Tertiary lymphoid structure drives allograft rejection via IFN-γ-JAK-STAT-dependent atypical memory B cell differentiation

Lijie Ma¹Shiming Gong¹Jianchen Fang²Mingxuan Feng^1*Zhenzhen Zhan^1*Qiang Xia^1*

• ¹Department of Liver Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China
• ²Shanghai Jiao Tong University School of Medicine Affiliated Renji Hospital, Shanghai, China

Background: Tertiary lymphoid structure (TLS) is a neogenized, ectopic lymphoid aggregate found in infected, autoimmune and tumour tissues with an activated adaptive immune response. However, a comprehensive understanding of the pathological role, function, and formation of TLS in allograft rejection remains incomplete. Methods: We enrolled two large retrospective cohorts of liver biopsy (LB) after pediatric living donor liver transplantation (LDLT) and developed a deep learning pathomics (DLP) model. Gene expression profiles and corresponding clinical information of 590 cases were enrolled from three transcriptomic databases, including cohort-GSE193135 (n=337), cohort-GSE145780 (n=235), cohort-Renji (n=18). ESTIMATE, CIBERSORT, XCELL and MCP analyses were performed to visualize the immune landscape. Single-cell RNA-sequencing (scRNA-seq) analysis of 11 LBs after LDLT and multiplexed immunohistochemistry (mIHC) were performed to validate the discoveries of bioinformatics analysis. Results: We provided evidence that increased TLS in the liver was closely correlated with allograft rejection, fibrosis, and declined liver function. ScRNA-seq and in vitro co-culture analysis revealed that TLS form through the differentiation of atypical memory B (AtM B) cells via the JAK-STAT signalling pathway, stimulated by IFN-γ from exhausted CD8+ T effector memory (TEM) cells. The presence of TLS resulted in IgG accumulation, initiating pathological antibody-dependent cell-mediated phagocytosis (ADCP) of apoptotic hepatocytes by CD68+ macrophages. Preclinically, blocking JAK1/3 or knocking out Lta in mice limited TLS formation and attenuated allograft rejection in mouse orthotopic liver transplantation models, informing novel therapeutics for allograft rejection pathology. Conclusion: We proposed an efficient DLP model for predicting allograft rejection, and revealed an unexpected immunological mechanism of TLS in allograft rejection livers and clarified an IFN-γ-JAK-STAT-dependent circuit that could be targeted with drugs and transformed AtM B cells into potent instigators of hepatocellular injury in allograft rejection.