The IL-33-induced p38-/JNK1/2-TNFα axis is antagonized by activation of β-adrenergic-receptors in dendritic cells

IL-33, an IL-1 cytokine superfamily member, induces the activation of the canonical NF-κB signaling, and of Mitogen Activated Protein Kinases (MAPKs). In dendritic cells (DCs) IL-33 induces the production of IL-6, IL-13 and TNFα. Thereby, the production of IL-6 depends on RelA whereas the production of IL-13 depends on the p38-MK2/3 signaling module. Here, we show that in addition to p65 and the p38-MK2/3 signaling module, JNK1/2 are essential for the IL-33-induced TNFα production. The central roles of JNK1/2 and p38 in DCs are underpinned by the fact that these two MAPK pathways are controlled by activated β-adrenergic receptors resulting in a selective regulation of the IL-33-induced TNFα response in DCs.


Results
JNK1/2 are essential for the IL-33-induced production of TNFα in BMDCs. Splenic DCs do not express the IL-33R 2 . In contrast to this, GM-CSF-generated BMDCs express the IL-33R and are thus sensitive to IL-33 stimulation 5,25 . Therefore we used BMDCs as an in vitro model to investigate IL-33-induced signaling pathways in DCs. As recently shown in BMDCs 5 , IL-33 induces a MyD88-NF-κB-mediated TNFα production ( Supplementary Fig. 1B-D) which also depends on the p38-MK2/3 signaling module ( Supplementary Fig. 1E,F). In addition, IL-33 activates JNK1/2 in BMDCs (Fig. 1A). Inhibition of JNK1/2 by SP600125 reduced the production of TNFα (Fig. 1B) but not of IL-6 (Fig. 1C). This demonstrates that beside the p38-MK2/3 signaling module 5 , JNK1/2 are dispensable for the IL-33-induced activation of IKK2 and p38. IL-33 induces a JNK-dependent TNFα response which also depends on IKKs and p38 5 . Therefore, we speculated that JNK1/2 activate IKK2 and/ or p38. However, neither the JNK1/2 inhibitor SP600125 (Supplementary Fig. 2A,B) nor JNK1 or JNK2 deficiency ( Supplementary Fig. 2C-F) influenced the IL-33-induced activation of IKK2 and p38. Next we determined the influence of JNK1 or JNK2 in the TNFα and IL-6 production. As shown in Supplementary  Fig. 2G neither JNK1 nor JNK2 deficiency reduced the IL-33-induced TNFα and IL-6 production in BMDCs. These data show that JNK1/2 are not involved in the IL-33-induced activation of IKK2 and p38 and that inactivation of all JNKs by SP600125, but not the specific inactivation of either JNK1 or JNK2, is prerequisite to reduce the production of TNFα.  Supplementary Fig. 5. (B,C) Wt BMDCs were treated with SP600125 (5 µM). Afterwards cells were stimulated with IL-33 (100 ng/ml) (n = 3). Supernatants were collected and analyzed for TNFα (B) or IL-6 (C) (n = 3). Shown is the mean ± SD; ***p < 0.001. IL-33 predominantly activates JNK2L in BMDCs. Next, we determined why neither JNK1 nor JNK2 deficiency influenced the IL-33-induced TNFα production. We hypothesized that a compensatory mechanism of the JNK isoforms in JNK deficient BMDCs. JNK1 and JNK2 are expressed as long (L) and short (S) isoforms (JNK1L/S and JNK2L/S) 32 . First, we evaluated the role of JNK1 isoforms by using jnk1 −/− BMDCs. Compared to the pJNK blots in wt BMDCs, the pJNK blots of the remaining JNK2L/S in jnk1 −/− BMDCs were reduced (Fig. 3A, Ai and Aii). However, when the controls in wt or jnk1 −/− BMDCs were set as 1, the fold activation of the JNK isoforms in wt BMDCs was similar to the fold activation of remaining JNK2L and JNK2S isoforms in jnk1 −/− BMDCs (Fig. 3A, Aiii and Aiv). This indicates, that inactivation of JNK1 reduced the total JNK activity in BMDCs without affecting the IL-33-induced activation of JNK2L/JNK2S in BMDCs. Next, we tested the role of JNK2 by using jnk2 −/− BMDCs. Compared to the pJNK blots in wt BMDCs, the activity of JNK1L was strongly reduced in jnk2 −/− BMDCs ( Fig. 3B and Bi). However, the activity of JNK1S in jnk2 −/− BMDCs was only slightly reduced compared to the pJNK blots in wt BMDCs (Fig. 3B and Bii). This indicates that with the deletion of JNK2L in jnk2 −/− BMDCs, BMDCs lose the main JNK L isoform which contributes to the total activity of large JNK1/2 (pJNK1/2 L) isoforms. In contrast to this, with the loss of JNK2S in jnk2 −/− BMDCs, BMDCs lose the short JNK isoform which slightly contributes to the total activity of the small JNK1/2 (pJNK1/2 S) isoforms in wt BMDCs. However, the fold induction of JNK1L ( Fig. 3B and Biii) and JNK1S ( Fig. 3B and Biv) in jnk2 −/− BMDCs is similar to the pJNK blots in wt and BMDCs indicating that JNK2 deficiency also does not influence the IL-33-induced activation of JNK1S and JNK1L. The barely detectable activation of the remaining JNK1L isoform in JNK2 deficient BMDCs indicates that IL-33 predominantly induces the activation of JNK2L. However, the deletion of one JNK isoform is compensated by the remaining JNK isoform.
Noradrenalin modulates the IL-33-induced cytokine production. Adrenergic receptors are negative regulators of TIR family member-mediated signaling 23,33 . We investigated, whether stimulation of adrenergic receptors influence the IL-33-induced cytokine response in BMDCs. As shown in Fig. 4A,B treatment with Noradrenalin strongly reduced the IL-33-induced production of TNFα, but did not affect the IL-6 production. Thereby, 1 µM and 10 µM Noradrenalin are equally effective to reduce the IL-33-induced TNFα response (Fig. 4A). Next, we tested the stimulation sequence with Noradrenalin and IL-33. Simultaneous or pre-incubation with Noradrenalin for 30 min most efficiently blocked the IL-33-induced TNFα production (Fig. 4C). Stimulation of adrenergic receptors mediates the activation of adenylate cyclases and thereby the production of the second messenger cAMP 34,35 . Forskolin, an activator of the adenylate cyclases, strongly increases the production of cAMP 34 . Treatment of BMDCs with Forskolin blocked the IL-33-induced production of TNFα (Fig. 5A), but not the IL-6 production (Fig. 5B). This indicates that adrenergic receptors via cAMP inhibit IL-33-induced signaling pathways. Noradrenalin is a non-selective agonist of adrenergic receptors. Treatment of BMDCs with Propranolol, an antagonist of β-adrenergic receptors, reverses the effects of Noradrenalin on the IL-33-induced TNFα production whereas the production of IL-6 was not altered (Fig. 5C,D) indicating that Noradrenalin via β-adrenergic receptors controls the IL-33-induced TNFα production.
Noradrenalin blocks the IL-33-induced activation of p38 and of JNK. Noradrenalin blocked the IL-33-induced production of TNFα but not of IL-6 in BMDCs, most likely by blocking essential signaling pathways involved in the IL-33-induced TNFα, but not IL-6 production. The production of TNFα but not of IL-6 depends on the p38-MK2/3 signaling module, and on JNK1/2. Thus we tested which of these signaling pathways are influenced by stimulation with noradrenalin. While Noradrenalin alone did not induce the activation of JNK1/2 and of p38 in BMDCs, treatment of BMDCs with noradrenalin equally reduced the IL-33-induced activation of JNK1 and JNK2 as well as of p38 (Fig. 5E). These data demonstrate that β-adrenergic receptors specifically regulate the production of TNFα by controlling the IL-33-induced activation of p38 and JNK1/2. www.nature.com/scientificreports www.nature.com/scientificreports/
Our data further indicate that the ligand-dependent mode of cooperation of different MAPK pathways mediates different cellular responses in DCs. Whereas the linked activation between the p38-MK2/3 signaling  . (A,B) Wt BMDCs were treated with Forskolin (10 µM) and stimulated with IL-33 (100 ng/ml). Supernatants were collected and analyzed by ELISA (shown is the mean ± SD; ***p < 0.001) (n = 3). (C,D) Wt BMDCs were treated with Propranolol (1 µM) and stimulated with Noradrenalin (10 µM) and IL-33 (100 ng/ml). Supernatant were collected and analyzed by ELISA (shown is the mean ± SD; ***p < 0.001) (n = 4). (E) Wt BMDCs were stimulated with either Noradrenalin (10 µM) or IL-33 (100 ng/ml) or both together. Lysates were analyzed by Western blotting. The unstimulated control was set as 1 (shown is the mean ± SD of n = 3 independent experiments; *p < 0.05, **p < 0.005 and ***p < 0.001). The original blots are shown in Supplementary Fig. 8 www.nature.com/scientificreports www.nature.com/scientificreports/ module and of JNK1/2 controls the GM-CSF-induced proliferation, the parallel activation of both MAPK pathways mediate the IL-33-induced cytokine production. Thereby the functional cooperation of JNK1/2 and the p38-MK2/3 signaling module together with NF-κB is essential for TNFα production in IL-33-activated BMDCs ( Supplementary Fig. 3B-D). In contrast to this, the IL-33-induced production of IL-6 neither depends on the p38-MK2/3 module nor on JNK1/2, but on NF-κB 5 ( Supplementary Fig. 3B-D), whereas the production of IL-13 in BMDCs depends on the p38-MK2/3 signaling module, but not on NF-κB 5 . This underpins the essential and central role of JNK1/2 for the IL-33-induced TNFα production in BMDCs. However, there is no preference for a JNK isoform which mediates the IL-33-induced TNFα production. Neither inactivation of JNK1, nor of JNK2, influenced the IL-33-induced TNFα production. Only the pharmacological JNK inhibition by SP600125 strongly reduced the TNFα production induced by IL-33. This is explained by the fact that in contrast to a pan JNK inhibitor, in jnk1 −/− BMDCs the activation of JNK2L, and in jnk2 −/− BMDCs the activation of JNK1S are still intact. Therefore, the remaining JNK isoform together with the p38-MK2/3 signaling module and NF-κB is sufficient to mediate the IL-33-induced TNFα production. We hypothesize that JNK1/2 cooperatively with NF-κB mediate the transcription of TNFα. In contrast to this, the p38-MK2/3 signaling pathway stabilizes the TNFα transcripts ( Supplementary Fig. 3B-D) 39 and further mediates the translation of the TNFα transcripts via the mTOR-RSK pathway 5 .
The importance of JNK1/2 and p38 is further supported by the fact that both MAPK pathways and the resulting TNFα response are inhibited by activated β-adrenergic receptors ( Supplementary Fig. 4A,B). The mechanism underlying the inhibitory effect is unknown. However, β-adrenergic receptors activate the cAMP-dependent protein kinase A (PKA)-CREB signaling pathway 40,41 which mediates the expression of MKPs and thus controls JNK1/2 and p38 37,38 . Therefore, we speculate that β-adrenergic receptors induce the expression of MKPs in BMDCs and thus limit the activation of p38 and JNK1/2 as well as the resulting TNFα production. In conclusion, by regulating JNK1/2 and p38, β-adrenergic receptors control the composition of IL-33-induced cytokine profiles of DCs ( Supplementary Fig. 4A,B) and thus regulate their mediated effector functions. Given the fact that GM-CSF-generated BMDC resemble to inflammatory DCs 21 , the regulatory function of β-adrenergic receptors on IL-33-induced MAPK activation might also be important for DCs in vivo. Interestingly, an in vivo relevance of a crosstalk between the signaling of the IL-33R and β-adrenergic receptors has recently been shown in ILC-2. In these cells the IL-33-induced and p38-dependent IL-13 production 14 is blocked by β 2 -adrenergic receptors and resulted in reduced inflammatory responses in vivo 42 . Together these data indicate that neuro-regulation of IL-33-induced effector functions on innate cells is a general mechanism to control and thus to avoid over-exuberant IL-33-induced inflammation. Therefore this provides novel therapeutic targeting strategies to modulate IL-33-induced inflammatory responses.

BMDC-generation.
For generation of BMDCs we used the protocol as recently published 5 . In brief, bone marrow cells were seeded (2 × 10 5 cells/ml) and after day 3, 6 and 8 medium [RPMI 1640 (Sigma Aldrich), with supplements and conditioned GM-CSF (20 ng/ml) supernatants from X63AG-GM-CSF cells] was refreshed. BMDCs were harvested (on day 9 or 10) and identified by surface expression of CD11c and CD11b (both from eBioscience) by flow cytometry.

Stimulation of BMDCs and lysis.
Prior to stimulation, BMDCs were starved for GM-CSF for 1 h.
Afterwards cells were pre-incubated for 30 min with inhibitors (as indicated in the Figures) (all Merck Millipore) and stimulated with IL-33 (Peprotech). In some experiments (as indicated in the Figures) BMDCs were treated with Noradrenalin (Sigma Aldrich) for 30 min and then stimulated with IL-33. Cell lyses was performed with a standard protocol 5 . Protein concentration was determined by using the BCA-assay (Pierce). Afterwards lysates were boiled in 6 × Laemmli buffer.