从相关到因果：SOX11 驱动一个协调的免疫抑制网络，抑制三阴性乳腺癌中的抗原提呈

Background: Triple-negative breast cancer (TNBC) is an aggressive subtype with no targeted therapies and limited response to immunotherapy. TNBC evades immune surveillance by repressing antigen presentation machinery (APM) genes, reducing MHC Class I expression. Using bioinformatics, we identified SOX11 as a potential transcriptional regulator associated with APM suppression and immune evasion in TNBC. Methods: We analyzed gene expression data from basal breast cancer (BRCA) tumors (n = 189) in TCGA and cancer cell lines (n = 1,189) in DepMap. Spearman correlation coefficients were used to evaluate associations between a T-cell infiltration signature and an APM signature. SOX11 was prioritized as it was strongly negatively correlated with both signatures. To identify downstream effectors, we performed RNA-seq analysis on SOX11 knockdown (KD) CAL-148 cells (high endogenous SOX11) and integrated these results with publicly available SOX11-overexpression and KD datasets. To validate key targets, SOX11 KD and MEX3A KD were performed using siRNA, followed by flow cytometry to measure surface MHC Class I (HLA-ABC) expression. Results: SOX11 was strongly inversely correlated (rho < -0.4, adjusted p < 0.01) with APM and T-cell infiltration signatures in basal BRCA and DepMap datasets. In basal BRCA tumors, SOX11 expression was enriched in T-cell cold tumors (p < 0.01). Analysis of our SOX11 KD RNA-seq data, integrated with public SOX11-modulated datasets, revealed a highly reproducible SOX11-driven gene signature. This signature was notably enriched for known suppressors of innate immunity. Among the most consistent SOX11 targets were genes implicated in destabilizing HLA-A (MEX3B) , degrading innate immune sensors (MEX3A) , activating immunosuppressive signaling (SBK1) , and facilitating epigenetic silencing of interferon pathways and APM (RCOR2, NELL2) . Consistent with this, SOX11 KD in CAL-148 cells significantly increased surface HLA-ABC expression. Furthermore, targeted knockdown of its effector, MEX3A , also resulted in a significant increase in APM gene expression. Conclusions: SOX11 suppresses HLA-A/B/C expression and contributes to immune evasion in TNBC. Its high expression in T-cell cold tumors and its regulation of multiple APM-suppressive targets highlight its potential as a therapeutic target. We hypothesize that SOX11-mediated APM repression is not driven by a single gene but by the concerted activity of this multi-effector network. Targeting SOX11 or its key effectors could restore antigen presentation, promote immune recognition, and enhance immunotherapy responses in TNBC.