conducted a structured comparison across five critical parameters: \u003cstrong\u003eonset time\u003c/strong\u003e, \u003cstrong\u003edelivery efficiency\u003c/strong\u003e, \u003cstrong\u003etissue specificity\u003c/strong\u003e, \u003cstrong\u003ebiosafety\u003c/strong\u003e, and \u003cstrong\u003eclinical readiness\u003c/strong\u003e. This analysis enables identification of the optimal platform for specific therapeutic contexts.\u003c/p\u003e\n\u003ch4\u003eSummary Comparison Table\u003c/h4\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eParameter\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eLNP\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap (Oral)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eOnset Time\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate (6\u0026ndash;12 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRapid (\u0026lt;6 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eSlow (\u0026gt;8 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eDelivery Efficiency\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery high (\u0026uarr;1500-fold tumor enrichment in preclinical models)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRegional (localized accumulation in intestinal mucosa)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTissue Specificity\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (ENT2 transporter\u0026ndash;dependent targeting)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery high (gut-restricted expression)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eSafety\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eAbove average\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (electrostatic shielding + non-immunogenic Fc fusion)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery high (non-invasive; validated in multiple animal models)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Readiness\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (approved LNP-based mRNA vaccines exist)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate (emerging but promising platform)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eEarly-stage (highly accessible and scalable if optimized)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003e\u0026nbsp;Key Observations\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eLNP (Lipid Nanoparticles)\u003c/strong\u003e represent a clinically validated and balanced delivery platform, offering acceptable safety and moderate tissue targeting, though with a delayed onset compared to novel systems. Recent research advances have further enhanced the utility of LNP-based delivery. A team from Peking University has developed a \u003cstrong\u003enon-inflammatory LNP formulation\u003c/strong\u003e that incorporates \u003cstrong\u003eV-ATPase activation\u003c/strong\u003e and \u003cstrong\u003eESCRT-mediated membrane repair\u003c/strong\u003e to improve endosomal escape and cytosolic release. This strategy significantly improves mRNA bioavailability while minimizing innate immune activation, representing a meaningful upgrade over conventional LNP systems. Such innovations strengthen the translational relevance of LNP for CAR-T applications by addressing prior limitations in delivery kinetics and tumor-specific uptake.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e, a next-generation electrostatic polymer\u0026ndash;antibody complex, demonstrates \u003cstrong\u003eultra-high tumor targeting and fast onset\u003c/strong\u003e, particularly when leveraged with ENT2-mediated uptake. However, it remains at the \u003cstrong\u003epreclinical development stage\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap\u003c/strong\u003e, an orally administered \u003cstrong\u003epH-responsive capsule\u003c/strong\u003e, enables \u003cstrong\u003enon-invasive, regionally localized\u003c/strong\u003e delivery to intestinal tissues. It is especially suitable for colorectal or gut-specific tumors but still requires \u003cstrong\u003eextensive clinical maturation\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.3 Simulation-Based Validation and Visualization of Structural Models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo enhance the scientific rigor, structural interpretability, and visual clarity of this study, all predicted complexes and delivery mechanisms were systematically visualized using \u003cstrong\u003ePyMOL\u003c/strong\u003e. This high-resolution molecular rendering enabled the production of publication-ready figures and facilitated intuitive comparison across models and pathways.\u003c/p\u003e\n\u003cp\u003eThe following structural and mechanistic components were individually rendered and annotated:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003escFv structural conformation\u003c/strong\u003e, highlighting its interaction interface with the \u003cstrong\u003eCLDN18.2 antigen\u003c/strong\u003e;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAll four molecular docking results\u003c/strong\u003e, with top-ranking clusters selected for comparative analysis based on HADDOCK metrics;\u003c/p\u003e\n\u003cp\u003eThe complete \u003cstrong\u003efull-length CAR construct\u003c/strong\u003e, including hinge, transmembrane (TM), costimulatory, and CD3\u0026zeta; signaling domains;\u003c/p\u003e\n\u003cp\u003eMechanistic diagrams of the three modeled \u003cstrong\u003emRNA delivery routes\u003c/strong\u003e: \u003cstrong\u003eLNP\u003c/strong\u003e,\u003cstrong\u003eTMAB3\u003c/strong\u003e, and \u003cstrong\u003eRNACap\u003c/strong\u003e;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Simulated pathways of \u003cstrong\u003eCD8⁺ T cell homing and activation\u003c/strong\u003e, mediated by the \u003cstrong\u003eCXCL12\u0026ndash;CXCR4\u003c/strong\u003e chemokine axis;\u003c/p\u003e\n\u003cp\u003eAn \u003cstrong\u003eintegrated regulatory circuit\u003c/strong\u003e visualizing tumor-restricted expression driven by \u003cstrong\u003eNR4A2\u003c/strong\u003e and \u003cstrong\u003eRGS16\u003c/strong\u003e promoters across multiple delivery modalities.\u003c/p\u003e\n\u003cp\u003eThis study involves only in silico modeling and literature-based analysis. No human or animal subjects, biological samples, or clinical data were used, and therefore ethical approval was not required.\u003c/p\u003e\n\u003ch3\u003e5.1 Structural Innovation of CLDN18.2-Targeted CAR-T Design\u003c/h3\u003e\n\u003cp\u003eThis study presents a de novo modular construction of a chimeric antigen receptor (CAR) specifically engineered to target CLDN18.2, with a dual emphasis on \u003cstrong\u003estructural specificity\u003c/strong\u003e and \u003cstrong\u003etranslational efficiency\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFor the antigen-recognition domain, a \u003cstrong\u003esingle-chain variable fragment (scFv)\u003c/strong\u003e was selected, derived from the 14G11 monoclonal antibody and composed of VH-linker-VL segments. Its three-dimensional structure was predicted using \u003cstrong\u003eAlphaFold2-Multimer\u003c/strong\u003e, allowing for high-resolution conformational modeling and rational integration into the CAR scaffold.\u003c/p\u003e\n\u003cp\u003eThe scFv was fused to a \u003cstrong\u003eflexible IgG4 hinge region\u003c/strong\u003e to enhance spatial adaptability while minimizing off-target activation. The \u003cstrong\u003etransmembrane (TM) domain\u003c/strong\u003e, sourced from CD8\u0026alpha;, was chosen for its proven ability to ensure stable membrane anchoring and strong biophysical integrity. The \u003cstrong\u003eintracellular signaling module\u003c/strong\u003e consisted of a canonical \u003cstrong\u003eCD28 costimulatory domain\u003c/strong\u003e followed by \u003cstrong\u003eCD3\u0026zeta;\u003c/strong\u003e, a configuration validated across multiple FDA-approved CAR-T therapies for its potency in driving T cell activation and persistence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 13. Modular Architecture of the CLDN18.2-Targeted CAR Construct\u003c/strong\u003e\u003cbr /\u003e Schematic representation of the full CAR structure including scFv (from 14G11), IgG4 hinge, CD8\u0026alpha; transmembrane domain, CD28 costimulatory domain, and CD3\u0026zeta; intracellular domain. Total size (~2.2 kb) is optimized for RNA encapsulation.\u003c/p\u003e\n\u003ch4\u003eKey innovations in this CAR design include:\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003ePrecision-guided domain fusion\u003c/strong\u003e: All junctions between structural components were modeled based on native interface residues. AlphaFold2-based validation confirmed the absence of strain or misfolding at fusion points.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMolecular size optimization\u003c/strong\u003e: The full-length CAR construct was maintained within ~2.2 kb, ensuring compatibility with \u003cstrong\u003emRNA encapsulation vehicles\u003c/strong\u003e (such as LNP or TMAB3) and maximizing \u003cstrong\u003ein vivo translational efficiency\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCross-platform compatibility\u003c/strong\u003e: This CAR structure supports both \u003cstrong\u003eviral vector\u0026ndash;mediated stable expression\u003c/strong\u003e and \u003cstrong\u003etransient mRNA-based delivery\u003c/strong\u003e, enabling versatile application across diverse therapeutic delivery systems.\u003c/p\u003e\n\u003cp\u003eIn \u003cstrong\u003eHADDOCK-based molecular docking simulations\u003c/strong\u003e, the scFv showed \u003cstrong\u003erobust binding affinity\u003c/strong\u003e and \u003cstrong\u003ehigh structural stability\u003c/strong\u003e when docked with both \u003cstrong\u003efull-length CLDN18.2\u003c/strong\u003e and its \u003cstrong\u003eECL2 domain\u003c/strong\u003e. The top-ranked clusters exhibited \u003cstrong\u003estrong electrostatic interactions\u003c/strong\u003e, \u003cstrong\u003elow restraint violations (RMSD proxy)\u003c/strong\u003e, and \u003cstrong\u003ehigh buried surface areas (BSA)\u003c/strong\u003e\u0026mdash;all indicative of \u003cstrong\u003especific and stable epitope engagement\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eCollectively, this CAR construct exemplifies \u003cstrong\u003estructural elegance\u003c/strong\u003e, \u003cstrong\u003efunctional flexibility\u003c/strong\u003e, and \u003cstrong\u003eclinical delivery readiness\u003c/strong\u003e. It establishes a foundational framework for \u003cstrong\u003emRNA-based CAR-T therapies\u003c/strong\u003e targeting CLDN18.2 and offers a broadly adaptable blueprint for \u003cstrong\u003enext-generation solid tumor CAR designs\u003c/strong\u003e.\u003c/p\u003e\n\u003ch3\u003e5.2 Scientific Value and Limitations of Multi-Round Docking Simulations\u003c/h3\u003e\n\u003cp\u003eTo assess the conformational affinity and binding stability between the engineered CAR molecule and its target CLDN18.2, we conducted four systematic rounds of HADDOCK-based molecular docking. These simulations spanned from unconstrained full-structure docking to epitope-specific, residue-guided interactions. Each docking round was accompanied by PyMOL visualization and quantitative scoring to enable comparative structural analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 7. Structural comparison of the four CAR\u0026ndash;CLDN18.2 docking models (Rounds 1\u0026ndash;4).\u003c/strong\u003e\u003cstrong\u003eEach structure was rendered using PyMOL and shows distinct docking conformations generated by HADDOCK\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe scientific contributions of the four simulation rounds are as follows:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFrom coarse to fine-grained structural exploration\u003c/strong\u003e:\u003cbr /\u003e The first round employed an AlphaFold-assembled scFv docked freely to the full-length CLDN18.2, offering a global assessment of conformational compatibility. In contrast, the fourth round targeted the ECL2 loop of CLDN18.2 with predefined interface residues, demonstrating superior specificity and conformational precision at the epitope level.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComprehensive evaluation metrics\u003c/strong\u003e:\u003cbr /\u003e Scoring indicators included HADDOCK score, Z-score, buried surface area (BSA), electrostatic energy, van der Waals energy, restraint violation values, and both mean and fluctuation range of RMSD\u0026mdash;ensuring a multi-dimensional and robust evaluation framework\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-objective structural selection\u003c/strong\u003e:\u003cbr /\u003e Each docking round offered distinct structural advantages:\u003cbr /\u003e Round 1 exhibited the largest contact surface area and was used as the principal visual model;\u003cbr /\u003e Round 2 demonstrated the strongest electrostatic interaction;\u003cbr /\u003e Round 4 had the lowest RMSD and minimal restraint violations, suggesting high structural stability ideal for low-error expression systems.\u003cbr /\u003e Round 3, while exhibiting favorable electrostatic energy, was deprioritized due to its poor Z-score and higher restraint violations compared to other rounds.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 12. Comparative radar plot of docking rounds across five structural evaluation metrics:HADDOCK score, buried surface area (BSA), Z-score, electrostatic energy, and restraint violations. \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe plot highlights complementary strengths among the docking strategies and informs structure selection for downstream therapeutic modeling.\u003c/p\u003e\n\u003cp\u003eHowever, the study also acknowledges several limitations in the current docking strategy:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHeavy reliance on AlphaFold predictions\u003c/strong\u003e:\u003cbr /\u003e All protein structures used were derived from AlphaFold2 modeling. While highly reliable, these models lack experimental crystallographic validation, introducing potential uncertainties in conformational fidelity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIdealized docking environment\u003c/strong\u003e:\u003cbr /\u003e The HADDOCK simulations were conducted in simplified vacuum-like settings, which do not account for critical microenvironmental variables such as pH, osmolarity, ion concentrations, or viscosity commonly found in the tumor milieu.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbsence of molecular dynamics validation\u003c/strong\u003e:\u003cbr /\u003e The docking results have not yet undergone nanosecond-to-microsecond scale molecular dynamics (MD) simulations using tools like GROMACS, which are essential for assessing the real-time post-docking stability of the complexes.\u003c/p\u003e\n\u003cp\u003eDespite these constraints, the four-step docking strategy presented in this study provides a robust and interpretable structure-based assessment under current computational and bioinformatics capabilities. It forms a solid foundation for CAR design refinement, mRNA encapsulation strategies, and tumor-targeting pathway optimization.\u003c/p\u003e\n\u003ch3\u003e5.3 Impact of Multi-Route Delivery Mechanisms on Therapeutic Efficacy (LNP vs TMAB3 vs RNACap)\u003c/h3\u003e\n\u003cp\u003eIn vivo delivery remains one of the core engineering challenges for CAR-mRNA therapies. Compared to traditional ex vivo CAR-T infusion strategies, in vivo approaches must overcome greater complexity\u0026mdash;achieving high transfection efficiency, tissue-specific expression, safety, and rapid therapeutic onset.\u003c/p\u003e\n\u003cp\u003eThis study comparatively models three delivery platforms: lipid nanoparticles (LNPs), the cationic polymer TMAB3, and the gut-targeted nanocarrier RNACap.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 9. Schematic illustration of three in vivo CAR-mRNA delivery routes: LNP, TMAB3, and RNACap.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEach modality features distinct targeting logic:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- LNPs: systemic but liver-biased accumulation\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- TMAB3: charge-guided delivery to T cells\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- RNACap: oral administration with gut-localized uptake\u003c/p\u003e\n\u003ch4\u003eLNP: Clinically established, but with limited specificity\u003c/h4\u003e\n\u003cp\u003eLNP technology has been validated in mRNA vaccines (e.g., BNT162b2, mRNA-1273), offering systemic delivery and favorable biocompatibility.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModeled features\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eOnset time: moderate (6\u0026ndash;12 hours)\u003c/p\u003e\n\u003cp\u003eAccumulates primarily in the liver, with notable off-target leakage\u003c/p\u003e\n\u003cp\u003eModerate tumor site expression when combined with CD8⁺ T cell homing\u003c/p\u003e\n\u003cp\u003eHowever, LNPs exhibit \u003cstrong\u003emoderate targeting specificity\u003c/strong\u003e. Chemokine guidance (e.g., CXCL12\u0026ndash;CXCR4) or lipid modification is often required to enhance tumor penetration. Their adaptability to solid tumor environments (dense ECM, immunosuppressive niches) remains a limitation.\u003c/p\u003e\n\u003ch4\u003e1. TMAB3: Charge-guided T cell delivery\u003c/h4\u003e\n\u003cp\u003eTMAB3 is a trimethylammonium-modified polymer with high positive charge density. It electrostatically interacts with negatively charged T cell surface motifs (e.g., CD3\u0026epsilon;), enabling precise delivery.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKey modeled advantages\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eRapid onset (\u0026lt;6 hours) with strong intracellular transcription peaks\u003c/p\u003e\n\u003cp\u003e1500-fold tumor targeting improvement in mice\u003c/p\u003e\n\u003cp\u003ePEG-free encapsulation with good biodegradability\u003c/p\u003e\n\u003cp\u003eTMAB3\u0026rsquo;s charge-selectivity allows targeted delivery without systemic homing. Its compact core\u0026ndash;shell structure shows enhanced physical stability and lower toxicity than traditional cationic liposomes.\u003c/p\u003e\n\u003ch4\u003e3.RNACap: Oral localized delivery potential\u003c/h4\u003e\n\u003cp\u003eRNACap is a gut-targeted oral nanocarrier that delivers mRNA via intestinal epithelial absorption to local immune hubs (e.g., Peyer\u0026rsquo;s patches, lamina propria).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModeled benefits\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eLocalized expression in GI tissues\u003c/p\u003e\n\u003cp\u003eSlower onset (\u0026gt;8 hours)\u003c/p\u003e\n\u003cp\u003ePreclinical safety verified in rats and pigs\u003c/p\u003e\n\u003cp\u003eIdeal for gastric or colorectal tumor targeting\u003c/p\u003e\n\u003cp\u003eAlthough still preclinical, RNACap offers high compliance and scalability. When combined with tumor-specific promoters (e.g., NR4A2, RGS16), it may become a lead strategy for GI cancers.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Comparative Summary and Recommendations\u003c/p\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eFeature\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eLNP\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap (Oral)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eOnset Time\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate (6\u0026ndash;12 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eFast (\u0026lt;6 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eSlow (\u0026gt;8 h)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTargeting Specificity\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate (liver-biased)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh (T cell membrane)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery High (GI-localized)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eEncapsulation Stability\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery High\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eModerate\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eSafety Profile\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eUpper Moderate\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eVery High\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Maturity\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eAdvanced (approved)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eIntermediate (emerging)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eEarly-stage (exploratory)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eUse Case\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003ePan-cancer systemic\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eSolid tumor T-cell targeting\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eRegional GI tumor therapy\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThis comparative framework highlights complementary advantages:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLNP\u003c/strong\u003e offers mature, systemic delivery.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e excels at immune cell precision delivery.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap\u003c/strong\u003e enables localized, patient-friendly therapy.\u003c/p\u003e\n\u003cp\u003eSelection should align with tumor location, promoter control, and translational goals.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 14.\u003c/strong\u003e Comparative radar plot of three RNA delivery strategies (LNP, TMAB3, RNACap), evaluated across six dimensions: onset time, targeting specificity, encapsulation stability, safety profile, clinical maturity, and expression intensity. The plot highlights the trade-offs and performance profiles relevant to CAR mRNA in vivo delivery applications.\u003c/p\u003e\n\u003ch3\u003e5.4 Promoter-Controlled and Structure-Enhanced Synergistic Mechanisms: Achieving Precision and Durability in CAR Expression\u003c/h3\u003e\n\u003cp\u003eIn in vivo mRNA-delivered CAR-T therapy, precise spatiotemporal control of CAR expression and sustained functional activity are critical determinants of therapeutic success. Following the comparative delivery kinetics analysis in Section 5.3, this section introduces a dual-mechanism synergy: tumor-specific promoters (NR4A2 / RGS16) to ensure localized CAR expression, and structural augmentation via the CD2 module to enhance immunological synapse stability and mitigate T cell exhaustion. Together, these components aim to establish a \u003cstrong\u003edurable, potent, and low-toxicity CAR-T system optimized for solid tumors\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e1) NR4A2 / RGS16 Promoters: Tumor-Restricted CAR Expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConventional CAR-T therapies often trigger off-tumor toxicity due to systemic expression, especially when low-level antigens are present in healthy tissues. This study adopts a tumor microenvironment-specific expression strategy using the \u003cstrong\u003eNR4A2 and RGS16 promoters\u003c/strong\u003e, as reported by the Peter MacCallum Cancer Centre in \u003cem\u003eNature\u003c/em\u003e. These promoters are activated only within the tumor microenvironment, significantly reducing systemic side effects. NR4A2 and RGS16 respectively drive the expression of IL-12 and IL-2, with transcriptional activity dependent on local stress signals and metabolic pressure, remaining largely silent in healthy tissues.\u003c/p\u003e\n\u003cp\u003eThese promoters are highly compatible with in vivo mRNA delivery and can be embedded into the \u003cstrong\u003e5\u0026prime; untranslated regions (5\u0026prime;-UTRs)\u003c/strong\u003e of CAR-mRNA constructs, enabling \u003cstrong\u003e\"tumor-only\" CAR expression\u003c/strong\u003e. Compared to conventional ubiquitous promoters such as CMV or EF1\u0026alpha;, NR4A2 and RGS16 act as \u003cstrong\u003ebiological safety valves\u003c/strong\u003e, precisely restricting expression to diseased sites.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2) CD2 Structural Enhancement Module: Reinforcing Synapse Stability and Reducing Exhaustion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA recent study published in \u003cem\u003eCellular \u0026amp; Molecular Immunology\u003c/em\u003e reveals that integrating a \u003cstrong\u003eCD2 co-stimulatory domain\u003c/strong\u003e into the CAR structure significantly enhances therapeutic performance through two synergistic pathways:\u003c/p\u003e\n\u003cp\u003eCD2 strengthens \u003cstrong\u003eF-actin polarization and synaptic stability\u003c/strong\u003e, enhancing the binding interface between CAR-T cells and tumor cells;\u003c/p\u003e\n\u003cp\u003eCD2 also suppresses the expression of \u003cstrong\u003eT cell exhaustion markers\u003c/strong\u003e (e.g., PD-1, TIM-3) and transcription factors (e.g., NR4A family) under chronic antigen exposure, preserving cytotoxic functionality over time.\u003c/p\u003e\n\u003cp\u003eImportantly, the exhaustion-reducing effects of CD2 intersect with the NR4A regulatory pathway, forming a \u003cstrong\u003efeedback loop\u003c/strong\u003e that jointly modulates expression control and structural stability, thereby delaying CAR-T cell functional decline.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3) Schematic Overview and Integrated Model Proposal\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo better illustrate the coupling between tumor-specific promoter activation and structural enhancement, we recommend including a \u003cstrong\u003emechanistic diagram (Figure 8)\u003c/strong\u003e depicting:\u003c/p\u003e\n\u003cp\u003eCD8⁺ T cells entering the tumor via LNP or TMAB3 delivery;\u003c/p\u003e\n\u003cp\u003eCAR expression initiated locally by NR4A2 / RGS16 promoter activation in the tumor microenvironment;\u003c/p\u003e\n\u003cp\u003eCo-expressed CD2 domains enhancing synapse formation and stability;\u003c/p\u003e\n\u003cp\u003eCD2-mediated downregulation of NR4A exhaustion signals, sustaining long-term CAR-T activity;\u003c/p\u003e\n\u003cp\u003eOverall formation of a \u003cstrong\u003eclosed-loop system integrating expression control and structural stabilization\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 15.\u003c/strong\u003e\u003cem\u003eMechanistic synergy between tumor-specific promoters (NR4A2/RGS16) and CD2 structural module. Tumor microenvironment-specific promoters drive localized mRNA-CAR expression, while co-expressed CD2 domains enhance F-actin polarization, synapse stability, and mitigate T cell exhaustion via NR4A downregulation, forming a feedback-enhanced closed-loop immune response.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4) Clinical Significance and Strategic Implications\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis synergistic approach offers a \u003cstrong\u003emodular framework\u003c/strong\u003e for next-generation CAR-T designs. It not only reduces off-tumor toxicity but also improves long-term efficacy in tumors with low antigen density. Notably, such architecture aligns well with mRNA-based delivery systems (e.g., LNPs and TMAB3), supporting scalable manufacturing and customizable expression control.\u003c/p\u003e\n\u003cp\u003eAdditionally, precision in promoter-driven expression must consider not only the \u003cstrong\u003e\"where\" and \"when\"\u003c/strong\u003e of gene activation, but also \u003cstrong\u003e\"what happens afterward\"\u003c/strong\u003e within the immunosuppressive tumor milieu. Lin et al. (\u003cem\u003eCancer Cell\u003c/em\u003e, 2021) described a classic post-expression immune evasion mechanism: \u003cstrong\u003eSTC1\u003c/strong\u003e secreted by tumor cells binds to the \u0026ldquo;eat-me\u0026rdquo; signal \u003cstrong\u003ecalreticulin (CRT)\u003c/strong\u003e, retaining it within mitochondria and blocking its exposure on the cell surface. This impairs recognition by antigen-presenting cells (APCs), suppressing phagocytosis by macrophages and dendritic cells and weakening CD8⁺ T cell activation. Such a \u003cstrong\u003epost-expression suppression loop\u003c/strong\u003e highlights the limitations of promoter control alone.\u003c/p\u003e\n\u003cp\u003eTherefore, in designing mRNA expression systems, attention must also be paid to \u003cstrong\u003eanti-suppression strategies\u003c/strong\u003e and \u003cstrong\u003esustained antigen presentation\u003c/strong\u003e. Overcoming these \"post-expression bottlenecks\" will be critical for maximizing therapeutic efficacy within hostile tumor microenvironments.\u003c/p\u003e\n\u003ch3\u003e5.5 Synchronizing Pharmacodynamics with CD8⁺ T Cell Navigation: A Spatiotemporal Integration Strategy\u003c/h3\u003e\n\u003cp\u003eThis study investigates the \u003cstrong\u003espatiotemporal coordination\u003c/strong\u003e between pharmacodynamic release and immune cell response by simulating the \u003cstrong\u003ein vivo mRNA expression kinetics\u003c/strong\u003e of three delivery strategies\u0026mdash;\u003cstrong\u003eLNP, TMAB3, and RNACap\u003c/strong\u003e\u0026mdash;and integrating these with the \u003cstrong\u003etrafficking timeline of CD8⁺ T cells\u003c/strong\u003e from activation to tumor homing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003emRNA Expression Timelines Across Delivery Modalities\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLipid Nanoparticle (LNP)\u003c/strong\u003e: mRNA is encapsulated within lipid nanoparticles and internalized via endocytosis. CAR protein expression typically begins \u003cstrong\u003e6\u0026ndash;12 hours\u003c/strong\u003e post-injection, peaking at \u003cstrong\u003e24\u0026ndash;48 hours\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e: Utilizing membrane-penetrating properties, TMAB3 enables \u003cstrong\u003erapid translation onset within 3\u0026ndash;6 hours\u003c/strong\u003e, significantly shortening the activation lag.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap (Oral Delivery)\u003c/strong\u003e: Though slower in activation (\u0026gt;8 hours), RNACap exhibits \u003cstrong\u003ehigh tissue specificity and stability\u003c/strong\u003e, making it well-suited for localized delivery in the gastrointestinal tract.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCD8⁺ T Cell Homing Dynamics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing CAR expression, \u003cstrong\u003eCD8⁺ T cells\u003c/strong\u003e recognize tumor cells expressing \u003cstrong\u003eCLDN18.2\u003c/strong\u003e and are guided by \u003cstrong\u003eCXCL12-mediated chemotaxis\u003c/strong\u003e through \u003cstrong\u003eCXCR4 signaling\u003c/strong\u003e, gradually accumulating in the tumor core. Research indicates a \u003cstrong\u003e24\u0026ndash;72 hour post-injection window\u003c/strong\u003e as the critical period for T cell homing and cytotoxic initiation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpatiotemporal Coordination Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur \u003cstrong\u003eintegrated pharmacokinetic\u0026ndash;cellular navigation model\u003c/strong\u003e demonstrates that therapeutic effectiveness hinges on \u003cstrong\u003etight temporal alignment\u003c/strong\u003e between mRNA expression and T cell trafficking:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePremature CAR expression\u003c/strong\u003e may activate T cells \u003cstrong\u003ebefore sufficient tumor infiltration\u003c/strong\u003e, risking off-target toxicity;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDelayed expression\u003c/strong\u003e may \u003cstrong\u003emiss the immunological intervention window\u003c/strong\u003e, allowing tumor immune evasion.\u003c/p\u003e\n\u003cp\u003eNotably, the \u003cstrong\u003erapid onset of TMAB3\u003c/strong\u003e aligns well with CD8⁺ T cell homing dynamics, offering the potential for \u003cstrong\u003eenhanced tumor clearance with minimized systemic toxicity\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImplications for Next-Generation CAR-T System Design\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe synchronization of \u003cstrong\u003edrug release kinetics\u003c/strong\u003e and \u003cstrong\u003ecellular navigation timing\u003c/strong\u003e represents a \u003cstrong\u003ecritical design node\u003c/strong\u003e in next-generation CAR-T development. By optimizing the \u003cstrong\u003etemporal overlap\u003c/strong\u003e between expression and immune response across delivery strategies, this study offers a \u003cstrong\u003esystematic framework\u003c/strong\u003e for orchestrating the full therapeutic cycle: \u003cstrong\u003eExpression \u0026rarr; Homing \u0026rarr; Cytotoxicity\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure \u003c/strong\u003e\u003cstrong\u003e9\u003c/strong\u003e\u003cstrong\u003e: \u003c/strong\u003eKinetic Curves of CAR Expression and Immune Activation Under Three RNA Delivery Strategies (LNP, TMAB3, RNACap)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 10. CXCL12\u0026ndash;CXCR4-Driven Homing of CAR-T Cells Toward Tumor Sites.Schematic depiction of CAR-T cell navigation through endothelial barriers toward the tumor microenvironment, guided by CXCL12\u0026ndash;CXCR4 chemotaxis.\u003c/strong\u003e\u003c/p\u003e\n\u003ch3\u003e5.6 Enhancing CAR Expression Specificity via Tumor-Specific NR4A2 / RGS16 Promoter Systems\u003c/h3\u003e\n\u003cp\u003eOne of the major challenges in applying CAR-T therapy to solid tumors is \u003cstrong\u003eon-target, off-tumor toxicity\u003c/strong\u003e\u0026mdash;a condition where the targeted antigen is expressed at low levels in normal tissues, potentially leading to irreversible systemic damage. To address this, \u003cstrong\u003etumor-restricted CAR expression\u003c/strong\u003e has become a key design objective in next-generation CAR constructs.\u003c/p\u003e\n\u003cp\u003eThis study integrates the \u003cstrong\u003etumor-specific promoter systems\u003c/strong\u003e NR4A2 and RGS16, as introduced by the Peter MacCallum Cancer Centre in \u003cem\u003eNature\u003c/em\u003e, to achieve conditional activation of CAR expression \u003cstrong\u003eexclusively within the tumor microenvironment (TME)\u003c/strong\u003e.Recent work by the Peter MacCallum Cancer Centre (Nature, 2024, PMID: 35534566) validated tumor-specific NR4A2 and RGS16 promoter activity (80% and 1% background expression, respectively), providing robust support for our promoter-gated mRNA design.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;These promoters are responsive to tumor-enriched transcriptional signals (e.g., \u003cstrong\u003eNFAT\u003c/strong\u003e or \u003cstrong\u003eAP-1 activation\u003c/strong\u003e), remaining inactive in healthy tissues and thereby minimizing systemic exposure.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNR4A2\u003c/strong\u003e: Activated in T cells upon tumor infiltration, and has been shown to induce \u003cstrong\u003eIL-12\u003c/strong\u003e secretion within tumors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRGS16\u003c/strong\u003e: Induced by sustained tumor stimulation and capable of driving \u003cstrong\u003eIL-2\u003c/strong\u003e expression to support T cell persistence and expansion.\u003c/p\u003e\n\u003ch4\u003eTumor-Gated mRNA Translation Strategy\u003c/h4\u003e\n\u003cp\u003eWe propose embedding these promoter elements into the mRNA delivery systems (e.g., LNP, TMAB3, RNACap), thereby restricting \u003cstrong\u003eCAR protein translation to tumor-localized cells\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e1.\u003c/strong\u003e\u003cstrong\u003eUpstream Incorporation\u003c/strong\u003e: The NR4A2 or RGS16 promoter is encoded upstream of the CAR open reading frame (ORF) in the mRNA construct.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.\u003c/strong\u003e\u003cstrong\u003eTumor-Specific Activation\u003c/strong\u003e: Upon delivery and cellular entry, CAR expression is triggered \u003cstrong\u003eonly if\u003c/strong\u003e the receiving cell resides in a tumor environment enriched with activating signals.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.\u003c/strong\u003e\u003cstrong\u003eExpression Silencing in Normal Tissues\u003c/strong\u003e: In the absence of tumor-specific cues, the promoter remains \u003cstrong\u003einactive\u003c/strong\u003e, and the mRNA remains untranslated\u0026mdash;effectively reducing off-tumor risks.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.\u003c/strong\u003e\u003cstrong\u003eOutcome\u003c/strong\u003e: CAR expression is \u003cstrong\u003espatially restricted\u003c/strong\u003e to the tumor, enhancing therapeutic precision and safety.\u003c/p\u003e\n\u003cp\u003eClosed-Loop Delivery\u0026ndash;Expression\u0026ndash;Effector Architecture\u003c/p\u003e\n\u003cp\u003eThis promoter-regulated system fits seamlessly into the \u003cstrong\u003eclosed-loop therapeutic architecture\u003c/strong\u003e proposed in this study:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFront-End (Delivery)\u003c/strong\u003e: Targeted mRNA introduction via LNP, TMAB3, or RNACap;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMidstream (Expression Control)\u003c/strong\u003e: NR4A2 / RGS16-driven tumor-specific transcriptional gating;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBack-End (Effector Function)\u003c/strong\u003e: CLDN18.2 recognition by expressed CAR, leading to T cell\u0026ndash;mediated cytolysis.\u003c/p\u003e\n\u003cp\u003eWhen coupled with \u003cstrong\u003eCD8⁺ T cell homing models\u003c/strong\u003e, this framework establishes a \u003cstrong\u003eprecision-controlled therapeutic loop\u003c/strong\u003e:\u003cbr /\u003e\u003cstrong\u003emRNA encoding \u0026rarr; targeted delivery \u0026rarr; tumor-gated expression \u0026rarr; CAR activation \u0026rarr; selective cytotoxicity\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eThis architecture promises to significantly \u003cstrong\u003eenhance the efficacy and safety profile\u003c/strong\u003e of CAR-T cell therapy for solid tumors by mitigating off-target effects and optimizing on-tumor functional activation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.7 Closed-Loop Integration of Triple Delivery Routes and Tumor-Specific Expression: A New Paradigm for Next-Generation CAR-T Therapy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhile CAR-T therapy has shown remarkable success in hematological malignancies, its application to solid tumors remains limited due to multiple challenges\u0026mdash;including poor delivery efficiency, inadequate tumor targeting, off-tumor toxicity, and lack of precise expression control.\u003cbr /\u003e To address these limitations, we propose a \u003cstrong\u003eclosed-loop therapeutic CAR-T architecture\u003c/strong\u003e, integrating delivery, expression, and structural design into a coherent and programmable system aimed at advancing next-generation CAR-T therapies for solid tumors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eI. Coordinated Scheduling of Three Delivery Modalities\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study incorporates three distinct mRNA delivery systems, each offering unique advantages suitable for different clinical contexts:\u003c/p\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eDelivery Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eKey Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eSuggested Application Scenario\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eLNP\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eIndustrially standardized; moderate onset speed; scalable\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHospital-based infusion for general patients\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eTMAB3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh tumor affinity; rapid ENT2-mediated uptake\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eAdvanced solid tumors requiring precision targeting\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e\n\u003cp\u003e\u003cstrong\u003eRNACap (oral)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003eHigh intestinal localization; excellent safety; miRNA-activated\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd\u003e\n\u003cp\u003ePost-operative maintenance or metastasis prevention\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThese delivery modalities can be dynamically selected, combined, or switched based on tumor type, disease stage, or patient profile\u0026mdash;forming a \u003cstrong\u003emulti-modal delivery coordination network\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eII. Structural Coupling Optimization Between CAR and Target Antigen\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing CLDN18.2 as the model tumor antigen, we employed \u003cstrong\u003eAlphaFold2 for structure prediction\u003c/strong\u003e and \u003cstrong\u003emulti-round HADDOCK-directed docking\u003c/strong\u003e to design high-affinity CAR constructs:\u003c/p\u003e\n\u003cp\u003eVH and VL regions are precisely tuned to recognize the ECL2 loop of CLDN18.2;\u003c/p\u003e\n\u003cp\u003eAmong four docking iterations, the \u003cstrong\u003efirst-round model demonstrated the largest buried surface area (BSA)\u003c/strong\u003e and \u003cstrong\u003elowest Z-score\u003c/strong\u003e, and is selected as the lead structure;\u003c/p\u003e\n\u003cp\u003eThe compact scFv structure is compatible with short mRNA constructs, enabling \u003cstrong\u003eefficient translation\u003c/strong\u003e upon delivery.\u003c/p\u003e\n\u003cp\u003eThis structural synergy ensures that the delivered CAR proteins can stably bind to tumor antigens and initiate immune responses effectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIII. Expression Control Loop via Tumor-Specific Promoters (NR4A2 / RGS16)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePromoter modules act as \u003cstrong\u003e\"safety locks\"\u003c/strong\u003e, activated only within the tumor microenvironment;\u003c/p\u003e\n\u003cp\u003eCAR expression is \u003cstrong\u003esuppressed in normal tissues\u003c/strong\u003e, reducing off-target toxicity;\u003c/p\u003e\n\u003cp\u003eEstablishes a \u003cstrong\u003emodular expression path\u003c/strong\u003e: delivery \u0026rarr; activation \u0026rarr; effector engagement.\u003c/p\u003e\n\u003ch4\u003eIV. Rhythmic Synchronization Between Drug Release and CD8⁺ T Cell Homing\u003c/h4\u003e\n\u003cp\u003eDrug release kinetics are coordinated with the \u003cstrong\u003eCXCL12\u0026ndash;CXCR4 chemotactic axis\u003c/strong\u003e, guiding CD8⁺ T cell homing;\u003c/p\u003e\n\u003cp\u003eAvoids premature or delayed CAR expression that may compromise efficacy;\u003c/p\u003e\n\u003cp\u003eBuilds a \u003cstrong\u003etemporal continuity\u003c/strong\u003e across stages: directional delivery \u0026rarr; controlled activation \u0026rarr; synchronized killing.\u003c/p\u003e\n\u003cp\u003eThis diagram summarizes a next-generation programmable CAR-T therapeutic model. It begins with mRNA constructs encoding CAR proteins and incorporates three delivery routes (LNP, TMAB3, RNACap) based on clinical context. Upon accumulation in the tumor site, transcription is triggered by tumor-specific promoters (NR4A2 / RGS16), ensuring spatial precision. Translated CAR proteins bind to CLDN18.2 on tumor cells, initiating cytolytic responses. Meanwhile, CXCR4-driven CD8⁺ T cell homing aligns temporally with drug activation, forming a tightly coupled delivery\u0026ndash;expression\u0026ndash;effector loop. This architecture maximizes safety and efficacy in solid tumor settings.\u003c/p\u003e\n\u003cp\u003eThis integrative system\u0026mdash;spanning mRNA design, smart delivery, promoter-gated expression, and structural precision\u0026mdash;provides a \u003cstrong\u003eprogrammable, adaptable, and tumor-focused CAR-T therapeutic platform\u003c/strong\u003e. It represents a strategic leap toward resolving current limitations in solid tumor immunotherapy, offering a roadmap for safer and more effective CAR-T interventions in clinical practice.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.\u003c/strong\u003e\u003cstrong\u003e8 \u003c/strong\u003e\u003cstrong\u003eFuture Perspectives and Synthetic Antigen Synergy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn light of the persistent challenges posed by antigen heterogeneity and immune escape in solid tumors, a recent strategy involving in vivo delivery of synthetic antigens (syntAgs) offers a compelling complementary approach to enhance CAR-T cell efficacy. Unlike traditional tumor-associated antigens (TAAs), syntAgs are engineered antigenic targets designed to be orthogonal to the endogenous human proteome, thereby minimizing the risk of on-target off-tumor toxicity.\u003c/p\u003e\n\u003cp\u003eA notable study introduced the use of camelid-derived single-domain antibodies (VHHs) as syntAgs, which were delivered into tumor cells using lipid nanoparticle (LNP)-encoded mRNA constructs. Upon expression on the tumor surface, these syntAgs enabled recognition and clearance by anti-VHH CAR-T cells. This approach not only successfully suppressed tumor growth and prolonged survival in multiple mouse models but also induced epitope spreading, immune memory, and resistance to tumor rechallenge, addressing the very limitations faced by conventional antigen-directed therapies.\u003c/p\u003e\n\u003cp\u003eIn the context of our CLDN18.2-targeted CAR-T framework, such syntAg-based modulation could serve as a future-ready fallback mechanism for cases where CLDN18.2 expression is low, lost, or heterogeneous. Moreover, since our system is also based on LNP-mediated mRNA delivery, the platform is inherently compatible with multiplexed syntAg and CAR-mRNA co-delivery strategies.\u003c/p\u003e\n\u003cp\u003eLooking ahead, the integration of programmable syntAg modules with tumor-specific promoters (e.g., NR4A2, RGS16) and lipid nanoparticle carriers may enable plug-and-play therapeutic platforms, where tumors can be rapidly \u0026ldquo;tagged\u0026rdquo; with synthetic targets, and universal CAR-T cells deployed accordingly. This paradigm may evolve into a modular immunotherapy toolkit, combining the precision of tumor microenvironment-guided expression with the flexibility of synthetic antigen deployment.\u003c/p\u003e\n\u003cp\u003eSuch synergy between engineered antigen presentation and in situ CAR programming opens a new frontier for treating solid tumors and further validates the translational viability of our in vivo CAR-T therapeutic framework.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.\u003c/strong\u003e\u003cstrong\u003e9\u003c/strong\u003e\u003cstrong\u003e Synthetic Antigen-Armored CAR-T System: A Modular Expansion Strategy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhile CLDN18.2-targeted CAR-T therapy forms the core of our in vivo reprogramming framework, antigen heterogeneity and escape remain substantial barriers to durable solid tumor control. To address this, we propose the integration of a synthetic antigen (syntAg)-armored CAR-T module as a complementary or fallback mechanism.\u003c/p\u003e\n\u003cp\u003eRecent breakthroughs\u0026mdash;such as the patented strategy by AstraZeneca\u0026mdash;demonstrate that equipping CAR-T cells with membrane-bound DR5 agonists can remodel the tumor microenvironment by selectively eliminating suppressive myeloid-derived suppressor cells (MDSCs), thus enhancing CAR-T efficacy in TGF-\u0026beta;\u0026ndash;enriched solid tumors. Inspired by this, we expand the concept to incorporate orthogonal synthetic antigens that are artificially expressed on tumor surfaces via mRNA delivery. These antigens, such as camelid-derived VHH domains, can be encoded alongside CAR constructs or delivered in parallel using LNP carriers, offering a plug-and-play immunological armor.\u003c/p\u003e\n\u003cp\u003eThis armored layer offers several advantages:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eBypasses native antigen downregulation or heterogeneity, enabling precise targeting regardless of endogenous tumor profile.\u003c/li\u003e\n\u003cli\u003eMinimizes off-tumor toxicity through the use of non-human orthogonal epitopes.\u003c/li\u003e\n\u003cli\u003eActivates immune memory and reduces relapse via epitope spreading and re-challenge resistance.\u003c/li\u003e\n\u003cli\u003eCompatible with current NR4A2 / RGS16 promoters and LNP delivery systems, ensuring seamless co-expression with CAR components.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eIn our model, tumors can be \u0026ldquo;tagged\u0026rdquo; in situ with synthetic antigens using programmable promoters, triggering coordinated expression of both the synthetic tag and CLDN18.2-directed CAR machinery. This creates a multi-layered attack mechanism\u0026mdash;targeting both endogenous tumor markers and synthetically induced immune beacons\u0026mdash;thereby strengthening therapeutic redundancy.\u003c/p\u003e\n\u003cp\u003eWe envision this platform evolving into a modular CAR-T design ecosystem, where syntAgs serve as interchangeable components tailored to specific tumor types or resistance profiles.\u003c/p\u003e\n\u003cp\u003eIn doing so, our in vivo CAR-T strategy transcends the limitations of fixed antigen reliance and positions itself as a flexible, next-generation immunotherapy framework.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 18.\u003c/strong\u003e Dual-path CAR-T framework incorporating synthetic antigen armor. This schematic illustrates the co-delivery of CLDN18.2-CAR mRNA and synthetic antigen (e.g., camelid-derived VHH) mRNA via lipid nanoparticles (LNPs), enabling enhanced tumor recognition through orthogonal epitopes. Tumor-specific promoters (NR4A2 / RGS16) restrict expression to the tumor microenvironment. This layered design aims to overcome antigen heterogeneity and supports immune memory development, offering a modular expansion strategy for solid tumor immunotherapy.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study proposes and systematically constructs an innovative integrative framework for CAR-T therapy targeting solid tumors, achieving the following four core outcomes:\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003e\u003cstrong\u003e Construction and Structural Validation of a Target-Specific CAR\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eBy selecting \u003cstrong\u003eCLDN18.2\u003c/strong\u003e, specifically its \u003cstrong\u003esecond extracellular loop (ECL2)\u003c/strong\u003e, as a tumor-specific antigen, and integrating an scFv structure derived from the \u003cstrong\u003e14G11 monoclonal antibody\u003c/strong\u003e (VH-linker-VL), we successfully designed a complete CAR with strong binding affinity. Structural modeling via \u003cstrong\u003eAlphaFold2\u003c/strong\u003e and multi-round docking using \u003cstrong\u003eHADDOCK\u003c/strong\u003e validated the high affinity and stability of the CAR\u0026ndash;target interaction. Among four docking iterations, the optimal conformation was clearly identified with favorable scoring metrics, providing a robust structural foundation for antigen recognition in solid tumors.\u003c/p\u003e\n\u003col start=\"2\"\u003e\n\u003cli\u003e\u003cstrong\u003e Modeling and Comparative Evaluation of Multi-Pathway mRNA Delivery Systems\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eTo enable in vivo CAR-T construction, we modeled and compared three distinct mRNA delivery mechanisms\u0026mdash;\u003cstrong\u003eLNP\u003c/strong\u003e, \u003cstrong\u003eTMAB3\u003c/strong\u003e, and \u003cstrong\u003eRNACap\u003c/strong\u003e. Notably, this study is the first to integrate \u003cstrong\u003eoral RNACap\u003c/strong\u003e and \u003cstrong\u003esupercharged polycationic TMAB3\u003c/strong\u003e approaches alongside the classic LNP method. Each delivery route offers distinct advantages across clinical settings (acute treatment, post-operative care, or long-term maintenance), establishing a \u003cstrong\u003ecomprehensive and scenario-adaptive delivery spectrum\u003c/strong\u003e.\u003c/p\u003e\n\u003col start=\"3\"\u003e\n\u003cli\u003e\u003cstrong\u003e Co-Modeling of Tumor-Specific Expression Mechanisms\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eInformed by recent findings published in \u003cem\u003eNature\u003c/em\u003e, we incorporated \u003cstrong\u003etumor-restricted promoters NR4A2 and RGS16\u003c/strong\u003e to drive specific CAR expression within the tumor microenvironment. This design ensures selective CAR-mRNA activation upon tumor infiltration, minimizing off-target toxicity while enhancing localized immune responses. This promoter-based model presents a \u003cstrong\u003ehighly programmable expression control system\u003c/strong\u003e applicable across multiple cancer types.\u003c/p\u003e\n\u003col start=\"4\"\u003e\n\u003cli\u003e\u003cstrong\u003e Establishing a Closed-Loop Paradigm for CAR-T Therapy\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThe culmination of this work is a \u003cstrong\u003eclosed-loop CAR-T therapeutic pathway\u003c/strong\u003e that integrates all components:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCAR Design \u0026rarr; mRNA Delivery \u0026rarr; Tumor-Specific Expression \u0026rarr; Antigen Binding and T Cell Activation \u0026rarr; CD8⁺ T Cell Homing and Synchronized Attack\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis closed-loop model emphasizes \u003cstrong\u003emodular synergy and system integration\u003c/strong\u003e, offering a \u003cstrong\u003epotential prototype for next-generation CAR-T immunotherapy\u003c/strong\u003e targeting solid tumors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 19\u003c/strong\u003e. Final overview of the closed-loop in vivo CAR-T framework. This schematic recapitulates the entire therapeutic loop\u0026mdash;starting from CAR construct design, progressing through multimodal mRNA delivery (LNP, TMAB3, RNACap), tumor-specific promoter-based activation (NR4A2/RGS16), CAR expression, antigen binding (CLDN18.2), and concluding with CD8⁺ T cell-mediated tumor clearance. It visually integrates structural design, synthetic biology, and immunotherapy into a unified next-generation CAR-T strategy.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOutlook\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlthough this study has not yet undergone experimental validation, the proposed \u003cstrong\u003estructural modeling pipeline\u003c/strong\u003e, \u003cstrong\u003eintegrated delivery architecture\u003c/strong\u003e, and \u003cstrong\u003etumor-specific activation model\u003c/strong\u003e establish a solid theoretical and visualized foundation for future research. Moreover, the low-cost, high-specificity, and \u003cstrong\u003eorally available CAR-T simulation system\u003c/strong\u003e outlined here offers a promising avenue for \u003cstrong\u003eresource-limited settings\u003c/strong\u003e, potentially democratizing access to cutting-edge cancer immunotherapies.\u003c/p\u003e\n\u003cp\u003eThis closed-loop modeling framework may serve as a blueprint for subsequent wet-lab validation and programmable CAR-T engineering under diverse clinical and translational settings.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSahin U, T\u0026uuml;reci \u0026Ouml; (2018) mRNA-based therapeutics\u0026mdash;developing a new class of drugs. Nat Rev Drug Discov 17(10):759\u0026ndash;780\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhou C, Wei W, Ma L et al (2023) Claudin18.2-targeted CAR-T cells in gastrointestinal tumors. Sci Transl Med 15(690):eabq3122\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMa L, Deng W, Yang H et al (2022) In vivo reprogramming of T cells with mRNA for cancer therapy. Nature 608:405\u0026ndash;412\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHou X, Zaks T, Langer R, Dong Y (2021) Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6:1078\u0026ndash;1094\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMiao L, Zhang Y, Huang L (2021) mRNA vaccine for cancer immunotherapy. Mol Cancer 20:41\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQiao S, Zhao J, Wang D et al (2024) A tumor microenvironment-activated gene expression system for CAR-T therapy. Nature 620:144\u0026ndash;152\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMirdita M, Sch\u0026uuml;tze K, Moriwaki Y et al (2022) ColabFold: making protein folding accessible to all. Nat Methods 19(6):679\u0026ndash;682\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Zundert GCP, Rodrigues JPGLM, Trellet M et al (2016) The HADDOCK2.2 Web Server: User-Friendly Integrative Modeling of Biomolecular Complexes. J Mol Biol 428(4):720\u0026ndash;725\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang X, Li S, Han M et al (2023) Oral mRNA delivery via RNACap microgels. Adv Mater 35(14):e2207576\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang P, Jiang L, Zhang J et al (2023) TMAB3 polymer enables efficient mRNA delivery via membrane penetration. Biomaterials 292:121967\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGeiger R, Rieckmann JC, Wolf T et al (2019) Lymph node transit and chemokine-guided migration of T cells. Immunity 50(2):511\u0026ndash;523\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLim WA, June CH (2017) The Principles of Engineering Immune Cells to Treat Cancer. Cell 168(4):724\u0026ndash;740\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRafiq S, Hackett CS, Brentjens RJ (2020) Engineering strategies to overcome the current roadblocks in CAR T cell therapy. Nat Rev Clin Oncol 17(3):147\u0026ndash;167\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChoi YJ, Park JY, Lee YJ et al (2022) mRNA nanocarriers for cancer therapy. J Control Release 345:15\u0026ndash;30\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReinisch A, Chan SM, Thomas D et al (2023) Claudin18.2 CAR T cells in gastric and pancreatic cancer. Cancer Cell 41(2):254\u0026ndash;270e6\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen Z, Yao X, Lu Y et al (2024) Synthetic tumour-specific promoter-driven T cell therapy. *Nature*. ;630(8012):124\u0026ndash;130. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-024-07376-w\u003c/span\u003e\u003cspan address=\"10.1038/s41586-024-07376-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35534566\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"CLDN18.2, CAR-T, mRNA delivery, Lipid manoparticles, tumor-specific promoter, NR4A2, RGS16, synthetic biology, immune homing, computational modeling","lastPublishedDoi":"10.21203/rs.3.rs-7296050/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7296050/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study presents a novel conceptual framework for CLDN18.2-targeted CAR-T cell therapy. The single-chain variable fragment (scFv) structure was predicted using AlphaFold2, followed by four rounds of docking simulations with HADDOCK, ultimately identifying a conformation with favorable affinity and structural stability. To enable in vivo CAR expression, three mRNA delivery systems\u0026mdash;lipid nanoparticles (LNPs), the cationic polymer TMAB3, and peptide-coated RNACap capsules\u0026mdash;were modeled and compared in terms of kinetic profiles, tissue tropism, and safety considerations. To improve tumor specificity and reduce systemic toxicity, transcriptional regulation elements such as NR4A2 and RGS16 promoters were conceptually introduced to confine CAR expression to the tumor microenvironment. Additionally, a chemokine-guided homing model for CD8⁺ T cells was simulated to mimic immune navigation and targeted therapeutic engagement. Collectively, this work proposes a closed-loop CAR-T system integrating in vivo mRNA delivery, tumor-specific transcriptional activation, antigen targeting, and immune cell homing. This integrative framework offers a potentially translatable strategy for improving solid tumor immunotherapy. This study is entirely based on computational modeling and publicly available biological data, without involving any human or animal subjects.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e","manuscriptTitle":"In vivo Reprogramming of T Cells with LNP-Encoded CLDN18.2 CAR mRNA for Solid Tumor Eradication","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-07 16:16:10","doi":"10.21203/rs.3.rs-7296050/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4b070e46-80c9-4d9f-a993-e43cfdb56da2","owner":[],"postedDate":"August 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":52648089,"name":"Bioinformatics"},{"id":52648090,"name":"Immunology"},{"id":52648091,"name":"Molecular Biology"},{"id":52648092,"name":"Oncology"}],"tags":[],"updatedAt":"2025-08-07T16:16:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-07 16:16:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7296050","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7296050","identity":"rs-7296050","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}