Quantitative and Qualitative Perturbations of CD8+ MAITs in Healthy Mycobacterium tuberculosis-Infected Individuals

Abstract

CD8 T cells are considered important contributors to the immune response against Mycobacterium tuberculosis, yet limited information is currently known regarding their specific immune signature and phenotype. In this study, we applied a cell population transcriptomics strategy to define immune signatures of human latent tuberculosis infection (LTBI) in memory CD8 T cells. We found a 41-gene signature that discriminates between memory CD8 T cells from healthy LTBI subjects and uninfected controls. The gene signature was dominated by genes associated with mucosal-associated invariant T cells (MAITs) and reflected the lower frequency of MAITs observed in individuals with LTBI. There was no evidence for a conventional CD8 T cell-specific signature between the two cohorts. We, therefore, investigated MAITs in more detail based on Vα7.2 and CD161 expression and staining with an MHC-related protein 1 (MR1) tetramer. This revealed two distinct populations of CD8⁺Vα7.2⁺CD161⁺ MAITs: MR1 tetramer⁺ and MR1 tetramer^-, which both had distinct gene expression compared with memory CD8 T cells. Transcriptomic analysis of LTBI versus noninfected individuals did not reveal significant differences for MR1 tetramer⁺ MAITs. However, gene expression of MR1 tetramer^- MAITs showed large interindividual diversity and a tuberculosis-specific signature. This was further strengthened by a more diverse TCR-α and -β repertoire of MR1 tetramer^- cells as compared with MR1 tetramer⁺ Thus, circulating memory CD8 T cells in subjects with latent tuberculosis have a reduced number of conventional MR1 tetramer⁺ MAITs as well as a difference in phenotype in the rare population of MR1 tetramer^- MAITs compared with uninfected controls.

FIGURE 1.. Transcriptomic profile of memory CD8 T cells comparing LTBI subjects with TB neg controls reveal changes in MAIT frequency.

(A) Volcano plot obtained from the DEseq2 analysis showing log2 fold change versus −log10 p value. The 41 differentially expressed genes are represented in black (adjusted p value <0.05 indicated by dotted line). Known genes expressed by CD8⁺Vα7.2⁺CD161⁺ T cells are shown in black circles. Genes upregulated in TB neg controls (left) and genes upregulated in LTBI (right). (B) Heatmap displaying regularized logarithm–transformed raw counts of the 41 differentially expressed genes from (A), genes are ordered by hierarchical clustering, and subjects are ordered by principal component 1 (PC1). TB neg (blue), LTBI (red). (C) Frequency of MR1tet⁺ of CD8⁺ T cells as measured by flow cytometry LTBI (n = 29), TB neg (n = 11). Two-tailed Mann–Whitney U test. Median ± interquartile range is indicated. (D) Correlation between the PC1 of the combined expression of the 41 differentially expressed genes, identified in (A), and the frequency of MR1tet⁺ of CD8⁺ T cells in corresponding subjects. Correlation is indicated by Spearman r and associated two-tailed p value.

FIGURE 2.. Minor differences between LTBI versus TB neg in M. tuberculosis–specific CD8⁺ T cell responses and transcriptomic analysis of memory CD8 T cells excluding Vα7.2⁺CD161⁺.

(A) Transcriptomic analysis of Vα7.2⁻CD161⁻ memory CD8 T cells. Volcano plot obtained from the DEseq2 analysis showing log2 fold change versus −log10 p value. The eight differentially expressed genes are represented in black (adjusted p value <0.05 and absolute log2 fold change >1 are indicated by dotted lines). Genes upregulated in TB neg controls (left) and genes upregulated in LTBI (right). (B and C) Magnitude of epitope pool responses measured as total spot-forming cells (SCFs) per 10⁶ PBMCs in an IFN-γ ELISPOT assay in individuals with LTBI and TB neg controls. Each dot represents one subject. Median ± interquartile range is indicated. Two-tailed Mann–Whitney U test. (B) CD8 megapool (LTBI, n = 28, red dots; TB neg, n = 12, blue dots), (C) CD4 megapool (LTBI, n = 28; TB neg, n = 27).

FIGURE 3.. Phenotypic and transcriptomic profiling of MAITs reveals four distinct subsets and overlapping gene expression programs between CD8⁺ MR1tet⁺ and CD8⁻ MR1tet⁺ cells.

(A) Percentage of the two dominant MR1 tetramer⁺ CD3⁺ T cell subsets expressing combinations of CD4, CD8, CD161, and/or Vα7.2 with a median expression above 3%. Each dot represents one subject (n = 40; 11 TB neg and 29 LTBI). Median ± interquartile range is indicated. Dotted line indicates arbitrary cut-off for median expression at 3%. (B) Percentage of the four dominant Vα7.2⁺CD161⁺CD3⁺ T cell subsets expressing or binding combinations of CD4, CD8, and MR1 tetramer with a median expression above 3%. Each dot represents one subject (n = 40; 11 TB neg and 29 LTBI). Median ± interquartile range is indicated. Dotted line indicates arbitrary cut-off for median expression at 3%. (C) Percentage of the two dominant MR1 tetramer⁺ CD3⁺ T cell subsets expressing combinations of CD4, CD8, CD161, and/or Vα7.2 with a median expression above 3% in 11 TB neg (blue) and 29 LTBI (red) individuals. Each dot represents one subject. Median ± interquartile range is indicated. Dotted line indicates arbitrary cut-off for median expression at 3%. Two-tailed Mann–Whitney U test. (D) Percentage of the four dominant Vα7.2⁺CD161⁺CD3⁺ T cell subsets expressing or binding combinations of CD4, CD8, and MR1 tetramer with a median expression above 3% in 11 TB neg (blue) and 29 LTBI (red) individuals. Each dot represents one subject. Median ± interquartile range is indicated. Dotted line indicates arbitrary cut-off for median expression at 3%. Two-tailed Mann–Whitney U test. (E) Volcano plot obtained from the DEseq2 analysis showing log2 fold change versus −log10 p value. The six differentially expressed genes are represented in black (adjusted p value <0.05 and absolute log2 fold change >1 are indicated by dotted lines). Genes upregulated in CD8⁺ MR1tet⁺ cells (left) and genes upregulated in CD8⁻ MR1tet⁺ cells (right).

FIGURE 4.. Distinct gene expression profile of MR1tet⁺ MAITs compared with memory CD8⁺ T cells.

(A) PCA plot illustrating differences between memory CD8⁺ T cells and MR1tet⁺ MAITs and between LTBI and TB neg individuals. (B) Volcano plot obtained from the DEseq2 analysis showing log2 fold change versus −log10 p value. The differentially expressed genes are represented in black (adjusted p value <0.01, absolute log2 fold change >1 are indicated by dotted lines). (B) MR1tet⁺ cells compared with memory CD8⁺ T cells, (C) CCR1, CXCR4, CXCR6, TIGIT, IL7R, and ABCB1 expression at the mRNA (upper panels: gene expression values in counts normalized by sequencing depth calculated by the DEseq2 package) and protein (lower panels: protein expression as percent frequency of subset) levels in memory CD8⁺ T cells and MR1tet⁺ MAITs. Gene expression data were derived from memory CD8⁺ T cells from 17 individuals and MR1tet⁺ cells (n individuals = 16) using an Illumina sequencing platform. Protein expression data were derived from memory CD8⁺ T cells from 20 individuals and MR1tet⁺ cells (n individuals = 20) using flow cytometry. Median ± interquartile range is shown. Two-tailed Mann–Whitney U test.

FIGURE 5.. Gene expression profile and M. tuberculosis–specific signature of MR1tet⁻ MAITs compared with memory CD8 T cells and MR1tet⁺ MAITs.

(A) PCA plot illustrating differences between memory CD8 T cells and MR1tet⁻ MAITs and between LTBI and TB neg individuals. (B, D, and E) Volcano plots obtained from the DEseq2 analysis showing log2 fold change versus −log10 p value. The differentially expressed genes are represented in black [adjusted p value <0.01 (B and E) and p < 0.05 (D), absolute log2 fold change >1 are indicated by dotted lines]. (B) MR1tet⁻ cells compared with memory CD8 T cells. (C) Venn-diagram showing overlap between the 226-gene signature identified in Fig. 4B and the signature in Fig. 5B, based on hypergeometric distribution test (considering the 18,315 transcripts detected within memory CD8 T cells as the total number of genes). (D) MR1tet⁻ cells comparing individuals with LTBI versus TB neg. (E) Volcano plot comparing MR1tet⁻ cells with MR1tet⁺ cells. (F) CCR1, CXCR4, CXCR6, TIGIT, IL-7R, and ABCB1 expression at the mRNA (upper panels: gene expression values in counts normalized by sequencing depth calculated by the DEseq2 package) and protein (lower panels: protein expression as percent frequency of subset) levels in memory CD8 T cells and MR1tet⁻ MAITs. Gene expression data were derived from memory CD8 T cells from 17 individuals and MR1tet⁻ cells (n individuals = 12) using an Illumina sequencing platform. Protein expression data were derived from memory CD8 T cells from 20 individuals and MR1tet⁻ cells (n individuals = 7) using flow cytometry. Median ± interquartile range is shown. Two-tailed Mann–Whitney U test.

FIGURE 6.. Expression of previously described MAIT signature genes in MR1tet⁺ and tet⁻ cells.

Expression at the mRNA level (gene expression values in counts normalized by sequencing depth calculated by the DEseq2 package) in memory CD8 T cells, MR1tet⁺, and MR1tet⁻ MAITs. Gene expression data were derived from memory CD8 T cells from 17 individuals, MR1tet⁺ (n individuals = 16), and MR1tet⁻ cells (n individuals = 12) using an Illumina sequencing platform. Median ± interquartile range is shown. Two-tailed Mann–Whitney U test.

FIGURE 7.. TCR β-chain gene segment usage.

(A) TRBV family usage in CD8⁺, MR1tet⁺, and MR1tet⁻ cells summing over J combinations. (B) Heatmap illustrating frequency of a particular V-J combination in CD8⁺, MR1tet⁺, and MR1tet⁻ cells. Averaging is done across the subjects. V gene segments are grouped by family by summing up family gene segments prior to plotting.