Interferon-γ couples CD8+ T cell avidity and differentiation during infection

Abstract

Effective responses to intracellular pathogens are characterized by T cell clones with a broad affinity range for their cognate peptide and diverse functional phenotypes. How T cell clones are selected throughout the response to retain a breadth of avidities remains unclear. Here, we demonstrate that direct sensing of the cytokine IFN-γ by CD8⁺ T cells coordinates avidity and differentiation during infection. IFN-γ promotes the expansion of low-avidity T cells, allowing them to overcome the selective advantage of high-avidity T cells, whilst reinforcing high-avidity T cell entry into the memory pool, thus reducing the average avidity of the primary response and increasing that of the memory response. IFN-γ in this context is mainly provided by virtual memory T cells, an antigen-inexperienced subset with memory features. Overall, we propose that IFN-γ and virtual memory T cells fulfil a critical immunoregulatory role by enabling the coordination of T cell avidity and fate.

Fig. 1. IFN-γR deletion in CD8⁺ T-cells improves effector responses against Influenza infection.

a, b CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with 4 × 10⁴ pfu X31-OVA. a Weight was measured every two days and quantified as relative weight loss (n = 12 Ctrl, 14 KO animals). b Lung viral titres were assessed by qRT-PCR on day 3/4 (n = 11 Ctrl, 12 KO animals), 6 (n = 11 animals) and 8 (n = 6 Ctrl, 5 KO animals). Expression levels were normalised to the housekeeping gene β-actin. c, d CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with 4 × 10⁵ pfu X31-OVA. Graphs show average weight (n = 9 Ctrl, 8 KO animals) (c) and survival (n = 9 animals) (d) over time. e CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA and injected with ad-mixed N4-loaded or unloaded target splenocytes 7 day post-infection to quantify in vivo cytotoxicity. The graph shows in vivo cytotoxicity normalized by the percentage of endogenous OVA-specific T-cells (n = 8 animals). f–h CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA and splenocytes were isolated after 7–9 days. f Splenocytes were re-stimulated with N4 peptide and LAMP-1 expression in CD8⁺ T-cells was analyzed by flow cytometry and normalized by the percentage of endogenous OVA-specific T-cells (n = 13 animals). Perforin (g) and Granzyme B (h) expression was analyzed by flow cytometry (n = 7 Ctrl, 8 KO animals). i–k CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with 4 × 10⁴ pfu X31-OVA. i Relative abundance of lung N4-tetramer⁺ CD8⁺ T-cells was analysed by flow cytometry on day 4, 6, 8 (n = 6 animals) or 9 (n = 14 Ctrl, 13 KO animals). Absolute numbers (n = 14 Ctrl, 13 KO animals) (j) and relative avidity (n = 12 Ctrl, 11 KO animals) (k) of N4-tetramer⁺ CD8⁺ T-cells were analyzed by flow cytometry between day 7 and 10 post-infection. Data are from ≥3 (a–f, i–k or 2 g, h) independent experiments. Two-tailed unpaired Student’s t-test (E-H, J-K), Two-way ANOVA with Šidák’s multiple comparison test (a–c) Mantel-Cox test (d). Error bars indicate the mean ± s.e.m.

Fig. 2. IFN-γR deletion in CD8⁺ T-cells increases the avidity of effector T-cells.

a–h CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA and splenocytes were isolated after 7–10 days. Relative abundance (n = 25 Ctrl, 28 KO animals) (a), representative N4-tetramer histogram (b) and normalized N4-tetramer MFI (n = 37 animals) (c) of N4-tetramer⁺ CD8⁺ T-cells, analyzed by flow cytometry. N4-tetramer gMFIs of KO samples were normalized to the mean Ctrl N4-tetramer gMFI within each independent experiment. d–h Splenocytes were re-stimulated in vitro with the indicated concentrations of N4 for 16 h. d Percent KO and Ctrl CD8⁺ T-cells expressing IFN-γ at the maximum peptide concentration, assessed by flow cytometry and normalized by % of N4-tetramer⁺ CD8⁺ T cells (n = 24 Ctrl, 25 KO animals). Analysis of relative IFN-γ expression (e) and corresponding EC₅₀ (f) by flow cytometry (n = 15 Ctrl, 16 KO animals). (g, h) Analysis of relative IFN-γ expression (g) and corresponding EC₅₀ (h) by ELISA (n = 9 animals). i, j CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA and N4-tetramer⁺ CD8⁺ T-cells were sorted after 9 days to measure TCR avidity by acoustic force spectroscopy (n = 3 experiments, 4 mice per experiments). i Representative graph showing KO (red) and Ctrl (grey) CD8⁺ T-cell binding to target cells during acoustic force measurement. j Quantification of KO and Ctrl CD8⁺ T-cell binding at maximum force. Data are from 6 (a–d) or ≥ 3 (e–j) independent experiments. Two-tailed unpaired Student’s t-test (a–d, h, j) and two-tailed Mann-Whitney test (f). Error bars indicate the mean ± s.e.m.

Fig. 3. IFN-γ-is by CD8⁺ T-cells at priming and is paracrine.

a–c WT mice were infected with LM-OVA and mice were either left untreated (Ctrl, grey) or treated with anti-IFN-γ 16-24 h (red) or at day 5 and 6 (orange) post-infection. Graphs show relative abundance (n = 13 Ctrl and 24h-anti-IFN-γ-treated animals, 12 day 5/6-anti-IFN-γ-treated animals) (a), normalized N4-tetramer gMFIs (n = 13 Ctrl/24h-anti-IFN-γ-treated animals, 12 day5/6-anti-IFN-γ-treated animals) (b), and CD3 gMFI (n = 6 Ctrl/24h-anti-IFN-γ-treated animals, 7 day5/6-anti-IFN-γ-treated animals) (c) of N4-tetramer⁺ CD8⁺ T-cells analyzed by flow cytometry 9 days post-infection. d–g WT mice were treated with depleting NK1.1 (brown) or control antibodies (grey) and infected with LM-OVA. d, e Relative abundance (d) and the normalized N4-tetramer gMFIs (e) of N4-tetramer⁺ CD8⁺ T-cells 9 days post-infection (n = 10 animals). f–g Splenocytes were re-stimulated in vitro with the indicated concentrations of N4 for 16 h. Quantification of relative IFN-γ expression (n = 5 animals) (f) and EC₅₀ of IFN-γ production (n = 10 animals) (g) by flow cytometry. h–k Control (Ctrl, blue) and CD8-IFN-γ^KO (pink) chimera mice were infected with LM-OVA. h, i Splenocytes were isolated after 24 hrs. h IFN-γ production by CD8⁺ T-cells and NK cells was quantified after ex-vivo PMA and Ionomycin stimulation (n = 3 animals). i STAT1 phosphorylation (pSTAT1) was quantified in activated CD8⁺ T-cells (n = 12 animals). pSTAT1 of BM chimeras was normalized to IFN-γKO pSTAT1 MFI by subtracting the background gMFI. j, k Splenocytes were isolated after 9 days. Relative abundance (n = 11 Ctrl, 12 CD8-IFN-γ^KO animals) (j) and normalized N4-tetramer gMFIs (n = 11 animals) (k) of N4-tetramer⁺ CD8⁺ T-cells. N4-tetramer gMFIs of IFN-γ- or NK-depleted samples were normalized to the mean Ctrl N4-tetramer gMFI within each independent experiment. Data are from ≥3 (a–e, i–k), 2 (f, g) or representative of 3 (h) independent experiments. Two-tailed unpaired Student’s t-test (d, e, g, j, k), one-way ANOVA with Tukey’s multiple comparison test (a–c) and two-way ANOVA and Šidák’s multiple comparison test (h). Error bars indicate the mean ± s.e.m.

Fig. 4. T_VM are the main CD8⁺ T-cell source of IFN-γ-during priming.

a–f CD8⁺ T-cells from naïve mice (Ctrl) or LM-OVA-infected mice (LM-OVA) were sorted after 24 h and subjected to scRNA-seq analysis (n = 3224 Ctrl, 3495 LM-OVA cells from 3 animals). a Graph-based clustering of the assembled clusters. b Heatmap shows the average expression of selected markers. c Volcano plot shows differentially expressed genes between virtual memory (T_VM) and other CD8⁺ T-cells (268 variables). Green dots: genes with log2 (fold-change) value > 0.5 or <−05; blue dots: genes with an adjusted p value < 0.05; red dots: genes with log2 (fold-change) value > 0.5 or <−05 and an adjusted p value < 0.05. Two-tailed Wilcoxon rank sum test. d Violin plot shows the expression of IFN-γ in Ctrl and LM-OVA infected mice. e Graph shows the relative frequency of naïve, T_VM or true memory (T_M) CD8⁺ T-cells within the IFN-γ⁺ T-cell fraction. f Dot plot shows the average expression of IFN-γ in naïve, T_VM or T_M cells at steady-state and 24 h post LM-OVA infection. g GREAT mice were infected with LM-OVA or LM-gp33. Relative abundance of naïve (red), T_VM (green) or T_M (blue) T-cells among IFN-γ⁺ splenocytes was analyzed by flow cytometry 24 hrs post-infection (n = 11 LM-OVA-, 8 LM-gp33-infected animals). h, i CD8⁺ T-cells from GREAT mice were stimulated in vitro with anti-CD3/−28 (TCR), and the indicated cytokines. IFN-γ production was analyzed by flow cytometry after 24 h (n = 3–6 animals). j, k WT mice were transferred with 2×10⁶ GREAT OT-I CD8⁺ T-cells. j IFN-γ expression by OT-I was quantified by flow cytometry 24 h post-infection with LM-OVA (red) or LM-gp33 (orange) (n = 6 naïve, 8 LM-OVA- and LM-gp-33-infected animals). k The relative abundance of naïve (red), T_VM (green) or memory T_M (blue) among GREAT OT-I IFN-γ⁺ splenocytes was analyzed by flow cytometry 24 h post-infection (n = 4 animals). l, m Sorted T_VM and naïve OT-I CD8⁺ T-cells were transferred into WT hosts. l Mice were treated with BFA 6hrs before harvest. IFN-γ expression of naïve (red) and T_VM (green) OT-I T-cells among splenocytes 24h post-LM-OVA infection (n = 9 for naïve transfer, 10 animals for T_VM transfer). m Relative abundance of naïve, T_VM or T_M CD8⁺ T-cells among naïve (red) or T_VM (green) OT-I IFN-γ⁺ splenocytes (n = 10 animals), analysed by flow cytometry. n GREAT mice were infected with LM-OVA and treated with anti-MHC class-I, and anti-IL-12/−18 as indicated. IFN-γ production was quantified in CD8⁺ T-cells 24h post-infection (n = 4 animals). Data are from 1 (a–f) or ≥2 (g–n) independent experiments. One-way ANOVA with Tukey’s multiple comparison test (h–l, n) and Two-way ANOVA and Šidák’s multiple comparison test (g, n, m). Bars indicate the mean ± s.e.m.

Fig. 5. IFN-γR deletion in CD8⁺ T-cells does not regulate overall differentiation and TCR diversity.

a–g CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA. Splenocytes were isolated 9 days post-infection. a–c, f, g N4-tetramer⁺ CD8⁺ T-cells were subjected to scRNA- and scTCR-seq (n = 5646 Ctrl, 4837 KO cells from 4 animals). a Graph-based clustering of the assembled differentiation states. b Violin plot shows the expression of selected markers. c Plot shows the average gene expression of Ctrl and KO clusters. Relative abundance of Ctrl and KO SLEC (d) and MPEC (e) cells(n = 22 Ctrl, 25 KO animals). f Relative TCR clonal diversity (Shannon index) by cell state and genotype extracted from scTCR-seq analysis (n = 4 animals). g Clonal overlap between the cycling cluster and the other clusters from scTCR-seq analysis (n = 4 animals). h WT mice were transferred with 50.000 ad-mixed WT and IFN-γR^KO OT-I T-cells and infected with LM expressing the indicated OVA peptides. OT-I T-cells expansion was analyzed by flow cytometry after 9 days. Data are expressed as the ratio between IFN-γR^KO and WT OT-I cell numbers (n = 6 animals). i OT-I T-cells were activated in vitro with 50 ng/mL N4 in the presence of increasing doses of IFN-γ. After 36 h, cell apoptosis was assessed by Annexin V staining and quantified by flow cytometry (n = 12 animals). j WT mice were transferred with OT-I T-cells, infected with LM-OVA and treated with anti-IFN-γ after 24 h. OT-I T-cell apoptosis was analyzed by Annexin V staining 6 days post-infection (n = 9 ctrl, 11 anti-IFN-γ-treated animals). k WT mice were transferred with 50.000 ad-mixed WT and IFN-γR^KO OT-I T-cells and infected with LM expressing the indicated OVA peptides. EdU was injected on day 6 and proliferation of OT-I CD8⁺ T-cells was analyzed 16 h later by flow cytometry (n = 9 animals). Data are from 1 (a–c, f, g), ≥ 2 (h, k) or 6 (d, e) independent experiments. Comparison between groups was calculated using the two-tailed unpaired Student’s t-test (d, e, j), One-way ANOVA with Tukey’s multiple comparison test (f, g, k) and a Two-way ANOVA and Šidák’s multiple comparison test (h, i). Error bars indicate the mean ± s.e.m.

Fig. 6. IFN-γR deletion in CD8^± T-cells lowers the avidity of CD8^± T-cell memory responses.

a–d WT mice were transferred with 50.000 ad-mixed WT and IFN-γR^KO OT-I T-cells, infected with LM expressing the indicated OVA peptides. Relative abundance of OT-I SLEC (a) and MPEC (b) cells, analyzed 9 days post-infection by flow cytometry (n = 4 for T4, 6–9 animals for other peptides). c Relative abundance of transferred OT-I T-cells ≥60 days post-infection (n = 10 for N4, 7 animals for other peptides). d Mice were re-challenged after ≥60 days with LM-OVA and OT-I T-cell expansion was analyzed 5 days later by flow cytometry (n = 5 for Q4H7, 6 animals for other peptides). Data are expressed as the ratio between IFN-γR^KO and WT OT-I cell number. e–h CD8-IFN-γR^KO (KO, red) and control (Ctrl, grey) mice were infected with LM-OVA. e–g Splenocytes were isolated after 9 days. e Representative histograms of N4-tetramer staining within the SLEC and MPEC populations. Normalized N4-tetramer-staining of SLEC (f) and MPEC cells (g) (n = 20 animals). (h) Mice were re-challenged 60 days post-infection with LM-OVA and normalized N4-tetramer gMFI was analyzed by flow cytometry after 5 days (n = 11 animals). i–M KO (red) and Ctrl (grey) mice were infected with X31-OVA. Mice were re-challenged after ≥60 days with PR8-OVA. Tissues were harvested after 5-7 days and stained for CD8⁺ T-cells and the indicated tetramer. i Graph shows normalized N4-tetramer gMFI of N4-tetramer⁺ CD8⁺ T-cells in the lung (n = 7 animals). Graph shows the percentage (n = 9 Ctrl, 10 KO animals) (J) and normalized NP₆₈-tetramer gMFI (n = 6 animals) (K) of NP₆₈-tetramer⁺ CD8⁺ T-cells in draining LNs. Graph shows the percentage (n = 11 Ctrl, 13 KO animals) (I) and normalized NP₆₈-tetramer gMFI (n = 8 animals) (M) of NP₆₈-tetramer⁺ CD8⁺ T-cells in the lung. KO tetramer gMFIs samples were normalized to the mean Ctrl tetramer gMFI within each independent experiment. Data are from ≥2 (a–d), 3 (h–M) or 6 (e–g) independent experiments. Two-tailed unpaired Student’s t-test (f–M). One-way ANOVA with Tukey’s multiple comparison test (c, d) and two-way ANOVA and Šidák’s multiple comparison test (a, b). Error bars indicate the mean ± s.e.m.