TREM2 mediates MHCII-associated CD4+ T cell response against gliomas

Abstract

Triggering receptor expressed on myeloid cells 2 (TREM2) was recently highlighted as a novel immune suppressive marker in peripheral tumors. The aim of this study was to characterize TREM2 expression in gliomas and investigate its contribution in glioma progression by using Trem2^-/- mouse line. Our results showed that higher TREM2 expression was correlated with poor prognosis in glioma patients. Unexpectedly, TREM2 deficiency did not have a beneficial effect in a pre-clinical model of glioma. The increased TREM2 expression in glioma was likely due to increased myeloid cell infiltration, as evidenced by our single-cell analysis showing that almost all microglia and macrophages in gliomas were TREM2⁺. Furthermore, we found that deficiency of TREM2 impaired tumor-myeloid phagocytosis and MHCII presentation, and significantly reduced CD4⁺ T cells in tumor hemispheres. Our results revealed a previously unrecognized protective role of tumor-myeloid TREM2 in promoting MHCII-associated CD4⁺ T cell response against gliomas.

Keywords: TREM2; glioma; tumor-associated macrophages.

Figure 1:. High TREM2 mRNA expression in human gliomas is associated with poor patient prognosis.

A. Top six types of tumors with elevated TREM2 expression in a descending order from left to right, which were glioblastoma (GBM), brain lower grade glioma (LGG), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), pancreatic adenocarcinoma (PAAD), breast cancer (BRCA). B. Gene expression of TREM2 obtained from TCGA_GBMLGG dataset (n = 620). TREM2 expression was increased along with glioma WHO grade (grade II, n = 226, median = 9.28; grade III, n = 244, median = 9.74; grade IV, n = 150, median = 10.67). C. TREM2 expression was relatively lower in the IDH mutant (mutant, n = 429, median = 9.32; WT, n = 233, median = 10.49). D. TREM2 expression was relatively lower in the chromosome 1p/19q codeletion group (codel, n = 169, median = 8.33; non-codel, n = 494, median = 10.20). E & F. Kaplan-Meier survival curves generated for TREM2 expression in glioma patients. Patients were divided in high- and low- expressing groups based on quantile of TREM2 expression. In the TCGA dataset, higher TREM2 expression correlated with worse overall (TREM2 high, n = 168, events = 92; median = 24.2; TREM2 low, n = 166, events = 31; median = 134.3). Similar result was observed from the CGGA dataset (TREM2 high, n = 159, events = 103; median = 30.1; TREM2 low, n = 158, events = 66; median = 112.1). Data were tested for normal distribution using Shapiro-Wilk test first. P-values were acquired using two-tailed Student t-tests if data were normally distributed, or Mann-Whitney test if not. Survival curves were analyzed using log-rank test.

Figure 2:. TREM2 deficiency accelerates glioma but not peripheral tumor progression.

A. A schematic illustration of establishing an immunocompetent glioma model using murine glioma GL261 cells. Tumor size was monitored by bioluminescence imaging (BLI) every 7 days from day 14 post-inoculation. When mice reached the humane endpoint, contralateral and tumor hemispheres were collected separately for further analysis. B. The weight of hemispheres of WT and Trem2^−/− when humane endpoints were reached. C. The mRNA levels of Trem2 in the contralateral and tumor hemispheres were quantified by qRT-PCR. D & E. Representative bioluminescence images and statistical analysis showed that brain tumor burden was relatively higher in Trem2^−/− mice compared to WT mice. F. A larger tumor size was observed in Trem2^−/− mice compared to the WT mice 21 days after tumor inoculation, as indicated by the increased weight of tumor hemisphere. G - I. The survival study using 26 WT (12 males and 14 females) and 23 Trem2^−/− (9 males and 14 females) mice showed Trem2 deficiency did not confer any survival benefit in glioma. J - M. MC38 subcutaneous tumor experiment using 9 WT and 8 Trem2^−/− mice showed a clear trend (P = 0.0575) towards an attenuated tumor progression in Trem2^−/− mice. The data were shown as mean ± SEM. The data were tested for normal distribution using Shapiro-Wilk test first. P-values were acquired using two-tailed Student t-tests or two-way ANOVA if the data were normally distributed, or Mann-Whitney test if they were not. Survival curves were analyzed using log-rank test.

Figure 3:. TREM2 deficiency dampens MHC class II expression. A & C.

UMAP plots displaying the immune cells in patients with GBM and in mice with glioma GL261. B & D. TREM2 transcription in different immune cell populations. E. The correlation between TREM2 expression and signature genes related to myeloid cell phagocytosis, immunosuppression t, and antigen presentation was evaluated in LGG and GBM patients. The Spearman’s correlation test produces both P-values and correlation coefficients (R²). All listed genes had a P-value less than 0.05. Genes with an R² value greater than 0.25 were considered to have a correlation (either positive or negative) with TREM2 expression. F. Diagram of investigating the role of TREM2 in regulation of phagocytosis and antigen presentation in the mouse GL261 model. G & H. The mRNA levels of Cd68 and H2Aa (encoding MHC class II) in the contralateral and tumor hemispheres were quantified by qRT-PCR. I. UMAP plots of myeloid clusters in the MCA tumors from Trem2^+/+ and Trem2^−/− male mice. J. Feature plots of selected cluster markers and feature genes involved in phagocytosis and antigen presenting. The data were shown as mean ± SEM. The data were tested for normal distribution using Shapiro-Wilk test first. P-values were acquired using two-tailed Student t-tests if the data were normally distributed, or Mann-Whitney test if they were not.

Figure 4:. TREM2 deficiency impairs myeloid cell uptake tumor debris and antigen presentation.

A. A schematic illustration of tumor inoculation and window surgery for in vivo two-photon imaging. B. In vivo imaging of CX3CR1^GFP (green) myeloid cells interacting with mCherry⁺ (red) tumor cells. Solid triangles indicated myeloid cells that uptake red tumor debris and hollow triangles indicating those that do not. Scale bar: 10 μm. C. Dissociated CX3CR1 ^GFP cells from brain tumor under EVOS microscope, with some containing red tumor debris. Scale bar: 10 μm. D. Flow cytometry gating showing the detection of mCherry⁺ tumor debris signal in F4/80⁺MHCII⁺ macrophages. E. The percentage of tumor debris signal in F4/80⁺MHCII⁺ macrophages. F. UMAP plots of CD45⁺ immune populations detected in the GL261 tumor hemispheres and selected cluster markers. G. Correlation between percentage of microglia in CD45⁺ and the weight of tumor hemispheres in WT and Trem2^−/−. H. UMAP plots of infiltrating myeloid clusters in WT and Trem2^−/− and selected myeloid cell markers. I. Correlation between percentage of Ly6C^negF4/80⁺ in infiltrating myeloid cells and the weight of tumor hemispheres in WT and Trem2^−/−. J. Percentage of Ly6C^negF4/80⁺ in infiltrating myeloid cells in the high tumor burden group (tumor hemisphere > 250 mg). The bar graphs were shown as mean ± SEM. The data were tested for normal distribution using Shapiro-Wilk test first. P-values were acquired using two-tailed Student t-tests if the data were normally distributed, or Mann-Whitney test if they were not. For the correlation study, the Spearman’s correlation test produces both P-values and correlation coefficients (R²). A correlation is considered significant when the R² is greater than 0.25 and the P-value is less than 0.05.

Figure 5:. TREM2 is necessary for accumulation of CD4⁺ T cells in brain tumors.

A. Quantification of CD4 mRNA levels in tumor hemispheres using qRT-PCR. B. Quantification of the number of CD4⁺ T cells per mm² in the tumor core using confocal microscopy on 5 μm thick brain slides. C. Representative images of CD4⁺ T cell and myeloid cell (Iba1⁺) interactions. Scale bar: 10 μm. D. Quantifications of different types of CD4⁺-myeloid cell interactions in WT and Trem2^−/−. E. UMAP plots of CD3⁺ T cells in WT and Trem2^−/−. F & I. Expression levels of T cell cluster markers and feature genes indicating T cell activation or immunosuppression. G & H. Correlation between percentage of T regulatory cell and CD8⁺ in total T cells (CD3⁺) and the weight of tumor hemispheres in WT and Trem2^−/−. Spearman’s correlation test was used to calculate P-values and correlation coefficients (R²). J. Projection of T cells from newly diagnosed GBM using the ProjecTIL package to reveal T cell subsets. K. Correlation between TREM2 expression and marker genes of T cell subtypes in LGG and GBM patients. All listed genes, except for CTLA4, had a P-value less than 0.05. L. Our working model proposes that TREM2-mediated phagocytosis of glioma debris by myeloid cells leads to further MHC class II presentation to CD4⁺ T cells, ultimately contributing to anti-tumor immunity in brain tumors. The bar graphs were shown as mean ± SEM. The data were tested for normal distribution using Shapiro-Wilk test first. P-values were acquired using two-tailed Student t-tests if the data were normally distributed, or Mann-Whitney test if they were not.