VU0155069 inhibits inflammasome activation independent of phospholipase D1 activity

The inflammasome is a specialized multiprotein oligomer that regulates IL-1β production. Although regulation of the inflammasome is related to crucial inflammatory disorders such as sepsis, pharmacological inhibitors that effectively inhibit inflammasome activity are limited. Here, we evaluated the effects of a phospholipase D1 (PLD1)-selective inhibitor (VU0155069) against sepsis and inflammasome activation. VU0155069 strongly enhances survival rate in cecal ligation and puncture (CLP)-induced sepsis by inhibiting lung inflammation, leukocyte apoptosis, and the production of proinflammatory cytokines, especially IL-1β. VU0155069 also significantly blocked IL-1β production, caspase-1 activation, and pyroptosis caused by several inflammasome-activating signals in the bone marrow-derived macrophages (BMDMs). However, VU0155069 did not affect LPS-induced activation of signaling molecules such as MAPK, Akt, NF-κB, and NLRP3 expression in the BMDMs. VU0155069 also failed to affect mitochondrial ROS generation and calcium increase caused by nigericin or ATP, and subsequent ASC oligomerization caused by several inflammasome-activating signals. VU0155069 indirectly inhibited caspase-1 activity caused by LPS + nigericin in BMDMs independent of PLD1 activity. We demonstrated that a PLD1 inhibitor, VU0155069, shows anti-septic activity as well as inflammasome-inhibiting effects. Our results suggest that VU0155069 can be considered a novel inflammasome inhibitor.

Phospholipase D (PLD) is an enzyme that hydrolyzes phosphatidylcholine into phosphatidic acid (PA) and choline 11 . Generated PA can regulate diverse signaling molecules including kinases, phosphatases and nucleotide binding proteins 12 . PLD involves the regulation of pathological progress of diverse diseases such as infection, cancer and some neurodegenerative disorders 13,14 . Although PLD, especially PLD2 has been known to mediate inflammation during bacterial infection 14 , it remains to be elucidated what is the role of PLD1 in inflammation and regulation of inflammasome. Previously, PLD isoform-selective inhibitors including a PLD1-selective inhibitor (VU0155069) were developed by Scott and colleagues 15 , which facilitates to analyze physiological and pathological function of PLD. In this study, we investigated the effects of a PLD1-selective inhibitor, VU0155069 on the regulation of inflammation and inflammasome.

Results
VU0155069 shows therapeutic effects against experimental polymicrobial sepsis. We investigated the effects of a PLD1 inhibitor (VU0155069) against a mouse model of sepsis. Administration of VU0155069 significantly increased survival rate in an experimental animal sepsis model, showing 77% survival comparing to 23% survival for the vehicle control group (Fig. 1A). Polymicrobial sepsis-induced mortality is associated with lung inflammation 14 . We also found that cecal ligation and puncture (CLP) surgery caused marked lung inflammation, showing increased lung congestion and thrombus formation. VU0155069 administration Figure 1. VU0155069 shows therapeutic effects against experimental polymicrobial sepsis. (A) C57BL/6 mice were subjected to CLP surgery. Vehicle (0.5% Tween 80 in PBS) or VU0155069 (10 mg/kg) was injected s.c. four times into mice 2, 14, 26, and 38 h after CLP. Survival was monitored for 10 days. **p < 0.01 by log-rank test. Sample size: n = 13 per group (A). (B-F) CLP mice were sacrificed 24 h after surgery. The lungs were stained with hematoxylin and eosin (magnification, ×200) (B left). Lungs were used to measure the wet/dry weight ratio 24 h after CLP (B right). The TUNEL assay was performed in the spleen and thymus from CLP mice (C left), and TUNEL-positive cells were quantified (C right). CFU were counted in peritoneal fluid from CLP mice (D). Inflammatory cytokines were measured in peritoneal fluid from CLP mice (E). Caspase-1 and IL-1β were detected in the spleen of CLP mice using immunohistochemistry (F). The data are representative of 3-5 mice per group from two independent experiments. Scale bar, 100 μm (B left, C left and F). Mean ± SEM (n = 3-5 for B right, C right, D,E). *p < 0.05, **p < 0.01, ***p < 0.001 by unpaired two-tailed t-test (E) or by one-way ANOVA followed by Tukey's multiple-comparisons test (B right, C right and D). strongly blocked this lung inflammation (Fig. 1B left). Lung wet-to-dry ratio was also increased by CLP, indicating that lung edema is generated, which was significantly decreased by VU0155069 (Fig. 1B right). Previously, widespread leukocyte apoptosis was known to be induced during the pathological progression of sepsis 16 . In our experiments, we also found that CLP operation induced cell death in the spleen and thymus. Administration of VU0155069 significantly decreased TUNEL-positive cells from the spleen and thymus (Fig. 1C). It has been known that septic mortality positively correlates with bacterial burdens 17 . Since we found that administration of VU0155069 significantly increased survival rate in CLP mice (Fig. 1A), we examined the effects of VU0155069 on bacterial burden. Administration of VU0155069 significantly decreased bacterial colony forming units in CLP mice (Fig. 1D).
During the pathological progression of sepsis, a cytokine storm is induced 18 . Measurement of the levels of inflammatory cytokines from the peritoneum showed that levels of several inflammatory cytokines (TNF-α, IL-1β, IL-6, and IL-22) were significantly decreased by VU0155069 administration in CLP mice (Fig. 1E). Among these inflammatory cytokines, IL-1β has been reported to play a crucial role in sepsis pathology 16 . Moreover, IL-1β is known to be processed by active caspase-1 in inflammatory conditions 4 . Immunohistochemical analysis showed that CLP operation elicited marked increase of caspase-1 and IL-1β from the spleen. Administration of VU0155069 almost completely decreased CLP-induced caspase-1 level and subsequent IL-1β level in the spleen (Fig. 1F).
Production of the inflammatory cytokine IL-1β was previously associated with pyroptosis, a specialized cell death that is mediated by activation of caspase-1 19 . We also observed that stimulation of BMDMs with inflammasome-activating stimuli (LPS + nigericin, LPS + MSU, LPS + dsDNA) increased cell death. Addition of VU0155069 significantly decreased cytotoxicity caused by the three different inflammasome-activating stimuli (Fig. 3). The results are consistent with our findings that VU0155069 significantly blocks inflammasome activation and subsequent IL-1β production caused by LPS + nigericin, LPS + MSU, or LPS + dsDNA.
VU0155069 does not affect LPS-stimulated signal 1, ASC oligomerization or membrane pore formation. For proper activation of the NLRP3 inflammasome and subsequent IL-1β production in macrophages, at least two different signals are required; signal 1 is induced by LPS to stimulate the signaling cascade including activation of MAPK and transcription factors such as NF-κB, and signal 2 is induced by particulates such as alum, uric acid crystals, etc. to induce lysosome damage, or by ATP to induce K + efflux 20 . Our finding that VU0155069 blocks NLRP3 inflammasome activation in BMDMs led us to test if VU0155069 affects signal 1 induced by LPS. Stimulation of BMDMs with LPS elicited the activation of MAPKs (ERK, p38 MAPK, JNK) and Akt. However, the addition of VU0155069 did not inhibit the activation of these kinases by LPS in BMDMs (Fig. 4A). The activation of STAT3, IκB degradation and nuclear translocation of NF-κB (markers for NF-κB activation) by LPS stimulation were also not inhibited by VU0155069 (Fig. 4A,B). One final outcome of signal 1 induced by LPS in macrophages is the upregulation of NLRP3 expression 20 . However, addition of VU0155069 did not inhibit the increased NLRP3 expression induced by LPS stimulation in BMDMs (Fig. 4C). The results suggest that VU0155069 does not inhibit LPS-induced signal 1.
Previous reports demonstrated that K + efflux and mitochondrial ROS generation are required for proper activation of the NLRP3 inflammasome complex 21,22 . Although it is controversial whether calcium influx is essential for the activation of NLRP3 inflammasomes by ATP 21 , calcium influx is also observed during the activation of NLRP3 inflammasomes by ATP. Here, we investigated if VU0155069 affect calcium influx caused by ATP, an important stimulus of signal 2. As shown in Fig. 5A, VU0155069 did not affect ATP-stimulated calcium influx in BMDMs. VU0155069 also failed to affect mitochondrial ROS generation caused by LPS + nigericin in BMDMs (Fig. 5B). ASC oligomerization is accompanied by NLRP3 inflammasome activation 23 . In this study, we also observed that stimulation of BMDMs with four different combinations of stimuli (LPS + ATP, LPS + nigericin, LPS + dsDNA, LPS + flagellin) elicited ASC oligomerization, which was detected by Western blot analysis using anti-ASC antibody. Addition of VU0155069 did not inhibit ASC oligomerization induced by the four different combinations of stimuli (Fig. 5C). Immuno-fluorescence staining experiments showed that stimulation of BMDMs with LPS + nigericin caused formation of ASC specks, which was not affected by VU0155069 (Fig. 5D). The results suggest that VU0155069 does not inhibit ASC oligomerization in response to LPS plus nigericin in BMDMs.
A previous report demonstrated that a pore-forming protein gasdermin regulates IL-1β secretion from living macrophages 24 . Since VU0155069 blocked IL-1β secretion caused by LPS + nigericin in BMDMs, we examined whether VU0155069 block gasdermin-dependent pore formation in BMDMs by PI staining. We found that stimulation of BMDMs with LPS + nigericin significantly increased PI + cells, suggesting gasdermin D is cleaved www.nature.com/scientificreports www.nature.com/scientificreports/ and pores are formed. Addition of VU0155069 did not decrease PI + cells caused by LPS + nigericin (Fig. 5E). The results suggest that VU0155069-induced inhibitory effects on IL-1β secretion is not mediated by blocking gasdermin-dependent pore formation.
VU0155069 indirectly inhibits caspase-1 activity. Next, we examined if VU0155069 affects caspase-1 activity. Stimulation of BMDMs with LPS + nigericin strongly increased caspase-1 activity. Addition of VU0155069 prior to the stimuli (LPS + nigericin) almost completely blocked caspase-1 activity (Fig. 6A). To further test if VU0155069 directly affects caspase-1 activity, we added VU0155069 to the supernatants harvested from LPS + nigericin-stimulated BMDMs. VU0155069 only slightly inhibited caspase-1 activity when added to the supernatants (Fig. 6B). As a positive control, we observed that addition of caspase-1 inhibitor to the BMDM supernatants stimulated by LPS + nigericin strongly inhibited caspase-1 activity (Fig. 6B). The results suggest that VU0155069 inhibit caspase-1 activity indirectly rather than directly.

PLD1 is not involved in inflammasome activation in BMDMs.
Our finding on the inhibitory effects of VU0155069 on inflammasome activation and subsequent IL-1β production in macrophages led us to examine the role of PLD1 in the process in BMDMs. At first, we verified that BMDMs generated from PLD1 −/− mice do not express PLD1 (Fig. 7A). BMDMs generated from PLD1 +/+ or PLD1 −/− mice were stimulated with several inflammasome-activating signals (LPS + ATP, LPS + nigericin, LPS + dsDNA, LPS + flagellin), and secreted IL-1β levels were measured. Unexpectedly, PLD1 deficiency did not affect IL-1β production induced by these inflammasome-activating signals (Fig. 7B). Moreover, VU0155069 still almost completely inhibited LPS + nigericin-induced IL-1β production in BMDMs generated from PLD1 −/− mice (Fig. 7C). The results suggest that the PLD1 protein itself is not involved in the regulation of inflammasome activation, and VU0155069-induced inhibitory effects on inflammasome activation is mediated in a PLD1-independent manner.

Discussion
In this study we found that administration of VU0155069, an inhibitor of PLD1, effectively induces therapeutic effects against experimental polymicrobial sepsis. The therapeutic effects of VU0155069 on sepsis are associated with decreased production of IL-1β, and VU0155069 strongly inhibited inflammasome activation and subsequent IL-1β production from BMDMs. On the mode of action of VU0155069 involved in inhibition of the inflammasome, we found that VU0155069 did not inhibit LPS-stimulated signal 1 or ASC oligomerization. However, VU0155069 effectively inhibited caspase-1 activity in response to LPS + nigericin, in an indirect manner. Based on the results obtained from this study, we suggest that VU0155069 is a novel regulator of the inflammasome and inflammatory cytokine IL-1β production.
Unlike other inflammatory cytokines such as TNF-α and IL-6, IL-1β production is regulated by using specialized machinery in form of the inflammasome, that requires at least two different signals (signal 1 and signal 2) 20 . Here, we observed that a PLD1 inhibitor VU0155069 strongly blocked inflammasome activation, IL-1β production, and pyroptosis from BMDMs (Figs 2 and 3). Through several experiments on the molecular mechanism involved in inhibition of the inflammasome by VU0155069, we observed that VU0155069 did not affect LPS-induced signal 1 (the activation of MAPK, Akt, and NF-κB), signal 2 caused by ATP (calcium influx), or ASC oligomerization and speck formation (Figs 4 and 5). However, VU0155069 strongly inhibited caspase-1 activity elicited by an inflammasome activating signal (LPS + nigericin) in BMDMs (Fig. 6A). In a separate experiment, we found that VU0155069 only slightly inhibited caspase-1 activity in vitro (Fig. 6B). VU0155069 inhibited caspase-1 activation and subsequent IL-1β production in response to not only NLRP3 but also AIM2 and NLRC4 (Fig. 2C,F). Although AIM2 and NLRC4 are distinct inflammasome to NLRP3, they are all controlled by caspase-1 activity 5 . Our results suggest that VU0155069 regulates the activity of inflammasome by indirect controlling caspasee-1, and VU0155069 can be used to inhibit these three types of inflammasome. Collectively, our results suggest that VU0155069 may act downstream of ASC oligomerization to induce inhibitory effects on the inflammasome, but not directly on caspase-1. Our findings on the action mode of VU0155069 additionally suggest that an unrevealed regulatory mechanism leading to caspase-1 activation may exist downstream of ASC oligomerization.
Although VU0155069 has been reported to specifically inhibit PLD1 in a previous report 15 and VU0155069 strongly inhibited inflammasome activation in BMDMs, we observed that PLD1 deficiency in BMDMs did not affect inflammasome activation (Fig. 7B). Moreover, inflammasome-dependent IL-1β production was also almost completely inhibited by VU0155069 from PLD1-deficient BMDMs (Fig. 7C). Our findings suggest that VU0155069 may have a different target from PLD1, that plays a role in regulation of the inflammasome. Since www.nature.com/scientificreports www.nature.com/scientificreports/ regulation of inflammasome activation is important in the regulation of diverse inflammatory disorders and VU0155069 shows strong inhibitory effects on inflammasome activation independent of PLD1, identification of the target molecule of VU0155069 should be important.
The inflammasome and inflammasome-activated IL-1β production are associated with diverse diseases such as Crohn's disease and gout 5 . Since we found that VU0155069 strongly inhibited IL-1β production induced by inflammasome-inducing signals by inhibiting caspase-1 activity, VU0155069 can be used to control inflammasome-associated diseases. In this study, we also observed that VU0155069 administration strongly induced therapeutic effects against polymicrobial sepsis, whose pathological progress is also mediated by increased IL-1β. Since we found that VU0155069 showed inhibitory effects on activation of the inflammasome with an unrevealed molecular mechanism to block caspase-1 activation, future studies on the mode of action of VU0155069 would provide novel insights to control inflammasome and inflammasome-related diseases. Thus, VU0155069 can be regarded as a new candidate to control inflammasome-associated disorders.   Lung histology and TUNEL assay. Vehicle or VU0155069-injected CLP mice were sacrificed at 24 h after surgery. Their lungs were fixed in neutral buffered formalin, sectioned with paraffin, and stained with hematoxylin and eosin for medical diagnosis. Isolated spleens and thymus were used to make frozen sections for the TUNEL assay (Roche, Basel, Switzerland). The sections were placed in Triton X-100 at 4 °C for 2 min for permeation and incubated with TUNEL reagent at 37 °C for 60 min. The percentage of TUNEL-positive cells was counted under a light microscope.

Methods
Quantification of pulmonary edema. Vehicle or VU0155069-injected CLP mice were sacrificed at 24 h after CLP. Whole lungs were harvested and then weighed. Afterwards, the lungs were placed in a 60 °C oven for 48 h. The dry lungs were weighed, and the wet/dry weight ratio was calculated.
Measurement of CFUs in CLP model. Vehicle or VU0155069-injected CLP mice were sacrificed at 24 h after surgery. Peritoneal fluid was harvested and cultured overnight on Trypticase soy agar plates at 37 °C incubator. CFUs were determined as described previously 14 .
Generation of mouse BMDMs. Mouse BMDMs were generated according to a previous report 25 . Briefly, femur and tibia were obtained from 5 week old mice. After flushing the bone marrow cells from the femur and tibia, red blood cells were lysed using ACK lysis buffer (Gibco, Waltham, MA, USA). The bone marrow cells were cultured in α-MEM containing 10% FBS for one day, and non-adherent cells were further cultured for three days with 30 ng/ml M-CSF (Peprotech, Rocky Hill, NJ, USA).

Measurement of mitochondrial ROS.
Confocal microscopy. Poly-L-lysin was coated on a slide glass overnight, and isolated BMDMs were placed on the slide. After washing with Opti MEM, 1 μg/ml of LPS was added for 4 h, and subsequently nigericin was added for 1 h, and then the samples were fixed with 4% paraformaldehyde for 20 min at room temperature. Next, permeabilization with 0.2% Triton X-100 was performed. After blocking for 1 h with 1% BSA in PBS, samples were stained with ASC antibody for 3 h and then stained for 1 h with anti-mouse Alexa Fluor 488. After staining the nuclei with Hoechst for 5 min, samples were observed with confocal microscopy, and photo images were analyzed with ZEN software.