Tumor cell–derived IFN spatially reprograms osteopontin-enriched macrophage niches to promote PARP inhibitor resistance
Dan Liu, Kangjia Tao, Cheng Xu, Wen Yang, Chujun Cai, Cui Feng, Kairong Xiong, Sisi Wu, Yaying Lin, Zikun Peng, Jianhua Chi, Wen Pan, Qing Zhong, Jiahao Liu, Xiong Li, Xingzhe Liu, Dongchen Zhou, Ding Ma, Guang-Nian Zhao, Yu Xia, Yong Fang, Qinglei Gao
Dan Liu, Kangjia Tao, Cheng Xu, Wen Yang, Chujun Cai, Cui Feng, Kairong Xiong, Sisi Wu, Yaying Lin, Zikun Peng, Jianhua Chi, Wen Pan, Qing Zhong, Jiahao Liu, Xiong Li, Xingzhe Liu, Dongchen Zhou, Ding Ma, Guang-Nian Zhao, Yu Xia, Yong Fang, Qinglei Gao
View: Text | PDF
Research Article Immunology Oncology

Tumor cell–derived IFN spatially reprograms osteopontin-enriched macrophage niches to promote PARP inhibitor resistance

Abstract

Poly (ADP-ribose) polymerase inhibitors (PARPis) benefit homologous recombination-deficient (HRD) malignancies, yet resistance remains a major challenge. Leveraging specimens from a prospective neoadjuvant niraparib monotherapy trial in treatment-naive, high-grade serous ovarian cancer, we integrated PhenoCycler-Fusion spatial profiling, scRNA-Seq, and multiplex immunohistochemistry to identify 2 therapeutic-modulated cellular neighborhoods: an IFN+ tumor cell–enriched niche that expands in resistant lesions and a niche enriched in tumor-associated macrophage (TAM) that persists but acquires enhanced immunosuppressive features. Mechanistically, sustained tumor cell–derived IFN induced osteopontin (SPP1) expression in TAMs via STAT signaling, creating immunosuppressive niches enriched in Tregs and myofibroblastic cancer–associated fibroblasts with intensified cell-cell interactions. SPP1 directly suppressed T cell signaling and effector function. High baseline SPP1+ cells predicted lower response rate (30.0% vs. 76.2%; P = 0.021) and shorter progression-free survival (median 13.5 vs. 28.3 months; P = 0.0006). In HRD mouse models, SPP1 blockade restored PARPi sensitivity, reversed acquired resistance, and enhanced T cell cytotoxicity—effects abrogated in immunodeficient mice, confirming immune dependence. These data establish a spatial IFN-SPP1 axis whereby persistent tumor cell IFN reprograms TAMs to promote PARPi resistance, position SPP1 as a key therapeutic target and prognostic biomarker for this therapy, and underscore therapeutic potential of microenvironment-targeted strategies to overcome PARPi resistance.

Authors

Dan Liu, Kangjia Tao, Cheng Xu, Wen Yang, Chujun Cai, Cui Feng, Kairong Xiong, Sisi Wu, Yaying Lin, Zikun Peng, Jianhua Chi, Wen Pan, Qing Zhong, Jiahao Liu, Xiong Li, Xingzhe Liu, Dongchen Zhou, Ding Ma, Guang-Nian Zhao, Yu Xia, Yong Fang, Qinglei Gao

×

Figure 2

IFN-responding macrophages are enriched in resistant lesions.

Options: View larger image (or click on image) Download as PowerPoint
IFN-responding macrophages are enriched in resistant lesions. (A) Integr...
(A) Integrated Uniform manifold approximation and projection (UMAP) visualization with heatmap showing color-coded cell subtype annotation of epithelial (epi) cells, TAMs and DCs, and CD4+ T cells derived from PCF analysis; right-side panel displays z-scored marker expression confirming lineage-specific signatures. (B) Quantitative box plots representing the quantified percentages of specific cell subsets in pretreatment (n = 25), post-niraparib responder (R) (n = 21) and post-niraparib nonresponder (NR) (n = 18) clinical specimens. (C) Heatmaps illustrating the log2-fold changes (Log2FC) in cell abundance between post-niraparib R (n = 21) versus pretreatment (n = 25) (left), post-niraparib NR (n = 18) versus pretreatment (n = 25) (middle), and post-niraparib NR (n = 18) versus R (n = 21) (right) across all cell subtypes. (D) UMAP projection of scRNA-Seq data annotated into major cell types. (E) Proportion quantification of TAM in pretreatment (n = 17), post-niraparib R (n = 8), and post-niraparib NR (n = 5) clinical specimens. (F and G) UMAP plots of major cell types with Milo differential abundance testing results overlaid in color gradient, depicting log fold changes of post-niraparib R (F) or NR (G) relative to pretreatment (Pre) samples. (H) Heatmap illustrating the log2 fold changes in cell abundance of specific TAM subclusters across 3 comparisons (R versus Pre; NR versus Pre; NR versus R). Adjacent columns display the gene set variation analysis (GSVA) scores for M1 and M2 polarization signatures across identified TAM subtypes. Box plot shows quantification of the stromal proportion of the IFN-responding TAM cluster across pretreatment (n = 17), post-niraparib R (n = 8), and post-niraparib NR (n = 5) clinical specimens. (I) Box plot quantifying the proportion of the indicated cell subsets across pretreatment (n = 17), post-niraparib R (n = 8), and post-niraparib NR (n = 5) clinical specimens. Wilcoxon rank-sum test. Statistical significance was determined by 2-tailed Wilcoxon rank-sum test, and the Padj values were calculated by Benjamini-Hochberg correction (B, C, E, and H). *Padj < 0.05, **Padj < 0.01, ***Padj < 0.001.