Fig 2: The miRNA–target gene and target gene–TF regulatory network. (A) Network diagram of miRNAs regulating TOX or RasGRP3. Bright red represents TOX or RasGRP3, green represents miRNAs. (B) Network diagram of TFs regulating 23 significant DEGs. Purple circles indicate 23 significant DEGs, green octagons indicate TFs with targets ≥ 10. (C) Network diagram of TFs simultaneously regulating TOX, RasGRP3, and TSPAN13. Purple circles indicate TOX, RasGRP3, or TSPAN13, green octagons indicate TFs.

Fig 3: HE staining and validation of the microarray data by IHC. (A) Representative figures of HE staining in group A without coronary stenosis. (B) Representative figures of HE staining in group B with coronary stenosis less than 25%. (C) Negative control of TOX in group A. (D) Negative control of TOX in group B. (E) Representative IHC figures of TOX in group A. (F) Representative IHC figures of TOX in group B. Compared to group A, the expression of TOX in group B was significantly upregulated (p = 0.031). (G) Negative control of RasGRP3 in group A. (H) Negative control of RasGRP3 in group B. (I) Representative IHC figures of RasGRP3 in group A. (J) Representative IHC figures of RasGRP3 in group B. Compared to group A, the expression of RasGRP3 in group B was significantly downregulated (p = 0.032). (K) Negative control of TSPAN13 in group A. (L) Negative control of TSPAN13 in group B. (M) Representative IHC figures of TSPAN13 in group A. (N) Representative IHC figures of TSPAN13 in group B. Compared to group A, the expression of TSPAN13 in group B was significantly downregulated (p = 0.046). Magnification of HE: 200×. Magnification of IHC: 400×. bar = 50 μm.

Fig 4: An overview of the spatial-omics framework to interrogate differences between EBV-positive and EBV-negative cHL TME(A) Study overview to spatially dissect the EBV-positive and EBV-negative cHL TME. (1) Biopsies from the initial cohort of EBV-positive (number of patients = 6, number of FOVs = 19) and EBV-negative (number of patients = 14, number of FOVs = 23) patients were collected, (2) stained with 30 lanthanide-conjugated antibodies, and acquired by the MIBI spatial proteomics platform to (3) generate multiplexed images for the identification of cell phenotypes and neighborhoods, where upon (4) computational analysis illustrated key differences between EBV-positive and EBV-negative cHL TME.(B) Representative MIBI images of lymph node (left), EBV-positive (middle), and EBV-negative (right) sections, with markers for nuclei (Histone H3), M2-like macrophages (CD163), HRS cells (Pax5lo), dendritic cells (CD11c), programmed death-ligand 1 (PD-L1), memory T cells (CD45RO), CD4 T cells (CD4), CD8 T cells (CD8), and T-cell-dysfunction-associated transcription factor (Tox). Scale bars: 100 μm.

Fig 5: Distinct immune components and T cell dysfunction features between EBV-positive and EBV-negative cHL TME(A) Relative Z score expression levels of phenotypic markers (left) for the annotated cell phenotypes in this MIBI dataset; a mean expression visualization is in Figure S3A. Representative MIBI images of EBV-positive (middle, top) and EBV-negative cHL FOVs (right, top), with markers for nuclei (Histone H3), T (CD3), CD4 T (CD4), CD8 T (CD8), B (Pax5hi), and HRS (Pax5lo) cells, along with the corresponding EBV-positive (middle, bottom) and EBV-negative cell phenotype maps (right, bottom). Phenotype maps for the other FOVs are in Figure S2. Scale bars: 400 μm (main) and 50 μm (inset).(B) Relative proportions of annotated cell types across EBV-positive and EBV-negative cHL FOVs.(C) Log2 fold enrichment plot of annotated cell types between EBV-positive and EBV-negative cHL FOVs. Significance stars are only shown for significant comparisons. Two-sided Wilcoxon tests were conducted for all cell type proportions; test results were adjusted for multiple comparisons using the Benjamini-Hochberg (BH) method with a targeted false discovery rate (FDR) at 0.05. Unadjusted p value and BH corrected test results are in Table S4.(D) Relative proportion infiltration of CD4 T, CD8 T, and HRS cells across all FOVs.(E) Top: relative expression of the MHC class I (B2M and HLA1) on annotated cell types, apart from antigen-presenting cells and HRS cells, in the cHL TME. Bottom: Relative expression of MHC class I and MHC class II (HLA-DR) on antigen-presenting cells and HRS cells in the cHL TME. Representative MIBI images of MHC class I differences are in Figure S3E.(F) Top: relative expression of memory (CD45RO) and T cell dysfunction (Tox, Lag3, and PD-1) markers on T cell populations, with the relative cell counts inscribed above. Bottom: relative expression of PD-L1 on antigen-presenting cells (DCs and macrophages) and HRS cells, with the cell counts inscribed above; both Z score and mean expression representations are here. Statistical test results for (E) and (F) are in Figures S3C and S3F. Mean expression visualization for (E) and (F) heatmaps are in Figures S3D and S3H. All statistical tests were performed by comparing EBV-positive to EBV-negative populations, and unadjusted p value and BH corrected test results are in Table S4. Significant stars: ∗p ≤ 0.05, ∗∗p ≤ 0.01.