Autophagy regulates cholesterol efflux from macrophage foam cells via lysosomal acid lipase

Abstract

The lipid droplet (LD) is the major site of cholesterol storage in macrophage foam cells and is a potential therapeutic target for the treatment of atherosclerosis. Cholesterol, stored as cholesteryl esters (CEs), is liberated from this organelle and delivered to cholesterol acceptors. The current paradigm attributes all cytoplasmic CE hydrolysis to the action of neutral CE hydrolases. Here, we demonstrate an important role for lysosomes in LD CE hydrolysis in cholesterol-loaded macrophages, in addition to that mediated by neutral hydrolases. Furthermore, we demonstrate that LDs are delivered to lysosomes via autophagy, where lysosomal acid lipase (LAL) acts to hydrolyze LD CE to generate free cholesterol mainly for ABCA1-dependent efflux; this process is specifically induced upon macrophage cholesterol loading. We conclude that, in macrophage foam cells, lysosomal hydrolysis contributes to the mobilization of LD-associated cholesterol for reverse cholesterol transport.

Figure 1. LDs Surround Neutral Lipid Rings and Colocalize with Lysosomes; Inhibition of Lysosomal Function Reduces CE Hydrolysis and Cholesterol Efflux

(A and B) BMDMs were loaded with AcLDL-derived cholesterol for 30 hr, equilibrated overnight in BSA media, and then incubated with media containing Bodipy (10 µg/mL) with or without LysoTracker Red (50 nM) for 30 min prior to visualization. (C) Cells were cholesterol loaded as above and then incubated with a BSA conjugate prior to labeling with Bodipy. (D) Colocalization between the macrophage LD coat protein adipophilin and LAMP-1-positive lysosomes in AcLDL-loaded macrophages. (E and F) Cellular CE (E) and CE hydrolysis (F) were measured in AcLDL-loaded cells treated with paraoxon or chloroquine for 24 hr in the presence of apoA-I, with or without ACATi. Variations in CE are expressed as fold change relative to control (E) or as a percent CE hydrolysed in 24 hr (F). ***p < 0.0001, **p < 0.001, or *p < 0.005 compared to ACATi, and ACATi was compared to control. (G) BMDMs were loaded with ³H-cholesterol-AcLDL for 30 hr and equilibrated overnight, and efflux to apoA-I was measured for 24 hr in the presence or absence of paraoxon or chloroquine.

Figure 2. Autophagy Is Implicated in Cytoplasmic LD Degradation

(A–C) In lipid-loaded BMDMs, direct association of autophagomes with LDs is observed by immunofluorescence (A), electron microscopy (B), and in an isolated LD fraction (C). (D) Vinblastine treatment inhibits autophagosome degradation, as shown by elevated LC3-II. (E) Inhibition of autophagy by vinblastine treatment during cholesterol efflux decreases efflux to apoA-I. (F) Quantification of LDs containing gold particles (cells immunostained with LC3 are compared to the secondary antibody alone negative control).

Figure 3. Autophagy Is Induced in Response to Lipid Loading, and LAL Mediates LD Catabolism

(A) Chloroquine inhibits cholesterol efflux in macrophage foam cells, but not in unloaded cells. (B) Chloroquine has no effect on cholesterol efflux at an early time point, but impairs cholesterol efflux upon prolonged inhibition of lysosomal function. (C) Autophagic flux is induced in lipid-loaded macrophages. (D and E) Effect of LAL inhibition on cholesterol efflux (D) and cellular CE mass (E) in unloaded versus lipid-loaded macrophages.

Figure 4. Ablation of Autophagy Impairs LD Delivery to Lysosomes for LAL-Mediated CE Hydrolysis in Macrophage Foam Cells

(A) Atg5 deletion in macrophages impairs LC3-I maturation to LC3-II, and autophagy cannot ensue. (B and C) Cholesterol loading is similar in WT and Atg5^−/− macrophages, as recorded by mass measurements (B) and ³H-cholesterol uptake (C). (D) Cholesterol efflux to apoA-I is impaired in Atg5^−/− lipid-loaded macrophages; LAL inhibition reduces cholesterol efflux in WT, but not Atg5^−/−, macrophages. ★p < 0.0001 relative to WT control, and #p < 0.05 relative to Atg5^−/− control. (E) CE hydrolysis in lipid-loaded WT and Atg5^−/− macrophages over 24 hr in the presence of apoA-I, expressed as a fold change relative to the WT control (cellular CE mass was measured before and after cholesterol efflux, and resulting percent hydrolysis was normalized to that of WT). (F) Autophagy-deficient macrophages exhibit impaired LD catabolism. Lipid-loaded WT or Atg5^−/− macrophages were immediately fixed for fluorescence microscopy after AcLDL loading or after a 24 hr incubation with media containing lipid-poor apoA-I in the presence of an ACAT inhibitor. Neutral lipids were stained with Nile Red. (G) AcLDL-derived ³H-cholesterol is esterified to ¹⁴C-oleic acid to the same extent in WT and ^Atg5−/− macrophages. In the absence of the ACATi, esterification of the ³H and ¹⁴C labels parallels each other and is equivalent in WT and Atg5^−/− cells. When the ER-resident ACAT is inhibited by ACATi, esterification of both the ³H and ¹⁴C labels is abolished. (H) Degradation of lipoprotein ³H-cholesteryl oleate occurs at the same rate in WT and Atg5^−/− macrophages. Macrophages were loaded with AcLDL or OxLDL containing ³H-cholesteryl oleate in the presence or absence of Lalistat 1 to inhibit LAL, followed by an O/N equilibration or not. All experiments were performed in the presence of the ACATi to prevent re-esterification of the liberated ³H-cholesterol.

Figure 5. Enhanced Autophagy in Response to Various Atherogenic Lipoproteins and in Peritoneal Macrophages from Hypercholesterolemic Mice

(A) LD genesis in bone marrow-derived macrophages exposed to AgLDL, VLDL, and OxLDL. (B) Elevated autophagic flux in response to atherogenic lipoproteins. (C) Impaired efflux to apoA-I in atg5^−/− macrophages loaded with AgLDL, VLDL, and OxLDL. (D) Atg5 ablation reduces cholesterol efflux to HDL in AcLDL-loaded macrophages. (E and F) Efflux to apoA-I (E) and HDL (F) in WT and Atg5^−/− macrophages with (control) or without (ACATi) cytoplasmic LDs. (C–F) ★★★p < 0.0001 or ★p < 0.05 relative to WT control, and ##p < 0.001 or #p < 0.05 for ACATi-treated WT cells relative to WT control, and ♦♦p < 0.001 or ♦p < 0.05 for ACATi-treated Atg5^−/− cells relative to ACATi-treated WT cells. (G) Autophagy mediates cholesterol efflux in peritoneal macrophages. (H) Peritoneal macrophages isolated from hypercholesterolemic apoe^−/− mice contain more LDs than their WT counterparts. (I) Autophagy levels are elevated in peritoneal macrophages lipid loaded in vivo.

Figure 6. Macrophage-Specific Deletion of Atg5 Reduces RCT In Vivo

(A) ³H-tracer in plasma at 24 and 48 hr postinjection. (B) ³H-tracer in livers at 48 hr postinjection. (C) ³H-tracer in gallbladders at 48 hr postinjection. (D) ³H-tracer in feces at 48 hr postinjection.

Figure 7. Model of Autophagy in LD-Associated CE Hydrolysis in Macrophage Foam Cells

(A) Cholesterol homeostasis in “normal,” unloaded macrophages; a model for CE hydrolysis that mainly involves the action of neutral CE hydrolases is presented. (B) Autophagy is induced in early macrophage foam cells, and delivery of LDs to lysosomes enhances LD-associated CE hydrolysis and cholesterol efflux.