Death-associated protein kinase 1 mediates interleukin-1β production through regulating inlfammasome activation in Bv2 microglial cells and mice

Interleukin-1β (IL-1β) plays a crucial role in mediating inflammation and innate immunity response in the central nervous system. Death-associated protein kinase 1 (DAPK1) was shown to be involved in several cellular processes. Here, we investigated the effects of DAPK1 on IL-1β production in microglial cells. We used a combination of in vitro (Bv2 microglial cell cultures) and in vivo (mice injected with amyloid-β (Aβ)) techniques to address the role of caspase-1 activation in release of IL-1β. DAPK1 involvement was postulated through genetic approaches and pharmacological blockade of this enzyme. We found that Aβ25–35 stimulation induced IL-1β production and caspase-1 activation in LPS-primed Bv2 cells and mice. DAPK1 knockdown and catalytic activity inhibition reduced IL-1β maturation and caspase-1 activation, nevertheless, DAPK1 overexpression attenuated these effects. Aβ25–35-induced lysosomal cathepsin B leakage was required for DAPK1 activation. Furthermore, repeated DAPK1 inhibitor treatment ameliorated the memory impairment in Aβ25–35-injected mice. Taken together, our findings suggest that DAPK1 facilitates Aβ25–35-induced IL-1β production through regulating caspase-1 activation in microglial cells.


Cathepsin B leakage is required for Aβ 25-35 -induced IL-1β production and caspase-1 activation in LPS-primed Bv2 cells.
To directly investigate the changes in lysosome membrane permeability in live microglia, we first stained cells with the acidity-dependent acridine orange (AO), which accumulates in acidic compartments such as intact lysosomes with red fluorescence and displays green staining in the less acidic environment including the cytoplasm and nucleus. The disruption of lysosomal integrity is visualized by decreased red staining and increased green fluorescence 24 . As shown in Fig. 2A, non-primed cells and LPS-primed cells contained red puncta signals for AO, whereas Aβ 25-35 treatment turned fluorescent signals largely green. We also stained cells with a cell-permeable, fluorescent cathepsin B substrate. As expected, we observed a marked increase of activated cathepsin B leakage into the cytoplasm after Aβ 25-35 exposure in LPS-primed cells ( Fig. 2A).
Next, we treated cells with CA-074Me, a specific inhibitor of cathepsin B. As shown in Fig. 2B-C, CA-074Me significantly lowered Aβ 25-35 -induced IL-1β production in a dose-dependent manner and blocked cleavage of caspase-1 in LPS-primed Bv2 cells. DAPK1 is required for Aβ25-35-induced IL-1β production and caspase-1 activation in LPS-primed Bv2 cells. To identify the role of DAPK1 in the regulation of IL-β production and NLRP3 inflammasome activation in Bv2 cells, an shRNA-mediated DAPK1 deletion experiment and DAPK1 ΔCaM , a constitutively active form of truncated DAPK1 construct (here referred to as cDAPK1), expression experiment was performed. As shown in Fig. 3A, DAPK1 expression was dramatically decreased in DAPK1 knockdown cells, and significantly increased in cDAPK1-expressing cells. As expected, p-MLC levels were remarkably up-regulated in cDAPK1-expressing cells.
Simultaneously, we observed that Aβ 25-35 -induced secretion of IL-1β and cleavage of caspase-1 was significantly reduced in DAPK1 knockdown cells and increased in cDAPK1-expressing cells (Fig. 3B,F,G). As a control, the secretion of other pro-inflammatory cytokines, tumor necrosis factor-α (TNF-α) and IL-6, was nearly identical between control and DAPK1 knockdown cells activated by LPS (Fig. 3C-D). Lactate dehydrogenase (LDH) assays were used to evaluate cell viability, and the results showed that DAPK1 knockdown did not affect LDH release, either in the presence or absence of stimulation (Fig. 3E). Notably, the levels of NLRP3 components (NLRP3, ASC, pro-caspase-1) and pro-IL-1β were comparable among all groups (Fig. 3F,H-J).

DAPK1 activation is dependent on cathepsin B.
Considering that DAPK1 interacts with cathepsin B 26 , of further interest was whether cathepsin B mediated the process of DAPK1-involved caspase-1 activation. To investigate whether DAPK1 activation is dependent on cathepsin B release, the effect of CA-074Me on DAPK1 activity was tested in LPS-primed cells. The results showed that CA-074Me treatment had no effect on DAPK1 expression, but abolished the increase in the protein levels of p-MLC induced by Aβ [25][26][27][28][29][30][31][32][33][34][35] (Fig. 5A), which is indicative of less DAPK1 activation. In addition, overexpression of DAPK1 restored Aβ 25-35 -induced IL-1β release in presence of CA-074Me in LPS-primed cells (Fig. 5B). On the contrary, DAPK1 inhibitor treatment had no influence on the expression (Fig. 5C) and leakage of cathepsin B (Fig. 5D).

Acute DAPK1 inhibitor treatment attenuates Aβ 25-35 -induced IL-1β production in mice.
To further test whether the essential role of DAPK1 in caspase-1/IL-1β signaling was also observable in vivo, an Aβ 25-35 injection rodent model was adopted. The experimental scheme of animal treatment and neurochemical analyses is explained in Fig. 6A. Increased microglia marker Iba1-/NLRP3-positive cells were observed in the CA1 and CA3 areas of hippocampus from Aβ 25-35 -injected mice (Fig. 6B). As shown in Fig. 6C-D, compared

Subchronic DAPK1 inhibitor treatment improves Aβ 25-35 -induced memory deficits in mice.
The time chart of animal treatment and behavioral assessments is explained in Fig. 7A. Compared with the sham mice, Aβ 25-35 injection decreased the recognition index ( Fig. 7B), thereby implying a cognitive impairment. Subchronic DAPK1 inhibitor treatment significantly increased the recognition index in Aβ 25-35 -injected mice. Furthermore, it is important to emphasize that the deficit in the ORT observed in Aβ 25-35 -injected mice cannot be explained by a reduced locomotor activity, as there was no major difference observed among groups in the total distances travelled in the open-field (Fig. 7C).
Next, we studied behavior in FCTs with respect to context-and cue-retention. Aβ 25-35 -injected mice demonstrated significantly less contextual-and cue-freezing responses compared with the sham mice (Fig. 7D,E), suggesting an impairment of associative memory. DAPK1 inhibitor administration significantly attenuated the impairment of contextual and cue freezing responses in Aβ 25-35 -injected mice.
Aβ could activate microglia and initiate the release of various neurotoxic inflammatory cytokines and chemokines 22,27,28 . We found that Aβ [25][26][27][28][29][30][31][32][33][34][35] which retains most of the neurotoxic properties of the full-length Aβ was able to induce robust IL-1β production and caspase-1 activation in LPS-primed cells in our current study 7,29 . This is consistent with a recent study which reported Aβ 25-35 peptide triggered IL-1β secretion through activating the NLRP3 inflammasome in microglia 30 . Lysosomal protease cathepsin B has been implicated in the activation of the inflammasome induced by multiple particulate stimuli 31,32 . We here observed an increase in the lysosomal membrane permeabilization after stimulating LPS-primed Bv2 cells with Aβ [25][26][27][28][29][30][31][32][33][34][35] , which in turn caused the leakage of cathepsin B into the cytosol. This process seems to be responsible for the initiation of Aβ 25-35 -induced activation of the caspase-1/IL-1β pathway since we found that IL-1β secretion and caspase-1 activation were decreased in the presence of cathepsin B inhibitor.  Supplementary Fig. S4). Data are expressed as mean ± SEM for at least three independent experiments. *** P < 0.001; # P < 0.05, ### P < 0.001. DI: DAPK1 inhibitor. Recent studies indicated a role of DAPK1 in several cellular processes and DAPK1 exerts its effects through an increase in its activity 33,34 . In the experimental setting here, LPS priming and Aβ 25-35 stimulation led to a significant increase of DAPK1 activity in Bv2 cells, as evidenced by increased levels of p-MLC.
Prior studies revealed the involvement of DAPK in inflammatory responses. It has been demonstrated that DAPK1 negatively regulated the activation of the TNF-α and IFN-γ-stimulated NF-κB signaling 35 . However, DAPK promoted the assembly of the NLRP3 inflammasome in macrophages 18 . To explore a mechanistic insight into the role of DAPK1 in IL-1β production in microglia, we examined the effects of DAPK1 knockdown and overexpression, as well as DAPK1 catalytic activity inhibition on Aβ 25-35 -induced caspase-1 activation. We found that DAPK1 silencing and activity inhibition largely abolished Aβ 25-35 -induced IL-1β secretion and caspase-1 cleavage, while cDAPK1 attenuated these effects. We also observed that the expression of NLRP3, ASC, pro-caspase-1 and pro-IL-1β was not changed in the presence of DAPK1 knockdown or DAPK1 overexpression before and after the treatment of Aβ [25][26][27][28][29][30][31][32][33][34][35] in LPS-primed Bv2 cells. DAPK1 knockdown and activity inhibition did not affect TNF-α and IL-6 release, as well as microglial viability. It appears that DAPK1 had little effect on the NF-κB-dependent priming stage of NLRP3 inflammasome activation. These results suggest that DAPK1 is a potent enhancer in the stage of caspase-1 activation, and are in keeping with a previous study in macrophages 18 .
In addition, DAPK1 was reported to be regulated by the lysosome. It is suggested that cathepsin B directly bind to DAPK1 in the TNFR-1-induced apoptosis, and that its deficiency increase the steady-state levels of DAPK1 26,36 . The region consisting of amino acids 836-947 in the DAPK1 was suggested to be essential for the binding of cathepsin B. Here, we observed a possibly telling connection between cathepsin B and DAPK1 by showing that cathepsin B inhibitor treatment dampened DAPK1 activation, while DAPK1 inhibitor had no effect on cathepsin B expression and release, which places cathepsin B upstream of DAPK1 activation. Our findings thus suggest a possibility that the cathepsin B-mediated increase in DAPK1 activation ensures that there are sufficient amounts of activated DAPK1 to participate in the activation of caspase-1 that will ultimately lead to IL-1β maturation.
Transfer of sensory information to the hippocampus mainly through the hippocampal trisynaptic circuitry (entorhinal cortex → dentate gyrus → CA3 → CA1 pathway) 37 . Despite the underlying neural circuitry supporting object recognition has not been clearly defined, there is some evidence so far to support the significant role of the hippocampus in the formation of objective memory [38][39][40] . ORT test session performance activated gene expression in the hippocampus (CA1 and CA3 sub-regions) and increased CA1 neurons firing rates 38,41 . More recent studies suggest that the amygdala and hippocampus operate in parallel during fear conditioning 42,43 . Hippocampal sub-regions CA1 and CA3 project to the amygdala via relays in the entorhinal cortex or through the ventroangular pathway directly 44 . A single intracerebroventricular (i.c.v.) administration of Aβ 25-35 into the rodent brain provides a useful model to study neuroinflammation and behavior features 45 . NLRP3 inflammasome is primarily expressed in microglia in the brain 30,46 . Here, we observed increases in the Iba1 immunoreactivity and NLRP3 expression in the stratum radium of hippocampal CA1 and CA3 areas of Aβ 25-35 -injected mice, whereas there are no obvious changes in the pyramidal layer. Of particular importance, high levels of IL-1β have negative impacts on the processes of hippocampal long-term potentiation and synaptic plasticity [47][48][49] , which ultimately results in a decline in the cognitive performance. Consistent with the previous reports 46,50 , the i.c.v. injection of Aβ 25-35 caused functional deficits in ORT and FCTs in our study. However, DAPK1 inhibitor significantly prevented Aβ 25-35 -evoked caspase-1 activation and IL-1β production, and rescued the decline in the object recognition and fear memory. Due to the fact that the inflammasome activation is closely associated with the neuronal injury 22,51 , we speculated that the neuroprotective effects of DAPK1 inhibitor, at least in part, may be related to its anti-inflammasome/caspase-1 activation effects.
Sufficient protein levels of NLRP3 are necessary to the formation and activation of the NLRP3 inflammasome. We observed an increase in NLRP3 expression in vivo and no alterations in LPS-primed Bv2 cells in response to Aβ 25-35. Aβ injection results in a pro-inflammatory milieu in rodent brain and activation of the NF-κB signaling pathway 52 . In Bv2 cells, however, 6 h of LPS priming had robustly activated the expression of NLRP3 through the NF-κB pathway and following Aβ 25-35 stimulation might mainly played a role in NLRP3 activation 18,31 . Based on the findings of our present study, it is plausible that DAPK1 functions as an endogenous enhancer for Aβ 25-35 -induced IL-1β production through regulation of caspase-1 activation in microglia. Considering the crucial role of microglial inflammasome/caspase-1/IL-1β axis in neuroinflammation 8,53 , we propose that the inhibition of DAPK1 represents a potential therapeutic application for IL-1β-associated neurological diseases.

DAPK1 knockdown and overexpression.
To generate the stable DAPK1 knockdown cell line, lentiviral particles encoding DAPK1-specific shRNA or scrambled shRNA (Santa Cruz Biotechnology, CA, USA) were used to infect the Bv2 cells and polybrene (5 μg/ml) was added to the culture medium. After 6 h, the medium was replaced with DMEM containing 10% FBS. Two days later, puromycin (5 μg/ml) was added to the medium for selection. DAPK1 knockdown efficiency was confirmed by western blotting.
The DAPK1 ΔCaM mutant (here referred to as cDAPK1) which lacks a calcium/calmodulin regulatory domain was received as a gift from Prof. Lu  The contents of LDH in samples of supernatants were quatitated by an LDH cytotoxicity assay (Nanjing Jiancheng, Nanjing, China) to evaluate the impact of different treatments on cell viability.
Protein extraction and western blotting. Stimulated cells and tissue samples were harvested for protein extraction and western blotting analysis with standard protocols as previously described 12,56 . Cytoplasmic fractionations were performed with NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, Rockford, IL, USA) according to the manufacturer' s protocol. Briefly, cells were harvested, washed and pelleted. Then added ice-cold CER I to the cell pellet and incubated the tube for 10 min on the ice. After that, added ice-cold CER II to the tube and incubated for 1 min on the ice. The volume ratios of cell pellets, CER I and CER II are 10: 100: 5.5. At last, centrifuged the tube for 5 min at 16, 000 × g and transferred the supernatant to a clean pre-chilled tube. Stored this tube in −80 °C until use. The following antibodies were used: anti-GAPDH (1:10000 Preparation of animal model. Mice were anesthetized with intraperitoneal (i.p.) admininistration of 2% pentobarbitone sodium (0.06 ml/10 g) and placed in a stereotactic apparatus (RWD, Shenzhen, China). Body temperature was maintained at 37 °C using a heating pad. Aβ [25][26][27][28][29][30][31][32][33][34][35] (3 nmol/mouse at 3 μl) was administered intracerebroventricularly (i.c.v.) using a microsyringe with a 28-gauge stainless-steel needle at a rate of 0.5 μl/min 57 . The injection site was confirmed by the injection of Evans blue dye (2%) in the preliminary experiments (see Supplementary Fig. S7). The mice i.c.v. injected with an equal volume of PBS served as the sham group. Before skull was exposed, the lidocaine cream was local administrated to each mouse to prevent pain.
For the acute treatment of DAPK1 inhibitor, mice were received a single i.c.v. injection of DAPK1 inhibitor (2.5 or 5 nmol/ mouse at 2 μl) 1 h before Aβ [25][26][27][28][29][30][31][32][33][34][35]  Immunofluorescence analysis. Mice were anesthetized with pentobarbitone sodium (0.06 ml/10 g, i.p.) and perfused with 4% paraformaldehyde at 48 h after surgery. Brains were post-fixed in paraformaldehyde for 24 h, cryoprotected with 30% sucrose for 48 h, embeded into OCT compound (Torrance, CA, USA) and frozen at −80 °C overnight. Coronal sections (10 μm) including the hippocampus were prepared uisng a cryostat and mounted on precoated glass slides. Brain sections were blocked with 10% normal goat serum in PBS containing 0.3% Triton X-100 and incubated overnight at 4 °C with primary antibodies [NLRP3, 1:300; Iba1 (Wako, Japan), 1:300]. After washing, sections were incubated with IFKine Green AffiniPure Donkey Anti-Rabbit IgG and Dylight 549 Goat Anti-Mouse IgG (both from Abbkine, Redlands, CA, USA) (1:200) for 60 min at 37 °C and counterstained with DAPI (Roche, Mannheim, Germany) for 10 min. Subsequently, confocal images were acquired using a Nikon A1 confocal laser scanning microscope (Nikon, Japan). The numbers of Iba1-/ NLRP3-double-positive cells were calculated in 10 coronal sections corresponding to the hippocampal CA1 and CA3 regions with the software Image-Pro Plus. Briefly, the images were analyzed by setting a threshold for all sections of a specific labeling. The stained area above the threshold was determined for each section. The co-localization of Iba1 and NLRP3 was then determined as the co-stained area and counted.

Behavior testing. Open-field test.
The open-field test (OFT) was used to measure the locomotor activity of the animals on day 7 after Aβ 25-35 injection. A detailed description of this method is provided in Supplementary Methods.
Object recognition test. The object recognition test (ORT) was performed on the days 7-8 after Aβ [25][26][27][28][29][30][31][32][33][34][35] injection in accordance with a standardized protocol described by Legar M and colleagues with some modifications 58  Fear conditioning test. On the days 9-10 after peptide injection, fear conditioning tests (FCTs) were carried out as detailed in a previous report 46 . A detailed description of this method is provided in Supplementary Methods.

Statistical analysis.
All variance values were represented as the mean ± SEM. When assumptions of normality and equal variance were met, group comparisons were evaluated using unpaired two-tailed Student's t-tests or one-way ANOVA followed by Bonferroni tests. Statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software Inc., San Diego, CA, USA). P < 0.05 was considered statistically significant.