Radio-chemotherapy and metformin selectively modulate the heterogeneous landscape of glioma with ribosome biogenesis, long non coding RNA and immune-escape markers as major player

Abstract

Glioblastoma multiforme (GBM) is the most common primary brain tumor in adults with a short survival time after standard therapy administration including radiotherapy (RT) associated with temozolomide (TMZ). Here, we investigated the effects of radiochemotherapy in association with metformin (MET), a drug targeting cell metabolism on a syngeneic GBM mouse model using Positron Emission Tomography imaging with [¹⁸F]FLT and [¹⁸F]VC701 and single-cell RNA-sequencing analysis. The addition of drugs to RT significantly increased survival and [¹⁸F]FLT showed an early predictive response of combined therapy. We identified the presence of heterogeneous tumor populations with different treatment sensitivity and a complex immune evasive microenvironment. Tumor cells surviving to treatments showed immune response, among the main differentially modulated biological functions and a potential role of long non-coding RNAs (lncRNAs) in treatment resistance. Association with TMZ or TMZ plus MET reduced the pro-tumor phenotype of immune reaction acting more on myeloid cells the first and on lymphocytes the latter. Off note, MET add-on counteracted the immune-evasive phenotype particularly of T cells suggesting a potential role of MET also in adopted immunity.

Keywords: Glioblastoma; LAG3; PET imaging; long-non coding RNA; metformin; radio-chemotherapy; single-cell RNA sequencing.

Figure 1

In vivo effect of radiotherapy treatment. A Kaplan-Meier survival curves of GL261 bearing mice treated with vehicle, radiotherapy (RT), Temozolomide (TMZ), Metformin (MET), TMZ+MET, RT+TMZ and RT+TMZ+MET. Statistical significance was determined by the log-rank test. The p values are indicated. n = number of mice. B Representative T2-weighed MRI images and their fusion with PET images for [¹⁸F]FLT- and [¹⁸F]VC701 of GL261 bearing mice treated with vehicle, RT alone, RT+TMZ and RT+TMZ+MET acquired before the beginning of treatment and after about 14 and 28 days from the beginning of the treatment. C At 2 weeks we detected a significant lower uptake of [¹⁸F]FLT in RT+TMZ treated mice. Radiotracers uptake was expressed as tumor to background ratio. * p<0.05 by ordinary one-way ANOVA analysis followed by Tukey's multiple comparison test. Each symbol represents one animal, bars and error bars indicate group mean±sd. Correlation curve indicated that only early [¹⁸F]FLT uptake correlated with overall survival for both RT plus TMZ (Pearson r = -0.986, R² = 0.972, p<0.0001) or RT plus TMZ and MET (Pearson r = -0.903, R² = 0.816, p = 0.0356). Each symbol represents one animal. D ROC analysis of [¹⁸F]FLT uptake for prediction of different response to therapy. Optimal cut-off point was defined for [¹⁸F]FLT as 2.965 (75.0% sensitivity; 70.0% specificity).

Figure 2

Post treatment expression of GFAP, Iba1 and TMEM. Immunofluorescence image of A) DAPI and B) GFAP, Iba1 and TMEM after treatment with vehicle, RT early, RT, RT+TMZ and RT+TMZ+MET. For each treatment condition, 3 samples were analyzed.

Figure 3

Evaluation of RT and drugs treatment on tumor microenvironment. Immunofluorescence image and quantification of selected markers. For each treatment condition, 3 samples were analyzed. A) GFAP, IBA1, CD201, CD16 and TMEM expression was quantified in the tumor (Tumor), in tumor-brain border (Peripheral) and in the brain region contralateral to the tumor (CL) and data were expressed as percentage of positive cells. B) CD31 expression was evaluated in all the sample and expressed as percentage of positive cells; CD3 expression was evaluated only in the tumor area and expressed as number of positive cells on tumor area and finally TSPO expression was evaluated in the tumor, in Peripheral and in CL and data expressed as ratio between the values in tumor and peripheral areas on CL. C) Graphs showed the expression of CD206 and TMEM on IBA1 positive cells. Data represent mean ± sd. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 for ordinary two-way or one-way ANOVA analysis followed by Tukey's multiple comparison test.

Figure 4

Landscape of cells in tumor microenvironment. A Uniform manifold approximation and projection (UMAP) plot representing cells of GBM and TME of the entire dataset obtained by merging all the samples by applying a Panglao set base panel. Each point represents one cell. B Genotype analysis identified a large exogenous region overlapping the unassigned area corresponding to GL261 tumor cells. C Tumor cells (grey), myeloid cells (pink), parenchymal cells (yellow), and lymphoid cells (light blue) were identified. D UMAP of the 30 clusters identified using Seurat tool. Each point represents one cells. Astr: astrocytes; Bcell: B lymphocytes; DC1: dendritic cells; DC2: dendritic cells; End: endothelial vascular cells; Epend: ependymal cells; mDC: migratory dendritic cells; MG: microglia; Mo: infiltrating monocytes; Neur: neurons; NK: natural killer cells; Olig: oligodendrocytes; Olig/neu prec: oligodendrocytes/neuron precursors; Per: pericytes; Teff: T effector lymphocytes; Treg: T regulatory lymphopcytes; Tum: tumor cells.

Figure 5

The four tumor subclusters displayed different radio and/or chemoresistance features. A Relative contribution of each cluster of cells in the different groups of treatment. B UMAP plots showing normal (light grey) and tumor (dark grey) cells. C Relative contribution of normal and tumor cells of UMAP plots showed in a. The numbers upon the bars indicate the percentage of tumor cells on the number of the cells of each sample. D UMAP plot of the tumor cluster where the 4 subclusters are showed: CL5, CL8, CL9 and CL10. E UMAP plots showing cells in the different cell cycle states (red: G1, green:G2M and blue: S) after treatment. The bar graphs indicated the relative contribution of tumor and normal cells in each cell cycle state. The contribution is normalized on the number of cells of each sample. F Relative and absolute expression of each tumor subclusters after treatment. G Heatmap of the common DEGs obtained from the comparison of untreated group with RT group and RT group with RT+TMZ group performed in the cluster CL5, CL8, CL9 and CL10.

Figure 6

Modification in biological functions after therapy in tumor clusters. Top upregulated and downregulated biological functions in tumor clusters 5, 8, 9 and 10 obtained by gene ontology (GO) analysis of significant DEGs previously identified among the groups: untreated versus RT alone, RT alone versus RT plus TMZ, RT plus TMZ versus RT plus TMZ and MET.

Figure 7

Modulation of non-tumor cell populations after RT and chemotherapy. The bar graphs indicate the relative A) and the absolute B) contribution of parenchymal, myeloid, and lymphoid clusters. For the relative graphs the contribution is normalized on the number of cells of each sample. Relative contribution of dendritic, microglia and peripheral monocytes in myeloid cells and in single clusters are shown in panel C).

Figure 8

RT and chemotherapy induced modifications in monocytes/tumor associated macrophages and microglia. A Heatmap of the common differentially expressed genes analysis between untreated versus RT, RT versus RT+TMZ and RT+TMZ versus RT+TMZ+MET in monocytes/TAMs clusters and the corresponding volcano plots of CL 0, 2 and 7. MoTAM: monocytes/tumor associated macrophages. B Heatmap of the common differentially expressed genes analysis between untreated versus RT, RT versus RT+TMZ and RT+TMZ versus RT+TMZ+MET in microglia clusters and the corresponding volcano plots of CL 3 and 15. Up-regulated genes of the first term of comparison are highlighted in blue, down-regulated genes of the first term of comparison are highlighted in green, and not significant genes are highlighted in orange (Log₂FC > ±0.3; Adj p<0.05).

Figure 9

Modification in biological functions after therapy in MoTAM clusters. Top upregulated and downregulated biological functions in monocytes/tumor associated macrophages clusters 0, 2, and 7 obtained by gene ontology (GO) analysis of significant DEGs previously identified among the groups: untreated versus RT alone, RT alone versus RT plus TMZ.

Figure 10

Modification in biological functions after therapy in microglia clusters. Top upregulated and downregulated biological functions in microglia clusters 3, 12, 15, 18, and 20 obtained by gene ontology (GO) analysis of significant DEGs previously identified among the groups: untreated versus RT alone and RT alone versus RT plus TMZ.

Figure 11

Modification in biological functions after therapy in T cells clusters. Top upregulated and downregulated biological functions in T cells clusters 1 and 4 obtained by gene ontology (GO) analysis of significant DEGs previously identified among the groups: untreated versus RT alone and RT alone versus RT plus TMZ.

Figure 12

RT and chemotherapy induced modifications in lymphocytes. Heatmap of the common differentially expressed genes analysis between untreated versus RT, RT versus RT+TMZ and RT+TMZ versus RT+TMZ+MET in lymphocytes and the corresponding volcano plots of CL 1, 4 and 22. Up-regulated genes of the first term of comparison are highlighted in blue, down-regulated genes of the first term of comparison are highlighted in green, and not significant genes are highlighted in orange (Log₂FC > ±0.3; Adj p<0.05).