A direct and nonredundant role for thymic stromal lymphopoietin on antiviral CD8 T cell responses in the respiratory mucosa

Abstract

Mucosally produced thymic stromal lymphopoietin (TSLP) regulates Th2 responses by signaling to dendritic cells and CD4 T cells. Activated CD8 T cells express the TSLP receptor (TSLPR), yet a direct role for TSLP in CD8 T cell immunity in the mucosa has not been described. Because TSLP shares signaling components with IL-7, a cytokine important for the development and survival of memory CD8 T cells in systemic infection models, we hypothesized that TSLP spatially and nonredundantly supports the development of these cells in the respiratory tract. In this study, we demonstrate that influenza infection induces the early expression of TSLP by lung epithelial cells with multiple consequences. The global loss of TSLP responsiveness in TSLPR(-/-) mice enhanced morbidity and delayed viral clearance. Using a competitive adoptive transfer system, we demonstrate that selective loss of TSLPR signaling on antiviral CD8 T cells decreases their accumulation specifically in the respiratory tract as early as day 8 after infection, primarily due to a proliferation deficiency. Importantly, the subsequent persistence of memory cells derived from this pool was also qualitatively and quantitatively affected. In this regard, the local support of antiviral CD8 T cells by TSLP is well suited to the mucosa, where responses must be tempered to prevent excessive inflammation. Taken together, these data suggest that TSLP uniquely participates in local immunity in the respiratory tract and modulation of TSLP levels may promote long-term CD8 T cell immunity in the mucosa when other prosurvival signals are limiting.

FIGURE 1

Influenza infection induces the expression of TSLP mRNA. The murine lung epithelial cell line (MLE-15) was infected with 0.5 MOI of x31 and cultured for the indicated times post infection before cells were harvested, RNA was extracted, and RT-qPCR was performed. For in vivo analysis, C57BL/6 mice were infected with 1,000 pfu x31 or mock infected with PBS i.n.. Tissues were collected at the indicated times post infection and TSLP mRNA was quantified using RT-qPCR. Data was normalized using an endogenous control and is displayed as relative quantification over mock-infected controls (1 or Ref) for each indicated time point, as determined by the ΔΔct method. Values are shown as mean RQ +/− SEM (n=3 samples/group). Data is representative of two independent experiments.

FIGURE 2

TSLP-R−/− mice develop more severe disease than WT mice. (A) WT B6 mice (n=10) and TSLP-R^−/− mice (n=9) were infected i.n. with 1,000 pfu of x31 and weighed daily. Data is shown as mean percent change in body weight over time +/− SEM. Overall significance in weight change between WT and TSLP-R^−/− mice was assessed using a two-way ANOVA test, p=0.0017. Significance between groups at individual time points was assessed using a two-tailed student t-test (*p=<0.05, **=p<0.01, ***=p<0.001). Data is representative of two independent experiments. (B) WT and TSLP-R^−/− mice were infected i.n. with 1,000 pfu x31. Viral titers were measured by plaque assay at days 2, 3, 4, 6, (left) and 8 (right) post infection and graphed as pfu/mg of lung tissue +/− SEM (n=3mice/group/day). Data is representative of two independent experiments. (C) Lymphocytes isolated from the BAL, lung and spleen of WT and TSLP-R^−/− mice (n=3/group) at 10 days post x31 infection were stimulated ex-vivo in the presence or absence of a class II-restricted HA peptide, and CD4 T cells were assessed for IFN-γ production. The top panels show representative IFN-γ staining in CD4⁺ (CD44^hi) lymphocytes. The bottom panel shows the quantification of this data. Data is representative of two independent experiments.

FIGURE 3

TSLP-R^−/− mice harbor similar frequencies of influenza-specific CD8 T cells as WT mice. Lymphocytes were isolated from the indicated tissues of x31 infected animals and analyzed by flow cytometry for tetramer reactivity at 10, 35, and 115 days p.i. (A) Representative dot plots from the BAL, Lung, Spleen, MdLN and non-draining inguinal LN (ILN) of WT and TSLP-R^−/− mice at day 10 p.i.. Cells were first gated on total CD8⁺ lymphocytes and analyzed for CD11a expression and tetramer reactivity. (B) Average percent NP-tetramer⁺ of CD8 T cells isolated from indicated tissues at days 10, 32, and 115 p.i. in WT and TSLP-R^−/− animals. Data is shown as the mean percent of tetramer positive cells of the CD8 T cells +/− SEM (n=3 mice/group), and is representative of two independent experiments. Data was analyzed for significance using a two-tailed students t-test; no significant differences were found.

FIGURE 4

TSLP-R^−/− OT-I cells are less prevalent at the site of infection than WT OT-I cells following influenza infection. (A) Competitive adoptive transfer scheme; Congenically mismatched WT OT-I and TSLP-R^−/− OT-I cells (1,000 each) were adoptively transferred via tail vein injection into congenically distinct WT mice that were infected 24 hours later i.n. with 1,000 pfu x31-OVA. (B) Representative flow from the indicated tissues at days 8 and 50 p.i. Single cell lymphocyte populations were first gated on double positive Vα2⁺, CD44^hi cells. (C) Frequencies of WT and TSLP-R^−/− OT-I cells at days 8 and 50 following infection as assessed by flow cytometry. Data is shown as pooled samples from 3 identical experiments and expressed as a ratio of TSLP-R^−/− OT-I cells/WT OT-I cells. Significance was determined using one sample t-test against a theoretical mean of 1 (**=p<0.01, ***=p<0.001).

FIGURE 5

TSLP-R^−/− OT-I cells express higher levels of CD62L than WT-OT-I cells. (A) Median fluorescent intensity was determined on the transferred populations of cells for staining against CD127 at day 50 p.i. Data is representative of three independent experiments (n=6). (B) CD62L expression was analyzed on the populations of adoptively transferred cells at day 50 p.i.. Data is pooled from three independent experiments and displayed as frequency of CD62L high cells among total WT and TSLP-R^−/− OT-I donors (with these populations within individual recipients connected by line). Data is shown as pooled samples from 3 identical experiments. Statistical significance was determined using a two-tailed student t-test comparing the mean frequency of CD62L high OT-I cells between groups (*=p<0.05, **p<0.01).

FIGURE 6

TSLP-R^−/− OT-I cells proliferate less following influenza infection than WT-OT-I cells. (A) 2,000 CD45.1 OT-I cells were transferred into CD45.2 recipient mice and infected i.n. with HKx31-OVA 24 hours later. Number of donor CD8 T cells in the BAL, Lung and Spleen were quantified using flow cytometry and displayed as mean +/− SEM (n= 4 mice per group). B and C, Congenically distinct TSLP-R^−/− OT-I and WT OT-I cells (1,000 each) were transferred into recipient mice that were infected i.n. with x31-OVA 24 hours following transfer. 100ug of BrdU was administered i.p. at 6 dpi (24 hours prior to sacrifice) and BrdU incorporation was assessed at 7 days post infection by intracellular staining followed by flow cytometric analysis. (B) The left panel depicts representative BrdU staining in either the WT or TSLP-R^−/− OT-I pools activated in the same animal. BrdU incorporation for the individual OT-I pools is quantified for all tissues analyzed on the right. Data is shown as % BrdU positive cells where each set of connected points represents the transferred populations of OT-I cells found within the same recipient mouse. Differences in the level of BrdU incorporation between the WT and TSLP-R^−/− OT-Is are depicted for the respiratory tract only (C). Significance in (B, C) was tested for using a paired student’s t-test (*=p<0.05). Data shown is representative of 3 experimental repeats. (D) Cell death was measured by antibody staining for Annexin-V and 7-AAD in the indicated tissues at 9 and 10 days p.i., each set of connected dots represents the transferred OT-I populations within the recipient mouse. Data shown is representative of two experimental repeats; significance was tested for using a paired student’s t test and no significant differences were found.