FOXP3+ Regulatory T Cells Require TBET to Regulate Activated CD8+ T Cells During Recovery from Influenza Infection

Abstract

FOXP3+ regulatory T (Treg) cells are necessary to coordinate resolution of lung inflammation and a return to homeostasis after respiratory viral infections, but the specific molecular requirements for these functions and the cell types governed by Treg cells remain unclear. This question holds significance as clinical trials of Treg cell transfer therapy for respiratory viral infection are being planned and executed. Here, we report causal experiments in mice determining that Treg cells are necessary to control the numbers of activated CD8+ T cells during recovery from influenza infection. Using a genetic strategy paired with adoptive transfer techniques, we determined that Treg cells require the transcription factor TBET to regulate these potentially pro-inflammatory CD8+ T cells. Surprisingly, we found that Treg cells are dispensable for the generation of CD8+ lung tissue resident-memory T (Trm) cells yet similarly influence the transcriptional programming of CD8+ Trm and activated T cells. Our study highlights the role of Treg cells in regulating the CD8+ T cell response during recovery from influenza infection.

Figure 1.. Treg cells regulate activated CD8+ cells in a TBET-dependent manner during recovery from influenza.

(A) Experimental outline. Foxp3^DTR mice were intra-tracheally inoculated with 6 plaque-forming units of influenza A/WSN/33 H1N1 virus. At day 14 post-inoculation, a group of inoculated Foxp3^DTR mice received a diphtheria toxin (DT) loading dose (50 μg/kg) and three DT maintenance doses (10 μg/kg) every other day (q.o.d.). Lung CD8+ T cell subsets were quantified in DT-untreated (control) and - treated Foxp3^DTR mice at days 22, 28, and 45 post-inoculation. (B) CD8+ CD69+ CD103+ Trm cell and (C) CD8+ CD69+ CD103- (activated) T cell absolute counts per pair of lungs at days 22, 28, and 45 post-inoculation in control and DT-treated Foxp3^DTR mice. nd no discovery, * q < 0.05, according to Mann-Whitney U test with two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli with Q = 5%. Summary plots show all data points with mean and SD from two independent experiments per time point. (D) Principal component analysis of RNA sequencing data generated with sorted double-negative (CD69- CD103-), activated (CD69+ CD103-), and Trm (CD69+ CD103+) CD8+ T cells from DT-untreated and -treated Foxp3^DTR mice at day 28 post-influenza virus inoculation. (E) K-means clustering of differentially expressed genes (q < 0.05) and selection of top GO processes from the comparison of activated CD8+ T cells sorted from DT-untreated and -treated Foxp3^DTR mice at day 28 post-influenza virus inoculation. (F) Experimental outline. Foxp3^DTR mice received either TBET-sufficient (WT) or -deficient (KO) Treg cells via retro-orbital adoptive transfer alongside an intra-peritoneal diphtheria toxin (DT) loading dose (50 μg/kg). These mice received a maintenance dose of DT (10 μg/kg) every other day (q.o.d.) thereafter. One day after adoptive transfer, Foxp3^DTR mice were intra-tracheally inoculated with 6 plaque-forming units of influenza A/WSN/33 H1N1 virus. Lung T cells were quantified on day 28 post-inoculation. (G) Absolute counts of total CD8+ cells and (H) CD8+ CD69+ CD103- T cells per pair of lungs at day 28 post-influenza virus inoculation in DT-treated Foxp3^DTR mice that received either TBET WT or KO Treg cells. Data points were normalized to the TBET WT Treg adoptive transfer group’s average. ns not significant, ** p < 0.01 according to Mann-Whitney U test.

Figure 2.. Characterization of T cell subsets during recovery from influenza in the absence of Treg cells.

(A) Treg, (B) CD8+ CD69+ CD103+ Trm, and (C) CD8+ CD69+ CD103- (activated) T cell frequencies and (D) absolute count (left) and frequency (right) of double-negative CD8+ T cells in the lungs from the experiment outlined in Figure 1A. nd no discovery, * q < 0.05 according to Mann-Whitney U test with two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli with Q = 5%. Summary plots show all data points with mean and SD. (E) Representative heatmap statistic plot showing IV CD45-BUV563 signal intensity of lung CD8+ T cell subsets in control and DT-treated Foxp3^DTR mice that received 50 µg of CD45-BUV563 antibody retro-orbitally before harvest. (F) K-means clustering and selection of top GO processes from the comparison of CD8+ CD69+ CD103+ Trm T cells sorted from DT-untreated and -treated Foxp3^DTR mice at day 28 post-influenza virus inoculation.

Figure 3.. Necessity of Treg cell TBET expression on Treg, Trm, and other lung T cell numbers during influenza recovery.

(A) Flow cytometry gating strategy for the quantification of T cell subsets, including CD8+ Trm cells (CD69+ CD103+), activated CD8+ T cells (CD69+ CD103-), double-negative CD8+ T cells (CD69- CD103-), endogenous Treg cells (CD4+ FR4+ CD25+ Foxp3-GFP+), and adoptive transfer Treg cells (CD4+ FR4+ CD25+ Foxp3-GFP-) in the experiment outlined in Figure 1F. (B) Frequency of Foxp3-GFP+ cells (endogenous Treg cells) of total CD4+ cells in DT-untreated Foxp3^DTR mice as well as DT-treated Foxp3^DTR mice that received either WT or TBET-deficient (KO) Treg cells. (C) Representative contour plots of total Treg frequency based on FR4 and CD25 staining in mice that received WT or TBET-deficient (KO) Treg cells. (D) Representative histograms of CXCR3+ Treg cell frequency of total Treg cells from DT-treated Foxp3^DTR mice that received either WT or TBET-deficient (KO) Treg cells. (E) Absolute counts of CD8+ CD69- CD103- (activated) T cells and (F) CD8+ CD69+ CD103+ Trm cells per pair of lungs at day 28 post-influenza virus inoculation from the experiment outlined in Figure 1F. Datapoints were normalized to the TBET WT Treg adoptive transfer group average. ns not significant, ** p < 0.01 according to Kruskal-Wallis test (B) or Mann-Whitney U test (E and F).