Tim-3-expressing CD4+ and CD8+ T cells in human tuberculosis (TB) exhibit polarized effector memory phenotypes and stronger anti-TB effector functions

Abstract

T-cell immune responses modulated by T-cell immunoglobulin and mucin domain-containing molecule 3 (Tim-3) during Mycobacterium tuberculosis (Mtb) infection in humans remain poorly understood. Here, we found that active TB patients exhibited increases in numbers of Tim-3-expressing CD4(+) and CD8(+) T cells, which preferentially displayed polarized effector memory phenotypes. Consistent with effector phenotypes, Tim-3(+)CD4(+) and Tim-3(+)CD8(+) T-cell subsets showed greater effector functions for producing Th1/Th22 cytokines and CTL effector molecules than Tim-3(-) counterparts, and Tim-3-expressing T cells more apparently limited intracellular Mtb replication in macrophages. The increased effector functions for Tim-3-expressing T cells consisted with cellular activation signaling as Tim-3(+)CD4(+) and Tim-3(+)CD8(+) T-cell subsets expressed much higher levels of phosphorylated signaling molecules p38, stat3, stat5, and Erk1/2 than Tim-3- controls. Mechanistic experiments showed that siRNA silencing of Tim-3 or soluble Tim-3 treatment interfering with membrane Tim-3-ligand interaction reduced de novo production of IFN-γ and TNF-α by Tim-3-expressing T cells. Furthermore, stimulation of Tim-3 signaling pathways by antibody cross-linking of membrane Tim-3 augmented effector function of IFN-γ production by CD4(+) and CD8(+) T cells, suggesting that Tim-3 signaling helped to drive stronger effector functions in active TB patients. This study therefore uncovered a previously unknown mechanism for T-cell immune responses regulated by Tim-3, and findings may have implications for potential immune intervention in TB.

Figure 1. TB infection led to up-regulation of Tim-3 expression and increases in numbers of Tim-3-expressing CD4⁺ and CD8⁺ T cells.

PBMC from 30 individuals with untreated active TB disease or 30 individuals with LTBI were stained directly or stimulated ex vivo first with pooled 15aa peptides overlapped by 12 spanning entire Mtb Ag85-b and ESAT-6 (termed Mtb peptide in figures), and then analyzed by polychromatic flow cytometry. (A) Representative flow cytometric dot plots showing Tim-3 expression in a healthy control (HC), a representative individual with untreated active TB disease, or a typical individual with LTBI. No Tim-3 expression was observed when we used isotype matched IgG to stain PBMC (see a representative panel on the right of Figure 1A). Values in the upper right quadrant indicate the percentages of Tim-3-expressing CD4⁺ and CD8⁺ T cells. Data were gated on CD3⁺CD4⁺ and CD3⁺CD8⁺, respectively. (B) and (C) are bar graph data showing that the percentages (%) of Tim-3 expression on CD4⁺ and CD8⁺ T cells from 30 patients with active TB disease are higher than those from 30 individuals with LTBI and 9 healthy control (HCs). Horizontal bars depict the mean percentage of Mtb-specific Tim-3 expression on CD4⁺ and CD8⁺ T cells. * p<0.05, ** p<0.01, *** p<0.001.

Figure 2. Tim-3^High CD4⁺ and CD8⁺ T-cell subsets showed greater effector functions for producing Mtb-specific IFN-γ and other cytokines (see Supporting Information, Figure S5, Figure S6) than their Tim-3^Low counterparts.

PBMC derived from individuals with active TB disease (n = 9) or with LTBI (n = 9) were cultured ex vivo in presence or absence of pooled Ag85-b/ESAT-6 peptides, stained for fluorochrome-conjugated mAbs and analyzed by polychromatic flow cytometry. To examine the association between the IFN-γ response and Tim-3 expression, we used a two-tiered gating system for analyzing IFN-γ response by Tim-3-expressing T cells (Tim-3^High) and Tim-3-negative cells (Tim-3^Low) subpopulations. (A) shows representative flow cytometric dot plots derived from an untreated active TB patient, indicating percentages of IFN-γ-producing cells in Tim-3^HighCD4⁺ (or CD8⁺) T-cell subset versus Tim-3^LowCD4⁺ (or CD8⁺) T-cell subset under the conditions with or without ex vivo stimulation with pooled Ag85B/ESAT-6 peptide. No significant intracellular staining of IFN-γ and other cytokines was seen when using isotype Ig control (see a representative panel in Figure 2A). (D) Similar flow cytometric dot plots show numbers of IFN-γ-producing cells in Tim-3^HighCD4⁺ (or CD8⁺) T cells versus Tim-3^LowCD4⁺ (or CD8⁺) T cells in an individual with LTBI. (B) shows bar graph data from individuals with active TB disease (n = 9) and demonstrates that percentages of IFN-γ⁺ T cells within Tim-3^HighCD4⁺ T-cell subsets are much greater than those within Tim-3^LowCD4⁺T cell subsets. (C) is similar to (B), except that data are IFN-γ⁺ T cells within Tim-3^HighCD8⁺ T-cell subsets. (E) are bar graph data from individuals with LTBI (n = 9) showing that percentages of IFN-γ⁺ T cells within Tim-3^HighCD4⁺ T-cell subsets are much greater than those within Tim-3^LowCD4⁺T cell subset. (F) is similar to (E), except that data are IFN-γ⁺ T cells within Tim-3^High (or Tim-3^Low)CD8⁺ T-cell subsets. Shown are data from at least three independent experiments. ** p<0.01, *** p<0.001. Error bars represent SD. Note that frequencies of cytokine-producing cells are expressed within Tim-3⁺ or Tim-3⁻IFN-γ⁺CD4⁺ or CD8⁺ T subsets, not total CD4⁺ or CD8⁺ T-cell population. Numbers of IFN-γ-producing cells within peptide-stimulated and unstimulated Tim-3⁺CD4⁺(or CD8⁺) T-cell subsets in active TB patients are significantly greater than those in healthy subjects with LTBI(p<0.001).

Figure 3. Tim-3^High CD4⁺ and CD8⁺ T-cell subsets also had greater effector functions of cytotoxic molecule production than their Tim-3^Low counterparts.

PBMC from individuals with active TB disease (n = 9) or with LTBI (n = 9) were cultured in presence or absence of pooled Ag85-b/ESAT-6 peptides, stained with fluorochrome-conjugated mAbs, and analyzed by polychromatic flow cytometry. (A) shows representative flow cytometric dot plots derived from an untreated active TB patient, indicating percentages of perforin-producing cells in Tim-3^HighCD4⁺ (or CD8⁺) T-cell subset versus Tim-3^LowCD4⁺ (or CD8⁺) T-cell subset under the conditions with or without ex vivo stimulation with pooled Ag85B/ESAT-6 peptide. (D) Similar flow cytometric dot plots show numbers of perforin-producing cells in Tim-3^HighCD4⁺ (or CD8⁺) T cells versus Tim-3^LowCD4⁺ (or CD8⁺) T cells in an individual with LTBI. (B) shows bar graph data from individuals with active TB disease (n = 9) and demonstrates that percentages of perforin⁺ T cells within Tim-3^HighCD4⁺ T-cell subsets are much greater than those within Tim-3^LowCD4⁺T cell subsets. (C) is similar to (B), except that data are perforin⁺ T cells within Tim-3^High (or Tim-3^Low)CD8⁺ T-cell subsets. (E) are bar graph data from individuals with LTBI (n = 9) showing that percentages of perforin⁺ T cells within Tim-3^HighCD4⁺ T-cell subsets are much greater than those within Tim-3^LowCD4⁺T cell subsets. (F) is similar to (E), except that data are perforin⁺ T cells within Tim-3^High(or Tim-3^Low)CD8⁺ T-cell subsets. Shown are data from at least three independent experiments. ** p<0.01, *** p<0.001. Error bars represent SD. Numbers of perforin-producing cells within peptide-stimulated and unstimulated Tim-3⁺CD4⁺(or CD8⁺) T-cell subsets in active TB patients are significantly greater than those in healthy subjects with LTBI (p<0.001).

Figure 4. Tim-3 knockdown by siRNA led to decreases in de novo production of IFN-γ and TNF-α by CD4⁺ and CD8⁺ T cells.

PBMC isolated from untreated active TB patients (n = 9) were transiently transfected with Tim-3-targeting (si-Tim-3) or control siRNA (si-control). Cells were then stimulated with or without pooled Ag85-b/ESAT-6 peptides for 6 days, stained by ICS with fluorochrome-conjugated mAbs and evaluated for the effects of Tim-3 silencing on T-cell effector functions using polychromatic flow cytometry. (A) Real-time PCR analysis of the expression of Tim-3 in PBMC from Mtb-infected subjects (n = 9) at 48 hrs after transient transfection of Tim-3-targeted siRNA(si-Tim-3) and non-targeted siRNA(si-control), with data presented as values relative to expression in PBMC treated with Lipofectamine medium. (B) Typical flow cytometric dot plots and (C) summary bar graphic data show that Tim-3-specific siRNA (si-Tim-3), but not non-targeted siRNA(si-control) or Lipofectamine medium, specifically reduces the de novo or Mtb-driven expression of Tim-3 in CD4⁺ and CD8⁺ T cells. Values in upper right quadrant of flow cytometric dot plots indicate the percentages of Tim-3^High cells. (B) Flow cytometric dot plots and (D, E) summary bar graphic data show that Tim-3-targeted siRNA (si-Tim-3), but not non-targeted siRNA(si-control) or Lipofectamine medium, significantly reduces the percentages of IFN-γ- and TNF-α-expressing T cells in Tim-3^HighCD3⁺CD4⁺ T cells. Dot plots in (B) were gated on Tim-3^HighCD3⁺CD4⁺ T cells. (F), (G), (H), and (I) show that similar results in Tim-3^HighCD3⁺CD8⁺T cells. Dot plots in (F) were gated on Tim-3^HighCD3⁺CD8⁺ T cells. Values in each of flow cytometric dot plots are percentages of IFN-γ- or TNF-α-expressing T cells in Tim-3^HighCD3⁺CD4(or CD8)⁺ T cells. Data are from at least three independent experiments. Error bars represent SD. *** p<0.001, NS, no statistical significance.

Figure 5. Adding a low concentration of soluble Tim-3 (s-Tim-3) to PBMC cultures reduced effector functions of IFN-γ and TNF-α production by Tim-3-expressing CD4⁺ and CD8⁺ T cells.

PBMC derived from untreated active TB patients (n = 9) were incubated with soluble form of Tim-3 (s-Tim-3) in a concentration of 2 µg/ml in presence or absence of pooled Mtb Ag85-b/ESAT-6 peptides for 6 days. Cells were then stained using the ICS protocol, and analyzed by flow cytometry. (A) and (E) show representative flow cytometric data demonstrating the effects of Tim-3 ligand competition on Mtb-driven IFN-γ and TNF-α responses of Tim-3^HighCD4⁺ T cells. (C) and (G) show bar graphic data demonstrating that ligand competition by adding 2 µg/ml of soluble Tim-3 (s-Tim-3) significantly inhibits Mtb-driven IFN-γ and TNF-α production by CD4⁺ T cells. (B), (D), (F), and (H) show similar results were also observed in Tim-3^HighCD8⁺ T cells. Data are from at least two independent experiments. Error bars represent SD. *p<0.05, ** p<0.01 *** p<0.001.

Figure 6. Purified Tim-3⁺CD3⁺T cells more effectively restricted Mtb replication in MΦs than Tim-3⁻CD3⁺T cells.

PBMC were derived from untreated active TB patients (n = 9). Monocytes-derived MΦs were infected with Mtb overnight, and extracellular Mtb were then removed by extensive PBS wash. Autologous Tim-3⁺CD3⁺T cells and Tim-3⁻CD3⁺T cells were purified by magnetic beads and co-cultured with Mtb-infected MΦs for 4 days in presence or absence of anti-IL-1β Ab and isotype control IgG. Cultured cells containing Mtb-infected cells were then lysed, and Mtb CFUs were examined (see details in Materials and Methods). CFU numbers were expressed as per 10⁶ MΦs. Data are from at least three independent experiments. Error bars represent SD. *** p<0.001,** p<0.01, NS, no statistical significance.

Figure 7. Tim-3-expressing CD4⁺ and CD8⁺ T cells from active TB patients exhibited stronger p38, stat3, stat5, and Erk1/2 signaling in cultures with or without Mtb peptide stimulation.

Phosphorylation status of p38, stat5, stat3, and Erk1/2 were analyzed by polychromatic flow cytometry in Tim-3^High versus Tim-3^Low T-cell subsets. PBMC from untreated active TB patients (n = 9) were stained directly without peptide stimulation or stained after 30 min ex vivo stimulation with pooled Ag85-b/ESAT-6 peptides, and assessed for expression of phosphorylated(P)-p38, P-stat5, P-stat3, or P-Erk1/2 cells in Tim-3^High versus Tim-3^Low T-cell subsets using flow cytometry analysis as described above. (A) and (D) shows a representative flow cytometric gating strategy for Tim-3^High and Tim-3^LowCD4⁺ and CD8+ T cells, respectively, to evaluate phosphorylated signaling molecules in cells cultured with or without ex vivo Mtb peptide stimulation. (B) shows representative flow cytometric dot plots demonstrating that expression levels of P-p38, P-stat5, P-stat3, and P-Erk1/2 were much higher in Tim-3^HighCD4⁺ T cells than Tim-3^LowCD4⁺ T cells in the absence (upper panels) and presence (lower panels) of Mtb peptide stimulation. Values in dot plots indicate the percentages of expression levels of each of phosphorylated signaling molecule. Similar results were seen when Tim-3⁺ T cells were enriched by immunonmagnetic beads, and then cultured with or without Mtb peptide prior to flow analyses (data not shown). (E) is similar to (B) except for comparisons between Tim-3^HighCD8+ T cells Tim-3^LowCD8⁺ T cells. (C) and (F) are bar graph data showing higher expression levels of P-p38, P-stat5, P-stat3, or P-Erk1/2 cells in Tim-3^HighCD4⁺ or Tim-3^HighCD8⁺ T cells than Tim-3^LowCD4⁺ or Tim-3^LowCD8⁺T cell subpopulations in the absence (upper panels) and presence (lower panels) of Mtb peptide stimulation (Lower panel). (C) and (F) show that percentages of expression of P-p38, P-stat5 and P-Erk1/2 are much higher than P-stat3 in Tim-3^HighCD4⁺ or Tim-3^HighCD8⁺ T cells. Data are from at least two independent experiments. *** p<0.001. Error bars represent SD.