IL-33 is an IL-1 family member recently identified as the ligand for T1/ST2 (ST2), a member of the IL-1 receptor family. ST2 is stably expressed on mast cells and Th2 effector T cells and its function has been studied in the context of Th2-associated inflammation. Indeed, IL-33 induces Th2 cytokines from mast cells and polarized mouse T cells and leads to pulmonary and mucosal Th2 inflammation when administered in vivo. To better understand how this pathway modulates inflammatory responses, we examined the activity of IL-33 on a variety of human immune cells. Human blood-derived basophils expressed high levels of ST2 receptor and responded to IL-33 by producing several pro-inflammatory cytokines including IL-1β, IL-4, IL-5, IL-6, IL-8, IL-13 and granulocyte macrophage colony-stimulating factor. Next, utilizing a human Th2-polarized T cell culture system derived from allergic donor blood cells, we found that IL-33 was able to enhance antigen-dependent and -independent T cell responses, including IL-5, IL-13 and IFN-γ production. IL-33 activity was also tested on Vα24-positive human invariant NKT (iNKT) cells. In the presence of α-galactosylceramide antigen presentation, IL-33 dose dependently enhanced iNKT production of several cytokines, including both IL-4 and IFN-γ. IL-33 also directly induced IFN-γ production from both iNKT and human NK cells via cooperation with IL-12. Taken together, these results indicate that in addition to its activity on human mast cells, IL-33 is capable of activating human basophils, polarized T cells, iNKT and NK cells. Moreover, the nature of the responses elicited by IL-33 suggests that this axis may amplify both Th1- and Th2-oriented immune responses.
ST2 (also known as T1, Fit-1 or DER4) is an IL-1 receptor family member implicated in regulating type 2-associated immune responses. The receptor exists as a full-length membrane-spanning molecule as well as a soluble ST2 (sST2), decoy variant that is expressed by fibroblasts and other cells. Full-length ST2 is constitutively expressed on mature mast cells as well as mast cell precursor cells in both mice and humans (1–3). Expression of ST2 is associated with a Th2-oriented memory effector phenotype and in vitro polarization studies have implicated a restricted expression pattern of ST2 on committed IL-4-expressing T cells (4–6). In a later study, ST2 expression on CD4+ T cells localized with a Th2-dominated immune response in the lungs of Schistosoma-infected mice. Upon ex vivo stimulation, some of the ST2+ T cells produced IL-2 and IFN-γ in addition to IL-4 and IL-5, suggesting that ST2 expression may not associate with a classical Th2 phenotype in vivo (7). Studies with ST2-deficient animals suggest that ST2 expression is not obligatory for the development of a normal Th2 response but may be required for effective primary responses, possibly mediated by cells other than antigen-specific memory cells (8, 9).
In two studies with human cells, ST2 expression on T cells was detected on Th2- but not on Th1-polarized blood cell lines or CD4+ T cell clones and only after activation (10, 11). In the first study, ST2+ lymphocytes were associated with IL-4, IL-5 and IFN-γ expression (10). The authors also reported a small percentage of circulating ST2+ cells that were positive for the surface markers CD3 and CD56, possibly representing NKT cells, which are distinguished by their ability to recognize glycolipid antigens presented by the MHC-like molecule CD1d (12).
Recently, IL-33 (previously identified as nuclear factor in high endothelial venules) was identified as the natural ligand for ST2 (13, 14). IL-33 is expressed in multiple tissues and by several cell types such as dermal fibroblasts and small airway epithelial and bronchial smooth muscle cells. IL-33 activated IL-1-like signaling responses in mast cells and enhanced IL-5 and IL-13 production from murine Th2-polarized splenocytes. Challenge of naive mice with IL-33 led to increased levels of pulmonary and circulating IL-5 and IL-13, elevated IgA and IgE, splenomegaly and histological signs of inflammation in the lungs, again indicating that ST2-expressing cells can mediate a significant inflammatory response even in a naive setting (14). Different groups have recently identified the IL-1R accessory protein (AcP), as the second subunit of the IL-33 receptor, obligatory for ST2-mediated responses (15–17).
Allakhverdi et al. (3) recently demonstrated that human mast cells and mast cell progenitors are activated by IL-33. Treatment with IL-33 led to increased levels of several cytokines including IL-5, IL-13, granulocyte macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor-α (TNF-α), and several of these responses were enhanced in the presence of the hematopoietin cytokine TSLP. In order to better understand the range of biological effects of IL-33, we determined the activity of IL-33 on various human hematopoetic immune cells. We found that the range of IL-33 target cells extends well beyond mast cells and Th2 cells and moreover that depending on the cellular and cytokine context, IL-33 is able to potently enhance production of IFN-γ as well as classical Th2 cytokines.
Cell culture and reagents
Human PBMC, invariant NKT (iNKT) and NK cells were cultured in RPMI 1640 medium containing L-glutamine, penicillin (100 U ml−1), streptomycin (100 μg ml−1) (Invitrogen, Carlsbad, CA, USA) and 5 (Th2-expanded PBMC) or 10% (iNKT, NK and PBMC) normal human serum (Gemini BioProducts, West Sacramento, CA, USA), heat inactivated at 56°C for 20 min prior to use. The cultures were maintained in a 37°C incubator with 5% CO2. Human IL-2, IL-3, IL-5, IL-12 and IFN-γ ELISA kits as well as human IL-23 were purchased from R&D Systems (Minneapolis, MN, USA). Human IL-18 was purchased from MBL (Woburn, MA, USA). All other cytokines were either obtained from R&D Systems or expressed and purified at Amgen. Human IL-33, specifically, was expressed and purified from Escherichia coli as an untagged mature cytokine (amino acids 112–270). IL-33 was formulated in PBS containing 1 mM dithiothreitol and 1 mM EDTA. Endotoxin levels were <0.1 EU per 1 μg of cytokine as determined by the LAL method. PE-conjugated anti-Vα24-JαQ mAb 6B11, huCD1d:Ig dimer, APCs-conjugated anti-CD3, Alexa-488-conjugated anti-CD4, PE-conjugated anti-CD56 and APC-conjugated anti-IFN-γ and isotype controls were purchased from BD Biosciences (San Jose, CA, USA). Anti-huST2 antibody as well as recombinant sST2 (either ST2-Flag-His or ST2–Fc fusion protein) was expressed and purified at Amgen. Both forms of sST2 were shown to effectively inhibit IL-33 in vitro (data not shown). House dust mite (HDM) preparation was obtained from Hollister Stier Laboratories (Spokane, WA, USA). APC-conjugated CD1d tetramer was purchased from Proimmune (Bradenton, FL, USA). A modified α-galactosylceramide (α-GalCer), PBS-57, as described in (18), was kindly provided by Paul Savage (BYU, Provo, UT, USA). Luminex multiplex immunoassay kit was purchased from Linco Research (Billerica, MA, USA). TaqMan gene expression assays for hypoxanthine guanine phosphoribosyltransferase (HPRT), IL-1RL1 (full-length ST2), IL-18R and IL-1R AcP were purchased from Applied Biosystems (Foster City, CA, USA).
Primary human basophil isolation, analysis and culture
Primary human basophils were purified from healthy donor PBMCs that were enriched from leukopheresis product (Puget Sound Blood Center, Seattle, WA, USA) using a CYTOMATE Cell Processing System (Baxter, Deerfield, IL, USA) and basophils were isolated using the Human Basophil Isolation Kit II (Miltenyi Biotec, Auburn, CA, USA). The purity of isolated basophils was assessed based on cell–surface marker expression using the following antibodies, anti-CD123-APC (Miltenyi Biotec), anti-CD203c-PE (Miltenyi Biotec) and anti-FcεRI-FITC (eBiosciences, San Diego, CA, USA). Basophils were cultured in RPMI 1640 + 5% heat-inactivated fetal bovine serum + 10 mM HEPES + 20 μg ml−1 gentamicin (Invitrogen) in the presence or absence of 100 ng ml−1 human IL-3, 100 ng ml−1 human IL-33, soluble human ST2-Flag-His or soluble human IL-18R-Flag-His (Amgen). Cell-free supernatants were assessed for IL-5 protein levels by ELISA and also submitted for multi-analyte profiling (MAP) (Rules-Based Medicine, Austin, TX, USA).
Human Th2 T cell expansion and activation
Fresh human PBMCs were purified by centrifugation over Isolymph separation medium (CTL Scientific Supply, DeerPark, NK, USA). Mononuclear cells were aspirated, washed in PBS and diluted to 1 × 106 cells ml−1 in RPMI 1640 + 5% heat-inactivated pooled human serum and 1/100 of HDM antigen preparation. PBMCs were incubated for 10 days and supplemented on days 1, 3, 5 and 7 with 10 ng ml−1 IL-4. At the time of PBMC isolation, autologous APCs were prepared by E-rosetting PBMC and centrifugation through Isolymph. The resulting T-depleted cells were washed in PBS and frozen in liquid nitrogen. On day 11, APC were thawed, washed and placed in 96-well plates overnight at 1 × 105 cells per well with 1/100 HDM. The next day, day 12 T cell cultures were centrifuged over Isolymph to remove dead cells and distributed into 96-well plates at 1 × 105 cells per well with or without the HDM-pulsed APC. To some samples, IL-2 or IL-33 was added at a final concentration of 10 ng ml−1. Cells were incubated for 18 h and supernatants were collected for cytokine quantification by ELISA. All treatments were performed in triplicate.
Isolation and expansion of iNKT cells
Human PBMCs from healthy donors were purified from leukopheresis product (provided by Puget Sound Blood Center) using a CYTOMATE Cell Processing System (Baxter). Approximately 2 × 109 PBMCs from one donor were used to generate ∼3 × 108 iNKT cells; iNKT cells were isolated from 109 PBMCs and the remaining 109 PBMCs were frozen to use as APC during expansion [monocyte-derived dendritic cells (DCs), described separately]. Magnetically activated cell sorting using PE-conjugated anti-Vα24-JαQ antibody 6B11 followed by anti-PE microbeads (Miltenyi Biotec) was employed to isolate iNKT cells from the PBMCs. For the first round of stimulation, 6B11+-positive cells were seeded at 1 × 104 cells per well of a 96-well round bottom plate in 0.2 ml media and stimulated with 100 ng ml−1 of PBS-57. The next day, IL-2 (50 ng ml−1) was added to the cultures. To expand the cells, subsequent stimulations were performed every 7–10 days for 4 weeks by seeding iNKT cells at 2–5 × 104 cells per well and stimulating with cesium-irradiated (3000 Rads), autologous monocyte-derived DC (isolated as described below) pulsed with 1 μg PBS-57 in 1 ml media for 1 h at 37°C prior to irradiation. The iNKT cells and DC were cultured at a ratio of 2:10 (iNKT cell:DC), depending on cell recovery throughout the expansion procedure. The next day, IL-2 (10 ng ml−1) was added to the cultures. iNKT cell phenotype was monitored by FACS for enrichment of 6B11+ cells at the end of every stimulation. In all, 2 × 105 cells per 0.1 ml were co-stained with PE-conjugated 6B11 mAb and APC-conjugated anti-CD3. For tetramer staining, cells were co-stained with human APC-conjugated CD1d tetramer loaded with PBS-57 and PE-conjugated 6B11. Cell events were acquired on a FACSCalibur using CellQuest software (BD Biosciences) and analyzed using FCS Express software (BD Biosciences).
Isolation of monocyte-derived DCs
Monocyte-derived DCs were isolated by incubating 5 × 107 PBMCs in X-Vivo 15 media (Lonza, Baltimore, MD, USA) in a T-75 flask. After 1–2 h, the non-adherent cells were removed and the adherent cells (primarily monocytes) were gently washed with PBS and then cultured in X-Vivo 15 supplemented with IL-4 (30 ng ml−1) and GM-CSF (50 ng ml−1) for 5–7 days. One day prior to harvest, CD40 ligand (1 μg ml−1) was added and the next day, adherent cells were lifted using cell dissociation buffer (Invitrogen), washed in media and used as APC as indicated above.
Isolation of NK cells
An NK Cell Isolation kit (Miltenyi Biotec) was used to indirectly isolate NK cells from human PBMCs. Briefly, PBMCs were magnetically labeled with a cocktail of biotin-conjugated antibodies to T cells, B cells, stem cells, DCs, monocytes, granulocytes and erythroid cells, which were then depleted using anti-biotin microbeads. FACS analysis of the isolated cells indicated that >97% of the cells were CD56 positive.
Loading of human CD1d tetramer and dimer
APC-conjugated human CD1d tetramer loaded with PBS-57 was used for FACS analysis of the iNKT cell line and was prepared according to the manufacturer's recommendation. Briefly, PBS-57 in dimethyl sulfoxide (DMSO) (0.2 mg ml−1) was incubated for 12 h at 37°C with gentle agitation and then sonicated for 10 min prior to use. A 12 molar excess of PBS-57 (0.1 μl per reaction) was added to tetramer (2 μl) and incubated at 37°C overnight in the dark. Human CD1d:Ig dimer loaded with PBS-57 was used to stimulate the iNKT cells and was also prepared according to the manufacturer's recommendation. Briefly, a 40 molar excess of PBS-57 (0.55 μg) in DMSO was mixed with the CD1d:Ig (4 μg) in PBS at 37°C overnight. To stimulate the iNKT cells, 4 μg of the CD1d:Ig was used per each condition.
Activation of iNKT cells, NK cells and PBMC
For stimulations, cells were cultured at 1–2 × 105 cells per well of a 96-well round bottom plate in 0.2 ml media, in triplicate. For TCR activation of the iNKT cells, cells were cultured with either nothing or anti-CD3 mAb (1 μg ml−1) plus recombinant IL-2 (10 ng ml−1), or in a separate experiment, with CD1d:Ig dimer loaded with DMSO or PBS-57. Increasing concentrations of huIL-33 were added to cultures that were stimulated with PBS-57-loaded CD1d:Ig. Supernatants were collected at 20 h and analyzed by multiplexed immunoassay. For non-TCR activation of iNKT, cells were stimulated with IL-12 (1 ng ml−1), IL-18 (10 ng ml−1) or IL-33 (10 ng ml−1) alone or combined as indicated. Supernatants were collected 18–20 h later and analyzed by IFN-γ and IL-4 ELISA. NK cells were also stimulated and analyzed in this manner. For IL-23 stimulation of NK cells, the cells were incubated with IL-2 (10 ng ml−1) for 7 days, washed in PBS and then stimulated with IL-23 (10 ng ml−1) alone or in combination with IL-33 (10 ng ml−1). For PBMC stimulations, frozen PBMCs from healthy donors isolated from leukopheresis product were thawed and seeded at 2 × 106 ml−1 and treated with nothing (medium), IL-12 (5 ng ml−1) or IL-12 in combination with IL-33 (10 ng ml−1) for 20 h. Supernatants were harvested at 18–20 h and analyzed by IFN-γ by ELISA as indicated. All treatments were performed in triplicate.
Intracellular cytokine staining
PBMC cultures and Th2-polarized T cells were activated as previously described. Four hours prior to harvesting the cells for intracellular cytokine (ICC) analysis, GolgiPlug (BD Biosciences) was added to half of the cultures; supernatants were collected from the remaining cultures for analyzing IFN-γ or IL-4 levels by ELISA. For ICC analysis, cells were stained simultaneously for CD4 and CD56 expression prior to fixation, permeabilization and washing using the Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's protocol. The cells were then stained on ice for intracellular IFN-γ and IL-4 expression as indicated. In total, 5 × 104 events were collected on a FACSCalibur using CellQuest software (BD Biosciences) and analyzed using FlowJo software (Tree Star).
iNKT and NK cells were stimulated with IL-12 p70 (10 ng ml−1) for various periods of time as indicated, and the cells were harvested and RNA was prepared from cell lysates using the Versagene RNA purification kit (Qiagen). cDNA was made using High-Capacity cDNA kit (Applied Biosystems) and TaqMan reactions performed using ∼100 ng cDNA and TaqMan buffer and gene expression assays (Applied Biosystems) using a PRISM 7900HT Sequence Detection System (Applied Biosystems). All reactions were performed in triplicate.
Human basophils express ST2 and respond to IL-33
Given that basophils and mast cells are thought to develop from common CD34+ progenitors (19) and the demonstrated ability of IL-33 to stimulate mature mast cells and mast cell progenitors (3), we hypothesized that basophils might also be a target cell for IL-33. Basophils were enriched from healthy donor PBMC to high levels of purity based on surface expression of CD123, CD203c and FcεRI (Fig. 1A). RNA message specific for full-length ST2 was found to be highly expressed in freshly isolated basophils compared with various primary human myeloid and lymphoid cells (Fig. 1B). Interestingly, the relative level of full-length ST2 message in basophils (as compared with HPRT) was very similar to the levels previously reported for mature human mast cells (3). We have not yet measured the level of sST2 messenger RNA (mRNA) that may also be expressed in these cells. Currently, there is not a quantitative PCR (qPCR) assay available specific for human sST2. Although we were unable to reproducibly detect ST2 expression by cell–surface staining of basophils from all donors (data not shown), we have previously shown that detection of surface expression by flow cytometry is not in itself a requisite for functional responses (3). We therefore tested the ability of freshly isolated basophils to respond to IL-33 stimulation alone or in the presence of IL-3, a known growth and survival factor for basophils (20). Basophils were stimulated for 48 h and cell-free supernatants were analyzed by MAP to determine the range of mediators that may be released. IL-33 stimulation alone induced modest production of multiple cytokines including IL-4, IL-5, IL-6, IL-8, IL-13 and GM-CSF (Fig. 1C) as well as some chemokines (MCP-1, MIP-1α and MIP-1, data not shown). Interestingly, when IL-33 was combined with IL-3, significant enhancement of most cytokines was observed. At the mRNA level, however, IL-3 treatment did not augment the expression of full-length ST2 (data not shown). The induction of IL-5 by IL-33 + IL-3 was confirmed separately by ELISA from basophil cultures derived from two additional donors and a soluble version of ST2, but not IL-18R, was able to inhibit the response, establishing specificity of the observed IL-33 activity (Fig. 1D).
IL-33 enhances antigen and non-antigen responses from Th2-skewed human T cells
We developed an in vitro culture system in which HDM-specific human Th2 T cells can be expanded from the peripheral blood T cell pool (D. Kaufman, M. R. Comeau and R. Armitage, in preparation). We used this culture system to examine the effect of IL-33 on human Th2-polarized T cells. Allergic donor PBMCs were cultured for 12 days in the presence of IL-4 and HDM to expand antigen-specific human Th2 cells. The cultures at this point were comprised primarily of CD4+ cells (75–85%) with minority populations of CD8+ (5–10%) and CD56+ (5–10%) cells present (data not shown and Fig. 3A). On day 12, the cells were activated in the presence or absence of IL-2 with autologous APC that had been pulsed overnight with HDM and intracellular and secreted cytokine levels were determined 24 h later. Depletion of CD8+ and CD56+ cells at day 0 had no significant effect on subsequent cytokine production, indicating that the antigen responses were elicited primarily by the CD4+ population (data not shown). Flow cytometry revealed a low level of ST2 surface expression on CD4+ cells at day 13 (Fig. 2A) and full-length ST2 mRNA was also detected in the day 13 culture by reverse transcription–PCR (data not shown). HDM-activated Th2 cultures routinely produced IL-5 and IL-13, two of the prototypical Th2 cytokines, and these responses were augmented by the addition of IL-2 (Fig. 2B). Addition of IL-33 routinely elevated the levels of IL-5 and IL-13 produced in response to HDM in experiments with T cell cultures from nine different donors and this effect was also observed for the antigen-independent responses (IL-2 alone) (Fig. 2B). Interestingly, activated cells from multiple donors also produced the Th1 cytokine IFN-γ and this response was significantly enhanced by the addition of IL-33. The specificity of these effects was established by the ability of sST2, but not a control protein, to block the IL-33-dependent enhancement in effector cytokine production (Fig. 2C). In separate experiments, sST2 by itself was shown to have no effect on HDM-induced activation (data not shown).
In order to further examine the cellular source of IFN-γ, we next performed ICC staining. As shown in Fig. 3(A), after addition of a 1:1 ratio of T cell-depleted APC, the cellular make up of the culture at day 13 comprised a mixture of CD4+ cells and CD4− cells as well as some CD56+ cells. Gating on CD4+ cells revealed that prior to activation the cells exhibited an IL-4+ and IFN-γ− phenotype indicative of their polarization to Th2 cells (Fig. 3B). Interestingly, 20 h after activation with HDM, most of the intracellular IL-4 was no longer present, whereas significant amounts of intracellular IFN-γ were now detected, indicating that the CD4+ T cells were indeed a predominant source of IFN-γ. Addition of IL-33 further enhanced the level of IFN-γ expressed in CD4+ cells. Activation with HDM also led to an increase in IFN-γ staining in the less abundant CD56+ population and this response too was augmented by the addition of IL-33. This result suggested that non-CD4 cells were contributing at least some of the measured IFN-γ, even though as mentioned earlier, depletion of CD56+ cells at initiation of culture does not significantly affect subsequent levels of extracellular cytokines following activation (data not shown). The identity of the CD56+ cells in the culture is undetermined but may be a mixture of NK and/or NKT cells. The reason for a loss of intracellular IL-4 following activation is unclear; however, this may reflect activation-induced release of preformed cellular stores of IL-4 and the ability to measure both IL-4 and IFN-γ in the supernatants following activation is consistent with this explanation (Fig. 3C).
IL-33 enhances type 1 and type 2 responses from human iNKT cells
In order to examine the potential activity of IL-33 on human iNKT cells, we established a primary cell culture via enrichment with a mAb against the invariant Vα24 TCR (6B11). After 5 weeks of expansion with autologous DCs and the prototypical iNKT ligand α-GalCer, we obtained a population that was ∼95% pure CD3+, Vα24+ iNKT cells (Fig. 4A). The population comprised ∼70% CD4+, 10% CD8+ and 20% double-negative cells and ∼15% were CD56+ (data not shown). Although not readily detectable by flow cytometry (data not shown), RNA specific for full-length ST2 was detected by qPCR (as shown in Fig. 7A). As shown in Fig. 4(B), recombinant CD1d bound to the iNKT cells only when loaded with α-GalCer, demonstrating both TCR specificity and effective antigen loading. Treatment with α-GalCer-loaded CD1d led to the production of IL-4, IL-5, IL-13, TNF-α, IFN-γ and IL-2. We found that all these antigen-driven responses were dose dependently augmented by the addition of IL-33 (Fig. 4C). These results indicate that in addition to traditional MHC-restricted memory T cells, IL-33 will also enhance the effector function of iNKT cells. Moreover, IL-33 treatment led to elevated levels of both Th1 (IFN-γ and TNF-α) and Th2 (IL-4, IL-5, IL-13) cytokines.
In cooperation with IL-12, IL-33 directly induces IFN-γ production from human iNKT and CD56-positive NK cells
In addition to TCR engagement, NKT cells can be activated directly by cytokines. A recent study demonstrated a role for mouse NKT-produced IFN-γ in amplifying innate responses to endotoxin challenge via a mechanism involving direct activation by IL-18 and IL-12 (21). Due to the similarity in receptor signaling between IL-33 and IL-18, we sought to determine if IL-33 or IL-18 could activate human iNKT cells in the absence of antigen. As shown in Fig. 5(A), neither IL-33 nor IL-18 was able to induce either IFN-γ or IL-4 when used alone and IL-12 by itself induced only a small amount of IFN-γ. However, when IL-12 was combined with either IL-33 or IL-18, a synergistic effect was seen leading to significantly increased levels of IFN-γ. ICC staining indicated that the CD4+ iNKT cells were the primary producers of IFN-γ (data not shown). Interestingly, neither IL-18 nor IL-33 alone or in combination with IL-12 induced IL-4 production although in the same experiment IL-4 was produced in response to TCR activation (Fig. 5A). The activity of IL-33 on the iNKT cells in this assay was specifically inhibited by either sST2 (data not shown) or a mAb against human ST2 but not by an isotype control antibody (Fig. 5B).
NK cells are major sources of IL-18-induced IFN-γ and the synergy between IL-18 and IL-12 in activating NK cells is well established (22). Having seen that IL-33 can synergize with IL-12 to induce IFN-γ from iNKT cells, we wished to see if this activity extended to NK cells. For these experiments, we utilized CD56+ NK cells directly purified from fresh human PBMC (greater than 97% purity, Fig. 6A). IL-33 or a sub-optimal concentration of IL-12 did not activate the NK cells by themselves, but when combined, led to production of greater than 10 000 pg ml−1 IFN-γ (Fig. 6B). Although the absolute level of IFN-γ that was produced varied from donor to donor, the ability of IL-33 to activate NK cells was observed using NK cells from all four donors examined. As shown in Fig. 6(C), the dose–response curves and level of IFN-γ induction from a single donor were virtually equivalent between IL-18 and IL-33, indicating that the potency of IL-33 on human NK cells is equivalent to IL-18. Moreover, the activity of IL-33 could be specifically inhibited by either sST2 (data not shown) or an antibody against ST2 but not a control antibody (Fig. 6D). We have also observed that IL-18 will cooperate with the other p40-containing IL-12 family member IL-23 to induce IFN-γ from human NK cells (data not shown), so we asked whether this was also true for IL-33. As shown in Fig. 6(E), IL-23 treatment alone did not induce IFN-γ from NK cells; however, when combined with IL-33, synergy led to production of IFN-γ levels that were nearly as high as that induced in the presence of IL-12. We also examined the effect of IL-33 on total human PBMCs and found that treatment with IL-33 + IL-12 resulted in the production of greater than 1 ng ml−1 IFN-γ after overnight stimulation (Fig. 7A). This activity was observed with PBMCs from four different individuals and ICC staining revealed that the predominant IFN-γ-producing cells were most likely NK cells, as evidenced by their CD4− and CD56+ staining pattern (Fig. 7B). Taken together, these results indicate that, like IL-12 and IL-18, IL-33 is a potent stimulator of NK cell-derived IFN-γ.
IL-12 treatment dramatically increases expression of ST2 in iNKT and NK cells
In order to better understand the potential mechanism underlying the synergy observed between IL-33 and IL-12 or IL-23, we examined the regulation of full-length ST2 mRNA. As mentioned earlier, low levels of ST2 expression were detected in the iNKT cells; however, following IL-12 treatment for 20 h, they were elevated nearly 100-fold (Fig. 8A). Steady-state IL-18Rα mRNA levels were higher than ST2 mRNA levels in the iNKT cells, yet were also further increased upon IL-12 treatment. Likewise, in NK cells, full-length ST2 mRNA expression was significantly augmented by IL-12 treatment and to a lesser extent by IL-23 (Fig. 8B). IL-18Rα expression was not as significantly affected by IL-12 or IL-23. We next assessed the kinetics of IL-12 induction and found that in NK cells from two different donors, IL-12 treatment led to an increase in ST2 levels within 2–6 h and these levels continued to increase at 24 h (Fig. 8C). We also detected mRNA expression of the IL-33 co-receptor AcP and the expression in resting NK or NKT cells was ∼10-fold higher than ST2 (data not shown). Interestingly, as shown in Fig. 8(C), the AcP message did not appear to be regulated by IL-12 treatment. Finally, in spite of the significant levels of mRNA expression, we were nonetheless unable to detect surface expression of ST2 by flow cytometry, even after IL-12 treatment (Fig. 8D). Taken together, however, these results suggest that positive and rapid regulation of full-length ST2 expression by IL-12 is one mechanism that may contribute to the ability of IL-33 and IL-12 to synergistically induce IFN-γ.
In this study, we found that IL-33 is capable of promoting a variety of effector responses through its activity on human basophils, Th2 cells, iNKT and NK cells. Previously, Allakhverdi et al. (3) demonstrated that IL-33 is capable of potently activating human mast cells. Similar to this observation, we report here for the first time that basophils express abundant mRNA levels of the IL-33 receptor ST2 and that IL-33 specifically activates human basophils, a cell population that, like mast cells, may arise from CD34+ progenitor cells (23). The inability to detect ST2 surface expression was surprising; however, the lack of a correlation between detectable surface staining and functional responsiveness is not without precedent (3, M. S. Smithgall, M. R. Comeau and D. E. Smith, unpublished observations). Flow cytometry-based surveys may not always be sufficient to determine cytokine-responsive cells. IL-33 alone induced modest levels of multiple cytokines and chemokines whose production was synergistically increased in the presence of IL-3, the main growth and differentiation factor for basophils (19). Interestingly, we did not observe up-regulation of the already high levels of ST2 mRNA in basophils by IL-3 treatment (data not shown), suggesting that other mechanisms may underlie the observed synergy. Like mast cells, basophils have the ability to rapidly produce and release abundant quantities of cytokines, chemokines and preformed mediators and have been proposed to play a key role in late-phase asthma and other allergic conditions (24–28). The finding that IL-33 is a potent activator of both mast cells and basophils suggests that localization of these cells in tissues reported to express IL-33 mRNA, such as bronchial smooth muscle and airway epithelial cells (14, M. R. Comcau and D. E. Smith, personal observations), provides an opportunity for IL-33 to profoundly influence the innate and adaptive arms of the inflammatory response in an antigen-independent manner.
ST2 expression on Th2-polarized T cells is one of the distinguishing features of this receptor and IL-33 has previously been shown to enhance effector cytokine release from mouse Th2 cells. Utilizing an HDM-specific T cell culture, we demonstrated that IL-33 will also enhance IL-5 and IL-13 production from human Th2-skewed cells. Interestingly, IL-33 treatment enhanced TCR-driven, as well as antigen-independent T cell activation. Moreover, although the HDM-specific T cells were skewed toward a Th2 phenotype with IL-4, cultures from several donors also produced IFN-γ upon activation and this response was significantly enhanced by the addition of IL-33. This observation is consistent with the notion that human Th cells are rarely a pure Th1 or Th2 phenotype as commonly observed in mouse systems and may also relate to earlier studies that found ST2 expression associated with IL-5 and IFN-γ co-expressing lymphocytes (7, 10). Similarly, IL-18, considered a prototypical Th1-promoting cytokine, has been shown capable of promoting both Th1 and Th2 cytokine production from Th1-polarized T cells in the presence of TCR activation (29). Recently, it was reported that IL-33 is capable of acting as a chemoattractant for human Th2 cells (30), which taken together with our results would imply that IL-33 may be able to promote both the recruitment and activation of Th2 cells, even in the absence of antigen.
We also found that human iNKT cells express ST2 and will respond to IL-33 treatment by producing elevated levels of both type 1 and type 2 cytokines. Pre-clinical models of autoimmune or allergic diseases such as asthma have implicated a role for iNKT cells in mediating destructive inflammation (recently reviewed in 31) and this lymphocyte population may represent a component of steroid-insensitive inflammation that is commonly associated with chronic asthma (32, 33). Moreover, the ability of iNKT cells to respond to putative endogenous ligands provides an additional means by which type 2 effector cytokines are produced in the absence of atopy. Along the same lines, the ability of IL-33 to act directly on various cells in the absence of an antigen-dependent response could explain the impaired primary immune response observed in ST2-deficient mice as well as the ability of IL-33 to induce robust inflammation in naive mice (9, 14). In addition to enhancing IFN-γ production from Th2-skewed cells from some donors, we observed that IL-33 will cooperate with IL-12 to directly induce IFN-γ production from both iNKT and circulating NK cells. In fact, we found that IL-33 was at least as potent as IL-18 in promoting this prototypical Th1-oriented effector response. These in vitro findings are consistent with a recent in vivo study that found a role for IL-33 in promoting hypernociception via an IFN-γ-dependent mechanism (34). The ability of IL-12 to up-regulate the expression of ST2 mRNA in NK cells is similar to observations of IL-18R regulation (35) and likely contributes to the observed synergy with IL-33. Finally, although we did observe synergy between IL-33 and IL-23 in our NK cell assay, in separate experiments we have been unable to demonstrate an activating effect of IL-33 or IL-23 in a human memory T cell IL-17 response (data not shown).
Our results, together with ST2’s well-characterized Th2-restricted T cell expression, suggest that ST2 (and IL-33) may act primarily through a Th2-oriented pathway, but can under certain conditions also promote Th1-type responses. IL-33 may therefore represent a general amplifier of inflammation with the outcome dependent on the local cellular and extracellular context, such as, for example, the presence or absence of IL-12. The disregulated expression of sST2 in a number of human diseases including asthma, fibrosis and arthritis is consistent with the hypothesis that this potent inflammatory pathway requires regulation and, moreover, is not limited to type 2-mediated inflammation (36, 37). It is unclear how these broad biological effects of IL-33 manifest in the context of allergic inflammation or chronic diseases such as asthma. Both ST2–Fc and an anti-ST2 mAb have demonstrated efficacy in reducing allergen-induced lung inflammation in short-term models, suggesting that the IL-33 pathway can influence pulmonary inflammation (4, 5). It would also be informative to examine the effects of IL-33 inhibition in Th1-oriented models as well as more chronic models of lung inflammation that have been demonstrated to involve pathological contributions from both Th2 (IL-5, IL-13) and Th1 cytokines (TNF-α, IFN-γ) (38) and, perhaps, be more representative of chronic human asthma (39, 40). Interestingly, chronic exposure of mice to HDM led to persistent airway inflammation and the infiltration of ST2+ T cells that was followed by irreversible remodeling changes in the lung (41). IL-33 is expressed by epithelial and smooth muscle cells in the lung and is well positioned to act on infiltrating inflammatory cells (14, M.C. and D.S., personal observations). Moreover, we have measured elevated levels of both IL-33 and full-length ST2 mRNA in human asthmatic lungs (personal observations), which suggest that this pathway is a potential source of chronic pulmonary inflammation in asthma.
granulocyte macrophage colony-stimulating factor
house dust mite
hypoxanthine guanine phosphoribosyltransferase
tumor necrosis factor-α
We thank D. Meininger for technical assistance with recombinant protein production and D. Farrer, C. Thor, and S. Lear for superb isolation of human PBMC and lymphocytes.