Cutting Edge: Direct Action of Thymic Stromal Lymphopoietin on Activated Human CD4⁺ T Cells

Thymic stromal lymphopoietin (TSLP)³ was identified as a growth-promoting activity produced by mouse Z210R.1 thymic stromal cells that supported the development of immature NAG8/7 B cells to the B220⁺/IgM⁺ stage (1), similar to the first description of another stromal factor, IL-7, as a pre-B cell growth factor (2). TSLP acts via a receptor containing the IL-7 receptor α-chain (IL-7Rα) and a TSLP-specific subunit, TSLPR (3, 4). Interestingly, TSLPR is most similar to the common cytokine receptor γ-chain, γ_c (3, 4), which is mutated in humans with X-linked severe combined immunodeficiency (5, 6). Thus, TSLP and IL-7 receptors not only share IL-7Rα but also have similar second components (TSLPR and γ_c, respectively).

IL-7 critically regulates T cell development, as revealed in mice lacking IL-7 or IL-7Rα (7, 8). Moreover, patients with IL-7Rα deficiency have a form of SCID in which T cells are greatly diminished but other lineages are intact (9). In mice (2, 7, 8) but not humans (6, 9), IL-7 signaling is vital for B cell development, demonstrating a key species variation.

In addition to its actions on B cells, like IL-7, mouse TSLP has roles in T cell biology (10). Specifically, TSLPR/γ_c double knockout (KO) mice have more impaired T cell development than γ_c KO mice, T cell recovery is defective in sublethally irradiated TSLPR KO mice, TSLP increases the proliferation and survival of CD4⁺ single positive thymocytes and peripheral T cells, and finally, peripheral TSLPR KO CD4⁺ T cells exhibit defective expansion in γ_c/Rag2 KO-irradiated hosts (10). Mouse TSLP acts primarily on naive rather than on memory CD4⁺ T cells (11).

TSLP has been implicated in the development of asthma and atopic dermatitis (11, 12, 13, 14, 15). TSLPR KO mice have a defective lung inflammatory allergic response, but this is reversed by adoptive transfer of WT CD4⁺ T cells (11). In humans, TSLP was shown to regulate Th2 allergic responses by acting directly on dendritic cells (DCs), but not on CD4⁺ T cells, with TSLP-activated DCs in turn stimulating CD4⁺ T cells (15, 16). Because of the surprising apparent difference that mouse but not human TSLP could directly act on CD4⁺ T cells, we decided to investigate if human CD4⁺ T cells expressed TSLP receptors and the responsiveness of these cells to TSLP.

CD4⁺ T cells from PBMC purified by negative depletion using a kit (Miltenyi Biotec) were cultured at 2 × 10⁶ cells/ml in RPMI 1640 medium containing 10% FBS, 2 mM glutamine, 100 U/ml penicillin and streptomycin. DCs were purified using the blood dendritic cell isolation kit II (Miltenyi Biotec).

RNA was extracted using TRIzol (Invitrogen Life Technologies) and RNeasy (Qiagen), reverse-transcribed using the iScript cDNA synthesis kit (Bio-Rad), and hTSLPR cDNAs identified by a fluorogenic 5′-nuclease PCR assay and an ABI Prism 7900HT sequence detection system (PerkinElmer). One microgram of RNA was used for RT-PCR, and 0.2 μg of this product was amplified using internal 5′ (5′-ATACCCAAGCGACTGGTCAGA-3′) and 3′ (5′-TGGGAGGCGTTGGTGTCT-3′) primers and the TaqMan FAM-TAMRA probe (5′-TGACATGCTGGCAGAGAGGCGAGATT-3′) (Operon Biotech). TSLPR mRNA levels were measured using a standard curve relative to 18S rRNA that was detected using TaqMan FAM-MGB primers (Applied Biosystems).

BALB/c mice were immunized with a human TSLPR/IL-7Rα-transfected mouse cell line, inguinal lymph node lymphocytes were fused with myeloma cells, and hybridoma supernatants evaluated. 2D10 (mouse IgG1) mAb stains TSLPR by flow cytometry and Western blotting. It was purified using Hi-Trap Protein G HP column (Amersham Bioscience). Biotin was conjugated using a kit (Invitrogen Life Technologies).

Clarified lysates were immunoprecipitated overnight at 4°C with anti-Stat5 (Santa Cruz Biotechnology) plus protein A-agarose (Upstate Biotechnology), resolved on NuPAGE 4–12% Bis-Tris gels (Invitrogen Life Technologies), transferred to polyvinylidene difluoride membranes (Invitrogen Life Technologies), and Western blotted.

PBMC or purified CD4⁺ T cells were stained with mAbs to CD3, CD4, CD8, CD56, CD11c, CD45R/B220 (all from BD Pharmingen) or CD19 (eBioscience) for 30 min in the presence of FcR-blocking mAb (Miltenyi Biotec). For TSLPR, cells were stained with biotinylated 2D10 or an IgG1 control mAb (BD Pharmingen) and FITC-streptavidin (BD Pharmingen). For IL-2Rα, cells were stained with PE-anti-CD25 (BD Pharmingen). Pre-activated CD4⁺ T cells were rested for 1 day and stimulated with 50 ng/ml hTSLP (R&D Systems) or 100 U/ml hIL-2 (Roche) in serum-free medium. Cells were fixed and permeabilized using Cytofix/Cytoperm kit (BD Pharmingen) and stained with PE-anti-phospho-Stat5 mAb (BD Phosflow) or PE-isotype-matched control and allophycocyanin-anti-CD4 (BD Pharmingen). In some experiments, cells were 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.

CD4⁺ T cells at 2 × 10⁵ cells/well were activated for 5 or 6 days with plate-bound anti-CD3 or IL-2, with or without 50 ng/ml TSLP, and pulsed with 1 μCi of [³H]thymidine for the final 16 h of culture. In some experiments, cells were pre-activated for 3 days and incubated in 96-well flat-bottom plates for 2 or 6 days in medium or with TSLP, IL-2, or IL-7, and then pulsed with 1 μCi of [³H]thymidine. Proliferation was also examined by labeling cells with 2.5 μM CFSE for 8 min at room temperature and monitoring CFSE dilution.

As noted above, whereas mouse CD4⁺ T cells respond directly to TSLP (10), human TLSP has been reported to only activate peripheral blood CD11c⁺ DCs but to not act directly on other DCs or T or B cells (15, 16). We analyzed TSLPR expression on resting and activated human CD4⁺ T cells and the responsiveness of these cells to TSLP to determine whether TSLP indeed exhibited as dramatic a species variation between humans and mice as is seen with IL-7, with mice but not humans requiring IL-7 for B cell development (6, 7, 8, 9). In freshly isolated human PBMC, as expected, TSLPR expression was readily detected on CD11c⁺ DCs (Fig. 1,A), but not on CD19⁺ B cells (Fig. 1,A), CD4⁺ T cells (Fig. 1,B), or CD8⁺ T cells (Fig. 1 B).

As TSLP production increases with inflammation (14, 15), we investigated whether cellular activation also induced TSLPR expression on rigorously purified human CD4⁺ T cells, which contained <1% CD11c⁺ DCs, approximately 1% B220⁺ B cells, and no detectable CD8⁺ T cells (Fig. 1,C). Although freshly isolated CD4⁺ T cells had little if any TSLPR mRNA, activation by anti-CD3 plus anti-CD28 (Fig. 1,D) or PHA-L (Fig. 1,E) induced TSLPR mRNA within 1 day. TSLPR mRNA levels were lower at days 2 and 3 (Fig. 1, D and E) but persisted for at least 14 days (Fig. 1,D). TSLPR mRNA levels in non-activated CD4⁺ T cells were, as expected, lower than in DCs (16), but increased after activation (Fig. 1,F). We also analyzed TSLPR expression on activated CD4⁺ T cells by flow cytometry (Fig. 1,G; see shift in 2nd and 3rd panels compared with the control mAb) and by Western blotting (Fig. 1,H, top panels, lanes 5–8 vs 1–4) using 2D10 mAb, revealing increased expression with activation that was stable after 1 day of rest (Fig. 1,G). 2D10 mAb is specific for TSLPR, as revealed by its Western blotting the proper sized band from TSLPR transfected but not from control Jurkat T cells (Fig. 1 I).

TSLP can activate Stat5 (17), so we evaluated whether TSLP induced Stat5 phosphorylation in human T cells. Freshly isolated CD4⁺ T cells responded to IL-2 and IL-7 but not to TSLP (Fig. 1,H, 2nd panel from top, lane 2–4), whereas all 3 cytokines induced Stat5 phosphorylation in pre-activated CD4⁺ T cells (lanes 6–8), correlating with TSLPR expression in these cells (Fig. 1 H, top panel, lanes 5–8 vs 1–4). Total Stat5 and actin expression were not affected by any of the treatments (bottom panels).

We next analyzed the kinetics of Stat5 phosphorylation in CD4⁺ T cells pre-activated with anti-CD3 plus anti-CD28 that were approximately 98.5% pure, with low to absent CD11c⁺, CD8⁺, CD56⁺ and B220⁺ cells (Fig. 2,A). TSLP rapidly induced tyrosine phosphorylation of Stat5, although it was less potent than IL-2 (Fig. 2,B). Similar results were obtained with cells pre-activated with PHA-L (Fig. 2,C). We confirmed TSLP-induced Stat5 activation by intracellular staining of phospho-Stat5 (Fig. 2,D). Using a higher dose of TSLP did not further increase Stat5 phosphorylation (data not shown). Because we simultaneously stained each sample in Fig. 2,D with anti-phospho-Stat5 and anti-CD4, Stat5 activation was indeed occurring in CD4⁺ T cells. Given the short stimulation period, the TSLP effect was a direct action on these cells. Because some cytokines activate more than one STAT protein (6), we also compared the ability of IL-2, IL-7, and TSLP to induce activate Stat1 and Stat3. In activated CD4⁺ T cells, TSLP activated Stat5 but mediated little if any phosphorylation of Stat1 and Stat3 (Fig. 2 E, 5th vs 1st and 3rd panels), in contrast to IL-2 and IL-7.

To further analyze the functionality of TSLP receptors on CD4⁺ T cells, cells were activated with anti-CD3 with or without TSLP for 5 days. Strikingly, the addition of TSLP substantially increased proliferation even at low concentrations of anti-CD3 that by themselves induced little if any cell division (Fig. 3,A). Based on a CFSE-labeling experiment, in the presence of TSLP, there was an increase in the number of responding cells, rather than simply an increase in the proliferation of a small population of cells (Fig. 3,B). We also evaluated whether TSLP affected cell viability, but found little if any effect on cells treated for 3 or 5 days with medium, anti-CD3, or anti-CD3 plus anti-CD28, as evaluated by staining with 7-AAD and annexin V (Fig. 3 C and data not shown).

Given that TSLP can augment the proliferation of CD4⁺ T cells after activation, we asked if TSLP could maintain this effect even after washing and removal of TCR stimulation. We activated cells and cultured them for 2 or 6 days in the medium, TSLP, IL-2, or IL-7 (Fig. 3, D). In each case, the presence of cytokine augmented proliferation.

Because TSLP induces Stat5 phosphorylation, we evaluated the induction of two genes, IL2RA (Fig. 4,A) and CIS (Fig. 4,B), known to be regulated by Stat5 (18, 19, 20). TSLP increased IL-2Rα and CIS mRNA expression, albeit less potently than seen with IL-2 or IL-7 (Fig. 4, A and B), consistent with the lower Stat5 phosphorylation induced by TSLP (Fig. 1,H). Flow cytometry confirmed a TSLP-induced increase in IL-2Rα expression, both on non-treated or anti-CD3 activated CD4⁺ T (Fig. 4,C); the effect was greater on the latter cells. Consistent with TSLP-induced IL-2Rα expression, which increases high-affinity IL-2R expression and responsiveness of cells to low concentrations of IL-2 (21, 22), TSLP increased IL-2-induced proliferation of CD4⁺ T cells even at low concentrations of IL-2 (Fig. 4 D). In contrast, TSLP did not augment IL-7-mediated proliferation of these cells (data not shown), indicating specificity to its effect.

TSLP has pleiotropic effects in both T and B cell biology and actions related to inflammatory/allergic/atopic disease. As noted above, TSLP and IL-7 both are stromal factors (1, 2) and share IL-7Rα as a receptor component (3, 4). IL-7 exhibits a marked species differences, being critical for T cell development in both humans and mice but for B cell development only in mice (7, 23). Much discussion has focused on species-specific actions of TSLP (11, 15, 24). TSLP was reported to have certain B cell and T cell-related actions in mice (10, 11, 25), whereas human TSLP was reported to activate DCs with only an indirect effect on CD4⁺ T cells (15). Subsequently, studies in mice revealed that mouse TSLP could also act on DCs (11, 12). We now show that activated human CD4⁺ T cells express TSLPR and that TSLP can rapidly activate Stat5 and induce expression of Stat5 target genes. The greater effect of TSLP on activated than on resting T cells correlates with its induction by inflammatory processes (14, 15) and the activation-induced expression of TSLPR that we now show. The ability of TSLP to induce IL-2Rα expression may explain the higher sensitivity of CD4⁺ T cells to low doses of IL-2 in the presence of TSLP, and it is possible that such an increase might augment an immune response even in a setting of low TCR engagement.

Overall, our findings demonstrate that human TSLP can act on activated CD4⁺ T cells as well as DCs, identifying TSLPR as a potential target for directly manipulating human CD4⁺ T cell responses. This may facilitate our better understanding the pathophysiological mechanisms underlying conditions associated with elevated TSLP expression, such as allergic asthma and atopic dermatitis, as well as having possible therapeutic implications.

We thank Dr. Rosanne Spoliski for valuable discussions.

The authors have no financial conflict of interest.