Thymic stromal lymphopoietin (TSLP) is a cytokine produced by stromal cells, epithelial cells, and basophils that acts on dendritic cells, mast cells, and CD4+ T cells. The receptor for TSLP contains a TSLP-specific receptor chain (TSLPR) and the IL-7R α-chain. Although IL-7 critically controls the expansion and survival of naive and memory CD8+ T cells, an action for TSLP on CD8+ T cells has not been reported. We now demonstrate that CD8+ T cells express TSLPR and that TSLP activates both STAT5 and Akt and induces Bcl-2 in these cells. Correspondingly, TSLP increases CD8+ T cell survival in vitro as well as in wild-type and T-depleted mice in vivo, without altering the homeostatic proliferation of these cells. Moreover, TSLP can maintain CD8+ T cells even in the absence of IL-7. Thus, our data reveal that TSLP contributes to CD8+ T cell homeostasis in both normal and lymphopenic conditions.
Thymic stromal lymphopoietin (TSLP)3 is a type I cytokine that was discovered as a factor secreted by Z210R.1 thymic stromal cells that could support the growth of NAG8/7 pre-B cells (1). In addition to its production by stromal cells, TSLP was subsequently shown to also be produced by epithelial cells and basophils (2, 3, 4, 5, 6). Human TSLP was first shown to act on dendritic cells (DC), and TSLP-activated DC could then augment the expansion and regulate the differentiation of CD4+ T cells (7, 8). Studies in the mouse revealed that CD4+ T cells were also direct targets for TSLP (9, 10), and then similar findings were observed in human cells as well (11). A variety of in vivo models indicated an important role for TSLP in allergic/atopic and inflammation responses (7, 12), including actions on mast cells (4). For example, TSLPR knockout (KO) mice did not develop inflammation in a mouse model of allergic asthma (13, 14), but adoptive transfer of wild-type (WT) CD4+ T cells into TSLPR KO mice reconstituted an inflammatory response, indicating a critical action of TSLP for CD4+ T cells (13). Interestingly, a previous report showed that irradiated TSLPR KO mice had a lower level of cellular restoration not only of CD4+ T cells, but also of CD8+ T cells, as compared with irradiated WT mice (9), which suggests a possible role for TSLP related to CD8+ T cells as well. We now demonstrate that TSLP indeed can also directly act on human and mouse CD8+ T cells and that it activates the STAT5 and Akt signaling pathways in these cells. Furthermore, we find that stimulation of CD8+ T cells by TSLP in vitro augments their survival and promotes up-regulation of the Bcl-2 antiapoptotic protein.
In lymphopenic mice following adoptive transfer, naive T cells have been reported to undergo homeostatic expansion that serves to restore/maintain the size of the T cell compartment. Homeostatic expansion requires both cell proliferation and survival. Homeostatic proliferation is dependent on the interaction of self-MHC ligands with TCR and is also regulated by IL-7 (15, 16, 17). The same mechanisms presumably help to maintain normal numbers of T cells with age (15, 18, 19, 20). Because IL-7Rα expression is essential for proliferation and survival of peripheral CD8+ T cells (19) and IL-7Rα is a shared receptor component for IL-7 and TSLP (21, 22), we investigated the role of TSLP in CD8+ T cell homeostasis. We found that TSLP does not affect homeostatic proliferation of CD8+ T cells, but instead is essential for their survival and contributes to CD8+ T cell homeostasis in both normal and lymphopenic conditions.
Materials and Methods
Mice and adoptive transfers
C57BL/6 and B6.Ly5.1 congenic mice were purchased from The Jackson Laboratory. Rag2 KO mice were obtained from Taconic Farms. Mice lacking TSLPR were described previously (9). CD8+ T cells from lymph nodes and spleens of mice were purified by positive selection using Miltenyi Biotec magnetic beads (94–98% purity). For adoptive transfer experiments, donor WT (CD45.1) and TSLPR KO (CD45.2) purified CD8+ T cells were labeled with 2.5 μM CFSE (Molecular Probes), mixed in a 1:1 ratio, and injected i.v. (1–3 × 106 cells/mouse) into Rag2 KO mice or into nonirradiated or whole body irradiated (600 cGy) WT or IL-7 KO B6 mice. All experiments with mice were performed in accordance with a protocol approved by the National Heart, Lung, and Blood Institute Animal Use and Care Committee.
CD8+ T cells that were freshly isolated or preactivated with anti-CD3 (2 μg/ml) plus anti-CD28 (1 μg/ml) for 2 days and then rested for 1 day were stimulated with 100 ng/ml TSLP or 10 ng/ml IL-7. At the indicated times after stimulation, cell supernatant was removed, and radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors (Sigma-Aldrich) was added to each sample. Cellular lysates were resolved on 4–12% Bis-Tris gels (NuPAGE) and immunoblotted with Abs to phospho-STAT5, phospho-Akt (Ser473) (Cell Signaling Technology), STAT5, or Akt (BD Transduction Laboratories). Before reprobing, blots were stripped with Re-Blot Solution (Chemicon International). To detect TSLPR on human T cells, lysates from purified nonactivated or anti-CD3 plus anti-CD28-activated CD8+ or CD4+ T cells were run on gels and blotted with Abs to human-TSLPR clone 2D10 (11) or β-actin (Sigma-Aldrich).
Collection of donor cells from recipient organs
Cells from lymph nodes, spleens, and livers were prepared using 40-μm filters (BD Biosciences). Lungs were cut to the small pieces and incubated with Liberase (Roche) and DNase I (Sigma-Aldrich) for 1 h at 37°C and then crushed through 40-μm filters. Blood was collected in tubes containing Alsever anticoagulant solution (Sigma-Aldrich) and treated with ACK to remove RBC.
CD8+ T cells were either untreated, activated with anti-CD3 (2 μg/ml) plus anti-CD28 (1 μg/ml), or activated and then rested, and then were stained with PE goat anti-mouse pAb to TSLPR (R&D) and biotinylated rat anti-mouse mAb to IL-7Rα (CD127; eBioscience), followed by allophycocyanin-streptavidin (BD Pharmingen) or isotype-matched control Abs. To analyze Bcl-2 expression, cells were fixed and permeabilized using Cytofix/Cytoperm kit (BD Pharmingen) and stained with PE anti-Bcl2 Ab or PE isotype-matched control Ab (BD Pharmingen). To detect CFSE-labeled donor cells, cells from lymph nodes, spleens, blood, lungs, or livers of recipient mice were stained with allophycocyanin anti-CD8 and PE anti-CD45.1 (BD Pharmingen). Anti-CD25, anti-CD62L (BD Pharmingen), anti-CD44, and anti-CD127 (eBioscience) fluorescence Abs were used to determine the phenotype of donor cells. For cytokine detection, cells were ex vivo stimulated with PMA (20 ng/ml) plus ionomycin (1 μM) in the presence of brefeldin A for 4 h. Then cells were fixed, permeabilized, and stained for IL-2 or IFN-γ (BD Pharmingen).
Staining with annexin V
To study CD8+ T cell survival in vitro, cells were washed in PBS, resuspended in 100 μl of annexin V buffer, and stained with FITC-annexin V plus 7-aminoactinomycin D (7-AAD; BD Pharmingen) for 15 min in room temperature. For in vivo experiments, CFSE-labeled donor CD8+ T cells were isolated from lymph nodes of recipients, stained with PE anti-CD45.1 for 30 min on ice, washed in PBS, and then stained with allophycocyanin-annexin V in annexin V buffer for 15 min.
Purified mouse CD8+ T cells at 2 × 105 cells/well were cultured without activation for 2 days, activated with anti-CD3 plus anti-CD28, or activated with anti-CD3 plus anti-CD28, and then rested for 2 days. During these conditions, cells were untreated or treated with 100 ng/ml TSLP or 10 ng/ml IL-7, and pulsed with 1 μCi of [3H] thymidine for the last 9 h of culture.
Mouse and human CD8+ T cells directly respond to TSLP stimulation
Both naive and memory CD8+ T cells express high levels of IL-7Rα, and IL-7 is an essential factor for the maintenance of these cells in vitro and in vivo (23, 24, 25). TSLP also uses IL-7Rα as part of its receptor, but whether it can stimulate CD8+ T cells has not been reported, and within T cells, the known biological effects of TSLP have largely been restricted to the CD4+ subpopulation. To evaluate the potential responsiveness of CD8+ T cells to TSLP, we first investigated whether these cells expressed TSLPR as well as IL-7Rα and compared its expression on different populations of cells. We found relatively similar levels of TSLPR expression on both CD4+ and CD8+ T cells, and higher expression of this receptor on DC (Fig. 1,A). All three populations of the cells showed high expression of IL-7Rα (Fig. 1,A). Interestingly, in contrast to the down-regulation of IL-7Rα expression on activated CD8+ T cells (19, 26), TCR activation instead increased TSLPR expression (Fig. 1,B), similar to what we previously observed in human CD4+ T cells (11), with a decline when cells were then rested for 2 days (Fig. 1 B).
Given constitutive expression of TSLPR on resting CD8+ T cells as well as its induction on these cells, we next investigated the ability of CD8+ T cells to respond to TSLP. Analogous to its effect on human and mouse CD4+ T cells (11, 27), TSLP rapidly induced the tyrosine phosphorylation of STAT5 in naive and preactivated mouse CD8+ T cells, and this was sustained for at least 1 h (Fig. 2, A and B). Moreover, TSLP induced phosphorylation of Akt (Fig. 2, A and B), indicative of its activation of the PI3K/Akt pathway, which is consistent with the presence of a YXXM motif in the IL-7Rα cytoplasmic domain, which mediates PI3K recruitment in cells stimulated with IL-7 (28, 29).
We also examined the responsiveness of human CD8+ T cells to TSLP. Freshly isolated human CD8+ T cells express little, if any, TSLPR, but activation of CD8+ T cells induced elevated TSLPR expression (Fig. 2,C) and thus augmented their sensitivity to TSLP (Fig. 2,D), analogous to our results on human CD4+ T cells (11). Although TSLP can promote tyrosine phosphorylation of STAT5 on these preactivated human cells, STAT5 activation was less potent than seen with IL-2 or IL-7 (Fig. 2 D). Together these experiments indicate that human and mouse CD8+ T cells can directly respond to TSLP.
TSLP regulates survival of CD8+ T cells in vitro and elevates expression of Bcl-2
We previously showed that TSLP increases the proliferation of TCR-activated human and mouse CD4+ T cells, but not of naive cells (9, 11). We therefore evaluated whether TSLP also affected the proliferative rate of naive and activated CD8+ T cells in vitro by thymidine incorporation and by using a CFSE dilution assay, but unexpectedly, TSLP did not affect the proliferation of these cells (Fig. 3, A and B) and only slightly increased proliferation of CD8+ T cells that were TCR activated following washing and resting in the absence of TCR stimulation (Fig. 3, A and B). As compared with TSLP, IL-7 had a more potent proliferative effect for naive and preactivated CD8+ cells, but neither cytokine significantly affected proliferation in activated cells (Fig. 3, A and B), where IL-7Rα expression is diminished.
We next assessed the effect of TSLP on CD8+ T cell viability. Based on 7-AAD and annexin V staining, TSLP augmented the percentage of viable nonactivated (Fig. 3,C) and preactivated (Fig. 3,D) CD8+ T cells in vitro, although the effect of TSLP was less potent than observed with IL-7 (Fig. 3, C and D). Given the importance of Bcl-2 in survival of T cells induced by IL-7 (30, 31, 32) and its mRNA up-regulation upon TSLP stimulation (data not shown), we decided to check expression of Bcl-2 protein in CD8+ T cells. Consistent with its effect on CD8+ T cell survival, TSLP increased expression of Bcl-2 protein in both nonactivated (Fig. 3,E) and preactivated (Fig. 3 F) CD8+ T cells, but again was less potent than IL-7.
Regulation by TSLP of CD8+ T cell homeostasis
Previously, it was reported that IL-7Rα expression is required for CD8+ T cell survival and homeostatic proliferation under lymphopenic conditions (19), and studies have focused on the role of IL-7 in this process (15, 23). To also evaluate the potential role of TSLP, we combined equal numbers of CFSE-labeled WT CD45.1 and TSLPR KO CD45.2 naive CD8+ T cells and injected these cells into Rag2 KO mice. The ratio of the WT CD45.1 and TSLPR KO CD45.2 subpopulations of CD8+ T cells was assessed in lymph nodes 7 days after adoptive transfer (Fig. 4,A). At this time, we observed that cells undergo fast and slow proliferation, and in both populations of cells, the percentage of WT cells was higher than percentage of TSLPR KO cells (Fig. 4,A). We also injected a mixed population of CD8+ T cells into C57BL/6 irradiated recipient mice and harvested cells from lymph nodes 7–9 days later (Fig. 4,B). Because fast proliferation is mostly dependent on MHC:Ag interaction with TCR (17, 33) and occurs in Rag2 KO mice, but not in irradiated mice 7–9 days after adoptive transfer, whereas slow proliferation is cytokine dependent (15, 17) and takes place in both lymphopenic conditions, we gated cells on the slow-dividing population. As shown in Fig. 4,A, right panel, and Fig. 4,B, the ratio of WT to TSLPR KO cells was ∼70:30, independent of whether the recipient mice were Rag2 KO or B6 irradiated mice. We asked whether the absence of TSLP signaling altered the rate of homeostatic proliferation of TSLPR KO CD8+ T cells, which could at least in part explain the reduction in the ratio of these cells in lymphopenic mice as compared with WT cells. Surprisingly, we did not observe a decrease in the proliferative rate of TSLPR KO cells as compared with WT CD8+ T cells within any individual host (Fig. 4,C, compare v with i, vi with ii, vii with iii, and viii with iv). As compared with nontreated Rag2 KO mice or WT irradiated mice, when we used either Rag2 KO mice injected with an anti-IL-7 Ab (Fig. 4,C, compare ii with i and vi with v) or IL-7 KO mice (Fig. 4,C, compare iv with iii and viii with vii) as recipients, there was a marked decrease in rate of homeostatic proliferation of both populations of CD8+ T cells, underscoring the critical role of IL-7. Also, we found only minor changes in the WT:TSLPR KO CD8+ T cell ratio in these mice as compared with their littermate (Fig. 4 B). Thus, TSLPR deficiency does not affect the sensitivity of CD8+ T cells to IL-7 and, conversely, a normal response to IL-7 cannot compensate for the absence of TSLP signaling.
Distribution of WT and TSLPR KO cells in different organs under lymphopenic conditions
Other possible explanations for the elimination of TSLPR KO CD8+ T cells from the lymph nodes could be the migration of these cells to other peripheral organs or their augmented death. To evaluate the effect of TSLP signaling on cell migration, we transferred WT and TSLPR KO cells in a 1:1 ratio into Rag2 KO mice and 7–9 days later evaluated the percentage of these cells in peripheral and mesenteric lymph nodes, spleen, blood, lung, and liver (Fig. 5,A). In each tissue, the percentage of TSLPR KO T cells undergoing slow homeostatic proliferation was significantly lower than the percentage of WT CD8+ T cells (Fig. 5,A), indicating that TSLP plays a role in the maintenance of CD8+ T cells in a broad range of organs. Because lymphopenic conditions can result in altered phenotypes of T cells (17), we further analyzed the expression of surface markers on WT and TSLPR KO CD8+ T cells (Fig. 5,B). Homeostatic proliferation was associated with increased expression of CD44 (i.e., those cells that underwent the greatest number of cell divisions based on CFSE dilution had the highest CD44 expression), but no such changes in expression of CD25 or CD62L were observed (Fig. 5,B). We also did not observe differences in IL-7Rα expression (CD127) (Fig. 5,B), suggesting that both populations could respond equally to IL-7. Furthermore, production of IL-2 and IFN- γ was relatively similar (Fig. 5 C).
TSLP is essential for survival of CD8+ T cells in lymphopenic and normal conditions
Finally, to determine whether TSLP contributes to the survival of CD8+ T cells in vivo, we isolated donor WT and TSLPR KO cells from peripheral lymph nodes of Rag2 KO or irradiated B6 WT host mice 7 days after adoptive transfer and stained the cells with annexin V. In both groups of recipient mice, we found a higher apoptotic rate in TSLPR KO than in WT CD8+ T cells (Fig. 6, A and B), indicating a role for TSLP in the survival of these cells. Moreover, when we analyzed WT and TSLPR KO CD8+ T cells in normal nonirradiated B6 hosts 7 days after adoptive transfer, we also observed a higher apoptotic rate in the TSLPR KO CD8+ T cells (Fig. 6,C). The increased apoptosis of these cells was also seen when they were adoptively transferred into IL-7 KO nonirradiated lymphopenic mice (Fig. 6,D). Correlating with the higher apoptotic rate of TSLPR KO CD8+ T cells after transfer into either WT or IL-7 KO nonirradiated recipients, we observed that the percentage and number of these cells were lower than seen with similarly treated WT CD8+ T cells (Fig. 7). Thus, TSLP is important for the survival of CD8+ T cells under both lymphopenic and normal conditions.
Previous studies have revealed TSLP to have important actions in allergy and inflammation (2, 3, 4, 5, 8, 13, 14), with major actions for this cytokine on DC (7, 8, 34), mast cells (4), and CD4+ T cells (9, 10, 11). However, only limited information has been available regarding a relationship of TSLP to CD8+ T cells. It was reported that human TSLP-activated DC promote activation and differentiation of CD8+ T cells into proallergic CTLs (35), but the action of TSLP in this case was indirectly mediated by DC, and direct actions of TSLP on CD8+ T cells have not been reported. In the current study, we now demonstrate that mouse and human CD8+ T cells express TSLPR and can directly respond to TSLP, as evidenced by the ability of TSLP to rapidly induce the tyrosine phosphorylation of STAT5 and serine phosphorylation of Akt. Moreover, we found that TSLP serves as survival factor for CD8+ T cells in vitro and increases expression of antiapoptotic protein Bcl-2, an essential factor in the regulation of T cell survival (36).
Development and homeostasis of human and mouse T cells strongly depend on the expression of IL-7Rα (19, 24, 25, 37), which is a component of the receptors for both IL-7 and TSLP. In humans, defective IL-7Rα expression results in a form of SCID in which T cells are profoundly diminished in number (37, 38), and Il7Ra-deficient mice also have markedly decreased T cells (19, 37). Although the phenotypes of Il7Ra and Il7 KO mice are similar, the absence of signaling by IL-7 in the latter mice can be partially compensated through TSLP overexpression, and this results in a partial restoration of T and B cells (9, 39). Additionally, it was also previously shown that TSLP injection into common γ-chain KO mice, which cannot respond to IL-7, results in a partial increase not only for CD4+ T cells and B cells, but also for CD8+ T cells (9). Our demonstration of functional receptors for TSLP on CD8+ T cells provides an explanation for this latter observation.
Given the known role of IL-7 for CD8+ T cell homeostasis, we compared the roles of IL-7 and TSLP in CD8+ T cell homeostasis. Unexpectedly, whereas IL-7 induces proliferation and survival of CD8+ cells, TSLP does not induce proliferation of these cells, but instead is required for their survival in both lymphopenic or nonlymphopenic hosts. Although it is unclear why IL-7 mediates a potent proliferative effect, whereas TSLP does not, there are signaling differences between IL-7 and TSLP, including the potent activation of Jak3 by IL-7, but not by TSLP (40).
Overall, our results demonstrate the direct action of TSLP on CD8+ T cells, and that it induces Bcl-2 expression and has an antiapoptotic effect. This has important broader implications for the role of this cytokine and suggests potential cooperative effects with other cytokines, such as IL-7 and IL-15, that are involved in the generation and maintenance of populations of CD8+ T cells (23). Finally, these studies underscore the pleiotropic actions of TSLP, extending its known roles in the biology of mast cells, DC, CD4+ T cells, and CD8+ T cells as well, and also revealing its actions as a survival factor. Thus, TSLP not only functions in pathological conditions, including inflammation and allergy, but it additionally supports T cell homeostasis, including the survival of CD8+ T cells.
We thank Dr. Ronald H. Schwartz, National Institute of Allergy and Infectious Diseases, for critical comments.
W.J.L. is a coinventor on issued patents and patent applications related to TSLP.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Division of Intramural Research, National Heart, Lung, and Blood Institute, National Institutes of Health.
Abbreviations used in this paper: TSLP, thymic stromal lymphopoietin; 7-AAD, 7-aminoactinomycin D; DC, dendritic cell; KO, knockout; WT, wild type.