狼疮性肾炎
免疫学
发病机制
效应器
自身免疫
系统性红斑狼疮
肾炎
免疫抑制
自身抗体
免疫系统
医学
表型
生物标志物
肾
疾病
信号转导
生物
炎症
肾脏疾病
先天免疫系统
免疫耐受
周边公差
细胞因子
表位
补体系统
巨噬细胞
红斑狼疮
免疫复合物
癌症研究
足细胞
自身免疫性疾病
遗传倾向
作者
Yachao Li,Xinyi Fang,Xiang Gao,Chi Liu
标识
DOI:10.1080/08820139.2026.2699369
摘要
BACKGROUND: The PD-1/PD-L1 axis is a key immune checkpoint that maintains peripheral tolerance by restraining T-cell activation. In lupus nephritis (LN), this pathway has long been viewed as protective, based on Pdcd1-deficient mice developing lupus-like nephritis and PD-L1 fusion protein efficacy in mouse models. However, recent clinical findings reveal a paradox: PD-1⁺ T cells - especially CD8⁺ and CD4⁺ subsets - are greatly expanded in blood and kidneys of LN patients, correlating with disease activity, proteinuria, and renal decline. OBJECTIVE: To reconcile the protective and pathogenic roles of PD-1/PD-L1 signaling and propose a unifying "signal balance model" for precision immunomodulation. FINDINGS: These PD-1⁺ cells are not exhausted but exhibit an activated, clonally expanded effector phenotype (granzyme B, perforin, IFN-γ) that drives tissue injury, while PD-1⁺ Tph and Tfh cells promote autoantibody production. The biphasic nature of PD-1/PD-L1 signaling arises from T-cell subset heterogeneity, inflammatory cytokines (IL-12, IFN-α) that override inhibition, costimulatory signals, and reverse signaling through PD-L1 on renal parenchymal cells. CONCLUSIONS: We propose that the net outcome - protection versus pathogenesis - depends on the relative weights of these four determinants, further influenced by genetic background (e.g., PDCD1 polymorphisms). PD-1 agonism (rosnilimab, PD-L1-Fc, nanoparticle delivery) offers promise, but the pathogenic activity of PD-1⁺ effector T cells argues against a uniform approach. Our framework provides a basis for patient stratification, biomarker development, and combination strategies, ultimately guiding a shift from broad immunosuppression to tailored therapy in LN and informing future translational research.
科研通智能强力驱动
Strongly Powered by AbleSci AI