Fig 2: SELP-knockdown inhibits tumor growth and prolongs survival in human GB mouse models.A SELP knockdown reduced tumor growth rate of PD-GB4 tumors in mice compared to control WT (control) or shNC GB tumors. Tumor volume was calculated using Radiant software. Data represent mean ± s.e.m. N = 8 control, 9 shNC, and 14 shSELP. One-way ANOVA, Dunn's method, p < 0.001. B Kaplan–Meier curve showing prolonged survival of shSELP PD-GB4 tumor-bearing mice compared to control WT and shNC. N = 8 control, 9 shNC, and 11 shSELP. P values were determined using two sided log rank test. C Representative T1 weighted MRI images of PD-GB4 tumors following Gd-DTPA administration, detected at day 22 post tumor inoculation. Representative images and quantification of immunostaining for proliferating cells (Ki-67), activated microglia (Iba1), and blood vessels (CD31) in the tumors showing reduced proliferation and blood vessel density, and activated microglia in shSELP tumors. Data represent mean ± s.d. N = 5 images per mouse. The graphs show data from a representative mouse per group out of two mice per group. Statistical significance was determined using one-way ANOVA test with multiple comparisons adjustment. Scale bars represent 100 µm. D Tumor growth of control WT (control), shNC, and shSELP U251 tumors in SCID mice demonstrating delayed tumor growth in shSELP U251 tumors. Tumors were detected by MRI imaging (MR solutions, T2 weighted). Tumor volume was calculated using Radiant software. Data represent mean ± s.e.m. N = 9 control, 8 shNC, and 9 shSELP. One-way ANOVA, Dunn’s method, p < 0.001. E SELP knockdown prolonged the survival of U251 tumor-bearing mice compared to control WT and shNC. N = 7 control, 6 shNC, and 6 shSELP. P values were determined using two-sided log rank test. F Representative T2 weighted MRI images of U251 tumors detected at day 19 post tumor inoculation. Representative images and quantification of immunostaining for proliferating cells (Ki-67), activated microglia (Iba1), and blood vessels (CD31) in the tumors showing reduced proliferation and blood vessel density, and activated microglia in shSELP tumors. Data represent mean ± s.d. Each dot represents the average of three images per mouse. N = 3 mice per group. Statistical significance was determined using one-way ANOVA test with multiple comparisons adjustment. Scale bars represent 100 µm. Source data are provided as a Source Data file.

Fig 3: SELP inhibits the microglia pro-inflammatory phenotype and promotes their immunosuppressive activity.A ELISA assay showing higher secretion of SELP by human microglia treated with PD-GB CM compared to naïve DMEM. Data represent mean ± s.d. Each dot represents a triplicate. The graph shows the average of three independent repeats. Statistical significance was determined using an unpaired, two-sided Student’s t test. B Real-time PCR showing higher expression SELP mRNA by human microglia treated with PD-GB4 CM compared to naïve DMEM. Data represent mean ± s.d. Each dot represents a triplicate. The graph shows the average of three independent repeats. Statistical significance was determined using an unpaired, two-sided Student’s t test. Representative image (C) and quantification (D) of flow cytometry analysis of PSGL-1 expression by human microglia, showing higher expression when treated with PD-GB4 GB CM compared to naïve DMEM. Data represent mean ± s.d. The graph shows the average of five independent experiments. Statistical significance was determined using an unpaired, two-sided Student’s t test. E tSNE plot of single-cell RNA-seq analysis of microglia and macrophages isolated from patients' GB tumors, showing the expression of PSGL-1 by microglia compared to macrophages. Data obtained from Darmanis et al.56. Treatment with rSELP resulted in elevated mRNA expression of ARG1 (F), reduced expression of iNOS (G) and elevated expression of IL-10 (H) and TGF-ß (I) by human microglia. SELPi or anti-PSGL-1 neutralizing antibody rescued rSELP effects while anti-CD44 or CD24 neutralizing antibody did not affect gene expression. Data represent mean ± s.d. Each dot represents a triplicate. The graphs show the average of three independent studies. Statistical significance was determined using one-way ANOVA test with multiple comparisons adjustment. Phagocytic activity (J) and NO release (K) by human microglia were reduced when treated with rSELP, and were restored when inhibiting SELP or PSGL-1 but not when neutralizing CD44 or CD24. Data represent mean ± s.d. of four wells per group. The graphs are representative of three independent repeats. Scale bars represent 200 µm. Statistical significance was determined using one-way ANOVA test with multiple comparisons adjustment. Source data are provided as a Source Data file.

Fig 4: Perturbation of the SELP-PSGL-1 interactions causes tumor cells to exhibit reduced tumorigenesis and the microglia/macrophages cells to express higher score of neurodegenerative and antigen presentation signatures.A Unsupervised clustering of the single-cell RNA-seq profiles of different cell types populations of interest identified in GL261 tumors. B Clusters distribution between the shNC (NC) and shSELP groups within the tumor cell population. C Heat map showing the most up-regulated differentially expressed genes in each cluster within the tumor cells population. D Representative genes from the tumor cells clusters and comparison of gene signature scores showing down-regulation of the invasion, proliferation, and angiogenesis signatures in shSELP tumors compared to shNC. Center of the box plots shows median values, boxes extent from 25% to the 75% percentile, whiskers show minimum and maximum values. Statistical significance was determined using an unpaired, two-sided Student’s t test. E Unsupervised clustering of the single-cell RNA profiles of microglia/macrophage cells. F Bar graphs showing the number of cells present in each cluster from microglia/macrophages isolated from shNC versus shSELP GL261 tumors. G Heat map showing the ten most up-regulated differentially expressed genes in each cluster within microglia/macrophages population. H Projection of a neurodegenerative-associated microglia signature42 and bar plots showing up-regulation in shSELP tumors compared to shNC (p value = 2.2 × 10-16, two-sided Wilcoxon rank sum test and p value = 4.4 × 10-9 CERNO enrichment test, see “Methods”). I Projection of an antigen presentation and chemokines/cytokines microglia signature and bar plots showing up-regulation in shSELP tumors compared to shNC (p value = 8.969 × 10-16, two-sided Wilcoxon rank sum test and p value = 6.4 × 10-8 CERNO enrichment test, see “Methods”). The contour marks the region of highly scored cells by taking into account only cells with a signature score above the 10th percentile.

Fig 5: SELP-knockdown inhibits tumor growth and prolongs survival in a murine GB mouse model.A Tumor growth of control WT (control), shNC, and shSELP GL261 tumors in C57BL/6 mice demonstrating delayed tumor growth in the shSELP group. Tumors were detected by MRI imaging (MR solutions, T1 weighted). Tumor volume was calculated using Radiant software. Data represent mean ± s.e.m. N = 29, two independent experiments. One-way ANOVA, Dunn’s method, P < 0.001. B Kaplan–Meier curve showing prolonged survival of shSELP GL261 tumor bearing mice compared to control WT and shNC. N = 24 Control, 18 shNC, 15 shSELP, two independent experiments. p values were determined using two-sided log rank test. C Representative T1 weighted MRI images of GL261 tumors following Gd-DTPA administration, detected at day 19 post tumor inoculation. Representative images and quantification of immunostaining for proliferative cells (Ki-67), apoptotic cells (caspase-3) and blood vessels (CD31). The results demonstrate reduced proliferation, blood vessel density and increased apoptosis in shSELP tumors. Data represent mean ± s.d. Each dot represents the average of three images per mouse. N = 3 mice per group. Statistical significance was determined using one-way ANOVA test with multiple comparisons adjustment. Scale bars represent 100 µm. Source data are provided as a Source Data file.