Smad4 promotes differentiation of effector and circulating memory CD8 T cells but is dispensable for tissue-resident memory CD8 T cells

Abstract

Tissue-resident memory CD8 T cells are a unique subset of virus-specific CTLs that bolster local immune responses after becoming lodged in previously infected tissues. These cells provide enhanced protection by intercepting returning pathogens before a new infection gets established. In contrast, central memory CD8 T cells circulate in the bloodstream and proliferate in secondary lymphoid organs before replenishing effector and memory CD8 T cell populations in remote parts of the body. Both populations of virus-specific memory CD8 T cells participate in immunity to influenza virus infection; however, the signaling pathways that instruct developing memory CD8 T cells to distribute to specific tissues are poorly defined. We show that TGF-β promotes the development of pulmonary tissue-resident memory T cells via a signaling pathway that does not require the downstream signaling intermediate Sma- and Mad-related protein (Smad)4. In contrast, circulating memory CD8 T cells have no requirement for TGF-β but show signs of arrested development in the absence of Smad4, including aberrant CD103 expression. These signaling pathways alter the distribution of virus-specific CTLs in the lungs but do not prevent robust cytokine responses. Our data show that Smad4 is required for normal differentiation of multiple subsets of virus-specific CD8 T cells. In normal circumstances, Smad4 may be activated via a pathway that bypasses the TGF-β receptor. Improved understanding of these signaling pathways could be used to augment vaccine-induced immunity.

FIGURE 1.

CTLs have reciprocal phenotypes in the absence of TGF-βRII and Smad4. Two groups of chimeric mice were made with mixed bone marrow cells from either TKO mice and WT littermates or SKO mice and WT littermates. Plots are representative of three independent experiments (three to four animals per group). (A) CD44 and CD62L expression on CD8 T cells in peripheral blood. (B) KLRG1 expression on CD8 T cells in peripheral blood, with percentages of cells in the marked region. TKO or SKO (dashed line) and WT (gray fill) are shown. (C) Chimeric mice were infected with X31-OVA. Total numbers of NP-specific CTLs are shown. TKO (hatched), SKO (black), and littermates (white) are shown. Error bars are means ± SD for three animals per group. *p ≤ 0.05, ***p ≤ 0.005. (D) Gated populations of NP-specific CTLs were analyzed for KLRG1 and CD103 expression.

FIGURE 2.

Pulmonary T_RM cells do not require Smad4. Congenically marked OTI-WT (top), OTI-TKO (middle), or OTI-SKO cells (bottom) were sorted for low CD44 expression and transferred to C57BL/6 mice before X31-OVA infection. Error bars are means ± SD from three to four animals per group. Three independent experiments gave consistent results. (A) Numbers of KLRG1⁺ CTLs (gray shading). (B) Numbers of CD69⁺ (hatched), CD69⁺/CD103⁺ (black shading, T_RM), CD103⁺ (gray shading), or no markers (white). Statistical comparisons are shown in Supplemental Table I.

FIGURE 3.

TGF-β prevents prolonged cell proliferation. OTI-WT, OTI-TKO, and OTI-SKO cells were sorted for low CD44 expression and transferred to C57BL/6 mice 48 h before infection with X31-OVA. (A) OTI-WT, OTI-TKO, and OTI-SKO cells were analyzed for CFSE dilution 3.5 dpi (three animals per group). OTI-WT (gray fill), TKO or SKO (dashed line), and uninfected (solid line). Two experiments gave similar results. BrdU was injected on the days shown, and CD45.I⁺ CTLs were analyzed between 3 and 6 h later. The gating strategy is shown in Supplemental Fig. 1. (B) Percentages of BrdU⁺ cells within gated populations of OTI-WT (white bars), OTI-TKO (hatched bars), and OTI-SKO (black bars). Means ± SD from three animals per group. Two experiments gave similar results. *p ≤ 0.05, ***p ≤ 0.005. (C) Overlaid histograms show gated populations of CD45.1⁺ CTLs analyzed for activated caspase-3/7. OTI-WT (gray fill), TKO or SKO (dashed line), and unstained (solid line). Pooled data are from three mice; two experiments gave similar results. (D) Gated populations of transferred cells were analyzed for Bcl2 expression. OTI-WT (gray fill), TKO or SKO (dashed line), and isotype control (solid line). Pooled data are from three mice; two experiments gave similar results.

FIGURE 4.

KLRG1⁺ and CD103⁺ CTLs maintain different distributions inside the lungs. Congenically marked OTI-WT (top), OTI-TKO (middle), or OTI-SKO cells (bottom) were sorted for low CD44 expression and transferred to C57BL/6 mice before X31-OVA infection. Fragments of lung tissue were stained with Abs to EpCAM for epithelial cells (red), CD31 for blood vessels (yellow), CD45.1 for transferred cells (blue), KLRG1 (green), and CD103 (magenta). Z-stack images were recorded at original magnification ×20 (scale bars, 80 μm). Subregions show (I) colocalization between CD45.1 and KLRG1 10 dpi and (II) colocalization between CD45.1 and CD103 10 and 30 dpi. Representative data are from three animals per group; two experiments gave similar results.

FIGURE 5.

KLRG1⁺ memory CD8 T cells persist in the vasculature. CD45.1⁺ donor cells from OTI-TKO, OTI-SKO, and their respective littermates were transferred 48 h before infection with X31-OVA. The mice were injected with Abs to CD8β 5 min before harvest. (A) Gated populations of CD45.1⁺ and KLRG1⁺ CTLs in the lungs were analyzed for CD8β⁺ (blood) and CD8β⁻ (tissue) subsets. The CTLs were analyzed 8 dpi (top) and 40 dpi (bottom); n = 3–4 animals/group. Two experiments gave similar results. (B) Lungs were harvested 30 dpi and gated populations of OTI-TKO and OTI-WT cells were analyzed for CD11a and KLRG1 expression; n = 3 animals/group. Two experiments gave similar results. (C–E) Enriched lymphocytes were recovered 40 dpi. Gated CD45.1⁺ CTLs were divided into CD8β⁺ and CD8β⁻ subsets. Representative data are from three to four animals per group. Two experiments gave similar results. CTLs in the (C) lungs and (D) spleens were analyzed for KLRG1 and CD103 expression. (E) CTLs in the lungs were analyzed for CD69 and CD103 expression.

FIGURE 6.

Reduced numbers of T_CM cells accumulate in the resting lymph nodes when Smad4 is not expressed. Congenically marked OTI-WT, OTI-TKO, and OTI-SKO cells were sorted for low CD44 expression and transferred to C57BL/6 mice before X31-OVA infection. (A) CD62L and CD103 expression on gated OTI-WT, OTI-TKO, and OTI-SKO cells 30 dpi. Dot plots show representative data from three to four animals per group. Three independent experiments gave similar results. (B) Numbers of CD62L⁺ T_CM (black), CD62L⁻ (white), CD62L⁺CD103⁺ (hatched), and CD62L⁻CD103⁺ transition (gray fill) cells 30 dpi. Statistical comparisons are shown in Supplemental Table II. (C) Inguinal lymph nodes were imaged 20 dpi. CD31 (yellow), CD45.1 (blue), B220 (red), CD103 (magenta). Subregions (white boxes) show enlarged images of transferred CTLs with total numbers of cells in a single slice (means ± SD; n = 3). Z-stack images were recorded at original magnification ×10. Representative data from three animals per group. Two experiments gave similar results. (D) Overlaid histograms show CCR7 expression on OTI-WT (gray fill), OTI-SKO or OTI-TKO (dashed line), and isotype control (solid line). Representative data are from three animals per group. Two experiments gave similar results. (E) Mean fluorescence intensity for CCR7 staining. OTI-TKO (hatched), SKO (black), and littermates (white). *p ≤ 0.05.

FIGURE 7.

Mice recover from IAV infection with slightly delayed kinetics when Smad4 is not expressed in peripheral T cells. Animals were infected with X31-OVA and weighed daily. Combined data from two independent experiments are shown. (A) Reduction in body weight as percentage of maximum. TGF-βRIIKO (○, n = 7) TGF-βRII^flox/flox (●, n = 6), Smad4KO (□, n = 9), Smad4^flox/flox (▪, n = 6). *p ≤ 0.05, **p ≤ 0.01. (B) Model illustrating the roles of Smad4 and TGFβ during the differentiation of virus-specific CTLs. Left panel, Normal CD8 T cells give rise to mixed populations KLRG1⁺ T_EFF cells, CD103⁺ T_RM cells, and circulating memory CD8 T cells. Middle panel, TGF-β signaling reduces the numbers of terminally differentiated T_EFF cells and is essential for T_RM development. Right panel, Smad4 dependent signaling is required for terminal differentiation of virus-specific T_EFF cells and T_CM cells.