IL-33 synergizes with TCR and IL-12 signaling to promote the effector function of CD8⁺ T cells

Abstract

The effector functions of CD8⁺ T cells are influenced by tissue inflammatory microenvironments. IL-33, a member of the IL-1 family, acts as a danger signal after its release during cell necrosis. The IL-33/ST2 axis has been implicated in various Th2 responses. Its role in CD8⁺ T cell-mediated immune response is, however, not known. Here we find that type 1 cytotoxic T (Tc1) cells cultured in vitro unexpectedly express high levels of the IL-33 receptor ST2. Interestingly, the expression of ST2 in Tc1 cells is dependent on T-bet, a master Th1/Tc1 transcription factor. In addition, IL-33 enhances TCR-triggered IFN-γ production. IL-33 together with IL-12 can stimulate IFN-γ production in Tc1 cells. Moreover, IL-33 synergizes with IL-12 to promote CD8⁺ T cell effector function. The synergistic effect of IL-33 and IL-12 is partly mediated by Gadd45b. Together, these in vitro data establish a novel role of IL-33 in promoting effector type 1 adaptive immune responses.

Introduction

Interleukin-33 is a member of the IL-1 family of cytokines which also includes IL-1 (α and β) and IL-18 [1]. IL-1β and IL-18 are expressed as prodomains containing polypeptide precursors which are proteolytically cleaved by caspase-1 to generate the active forms of these cytokines [2]. IL-33 is different from IL-1β and IL-18 in that it cannot be processed by caspase-1; instead, IL-33 is cleaved by caspase-7 and -3 during apoptosis to inactivate IL-33 function [3-5]. IL-33 is constitutively expressed in the nuclei of blood vessel endothelial cells, fibroblastic reticular cells of lymphoid tissues, and tissue cells exposed to the external environment such as skin keratinocytes and stomach epithelial cells [6]. During necrotic processes, full-length but biologically active IL-33 can be released [3, 5]. The fact that IL-33 may be released by necrotic cells during infection or trauma suggests it may serve as an endogenous danger signal or ‘alarmin’ [6].

Ample evidence supports an important role of IL-33 in Th2 cell-mediated immune responses [1]. The IL-33 receptor complex consists of ST2 and IL-1RAcP, both of which are members of the IL-1 receptor family [7-8]. ST2 is expressed by a number of cells involved in Th2 type responses such as Th2 cells [9-10], dendritic cells [11-12], mast cells [13-14], basophils [15-16], and eosinophils [17]. IL-33 enhances IL-5 and IL-13 production by Th2 cells independently of IL-4 [7, 18]. Administration of either an antibody against ST2 or recombinant ST2 fusion protein inhibits eosinophilic airway inflammation and induces resistance to Leishmania major infection in BALB/c mice [9-10]. In mice, IL-33 induces anaphylactic shock, in a T cell-independent, mast cell-dependent manner [13]. Interestingly, IL-33 induces IL-13-dependent cutaneous fibrosis [19]. In humans, the level of IL-33 is greatly increased in the blood of atopic patients during anaphylactic shock. Besides its expression on effector cells of Th2 immune responses, ST2 is also found on NK and NKT cells, which respond to IL-33 with increased IFN-γ production, suggesting a role for IL-33/ST2 in innate Th1 type immune responses [15, 20]. Whether IL-33 plays a role in adaptive Th1 type immune response is not known.

Here, we reveal that ST2 is highly expressed on Tc1 cells. Its expression on Tc1 cells is mainly dependent on T-bet, a master transcription regulator of Th1 and Tc1 cells [21]. We have further found that IL-33 synergizes with TCR, IL-12 signaling, or both to drive IFN-γ production in Tc1 cells and promote features of effector CD8⁺ T cells. Our study establishes a novel role of IL-33 in driving effector function of CD8⁺ T cells.

Results

IL-33 receptor ST2 is highly expressed in effector Tc1 cells

In order to understand the molecular characteristics of Tc1 cells, we performed gene profiling studies and found that IL-33 receptor ST2 was highly expressed in these cells (data not shown). To confirm our result, we performed a quantitative RT-PCR (qRT-PCR) analysis on naïve CD8⁺ T cells polarized in Tc0, Tc1, Tc2, and Tc17 conditions for 4 days. The ST2 mRNA was not expressed in naïve CD8⁺ T cells (data not shown) and could be induced in CD8⁺ T cells cultured in Tc0 and Tc17 conditions. The level of ST2 mRNA was four fold higher in Tc1 cells compared to Tc0 cells, confirming our microarray analysis (Fig 1A). Surprisingly, CD8⁺ T cells cultured in the Tc2 condition showed minimum ST2 expression compared to those cultured in other conditions. These results suggest that IL-12 further increases ST2 expression whereas IL-4 suppresses ST2 expression in CD8⁺ T cells at the mRNA level. The ST2 protein could also be detected on the surface of Tc1 cells by flow cytometry (Fig 1B). Additionally, we also found that TCR signaling further increased levels of ST2 mRNA in Tc1 cells (Fig 1C). In contrast, IL-1RAcP mRNA is expressed at similar levels among all subsets of effector CD8 T cells (supplementary Fig 1).

Figure 1. Expression of ST2 in effector CD8⁺ T cells.

(A and B) Naïve CD8⁺T cells were cultured in Tc0, Tc1, Tc2, and Tc17 conditions for 4 days. (A) Total RNA was made and subject to real time RT-PCR analysis of ST2 mRNA levels. RQ refers to relative quantity of ST2 mRNA, calculated by software provided by Applied Biosystem. RQ for Tc2 was set as 1. (B) Surface expression of ST2 was determined by flow cytometry. Isotype control is presented as the shaded area. (C) CD8⁺ T cells were cultured in Tc1 condition for 4 days (4d). These cells were either stimulated with anti-CD3 for 4h (4d + 4h), or cultured with the addition of fresh IL-2 (10 units/ml) for 3 more days (7d). At the end of 7 day culture, cells were extensive washed and cultured in the Tc1 polarizing condition for an additional 4 days (2° 4d). Total RNA was made from cells in each group and subjected to real time RT-PCR analysis of ST2 mRNA. RQ for Tc1 day 4 was set as 1. (D and E) Naïve CD8⁺ T cells were cultured in Tc1 or Tc2 polarizing conditions for 7 days. The cells were then subject to a second round of Tc1 or Tc2 polarization for 4 days. (D) Total RNA was made and analyzed by real time RT-PCR for levels of ST2 mRNA. RQ for Tc2 was set as 1. (E) Surface expression of ST2 was determined by flow cytometry. Isotype control is presented as the shaded area. Data in A, C and D are presented as the mean ±SEM of triplicate samples and are representative of three independent experiments. **P < 0.001, two-tailed unpaired Student’s t-test.

In order to examine whether ST2 expression on Tc1 cells is stable, we performed a second round of Tc1 polarization by stimulating Tc1 cells with anti-CD3 and anti-CD28 in Tc1 culture conditions for another 4 days. We found even higher levels of ST2 expression on CD8⁺ T cells following two rounds of Tc1 polarization (Fig 1C, D, E). In contrast, CD8⁺ T cells cultured in the Tc2 condition for two rounds expressed lower levels of ST2 mRNA and protein compared to those in the Tc1 condition (Fig 1C, D, E). In addition, we have found similar differences in ST2 expression between human CD8⁺ T cells cultured in Tc1 and Tc2 conditions (supplementary Fig 2). Finally, important differences in ST2 expression were also observed between effector CD4⁺ T cells (Supplementary Fig 3).

T-bet is required for ST2 expression in Tc1 cells

T-bet and Eomes are two Tc1 master transcription regulators expressed in effector/memory CD8⁺ T cells [21-22]. T-bet is highly expressed in effector CD8⁺ T cells and plays a predominant role in the function of effector CD8⁺ T cells [23-24]. In contrast, Eomes is highly expressed in both effector and memory CD8⁺ T cells and plays a critical role in the survival of memory CD8⁺ T cells [25]. Since ST2 is expressed in Tc1 cells, we decided to determine whether T-bet and/or Eomes regulate its expression. Naïve CD8⁺ T cells were isolated from WT, T-bet ^−/−, Eomes ^−/−, and T-bet/Eomes doubly deficient (DKO) mice and were polarized in vitro in Tc1 and Tc2 conditions. The levels of ST2 in cells cultured in these conditions were measured by qRT-PCR. ST2 mRNA was highly expressed in WT effector Tc1 cells (Fig 2A, B). Its level was unchanged in Eomes ^−/− effector Tc1 cells (Fig 2A). In contrast, ST2 mRNA was greatly reduced in T-bet ^−/− Tc1 cells and was also reduced in DKO Tc1 cells (Fig 2A). The surface expression of ST2 protein was also reduced in T-bet^−/− Tc1 cells compared to WT Tc1 cells (Fig 2B). These data suggest that T-bet is primarily required for ST2 expression in Tc1 cells and Eomes seems not involved in regulating ST2 mRNA. We have shown previously that DKO CD8⁺ T cells cultured in Tc2 conditions showed drastically increased Tc2 characteristics [26]. Because ST2 has been shown to be highly expressed in Th2 cell, we examined whether ST2 can be upregulated in DKO Tc2 cells. Surprisingly, ST2 expression was absent in DKO CD8⁺ T cells cultured in the Tc2 condition (Fig 2A). It has been shown that IFNγ produced by T cells cultured in Tc1 conditions is involved in driving Th1-specific gene expression [27]. However, when we neutralized the function of IFNγ in Tc1 culture, ST2 expression was unchanged (Fig 2C). Therefore, ST2 expression is not regulated by IFNγ in Tc1 cells.

Figure 2. Role of Tbet and Eomes in ST2 expression in Tc1 cells.

(A, B) Naive CD8⁺ T cells from C57BL/6WT (WT), T-bet ^−/− (TKO), Eomes ^−/−(EKO), and T-bet/Eomes doubly deficient (DKO) mice were cultured in Tc1 or Tc2 for 4 days. (A) Total RNA was made and subjected to real time RT-PCR analysis for ST2. (B) Surface ST2 expression was analyzed by flow cytometry. (C) CD8⁺ T cells cultured in Tc1 condition with or without anti-IFN-γ mAb (10μg/ml) for 4 days. Total RNA was made and subjected to real time RT-PCR analysis for ST2. Data in (A) and (C) are presented as the mean±SEM of triplicate samples. Results are representative of three independent experiments.

IL-33 synergizes with TCR signaling and/or IL-12 in promoting IFNγ production

We have shown that ST2 is expressed in Tc1 cells, and ST2 mRNA level is further increased in Tc1 cells upon TCR stimulation. These data suggest that IL-33 might be involved in the function of Tc1 cells. Since IFNγ is a hallmark cytokine for Tc1 cells, we investigated whether IL-33 synergizes with TCR signaling in the production of IFNγ by Tc1 cells. We first stimulated Tc1 cells with different dose of anti-CD3 for 16h in the presence or absence of IL-33. Addition of IL-33 toTc1 cells stimulated with 1 μg/ml and 0.5 μg/ml anti-CD3 mAbs, resulted in higher frequencies of IFNγ producers and greater levels of secreted IFNγ protein (Fig 3A and B). These data suggest that IL-33 synergizes with TCR signaling in the induction of IFNγ production in Tc1 cells.

Figure 3. IL-33 synergizes with TCR signaling or IL-12 to promote IFN-γ production.

Naive CD8⁺ T cells were cultured in Tc1 conditions for 4 days and were then stimulated with IL-33 and anti-CD3 for 16h alone or in combination. IFN-γ production was detected by (A) flow cytometry and (B) ELISA. Naive CD8⁺ T cells were cultured in Tc1 condition for 4 days and were subsequently stimulated with IL-33 or IL-12 alone or in combination for 24h. IFNγ production was measured by (C) flow cytometry or (D) ELISA. (E) Naive CD8⁺ T cells cultured in Tc1 condition for 4 days and were subsequently stimulated with IL-33 or IL-12 alone or in combination for 48h. Cell viability was determined by Trypan blue exclusion.

(F) Naive CD8⁺ T cells from Pmel-1 WT and Pmel-1 T-bet^−/− (TKO) mice cultured with APCs and the gp100 peptide in Tc1 polarizing condition for 4 days. The activated cells were then stimulated with IL-33 or IL-12 alone or in combination for 24h. WT Tc1 cells stimulated by anti-CD3 for 16h were used as a positive control. IFN-γ production was measured by ELISA.

ND, non-detectable. Data are presented as the mean ±SEM of triplicate samples and representative of three independent experiments. **, P < 0.01, two-tailed unpaired Student’s t-test.

IL-18, a member of the IL-1 family protein has been shown to synergize with IL-12 to stimulate IFNγ production in Th1 cells [28-29], we decided to determine whether IL-33 synergizes with IL-12 to induce IFNγ in Tc1 cells. Tc1 cells were cultured with IL-12, or IL-33 or IL-12 plus IL-33 for 24h. IL-12 or IL-33 alone induced very few IFNγ⁺ Tc1 cells (less than 1%) and low levels of IFNγ protein (Fig 3C and D). This result is consistent with prior publication and our own observation which showed non-detectable levels of IFNγ when Th1 cells were stimulated with IL-12 alone [29]. In contrast, combining IL-12 and IL-33 drastically increased the frequency of IFNγ⁺ Tc1 cells (to about 28%) (Fig 3C). IFNγ protein levels were also greatly increased when both IL-12 and IL-33 were added to the culture (Fig 3D). Besides IFNγ production, IL-12 and IL-33 also significantly increased the number of cultured Tc1 cells (Fig 3E). These data suggest that IL-12 and IL-33 synergize to drive effector function in Tc1 cells. Interestingly, when anti-CD3, IL-12, and IL-33 added together, they synergize in inducing IFNγ production (supplementary Fig 4). Thus, IL-33 synergizes with TCR signaling and IL-12 in promoting IFNγ production.

Since T-bet is required for ST2 expression in Tc1 cells (Figure 2), we determined whether this transcription factor is required for the IFNγ production driven by IL-12 plus IL-33. Naïve CD8⁺ T cells were isolated from WT Pmel-1 TCR transgenic mice [30] and T-bet KO Pmel-1 TCR transgenic mice and were stimulated either with cognate peptide plus antigen presenting cells (APCs) (Fig 3F) in Tc1 conditions for 4 days. These cells were incubated with IL-12, or IL-33, or IL-12 plus IL-33 for an additional 24 h. Supernatants were analyzed for levels of IFNγ by ELISA. We observed that IL-12 plus IL-33 induced high amounts of IFNγ in WT Pmel-1 Tc1 cells. In contrast, no IFNγ was produced in T-bet KO Pmel-1 Tc1 cells cultured with IL-12 plus IL-33 (Figure 3F). These results demonstrate that Tc1 cells generated via stimulation by a cognate peptide plus APCs are similarly subject to synergistically effect of IL12 plus IL-33, and T-bet is required for IFNγ production driven by IL-12 plus IL-33, likely via regulation of ST2. Similar results were obtained using Tc1 cells stimulated with plate bound anti-CD3 and anti-CD28 without APCs (data not shown). In addition, the synergy between anti-CD3 and IL-33 was also dependent on T-bet, and IL-33 could not further increase IFNγ production induced by anti-CD3 in T-bet ^−/− CD8 T cells (supplementary Fig 5).

IL-33 promotes effector signatures but inhibits characteristics of resting CD8⁺ T cells

We have shown that IL-12 and IL-33 induced IFNγ production in Tc1 cells. Thus, we further investigated whether IL-12 and IL-33 regulate IFNγ gene expression. We performed real time RT-PCR analysis on IFNγ mRNA. We found that IL-33 or IL-12 alone induced small amounts of IFNγ mRNA. IL-12 plus IL-33 induced a large amount of IFNγ mRNA in Tc1 cells (Fig 4A). Therefore, IL-12 and IL-33 regulate IFNγ at the mRNA level.

Figure 4.

IL-33 and IL-12 synergistically drive the effector fate of Tc1 cells.

Naive CD8⁺ T cells cultured in Th1 condition for 4 days were then stimulated with IL-33 or IL-12 alone or combined together for (A-D) 4h or (E-G) 24h. mRNAs for (A) IFNγ, (B) T-bet, (C) Blimp1, (D) Eomes, (E) Lef1, (F) TCF-1, and (G) IL7R were quantified by Real-time PCR. Data are presented as mean ±SEM of triplicate samples and are representative of three independent experiments. *P<0.05, **P<0.01, ***P< 0.001, two-tailed unpaired Student’s t-test.

Since T-bet and Eomes are involved in IFNγ production in Tc1 cells, we also examined the levels of T-bet and Eomes mRNAs in Tc1 cells cultured in IL-12, or IL-33, or IL-12 plus IL-33 conditions. IL-12 and IL-33 individually did not further induce significant amounts of T-bet mRNA in Tc1 cells. However, IL-12 plus IL-33 induced much greater levels of T-bet mRNA (Fig 4B). In contrast, both IL-12 and IL-33 modestly diminished Eomes mRNA (Fig 4D). Together, IL-12 and IL-33 further reduced the level of Eomes mRNA (Fig 4D). Therefore, IL-12 and IL-33 differentially regulate T-bet and Eomes in Tc1 cells. Besides T-bet and Eomes, Runx3 has also been implicated in regulating levels of IFNγ [23]. However, IL-12 and IL-33 did not seem to significantly regulate mRNA levels of Runx2 nor Runx3 in Tc1 cells (data not shown).

Recent data showed that Blimp1 is a master regulator of the terminal differentiation of CD8⁺ effector T cells [31-33]. In addition, we have found also T-bet and Eomes are involved in Blimp1 expression in CD8⁺ T cells (supplementary Fig 6). Because IL-12 and IL-33 synergistically induce T-bet, we therefore examined whether IL-12 and IL-33 induced Blimp1 in Tc1 cells. IL-33, but not IL-12, induced small amounts of Blimp1. IL-12 and IL-33 together, greatly elevated levels of Blimp1 (Fig 4C). It is worth noting that IL-12 and IL-33 don’t increase levels of granzyme B or perforin (supplementary Fig 7). Thus, IL-12 and IL-33 together promote the effector function of Tc1 cells by inducing IFNγ, T-bet, and Blimp1.

TCF-1 and LEF-1 are known transcription factors mediating WNT signals. We have shown that TCF-1 and LEF-1 are associated with naïve and resting activated T cells and are drastically down-regulated upon TCR stimulation in both naïve and effector CD4⁺ T cells [34]. Recently, TCF-1 was shown to be associated with central memory T cells and required for the maintenance of these cells via its regulation of Eomes [35]. IL-33, or IL-12, or IL-12 plus IL-33 down-regulated both TCF-1 and LEF-1 mRNAs in Tc1 cells (Fig 4E and F). TCF-1 and LEF-1 down-regulation may contribute to the reduction of Eomes mRNA when Tc1 cells were cultured with IL-12 and IL-33.

IL7 is required for the survival of long term memory T cells [36]. IL7R can also identify a population of effector T cells that have a substantially greater potential to form memory CD8⁺ T cells [37]. We thus examined IL7R mRNA levels in Tc1 cells cultured with or without IL-33 and IL-12. Both IL-12 and IL-33 individually induced down-regulation of IL7R (Fig 4F). Together, they further diminished levels of IL7R mRNA and protein (Fig 4G, supplementary Fig 8). Collectively, these data support the idea that, IL-12 and IL-33 synergistically drive characteristics of effector Tc1 cells.

Gadd45b mediates IL12/IL-33-driven IFNγ production in Tc1 cells

We and others have previously shown that Gadd45b regulates activities of the p38 MAP kinase in Tc1 cells and mediated the synergistic effect of IL12/TCR as well as IL12/IL18 in Th1 cells [38-40], thus, we investigated whether Gadd45b was involved in IFNγ production driven by IL-33 plus IL-12 in Tc1 cells. We performed real time PCR analysis and found that IL-33 induced Gadd45b expression and IL-12 plus IL-33 further increased Gadd45b mRNA levels in Tc1 cells (Fig 5A). Gadd45b was partially involved in regulation of ST2, and 2 fold reduced levels of ST2 mRNA were observed in Gadd45b ^−/− Tc1 cells when compared to WT Tc1 cells (Supplementary Fig 9). In WT Tc1 cells, IL-12 and IL-33 can further enhanced p38 activation (Fig 5B and C). In contrast, Gadd45b deficiency reduced levels of detectable active p38 in Tc1 cells (Fig 5B and C). IL-12 and IL-33 failed to further increase levels of active p38 in Gadd45b deficient Tc1 cells (Fig 5B and C). In addition, Gadd45b deficiency also dampened IL-12/IL-33-induced IFNγ mRNA (Fig 5D) and protein levels in Tc1 cells (Fig 5E) as well as anti-CD3/IL-33-induced IFNγ mRNA (Supplementary Fig 10). Consistent with a role of the p38 MAP kinase in IL-12/IL-33 stimulated IFNγ production, addition of 10 μM of p38 inhibitor almost completely abolished IFNγ production (Fig 5F). The cell viabilities were not significant different in these culture conditions. Collectively these data demonstrate an important role of the Gadd45b/p38 axis in mediating IL-12/IL-33 synergistic effect on IFNγ production in Tc1 cells.

Figure 5. Role of Gadd45b in IL-33/IL-12-stimulated IFN-γ production in Tc1 cells.

(A) Naive CD8⁺ T cells were purified from WT mice and cultured in Tc1 condition for 4 days. Then these cells were stimulated with IL-33, or IL-12, IL-12 plus IL-33 for 4h. Gadd45b mRNA levels were measured by Real-time PCR.

(B and C) Naive CD8⁺ T cells were purified from WT and Gadd45b deficient mice and cultured in Tc1 condition for 4 days. Cells were then stimulated with IL-33 plus IL-12 for 15 minutes, 30 minutes and 60 minutes. Phosphorylated (phospho)-P38 (pP38) level was measured by (B) Western blot and (C) the mean relative intensity of pP38 was calculated.

(D and E) TC1 cells generated as in (B) were stimulated with or without IL-33 or IL-12 for 24h. (D) IFN-γ mRNA was detected by Real-time PCR, and (E) IFN-γ protein was measured by ELISA. (F) Naive CD8⁺ T cells were cultured in Tc1 conditions for 4 days. These cells were then cultured with IL-12 and IL-33 in the presence or absence of P38 inhibitor (50 μM, 10 μM, 1 μM) for 24 h; IFN-γ production was detected by ELISA. ND, undetectable. KO, Gadd45b ^−/−.

Data are presented as mean ±SEM of triplicate samples and are representative of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-test.

Discussion

The IL-33 receptor ST2 was originally reported as a Th2 marker that was expressed on Th2 cell lines but not Th1 cell lines [9-10]. Administration of either a mAb against ST2 or ST2 fusion protein greatly inhibited the induction of a lung mucosal Th2 immune response [10]. Recently, NK cells and NKT cells were shown to express ST2 and respond to IL-33. In addition, IL-12 and IL-33 stimulate IFNγ production in NK cells and NKT cells. Therefore, the innate arms of Th1 immune response are also driven by IL-33. Consistent with this idea, NKT cells are important for suppressing IL-33-driven allergic responses [15, 20, 41]. Our study showed that ST2 is expressed at high levels in CD8⁺ T cells cultured in Tc1 polarizing conditions. Strikingly, CD8⁺ T cells cultured in Tc2 conditions did not express ST2. Furthermore, ST2 expression in Tc1 cells is regulated by T-bet, a master transcription regulator of Th1 effector functions. Therefore, our data reveal a new role of IL-33/ST2 axis in promoting adaptive effector Tc1 cell-mediated immune responses. IL-33 was originally named as NF-HEV (nuclear factor from high endothelial venules), as it was known to interact with nuclear chromatin [42]. Similar to IL-1a and HMGB1, IL-33 protein is mainly expressed in cell nuclei, and is not normally secreted but can be released during cellular necrosis [3, 5]. Therefore, IL-33 is considered as a novel alarmin, an endogenous ‘danger’ signal to alert the immune system upon tissue damage during trauma or infection [6]. By itself, IL-33 induces a small amount of IFNγ. However, IL-33 synergizes with IL-12 and TCR signaling to drive large amounts of IFNγ production. Therefore, the biological function of IL-33 is dependent on other cytokines such as IL-12, and thus IL-33 acts to amplify inflammatory responses rather than to determine the nature of the inflammation.

It has been reported that IL-12, a signature cytokine of cell-mediated immune responses, inhibits Eomes while upregulating T-bet in effector CD8⁺ T cells during the peak of infection with Listeria monocytogenes. After resolution of infection, Eomes levels rise whereas T-bet expression declines in resting memory CD8⁺ T cells [24]. Therefore T-bet plays a dominant role in effector CD8⁺ T cells and Eomes is more important for homeostasis of memory CD8⁺ T cells. Our data showed that IL-12 and IL-33 synergistically increased T-bet expression whereas it suppressed levels of Eomes. In addition, IL-12 and IL-33 synergistically increased Blimp1 expression, another key transcription factor critical for effector fate of CD8⁺ T cells. In order to address whether IL-33 is involved in Tc1 differentiation, we cultured naïve CD8⁺ T cells in the presence of IL-33 and compared to CD8⁺ T cells cultured in the neutral condition for four days. We then washed the cultured cells extensively and subsequently re-stimulated these T cells with anti-CD3 or PMA/ionomycin without IL-33. We have found no difference in IFNγ production between cells from these cultures (data not shown). Therefore, IL-33 alone does not seem to promote early Tc1 differentiation. This is likely due to the fact that ST2 is induced around 72h to 96h after start of the Tc1 or Tc0 culture and is express at negligible levels during earlier T cell differentiation (data not shown). Collectively, these data are consistent with the idea that IL-33 promotes the effector characteristics of Tc1 cells directly.

It is intriguing that we observed lower levels of ST2 expression in CD8⁺ T cells cultured in the Tc2 condition. In humans, ST2 has been shown to be expressed on the Tc2 cell lines but not on Tc1 cell lines [43]. It is very likely that the difference is due to the culture conditions. We used plate bound anti-CD3 and anti-CD28 plus IL-12 and anti-IFNγ to activate and differentiate Tc1 cells. In Chan’s paper, allogeneic APCs were used to stimulate human T cells. In addition, we stimulated T cells for 4 days. In Chan’s paper, they stimulated CD8⁺ T cells for many rounds. It is possible that ST2 levels increase more after several rounds of T cell stimulation. Consistent with this idea, it has recently been reported that ST2 is only highly expressed in Th2 cells after three rounds of stimulation [44]. Similarly, we have found that ST2 expression was quite low in Th2 cells polarized for one round (supplementary Fig 3). The reason Chan et al did not observe ST2 expression on Tc1 cells might have been due to the way their Tc1 lines were generated. Multiple rounds of stimulation might lead to outgrowth of Tc1 lines that are different from our Tc1 lines due to massive apoptosis of Tc1 cells.

Our study demonstrates that IL-33 can promote Tc1 immune response in vitro. In vivo, potentially through acting on Th2 cells or Th2-supporting cells, such as mast cells and basophils, IL-33 plays an important role in Th2 responses against pathogens such as Pneumocystis murina, nematode Trichuris muris, Toxoplasma gondii, and respiratory syncytial virus [45-48]. Upon L. major infection mice treated with anti-ST2L antibody had enhanced Th1 responses and developed significantly smaller lesions compared with mice treated with control IgG. ST2 deficient mice show a normal host defense against lung infection with Mycobacterium tuberculosis [49]. By contrast, based on our study, it is possible that IL-33 is required for Tc1-driven immune responses against virus, intracellular bacteria, and tumors. In addition, IL-33 has the potential to be used as an adjuvant in vaccines to boost Tc1 immune responses. Our future studies will focus on the in vivo role of IL-33 in various Tc1-mediated diseases.

Materials and Methods

Mice

CD4-cre Eomes fl/fl /Tbet doubly deficient mice and CD4-cre Eomes fl/fl mice were described [26]. Pmel-1 TCR transgenic mice were purchased from the Jackson laboratory. All animals were maintained under specific pathogen-free conditions. All animal work has been approved by the Institution Animal Care and Use Committee at University of Pittsburgh.

CD8⁺ T cells culture

Lymphocytes were collected from spleens and lymph nodes obtained from C57BL/6WT, T-bet ^−/− (TKO), Eomes ^−/−(EKO), T-bet/Eomes doubly deficient (DKO) mice. Naïve CD62L⁺ CD44⁻ CD8⁺ T cells were purified by FACS or magnetic beads based methods. The naïve CD8 T cells are more than 98% and cultured in Tc1, Tc0, Tc2, and Tc17 conditions as indicated. Cells were stimulated with 5 μg/ml plate-bound anti-CD3 (clone 145-2C11) and 5 μg/ml plate-bound anti-CD28 mAbs (clone 37.51) in complete RPMI (cRPMI, RPMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 50 μM 2-ME, 100 U/ml penicillin, 100 μg/ml streptomycin) in the presence of huIL-2 (20 U/ml, obtained from the BRB Preclinical Repository), IL-12 (3.4 ng/ml) plus anti–IL-4 (10 μg/ml, clone 11B11, from the BRB Preclinical Repository) for the Tc1 condition, or huIL-2 (20 U/ml) for the Tc0 condition, or huIL-2 (20 U/ml), IL-4 (2 ng/ml), and anti-IFN-γ (10 μg/ml, clone XMG 1.2) for Tc2, or anti-IL-2Rα (10 μg/ml, clone PC61, ATCC), IL-23(10 ng/ml), IL-6 (10 ng/ml), TGF-β1 (1 ng/ml), anti–IFN-γ (10 μg/ml, clone XMG 1.2), and anti–IL-4 (10 μg/ml, clone 11B11) for the Tc17 condition. After 48h, cells were re-plated to new wells without anti-CD3 and anti-CD28 and with freshly added IL-2 (20 U/ml) for another 2 days or 5 days. Alternatively, naive CD8⁺ T cells were cultured with anti-CD3 in the presence of T cell–depleted cell cycle–arrested splenocytes as antigen-presenting cells. Naive Pmel-1 TCR transgenic CD8⁺ T cells were cultured with 1 μM gp100_25-33 in the presence of antigen presenting cells for 4 days in various polarizing conditions as mentioned above.

Upon being cultured in various polarization conditions, CD8⁺ T cells were washed once with cRPMI and were subsequently stimulated for various time points with plate-bound anti-CD3 with or without IL-33 (10ng/ml, Peprotech), or in the presence of IL-12 (3.4 ng/ml) with or without IL-33 (10 ng/ml). For the treatment with inhibitors, cells were cultured with P38 inhibitor (Enzo Life Sciences), MEK1 inhibitor PD 098059 (Calbiochem), NFκB inhibitor (Sigma) and JNK inhibitor (Calbiochem).

Antibodies

For flow cytometry, anti-CD4 (GK1.5), anti-CD8 (53-6.7), and anti-IFN-γ Ab (clone XMG1.2) were all purchased from eBioscience (San Diego, CA), and anti-ST2 (B4E6) was from MD Bioproducts. Flow cytometric analysis was performed using a FACS flow cytometer (BD Biosciences, San Jose, CA). For Western blot, anti-phosphorylated (phospho)-p38, anti-phospho-JNK, and anti-phospho-ERK were obtained from Cell Signaling.

Real-time PCR

Cells were lysed in Trizol (Invitrogen) and total RNA was extracted following manufacturer’s instructions. RNA was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems).The real-time PCR was performed on with SYBR green kit (Applied Biosystems). Cycling conditions were 10 min at 95°C, followed by 40 repeats of 95°C for 15 s and 60°C for 60 s. The primers were: T-bet sense, 5′-CGGGAGAACTTTGAGTCCATGT-3′,Tbet antisense,5′-GCTGGCCTGGAA GGTCG-3′; Eomes sense, 5′-GGCCTACCAAAACACGGATA-3′,Eomes antisense,5′-GACCTCCAGGGACAATCTGA-3′; ST2 sense, 5′-CAAGTAGGACCTGTGTGCCC-3′ST2 antisense,5′-CGTGTCCAACAATTGACCTG-3′; Lef-1 sense,5′-AGAAATGAGAGCGAATGTCGTAG-3′,Lef-1 antisense,5′-TTTGCACGTTGGGAAGGA-3′;Tcf-1 sense,5′-AGCTTTCTCCACTCTACGAACA-3′,Tcf-1 antisense,5′-AATCCAGAGAGATCGGGGGTC-3′;Runx2 sense,5′-ATGCTTCATTCGCCTCACAAA-3′,Runx2 antisense, 5′-GCACTCACTGACTCGGTTGG-3′;Runx3 sense,5′-GGTCACCACCGTTCCATC-3′,Runx3 antisense,5′-ACTTCCTCTGCTCCGTGCT-3′;IL7R sense,5′-TATGTGGGGCTCTTTTACGAGT-3′,IL7R antisense,5′-GCCTCGGCTTTAACTATTGTGT-3′; Gadd45b sense,5′-CAGATTCACTTCACCCTGATCC-3′,Gadd45b antisense,5′-GTTGTGCCCAATGTCTCCG-3′;Blimp1 sense,5′-CATGGAGGACGCTGATATGAC-3′ Blimp1 antisense,5′-ATGCCTCGGCTTGAACAGAAG-3′;IFNγ sense,5′-TCAAGTGGCATAGATGTGGAAGAA-3′,IFNγ antisense,5′-TGGCTCTGCAGGATTTTCATG-3′

Statistical analysis

We used the two-tailed unpaired Student’s t-test. We considered p values less than 0.05 as being significant.

Abbreviations used

T/E DKO

T-bet/Eomes double knockout

EKO

Eomes^−/−

TKO

T-bet^−/−

WT

wild-type