Spatial transcriptomics reveals prognostically LYZ⁺ fibroblasts and colocalization with FN1⁺ macrophages in diffuse...

Diffuse large B-cell lymphoma (DLBCL) is a clinically heterogeneous malignancy with diverse patient outcomes, largely influenced by the tumor microenvironment (TME). Understanding the roles of fibroblasts and macrophages within the TME is essential for developing personalized therapeutic strategies in DLBCL.

This study is a multi-omics approach, integrating spatial transcriptomics (n = 11), bulk transcriptomics (n = 2,499), immunohistochemistry (IHC, n = 37), multiplex immunofluorescence (mIF, n = 56), and plasma samples (n = 240) to identify and characterize fibroblast and tumor-associated macrophage subtypes in the TME. Hub genes for LYZ⁺ fibroblasts and FN1⁺ macrophages were selected through univariate Cox regression and random forest analyses. Their prognostic significance was validated using IHC, mIF, and autoantibody assays in DLBCL patients treated with R-CHOP and in non-small cell lung cancer (NSCLC) patients receiving immune checkpoint inhibitors (ICIs).

Fibroblasts and macrophages were classified into two distinct subtypes. Patients with higher LYZ⁺ fibroblasts infiltration demonstrated superior prognosis, which was associated with increased infiltration of FN1⁺ macrophages. Key hub genes identified for LYZ⁺ fibroblasts included LYZ, ANPEP, CSF3R, C15orf48, LILRB4, CLEC7A, and COL7A1, while hub FN1⁺ macrophages genes included COL1A1, FN1, APOE, DCN, MMP2, SPP1, COL3A1, and COL1A2. Independent prognostic markers in DLBCL treated with R-CHOP and NSCLC treated with ICIs were identified, including LYZ and LILRB4 at both protein and mRNA levels, and COL1A2 autoantibodies (p < 0.05). In DLBCL patients treated with R-CHOP, FN1 mRNA and autoantibody levels were also prognostic markers (p < 0.05). In NSCLC treated with ICIs, COL3A1 autoantibody was prognostic marker (p < 0.05).

This study identified a prognostically relevant LYZ⁺ fibroblasts and FN1⁺ macrophages in DLBCL. The hub genes associated with these subtypes represent potential biomarkers, providing insights into improving patient outcomes in DLBCL.