The meninges host a distinct compartment of regulatory T cells that preserves brain homeostasis

Abstract

Our understanding of the meningeal immune system has recently burgeoned, particularly regarding how innate and adaptive effector cells are mobilized to meet brain challenges. However, information on how meningeal immunocytes guard brain homeostasis in healthy individuals remains limited. This study highlights the heterogeneous, polyfunctional regulatory T cell (T_reg) compartment in the meninges. A T_reg subtype specialized in controlling interferon-γ (IFN-γ) responses and another dedicated to regulating follicular B cell responses were substantial components of this compartment. Accordingly, punctual T_reg ablation rapidly unleashed IFN-γ production by meningeal lymphocytes, unlocked access to the brain parenchyma, and altered meningeal B cell profiles. Distally, the hippocampus assumed a reactive state, with morphological and transcriptional changes in multiple glial cell types. Within the dentate gyrus, neural stem cells underwent more death and were blocked from further differentiation, which coincided with impairments in short-term spatial-reference memory. Thus, meningeal T_regs are a multifaceted safeguard of brain homeostasis at steady state.

Figure 1.. A population of Foxp3⁺CD4⁺ T cells within the meningeal dura at steady-state.

(A) Flow cytometry of Foxp3⁺CD4⁺ T cells in the dural meninges of 6wk-old mice. Left panel: representative flow-cytometric plot; Right panels: summary data. n=17 (B) Confocal imaging of transverse sections of the dura mater. Left panel: graphical representation of the meningeal sinuses, with the region of interest delineated by a square; right panels: representative images. (C) Quantification of Treg locations. n=6 (D) Representative confocal image of a meningeal Treg cluster on a transverse section stained for the indicated marker proteins. (E) Tile-scan images of multiple duras registered to a reference map (left), plotting average densities of the Foxp3-GFP and MHC-II signals across 6 tissues (right). (F) Quantification of Treg cells in the dura mater across age. n≥4 Each data-point is from an individual mouse. iv, intravenous; SMA, smooth-muscle actin; MHC, major histocompatibility complex. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****, p<0.0001 by paired Student’s t test [(C)] or one-way ANOVA [(F)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 2.. Heterogeneity and clonal expansions within the meningeal Treg compartment.

(A-C) Merged scRNA-seq data from dura mater Tregs of 40 mice (15 weeks: n=15, 36 weeks: n=25). (A) UMAP plot. Percentages indicate the fractional representation of each cluster. (B) Heatmap of the 50 genes most differentially expressed by each cluster. The 10 genes (excluding Rps genes in cluster 2) most specific are indicated to the right. (C) Expression density plots of the indicated signatures. (D-F) Flow cytometry of the three meningeal Treg-subtypes’ key marker proteins. Left: representative dot plots; right: summary data. For (E) and (F) representative flow plots depict 3 samples concatenated. (G-I) scTcr-seq data from the meningeal Treg-subtypes. (G) Pie-charts showing the proportion of clonally expanded cells in each cluster. Individual clones are depicted by different colors; non-expanded clones are in gray. (H) Bar plot showing the clonal overlap between the various clusters. x-axis: clone names: y-axis: number of Tregsof that particular clone. Bar colors correspond to the clusters of (A). (I) Expanded TCR clones situated on the UMAP plot from (A). UMAP, Uniform Manifold Approximation and Projection; Tfr, T follicular regulatory cell; Th1, T helper1; M, meninges; S, spleen; IC, isotype control; FMO, fluorescence minus one. Mean ± SEM, p-values as per Fig. 1. Mean ± SEM; p-values as per Fig. 1 by unpaired Student’s t test [(D-F)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 3.. Ablation of Tregs triggers the production of IFNγ by meningeal lymphocytes.

((A) Treg depletion protocol. 6wk-old male DTR⁺ and DTR⁻ littermates were ip-injected with DT as indicated, and assays were performed on day 5 (D5) after the first injection. (B) Percentage of meningeal Treg cells. n≥15 (C) Total meningeal immunocyte counts. n≥15 (D) Percentage (left) and numbers (right) of select immunocyte populations. n≥8 (E) Percentage (upper) and numbers (bottom) of the major T cell subsets. n≥15 (F) Imaging of meningeal sinuses from DT-treated DTR⁻ and DTR⁺ mice. Graphical representation showing the region of interest is shown to the left. Then a representative confocal image of the three indicated stains. Higher-power images of the white-squared region stained with, in order: αSMA for sinus structure, αCD3 for T cells and anti-MHCII for antigen presenting cells. (G) Percentage (left) and numbers (right) of meningeal B cell populations. n=9 (H) Representative immunofluorescence image of Treg and B cells in the meninges. (I) Pathway enrichment analysis via Metascape (127) of the genes differentially expressed in panel J [fold-change (FC) >1.5, p-value <0.05]. (J) Volcano plot of population-level RNA-seq comparing meninges in the presence and absence of Treg cells. Annotated transcripts are those involved in the pathways highlighted in Fig. 3I (interferon signaling and brain development pathways). Triplicate samples. (K) Percentage (left) and numbers (center) of total cytokine-positive immunocytes from the dura mater. (L) Representation of the IFNγ producers from (K). (M) Percentage of major cytokine-producing cells from the dura mater. DT, diphtheria toxin; DTR, DT receptor; MF, macrophages; NF, neutrophils; DCs, dendritic cells; ILC, innate lymphoid cells; Fol, follicular; other abbreviations as per Fig. 1. Mean ± SEM; p-values as per Fig. 1, by unpaired Student’s t test [(B) to (E), (G), (K), (L), (M)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 4.. T cells invade the brain parenchyma in the absence of Treg cells.

(A-B) Flow-cytometric analysis of the brain parenchyma of mice depleted of Tregs or not, as per Fig. 3A. Males mice aged 6 to 7 wks old. (A) Total immunocyte numbers. n ≥ 18 (B) Frequency (left) and numbers (right) of select immunocyte populations. n ≥ 11 (C-E) T cell invasion into the brain parenchyma after depletion of Tregsin DT-treated DTR⁺ mice, as per Fig. 3A. (C) Diagrammatic representation of the brain with regions of interest delineated. Coronal section. (D) Quantification of the average CD3⁺ T cells. n=4 (E) Representative confocal images of coronal sections of the whole hippocampus. Top row: regions of interest. The quartet of higher power images below show anti-CD3 mAb (+DAPI) staining for T cells in each region of interest, framed in the colors corresponding to those of the top panel. Arrows point to T cells. (F) Left: representative confocal images of the hippocampus’ cFos staining pattern in the presence (left) or absence (right) of Treg cells. The dotted line delimits the DG. Right: quantification of cFos⁺ cells in the hippocampus and DG. n≥8 (G-J) Whole-tissue RNA-seq analysis of brain regions from mice depleted of Tregsor not, as per Fig. 3A. (G) Diagram showing regions of interest. (H) Pathway-enrichment analysis via Metascape (127) of the differentially expressed transcripts from panel I [FC>1.5; p-value<0.05]. (I) Volcano comparing DTR⁺ and DTR⁻ hippocampus. Annotated genes are those involved in pathways highlighted in (H) (interferon signaling and response to IFNγ). Triplicate samples. (J) Specific analysis via Gene Ontology (GO) and GSEA (MsigDB) database of pathways related to glial activation and hippocampal functions. DG, dentate gyrus; CA, Cornu Ammonis; other abbreviations as per Fig. 3. Mean ± SEM; p-values as per Fig. 3. Mean ± SEM; p-values as per Fig. 1, by unpaired Student’s t test [(A), (B), (D), (F)] or chi-square test [(H), (J)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 5.. Tregscontrol hippocampal glial cells.

(A) Left: representative confocal images of DG microglia (Iba1+); right: summary quantification of classical morphological changes in activation. n= 5 (B) “Activation upon demyelination” (56) (top) and “Disease-associated” (55) (bottom) microglia signatures are red-highlighted on a volcano plot comparing transcripts expressed in sorted hippocampal microglia from DT-treated DTR- versus DTR+ mice. Duplicate samples. (C) Same as panel A, but with images of astrocytes (GFAP+). (D-I) scRNA-seq analyses of dissected hippocampal cells depleted of mature neurons (Thy1+) and microglia (CD45+). Duplicate samples. (D) UMAP representation. (E) Normalized local cell densities on the UMAP space from DTR− (left) or DTR+ (center) mice. Density differential map between the two genotypes (right). (F) UMAP representation of re-clustered astrocytes (light green) from panel D. (G) Same as panel E except specifically for the astrocyte cluster (light green) in panel D. (H-I) Same as panels F and G except the OPCs (blue) from panel D were re-clustered. OPCs, oligodendrocyte precursor cells; SVMCs, smooth vascular muscle cells; other abbreviations as per Fig. 3. Mean ± SEM; p-values as per Fig. 1, by unpaired Student’s t test [(A), (C)] or chi-square test [(B)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 6.. Treg-mediated protection of RGL cells from IFNγ-induced death and against impairments to short-term memory.

(A-N) DTR⁺ and DTR⁻ littermates (6 – 7wks old) were injected with DT as per Fig. 3A and analyzed on day 5 after the first injection. n=18 for panels B-F. (A) Representative image of Nissl staining of the hippocampus. Relevant DG anatomical structures: SGZ (solid magenta line) and GCL (dotted yellow line). (B) GCL volume. (C) SGZ area. (D) Histological quantification of immature (DCX⁺) neurons. (E) Histological quantification of granular neurons. (F) Left and center: representative confocal images of RGL cells (arrows) from DTR⁻ and DTR⁺ littermates. All markers are shown in the upper panels, while DAPI staining has been removed in the lower panels. Right: summary quantification of RGL cells per SGZ. (G) Histological quantification of Fractin+ cells per SGZ. n≥8 (H) Histological quantification of pH3+ cells per SGZ. n≥18 (I) Violin plot of Bax and Bcl2 transcription expression from the clusters delimited in Fig S8D. (J) Flow-cytometric quantification of the frequency of RGL cells (CD45⁻SOX2⁺) expressing death markers: Annexin-V⁺ (left) and cleaved-caspase-3⁺ (right). n≥10 (K) Percentage of Ki-67⁺CD45⁻SOX2⁺ cells. n=11 (L) Correlation between the meningeal Treg frequency and the frequency of death markers on the RGL cells (CD45⁻SOX2⁺): Annexin-V⁺ (left) and cleaved-caspase-3⁺ (right). n=21 (M) Flow-cytometric analysis of Annexin-V⁺ RGL (CD45⁻SOX2⁺) cells from DT-treated DTR⁺ and DTR⁻ mice treated with an anti-IFNγ or isotype-control mAb. n≥8 (N) Frequency of entries by individual mice into the familiar (red dots) versus novel (grey dots) arms of the Y-maze 5 days after the initial DT injection. n≥12 (O) Same as panel M, except the readout was 56 days after the initial DT injection. n≥11 DG, dentate gyrus; GCL, granular cell layer; SGZ, subgranular zone; IFN, interferon; other abbreviations as per Fig. 6. Mean ± SEM; p-values as per Fig. 1, by unpaired Student’s t test [(B) to (H), (J), (K), (N), (O)], one-way ANOVA [(M)], simple linear regression [(L)], or Wilcoxon test [(I)]. Summary plots show data pooled from ≥2 independent experiments.

Figure 7.. Dura mater IL-2 availability is fine-tuned by Treg cells.

(A) Frequency of pSTAT5+ select immunocyte populations in mice with or without Treg depletion. n ≥ 8 (B-F) Flow-cytometric quantification of DTR⁺ and DTR⁻ male littermates (6-7wks old) ip-injected with DT and anti-IL2 or isotype control mAb. n ≥ 5 (B) Anti-IL2 blocking protocol and schematic of the different conditions. (C) Numbers of select meningeal immunocyte populations (top) and T cell subsets (bottom). (D) Same as panel C except for the brain parenchyma. (E) Frequency of Annexin-V⁺ hippocampal RGL cells. (F) Frequency of Ki-67⁺ hippocampal RGL cells. (G-P) Flow-cytometric analysis of the duramater and brain parenchyma of B6 WT littermate males (6 – 7wks old) after ip-injection of PBS (black), JES6-containing complex IL2 (cIL2) (gray) or S4B6-containing cIL2 (white). n ≥ 14 (G) Schematic representation of the both cIL2 and their target cells (H) Percentage of meningeal Tregs. (I) Total meningeal immunocyte counts. (J) Percentage of meningeal select immunocyte populations (left) and T cell subsets (right). (K) Percentage of IFNγ⁺ immunocyte populations. n ≥ 10 (L) Total brain immunocyte counts. n ≥ 10 (M) Same as in panel I except for the brain. n ≥ 10 (N) Same as in panel J except for the brain. n ≥ 10 (O) Frequency of Annexin-V⁺ hippocampal RGL cells. n ≥ 10 (P) Frequency of Ki-67⁺ RGL cells. n≥10 Mean ± SEM, p-values as per Fig. 1, by unpaired Student’s t test [(A)], one-way ANOVA [(E), (F), (H), (I), (L), (O), (P)], two-way ANOVA [(C), (D), (J), (K), (M), (N)]. For simplicity, only relevant statistical differences are shown. Summary plots show data pooled from ≥2 independent experiments.