Subthreshold IKK activation modulates the effector functions of primary mast cells and allows specific targeting of transformed mast cells.

Mast cell differentiation and proliferation depends on IL-3. IL-3 induces the activation of MAP-kinases and STATs and consequently induces proliferation and survival. Dysregulation of IL-3 signaling pathways also contribute to inflammation and tumorigenesis. We show here that IL-3 induces a SFK- and Ca2+-dependent activation of the inhibitor of κB kinases 2 (IKK2) which results in mast cell proliferation and survival but does not induce IκBα-degradation and NFκB activation. Therefore we propose the term “subthreshold IKK activation”. This subthreshold IKK activation also primes mast cells for enhanced responsiveness to IL-33R signaling. Consequently, co-stimulation with IL-3 and IL-33 increases IKK activation and massively enhances cytokine production induced by IL-33. We further reveal that in neoplastic mast cells expressing constitutively active Ras, subthreshold IKK activation is associated with uncontrolled proliferation. Consequently, pharmacological IKK inhibition reduces tumor growth selectively by inducing apoptosis in vivo. Together, subthreshold IKK activation is crucial to mediate the full IL-33-induced effector functions in primary mast cells and to mediate uncontrolled proliferation of neoplastic mast cells. Thus, IKK2 is a new molecularly defined target structure.


INTRODUCTION
Mast cells are located in peripheral tissues and regulate innate and adaptive immune responses [1] by producing mediators (e.g., histamine, proteases, leukotrienes or cytokines) that recruit and activate, granulocytes, dendritic cells, T-lymphocytes and other cells [1][2][3][4][5]. They are critical in type I hypersensitivity and therefore central to the pathogenesis of allergic diseases [6]. The participation of mast cells in the pathogenesis of autoimmune diseases has been reported [7] and refuted [8]. Mast cells can produce pathogenetically relevant cytokines such as IL-1β, IL-6, IL-13, IL-17 and TNFα [9]. It has also been demonstrated that mast cells are critical regulators of the tumor microenvironment [10] and that expression of constitutively active Ras-or c-Kit-mutants leads to development of mast cell tumors [11,12].
The most extensively characterized trigger for mast cell activation is crosslinking of the Fcε-receptor-I (FcεRI) resulting in the release of mediators including histamine, proteases, cytokines, and chemokines [13].
Recent publications reported a crosstalk between FcεRI either with TLRs (including TLR4) or c-Kit resulting in increased NFκB-, NFAT-or JNK activation and cytokine production [25,26]. We found an crosstalk between activated c-Kit and the IL-1R or the IL-33R [27,28] resulting in potentiated cytokine production in response to IL-1 or IL-33. Here, we identify a novel mechanism by which IL-3 induces IKK activation which mediates mitogenic signaling in primary-and tumor mast cells and modulates the full biological response of IL-33.

IL-3-induced IKK activation in BMMC
IL-3 induced IKK activation results in anti-apoptotic signaling in hepatocytes [29]. We found that IL-3 also induces IKK activation, IκBα phosphorylation but not degradation in BMMCs ( Figure 1A). To exclude defective IKK-IκBα signaling we used IL-33, a known activator of IKK-dependent IκBα degradation. In contrast to IL-3, IL-33 induced IκBα degradation and cytokine response in BMMCs which was blocked by treatment with the IKKinhibitor VII ( Figure 1B and 1C), the most efficient IKK inhibitor tested (Supplementary Figure S1A-D).
Next we used NFκB-EGFP-MC/9 [30] mast cells to examine whether IL-3 induces NFκB activation. Confirming the results in BMMCs, IL-3 did not induce IκBα degradation ( Figure 1D), EGFP expression ( Figure  1E) or an effective cytokine-production ( Figure 1F). Again, IL-33 elicited all these effects in an IKK-dependent manner ( Figure 1D-1F). Thus, IL-3 does not induce the canonical NFκB signaling. Why does the IL-3-induced IKK activation fail to activate NFκB? When cells were cultured with cycloheximide, an inhibitor of protein biosynthesis, IL-3 induced a mild but detectable IκBα degradation (Supplementary Figure S1E). This finding indicates that IL-3 can induce some IκBα degradation. In presence of ongoing IκBα re-synthesis, this IL-3induced IκBα degradation is quantitatively not sufficient to result a net loss of IκBα. Therefore, we propose the term "subthreshold activation" to describe IKK activation which suffices to phosphorylate IκBα but is insufficient to trigger NFκB activation.

Subthreshold IKK2 activation is critical for IL-3-induced proliferation
Having shown that the IL-3-induced subthreshold IKK activation does not result in efficient cytokine production, we asked if IKKs mediate the activation of the mitogenic JNK signaling [24]. Indeed, the IKKinhibitor VII potently impaired the IL-3-induced activation of JNKs and the phosphorylation of IκBα (Figure 2A). Consequently, the IKK-inhibitor VII blocked the IL-3induced proliferation ( Figure 2B) and BMMC expansion ( Figure 2C) without inducing cell death ( Figure 2D). When IKK-inhibitor VII-treated BMMC were washed and re-stimulated with IL-3, their proliferative capacity was restored ( Figure 2E).
To confirm the results obtained by pharmacological inhibition of IKKs, we induced IKK2-deficiency by injection of Ikk2 ∆ -mice with poly(I:C) [31]. Consequently, the IL-3-induced proliferation and JNK activation was reduced in Ikk2 ∆ -BMMC compared to Ikk2 F/F -BMMCs ( Figure 2F and 2G).

Subthreshold IKK activation is SFK-dependent and primes mast cells for NFκB-dependent effector functions
Next we investigated which pathway mediates subthreshold IKK activation. Given that the Malt/Bcl10complex [32] and MyD88 (data not shown) are not involved www.impactjournals.com/oncotarget  we examined SFKs, critical for IKK2 activation and for mitogenic signaling [33][34][35][36][37]. The SFK inhibitor SU6656 blocked the IL-3-induced proliferation and inhibited the IL-3-induced JNK activation and IκBα phosphorylation ( Figure  3F and 3G). In contrast, the IL-33-induced IKK activation was not affected by SU6656 (Supplementary Figure S3A) indicating that the SFK-dependent IKK activation is specific for the IL-3-induced signaling.
SCF potentiates the IL-33-induced cytokine response in BMMCs [27]. Hence, we tested whether the IL-3induced subthreshold IKK activation primes BMMCs for stronger NFκB activation upon IL-33R-signaling. Indeed,  (C, D) BMMCs were single stimulated with IL-33 or IL-33 in combination with IL-3. Supernatants were collected and analyzed for TNFα (C) (p < 0,001) or IL-13 (D) (p < 0,001). (E, F) BMMCs were pre-treated with the IKK-inhibitor VII (E) or SU6656 (F). Cells were single stimulated with IL-33 or IL-33 in combination with IL-3. Collected supernatants were analyzed for IL-6 (E, F; p < 0,001). (G) Wt or Jnk1 -/-BMMCs were single stimulated with IL-33 or IL-33 in combination with IL-3. Supernatants were collected and analyzed for IL-6. co-stimulation with IL-3 and IL-33 increased the IκBα phosphorylation, accelerated its degradation ( Figure 4A) and potentiated the IL-6 mRNA production (Supplementary Figure S3B) compared to IL-33 alone. Consequently, IL-6 production after co-stimulation was much stronger than in response to IL-33 alone ( Figure 4B). Notably, TNFα and IL-13 were only produced when BMMCs were co-stimulated with IL-3 and IL-33 but were not detectable upon stimulation with IL-33 alone ( Figure 4C and 4D). Confirming the priming effect of IL-3, the full potentiated cytokine response was only detectable, when cells were first stimulated with IL-3 followed by exposure to IL-33. Pre-stimulation with IL-33 or simultaneous stimulation with IL-3 and IL-33 induced only a partial co-stimulatory effect (Supplementary Figure S3C).
Ca 2+ mobilization induces NFAT activation. Thus, we tested whether stimulation with IL-33 alone or in combination with IL-3 induces NFAT activation. Neither IL-33 nor IL-33 in combination with IL-3 induced NFAT activation in NFAT-EGFP-MC/9 cells compared to ionomycin (Supplementary Figure S4D-F). These data show that Ca 2+ is critical for the induced cytokine production independently of NFAT.
The IL-3-induced Ca 2+ mobilization depends on PLCγ [38]. In contrast, IL-33 does not induce PLCγ activation (data not shown) but was reported to induce a PLD1dependent Ca 2+ mobilization resulting in NFκB activation [42]. Neither treatment of NFκB-EGFP-MC/9 cells with the PLD1 inhibitor CAY10594 (Supplementary Figure  S4H) nor PLD1 deficiency ( Figure 5H) influenced NFκB activation or cytokine production. As expected the PLCγ inhibitor U-73122 did not affect the IL-33-induced cytokine response but blocked the potentiated cytokine response induced by co-stimulation with IL-3 and IL-33 ( Figure 5I).
These data indicate that the IL-3-induced PLCγ activation and the resulting Ca 2+ mobilization are crucial for mast cell priming and the resulting potentiated cytokine production.

Subthreshold IKK activation mediates survival of tumor mast cells
In primary mast cells, IKK inactivation reduced the IL-3-induced mitogenic signaling. IL-3, c-Kit and IKKs are involved in pathogenesis of a number of malignancies [43][44][45]. Therefore, we determined the activation status of IKKs and their relevance for survival and proliferation in tumor mast cells. The v-HA-Ras-transformed V2D1 murine tumor mast cells constitutively produces IL-3, resulting in autocrine stimulation, proliferation and survival [46]. In these cells we found an increased JNK activation and IκBα phosphorylation but no degradation excluding the production of cytokines that activate NFκB ( Figure 6A).
The IKK-inhibitor VII blocked the activation of JNKs, the phosphorylation of IκBα ( Figure 6B) and consequently reduced cell proliferation ( Figure 6C) by inducing cell death ( Figure 6D). This demonstrates the relevance of IKKs in V2D1 cells. Given that survival of V2D1 cells depends on IL-3 production we speculated that IKKs are involved in IL-3 production and that exogenous IL-3 rescues V2D1 cells from cell death induced by IKK inhibition. As shown in Figure 6E IL-3 production is reduced by IKK inhibition. Moreover, the IKK-inhibitor VII-induced cell death can be reversed by exogenous IL-3 ( Figure 6F). These data show that subthreshold IKK activation is important for IL-3 production to mediate mitogenic signaling in V2D1 cells. www.impactjournals.com/oncotarget

Inhibition of IKKs reduces tumor growth in vivo
To analyze tumor growth in vivo, we used DBA/1-Rag1 -/--mice ( Figure 7A). We injected 1 x 10 6 V2D1 mast cells subcutaneously. After 7 days, a 25 μM IKK-inhibitor VII solution or vehicle was injected intratumorally for 6 weeks. As shown in Figure 7B and 7C tumor size was significantly decreased in mice treated with the IKKinhibitor. Cells obtained from an explanted tumor still expressed c-Kit, IL-3Rα, and IL-33R ( Figure 7D) and showed IKK-dependent IL-3 production ( Figure 7E) and proliferation ( Figure 7F). These data verify that the explanted cancer cells are still dependent on IL-3 production for autocrine stimulation and survival.

DISCUSSION
We identified IKKs as important for IL-3-induced mitogenic signaling in BMMCs. Hitherto, IKK2 has been known as an important component in the signaling pathways emanating from receptors such as antigen receptors, TIR-and TNFR-superfamily family members [32,47]. In all of these cases, IKK2 activation results in IκBα degradation and NFκB activation [48]. In contrast to the canonical pathway, we found that IL-3 induced only a weak and transient subthreshold IKK activation, which resulted in IκBα phosphorylation without IκBα degradation and without NFκB activation.
The major questions are (i) how does IL-3 induce subthreshold IKK2 activation?; and (ii) why is there no IκBα degradation and NFκB activation? Neither MyD88-nor Malt1-or Bcl10-deficiency influence the IL-3-induced mast cell proliferation [32]. Instead, our data indicate a critical role of SFKs and Ca 2+ for the IL-3-induced subthreshold IKK activation. We hypothesize that the combined activation of SFKs, PLCs and a Ca 2+dependent PKC-isoform mediates subthreshold IKK2 activation.
These data indicate that components (e.g., the MyD88-IRAK-TAK1-signaling module) critical for effective IκBα phosphorylation and degradation are not activated. Therefore the IL-3-induced IKK activation is only sufficient to induce mitogenic signaling, but is below the threshold to induce IκBα degradation and NFκB activation. Although the reason for the missing IκBα degradation is still unknown we suggest that IκBα degradation occurs to some extent upon IL-3 stimulation but is quantitatively not sufficient to induce a net loss of IκBα in the presence of ongoing IκBα re-synthesis.
Additionally, the IL-3-induced subthreshold IKK activation primes mast cells for enhanced NFκB activation in response to IL-33. This shows that mast cells integrate signals from different receptors which activate IKKs. Thereby, both, the IL-3-induced and SFK-mediated, and the IL-33-induced but MyD88-IRAK-TAK1-dependent pathways are crucial to facilitate full IKK2 activity. Therefore the signal strength determines the effector functions resulting from IKK activation. Weak, subthreshold, IKK activation as induced by IL-3 suffices to induce proliferation but not NFκB activation. Stronger IKK activation as provided by IL-33 signaling results in NFκB activation and production of cytokines. Combined signaling via IL-3 and IL-33 results in proliferation and strongly enhanced cytokine production. Under homeostatic conditions IL-3 serves as growth and survival factor for mast cells. Under pathological conditions (tissue damage) the presence of the alarmin IL-33 dramatically alters the mast cells' response to IL-3 which now becomes much more pro-inflammatory (IL-6, TNFα, IL-13).
In tumor cells, dysregulated expression of IL-3 is the molecular basis for survival and proliferation [11,43,46,49] leading to malignancies, including mast cell leukemia (MCL), acute myeloid leukemia (AML), and chronic myeloid leukemia (CML). In such cells subthreshold IKK activation might be a prerequisite for permanent proliferation. Indeed, pharmacological inhibition of subthreshold IKK activation specifically reduces tumor growth in vivo. Since IKK inhibition induced cell death in tumor-but not in primary-cells, IKK-inhibition could be a powerful approach to specifically eliminate certain tumor cells. In summary, our data indicate that the IL-3-induced subthreshold IKK activation is an important mechanism that mediates pathological effector functions in inflamed tissues after infection, allergy, necrosis and supports survival of tumor cells. Therefore, inhibition of subthreshold IKK activation might be a potential and selective tool to treat such mast cell driven diseases.

q PCR
Cells stimulated with IL-3 and/or IL-33 were pelleted and lysed with TRIzol (life technologies). RNA was extracted according to the manufacturer's protocol. Total RNA (2 μg) was reverse transcripted using oligo(dT)-primers, the M-MLV reverse transcriptase (affymetrix), RNase inhibitor (Promega) and the PCR thermocycler (Biometra). For the quantitative IL-6 real-time PCR the 5′ primer TCTCTGCAAGAGACTTCCATCCAGT and the 3′ primer AGCCTCCGACTTGTGAAGTGGT were used. The quantitative real-time PCR was performed with the KAPA SYBR fast kit (peqlab) according to the manufacturer's guidelines in an ABI Step OnePlus Real-Time PCR System (life technologies). GAPDH real-time PCR was performed by using the 5′ primer TTGGCCGTATTGGGCGCCTG and the 3′ primer CACCCTTCAAGTGGGCCCCG. Relative gene expression was determined using the conventional ßßCT method setting the control (unstimulated samples) as 1.
For cell counting BMMCs (10 6 cells/ml) were seeded in IL-3-free media. After 1 h cells were treated with IKK-inhibitor VII (for 30 min) and were stimulated with IL-3 as indicated. Cells were 1:1 mixed with trypan blue solution and trypan blue-negative cells were counted by using Neubauer counting chamber.

ELISA and proliferation assays
BMMCs (10 6 cells/ml) or V2D1 (10 5 cells/ml) were seeded in IL-3-free media. Cells were incubated with vehicle (DMSO) or inhibitors. BMMC were stimulated with IL-3, IL-33 or both (Peprotech). Supernatants were analyzed for IL-3, IL-6, TNFα and IL-13 using matched pair antibodies (eBioscience) by ELISA. For proliferation assays cells were cultured for 54 h. [H 3 ]thymidin (1 μCi) was added for additional 18 h. Incorporated radioactivity was measured by using a β-scintillation counter (Perkin Elmer). The increase of mast cell proliferation is indicated as the stimulation index (SI). Thereby the basal CPM values were set to 1.

Statistics
All experiments were performed at least three times (shown is one representative experiment). Proliferation assays, and ELISAs were performed three times in at least a 6-fold determination. Cytokine concentration is indicated as the mean of measurements ± standard deviation. For proliferation assays and ELISA one representative experiment is shown. The statistical analysis was performed with IBM SPSS Statistics version 20.0 (IBM). Statistical significance was assessed by Mann-Whitney-U test. Statistical significance was accepted for p < 0,05 (*, p < 0.05; **, p < 0.01; ***, p < 0.001).