Toxoplasma gondii infection supports the infiltration of T cells into brain tumors

Abstract

Few T cells infiltrate into primary brain tumors, fundamentally hampering the effectiveness of immunotherapy. We hypothesized that Toxoplasma gondii, a microorganism that naturally elicits a Th1 response in the brain, can promote T cell infiltration into brain tumors despite their immune suppressive microenvironment. Using a mouse genetic model for medulloblastoma, we found that T. gondii infection induced the infiltration of activatable T cells into the tumor mass and led to myeloid cell reprogramming toward a T cell-supportive state, without causing severe health issues in mice. The study provides a concrete foundation for future studies to take advantage of the immune modulatory capacity of T. gondii to facilitate brain tumor immunotherapy.

Keywords: Brain tumors; IFNγ; Toxoplasma gondii; Tumor associated myeloid cells; Tumor-infiltrating T cells.

Figure 1:. Medulloblastoma does not impact T. gondii burden and infection outcome in mice.

A) Schematic of T. gondii infection and tissue harvesting procedures to evaluate parasite burden in medulloblastoma-bearing mice (MB) and wild-type (WT) littermates. Briefly, mice were given an intraperitoneal injection (I.P.) of 100 Me49 tachyzoites and harvested at either 12 days post-infection (DPI) or 27 DPI. At 12 DPI, when T. gondii are primarily in the periphery organs, the heart and lung were collected. At 27 DPI, when T. gondii has disseminated to the brain, the forebrain was collected. Additionally, to assess parasite burden in medulloblastoma, the tumor mass and nearby tumor-free cerebellar regions (CB) were collected, respectively. B - C) Parasite burden in the periphery was equivalent between WT and MB mice. At 12 DPI, hearts (B) and lungs (C) from infected WT (n=12) and infected MB mice (n=12) were collected. Genomic DNA was extracted and subjected to qPCR. Signals from T. gondii specific 529-bp repeat element were normalized to a standard curve of 3 to 300,000 T. gondii genome copies to calculate the actual T. gondii genomic copies in the samples. Unpaired two-tailed T-test. D) Parasite burdens in the forebrains were equivalent between WT and MB mice. At 27 DPI, forebrain samples from infected WT (n=18) and infected MB mice (n=18) were collected. Extracted genomic DNA was subjected to qPCR evaluation as described above. Unpaired two-tailed T-test. E) Parasite burdens were equivalent between tumor and tumor-free cerebellar regions in MB mice. At 27 DPI, tumors and adjacent tumor-free cerebella from infected MB mice (n=10) were isolated and measured for parasite burden. Paired two-tailed T-test. F) T. gondii infection did not exacerbate sickness in MB mice. Sickness was monitored in uninfected PBS-injected WT mice (green, n=15), uninfected MB mice (blue, n=22), infected WT mice (black, n=37), and infected MB mice (red, n=36). The dashed line denotes the euthanasia-reaching point. Each dot represents one mouse. The total sickness score was analyzed using a Poisson mixed-effects model explaining the sickness symptoms by treatment group. The main model was followed by Tukey pairwise differences. *p < 0.05, **p < 0.005, *** p < 0.001, ns=not significant. Images Created with BioRender.com

Figure 2:. T cells were recruited with similar efficiency into the tumor mass and the tumor-free cerebellum of infected mice.

A) Diagram of samples collected to evaluate T cell abundance and distribution. At 27 DPI, cerebella (CB) from WT mice and tumors from MB mice were collected for downstream analysis via flow and immunofluorescence staining. B – C) At 27 DPI, similar amounts of T cells were detected in the cerebella of infected WT mice as in tumors of infected MB mice. The percentage of CD45⁺TCRβ⁺ T cells among live cells (B) and T cell number relative to tissue weight (C) in the cerebella of uninfected WT mice (n=10), the cerebella of infected WT mice (n=10), tumors of uninfected MB mice (n=5), and tumors of infected MB mice (n=10) were evaluated via flow cytometry as described in Supp. Fig 1A. D-G) At 27 DPI, T cells were found in equivalent amounts at the borders (Laminin+ vessels) and cerebellar parenchyma of infected WT mice (n=4) and tumors of infected MB mice (n=4). Representative confocal images of the spatial distribution of T cells (CD3+, yellow) relative to Laminin positive vessels (blue) within WT cerebella and tumor regions (tdTomato+, red) of uninfected and infected mice (D). The lower panels are zoom-ins of the boxed areas of the upper panels. Scale bar = 50 μm. Laminin-positive vessels demarcate leptomeninges and blood vessels (E). CD3+ T cell colocalization with laminin (F) or the brain parenchyma (G) was quantified. Data represents three images from each of 4 mice per group. Statistical significance was determined by ordinary 2-way ANOVA with main effects only, followed by Tukey pairwise comparisons (B, C, F, and G). * p < 0.05, ** p < 0.005, *** p < 0.001. Images Created with BioRender.com

Figure 3:. T cell composition and activation signatures induced by T. gondii infection were similar between tumor regions of MB mice and cerebella of WT mice.

A – C) At 27 DPI, T cells recruited to the tumors of infected MB mice shared a similar composition with those in the cerebella of infected WT mice. Flow cytometric quantification of the frequency of CD4+ T cells (A), CD8+ T cells (B), and CD4+Foxp3+ regulatory T cells (C) among CD45+TCRβ+ T cells from the cerebella of infected WT mice (n=10) and tumors of infected MB mice (n=10). The gating strategy is described in Supp. Fig. 1A. Unpaired two-tailed t-test. D – G) T cells from tumors and adjacent non-tumor cerebella of infected MB mice (n=5) were evaluated for their ability to produce effector cytokines ex-vivo. After tissue dissociation, samples underwent 6 hours of PMA/Ionomycin restimulation in the presence of Brefeldin A. Flow cytometric evaluation detected no significant differences in the frequency of IFNγ+CD4+ (D) and IFNγ+CD8+ cells (E) among total CD4+ and CD8+ T cells, between the cerebella and tumor regions. The gating strategy is described in Supp. Fig. 5. Paired two-tailed t-test. Flow cytometric quantification for Median Fluorescent Intensity (MFI) of IFNγ signals among CD4+ T cells (F) and CD8+ cells (G). Paired two-tailed t-test. * p < 0.05, ** p < 0.005, *** p < 0.001, ns = not significant

Figure 4:. T. gondii infection leads to a pro-inflammatory myeloid cell population in medulloblastoma at 27 DPI.

A - B) IFNγ was elevated in the tumor mass of MB mice infected with T. gondii. Via q/RT-PCR of bulk-tissue RNA, an elevated transcript level of IFNγ was detected in the tumors of infected MB mice (N=9 per group) (A). Correspondingly, via Luminex, IFNγ protein level was significantly higher in the protein lysate of tumors from infected MB mice (n=7) relative to those from uninfected MB mice (n=5) (B). Unpaired two-tailed T-test. C - D) IFNγ-induced transcripts were detected in the CD11b+ myeloid cells isolated from tumors of infected mice. CD11b+ myeloid cells were enriched from tumor samples of uninfected MB mice (n=5) and infected MB mice (n=5) using CD11b+ microbeads. Sample purity was validated as shown in Supp. Fig. 7. Gene Set Enrichment Analysis of RNA-seq data identified ‘Interferon gamma response’ as the top hallmark of significantly changed genes in enriched myeloid cells from T. gondii infected mice (C). Heatmap shows the z-score value for differentially enriched genes (padj<0.05) belonging to the ‘Interferon gamma response’ hallmark (D). The Z-score was calculated using log2(FPM). E - F) Immunostaining validates the increased expression level of MHC-II protein in myeloid cells within TME. Representative confocal image from tumor regions that were immunostained with Iba1 (green) to label myeloid cells and MHC-II (red) (E). The number of MHC-II+ Iba1+ cells in tumors from uninfected (n=4) or T. gondii-infected mice (n=5) was quantified (F). Unpaired two-tailed T-test. * p < 0.05, ** p < 0.005, *** p < 0.001.