Robust Control of a Brain-Persisting Parasite through MHC I Presentation by Infected Neurons

Abstract

Control of CNS pathogens by CD8 T cells is key to avoid fatal neuroinflammation. Yet, the modalities of MHC I presentation in the brain are poorly understood. Here, we analyze the antigen presentation mechanisms underlying CD8 T cell-mediated control of the Toxoplasma gondii parasite in the CNS. We show that MHC I presentation of an efficiently processed model antigen (GRA6-OVA), even when not expressed in the bradyzoite stage, reduces cyst burden and dampens encephalitis in C57BL/6 mice. Antigen presentation assays with infected primary neurons reveal a correlation between lower MHC I presentation of tachyzoite antigens by neurons and poor parasite control in vivo. Using conditional MHC I-deficient mice, we find that neuronal MHC I presentation is required for robust restriction of T. gondii in the CNS during chronic phase, showing the importance of MHC I presentation by CNS neurons in the control of a prevalent brain pathogen.

Keywords: CD8 T cell; Toxoplasma gondii; antigen presentation; brain infection; encephalitis; neuroinflammation; neuron; parasite.

Figure 1.. Expression of GRA6-OVA antigen by T. gondii leads to efficient parasite control and lower CNS inflammation

(a) Schematics of GRA6-OVA and vacOVA antigenic constructs expressed in the Tomato⁺ Prugnaud (Pru) parental strain. GRA6-OVA: fusion protein between GRA6(II) and LEQLE-SIINFEKL sequence, driven by the GRA6 promoter. vacOVA: fusion protein between SAG1ΔGPI and amino acids [140–386] of OVA, which contain the LEQLE-SIINFEKL sequence, driven by the tubulin promoter. Representative immunofluorescent images of a parasitophorous vacuole and an ex vivo cyst to illustrate activity of both promoters in tachyzoites and bradyzoites. Red: intrinsic Tomato fluorescence. Green: anti-SIINFEKL staining for GRA6-OVA, anti-full length OVA staining for vacOVA. (b) Schematics of experimental infections in mice infected i.p. with either of the 3 parasite strains. Analyses of brain parasite load and immunological status 3 weeks post-infection. (c, d) Number of brain cysts enumerated microscopically (mean +/− SEM) in H-2^b C57BL/6 mice (c) and H-2^k CBA mice (d). (e) Flow cytometry gating strategy to analyze myeloid cells in the CNS. (f) Analysis of inflammatory CD45^hi CD11b⁺ myeloid cells and resident CD45^int CD11b⁺ (‘microglia’) cells. Numbers on the representative contour plots show the mean percentage +/− SD of each subset out of single live Ly6G⁻ NK1.1⁻ cells. Graph shows the ratio (mean +/− SEM) of CD45^hi over CD45^int cells. (g) Representative brain cortical sections from uninfected and infected mice, stained for Iba1. Scale bar: 100 μm for 5X images, 25 μm for 20X images (h) Analysis of Ly6C^hi inflammatory monocytes. Numbers on the representative contour plots show the mean percentage +/− SD of Ly6C^hi cells out of single live Ly6G⁻ NK1.1⁻ CD45^hi CD11b⁺ cells. Graph shows absolute numbers (mean +/− SEM). (i) Proportion of MHC II⁺ cells (DC) among Ly6C^hi monocytes (mean percentage +/− SD). (j) IFNγ production by CNS-isolated CD4 T cells after incubation with brefeldin A. Numbers on the representative contour plots show the mean percentage +/− SD of IFNγ⁺ out of CD4⁺ T cells. Graph shows absolute numbers (mean +/− SEM). N = 4–8 mice / group. (c, j) Two experiments pooled. (d) Three experiments pooled. (f, h, i) One experiment representative of 3 independent experiments. (g) Representative field of view taken from a brain per condition. See also Figure S1.

Figure 2.. CD8 T cell responses against GRA6-OVA antigen are comparable whether or not bradyzoites express the antigen

(a) Schematics of the antigenic constructs introduced in GFP⁺ Prugnaud strain (Tg.GFP). GRA6-OVA is driven either by the GRA6 promoter, active at the tachyzoite and bradyzoite stages, or by the tachyzoite-restricted SAG1 promoter. (b) Immunofluorescence staining of ex vivo-isolated cysts from CBA brains. Green: intrinsic parasite fluorescence and lectin-stained cyst wall. Red: GRA6-OVA detected with anti-SIINFEKL antibody. Blue: GRA2. Scale bar 10 μm. Right panel: quantification of SIINFEKL fluorescence within cyst. Each dot represents one cyst. (c) MHC I K^b presentation of GRA6-OVA-derived SIINFEKL peptide by MutuDC infected with the indicated parasite lines, assessed by absorbance measurements following incubation with LacZ-inducible OVA-specific B3Z CD8 T cell hybridomas. (d) Schematics of experimental infections in C57BL/6 mice infected i.p. with either of the 3 parasite strains. Evaluation of CD8 T cell responses 2 months post-infection. (e-j) IFNγ-producing CD8 T cells from spleen (e, f, g) or brain (h, i, j) following in vitro restimulation with OVA-derived SIINFEKL peptide (f, i) and Tgd057-derived SVLAFRRL peptide (g, j). (e, h) Numbers on the representative contour plots show the mean percentage +/− SD of IFNγ⁺ cells out of CD8⁺ T cells. (f, g, i, j) Absolute numbers (mean +/− SEM) of IFNγ⁺ CD8⁺ T cells. N = 5 mice / group. Representative of two independent experiments. See also Figure S2.

Figure 3.. Tachyzoite-restricted GRA6-OVA antigen confers sustained parasite control and protection against T. gondii encephalitis

(a) Schematics of experimental infections in C57BL/6 mice infected i.p. with either of the 3 parasite strains. Analyses of brain parasite load and immune infiltrates 2 months post-infection. (b) Brain cysts enumerated microscopically in infected C57BL/6 mice (mean +/− SEM). (c, d) Parasite burden measured by qPCR on genomic DNA from infected C57BL/6 (c) or CBA (d) brains (mean +/− SEM). (e, f) Analysis of inflammatory CD45^hi CD11b⁺ myeloid cells and resident CD45^int CD11b⁺ (‘microglia’) cells. Numbers on the representative contour plots show the mean percentage +/− SEM of each subset out of single live Ly6G⁻ NK1.1⁻ cells. Graphs show the number of CD45^int CD11b⁺ microglia (mean +/− SEM) (e) or the ratio of CD45^hi over CD45^int cells (mean +/− SEM) (f). (g, h) Expression level of MHC II (g) and CD86 (h) on the surface of microglia. Numbers on the representative histograms show the geomean +/− SD. Graphs display the fold-increase of each marker with respect to microglia from uninfected mice (mean +/− SEM). (i) Analysis of Ly6C^hi inflammatory monocytes. Numbers on the representative contour plots show the mean percentage +/− SD of Ly6C^hi cells out of single live Ly6G⁻ NK1.1⁻ CD45^hi CD11b⁺ cells. Graph shows absolute numbers (mean +/− SEM). (j) Expression level of CD86 on the surface of Ly6C^hi monocytes. Numbers on the representative histograms show the geomean +/− SD. Graph displays the fold-increase over uninfected (mean +/− SEM). (k) Proportion of MHC II⁺ cells (DC) among Ly6C^hi monocytes. Numbers on the representative histograms and the graph show mean percentages +/− SD. (l) IFNγ production by CNS-infiltrating CD4 T cells after incubation with brefeldin A. Numbers on the representative contour plots show the mean percentage of IFNγ⁺ out of CD4⁺ T cells +/− SD. Graph shows absolute numbers (mean +/− SEM). For all panels, N = 9 mice / group with 2 experiments pooled, except for (l) where N = 5 mice / group from one experiment.

Figure 4.. Defective control of vacOVA-expressing T. gondii is not due to improper mobilization and effector differentiation of CD8 T cells in CNS

(a) Schematics of experimental infections in C57BL/6 mice infected i.p. with either of the 3 parasite strains. Evaluation of CD8 T cell responses in spleen and brain 3 weeks post-infection. (b, d) K^b-SIINFEKL dextramer labeling of spleen (b) and brain-infiltrating (d) CD8 T cells. Numbers on the representative contour plots show the mean percentage of dextramer⁺ out of CD8⁺ T cells +/− SD. Graph shows mean +/− SEM of absolute numbers. (c, e) Absolute numbers of IFNγ-producing CD8 T cells from spleen (c) and brain (e) following in vitro restimulation with OVA-derived SIINFEKL peptide (mean +/− SEM). (f) Analysis of triple-producing IFNγ⁺ TNF⁺ granzyme B⁺ brain CD8 T cells after restimulation with SIINFEKL peptide. Numbers on the representative contour plots show mean percentage of TNF⁺ granzyme B⁺ out of IFNγ⁺ CD8 T cells +/− SD. Graph shows absolute numbers of triple-producing CD8 T cells (mean +/− SEM). N=5 mice/ group. (b, c, d, e) Representative of 4 independent experiments. (f) Representative of 3 independent experiments.

Figure 5.. Poorer K^b-SIINFEKL presentation by neurons infected with T. gondii expressing vacOVA compared to neurons infected with T. gondii expressing GRA6-OVA

(a) Timeline of antigen presentation assay with Tg-infected primary neuronal cultures. AraC was added to inhibit growth of glial cells. (b, c) Tropism and infection rate of the two Tomato⁺ parasites following co-staining with MAP2 (green) and GFAP (blue). Pictures represent maximum intensity projections. Out of 24 fields, no vacuole was detected in a glial cell. Scale bar 10 μm. (d) K^b–SIINFEKL presentation by primary neurons infected for 24 h with the indicated tachyzoites, assessed by absorbance measurements following incubation with SIINFEKL-specific LacZ-inducible B3Z CD8 T cell hybridomas. (b, c, d) Representative of 2 independent experiments. See also Figure S3.

Figure 6.. MHC I presentation by CNS neurons is required for efficient brain control of T. gondii at chronic phase

(a) Schematic representation of the L^dLox DNA cassette introduced in C57BL/6 fertilized eggs to generate the B6.L^dLox mice. The cassette is a 12-kb genomic sequence comprising the L^d gene modified with 2 LoxP sites flanking exons 1 to 3 (domains L, N and C1 according to nomenclature from (Evans et al., 1982)) and the endogenous 5’ and 3’ UTR regulatory sequences from BALB/c. (b, c) Flow cytometry labeling of H-2 K^b (b) and H-2 L^d (c) on the surface of CD19⁻ CD3⁺ T cells, CD3⁻ CD19⁺ B cells and CD3⁻ CD19⁻ CD11c⁺ DC in the spleen of the indicated mouse strains. Numbers on the histograms show the geomean +/− SD. (b, c) Representative of two independent experiments. N=2–4 mice / group. (d) Schematics of experimental infections in C57BL/6, B6.L^dLox and congenic B6.H-2^d mice infected i.p. with Tg.pTUB/vacOVA, which naturally expresses GRA6. Evaluation of CD8 T cell responses and parasite load 3 weeks later. (e, f) Absolute numbers of IFNγ-producing CD8 T cells isolated from spleen (e) and brain (f), following in vitro restimulation with L^d-restricted GRA6-derived HF10 peptide (mean +/− SEM). (g) Number of brain cysts in mice of the indicated genotypes (mean +/− SEM). (e-g) Two experiments pooled with N=3–4 mice / group. (h) Schematics of the breeding strategy and outcome in terms of MHC I molecules expressed by neurons vs. non-neuronal cells in Cre⁻ and Cre⁺ mice. (i) Schematics of experimental infections in C57BL/6, B6.L^dLox.Cre⁺ or Cre⁻ infected per os with 15 cysts of the 76K strain. Evaluation of CD8 T cell responses and brain parasite load 3 weeks later. (j) Number of brain cysts enumerated microscopically. (k) Brain parasite burden measured by qPCR on genomic DNA. (l) Absolute number of total CD8 T cells isolated from infected brains. (m) Absolute numbers of brain-isolated IFNγ-producing CD8 T cells, following in vitro restimulation with L^d-restricted GRA6-derived HF10 peptide. (j-m) Graphs show the mean +/− SEM. N=5–6 mice / group. Representative of 2 independent experiments. See also Figure S4, S5 and S6.