Tumor suppressor death-associated protein kinase 1 inhibits necroptosis by p38 MAPK activation

Death-associated protein kinase 1 (DAPK1, DAPk, DAPK) is known for its involvement in apoptosis and autophagy-associated cell death. Here, we identified an unexpected function of DAPK1 in suppressing necroptosis. DAPK1-deficiency renders macrophages and dendritic cells susceptible to necroptotic death. We also observed an inhibitory role for DAPK1 in necroptosis in HT-29 cells, since knockdown or knockout of DAPK1 in such cells increased their sensitivity to necroptosis. Increased necroptosis was associated with enhanced formation of the RIPK1–RIPK3–MLKL complex in these DAPK1-deficient cells. We further found that DAPK1-deficiency led to decreased MAPK activated kinase 2 (MK2) activation and reduced RIPK1 S321 phosphorylation, with this latter representing a critical step controlling necrosome formation. Most TNF signaling pathways, including ERK, JNK, and AKT, were not regulated by DAPK. In contrast, DAPK bound p38 MAPK and selectively promoted p38 MAPK activation, resulting in enhanced MK2 phosphorylation. Our results reveal a novel role for DAPK1 in inhibiting necroptosis and illustrate an unexpected selectivity for DAPK1 in promoting p38 MAPK-MK2 activation. Importantly, our study suggests that modulation of necroptosis and p38/MK2-mediated inflammation may be achieved by targeting DAPK1.

Recent studies have further revealed several cell death suppressing checkpoints centered on RIPK1. Complex Iassociated RIPK1 is subjected to phosphorylation by IKK that inhibits the transition into complex II 20 . In a different inhibitory process, TNF-induced activation of p38 MAPK leads to phosphorylation of MAPKAPK2 (MK2), with this latter directly phosphorylating RIPK1 at the S321 and S326 positions (in mouse; S320 and S325 in human) to prevent binding of RIPK1 with FADD/caspase-8 to form complex II [21][22][23] . Alternatively, TBK1-mediated phosphorylation of RIPK1 inhibits RIPK1 activation and the conversion into complex II 24,25 .
Necroptosis participates in numerous pathological events [3][4][5][6]26 . It often leads to inflammation and is one of the mechanisms involved in counteracting specific viral infections, while expression of viral molecules can result in evasion of the necroptotic process. Necroptosis also contributes to ischemia-reperfusion injury, transplantation rejection, and cancer inhibition or progressing [3][4][5][6]26 .
Death-associated protein kinase 1 (DAPK, DAPk, DAPK1) is a multi-domain serine/threonine kinase regulated by calcium [27][28][29] . It was first identified for its role in mediating IFN-γ-induced cell death 30 , but subsequent evidence demonstrated involvement in apoptotic cell death induced by Fas 31 , TGF-β 32 , ceramide 33 , matrix detachment 34 , unliganded Netrin-1 receptor uncoordinated protein 5 homolog 2 (UNC5H2) 35 , or ER stress 36 . Although how DAPK1 regulates apoptosis remains incompletely understood 37 , it is known that it interacts with p53 and promotes p53-dependent cell death 38,39 . In addition, DAPK1 interacts with Fas-associated protein with death domain (FADD) 38 . DAPK1 is a tumor suppressor and is specifically downregulated in many types of cancer 40,41 . Apart from its role in apoptosis, DAPK1 participates in a wide variety of cellular events including autophagy, membrane blebbing, and stress fiber formation that all contribute to its tumor-suppressing functions. In T lymphocytes, DAPK1 inhibits T cell activation by suppressing T cell receptor-induced NF-κB activation 42 .
In the present study, we found that DAPK1 negatively regulates necroptosis, unlike its active involvement in other forms of cell death. DAPK1-deficiency enhances the sensitivity of myeloid and HT-29 cells to necroptotic induction. Knockdown or knockout of DAPK1 in HT-29 cells increases their sensitivity to necroptosis. The increased necroptosis in DAPK1-deficient cells was associated with enhanced formation of the RIPK1-RIPK3-MLKL complex. We further found that DAPK1 selectively increased TNF-induced p38 MAPK and MK2 activation, leading to phosphorylation of RIPK1 at position S321 and inhibition of necrosome formation. Our results reveal a novel role for DAPK1 in inhibiting necroptosis and illustrate the diverse death-associated physiological functions regulated by DAPK1.
In contrast, we found that Dapk1 −/− BMDMs were more resistant to thapsigargin-triggered apoptosis than WT BMDMs ( Supplementary Fig. 5a), consistent with the pro-apoptotic role of DAPK1 in ER stress-induced cell death 36 . In Jurkat cells, a cell line sensitive to Fasinitiated apoptosis, DAPK1 knockdown did not affect surface Fas expression but it did reduce Fas ligand (FasL)-triggered cell death ( Supplementary Fig. 5b, c). BMDMs are moderately sensitive to FasL-induced apoptosis, and we found that DAPK1-deficiency reduced the extent of cell death mediated by FasL in such cells ( Supplementary Fig. 5d). Therefore, consistent with the known involvement of DAPK1 in apoptosis, DAPK1-deficiency attenuates ER stress-and FasLinduced cell death. The enhanced susceptibility of DAPK1-deficient myeloid cells to necroptosis reveals a selective inhibitory role for DAPK1 in necroptosis.

Necroptosis is increased upon downregulation of DAPK1 in HT-29 cells
The enhanced sensitivity to necroptosis was not restricted to Dapk1 −/− myeloid cells. A similar effect was found in the human colon adenocarcinoma cell line HT-29. HT-29 cells were previously shown to be susceptible to necroptotic induction by treatment with zVAD plus SMAC mimetics 9 . We knocked down DAPK1 by shRNA in HT-29 cells, which did not affect expression of RIPK1 or RIPK3 (Fig. 2a). Treatment of WT HT-29 cells with zVAD or BV6 alone did not trigger cell death, as measured by propidium iodide (PI) staining (Fig. 2b). However, a combination of zVAD plus BV6 did induce cell death in WT HT-29 cells, and this outcome was significantly enhanced upon DAPK1 deficiency (Fig. 2b). The necroptotic nature of the cell death outcome was confirmed by addition of Nec-1, which suppressed the effect in both WT and DAPK1-knockdown HT-29 cells (Fig. 2b).

Sensitization to TNF-induced septic shock in DAPK1knockout mice
We also examined whether DAPK1-deficiency increases necroptosis in vivo. In RIPK1-and RIPK3-dependent processes, necroptosis mediates TNF-induced systemic inflammatory response syndrome. The administration of TNF-induced hypothermia but did not trigger death in WT mice (Fig. 5a, b). DAPK1-deficient mice were highly sensitive to severe hypothermia and lethality triggered by TNF (Fig. 5a, b). Therefore, DAPK1-deficiency conferred sensitivity to necroptosis in cultured cells and in vivo.

DAPK1 binds p38 MAPK and promotes p38 activation
MK2 is activated by p38 MAPK in the nucleus, followed by cytoplasmic entry to phosphorylate target substrates 44,45 . We determined the distribution of MK2 in Dapk1 −/− macrophages before and after TNF stimulation. The overall quantity of MK2 in the cytosol and nucleus was comparable between WT and Dapk1 −/− macrophages ( Supplementary Fig. 11). The reduced phosphorylation of MK2 in Dapk1 −/− macrophages was thus not associated with altered cytoplasmic presence of MK2.
We found that the kinase activity of DAPK1 was not involved in necroptosis inhibition. The introduction of kinase-dead DAPK1 mutant (DAPK1[K42A]) was as effective as WT DAPK1 in enhancing the survival of DAPK1-null HT-29 cells to necroptosis (Fig. 6b), as well as increasing the activation of p38 MAPK (Fig. 6c). Collectively, the kinase activity of DAPK1 is not required for conferring the resistance to necroptosis.
Overexpression of MKK3 alone, however, did not increase the survival of DAPK1-null HT-29 cells to necroptosis. In addition, DAPK1 was found to interact with p38 MAPK in HEK293T cells, in which p38 MAPK-HA was pulled down by DAPK1-FLAG, and vice versa (Fig. 6e). Recombinant p38 MAPK also bound purified DAPK1 protein (Fig. 6f), suggesting that DAPK1 may affect p38 MAPK activation through binding p38 MAPK. This was The extents of RIPK1(S321) and p38 MAPK activation from three independent experiments in (d) were quantitated using normalized intensity of pRIPK1(S321) and p-p38 in WT BMDMs at 15 min as 1. Mean ± SD are shown. *P < 0.05, ***P < 0.001 for two-way ANOVA followed by a Sidak's multiple comparison test. (g) Inhibition of p38 MAPK or MK2 confers susceptibility to necroptosis in WT macrophages. WT BMDMs were treated with zVAD, AT-406, SB203589 (1 μM) and PF3644022 (2 μM), as indicated, and the extent of necroptosis determined. Values are mean ± SD of triplicates in a single experiment. **p < 0.01, ***p < 0.001 for unpaired t-test. Results have been repeated in two independent experiments.
confirmed in an in vitro kinase analysis, in which recombinant MKK3 was used to activate recombinant p38 MAPK. The addition of increasing amounts of recombinant DAPK1 increased MKK3-directed p38 MAPK phosphorylation (Fig. 6g). In contrast, DAPK1 [ΔCyt], the DAPK1 mutant lacking cytoskeleton- Recombinant human p38 MAPK (200 ng) was incubated with purified DAPK (100 ng), as indicated. p38 MAPK was pulled down by anti-p38 MAPK (ab170099, Abcam), and the presence of DAPK and p38 MAPK in immunoprecipitants determined by anti-FLAG and anti-His, respectively. Results have been independently confirmed using another anti-p38 MAPK (7218, Cell Signaling). g DAPK1 promotes MKK3-directed p38 MAPK phosphorylation. Recombinant p38 MAPK (50 ng), MKK3 (100 ng), and DAPK1-FLAG (25 ng or 75 ng) was incubated as indicated in ATP-containing kinase buffer at 30°C for 1 h. The amounts of p38 MAPK, MKK3, DAPK1 and phospho-p38 MAPK was determined by immunoblots. h Binding of different DAPK1 mutants to p38 MAPK. HEK293T cells were transfected with different mutants of DAPK1-FLAG and p38-HA as indicated. The presence of p38-HA in anti-FLAG precipitates was determined. i Failure of DAPK1(ΔCyt) to increase p38 MAPK activation. DAPK1, DAPK1(ΔCyt) or DAPK1(ΔCAM) was evaluated for its ability to increase MKK3-directed p38 MAPK activation described in (g). Values are mean ± SD of triplicates in a single experiment. *P < 0.05, ***P < 0.001 (b, d) for two-way ANOVA followed by a Tukey's multiple comparison test. Results have been repeated in three (a-c, e, f) or two (d, g-i) independent experiments. interacting domain, failed to bind p38 MAPK (Fig. 6h), and was less effective to activate p38 MAPK than WT DAPK1 (Fig. 6i). Therefore, DAPK1 selectively increases p38 MAPK activation likely by its specific interaction with p38 MAPK.
Together, our results illustrate that DAPK1 inhibits necroptosis induction through enhanced activation of p38-MK2-RIPK1 cascade. DAPK1 specifically targets the activation of p38 MAPK by binding p38 MAPK, with consequently increased activation of MK2 and elevated RIPK1 S321 phosphorylation, resulting in suppression of necroptosis.

Discussion
In this study, we identified an unexpected role for DAPK1 in necroptosis. DAPK1 has mostly been shown to promote cell death. It mediates the apoptotic cell death induced by stimuli as diverse as Fas, TGF-β, ceramide, and ER stress [30][31][32][33]36 . This was also confirmed in the present study in which Dapk1 −/− BMDMs and DAPK1knockdown Jurkat T lymphoma cells were more resistant than their WT counterparts to FasL-or ER stresstriggered death (Supplementary Fig. 5). Furthermore, DAPK1 mediates cell death induced by oxidative stress through phosphorylation of PKD and subsequent activation of JNK 43 , and during ischemic injury, by activation of the N-methyl-d-aspartate (NMDA) glutamate receptor through interaction and phosphorylation of the NR2B subunit 46 . DAPK1 has also been implicated in mediating autophagic cell death 36,47,48 . In contrast, DAPK1deficiency rendered myeloid cells and HT-29 cells more sensitive to necroptosis triggered by treatment with zVAD plus either SMAC mimetics, TNF, or IFN-β (Figs. 1 and 2, Supplementary Figs. 2-4). The ability of DAPK1 to antagonize necroptosis therefore stands out among the capacity of DAPK1 to promote various types of cell death.
In addition to phosphorylation of the target proteins, DAPK1 has also been shown to regulate target protein function though protein-protein interaction that is independent of DAPK1 catalytic activity. DAPK1 binds and activates pyruvate kinase M2 (PKM2) in the absence of DAPK1 kinase domain 49 . The interaction of DAPK1 death domain with microtubule-affinity regulating kinase (MARK) activates MARK1/2 independent of DAPK1 kinase activity 50 . DAPK1 binds NLRP3 to promote NLRP3 inflammasome activation without the involvement of DAPK1 catalytic activity 51 . In the present study, we also found that the kinase activity of DAPK1 was not essential for repressing necroptosis (Fig. 6b, c). We further found that DAPK1-deficiency specifically attenuated the activation p38 MAPK, but not of ERK and JNK (Fig. 5c-e). In addition, the activation of MKK3 was normal in Dapk1 −/− BMDMs (Fig. 6a), suggesting the process regulated by DAPK1 is at the stage of p38 MAPK. Overexpression of p38 MAPK, but not MKK3, restored the viability of DAPK1 −/− HT-29 cells to necroptosis induction (Fig. 6d). A possible scaffold role of DAPK1 was suggested by the direct interaction of DAPK1 with p38 MAPK (Fig. 6e, f). We also mapped the cytoskeletonbinding domain of DAPK1 as the region interacting with p38 MAPK (Fig. 6h). Furthermore, in the in vitro kinase analysis containing only p38 MAPK, MKK3 and ATP, the inclusion of DAPK1 enhanced the phosphorylation of p38 MAPK (Fig. 6g). DAPK1 mutant (DAPK1ΔCyt) that did not bind p38 MAPK failed to enhance p38 MAPK activation in the same assay (Fig. 6h, i), suggesting that association of p38 MAPK with DAPK is required for the full TNFR-induced p38 MAPK activation. Therefore, our results provided the most direct evidences to illustrate on how DAPK1 increases TNF-directed p38 MAPK activation.
DAPK1 displays anti-inflammatory and proinflammatory activities in different cell types 52 . We previously demonstrated that TCR-induced NF-κB activation is enhanced in Dapk1 −/− T cells 42,53 , whereas LPS-triggered NF-κB activation is modestly reduced in Dapk1 −/− BMDMs 51 . In the present study, TNF-induced NF-κB activation and IKK activation in myeloid cells was not affected by DAPK1-deficiency (Fig. 5c), suggesting that increased necroptosis in Dapk1 −/− BMDMs is not linked to the processes of NF-κB activation. As another example, TCR-induced p38 MAPK activation is normal in Dapk1 −/− T cells 42 , while TNFR-initiated p38 MAPK phosphorylation was impaired in Dapk1 −/− BMDMs (Fig. 5d). These results further support the notion that signaling modulation by DAPK1 is cell type-and surface receptor-dependent 52 .
Necroptosis plays a dual role in cancer. Induction of necroptosis is a legitimate approach to killing tumor cells, especially for those that are resistant to apoptotic death, as confirmed by various in vitro studies 26,54,55 . Notably, RIPK1, RIPK3 and/or MLKL are downregulated in various types of cancer [56][57][58] , reflecting the necessity for tumor cells to circumvent necroptosis 26,59 . Our observation that DAPK1-deficiency increases necroptosis may seem incompatible with DAPK1's tumor-suppressing role. However, our results also suggest that some of DAPK1's inhibitory activity lies at the apex of the necroptotic pathway, i.e., RIPK1 S321 phosphorylation, but many tumors evade necroptosis at the effector stages, e.g. by mutations in RIPK3 or MLKL 26,54,55 . In addition, necroptosis may contribute to tumorigenesis 54,59 . RIPK1-RIPK3 necrosomes promote oncogenesis and immune suppression in pancreatic ductal adenocarcinoma 60 . Melanoma cells also trigger necroptosis of endothelial cells to promote tumor extravasation and metastasis 61 . Therefore, the necroptosis-inhibitory activity of DAPK1 does not necessarily contradict its tumor-suppressing function. The application of DAPK1-mediated necroptosis suppression is likely tumor-type and tumor stage dependent. Whether downregulation of DAPK1 in several tumor types may provide an opportunity to target these cancers with necroptosis-inducing agents warrants further exploration.
Necroptosis plays a critical role in the regulation of infection, inflammation and carcinogenesis, and the therapeutic applications of precise necroptosis regulation are well recognized. Recent studies reveal a specific role for RIPK1 phosphorylation at S321/326 in the control of necroptosis [21][22][23] . Our study suggests the possibility of regulating phosphor-S321/326 RIPK1-mediated necroptosis by modulating DAPK1 levels. Enhanced expression of DAPK1 inhibits necroptosis (Fig. S7), whereas reduced DAPK1 promotes necrosome formation (Figs. 1 and 2). Thus, DAPK1-mediated controlled necroptosis represents a potential therapeutic approach, but future research is needed to identify a reliable means of DAPK1-based necroptotic regulation.
The identification of p38 MAPK-MK2 as target signal molecules by DAPK1 reveals an additional unexpected regulatory pathway. MK2 has been implicated for its tumorigenic role in various cancers 62 . In addition, MK2 directly promotes autoimmune and inflammatory diseases including rheumatoid arthritis, chronic obstructive pulmonary disease, cardiovascular diseases and diabetes 63,64 . The selective association of DAPK1 to p38 MAPK-MK2 activation in macrophages provides a prospect in the regulation of MK2 activation and MK2associated inflammatory pathology. Whether DAPK1 downregulation could be used to treat diseases mediated by p38-MK2 over-activation also deserves further investigation.

DAPK1 expression and knockout
Human DAPK1-specific siRNA (siDAPK1; UCU GGG AAG CGG AGC UGA AUU) and siRNA control (siCtrl) were purchased from GE Dharmacon (Lafayette, CO, USA). Dapk1 −/− mice (in a C57BL/6 background) were previously described 36 . Mice were maintained in the SPF mouse facility of the Institute of Molecular Biology, Academia Sinica. All mouse experiments were conducted with approval from the Institutional Animal Care & Utilization Committee, Academia Sinica.
All mice used in this study were 8-12-week old. The same sex (male or female) mice were used in the same experiment, but opposite sex mice could be used in the repeat of the given experiment. No difference was observed between male and female mice in the analyses conducted in this study. Experimental groups were assigned randomly. Five or more mice in each experimental group was planned, but four mice in some experimental groups, that have been examined in previous studies, were used due to the knockout-mice availability. No blinding was done because the readouts of the mouse experiments in this study were clear-cut (body weight loss, death). No mice were excluded from scoring.

Cell viability assay
Necroptosis was induced by pretreating cells with z-VAD for 0.5 h, followed by stimulation with AT406, BV6, LPS, TNF or IFNβ for the indicated periods. The necroptosis inhibitor Nec-1 was added 0.5 h prior to stimulation for certain experiments. Cell viability of BMDMs or BMDCs was assessed by measuring ATP levels upon adding an equal volume of Cell Titer-Glo reagent (Promega) and incubating for 30 min. Luminescence was determined using a Victor3 1420 Multilabel Counter (PerkinElmer, Shelton, CT). Alternatively, cell viability was determined via reduction of MTT by mitochondrial reductase into purple formazan. The intensity of colored product was measured by absorbance at 490 nm on an Emax microtiter plate reader (Molecular Device, Sunnyvale, CA). Necroptotic HT-29 cells were determined by staining with PI (10 μg/ml) in PBS, and PI + cells were analyzed using flow cytometry. Apoptosis of Jurkat cells was induced by treating with FLAG-FasL and assessed by staining with Annexin V-Cy5 (BD-Biosciences) and subsequent Annexin V + cell quantitation by flow cytometry.

Western blot and immunoprecipitation
For immunoblotting and immunoprecipitation, cells were lysed by whole-cell extract (WCE) buffer (25 mM HEPES pH 7.9, 300 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM DTT and 0.1% Triton X-100) on ice for 30 min. Supernatants were separated by centrifugation at 13,200 rpm for 10 min at 4°C. The protein concentrations were determined by Bio-Rad protein assay (#500-00006). For immunoprecipitation, 0.5 mg total cell lysates were incubated with 1 μg specific antibody, and the mixtures were rotated overnight at 4°C. The immune-complexes were captured by Protein G Mag Sepharose (GE, 28-9670-70, 10 μl per sample). The beads were washed three times with WCE buffer and denatured by 4X SDS sample buffer (200 mM Tris-HCl pH 6.8, 1.2 M β-mercaptoethanol, 40% glycerol, 8% SDS, and 0.4% bromophenol blue). For immunoblots, samples were denatured and analyzed by SDS-PAGE with running buffer (0.025 M Tris, 0.192 M glycine, and 0.1% SDS). The PAGE was transferred to PVDF membrane (Millipore) with transfer buffer (25 mM Tris, 192 mM glycine, and 20% methanol) at 400 mA for 100 min at 4°C. Membranes were blocked with SuperB-lock™ T20 (Sigma) to detect anti-phosphate antibody or blocking buffer (5% non-fat milk and 0.1% Tween-20 in a TBST buffer of 50 mM Tris-HCl pH 7.4 and 150 mM NaCl) at room temperature for 30 min, before being incubated with specific primary antibodies at indicated dilutions at 4°C overnight. The membranes were washed three times with wash buffer (0.1% Tween-20 in TBST buffer) at room temperature for 10 min before incubating them with horseradish peroxidase-conjugated secondary antibodies in blocking buffer at room temperature for 1 h. After washing, the membranes were developed with ECL Western blot detection reagents (Advansta, K-12045-D50), with signals detected by X-ray film (Fujifilm).

TNF-induced septic shock
Wild-type littermates or Dapk1-knockout C57BL/6 mice of 6-8 weeks and same sex were used for TNFinduced septic shock. No randomization was used due to the availability of knockout mice. Mice were anesthetized by Avertin (0.25 ml of 20 mg/ml, Sigma) and mouse TNF (1.0 μg/g) was intravenously administered in total volume of 200 μl endotoxin-free PBS. Body temperatures were monitored from rectal by industrial electric thermometer (Kane-may) for 30 h, with simultaneous recording of mice death. Mice were sacrificed when body temperature fell below 22°C. No blinding was done because the readouts of septic shock were straightforward (temperature drop, death). No mice were excluded from scoring.

Statistics
Data were randomly collected, but not blindly. We did not exclude any data from this study. Data met the assumptions of applied statistical tests (i.e. normal distributions). Microsoft Office Excel and Prism 5.0 (GraphPad software) were used for data analysis. Unpaired two-tailed Student's t tests were used to compare most of the viability or death results between two groups. Two-way ANOVA followed by a Tukey's multiple comparison tests, or two-way ANOVA followed by a Sidak's multiple comparison test were used to compare other results, as indicated in the respective figure legends. Long-rank (Mantel-Cox) test was used to compare the survival of mice. Data are presented as means with standard deviation (SD).