An immune-based predictive model for HBV clearance: validation in multicenter cohorts and mechanistic insights from in vivo studies.

Background: Chronic HBV infection is a major risk factor for hepatocellular carcinoma, posing a significant global health burden. However, predictive models for HBV clearance based on immune biomarkers remain limited.

Methods: We systematically developed a predictive tool by quantifying mRNA expression levels of CD4⁺ T-cell subset transcription factors, cytokines, and immune checkpoints in PBMCs from chronic HBV patients and resolved HBV individuals using RT-qPCR. A binary logistic regression model was constructed in the training cohort, with performance evaluated by ROC and calibration curves, followed by internal and external validation in independent cohorts. For in vivo validation, an HBV-transfected mouse model was established via rapid tail vein injection of pGL3-CP-Fluc-HBV1.2_C2 plasmid. Outcomes included body weight, HBsAg/HBV DNA levels, and luciferase activity. Kaplan-Meier analysis assessed cumulative clearance rates, while RT-qPCR tracked model-related mRNA dynamics in PBMCs.

Results: The model identified GATA3, FOXP3, IFNG, TNF, and HAVCR2 as key genes, demonstrating robust predictive accuracy for HBV clearance. Dose-specific temporal patterns of immune gene regulation were observed, revealing distinct immunomodulatory mechanisms between groups.

Conclusion: This study establishes a reliable immune-based predictive model for HBV clearance and highlights divergent immune responses in chronic versus resolved infection.