Brain-infiltrating CD4 T cells drive inflammatory microglia proliferation during cryptococcal meningitis in mice

Abstract

Cryptococcal meningitis is a fungal infection in patients with compromised CD4 T cell function. CD4 T cells provide killing signals to macrophages, principally IFNγ, to limit intracellular fungal replication. However, CD4 T cells may also drive inflammatory tissue damage. Yet, it is not fully understood how fungal-specific CD4 T cells infiltrate the brain and how they influence functional phenotypes of CNS-resident myeloid cells. In the current work, we develop a mouse model to track fungal-specific CD4 T cells and determine their influence on microglia. We found IFNγ+ fungal-specific CD4 T cells have limited TCR signalling and characterise a population of inflammatory microglia that upregulate MHCII and IFNγ-regulated genes during infection. Inflammatory microglia have poor fungicidal capacity and significantly expand during infection, a process that depends on CD4 T cell infiltration. Taken together, these data identify the early inflammatory consequences of fungal-specific CD4 T cell infiltration and identify proliferating microglia as important drivers of brain inflammation during infection.

Fig. 1. CD4 T cells infiltration to the C. neoformans infected brain correlates with fungal burden and accumulate near sites of yeast growth.

a Fungal burdens in the brain at days 0, 3, 6 and 9 post-infection. Data from a single experiment, n = 3 mice (day 3 and 6) or 4 mice (day 9). Data presented as mean +/- SEM. b Weight loss, relative to starting weight, following i.v. infection with C. neoformans H99 in C57BL/6 mice. Data pooled from two independent experiments, n = 11 mice per time point. Data show as mean +/- SEM. c Total number of CD4 T cells per brain at days 0 (n = 5 mice), 3 (n = 4 mice), 6 (n = 7 mice) and 9 (n = 6 mice) post-infection. Data pooled from 1 to 2 independent experiments. Whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean. d Number of brain-infiltrating CD4 T cells relative to brain fungal burden in mice infected with low (open symbol), medium (grey symbol) and high (filled symbol) dose of C. neoformans. Each point represents an individual animal. Data analysed by simple linear regression. e Example flow cytometry plots of peripheral blood and brain samples (matched from same animal) from mice pre-injected with cell labelling dye CFSE 10 min prior to euthanasia and analysis. Graph shows quantification where each point represents an individual animal. Data analysed by paired t-test. *P = 0.0206. f Example confocal microscopy images of CD4 staining in the infected brain at day 7 post-infection (representative of 3 animals analysed). Wild-type mice were infected with GFP-expressing C. neoformans and brain sections stained with DAPI (blue) and anti-CD4 (white). g Example Cda2-MHCII tetramer staining of leukocytes isolated from uninfected (open symbol) or C. neoformans-infected brains (at day 7 post-infection; closed symbol). Example dot plots are pre-gated on CD45^hiCD44⁺CD4⁺TCRβ⁺ live singlets. Frequency and the number of cells acquired were quantified in uninfected (n = 5 mice) and infected (n = 11 mice) animals. Data pooled from two independent experiments and analysed by Mann-Whitney U-tests. *P = 0.0389, **P < 0.0094.

Fig. 2. Fungal-specific CD4 T cells are recruited to the brain late post-infection.

a Schematic of adoptive transfer model used in this study. b Frequency of transferred fungal-specific CnT.II CD4 T cells in the brain (n = 6 mice day 0, n = 9 mice day 7, **P = 0.0021), spleen (n = 7 mice day 0, n = 9 mice day 7, **P = 0.0021), lung (n = 7 mice day 0, n = 9 mice day 7, *P = 0.0212) and cLN (n = 6 mice day 0, n = 9 mice day 7, **P = 0.0052), c the frequency of divided (CFSE^low) CnT.II cells in the brain (n = 6 mice day 7), spleen (n = 5 mice day 0, n = 6 mice day 7, **P = 0.0027), lung (n = 5 mice day 0, n = 6 mice day 7, ****P < 0.0001) and cLN (n = 6 mice day 0, n = 6 mice day 7, **P = 0.0024) and (d) frequency of CD44 expression within CnT.II cells in the brain (n = 6 mice day 7), spleen (n = 5 mice day 0, n = 6 mice day 7, ****P = 0.0001), lung (n = 5 mice day 0, n = 6 mice day 7, ****P = 0.0001) and cLN (n = 3 mice day 0, n = 6 mice day 7, ***P = 0.0002). Data is pooled from two independent experiments and analysed by two-way ANOVA. e Number of brain-infiltrating CnT.II CD4 T cells relative to brain fungal burden in mice infected with low (open symbol), medium (grey symbol) and high (closed symbol) dose of C. neoformans. Each point represents an individual animal. Data analysed by simple linear regression. f Total number of CD4 T cells and fungal-specific CnT.II T cells in the meninges of uninfected (n = 3 mice; open symbol) and infected (n = 10 mice; day 7 post-infection, closed symbol) mice, Data pooled from three independent experiments (one uninfected mouse per experiment) and analysed by unpaired t-test. *P = 0.0127.

Fig. 3. Brain-infiltrating fungal-specific CD4 T cells are IFNγ-producing effector cells.

a Frequency of CnT.II cells that had diluted out cell proliferation dye indicating a high number of cell divisions shown by dark grey gate shown at far left of example flow plot, n = 6 mice analysed per organ, b the production of IFNγ by CnT.II cells (n = 6 mice analysed per organ) and c the proportion of CnT.II cells with an effector (CD44⁺CD127^-CD62L^-; open symbol), central memory (CM; CD44⁺CD127⁺CD62L⁺; grey symbol) or effector memory (EM; CD44⁺CD127⁺CD62L^-; closed symbol) phenotype in the indicated organs at day 7 post-infections (n = 6 mice analysed per organ). CLN cervical lymph nodes. Data pooled from 2 independent experiments. Each point represents an individual mouse, with the line representing the mean. Data analysed by two-way ANOVA. ***P = 0.0007, ****P < 0.0001. d Example flow cytometry plots of IFNγ-YFP and IL-17A-CD271 reporter expression in wild-type (no reporter genes), and uninfected and infected (day 7 post-infection) reporter animals. Similar results were also confirmed in a separate IL-17A-dTomato reporter mouse line, and example flow cytometry plots of IL-17A-dTomato reporter expression is shown for uninfected and infected (day 7 post-infection) reporter animals. Results are representative of 9 animals, analysed in three independent experiments.

Fig. 4. Fungal-specific CD4 T cells have limited TCR engagement in the brain.

a Fungal burdens in brain and lung at day 7 post-infection of mice (n = 6 mice) intravenously infected with C. neoformans and adoptively transferred with CnT.II cells as outlined in Fig. 2a. Data pooled from two independent experiments and analysed by Mann Whitney U-test (*P = 0.0189). Whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean. b Frequency of CnT.II cells expressing Nr4a3-Tocky (blue, red or both forms) in the indicated organs at day 7 post-infection (n = 6 mice). Data pooled from two independent experiments, presented as mean +/- SEM and analysed by two-way ANOVA (comparing total Nr4a3-Tocky⁺ cells), ****P < 0.0001. c Frequency of Nr4a3-Tocky⁺ CnT.II cells that were cultured in vitro either with splenocytes alone (unstimulated), splenocytes loaded with Cda2 peptide or cultured onto plates coated with anti-CD3/28 antibody. Each point represents a technical replicate (n = 3 wells). Data is a representative experiment from two independent repeats (each with at least 3 technical replicates). d Mice were treated as in Fig. 2a, and then injected with indicated amounts of Cda2 peptide intraperitoneally 4 h prior to analysis (n = 3 mice for 80 µg [closed symbol], n = 4 mice for 200 µg [open symbol], with line representing the mean). Each point represents an individual animal. Data analysed by unpaired t-test. *P = 0.0479, ****P < 0.0001. e Nr4a3-Tocky mice were infected intravenously with C. neoformans and the proportion of infiltrating CD4 T cells in the brain (closed symbol), lung (grey symbol) and spleen (open symbol) expressing Nr4a3-Tocky (all forms) determined at days 3, 7 and 9 post-infection. Example plots are gated on total CD3⁺CD4⁺ cells in the brain. Data is pooled from two independent experiments, n = 3 mice per time point, shown as mean +/- SEM. f Mean fluorescence intensity (MFI) of TCRβ expression of CD4 T cells in the brain, lung and spleen of mice (n = 10 mice analysed, with line representing the mean) at day 7 post-infection. Data pooled from two independent experiments; each point represents an individual animal and analysed by one-way ANOVA. ****P < 0.0001.

Fig. 5. Inflammatory microglia expand during late C. neoformans infection.

Analysis of single-cell RNA sequencing dataset from uninfected (n = 2) and infected (day 3, n = 2; day 6, n = 5) animals. a tSNE plot showing graph-based clustering results and 14 unique clusters identified within the data. b tSNE plots coloured by microglia, monocyte/macrophage or neutrophil signature genes. c tSNE plot coloured by time point analysed (day 0, grey; day 3, blue; day 6, red). d Proportion of sequenced cells from each of the analysed time points within microglia cluster 10 and all other identified microglia clusters in the dataset. e Heatmap of selected genes, grouped by function/similarity, showing mean expression by microglia cluster 10 or other microglia clusters. A full list of differentially-expressed genes is given in Data S1. f KEGG pathway analysis of upregulated genes by microglia cluster 10 (compared to other microglia clusters), showing gene ratio of detected genes within indicated pathways, P-value and enrichment scores. Statistics derived from Fisher’s exact test, with FDR multiple correction.

Fig. 6. Inflammatory microglia (IAMs) derive from resident microglia and co-express MSR1 and CD206.

a Heatmap of mean expression of indicated genes between IAMs and non-IAMs. b Example flow cytometry of IAMs gating. c Total number of MSR1⁺CD206⁺ microglia in uninfected or infected (day 6 post-infection) brains. Data pooled from 2 independent experiments, n = 9 mice per time point. Data analysed by unpaired t-test, *P = 0.0166. Whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean. d Quantification of Tmem119 expression by qRT-PCR on CD206⁺MSR1⁺ and CD206^-MSR1^- microglia (defined as CD45^intCX3CR1^hi), compared to macrophages (defined as CD45^hiCX3CR1⁺) at day 6 post-infection. Each point represents an independent sort experiment pulled from a batch of mice (10 mice per batch, n = 3 sorts in total), presented as mean +/- SEM. e Normalised counts of microglia signature genes in IAMs (red violin) and non-IAMs (grey violin), calculated from the sequencing dataset. Data analysed by Mann Whitney U-tests. *P < 0.0001. f Example histogram of Sall1-dTomato expression by microglia (CD45^intCX3CR1^hi) and macrophages (CD45^hiCX3CR1⁺) in Sall1^CreERRosa26^Ai14 mice. Graph shows quantification for 8 different animals, from 2 independent experiments. g Expression of IAMs markers CD206 and MSR1 within the Sall1-dTomato⁺ population and Sall1-dTomato^- population^. The graph shows quantification from 2 independent experiments, where each point represents an individual animal (n = 8 mice total). Data in (f, g) analysed by paired t-tests. **P = 0.0055, ***P = 0.0004.

Fig. 7. IAMs have poor fungicidal capacity.

a Frequency of IAMs (MSR1⁺CD206⁺) and non-IAMs (MSR1^-CD206^-) associating with GFP-expressing C. neoformans at day 6 post-infection as determined by flow cytometry and b the MFI of the GFP signal within each population. Each dot represents an individual mouse (n = 10 mice analysed), with lines denoting paired samples within the same mouse. Data pooled from 3 independent experiments and analysed by paired t-test, ***P < 0.0007. c Example imaging of FACS-purified uninfected microglia, C. neoformans-infected IAMs and C. neoformans-infected non-IAMs using an imaging flow cytometer (ImageStream). Mice were infected with GFP-expressing C. neoformans and microglia isolated at day 7 post-infection for sorting. Sorted populations were fixed and imaged in the brightfield and GFP channels to quantify number of yeast per microglia cell. d Quantification of number of yeast per IAM (n = 78 microglia counted) and non-IAM (n = 53 microglia counted) microglia. Data pooled from two independent sorts/infection experiments and analysed by Mann Whitney U-test. **P = 0.0071. e Frequency of live fungal cells within IAMs and non-IAMs. Mice were infected with GFP-expressing C. neoformans and infected IAMs or non-IAMs (GFP+) were FACS purified and 500-2000 total cells plated onto YPD agar plates. The number of colonies counted was then expressed as a proportion of the total number of plated cells to calculate viability. Data pooled from 3 independent experiments and analysed by paired t-test, *P = 0.0263. f Mean fluorescence intensity of arginase expression (as determined by intracellular flow cytometry) in IAMs and non-IAMs at day 6 post-infection. Each dot represents an individual mouse, lines denote pairing within the same animal. Data pooled from 2 independent experiments and analysed by paired t-test, ****P = 0.0006. g Example confocal microscopy of brains from uninfected mice (i, with close up insert) and an uninfected mouse treated with microglia-depleting drug PLX5622 as a staining control (panel ii), stained with DAPI (blue) and anti-Iba1 (red). h Confocal microscopy of C. neoformans infection in the brain at day 7 post-infection. Panels iii and iv show areas of extracellular yeast growth in the cerebral cortex (panel iii) and in the cerebellum (panel iv). White arrows denote rounded, activated microglia. Insert shows close up from panel iv. Panels v and vi show examples of microglia within areas of infection. Example images are representative sections from at least three individual animals in which 3–5 sections per brain were analysed.

Fig. 8. IAMs are a proliferating microglia subset.

a Pie chart showing proportion of cell-cycle genes expressed by IAMs that are highly expressed at specific stages of the cell cycle. Genes were assigned to specific stages of the cell cycle using annotations published in reference. See Supplemental Data 2 for a full list of genes included in the analysis and their cell cycle stage assignment. b Example flow cytometry plots of Ki67 staining within total microglia, and staining for IAMs markers CD206 and MSR1 in the Ki67^- and Ki67⁺ microglia populations. c Graph of frequency of total Ki67⁺ microglia at various time points post-infection (day 0, n = 7 mice; day 3, n = 3 mice; day 6, n = 6 mice; day 9, n = 5 mice). Data shown as mean +/- SEM and pooled from 2 independent experiments, with n = 5–8 mice analysed per time point. d Graph of frequency of Ki67⁺ positive cells within MSR1⁺CD206⁺ microglia (IAMs; filled bars) and MSR1^-CD206^- microglia (non-IAMs; open bars) at day 6 (n = 8 mice) and day 8 (n = 5 mice) post-infection. Data pooled from 2 independent experiments, with n = 8 (day 6 post-infection) or 5 (day 8 post-infection) mice analysed. Whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean. Data analysed by two-way ANOVA, ****P < 0.0001, *P = 0.0103. e Example confocal images of Ki67 labelling of sections from uninfected or infected (day 8 post-infection) brains. An area of fungal infection is outlined with dashed white lines. White arrows denote Ki67⁺ proliferating cells. f Quantification of Ki67 staining relative to DAPI. Pixel density of each stain was calculated using ImageJ relative to section size. Each point represents a section analysed from individual mice (n = 2 mice analysed per group). g Brain and h lung fungal burdens from wild-type mice infected intranasally with C. neoformans at day 14 (n = 12 mice) and day 21 (n = 12 mice) post-infection. Data analysed by Mann Whitney U-test, **P = 0.0088. i Number of IAMs and (k) Ki67⁺ microglia in brains of wild-type mice infected intranasally with C. neoformans at day 14 (n = 12 mice) and day 21 (n = 12 mice) post-infection. j Frequency of Ki67⁺ cells within the IAMs and non-IAMs populations at day 14 post-infection (n = 12 mice). Data analysed by paired t-test, ****P < 0.0001. Each point represents an individual mouse, with lines denoting matched samples from the same animal. Intranasal data pooled from two independent experiments and analysed by unpaired t-tests. In all box-and-whisker plots, whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean.

Fig. 9. Brain-infiltrating CD4 T cells drive microglia proliferation during infection.

a Number of Ki67⁺ microglia and b IAMs (CD206⁺MSR1⁺ microglia) at day 0 or day 8 post-infection in wild-type (n = 6 mice; filled symbols/bars) and Rag2^−/− (n = 6 mice; open symbols/bars) mice. Data pooled from two independent experiments, presented as mean +/- SEM and analysed by two-way ANOVA. **P = 0.002, ****P < 0.0001. c Fungal brain burdens in wild-type and Rag2^−/− mice at day 8 post-infection in mice infected with C. neoformans strain H99 (n = 10 wild-type mice, n = 7 Rag2-deficient mice) or C. neoformans strain Zc1 (n = 10 wild-type mice, n = 9 Rag2-deficient mice), and d number of Ki67+ microglia at day 8 post-infection in mice infected with C. neoformans strain H99 (n = 10 wild-type mice, n = 7 Rag2-deficient mice) or C. neoformans strain Zc1 (n = 10 wild-type mice, n = 9 Rag2-deficient mice). Data pooled from two independent experiments and analysed by two-way ANOVA. **P = 0.0052, ***P < 0.0001. e Example immunofluorescence of Ki67 and DAPI labelling in brain sections from wild-type and Rag2^−/− mice at day 8 post-infection. Infection lesions are outlined with white dashed lines. Images were quantified by randomly selecting three parenchymal regions near sites of infection within each brain section, and dividing the pixel intensity of the Ki67 stain by the pixel density of the DAPI stain for each analysed region to determine the Ki67/DAPI ratio. Three mice were analysed for wild-type mice (9 regions sampled) and 4 Rag2^−/− mice analysed (12 regions sampled). Different symbol shapes show matched regions for each mouse. Data analysed by unpaired t-test, presented as mean +/- SEM. **P < 0.008. f Total number of CD4 T cells, Ki67⁺ microglia and IAMs in the brains of mice at day 8 post-infection, treated with either isotype control antibody (n = 6 mice; filled symbols/bars) or anti-CD4 depleting antibody (n = 6 mice; open symbols/bars). Data pooled from two independent experiments and analysed by unpaired t-tests. *P = 0.0447 (IAMs), *P = 0.0364 (Ki67+ microglia), **P = 0.0059. g Weight loss, relative to weight at start of experiment, of isotype control or CD4-depleted mice (n = 9 per group) during infection. Data shown as mean +/- SEM, pooled from three independent experiments. Data analysed by two-way ANOVA with Bonferroni correction, **P = 0.0045. h Fungal brain burdens at day 8 post-infection in isotype or CD4-depleted mice (n = 6 mice per group). Data pooled from two independent experiments. In all box-and-whisker plots, whiskers refer to the maximum/minimum values, the box refers to interquartile ranges, the centre line refers to the mean.