Abstract
In advanced human plaques and aged patients, athero-protective pathways such as autophagy and reverse cholesterol transport (RCT) become dysfunctional, while atherogenic pathways such as NLRP3 inflammasome and TLR2/4 are induced. Here, we report that Miltefosine, an FDA approved drug for treating leishmaniasis, increased cholesterol efflux, induced membrane remodeling, and induced autophagy in macrophages. Macrophages treated with Miltefosine exhibited markedly increased ABCA1 mediated cholesterol efflux and decreased phosphatidylserine flip from the cell-surface. Miltefosine treatment of macrophages led to redistribution of phosphatidylinositol 4,5-bisphosphate (PIP2) from plasma membrane to actin rich regions of the cell. RAW264.7 macrophages treated with Miltefosine showed marked increase in p62 and LC3 puncta staining vs. control cells. The Lipid droplet degradation was induced by Miltefosine leading to ~ 50% decrease in the CE:FC (cholesterol ester: free cholesterol) ratio. The TLR4 signaling pathway in LPS primed bone marrow derived macrophages was blunted by Miltefosine treatment, leading to ~75% reduction in pro-IL-1β mRNA levels. Miltefosine pretreatment of macrophages potently inhibited NLRP3 inflammasome assembly induced by LPS/ATP treatment, exhibiting ~70% reduction in ASC1 speck forming cells. Gasdermin D mediated release of mature IL-1β was reduced by ~80% in Miltefosine treated vs. control cells. The qRT-PCR and western blot analysis showed no changes in basal or LPS induced levels of inflammasome components (NLRP3, ASC1, procaspase1), while the LPS mediated induction in ROS levels was significantly blunted in Miltefosine treated vs. control macrophages. Miltefosine did not alter the AIM2 inflammasome activity indicating specific targeting of Nlrp3 inflammasome pathway. Overall, this study showed that Miltefosine targets lipid trafficking, cell migration, autophagy, and Nlrp3 inflammasome activity in macrophages.
- Non-standard abbreviations and acronyms
- AMPK,
- AMP-activated protein kinase
- apoA1,
- apolipoprotein A-I
- apoE,
- apolipoprotein E
- ABCA1,
- ATP-binding cassette transporter A1
- ASC1,
- apoptosis-associated speck protein containing a CARD (Caspase activation and recruitment domain)
- DMPC,
- 1,2-dimyristoyl-sn-glycero-3-phosphatidylcholine
- Gsdmd;
- Gasdermin D
- HDL-C,
- high-density lipoprotein-cholesterol
- IL1-β,
- Interleukin 1-beta
- LXR,
- liver X receptor
- NLRP3,
- NOD-like receptor family pyrin domain-containing 3
- NF-kB,
- nuclear factor kappa-light-chain-enhancer of activated B cells
- PS,
- phosphatidylserine
- PC,
- phosphatidylcholine
- POPS,
- 1-palmitoyl-2-oleoyl-sn-glycero-3-serine
- SR-BI,
- scavenger receptor-BI
- TLR,
- toll like receptor
Significance Statement Atherosclerosis is a progressive inflammatory disease. In advanced human plaques, the athero-protective pathways such as reverse cholesterol transport (RCT) and autophagy become increasingly dysfunctional while atherogenic pathways such as NLRP3 inflammasome are aberrantly induced. The cholesterol efflux via RCT prevents atherosclerosis and inflammation by reducing lipid loads in foam cells. Autophagy, in addition to playing a pivotal role in removing stored cholesterol from macrophages, promotes removal of inflammasome activating stimuli from cytosol (such as damaged mitochondria), thus helping in sequestering IL-1β. Failure of foam cells to egress from plaques and prolific engulfment of modified lipids results in formation of cholesterol crystals, leading to induction of NLRP3 inflammasome assembly. The release of IL-1β from foam cells further amplify the inflammation and promote further infiltration of immune cells to plaque. How RCT, autophagy, and inflammation pathways coordinate with each other to maintain cellular homeostasis is not clear; thus, the interplay between these pathways needs to be investigated thoroughly. Increased cholesterol efflux and induced autophagy with simultaneous dampening of Nlrp3 inflammasome can prevent atherosclerosis. Here, we show that Miltefosine can target multiple pathways involved in lipid homeostasis and inflammation. The detailed investigation of mechanisms involved in Miltefosine’s action may led to novel therapeutic targets for preventing and treating atherosclerosis and CVD.
Introduction
Based on the American Heart Association’s 2016 Heart Disease and Stroke Statistics update; heart disease and stroke were the No. 1 and No. 2 killer worldwide23. Atherosclerosis, a sterile inflammatory disease, is the driving force behind progression of Coronary artery disease (CAD) and cardiovascular disease (CVD). The excess levels of low-density lipoproteins (LDLs) in blood can be deposited and modified in arterial intima, where macrophages can uptake these lipids. The continuous uptake of toxic lipids and failure of macrophages to regress results in foam cell formation. The foam cells over time can become dysfunctional and undergo necrosis, leading to increased recruitment of immune cells to plaque area1. Accumulation of cholesterol crystals and oxidized lipids in plaques can also activate NLRP3 inflammasome to further promote atherosclerosis2, 3. The NLRP3 inflammasome is reported to be activated in aortas of advanced atherosclerosis patients as well as in other metabolic diseases such as obesity4, 5, 6. The NLRP3 inflammasome plays a role in processing procaspase 1 resulting in subsequent caspase 1 mediated processing of pro IL-1β to generate active Interleukin-1β (IL-1β). The role of IL-1β in promoting human CVD was highlighted by the recently concluded Cantos trial, showing that anti-IL-1β therapy met the primary trial endpoint, a reduction in a composite of heart attack, stroke and cardiovascular death7. A newly discovered substrate of caspase1 and caspase 11, Gasdermin D, plays an essential role in pyroptosis (necrotic cell death) and IL-1β release 8–10. Gasdermin D is also necessary for IL-1β release following LPS and nigericin treatment under conditions that prevented cell membrane rupture11. Modified cholesterol promotes recruitment of Toll-like receptors 2 and 4 (TLR2/TLR4) to plasma membrane of foam cells, leading to activation of NF-kB signaling and increased synthesis of pro IL-1β mRNA transcript. Circulating monocytes from human patients with atherosclerosis exhibit increased expression of TLR2/4 and Apo E-/- null mice carrying large plaque burden displayed increased surface expression of TLR2 and TLR4 on circulating monocytes12. The induction of TLR (for production of IL-1β mRNA) and NLRP3 (for processing and release of IL-1β) in advanced atherosclerotic plaques can lead to amplification of immune responses and plaque instability. In contrast to atherogenic pathways, the atheroprotective pathways such as autophagy and cholesterol efflux become increasingly dysfunctional in advanced atherosclerotic plaques, aged patients, and in arteries of atherosclerosis and diabetes animal models3, 13, 14. During lipid-droplet autophagy, lipid droplets are engulfed by autophagosomes that then fuse with lysosomes where cholesterol esters are hydrolyzed to generate free cholesterol13, 14. Many lines of evidence have shown that inhibition of autophagy by drugs or genetic methods is associated with increased atherosclerotic plaque burden3, 13, 15. Once the free cholesterol is generated by lipid-droplet autophagy it can be transported back to liver for excretion via reverse cholesterol transport (RCT) 16–20. In addition to removing cholesterol, Abca1 expression can also disrupt lipid-rafts and dampen TLR recruitment to cell membrane21. Thus, the simultaneous induction of athero-protective pathways, such as cholesterol efflux and autophagy, along with dampening of athero-promotive pathways such as TLR signaling and Nlrp3 inflammasome assembly, may serve as a novel therapeutic treatment for CVD patients. Here, we report that Miltefosine, an FDA approved drug for treating visceral and cutaneous leishmaniasis, promoted cholesterol efflux, disrupted lipid-rafts and TLR4 signaling, increased cell-surface PS exposure, induced PIP2 trafficking from PM to cell periphery, induced basal autophagy, reduced LPS induced mitochondrial ROS generation and blunted NLRP3 inflammasome assembly and IL-1β release.
Results
Miltefosine induced Abca1 mediated cholesterol efflux from cells
Previous study have shown that Miltefosine and other alkylphospholipids induces cholesterol efflux from the HepG2 cells22. We tested effect of Miltefosine treatment on cholesterol efflux in RAW 264.7 macrophages. The RAW264.7 macrophages were radiolabeled with tritium cholesterol and induced with 8Br-cAMP to induce Abca1 expression. The control and ABCA1 expressing cells were incubated with different doses of Miltefosine for 1h or 4h and cholesterol efflux was determined. As shown in Fig. 1A, the Miltefosine treatment of control RAW cells led to ~1.45% increase (with 7.5μM Miltefosine) in cholesterol efflux to media (with ANOVA, mean± SD, n=6, p<0.05) and ~3.1% increase with 10μM Miltefosine (with ANOVA, mean± SD, n=6, p<0.005). The RAW cells induced for Abca1 expression showed much more robust cholesterol efflux to media in presence of Miltefosine with ~5.0 % increase with 7.5μM Miltefosine treatment (with ANOVA, mean± SD, n=6, p<0.05) and ~10 % increase in cholesterol efflux to media with 10 μM Miltefosine treatment (with ANOVA, mean± SD, n=6, p<0.005). To test the specificity of Miltefosine in inducing Abca1 mediated cholesterol efflux, we used Abca1 expressing non-macrophage HEK293 and BHK cell lines. In both cases, addition of Miltefosine markedly increased Abca1 mediated cholesterol efflux to the media in absence of any cholesterol acceptor (Fig. S1, S2). During 1h chase period with Miltefosine, the addition of cholesterol acceptor apoA1 further increased cholesterol efflux with Abca1+apoA1+7.5 μM Miltefosine treated cells vs. Abca1+apoA1 cells showing~4.8 % increase (with ANOVA, mean± SD, n=6, p<0.005) and Abca1+apoA1+10 μM Miltefosine treated cells vs. Abca1+apoA1 cells showing ~6.45 % increase in cholesterol efflux (with ANOVA, mean± SD, n=6, p<0.005). (Fig 1A). At higher concentration of 10μM Miltefosine and during 6h chase time period, the Abca1 mediated cholesterol efflux to media was more than efflux to apoA1 (Fig 1A, S3), indicating that Miltefosine may promotes ABCA1 mediated cholesterol efflux via microparticle generation in an acceptor independent manner. To determine the mechanism by which Miltefosine increased cholesterol efflux via Abca1, we performed western blot analysis of the total and cell-surface Abca1 levels. RAW264.7 macrophages were treated cAMP to induce Abca1 expression, followed by treatment with ± 7.5 μM Miltefosine for 4h and cell extract preparation for total and cell-surface protein fractions and western blot analysis. As shown in Fig. 1B, no significant increase was observed in either total or cell-surface Abca1 expression levels in control vs. Miltefosine treated cells, indicating that Miltefosine induced cholesterol efflux is not due to increased Abca1 levels. We also tested if Miltefosine affected apoA1 cellular binding by performing a flow-cytometry based assay using Alexa-647 labeled apoA1 as described in earlier study23. RAW cells ± ABCA1 expression were incubated with Alexa-647 labeled apoA1 for 45 min at 37°C ± 7.5μM Miltefosine. As expected, Abca1 expressing cells showed increased Alexa-647 apoA1 binding, while no significant differences were observed in apoA1 cell membrane binding in Miltefosine treated vs. non treated RAW264.7 macrophages (Fig. 1C). Next, we tested if Miltefosine modulated apoA1 mediated lipid-solubilization. ApoA1 in lipid-free form can spontaneously interact with DMPC liposomes and solubilize them, resulting in decreased turbidity of the liposome suspension, and generation of reconstituted HDL (rHDL). We performed apoA1 mediated liposome clearance assays at room temperature, using liposomes constituted of either DMPC alone or DMPC liposomes containing 5% Miltefosine. As shown in Fig. 1D, Miltefosine markedly enhanced the DMPC clearance by lipid-free apoA1. To further probe the role of Miltefosine in mediating enhanced cholesterol extractability from membranes that are more physiological and have similar lipid profile as plasma membrane, we performed lipid-solubilization assays using liposomes comprised of POPC, cholesterol and, PS. The POPC:PS:cholesterol MLVs were previously shown to be slowly solubilized by apoA1 at acidic pH 24. The addition of 5% Miltefosine to POPC:Cho:PS liposomes further increased rate of liposome clearance by apoA1 at lower pH (Fig 1E). Next, we tested Miltefosine analogues as well as lyso-PC for apoA1 mediated lipid solubilization. Perifosine and edelfosine promoted apoA1 mediated lipid solubilization at rate either equal or higher than Miltefosine while the lyso-PC also increased lipid solubilization by apoA1 but at lower rate than Miltefosine (Fig S4). These data indicated that Miltefosine promoted structural changes in lipid membrane that may facilitate the apoA1 mediated liposome solubilization.
Miltefosine increased PS exposure by inhibiting PS flip in macrophages
Lipid rafts are comprised mainly of PC, cholesterol and sphingomyelin25. Disruption of lipid-rafts by Abca1 is proposed to provide free cholesterol pool for efflux and hamper TLR signaling pathway via reduced recruitment of TLR2/4 to lipid-rafts26, 21, 27. Previous studies have also shown that Miltefosine use lipid-rafts as entry portals to cells28 and can also act as an lipid-raft disrupting agent to inhibit human mast cell activation29. We determined the status of lipid-rafts in Miltefosine treated RAW264.7 macrophages by staining for the ganglioside GM1 with Alexa 647-labeled cholera toxin B, followed by either fluorescent microscopy or flow cytometry. As shown in Fig. 2A, RAW 264.7 macrophages treated with ± 7.5 μM Miltefosine for 16h displayed remarkably lower GM1-CTB staining as observed by fluorescent microscopy, indicating disruption of lipid-rafts. For quantification of lipid-rafts, the flow-cytometry based binding analysis of cells treated with ± 7.5 μM Miltefosine for 16h was carried out using Alexa-647 labeled CTB. The Miltefosine treated cells showed ~26% less binding as compared to control cells (with two tailed t-test, mean± SD, n=3, p<0.004) (Fig. 2B). Disruption of lipid rafts may increase the availability of cholesterol at plasma membrane. In addition, other mechanisms such as disruption in cholesterol trafficking from plasma membrane to endoplasmic reticulum can also increase cholesterol on plasma membrane30. We have previously shown that membrane remodeling resulting in lipid raft disruption can also increase cell-surface PS, that can promote cholesterol extraction from cell membrane23. We investigated if Miltefosine increased PS exposure at the cell-surface by performing a flow-cytometry based Annexin Cy5 binding assay. Cells expressing Abca1 were used as positive control as PS exposure is increased by Abca1 PS floppase activity31, 32. As shown in Fig. 2C, cells treated with 7.5μM Miltefosine for 16h showed increased PS exposure at cell surface. The cells without Abca1 expression showed ~24% increase in PS exposure vs. control cells with ANOVA, mean± SD, n=4, p<0.005. The cells expressing Abca1 showed increased PS exposure (~23% vs control increase with ANOVA, mean± SD, n=4, p<0.05). The PS exposure was further increased in Abca1 expressing cells treated with Miltefosine (~39% increase vs. control with ANOVA, mean± SD, n=4, p<0.0005). We have previously shown that the incremental increased PS exposure observed upon Abca1 expression is well below than PS exposure observed in cells undergoing apoptosis23. These data indicated that Miltefosine increased PS exposure by a mechanism different than increased Abca1 mediated PS flop from inner leaflet to the outer leaflet of plasma membrane. The increased levels of cell surface PS upon Miltefosine treatment could be due to decreased rate of PS flipping (inward translocation). To qualitatively observe the PS flipping across plasma membrane by fluorescent microscopy, we used head-group labeled NBD-PS. Control RAW264.7 cells and cells treated with 7.5uM Miltefosine for 16h were incubated with exogenously added NBD-PS. In control cells, majority of the exogenous NBD-PS moved to intracellular compartments over a 15 min incubation period at 37°C (Fig. 2D). However, in Miltefosine treated cells, much less NBD-PS was observed inside the cells. To quantitatively measure flipping of NBD-PS, we used a previously described flow-cytometry based method using membrane-impermeable NBD quenching reagent sodium dithionite23. RAW cells treated with ± 7.5uM Miltefosine for 16h were incubated with NBD-PS for 15 min at 37 °C, sodium dithionite was added to selected aliquots, followed by flow cytometry analysis. The amount of NBD-PS flipped into the cells was calculated by measuring the % of NBD fluorescence resistant to quenching by sodium dithionite. As shown in Fig. 2E, Miltefosine treated cells flipped significantly lower amounts of NBD-PS as compared to control cells (~44% vs. ~70% by ANOVA, mean± SD, n=4, p<0.0002). These data indicated that Miltefosine inhibited PS flip from the cell surface, resulting in increased PS exposure at the plasma membrane.
PIP2 is redistributed from plasma membrane upon Miltefosine treatment
Similar to PS, PIP2 is also sequestered to inner leaflet of PM. To determine the effect of Miltefosine on PIP2 trafficking and localization, RAW264.7 cells stably transfected with plasmid containing 2X-PH-PLC-eGFP construct were treated with ± 7.5 μM Miltefosine for 16h. The cells were washed with serum free media and PIP2 reporter localization was determined by fluorescent microscopy. As shown in Fig. 3A, control cells showed uniform localization of PH-PLC-eGFP at the plasma membrane while the cells treated with Miltefosine showed delocalization of PIP2 from plasma membrane and enrichment at cytoplasm near the periphery of cells. This kind of PIP2 enrichment is usually observed in cells undergoing migration, where actin and PIP2 shifts towards the back of migrating cells. To test the colocalization of PIP2 and actin, RAW264.7 cells treated with7.5 μM Miltefosine for 16h were fixed, permeabilized and incubated with beta-actin antibody, followed by counterstaining with Alexa 568 labeled goat anti-mouse antibody and fluorescent microscopy. As shown in Fig. 3B, PIP2 reporter GFP signal colocalized with beta actin, indicating that Miltefosine may be promoting cell migration in macrophages.
Miltefosine induced basal and lipid-droplet autophagy
Autophagy plays a protective role in atherosclerosis14 and it is impaired in advanced human plaques13. Autophagy is required for degradation of lipid-droplets in foam cells and generating free cholesterol from cholesterol esters for efflux from the cells. Given that Miltefosine increased cholesterol efflux from the cells, we speculated that Miltefosine may induced autophagy of cholesterol esters to generate free cholesterol. Upon autophagy induction, the cytosolic form of LC3 (LC3-I) is conjugated to phosphatidylethanolamine to form lipidated LC3-II. The PE-conjugated LC3-II is recruited to autophagosome membrane and can be observed as cytoplasmic puncta by microscopy while the adaptor protein p62 is responsible for chaperoning damaged proteins and organelles to autophagophore. To test if Miltefosine can induce autophagy in macrophages, we probed the RAW 264.7 macrophages treated with ± 7.5 μM Miltefosine for 16h with autophagy markers LC3 and p62. The Miltefosine treated cells showed multiple p62 puncta in cytoplasm, while control cells had evenly distributed weak cytoplasmic signal with almost no p62 puncta signal (Fig. 4A, upper panel). Miltefosine treatment also increased lipidated form of LC3-II as observed by puncta formation in macrophages transfected with LC3-GFP plasmid and by microscopy (Fig. 4A lower panel). Next, we evaluated whether the LC3 puncta was due to increased autophagic initiation or due to decreased autophagic flux leading to an accumulation of autophagic markers. The levels of LC3-II were determined by western blotting in presence of chloroquine. Chloroquine inhibits the autophagic flux leading to accumulation of LC3-II. Western blot analysis of LC3-II showed that the LC3-II protein levels were increased by 4.3-fold in Miltefosine cells vs. control cells (Fig. 4B, 4C), while there was a 3.1 fold increase in Miltefosine + chloroquine cells as compared to cells treated with chloroquine alone. These data indicated that Miltefosine induced autophagy and had no effect on autophagic degradative flux. Next, we tested if autophagy induced by Miltefosine could degrade cholesterol ester rich lipid droplets to generate free cholesterol for efflux. RAW264.7 macrophages were cholesterol loaded by incubation with 100 µg/ml acetylated LDL (AcLDL) for 24h to generate foam cells. These foam cells were then treated with ± 7.5 µM Miltefosine and ± acat inhibitor (Acati) for 4h. Aacti prevents the esterification of free cholesterol to cholesterol esters. The total and free cholesterol levels were determined by enzymatic assay described previously33. As expected, the AcLdL loaded cells had higher ratio of cholesterol ester to free cholesterol (Fig. 4D). The cells treated with actai had higher free cholesterol levels and thus lower CE::FC ratio, while the Miltefosine treated cells showed further lowering of CE::FC ratio, indicating increased levels of free cholesterol. This data indicated that Miltefosine can induce degradation of lipid-droplets generating free cholesterol (Fig. 4D). To decipher the mechanism by which Miltefosine was inducing lipid droplet degradation in macrophages, we determine the phosphorylation status of AMPK in control vs Miltefosine treated cells. AMPK phosphorylation is known to induce autophagy RAW264.7 macrophages loaded with 100μg/ml AcLdL were treated with 7.5μM Miltefosine for 4h followed by cell extracts preparations. As shown in Fig. 4E, AMPK phosphorylation levels were higher in Miltefosine treated macrophages.
Miltefosine decreased LPS mediated induction in IL-1β levels
Lipid-rafts plays an essential role in AKT signaling, TLR2/4 signaling pathway and atherosclerosis progression21, 27, 34–36. TLR4 activation leads to plasma membrane recruitment of MyD88, followed by phosphorylation of downstream kinases such as IkappaB Kinase alpha (IKBα) and NF-kB nuclear translocation. Since Miltefosine treatment led to lipid-raft disruption, we investigated status of TLR4 signaling pathway in Miltefosine treated bone-marrow derived macrophages (BMDMs). Control and Miltefosine pretreated BMDMs were incubated with 1µg/ml LPS for 30 min, followed by western blot analysis of phosphorylated form of IKBα. A shown in Fig. 5A, the LPS mediated increase in phosphorylation of IKBα was reduced in Miltefosine treated cells. To determine the levels of NF-kB translocating inside the nucleus, macrophages were treated with LPS for 30 min, followed by fixing the cells, probing with anti-NF-kB antibody and Alexa-568 labeled anti-rabbit secondary antibody and imaging using fluorescent microscopy. The Miltefosine treated cells showed reduced NF-kB translocation inside the nucleus vs. control cells (Fig. 5B). Next, we tested if LPS mediated synthesis of pro IL-1β messenger RNA was affected by Miltefosine by performing a qRT-PCR analysis using IL-1β specific primers and control actin b primers. As expected, the mRNA levels of pro IL-1β were significantly induced in LPS treated control cells as compared to cells not treated with LPS (Fig. 5C). Though the LPS induced increase in levels of IL-1β levels were observed in both control and Miltefosine treated BMDMs, the Miltefosine pretreated macrophages showed ~4.5 fold decrease in mRNA levels of pro-IL-1β (with two tailed t-test, mean± SD, n=3, p<0.034) (Fig. 5C). Next, we performed western blot analysis of protein levels of pro IL-1β in control vs. Miltefosine treated cells. A shown in Fig. 5D, the LPS mediated induction in pro IL-1β protein levels was decreased in Miltefosine treated cells. Taken together, these data indicated that Miltefosine disrupted lipid-rafts, dampened TLR signaling pathway and decreased the LPS mediated induction in IL-1β mRNA and protein levels in macrophages.
Miltefosine inhibited NLRP3 inflammasome assembly and IL-1β release
Cholesterol crystals can cause disrupt autophagolysosome membrane, leading to NLRP3 inflammasome assembly2, 37. We tested effects of Miltefosine treatment on inflammasome assembly and IL-1β release in bone-marrow derived macrophages. Control and BMDMs pretreated with 5μM Miltefosine for 16h, were primed with LPS for 4h and incubated with ATP for 20 min, followed by probing with anti-ASC1 antibody for visualization of Nlrp3 inflammasome assembly or IL-1β ELISA. The BMDMs pretreated with Miltefosine showed remarkably decreased Asc1 specks formation upon LPS+ATP stimulation (~40% as compared to 12% in control vs. Miltefosine treated cells (Fig. 6A, 6B, S5, S6). Next, we performed IL-1β ELISA on media obtained from control and Miltefosine treated macrophages and found that cells treated with Miltefosine had ~80% decreased IL-1β release as compared to control macrophages (with two tailed t-test, mean± SD, n=3, p<0.0039) (Fig. 6C). To decipher the mechanism by which Miltefosine is preventing Nlrp3 inflammasome assembly, we determined the mRNA and protein levels of various Nlrp3 inflammasome complex members in BMDMs with or without LPS priming. The qPCR analysis was performed on control and Miltefosine pretreated BMDMs using actin b as control. The LPS mediated increase in transcript levels of, NLRP3 (3.34 fold increase in LPS treated and 3.36 fold increase in LPS+ Miltefosine treated cells by ANOVA, mean± SD, n=4, p<0.0001), procaspase 1 (2.74 fold increase in LPS treated and 3.31 fold increase in LPS+ Miltefosine treated cells by ANOVA, mean± SD, n=3, p<0.0001), and Gasdermin D (1.3 fold increase in LPS treated and 2.25 fold increase in LPS+ Miltefosine treated cells by ANOVA, mean± SD, n=4, p<0.0001), were observed(Fig. 6D). These data indicated that Miltefosine treatment did not affect the transcription of these genes. In agreement with earlier published studies38, the western blot analysis showed robust induction in protein levels of NLRP3 upon LPS treatment (Fig. 6E), while there was no change in protein levels of Asc1 or procaspase 1 (Fig. 6F, 6G). The release of mature IL-1β from cells is regulated by Gasdermin D mediated pore formation on plasma membrane9, 10. We tested if Miltefosine effected Gasdermin protein levels or Gasdermin D activity by performing western blot analysis and IL-1β ELISA in control vs. Miltefosine treated BMDMS. There were no alteration in Gsdmd protein levels by Miltefosine treatment (Fig. S7). Taken together, these data indicated that Miltefosine dampened the IL-1β release by targeting Nlrp3 inflammasome pathway.
Miltefosine blocks Nlrp3 inflammasome in cholesterol independent manner and reduces mitochondrial ROS
We determined if defects in IL-1β release in Miltefosine treated cells is indeed due to disruption of Nlrp3 inflammasome by using an Nlrp3 specific inhibitor MCC9550. As shown in Fig. 7A, the cells treated with Miltefosine and MCC9550 did not show additive effects on the levels of IL-1β release as compared to cells treated with Miltefosine or MCC590 alone (with ANOVA, mean± SD, n=3, n.s.). Next, to test if Miltefosine specifically inhibited ATP/LPS induced NLRP3 inflammasome assembly and IL-1β release, we tested effects of Miltefosine on AIM2 inflammasome. The control and Miltefosine treated macrophages were primed with LPS (1mg/ml) for 4hrs followed by lipofectamine mediated transfection of poly (dA:dT) for 3h. The media was collected and subjected to IL-1β ELISA and as shown in Fig. 7B, there was no difference in levels of released IL-1β in control vs. Miltefosin treated macrophages (with two tailed -test, mean± SD, n=3, n.s.). These data indicated that Miltefosine did not alter AIM2 inflammasome activity. Previous studies have shown that deficiency of abca1/abcg1 increases inflammation in macrophages and ABCA1 mediated cholesterol efflux in dendritic cells dampen inflammasome assembly 39, 40. Though we have used cholesterol unloaded BMDMs for inducing inflammasome, there is a possibility that Miltefosine is inhibiting Nlrp3 inflammasome assembly via cholesterol depletion. To directly test the role of cholesterol depletion on Nlrp3 inflammasome assembly, we treated unloaded BMDMs with 5μM Miltefosine for 16h or with 1mM cyclodextrin for 2h at 37°C, followed by LPS/ATP incubation. As shown in Fig. 7C, the cholesterol levels in BMDMs treated with cyclodextrin were significantly lower as compared to Miltefosine treated cells (with ANOVA, mean± SD, n=3, p<0.01), but Miltefosine was much more effective in inhibiting inflammasome activity (Fig. 7D) as assayed by ASC1 specks per cell (Fig. S8) and mature IL-1β release (with ANOVA, mean± SD, n=3, p<0.0016 for control vs. Miltefosine, n.s. for control vs. cyclodextrin). These data indicated that Nlrp3 inhibition by Miltefosine is not solely due to cholesterol depletion.
Since Miltefosine treatment did not alter the expression levels of Nlrp3 inflammasome pathway components such as Nlrp3, Asc1, Caspase 1 or Gasdermin D, we focused on role of mitochondria in Nlrp3 inflammasome assembly. determined the status of mitochondrial ROS, Mitochondria plays an important role in induction and assembly of Nlrp3 inflammasome with mitochondrial ROS responsible for driving NLRP3 inflammasome assembly41 while mitochondrial cardiolipin providing a docking site for Nlrp3 protein binding and inflammasome assembly42. We tested the mitochondrial ROS production by staining with MitoSox reagent in control and Miltefosine treated BMDMs with or without LPS treatment. As shown in Fig. 7E, the LPS treatment induced ROS production in cells as evident by increased staining with MitoSox while the Miltefosine pretreated cells showed blunted induction in MitoSox staining upon LPS treatment. To quantify the MitoSox binding, we performed a flow-cytometry based binding assay. A shown in Fig. 7F, Miltefosine treated cells showed significantly reduced binding to MitoSox, indicating decreased levels of LPS mediated ROS generation in Miltefosine treated cells vs. control cells. Collectively, our data suggested that Miltefosine may be preventing Nlrp3 inflammasome assembly by eliminating damaged mitochondria by promoting autophagy and thus reducing ROS levels in cells treated with LPS.
Discussion
Given the multifaceted nature of pathways involved in promoting atherosclerosis and CVD, the drugs that can simultaneously induce atheroprotective pathway and dampens atherogenic pathways may be used as stand-alone or as an adjuvant therapy with current CVD therapies. Miltefosine is on WHO model list of essential medicines and is was approved by FDA in 2014 for treating cutaneous and visceral leishmaniasis. Miltefosine is also used to treat free-living amoeba infections such as Naegleria fowleri43. The structure of Miltefosine is similar to lyso-PC and it is known to integrate in the cell membrane and redistribute in intracellular membranes of ER, Golgi and mitochondria. Miltefosine and other similar alkylphospholipids are proposed to perturb phospholipid and sphingolipid homeostasis and also cause a deregulation of cholesterol homeostasis by impairing cholesterol transport from plasma membrane to the ER30. It’s important to note that the mechanism of action of Miltefosine, whether regarding killing of parasites or cancer cells, is not completely clear and It is widely believed that most of the downstream effects of Miltefosine are dependent on cell-type with reported range of activities such as induction of apoptosis in cancer cells and inhibition of Mast cell activation29, 44.
We found that Miltefosine has multiple activities in macrophages; inducing Abca1 mediated cholesterol efflux from cells, promoting lipid-raft disruption, inhibition of PS flip from the cell-surface, redistribution of PIP2 to actin-rich regions indicating induction of cell migration, induction of basal and lipid droplet degradation, and inhibiting NLRP3 inflammasome assembly and IL-1β release. As shown before in HepG2 cells22, Miltefosine increased the cholesterol efflux from the control RAW264.7 murine macrophages. Interestingly, cells expressing Abca1 effluxed cholesterol at much higher levels to media in the absence of any cholesterol acceptor (Fig. 1A), though no significant changes in Abca1 levels were observed in Miltefosine treated vs. control macrophages. Miltefosine can effect plasma membrane fluidity and membrane fluidity plays a role in ABCA1 mediated microparticle generation45. Thus combination of Miltefosine and Abca1 may lead to higher cholesterol efflux through release of lipids/cholesterol in form of microparticles. Miltefosine treatment led to disruption of lipid-raft that can increase the plasma membrane pool of free cholesterol and increased PS exposre conferred by Miltefosine can promote cholesterol extraction from the cells. We have shown before that limited increased exposure of PS on cell-surface can promote cholesterol extraction from membranes 23. The rate of PS flips from outer leaflet of plasma membrane to inner leaflet was decreased in Miltefosine treated cells, indicating that the PS exposure is mostly due to decreased PS flip rather than increased PS flop by Abca1’s lipid floppase activity. Previous study have shown that short-term treatment of HePG2 cells with 50 μM Miltefosine and other alkylphospholipids induces cholesterol efflux from the cells to extracellular media22, while prolonged treatment with Miltefosine can cause defects in apoA1 mediated cholesterol efflux in HepG2 and TPA-treated THP1 cells44. Though, we did not see significant increase in cholesterol efflux to apoA1 with higher dose of Miltefosine or during longer chase time-periods with Miltefosine, we attribute this effect to robust activation of Abca1 mediated cholesterol efflux to media. We tested and found that apoA1 mediated lipid-solubilization was increased in presence of Miltefosine. At the lipid phase transition temperature, the membrane lattice defects at phase boundaries facilitates the increased rate of solubilization of phospholipid liposomes by apoAI 46. Miltefosine may increase the packing defects at the phase transition temperature, leading to increased insertion of apoA1 into liposomes.
The lipid-rafts play an essential role in recruiting TLRs to plasma membrane to initiate a signaling cascade that allows cytoplasmic NF-kB to shuttle inside the nucleus and upregulate the transcript levels of pro IL-1β. Resting macrophages express low levels of this genes but macrophages primed with primary signal such as LPS can robustly induce mRNA levels of pro IL-1β as well as NLRP3. Miltefosine treated cells showed significantly lowered levels of LPS induced pro IL-1β transcript levels. Surprisingly, the NLRP3 transcript levels in Miltefosine treated cells were induced similar to untreated cells, though both IL-1β and NLRP3 transcripts are induced by NF-kB. The LPS induced nuclear localization of NF-kB was partially inhibited in Miltefosine treated cells, indicating delayed TLR4 mediated cell signaling rather than a complete blockage. Since the levels of NLRP3 transcripts were modestly increased as compared to robust induction seen in case of pro IL-1β transcript levels, the decreased shuttling of NF-kB may not hinder NLRP3 transcription as compared to IL-1β transcription. Given the sharp decrease in levels of pro IL-1β transcript levels, in addition to decrease synthesis of mRNA, Miltefosine treatment may also be destabilizing IL-1β mRNA leading to increased degradation of IL-1β transcripts.
Autophagy plays a protective role in atherosclerosis14 and it is impaired in advanced human plaques13. Autophagy is controlled by multiple pathways responding to stimuli such as the status of cellular energy (AMP-dependent protein kinase, AMPK) or amino acid availability (target of rapamycin, TOR). Miltefosine induced basal autophagy in macrophages as evident by increased cytoplasmic p62 and LC3-GFP puncta. Miltefosine treated cells showed increased degradation of cholesterol esters as compared to control cells. The cells treated with ACATi showed higher free cholesterol levels in control cells but the Miltefosine treatment further increased the levels of free cholesterol. This data along with electron microscopy data showing double membrane bound autophagosomes containing lipid-droplets, mitochondria, and other organelles, supports the notion that Miltefosine may induce degradation of lipid-droplets via lipophagy. Increased phosphorylation of AMPK observed in Miltefosine treated cells as compared to control cells can explain induced basal autophagy observed in Miltefosine treated cells.
The components of inflammasome complex, NLRP3, ASC1, Caspase1, and, IL1-β are significantly increased in atherosclerotic plaques compared to healthy arteries. Blocking of NLRP3 by a chemical drug MCC9550 reduces atherosclerosis in mouse and infarct size in pigs 47, 48 and recent trials showed the reduced cardiovascular events upon treatment with IL-1β antibodies 7. The most striking activity of Miltefosine was its ability to strongly block the assembly of Nlrp3 inflammasome upon LPS and ATP stimuli and reduce release of mature IL-1β from macrophages. Nlrp3 inflammasome assembly is controlled via transcriptional as well as via posttranslational mechanisms. The protein levels of NLRP3 are boosted by a non-transcriptional priming pathway, involving TLR4 and Myd88, via NLRP3 deubiquitination38. The basal levels of NLRP3 inflammasome components (ASC1, NLRP3, and Caspase 1) in resting as well as in LPS induced macrophages were unaltered in control vs. Miltefosine treated cells. LPS mediated induction in protein levels of NLRP3 was also not altered by Miltefosine treatment. The levels of pro IL-1β were lower in Miltefosine treated cells as compared to control cells but the LPS mediated induction was still intact. The mature IL-1β is released via Gasdermin D pores11, but Miltefosine treatment did not alter the levels of Gasdermin D. The intriguing question raised by these data is that if all inflammasome components are present and more or less equally induced by LPS in control vs. Miltefosine treated macrophages, why the Nlrp3 inflammasome is not assembled in Miltefosin treated cells. Mitochondria do play an essential role in activating Nlrp3 inflammasome41, 49. Miltefosine treated cells showed reduced levels of ROS upon LPS treatment (Fig. 7E, F). Based on our data, we speculate that Miltefosine treated macrophages failed to assemble Nlrp3 inflammasome due to lower levels of oxidative stress and efficient removal of damaged mitochondria (mitophagy), thus blocking Nlrp3 inflammasome assembly. Recent studies have shown that a serine threonine kinase Nek7 binds directly to Nlrp3 to promote inflammasome assembly, though kinase activity of Nek7 is not required for this function50–52. Nek7-Nlrp3 interaction is required for Nlrp3 inflammasome assembly and is dependent on potassium efflux 51. Though Nek7 and Nlrp3 protein levels are not affected by Miltefosine treatment, Miltefosine can prevent potassium efflux and inhibit Nek7-Nlrp3 binding and prevent Nlrp3 inflammasome assembly.
Based on our data, Miltefosine or compounds with similar activities can be combined with statins to prevent CVD. Given the effects of Miltefosine on NLRP3 inflammasome pathway and IL-1β release, it may also serve as potential therapeutic for metabolic diseases where Nlrp3 inflammasome is hyper activated; such as diabetes, non-alcoholic fatty liver disease (NAFLD) and high fat diet induced inflammation. Further work is required to fully understand the underlying molecular mechanisms responsible for spectrum of Miltefosine’ s actions on macrophage cellular functions.
Material and Methods
Cell culture and reagents
All cell culture incubations were performed at 37°C, unless otherwise indicated, in a humidified 5% CO2 incubator. The growth media was Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal calf serum, 100 µg/mL penicillin, 100 µg/mL streptavidin. Drugs were added to growth media at indicated concentrations and an equivalent amount of vehicle was added as control. Miltefosine, 18:1 NBD-PS, perifosine, edelfosine were obtained from Avanti polar and Echelon biosciences, methyl-β-cyclodextrin is obtained from Sigma Aldrich. 3H-cholestreol (NET13900) was obtained from Perkin- Elmer. The primers for RT-PCR were obtained from ThermoFisher and siRNAs were obtained from Dharmacon (see supplementary material and methods). The RAW264.7 cells were from ATCC. Mifepristone ABCA1-inducible BHK cells, as previously described were obtained from Chongren Tang, University of Washington53. Mifepristone SR-BI-inducible BHK cells, as previously described54, were obtained from Alan Remaley, NIH54. ABCA1-GFP and the mutant isoform stably transfected HEK cells were as previously described55. Primary and secondary antibodies were purchased from various sources (see supplementary material and methods).
Mice
WT C57BL6 or C57BL6 ApoE-/- mice were maintained on chow diet and sacrificed at 16 weeks of age. Femurs were collected to isolate and culture bone marrow macrophages using conditioned L-cell media.
Isolation of total RNA from BMM cell pellets
Total RNA was isolated from cell pellets using the RNeasy Mini Kit (Qiagen, Valencia, CA), following the manufacturer’s protocol. On-column digestion with RNase-free DNase (Qiagen, Valencia, CA) was performed to remove genomic DNA. The qRT-pCR assays were performed using TaqMan reagent from ThermoFisher following the manufacturer’s protocol.
Cholesterol efflux assay
The HEK293-Abca1-GFP cells, BHK cells, or RAW 264.7 murine macrophages were plated in 24-well plates at a density of 300–400,000 cells per well. The cells were labeled with 3H cholesterol in 1% FBS in DMEM for 24h. The labeled cells were induced for ABCA1 expression for 16h with 0.3 mM 8Br-cAMP (for RAW264.7 cells) or with 10nM Mifepristone (for BHK cells). Inducers were included in the media during subsequent assays. ABCA1 expression was confirmed by western blot using the AC10 antibody. The cells were washed with serum free media and chased for indicated time period in serum-free DMEM in the presence or absence of indicated treatments. The radioactivity in the chase media was determined after brief centrifugation to pellet any residual debris. Radioactivity in the cells was determined by extraction in hexane:isopropanol (3:2) with the solvent evaporated in a scintillation vial prior to counting. The percent cholesterol efflux was calculated as 100 × (medium dpm) / (medium dpm + cell dpm).
Lipid-Raft quantification
RAW cells were treated with 7.5 μg/ml Miltefosine at 37°C for 16h. For imaging, cells were treated with 1 μg/ml Alexa647-CT-B labeled cells and images were captured using epifluorescent microscopy. For quantification, the cells were washed twice with PBS and gently scraped in 1.5 ml PBS. The Alexa647-CT-B (1μl of 1.5 mg/ml stock in PBS was added to 1.5 ml cells) was added at final concentration of 1 μg/ml to cells in 5 ml polystyrene round-bottom tubes (12 × 75 mm). The samples were subjected to Flow analysis with BD Biosciences LSRII cytometer using laser (Alexa 647 (Red) Ex: 639, Em: 650-670, Filter 660/20) and data was analyzed by Flowjo software.
Total/FC measurements
RAW264.7 cells grown in 12-well plates were loaded with 50 µg/ml cholesterol-CD or with 100 µg/ml AcLDL for 16 h at 37 °C with or without 2 µg/ml ACATi treatment for 2h. The total cholesterol and free cholesterol levels were determined by using enzymatic assay as described earlier33. Detailed protocol is described in SI.
ApoA1 lipid solubilization assay
DMPC dissolved in chloroform:methanol (2:1, v:v) plus 5 mole% Miltefosine was dried in a stream of nitrogen onto the sides of a glass culture tube and kept in vacuum overnight. The DMPC film was rehydrated at 5 mg/mL in PBS by extensive vortexing and alternating freeze/thaw in a dry ice/ethanol and 37°C water bath. The resulting DMPC-DPH multi-lamellar vesicles (MLV)s were warmed to 37°C and extruded 11 times through a polycarbonate membrane using a mini-extruder (Avanti Polar Lipids) to derive large unilamellar vesicles (LUVs) of 100 nm diameter. 20 μg of these MLVs in Tris-buffered saline-EDTA (pH7.5) were incubated with 1 μg of human apoAI at 25 °C in a final volume of 200 μl. MLV solubilization by human apoAI was monitored by measuring sample turbidity (absorbance) at 325 nm using a plate spectrophotometer.
Flow-Cytometry analysis
To measure cell surface PS, cells were cultured in 6 well plates in growth media, washed twice with PBS, and were resuspended in 500 µl Annexin V binding buffer plus 1 µl of Annexin V-Cy5 (Biovision). The samples were incubated at room temperature for 5 minutes in the dark. Flow cytometry analysis was performed using a BD Biosciences LSRFortessa cytometer using a 639 nm excitation laser and emission at 650–670 nm; and, data was analyzed using Flowjo software. To measure the inward translocation of various NBD labeled PS, RAW cells were incubated with phenol-red free DMEM containing 25 μM NBD-PS for 15 min at RT. The media was then replaced with PBS. The cells were gently scraped and were subjected to flow cytometry analysis first without sodium dithionite to calculate total NBD fluorescence, and then 5 min after the addition of 30 μl of a freshly prepared 1M sodium dithionite solution to determine the fraction of the signal resistant to quenching (translocated signal). Flow cytometry analysis was performed as above using a 488 nm laser for excitation and emission at 505–525 nm. To measure cell surface lipid rafts, a flow-cytometry based CTB Alexa 647-GM1 binding assay was used as described previously 23. RAW cells were treated with either DMSO vehicle or 7.5 µM Miltefosine (16 hr pretreatment) at 37°C. These cells were then washed twice with PBS and gently scraped in 1.5 ml PBS. Alexa 647-CTB was added at final concentration of 1 μg/ml and incubated for 1 min prior to flow cytometry as described above using a 639 nm excitation laser and emission at 650–670 nm.
Western blotting
RAW264.7 cells or BMDMs were grown treated as indicated. The PBS-washed cell pellet was lysed in 200 µL of NP40 lysis buffer. After discarding the nuclear pellet, the protein concentration was determined using the BCA protein assay (Pierce). 10–50 µg of cell protein samples were run on Novex 4–20% Tris-Glycine Gels (Invitrogen) and transferred onto polyvinylidene fluoride membranes (Invitrogen). Blots were incubated sequentially with 1:1000 rabbit polyclonal antibody raised against Nlrp3 (cell signaling), or 1:5000 rabbit polyclonal antibody raised against P-AMPK (cell signaling), or 1:1000 rabbit polyclonal antibody raised against IL1-b (cell signaling), 1:1000 rabbit polyclonal antibody raised against Gasdermin D, 1:1000 rabbit polyclonal antibody raised against ASC1. The signal was detected with an enhanced chemiluminescent substrate (Pierce). For LC3 western blot, RAW264.7 cells were incubated with or without 7.5 μM Miltefosine for 16 h, indicated wells received 30 µM chloroquine during the last 2 h of incubation. Blots were sequentially incubated with 1:1,000 rabbit polyclonal antibody raised against LC3B and 1:10,000 horseradish peroxidase-conjugated goat anti-rabbit as described above. GAPDH was used as a loading control.
Indirect immunofluorescence/fluorescent microscopy
RAW264.7 cells or BMDMs were grown in 4-chamber slides in indicated growth media. Cells were washed with 1xPBS, fixed with paraformaldehyde, and then incubated with permeabilization solution (1xPBS/0.2% Triton X-100) for 10 min at RT. Cells were washed with 1xPBS three times and blocked with normal goat serum (1xPBS/1%BSA/5% normal goat serum) at 37οC for 30 min. The cells were probed with p62 antibody and Alexa 647 labeled secondary antibody. For LC3 staining, RAW264.7 cells were transfected with LC3-GFP (Addgene) using lipofectamine LTX according to manufacturer protocol (Life Technologies). Cells were treated with or without 7.5 μM Miltefosine for 16 h. For ASC1 staining, Bone marrow derived macrophages from WT C57BL6 or C57BL6 apoE-/- KO mice were treated ± 5 μM Miltefosine for 16 h or ± 1mM cyclodextrin for 2h. All images were captured using an Olympus IX51 inverted epifluorescent microscope, Olympus LUCPlanFl ×40/0.6 lens, with a Q-Image EXi aqua camera and Olympus cellSens Dimension version 1.7 software. Post-imaging processing was performed in Adobe Photoshop CS2.
Total and cell surface ABCA1 levels
RAW 267.4 cells were cultured in DMEM with 10% FBS and were treated with 300 μM 8Br-cAMP to induce ABCA1 for 16 hr. Cells were incubated with incubated with or without 7.5μM Miltefosine for 4h. The cells were then incubated for 30 minutes on ice with phosphate-buffered saline (PBS) containing 1 mg/mL sulfo-NHS-biotin (Pierce). The PBS-washed cell pellet was lysed in 200 μl of lysis buffer (150 mM sodium chloride, 5 mM EDTA, 50 mM Tris-phosphate pH 8.0, 1% NP40, and 10% protease inhibitor). Cell-surface ABCA1 was determined by purifying biotinylated proteins by overnight incubation of 250 ug of cell protein with 75 uL of UltraLink Plus Immobilized Streptavidin gel (Pierce) at 4°C. The beads were spun down, washed, and resuspended by boiling in 100 μl 1x NuPage LDS Sample buffer (Invitrogen) and 20 µl of the bound protein samples along with 20 µg of total protein were run on NuPAGE 3–8% Tris-Acetate gels (Invitrogen) and transferred onto polyvinylidene fluoride membranes (Invitrogen). Blots were incubated sequentially with 1:500 mouse monoclonal antibody raised against ABCA1 (AC10) and 1:15 000 horseradish peroxidase-conjugated goat anti-mouse secondary antibody (Biorad). The signal was detected with an enhanced chemiluminescent substrate (Pierce).
PIP2 cellular reporter assay
RAW264.7 macrophage cell lines stably transfected with 2PH-PLCδ-GFP plasmid was described earlier32. Live or fixed cells were visualized by epifluorescent microscopy. Images were taken using the same exposure time.
Statistical analyses
Data are shown as mean ± SD. Comparisons of 2 groups were performed by a 2-tailed t test, and comparisons of 3 or more groups were performed by ANOVA with Bonferroni posttest. All statistics were performed using Prism software (GraphPad).
Author Contributions
A.J.I., and H.L. performed the experiments. J.H. performed the PS translocation assay. G.B., and B. A. R., performed the liposome solubilization assays. J.D.S. provided help with data analysis. K. G. conceived of the project idea, designed and directed study, performed experiments, performed data analyses and drafted the manuscript. All authors critically reviewed the manuscript.
Acknowledgments
We thank Dr. Brian Ritchey for the technical assistance and Dr. Shuhui Wang Lorkowski for providing reagents. This research was supported by NIH grants R01 HL128268 and R01 HL130085R01/PO1 to J.D.S. and American Heart Association scientist development grant SDG25710128 to K.G.