Active Tuberculosis Is Characterized by Highly Differentiated Effector Memory Th1 Cells

Abstract

Despite advances in diagnosing latent Mycobacterium tuberculosis infection (LTBI), we still lack a diagnostic test that differentiates LTBI from active tuberculosis (TB) or predicts the risk of progression to active disease. One reason for the absence of such a test may be the failure of current assays to capture the dynamic complexities of the immune responses associated with various stages of TB, since these assays measure only a single parameter (release of IFN-γ) and rely on prolonged (overnight) T cell stimulation. We describe a novel, semi-automated RNA flow cytometry assay to determine whether immunological differences can be identified between LTBI and active TB. We analyzed antigen-induced expression of Th1 cytokine mRNA after short (2- and 6-h) stimulation with antigen, in the context of memory T cell immunophenotyping. IFNG and TNFA mRNA induction was detectable in CD4+ T cells after only 2 h of ex vivo stimulation. Moreover, IFNG- and TNFA-expressing CD4+ T cells (Th1 cells) were more frequent in active TB than in LTBI, a difference that is undetectable with conventional, protein-based cytokine assays. We also found that active TB was associated with higher ratios of effector memory to central memory Th1 cells than LTBI. This effector memory phenotype of active TB was associated with increased T cell differentiation, as defined by loss of the CD27 marker, but not with T cell exhaustion, as determined by PD-1 abundance. These results indicate that single-cell-based, mRNA measurements may help identify time-dependent, quantitative differences in T cell functional status between latent infection and active tuberculosis.

Keywords: FISH-Flow; T cell activation; cytokine; flow cytometry; immunophenotyping; memory T cells; single-cell gene expression.

Figure 1

Analysis of cytokine expression and cell activation in CD4+ T cells from LTBI and active TB donors. (A) Representative dot plots of IFNG and TNFA cytokine expression in CD4+ T cells after 2 and 6 h in vitro stimulation. PBMC were stimulated for 2 and 6 h with PPD or left unstimulated. After stimulation samples were stained for surface markers and hybridized with IFNG or TNFA mRNA probes. Frequencies of CD4+ cytokine mRNA+ cells are reported above each gate. (B) Cytokine expression in CD4+ cells. Data are expressed as the frequencies of CD4+ T cells expressing IFNG or TNFA mRNA at each time point. LTBI n = 47, TB n = 34. (C) CD69 expression in unstimulated and PPD stimulated CD4+ T cells for 2 h. Data were expressed as frequency of CD4+ T cells. LTBI n = 24, TB n = 22. Gating strategy is described in Supplementary Figure 1C (upper panel). Outliers above 20% (one per group) were removed to allow for better data representation. (D) CD69 expression in IFNG and TNFA mRNA+ CD4+ T cells. Data were expressed as frequency of total IFNG+ or TNFA+ cells after 2 h of PPD stimulation. LTBI n = 24, TB n = 22. Gating strategy is described in Supplementary Figure 1C (lower panel). The Wilcoxon two-sample test was used to analyze differences between LTBI and active TB donors. The box plots show lower quartile, median, and upper quartile of the distribution. The upper and lower whiskers represent the 90th and 10th percentile. Extreme values (below the 10th percentile and above the 90th percentile) are shown as (•) symbol.

Figure 2

Analysis of memory CD4+ T cell subsets in LTBI and active TB donors. (A) Frequencies of CD4+ T cell subsets in LTBI and active TB patients. Subsets were defined according to surface expression of CD45RA and CCR7 markers. Data were expressed as the mean of the frequency of CD4+ T cells in unstimulated and stimulated samples at 2 and 6 h. LTBI n = 39, TB n = 31. (B) Cytokine expression in effector and central memory T cells from LTBI and active TB donors. Data were expressed as frequency of the total memory T cells expressing IFNG or TNFA after 2 h of PPD stimulation. LTBI n = 37, TB n = 29. (C) EM/CM ratio. Data are expressed as the ratio of the frequencies of EM and CM T cells expressing IFNG or TNFA mRNA from LTBI and active TB patients after 2 h of PPD stimulation. LTBI n = 37, TB n = 29. Similar results were observed at the 6-h time point (data not shown). Gating strategy utilized for memory markers analysis is described in Supplementary Figures 5A,B. The Wilcoxon two-sample test was used to analyse differences between LTBI and active TB donors. The box plots show lower quartile, median, and upper quartile of the distribution. The upper and lower whiskers represent the 90th and 10th percentile. Extreme values (below the 10th percentile and above the 90th percentile) are shown as (•) symbol.

Figure 3

Analysis of CD27 and PD-1 expression in CD4 EM memory T cells. (A) CD27 expression in EM CD4 T cells. Data from each patient were plotted as the mean of the frequency of total EM CD4 T cells in unstimulated and stimulated samples at 2 and 6 h. LTBI n = 16, TB n = 15. (B) Frequency of CD27- memory T cells in IFNG and TNFA expressing cells. Data were expressed as frequency of the total IFNG or TNFA expressing cells in EM CD4 T cells from donors after 2 and 6 h of PPD stimulation. LTBI n = 16, TB n = 15. (C) PD-1 expression in EM CD4 T cells. Data were expressed as the mean of the frequency of total EM T cells in unstimulated and stimulated samples at 2 and 6 h. LTBI n = 10, TB n = 11. (D) PD-1 expression in EM CD4 T cells expressing IFNG and TNFA mRNA. Data were expressed as the frequency of the total IFNG or TNFA expressing cells in EM subsets at 6 h. Similar results were observed at the 2-h time point (data not shown). LTBI n = 10, TB n = 11. Gating strategy utilized for CD27 and PD-1 analysis is described in Supplementary Figures 5C,D. The Wilcoxon two-sample test was used to analyze differences between LTBI and active TB donors. The box plots show lower quartile, median, and upper quartile of the distribution. The upper and lower whiskers represent the 90th and 10th percentile. Extreme values (below the 10th percentile and above the 90th percentile) are shown as (•) symbol.

Figure 4

CXCR3, CCR6 T memory subsets in LTBI and active TB donors. (A) Frequencies of CXCR3 and CCR6 subsets in CM and EM CD4+ T cells. Data from each patient were plotted as the mean of the frequency of total CM or EM CD4+ cells in unstimulated and stimulated samples at 2 and 6 h. (B) Cytokine expression in CXCR3 and CCR6 subsets in CM and EM CD4+ T cells. Data are expressed as the frequency of total IFNG or TNFA expressing cells after 6 h of PPD stimulation. Similar results were observed at the 2-h time point (data not shown). LTBI n = 10, TB n = 10. Gating strategy is reported in Supplementary Figures 7A,B. The Wilcoxon two-sample test was used to analyze differences between LTBI and active TB donors. The box plots show lower quartile, median, and upper quartile of the distribution. The upper and lower whiskers represent the 90th and 10th percentile. Extreme values (below the 10th percentile and above the 90th percentile) are shown as (•) symbol.