Tumour-associated macrophage infiltration differs in meningioma genotypes, and is important in tumour dynamics

Abstract

Background: Meningiomas are the most common primary intracranial tumours, with clinical behaviours ranging from benign to highly aggressive forms. The World Health Organisation classifies meningiomas into various grades, guiding prognosis and treatment. While surgery is effective for low-grade meningiomas, certain grade 1 tumours, as well as grade 2, 3, and recurrent cases are more aggressive and require new therapeutic approaches. Immunotherapy shows promise, with early-stage clinical trials demonstrating encouraging results. The tumour microenvironment (TME), particularly tumour-associated macrophages (TAMs), plays a pivotal role in tumour progression. TAMs influence tumour growth, metastasis, and immune evasion. However, their role in meningiomas, especially in relation to genomic mutations, remains poorly understood. Understanding how genetic alterations affect the TME is critical for developing targeted immunotherapies.

Methods: This study employed multiplex immunohistochemistry and bulk RNA sequencing to explore immune infiltration in genetically stratified meningioma tissues and matched three-dimensional (3D) spheroid models. We compared immune cell populations across parental tissues, two-dimensional (2D) monolayer cultures, and 3D spheroid models. In addition, co-culture experiments were conducted, introducing M2-polarised macrophages derived from peripheral blood mononuclear cells to study the interactions between immune cells and tumour cells.

Results: Our findings revealed significant differences in the immune infiltration patterns associated with specific genotypes and methylation classes, especially M2-like TAMs. Notably, the 3D spheroid models more closely replicated the TME observed in parental tissues compared to traditional 2D monolayer cultures, offering a superior platform for immune infiltration studies. Furthermore, co-culture experiments demonstrated that M2-polarised macrophages could effectively infiltrate tumour cells, promote tumour cell proliferation while inhibiting invasion, suggesting IL-6-mediated signalling in tumour progression.

Conclusions: These findings suggest that 3D co-culture models offer an excellent system for studying the role of immune cells, specifically TAMs, in meningioma progression. By providing a more accurate representation of the TME, these models can help identify novel immunotherapy strategies aimed at modulating the immune response within meningiomas. Ultimately, this approach may improve therapeutic outcomes and quality of life for patients with meningioma by enhancing the effectiveness of existing treatments or by offering new immunotherapeutic options.

Keywords: 3D co-culture model; Genotype; Macrophage-to-microglia ratio; Meningioma; Methylation class; Tumour microenvironment; Tumour progression; Tumour-associated macrophages.

Fig. 1

TME landscape in meningioma tissue across different WHO grades. (a) Representative images displaying the expression and spatial distribution of five immune markers in grade 1 (n = 35), grade 2 (n = 10) and grade 3 (n = 4). FFPE tissue sections were stained using mIHC for CD68 (pan macrophages, red), CD163 (M2 macrophages, green), TMEM119 and P2RY12 (microglia, white and yellow, respectively), and CD3 (T cells, cyan). The images were captured at 200× and 630× magnification using Leica Stellaris. Coloured arrows indicate specific cell types: red arrow: total TAMs (CD68 +), yellow arrow: M2-like TAMs (CD68 + CD163 +), green arrow: macrophages (CD68 + TMEM119 - P2RY12 -), orange arrow: microglia (CD68 + TMEM119 + or CD68 + P2RY12 + or CD68 + TMEM119 + P2RY12 +), white arrow: M2-like macrophages (CD163 + macrophages), purple arrow: M2-like microglia (CD163 + microglia), and cyan arrow: T cells (CD3 +). (b) Quantification of the proportion of each immune cell type in total cells. (c) Quantification of the percentage of TAM subtypes in total TAMs cell population. (d) Percentage of M2-like TAM subtypes within M2-like TAMs. (e) Ratio of cell number among different subtypes of TAMs and M2-like TAMs. Analysis of cell counts was conducted using QuPath software based on the distinct marker expressions. Statistical analyses were performed using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01

Fig. 2

TME landscape in meningioma tissue across different genotypes. (a) Representative images illustrating the expression and spatial distribution of five specific immune markers in NF2 (n = 15), AKT1 E17K (n = 10) and KLF4 K409Q (n = 9). FFPE tissue sections were processed using the same mIHC protocol as Fig. 1a. Whole-tissue images were captured using PhenoImager™ HT system (Akoya). Additional high-magnification images at 200× and 630× were captured using Leica Stellaris. Coloured arrows indicate specific immune cell types: red arrow: total TAMs (CD68 +), yellow arrow: M2-like TAMs (CD68 + CD163 +), green arrow: macrophages (CD68 + TMEM119 - P2RY12 -), orange arrow: microglia (CD68 + TMEM119 + or CD68 + P2RY12 + or CD68 + TMEM119 + P2RY12 +), white arrow: M2-like macrophages (CD163 + macrophages), purple arrow: M2-like microglia (CD163 + microglia), and cyan arrow: T cells (CD3 +). (b) Quantification of each immune cell type in the total cell population. (c) Percentage of TAM subtypes within total TAMs. (d) Quantification of M2-like TAM subtypes within M2-like TAMs. (e) Ratio of cell number between different subtypes of TAM and M2-like TAMs. Analysis of cell counts was conducted using QuPath software, based on the distinct marker expressions. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 3

TME landscape in meningioma tissue across different methylation classes (MCs). (a) Representative images showing the expression and spatial distribution of five specific immune markers in MC ben-1 (n = 7), MC ben-2 (n = 6), MC ben-3 (n = 4), MC int-A (n = 6) and MC mal (n = 8). FFPE tissue sections were stained as described in Fig. 1a. High-magnification images at 200× and 630× were captured using Leica Stellaris. Coloured arrows indicate specific immune cell types: red arrow: total TAMs (CD68 +), yellow arrow: M2-like TAMs (CD68 + CD163 +), green arrow: macrophages (CD68 + TMEM119 - P2RY12 -), orange arrow: microglia (CD68 + TMEM119 + or CD68 + P2RY12 + or CD68 + TMEM119 + P2RY12 +), white arrow: M2-like macrophages (CD163 + macrophages), purple arrow: M2-like microglia (CD163 + microglia), and cyan arrow: T cells (CD3 +). (b) Quantification of the proportion of each immune cell type in the total cell population. (c) Percentage of TAM subtypes within total TAMs. (d) Quantification of M2-like TAM subtypes within M2-like TAMs. (e) Ratio of cell number between different subtypes of TAM and M2-like TAMs. QuPath was used to analyse the cell number base on distinct marker expressions. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 4

TME landscape in meningioma tissue across different genotypes and matched 3D spheroids models (n = 3). (a) Representative images displaying the expression and spatial distribution of five specific immune markers in NF2, AKT1 E17K and KLF4 K409Q. FFPE sections were stained following the same protocol as in Fig. 1a. High-magnification images at 200× and 630× were captured using Leica Stellaris. (b) Quantification of the proportion of each immune cell type in the total cell population in parental tissue. (c) Proportion of each immune cell type within total cells in matched 3D. Statistical analyses were conducted using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, ****p < 0.0001

Fig. 5

TME landscape in meningioma tissue across different MCs and matched 3D spheroids models (n = 3). (a) Representative images showing the expression and spatial distribution of five specific immune markers in MC ben-1, MC ben-2, MC ben-3 and MC int-A (without MC mal samples). FFPE sections were stained as described in Fig. 1a. High-magnification images at 200× and 630× were captured using Leica Stellaris. (b) Quantification of the proportion of each immune cell type in the total cell population in parental tissue. (c) Proportion of each immune cell type within total cells in matched 3D. Statistical analyses were performed using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 6

Comparison of immune features in matched tissue, 2D and 3D meningioma samples (n = 23). (a) Proportion of major immune cells in parental tissue and matched 2D and 3D models. Immune cell representation in the TME was analysed using deconvolution of bulk RNA-seq data with CIBERSORTx. Statistical difference was analysed by mixed-effects analysis with Dunnett’s multiple comparisons test. (b) Comparison of the proportion of two polarised macrophage subtypes across different sample types. Statistical difference was analysed using mixed-effects analysis with Sidak’s multiple comparisons test. (c) Comparison of the immune score across matched sample types. Statistical analysis was performed using a nonparametric test with Dunn’s multiple comparisons test. (d) Heatmap illustrating the expression differences of key immune markers among matched sample types. Centralised data based on RNA-seq FPKM values was visualised to depict the expression patterns of immune markers referenced to Human Immune Cell Marker Guide (https://www.cellsignal.com/pathways/immune-cell-markers-human) and Neuronal and Glial Cell Marker Atlas (https://www.cellsignal.com/pathways/neuronal-and-glial-cell-markers) posters by Cell Signaling Technology. Statistical difference was analysed using a nonparametric test with Dunn’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 7

Co-culture models demonstrating the role of M2-like macrophages on tumour cells (n = 3). (a) Western blotting analysis of immune markers (CD68, CD163 and TMEM119) in 2D and 3D primary cells derived from NF2-negative meningioma samples across different culture passages. KT21-MG1 (a grade 3 meningioma cell line) was used as a Merlin-negative control. IOMM-Lee (another grade 3 meningioma cell line) was the Merlin-positive control. (b) Double colour cell tracking illustrating the infiltration of M2-polarised macrophages (red) into tumour cells (green) during the co-culture process over time. Tumour cells were labelled with CellTrace™ CFSE (green), and M2-polarised macrophages with CellTracker™ Orange CMTMR (red). (c) Tumour cell proliferation in the co-culture model over time. ICC was used to detect M2-like macrophages marker CD163 (red) and the proliferation marker Ki67 (green), with nuclei stained by DAPI (blue). Statistical difference of Ki67 + cell count was conducted using two-way ANOVA with Sidak’s multiple comparisons test. (d) Tumour cell invasion in the co-culture model over time. Matrigel assay was performed to observe the cell protrusions of spheroids over time and stained using ICC with Phalloidin (green) for actin filaments and DAPI (blue) for nuclei. Invasion areas were quantified using ImageJ at 24 h and 48 h and compared between untreated and treated M2-polarised macrophages in 3D models. Statistical difference was analysed using two-way ANOVA with Sidak’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 8

Expression of macrophage-related cytokines in co-culture models (n = 3), and correlations between IL6 expression and recurrence. (a-g) Cytokine gene expression in cell models. (a) IL6 relative to GAPDH, (b) IL10 relative to GAPDH, (c) TNF relative to GAPDH, (d) IL10 relative to CD68, (e) TNF relative to CD68, (f) CSF1 relative to GAPDH, (g) TGFB1 relative to GAPDH. Statistical differences were analysed using an unpaired Student’s t-test for normal data or a nonparametric test for non-normal data. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (h-i) Correlations between IL6 expression and tumour recurrence in patients of (h) All WHO grades (p = 0.0063) and (i) WHO grade 1 (p = 0.0462). Fisher’s exact test was used for statistical analysis