MET overexpression contributes to STAT4-PD-L1 signaling activation associated with tumor-associated, macrophages-mediated immunosuppression in primary glioblastomas

Abstract

Background: Dysregulated receptor tyrosine kinases, such as the mesenchymal-epidermal transition factor (MET), have pivotal role in gliomas. MET and its interaction with the tumor microenvironment have been previously implicated in secondary gliomas. However, the contribution of MET gene to tumor cells' ability to escape immunosurveillance checkpoints in primary gliomas, especially in glioblastoma (GBM), which is a WHO grade 4 glioma with the worst overall survival, is still poorly understood.

Methods: We investigated the relationship between MET expression and glioma microenvironment by using multiomics data and aimed to understand the potential implications of MET in clinical practice through survival analysis. RNA expression data from a total of 1243 primary glioma samples (WHO grades 2-4) were assembled, incorporating The Cancer Genome Atlas, Chinese Glioma Genome Atlas, and GSE16011 data sets.

Results: Pearson's correlation test from the three data sets indicated that MET showed a robust correlation with programmed death-ligand 1 (PD-L1) and STAT pathways. Western blot analysis revealed that in GBM cell lines (N33 and LN229), PD-L1 and phosphorylated STAT4 were upregulated by MET activation treatment with hepatocyte growth factor and were downregulated on MET suppression by PLB-1001. Tumor tissue microarray analysis indicated a positive correlation between MET and PD-L1 and macrophage-associated markers. Chromatin immunoprecipitation-PCR assay showed enrichment of STAT4 in the PD-L1 DNA. Transwell co-culture and chemotaxis assays revealed that knockdown of MET in GBM cells inhibited macrophage chemotaxis. Moreover, we performed CIBERSORTx and single-cell RNA sequencing data analysis which revealed an elevated number of macrophages in glioma samples with MET overexpression. Kaplan-Meier survival analysis indicated that activation of the MET/STAT4/PD-L1 pathway and upregulation of macrophages were associated with shorter survival time in patients with primary GBM.

Conclusions: These data indicated that the MET-STAT4-PD-L1 axis and tumor-associated macrophages might enforce glioma immune evasion and were associated with poor prognosis in GBM samples, suggesting potential clinical strategies for targeted therapy combined with immunotherapy in patients with primary GBM.

Keywords: biomarkers; brain neoplasms; gene expression profiling; immunity; tumor.

Figure 1

Correlation between MET and immune checkpoints in RNA expression level in primary gliomas. (A) MET expression is positively associated with expression of PD-L1 in TCGA and CGGA data sets. (B) The scatter diagrams show the coexpression patterns of MET and PD-L1/PD1/TIM-3. (C) With the increase of MET expression, the expression of PD-L1 increased gradually. (D) MET/PD-L1 expression positively correlated with STAT pathway in the TCGA data set. Corrgrams are derived according to Pearson’s r value between MET/PD-L1 and three pathways (STAT, AKT, and MAPK). In both the lower shade charts and the upper pie charts, positive correlations are displayed in red and negative correlations in green. Color intensity and the size of the circle are proportional to the correlation coefficients. CGGA, Chinese Glioma Genome Atlas; GBM, glioblastoma; LGG, lower grade glioma; MET, mesenchymal-epidermal transition factor; PD1, programmed cell death protein 1; PD-L1, programmed death-ligand 1; TCGA, The Cancer Genome Atlas; STAT, signal transducer and activator of transcription; AKT, protein kinase B; MAPK, mitogen-activated protein kinase.

Figure 2

MET triggered an increase in PD-L1 protein expression through STAT4 pathway. (A–C) Western blot of the indicated proteins in N33 and LN229 cell lines on treatment with HGF (200 ng/mL), and HGF/PLB-1001 (100 µM) combined, for 24 hours. GAPDH, protein-loading controls. Quantitative results of western blot analysis and relative expression difference are shown on the right panel. Fisher’s exact test, *p<0.05, **p<0.01, ***p<0.001. (D) In the IHC analysis of the tissue microarray, the scatter plot shows the correlation between the expression of MET and PD-L1 proteins in GBM. (E) Photographs of IHC staining of two representative WHO 4 (IDH-wildtype) primary glioma. Positive cells are stained brown. Magnification, 400×. (F) The statistical result of chromatin immunoprecipitation assay confirming the binding of STAT4 to the PD-L1 DNA. Student’s t-test, ***p<0.001. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GBM, glioblastoma; HGF, hepatocyte growth factor; IHC, immunohistochemistry; MET, mesenchymal-epidermal transition factor; PD-L1, programmed death-ligand 1; pMET, phosphorylated MET; pSTAT4, phosphorylated STAT, signal transducer and activator of transcription; PLB-1001, Bozitinib; NC, negative control.

Figure 3

Kaplan-Meier survival analysis of MET, STAT4, and PD-L1 in primary GBM of TCGA (A) and CGGA (B) data sets. Survival analysis was performed using Kaplan-Meier curve method in conjunction with two-sided log-rank test. CGGA, Chinese Glioma Genome Atlas; GBM, glioblastoma; MET, mesenchymal-epidermal transition factor; OS, overall survival; PD-L1, programmed death-ligand 1; TCGA, The Cancer Genome Atlas.

Figure 4

MET was closely related to glioma immunity, especially macrophage immunity. (A) GSEA results show that GO terms related to immunity are enriched in the MET^high group. (B) Volcano plot shows differently expressed genes between MET^high and MET^low groups. Red dots: significantly upregulated genes in the MET^high group; blue dots: downregulated genes in the MET^high group. (C) Bubble plot shows upregulated biological process in the MET^high group. (D and E) Macrophage abundance was determined with digital cytometry in CIBERSORTx. The total fraction of macrophages and subtypes (M0, M1, M2) in the MET^high and MET^low groups of TCGA and CGGA data set. (F and G) Visualizations of the resulting GEPs with tSNE plots which show macrophage and M2 macrophage expression difference between the MET^high and MET^low groups of TCGA and CGGA data sets. CGGA, Chinese Glioma Genome Atlas; GEPs, gene expression profiles; GO, gene ontology; GSEA, gene set enrichment analysis; MET, mesenchymal-epidermal transition factor; TCGA, The Cancer Genome Atlas; tSNE, t-distributed stochastic neighbor embedding; DEG, differently expressed genes; FDR, false discovery rate.

Figure 5

Tissue microarray of primary GBM verified the association between MET and macrophage-associated markers. (A) Macrophage-associated markers (IBA1 and TMEM119) and M2-like polarization markers (CD14 and IL-10). (C) Blood derived-like marker (TGFBI) and resident-like marker (BIN1). (B and D) Photographs of immunohistochemical staining of two representative primary IDH-wildtype GBM. Positive cells are stained brown. Magnification, 400×. GBM, glioblastoma; IL, interleukin; MET, mesenchymal-epidermal transition factor; TMA, tissue microarray.

Figure 6

MET recruited macrophages and decreased survival in primary GBM. (A) Western blotting of MET and PD-L1 protein in N33 or LN229 cells infected with the MET shRNA lentiviral vector or a negative control. Quantitative results of western blot analysis and relative expression difference are shown below. Fisher’s exact test, *p<0.05, **p<0.01, ***p<0.001. (B) A schematic diagram of the co-culture of macrophages and glioma cells in the Transwell system. Created with BioRender. (C) Transwell assays reveal that the knockdown of MET inhibits the chemotaxis of M2-like macrophages by GBM cells (N33 and LN229). Student’s t-test, **p < 0.01. (D) Patients with GBM with macrophage enrichment lived significantly shorter than other patients in all data sets. CGGA, Chinese Glioma Genome Atlas; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GBM, glioblastoma; MET, mesenchymal-epidermal transition factor; PD-L1, programmed death-ligand 1; OS, overall survival; TCGA, The Cancer Genome Atlas; sh, short hairpin; NC, negative control.

Figure 7

Single-cell transcriptome analysis of primary IDH-wildtype GBM. Two-dimensional UMAP shows dimensional reduction of data from single cells. (A) Nine samples are integrated and cells from each sample are differently colored. (B) Cells from nine samples are clustered into four groups including GAMs. (C) The violin plots show the expression of marker genes for the four groups of cells. (D) Nine samples are integrated for InferCNV analysis. The upper part of the heatmap shows the normal cells and the lower part shows the malignant cells. The red and blue colors refer to gains or deletions of chromosomes, respectively. (E) The UMAP plot shows tumor cells from the MET^high and MET^low groups. (F) The boxplot shows the level of MET expression in tumor cells from the MET^high and MET^low groups. (G) Proportion of GAMs in the MET^high group (42.33%) and the MET^low group (28.83%). Expression of proinflammatory mediators (H), anti-inflammatory mediators (I), and angiogenesis-associated genes (J) in GAMs from the MET^high and MET^low groups. GAMs, glioma-associated microglia, monocytes and macrophages; GBM, glioblastoma; MET, mesenchymal-epidermal transition factor; UMAP, uniform manifold approximation and projection.