CD4+ T cells re-wire granuloma cellularity and regulatory networks to promote immunomodulation following Mtb reinfection

Abstract

Immunological priming-in the context of either prior infection or vaccination-elicits protective responses against subsequent Mycobacterium tuberculosis (Mtb) infection. However, the changes that occur in the lung cellular milieu post-primary Mtb infection and their contributions to protection upon reinfection remain poorly understood. Using clinical and microbiological endpoints in a non-human primate reinfection model, we demonstrated that prior Mtb infection elicited a long-lasting protective response against subsequent Mtb exposure and was CD4⁺ T cell dependent. By analyzing data from primary infection, reinfection, and reinfection-CD4⁺ T cell-depleted granulomas, we found that the presence of CD4⁺ T cells during reinfection resulted in a less inflammatory lung milieu characterized by reprogrammed CD8⁺ T cells, reduced neutrophilia, and blunted type 1 immune signaling among myeloid cells. These results open avenues for developing vaccines and therapeutics that not only target lymphocytes but also modulate innate immune cells to limit tuberculosis (TB) disease.

Keywords: CD4 T cells; CD4 depletion; CD8 T cells; Mycobacterium tuberculosis; concomitant immunity; granuloma; macaque; non-human primate; reinfection; single-cell RNA sequencing.

Graphical abstract

Figure 1

Anti-CD4 antibody infusion depleted CD4⁺ T cells across anatomic compartments in cynomolgus macaques (A) Experimental design (BioRender.com). (B) PET-CT scan of representative NHPs pre- and post-reinfection. Old granulomas: blue arrows; new granulomas: green arrows. Left: IgG; middle: αCD4; right: naive. (C) Fraction of CD3 that were CD4⁺ post-antibody infusion in blood (number of animals: naive n = 6; IgG n = 6; αCD4 n = 7). Median and range shown (^∗∗∗∗p < 0.0001; mixed-effects model with Dunnett’s multiple comparisons test). (D) CD3⁺, CD4⁺ cells in new library S granulomas (number of granulomas: naive n = 37; IgG n = 21; aCD4 n = 38; number of animals: naive n = 6; IgG n = 5; αCD4 n = 7). (E) CD3⁺, CD4⁺ cells in uninvolved lung tissue from Mtb infected macaques (number of animals: naive n = 4, IgG n = 6; αCD4 n = 5). (F) CD3⁺, CD4⁺ cells in spleen (number of animals: naive n = 5, IgG n = 6, αCD4 n = 7). (G) CD3⁺, CD4⁺ cells from CFU⁺ LNs (number of animals: naive n = 6, IgG n = 5; αCD4 n = 7; number of LNs: naive n = 14; IgG n = 6; αCD4 n = 11). (D–G) Transparent smaller dots represent granulomas, colored by animal. Larger dots represent mean per animal, and lines represent medians. One-way ANOVA with Dunnett’s multiple comparisons test. See also Figure S1 and Table S1. Experiment was performed once.

Figure 2

Reinfection with Mtb reduced granuloma formation, as well as bacterial burden and dissemination in a CD4⁺ T cell-dependent manner (A) Number of new granulomas identified using PET-CT following infection with Mtb library S. Lines represent medians. One-way ANOVA with Dunnett’s multiple comparison test, adjusted p values reported. (B) Median number of viable Mtb colony-forming units (CFUs) per library S granuloma (Kruskal-Wallis with Dunn’s multiple correction). Solid dots represent the median CFU/granuloma per animal; lines represent medians. Transparent dots represent the median CFU of individual granulomas. Number of granulomas: naive n = 70, IgG n = 34; αCD4 n = 65. (C) Median number of chromosomal equivalents (CEQs) per macaque (Kruskal-Wallis with Dunn’s multiple correction). Solid dots represent the median CFU per animal; lines represent medians. Transparent dots represent the median CFU of individual granulomas. (D) CFU:CEQ ratio, a proxy for bactericidal activity (Kruskal-Wallis with Dunn’s multiple correction). Solid dots represent the median CFU per animal; lines represent medians. Transparent dots represent the median CFU of individual granulomas. (C and D) Number of granulomas: naive n = 59, IgG n = 26; αCD4 n = 51. (E) Total CFU from granulomas, uninvolved lung tissue, and thoracic lymph nodes. (F) Lung CFU from granulomas and uninvolved lung tissue. (G) Thoracic lymph node CFU. (E–G) Lines represent medians. One-way ANOVA with Dunnett’s multiple comparisons test. (A–G) Number of macaques: naive n = 6; IgG n = 6; anti-CD4 n = 7. (H) Percent of all sampled tissues that shared library S barcodes (Kruskal-Wallis with Dunn’s multiple comparison test). Data points in each treatment arm reflect number of distinct barcodes (naive, n = 77; IgG, n = 40; anti-CD4, n = 87) identified by sequencing; these barcodes were aggregated across independent animals (naive, n = 6; IgG, n = 6; anti-CD4, n = 7). (I) Individual granuloma Mtb CFU for single granulomas subjected to Seq-Well S³ scRNA-seq (Mann-Whitney U test). Naive n = 10, IgG n = 8, αCD4 n = 15 granulomas from 2 naive, 2 IgG, and 3 αCD4 NHPs. Experiment was performed once. See also Figure S2 and Table S1.

Figure 3

Mtb reinfection promoted cellular remodeling of the TB granuloma microenvironment (A) UMAP embedding of Seq-Well S³-derived granuloma transcriptomes colored by coarse cell type. (B) Coarse cell type frequencies colored by experimental group. Differentially abundant IgG vs. naive (purple) and IgG vs. αCD4 (green) marked with colored square. Cell types were differentially abundant if significant using two of three methods: Mann-Whitney U test; scCODA, and Fisher’s exact test. (C) Fraction of granuloma T, NK cells expressing CD4 from Seq-Well S³-derived transcriptomes (Mann-Whitney U test). (D) Fraction of granuloma T, NK cells expressing CD8A from Seq-Well S³-derived transcriptomes (Mann-Whitney U test). Individual dots in (B)–(D) represent single granulomas. Naive n = 10, IgG n = 8, αCD4 n = 15 granulomas from 2 naive, 2 IgG, and 3αCD4 NHPs. Experiment was performed once. See also Figure S3 and Tables S2, S3, and S4.

Figure 4

CD4⁺ T cells regulated T cell cellularity, cytokine flux, and immune tone in the TB granulomas following Mtb reinfection (A) UMAP embedding depicting T, NK cell subpopulations identified by sub-clustering. (B) Heatmap depicting gene expression (mean Z score) of T cell lineage markers CD4, CD8A, and CD8B. Columns represent gene expression in individual NHP groups—naive (light blue), IgG (yellow), and αCD4 (red). Bar plot of T, NK subpopulation frequencies among all granuloma cellular subpopulations colored by experimental group. Differentially abundant IgG vs. naive (purple) and IgG vs. αCD4 (green) marked with colored square. Cell types are differentially abundant if significant using two of three methods: Mann-Whitney U test; scCODA, and Fisher’s exact test. (C and D) T, NK cell pseudobulk log₂(CPM + 1) for naive (light blue), IgG (yellow), and αCD4 (red) NHP granulomas (^∗∗∗p < 0.001,^∗∗p < 0.01, ^∗p < 0.05; Wilcoxon rank-sum test). Heatmap depicting log_eFC (calculated using MAST^38,39) of lineage markers, cytolytic molecules, select transcription factors, immunoregulatory molecules, and chemokines, and cytokines (rows) for each cell type (columns) in NHP granulomas IgG vs. naive (C) or IgG vs. αCD4 lesions (D). White circles indicate log_e|FC| > log_e(1.3), relative to naive or αCD4 granulomas. Black rectangles indicate 0.05 > false discovery rate (FDR) and log_e|FC| > log_e(1.3), relative to naive or αCD4 granulomas. Individual dots in (B)–(D) represent single granuloma. Naive n = 10, IgG n = 8, αCD4 n = 15 granulomas from 2 naive, 2 IgG, and 3 αCD4 NHPs. See also Figure S3 and Tables S5, S6, S7, and S8.

Figure 5

Monocyte-derived transcriptomes featured attenuated type 1 immunity in Mtb reinfection granulomas (A) UMAP embedding depicting monocyte-derived cells identified by sub-clustering. UMAP embeddings depicting monocyte-derived cell subpopulation densities, split by NHP cohort. (B) Bar plot of monocyte-derived progenitor frequencies, among all granuloma cell subpopulations, colored by experimental group. Individual dots represent single granulomas. Differentially abundant IgG vs. naive (purple) and IgG vs. αCD4 (green) marked with colored square. Cell types are differentially abundant if significant using two of three methods: Mann-Whitney U test; scCODA, and Fisher’s exact test. (C and D) Enriched pathways from identified using differentially expressed genes (Mann-Whitney U test (Wilcoxon rank-sum) (p value < 0.05)) from naive, IgG, and αCD4 sizes. Circle size represents the number of genes in Hallmark Geneset, and color (red-blue) represents geneset enrichment score. Genesets that are “up” (x axis) are enriched among IgG granulomas, whereas “down” genesets are enriched among naive (C) and αCD4 (D) granulomas, respectively. (E and F) Monocyte-derived pseudobulk log₂(CPM + 1) for naive (light blue), IgG (yellow), and αCD4 (red) NHP granulomas (^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05; Wilcoxon rank-sum test). Heatmap depicting log_eFC of select transcription factors, immunoregulatory molecules, chemokines, and cytokines (rows) for each cell type (columns) in NHP granulomas IgG vs. naive (E) or IgG vs. αCD4 lesions (F). White circles indicate log_e|FC| > log_e(1.3), relative to naive or αCD4 granulomas. Black rectangles indicate 0.05 > FDR and log_e|FC| > log_e(1.3), relative to naive or αCD4 granulomas. Individual dots in (B), (E), and (F) represent single granulomas. Naive n = 10, IgG n = 8, αCD4 n = 15 granulomas from 2 naive, 2 IgG, and 3 αCD4 NHPs. See also Figure S4 and Tables S5, S6, S7, and S8.

Figure 6

Prior infection influenced neutrophil response and dampened induction of a TB susceptibility type 1 interferon gene module (A) UMAP embedding depicting neutrophil cell subpopulations identified by sub-clustering. (B) Bar plot of neutrophil subset frequencies, among all granuloma cell subpopulations, colored by experimental group. Individual dots represent single granulomas. Differentially abundant IgG vs. naive (purple) and IgG vs. αCD4 (green) marked with colored square. Cell types are differentially abundant if significant using two of three methods: Mann-Whitney U test; scCODA, and Fisher’s exact test. (C) Volcano plot depicting pseudobulk differential gene expression (DESeq2) ICAM1^hi, NBN^hi vs. SORL1^hi, CFD^hi neutrophils (for all NHP experimental groups). Volcano plot x axis indicates the log₂FC, and y axis indicates the −log₁₀(p value). (D and E) Neutrophil pseudobulk log₂(CPM + 1) for naive (light blue), IgG (yellow), and αCD4 (red) NHP granulomas (^∗∗∗p < 0.001,^∗∗p < 0.01, ^∗p < 0.05; Wilcoxon rank-sum test). Heatmap depicting log_eFC of select transcription factors, immunoregulatory molecules, chemokines, and cytokines (rows) for each cell type (columns) in NHP granulomas IgG vs. naive (D) or IgG vs. αCD4 lesions (E). White circles indicate log_e|FC| > log_e(1.3), relative to naive or αCD4 granuloma. Black rectangles indicate 0.05 > FDR and log_e|FC| > log_e(1.3), relative to naive or αCD4 granulomas. (F) Violin plots of IFN-inducible neutrophil module scores by NHP group. Mann-Whitney U test. Individual dots in (B), (D), and (E) represent single granulomas. Naive n = 10, IgG n = 8, αCD4 n = 15 granulomas from 2 naive, 2 IgG, and 3 αCD4 NHPs. See also Figure S4 and Tables S5, S6, S7, and S8.

Figure 7

Differential cell-cell interactions occurred in immunologically primed granulomas (A) Heatmap depiction of differential (naive vs. IgG) cell-cell interaction pairs among coarse cell types. Columns represent cell-cell interactions from the top-prioritized links—“sender” ligands and receptors differential L-R pairs specific to IgG or naive granulomas. Heatmap and dot size represent L-R interactions from the 50 top-prioritized links. Black rectangles indicate top 5 interactions, based on number of interactions between two cell types per NHP group. Green rectangles depict putative T, NK-T, NK interactions. (B) Cartoon depiction of (A) with differential L-R in naive granulomas. (C) Bar plot depiction of differential cell-cell interactions among naive granulomas. Left: IL10 sender cellular subpopulations; right: IL10RA/RB cell subpopulations. Receptor-ligand and inferred interaction pairs are derived from the top 200 top-prioritized linkages. (D) Similar heatmap to that of (A), highlighting linkages specific to IgG (vs. naive) granulomas. (E) Schematic representation of the differential L-R pairs in IgG granulomas from (D). (F) Bar plot representation of differential IL10-IL10RA/RB interactions among IgG lesions, similar to that of (C). (G) Heatmap of αCD4 (vs. IgG) granulomas. (H) Schematic representation of the differential L-R pairs in αCD4 granulomas from (G). (I) Heatmap of IgG (vs. αCD4) granulomas. (J) Schematic representation of differential L-R pairs specific to IgG granulomas from (I). (K) Bar plot representation of differential IL10-IL10RA/RB interactions among IgG lesions, similar to that of (C). See also Figure S5 and Table S9.