5, 6-Dimethylxanthenone-4-acetic acid (DMXAA, now named Vadimezan) is a cell permeable, small molecule in the xanthone class that shows potent antitumor activity in several mouse models. In human, the clinical benefits of this drug have also been demonstrated in both phase I and phase II clinical trials [31
]. The mechanism by which DMXAA exerts its anti-tumor actions has been of significant interest ever since its initial discovery in the 1990’s. In the murine system, actions on both endothelial cells and leukocytes, in particular macrophages and dendritic cells, have been demonstrated [1
]. With regard to macrophages, it is known that DMXAA can induce the secretion of large amounts of a number of proinflammatory cytokines and chemokines (i.e. TNFα, IL-6, IP10, and interferon-β). One particularly interesting aspect of DMXAA is that this macrophage activation appears to involve multiple signaling pathways. There are data to implicate the NF-κB pathway [5
], the TBK1-IRF3 signaling axis [6
], and the NOD pathways [7
]. To date, however, to our knowledge, there is only one previous study that indirectly implicated a role for p38 MAP kinase in the actions of DMXAA [8
]. In this study, the formation of DMXAA-induced tumor cell networks on matrigel was blocked by addition of the p38 inhibitor SB203580. Given the biological significance of MAPK pathways in inflammation and the multi-faceted activation events induced by DMXAA, we explored the role of the MAPKs signaling family in DMXAA-induced proinflammatory cytokine production.
Our results clearly show that DMXAA can induce rapid phosphorylation (beginning at 30 minutes) of all three MAPKs pathways in murine macrophages ( and ). Given the fact that DMXAA can stimulate large amounts of type I interferon secretion from macrophages [6
] and type I interferons can activate MAPKs pathway [23
], it was necessary to determine if the phosphorylation of MAPKs by DMXAA was due to a direct effect or due to secondary autocrine signaling via the type I interferons induced by DMXAA. To examine this issue, we used two approaches. First, we quantified the kinetics of IFNβ production (). In MHS cells, we were unable to detect any IFNβ in the cell supernatants at 2 hrs after DMXAA stimulation, as compared to large amounts at the 5hrs time point (4.5ng/ml). In peritoneal macrophages, we detected only minute amounts (< 10 pg/ml) of IFNβ at 2 hrs. We believe that the delayed appearance of IFNβ protein makes IFN-induced MAPKs phosphorylation highly unlikely. Second, we stimulated the MHS cells with DMXAA in the presence or absence of a blocking antibody against the type I interferon receptor (IFNAR-1) to inhibit any potential secondary autocrine effects of type I interferons. Robust phosphorylation of p38 induced by DMXAA was seen in a concentration of the IFNAR-1 antibody that blocked IFN-mediated induction of IP-10 (, Supplemental Fig. S1
). Taken together, these results indicate that phosphorylation of MAPK induced by DMXAA is a primary effect that is independent of the Interferon/IRF-3 pathway.
Our group has recently shown that DMXAA can also activate the intracellular NOD signaling pathway. Since it is known that NOD1 and NOD2 can activate MAPKs through interacting with a protein kinase named RIP2 (receptor-interacting protein 2; also called RICK, RIPK2 and CARDIAK) [27
], we wanted to determine if DMXAA-induced MAPK activation worked through this signaling pathway. To evaluate this, we exposed macrophages derived from RIP2 knock-out mice to DMXAA. Our data () showed that DMXAA was still able to induce phosphorylation of all three MAPKs independently of the NOD/RIP2 pathway.
Studies showing the induction of phosphorylation of key members of each of the three MAPK pathways by DMXAA were followed by experiments evaluating the functional significance of each pathway using specific pharmacologic inhibitors. We first tested a commonly used p38 MAPK inhibitor, SB203580, and were able to detect inhibition of proinflammatory cytokine secretion induced by DMXAA (data not shown). However, recent reports showed that this inhibitor was found to also inhibit RIP2 kinase with even greater potency than p38 MAPK [27
]. Although we had evidence to suggest that RIP2 was not involved (), we repeated the inhibitory experiments using a much more specific p38 MAPK inhibitor, BIRB796, at 0.1µM, a concentration at which the drug is reported to inhibit p38 MAPK specifically, without blocking RIP2, ERK1/2, JNKs or many other protein kinases [29
]. Our results demonstrated that both TNF-α and IL-6 protein secretion stimulated by DMXAA was reduced by pre-treatment of macrophages with BIRB796 ( and Supplemental Fig. S2
). We also applied an ERK- selective inhibitor, FR180204 at 5µM. At this concentration, the inhibitor is reported to show much higher specific inhibitory effect on the kinase activity of ERK1 and ERK2 over p38 and other kinases [28
]. Again, we saw reduction of DMXAA-stimulated cytokine secretion ( and Supplemental Fig. S3
). Finally, we used a SAPK/JNK1,2/3- selective inhibitor, 420135, which has 1000-fold selectivity for JNKs over other MAP kinases including ERK, p38 and little inhibitory activity against another 74 kinases [30
]. In this case, the inhibitor had no effect on DMXAA-induced cytokine production (). Our data thus show that the p38 and ERK1/2 MAPKs play a functional role in the regulation of proinflammatory cytokine production induced by DMXAA, but interestingly, MAPK pathways seem to play little role in the induction of chemokines by DMXAA ().
We also explored the mechanism by which MAPK signaling affected the production of cytokines after DMXAA stimulation. There are a number of reports showing effects of p38 due to transcriptional regulation. p38 MAPK is known to regulate various transcription factors, such as CHOP [37
], and ATF-2, Elk-1 [38
] by phosphorylation. Studies have shown that p38 MAPK regulates NF-κB-dependent gene transcription via regulating DNA binding of the TATA-binding protein (TBP) to the TATA box [39
]. ERK pathway has been reported to be involved in upregulation of TNF-α production by increasing TNF-α promoter activity via increased DNA binding activity of Egr-1 and NF-κB to TNF-α promoter [20
]. However, there are also previous studies demonstrating that the p38-MK2 signaling pathway plays an important role in the post-transcriptional regulation of both TNF-α and IL-6 protein secretion by regulating either translation or mRNA stability. The mechanism appears to be dependent on the binding status of phosphorylated adenine/uridine-rich element (ARE)-binding proteins, such as Tristetraprolin (TTP), to the ARE in the 3’ untranslated region of TNF-α and IL-6 mRNA [15
]. Similarly, in addition to transcriptional regulation, Tpl2/ERK signaling pathway also play an important role in post-transcriptional regulation of TNF-α production through regulating nucleocytoplasmic mRNA transport via a mechanism that targets the ARE in the 3’UTR of the TNF-α mRNA [22
]. Another possibility is that MNK/eIF4E pathway may be involved in the regulation of TNF-α production. MNK kinases are downstream targets of both the p38 and ERKs MAP kinases [42
], and have been shown to be involved in the, regulation of TNF and other pro-inflammatory cytokines in response to various TLR agonists [44
]. We have no direct evidence for this pathway yet, however, we plan to study the MAPK-MNK/eIF4E signaling axis in the future.
It was therefore of interest to analyze changes in both protein production and mRNA expression. We found that DMXAA stimulated large increases in both mRNA and protein secretion of macrophage TNFα and IL-6 ( and ). However, although the blockade of either p38 MAPK or ERK1/ERK2 significantly inhibited the production of TNFa and IL-6 protein secretion, the inhibitors had no effect on the mRNA expression levels ( and ). These data suggest that the activated p38 and ERK1/ERK2 signaling pathways induced by DMXAA affect the secretion of TNF-α and IL-6 primarily through regulating either translation or mRNA stability, not cytokine transcription.
Finally, we explored the role of the MAPKs in the phenomenon called “macrophage priming”. This is a well-known process by which low levels of INF-γ, that has very little effect on their own, but can synergize with LPS or microparticulate β-glucan to enhance the production of TNF-α, IL-6 or nitric oxide [46
]. The mechanisms underlying the IFN-γ priming action are complicated and still not known for certain, however, some studies have suggested IFN-γ priming can augment DNA binding of NF-κB in response to LPS [46
] or can cause up-regulation of LPS uptake and expression of the intracellular TLR4-MD-2 complex [49
]. Other studies have demonstrated that the priming of macrophages by IFN-γ is highly dependent on glycogen synthase kinase-3 [47
] or is dependent on the presence of functional NOD2 protein [50
]. The MAPKs may also be involved. At least one study suggested that IFN-γ can rapidly induce ERK1/2 phosphorylation which then plays a vital role in IFN-γ- inducible macrophage nitric oxide generation besides the involvement of JAK2-STAT1 pathway activated by IFN-γ [51
]. Another study suggested that activation of p38 MAPK by IFN-γ participates in the regulation of cytokines such as TNF-α, and INOS gene expression [52
Priming of the DMXAA effect by LPS in leukocytes has been previously described [53
]. We found that exposure of macrophages to a low dose of IFN-γ alone for 5h did not increase TNF-α protein secretion, but a clear priming effect could be seen by pretreatment with IFN-γ (). When we explored the involvement of MAPKs in this process, we saw a small increase of phosphorylation of p38 and ERK1/2 after IFN-γ treatment alone, but increased phosphorylation of all three MAPKs, especially ERK1/2, after DMXAA exposure in IFN-γ primed cells. We also saw clear blunting of TNF-α production in the primed cells after blocking p38 and ERK1/2. Thus, our data show that one of the mechanisms of synergistic effect between IFN-γ and DMXAA in augmenting TNF-α production in mouse macrophages is via post-transcriptional regulation due to enhancement of DMXAA-induced p38 and ERK1/2 MAPK activation through IFN-γ priming.
One important caveat to consider in our studies was that we used pharmacologic inhibitors of MAPKs. Since the three MAP kinases share 60–70% amino-acid sequence identity, we took care to use the most specific inhibitors available for each MAPK. In our p38 MAPK inhibitory experiments, we used a p38 MAPK inhibitor, BIRB796, which is much more specific than the most commonly used p38 MAPK inhibitor, SB203580, which also inhibits RIP2. In our hands, BIRB796 showed very strong inhibitory effects on the DMXAA-induced phosphorylation and activation of p38 MAPK at 0.1µM with no inhibitory effects on the phosphorylation of ERK1/2, JNKs MAPKs (data not shown). This supports the conclusion that the effects of BIRB796 are primarily due to p38 MAPK inhibition. For the ERK1/2 and JNKs MAPK blocking studies, we used recently developed inhibitors that directly inhibit either ERK1/2 or JNKs MAPKs activity instead of the more commonly used MEK1/2 or MKK4/7 inhibitors. The two inhibitors used in this study, FR180204 and 420135 are reported to have much more specific inhibitory effects on the kinase activity of ERK1/2 or JNKs (respectively) than of p38 MAPK and other kinases [28
]. It is recognized, however, that no inhibitor is completely specific. Although we would have liked to use genetic approaches to study these pathways (as we did with the RIP2 KO mice), attempts to achieve complete knockdown p38 or ERK1/2 MAPK using knockout mouse technology has been limited by either the critical role of p38α or ERK2 MAPK in mouse development, causing embryonic lethality or compensation by different isoforms of MAPKs [10
A major unanswered, but important question in these experiments is how a small, cell permeable molecule like DMXAA is able to activate all three MAPKs, in addition to the NF-κB, TBK1/IRF3, and NOD/RIP2 signaling networks. This question has been pursued for more than a decade by our labs and others without a clear positive answer. It has been shown, however, using knockout mice and other approaches, that DMXAA does not function through all known TLRs, cytosolic helicase receptors, or MyD88 [4
]. Our studies have supported these findings. Finding the “DMXAA receptor” remains a goal that may be especially valuable in correlating findings in murine versus human cells.
In conclusion, this paper provides strong evidence that DMXAA is able to induce phosphorylation of all three MAPKs pathways in murine macrophages and that at least two of these subfamilies (p38 and ERK1/2) play a role in the increased production of proinflammatory cytokines, such as, TNF-α and IL-6 through a post-transcriptional mechanism. These data further expand our knowledge on mechanism how DMXAA acts as a potent anti-tumor agent and will hopefully be helpful in understanding its potential efficacy in human cancer therapy.