Exploiting macrophage autophagy-lysosomal biogenesis as a therapy for atherosclerosis

Abstract

Macrophages specialize in removing lipids and debris present in the atherosclerotic plaque. However, plaque progression renders macrophages unable to degrade exogenous atherogenic material and endogenous cargo including dysfunctional proteins and organelles. Here we show that a decline in the autophagy-lysosome system contributes to this as evidenced by a derangement in key autophagy markers in both mouse and human atherosclerotic plaques. By augmenting macrophage TFEB, the master transcriptional regulator of autophagy-lysosomal biogenesis, we can reverse the autophagy dysfunction of plaques, enhance aggrephagy of p62-enriched protein aggregates and blunt macrophage apoptosis and pro-inflammatory IL-1β levels, leading to reduced atherosclerosis. In order to harness this degradative response therapeutically, we also describe a natural sugar called trehalose as an inducer of macrophage autophagy-lysosomal biogenesis and show trehalose's ability to recapitulate the atheroprotective properties of macrophage TFEB overexpression. Our data support this practical method of enhancing the degradative capacity of macrophages as a therapy for atherosclerotic vascular disease.

Conflict of interest statement

The authors declare no competing financial interests.

Figures

Figure 1. Mouse and human atherosclerotic plaques develop features of a progressive autophagy dysfunction.

(a–c) Representative immunofluorescence images of early-stage and more advanced atherosclerotic (ApoE-KO) aortic roots co-stained with antibodies against LC3 and p62. Early and advanced lesions were obtained from ApoE-KO mice fed a western diet for <2 months and 3–4 months, respectively (scale bar, 50 μm (a)). The mean intensity for LC3 and p62 stainings were analysed (n=5 mice for each group; b,c). (d,e) Co-localization of LC3 and p62 was also analysed in the same aortic roots. Representative co-localization images are shown from early and more advanced lesions (green indicates LC3/p62 co-localized, red indicates LC3-positive, and blue indicates p62-positive areas) (d). LC3/p62 co-localization is quantified as per cent of total signal (e). (f–k) Immunofluorescence analysis of human carotid endarterectomy specimens (n=8), which are separated as maximally- and adjacent minimally diseased regions (scale bar, 100 μm). Specimens were co-stained with LC3 and p62 (f), and co-localization (g) as well as correlation of staining intensity (h) between maximally and minimally diseased regions are quantified. Maximally diseased human atherosclerotic regions were co-stained for p62 and polyubiquitinated proteins (FK-1 antibody; i), and co-localization quantified (j). (k) Graph represents a comparison of the staining correlation between p62/LC3 versus p62/ubiquitin(FK-1) in maximally atherosclerotic regions. For all graphs, data are presented as mean±s.e.m. **P<0.01, ***P<0.001, two-tailed unpaired t-test. Max, maximum; Min, minimum; Ubiq., ubiquitination.

Figure 2. TFEB overexpression induces autophagy and autophagy–lysosomal biogenesis in macrophages.

(a–c) Control and TFEB-overexpressing (TFEB-TG) thioglycollate-elicited peritoneal macrophages (hereafter referred to as macrophages) were assessed as follows: transcript levels of (a) TFEB and (c) several autophagy–lysosome markers were evaluated by quantitative polymerase chain reaction (qPCR, n≥3 independent wells). (b) TFEB nuclear localization was assessed by immunofluorescence staining and quantified as percentage of TFEB-positive nuclei (n≥40 cells per group, scale bar, 20 μm). (d) Western blot analysis of p62 and LC3 in TFEB-TG macrophages after bafilomycin (200 nM) treatment for indicated times (C, control; T, TFEB-TG). (e–g) Control and TFEB-TG macrophages also co-expressing GFP-LC3 were evaluated by live imaging (every 30 s for the indicated times) while being incubated with either (e) DMEM, (f) bafilomycin (200 nM) or after staining with (g) Lysotracker-red. Graphs represent (e,f) GFP-LC3-positive areas or (g) per cent of GFP-LC3 co-localized with Lysotracker-red over the indicated times. For e,f each time point is compared with the control GFP-LC3 group (n≥10 cells for each treatment). (h) LC3 levels and the intracellular pattern were analysed by immunofluorescence staining of baseline (NoTX) or after 3 h of 10 μM chloroquine incubation. Graphs represent the mean LC3 intensity (n=16–46 cells, scale bar, 5 μm). (i) FACS analysis of peritoneal and splenic macrophages from control or macrophage-specific TFEB-TG mice for LAMP1 expression. For all graphs, three independent experiments were performed; data presented as mean±s.e.m. *P<0.05, ***P<0.001, two-tailed unpaired t-test. Baf,bafilomycin; CHQ, chloroquine; Ctrl, control; DAPI, 4,6-diamidino-2-phenylindole.

Figure 3. TFEB overexpression in macrophages induces the autophagy markers LC3 and p62 and restores their co-localization in atherosclerotic aortic roots.

(a,b) Representative immunofluorescence images of atherosclerotic aortic roots (2 months' western diet) from control and mφTFEB-TG mice (ApoE-null background) stained with antibodies against TFEB (a), TFEB and MOMA-2 (b; scale bar, 50 μm). (c) Quantification of the average TFEB intensity and co-localization with nuclear marker DAPI (n=4-5 mice per group). (d) Representative immunofluorescence images of atherosclerotic aortic roots from control and mφTFEB-TG mice stained with p62 and LC3 (scale bar, 50 μm). (e) Quantification of the p62 and LC3 average intensity from control and mφTFEB-TG-stained roots (n=13–14 mice per group). (f) Representative pseudocolour image of these p62/LC3 images (green represents co-localization) and graph depicting the increased p62/LC3 correlation seen in a representative mφTFEB-TG as compared to a control lesion (scale bar, 50 μm). (g) Quantification of the p62/LC3 co-localization from control and mφTFEB-TG-stained roots shown (n=13–14 mice per group). (h,i) FACS analysis of aortic macrophages isolated from atherosclerotic aortas of Control or mφTFEB-TG mice (western diet-fed ApoE-KO background, n=3–4 pooled aortas) and stained for either (h) p62 and LC3, or (i) Lamp2 and LC3 antibodies (per cent of macrophages expressing each marker is shown below plots). For all graphs, data are presented as mean±s.e.m. *P<0.05, ***P<0.001, two-tailed unpaired t-test.

Figure 4. Macrophage-specific TFEB overexpression is atheroprotective.

(a) Experimental protocol and mouse cohorts used for assessment of atherosclerosis. (b–g) Control and mφTFEB-TG mice (all on ApoE-KO background) were fed a western diet for 2 months for lesion development. (b) Serum cholesterol levels at 2 months (n≥14 mice per group). (c) Quantification of atherosclerotic plaque burden using Oil Red O-stained aortic root sections with representative roots shown on right (scale bar, 0.4 mm) and (d) en face analysis of whole aorta (statistical significance of differences calculated using Mann–Whitney U-test). (e) Macrophage content in aortic root sections was analysed by MOMA-2-positive area (n≥12 mice per group). (f) Apoptotic, necrotic core and combined apoptotic–necrotic core areas of aortic root sections were determined by quantifying TUNEL immunofluorescence staining, acellular areas or the combined area, respectively (n≥12 mice per group). (g) Serum IL-1β concentration was measured from n≥6 independent samples derived by pooling serum from two to three mice per sample (>12 mice per group). (h) Control and TFEB-TG macrophages treated with or without cholesterol crystals for 24 h and stained using DAPI (nuclei) and antibodies against polyubiquitinated proteins (FK-1) and p62 (scale bar, 5 μm). (i) Quantification of average p62/ubiquitin-positive dots and average p62 intensity per cell from immunofluorescence experiment described in h (numbers of cells under each bar). (j) Control and TFEB-TG macrophages were incubated with cholesterol crystals and per cent of caspase 3/7-positive cells quantified in three independent experiments. Representative immunofluorescence images are shown on left and numbers of cells shown under each bar (scale bar, 50 μm). (k) Control and TFEB-TG macrophages were treated with LPS (lipopolysaccharide)+cholesterol crystals (hereafter referred to as LPS+CC) for 24 h or with LPS followed by ATP for 3 h. Culture media were assayed for IL-1β by ELISA (n=3 independent wells for each treatment). (l) Control and TFEB-TG macrophages were treated with DiI-acetylated LDL for 12 h and intracellular lipid accumulation quantified by immunofluorescence microscopy (n≥187 cell per group, scale bar, 5 μm). For all graphs, data are presented as mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001, two-tailed unpaired t-test, except c,d.

Figure 5. Macrophage-specific TFEB overexpression requires an intact autophagy pathway for atheroprotection including efficient clearance of polyubiquitinated protein aggregates and reductions in macrophage apoptosis.

(a) Cohorts of control and mφTFEB-TG mice (all on mφATG5-KO and ApoE-KO background) were fed a western diet for 2 months to develop lesions—exact genotypes are provided at the top of the graph. Atherosclerotic plaque burden was quantified by computer image analysis of Oil Red O-stained aortic root sections with representative Oil Red O-stained aortic roots shown on right (scale bar, 0.4 mm; statistical significance of differences was calculated using Mann–Whitney U-test). (b) Macrophage content in aortic root sections was analysed by immunofluorescence staining using an antibody against MOMA-2 (n≥10 mice per group). (c) Apoptotic, necrotic core and combined apoptotic–necrotic core areas of aortic root sections were determined by quantification of TUNEL immunofluorescence staining, acellular areas or a combination of the two, respectively (n≥10 mice per group). (d) Immunofluorescence images of ATG5-KO and ATG5-KO/TFEB-TG macrophages treated with or without cholesterol crystals for 24 h and stained using DAPI (nuclei) and antibodies against polyubiquitinated proteins (FK-1) and p62 (scale bar, 5 μm). (e) Quantification of average p62/ubiquitin-positive dots and average p62 intensity per cell from immunofluorescence experiment described in d (numbers of quantified cells are shown under each bar). (f) ATG5-KO and ATG5-KO/TFEB-TG macrophages were incubated with cholesterol crystals and the per cent of caspase 3/7-positive cells was quantified in three independent experiments (numbers of quantified cells are shown under each bar). (g) ATG5-KO and ATG5-KO/TFEB-TG macrophages were treated with LPS+CC for 24 h or with LPS, followed by ATP for 3 h. Culture media were assayed for IL-1β by ELISA (n=3 independent wells for each treatment). For all graphs, data are presented as mean±s.e.m. *P<0.05, **P<0.01, NS: not significant, two-tailed unpaired t-test, except a.

Figure 6. The atheroprotective effect of macrophage-specific TFEB overexpression is also p62-dependent.

(a) Cohorts of control and mφTFEB-TG mice (all on p62-KO and ApoE-KO background) were fed a western diet for 2 months to develop lesions—exact genotypes are provided at the top of the graph. Atherosclerotic plaque burden was quantified by computer image analysis of Oil Red O-stained aortic root sections with representative Oil Red O-stained aortic roots shown on right (scale bar, 0.4 mm; statistical significance of differences was calculated using Mann–Whitney U-test). (b) Macrophage content in aortic root sections was analysed by immunofluorescence staining using an antibody against MOMA-2 (n≥11 mice per group). (c) Apoptotic, necrotic core and combined apoptotic–necrotic core areas of aortic root sections were determined by quantification of TUNEL immunofluorescence staining, acellular areas or a combination of the two, respectively (n≥11 mice per group). (d) Immunofluorescence images of p62-KO and p62-KO/TFEB-TG macrophages treated with or without cholesterol crystals for 24 h and stained using DAPI (nuclei), and antibodies against polyubiquitinated proteins (FK-1) and p62 (scale bar, 5 μm). (e) Quantification of average ubiquitin intensity per cell from immunofluorescence experiment described in d (numbers of quantified cells are shown under each bar). (f) p62-KO and p62-KO/TFEB-TG macrophages were incubated with cholesterol crystals and the per cent of caspase 3/7-positive cells was quantified in three independent experiments (numbers of quantified cells are shown under each bar). (g) p62-KO and p62-KO/TFEB-TG macrophages were treated with LPS+CC for 24 h or with LPS followed by ATP for 3 h. Culture media were assayed for IL-1β by ELISA (n=3 independent wells for each treatment). For all graphs, data are presented as mean±s.e.m. *P<0.05, ***P<0.001, NS: not significant, two-tailed unpaired t-test, except a.

Figure 7. Trehalose induces autophagy and the transcription of autophagy–lysosomal genes in macrophages.

(a,b) Time course of serum and tissue trehalose levels from wild-type mice (n≥4) after trehalose administration (3 g kg−1 i.p.) by colorimetric method and mass spectrometry, respectively. (c) Macrophages were treated with vehicle or trehalose at indicated concentrations for 3 h and intracellular trehalose levels measured by mass spectrometry (n=2 independent wells). (d) GFP-LC3-expressing macrophages were imaged live every 30 s while being incubated in DMEM (control)±bafilomycin (200 nM), PBS (starvation) or trehalose (1 mM, 100 μM) for 20 min. GFP-LC3-positive area was quantified (n≥10 cells for each treatment) and plotted relative to 0 min. No significant difference seen for DMEM and 100 μM trehalose treatments. Trehalose (1 mM), bafilomycin and PBS significance is demarcated by *, # and ‡, respectively (P<0.05 for all cases). (e) Protocol as in d but macrophages were imaged live for 1 h (DMEM and 1 mM trehalose) or 2 h (100 μM trehalose). No significant difference seen for DMEM and 100 μM trehalose treatments. *P<0.05 for 1 mM trehalose treatment time points. (f) Wild-type macrophages were incubated with trehalose (1 mM; 3 and 6 h), bafilomycin (200 nM; 6 h) or both (6 h trehalose pretreatment and 6 h co-incubation) and stained with LC3 antibody and DAPI. Representative images are shown at left and quantification of average LC3 intensity with each condition at right (number of cells under each bar). #shows significant difference compared to vehicle (using analysis of variance (ANOVA) followed by Tukey's multiple comparison test; scale bar, 5 μm). (g) Wild-type macrophages were treated with 1 mM trehalose for indicated time points and transcripts of autophagy–lysosomal genes detected by qPCR (n≥3 independent wells for each gene). (h) Western blot analysis of Cathepsin D, Lamp1, p62 and LC3 in macrophages after 1 mM trehalose treatment for indicated times. Ponceau S staining used as loading control. For graphs in c,f,g, data are presented as mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001, NS: not significant, two-tailed unpaired t-test compared to zero time point or vehicle, except f.

Figure 8. Trehalose induces TFEB nuclear localization and protects from atherogenic lipid-induced protein aggregation and related sequelae of apoptosis and inflammasome activation.

(a) Western blot analysis of TFEB in macrophages after trehalose treatment for indicated times. Ponceau S staining is shown as loading control and densitometric quantification from three separate experiments is shown below. (b) TFEB nuclear localization is analysed by immunofluorescence staining after trehalose treatment and graphed as nuclear TFEB intensity (n≥40 cells per group, scale bar, 10 μm). (c) Western blot analysis of other MiTF transcriptional family members (TFE3 and MiTF) after trehalose or chloroquine (CHQ, 10 μM) treatments for indicated doses and times. Ponceau S staining is shown as loading control. (d,e) TFE3 and MiTF nuclear localization was analysed by immunofluorescence staining after trehalose treatment for indicated times and quantified by the intensity of nuclear staining (n≥500 cells per group). (f) Western blot analysis of polyubiquitinated proteins (FK-1 antibody) in detergent-soluble and detergent-insoluble lysate fractions of vehicle (V) or trehalose (T) treated wild-type macrophages. Lanes 3 and 4 were either vehicle or trehalose pretreated for 3 h, and then co-treatment with cholesterol crystals is performed for 12 h. Lanes 5 and 6 were cholesterol crystal-treated for 6 h and then either treated with vehicle or trehalose alone for another 6 h. (g) Densitometric quantification of f from three similar separate experiments. (h) Immunofluorescence images of wild-type macrophages after indicated treatments using DAPI and antibodies against polyubiquitinated proteins (FK-1) and p62 (scale bar, 5 μm). (i) Graphs represent average p62/ubiquitin+ dot numbers and average p62 intensity per cell for immunofluorescence images in h (numbers of quantified cells are shown under each bar). (j) Wild-type macrophages were co-incubated with cholesterol crystals and trehalose (or vehicle); the per cent of caspase 3/7-positive cells was quantified in three independent experiments (numbers of quantified cells are shown under each bar). (k) Wild-type macrophages were treated as indicated and cell culture media were assayed for IL-1β by ELISA (n=3 independent wells for each treatment). For all graphs data are presented as mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001, two-tailed unpaired t-test compared to zero time point or vehicle treatment group.

Figure 9. Trehalose administration in mice is atheroprotective.

(a,b) GFP-LC3 mice (ApoE-KO background) were fed western diet for 2 months and administered vehicle or trehalose (2 g kg−1 given five times per week i.p.) in the final 2 weeks of diet (n=4 mice per group). Representative GFP fluorescence (scale bar, 100 μm; a) and quantification of GFP intensity in aortic roots by confocal microscopy (n=4 mice per group; b) is shown. (c–e) TFEB intensity and nuclear localization in aortic roots of the same cohort used in a was detected by immunofluorescence. Shown are (c) representative aortic root TFEB staining (scale bar, 100 μm), (d) average aortic root TFEB intensity; and (e) TFEB-DAPI co-localization. (f) Diagram summarizing experimental protocol and mice cohorts used for in vivo assessment of trehalose in atherosclerosis. Mice were fed a western diet for 2 months while being administered either vehicle, trehalose or sucrose (disaccharides given both i.p. 2 g kg−1 for three times per week and orally 3% ad libitum in drinking water). (g) Serum cholesterol levels at 2 months of western diet (n=7 mice per group). (h) Quantification of atherosclerotic plaque burden by computer image analysis of Oil Red O-stained aortic root sections in the experiment summarized in f (statistical significance of differences was calculated using Mann–Whitney U-test). (i) Serum trehalose levels in wild-type mice (n=4 per group) after administration of trehalose (3 g kg−1) either by i.p. injection or oral gavage at indicated time points. (j) Quantification of atherosclerotic plaque burden by computer image analysis of Oil Red O-stained aortic root sections in mice fed 2 months of western diet while being administered only oral trehalose (statistical significance of differences was calculated using Mann–Whitney U-test). All data are presented as mean±s.e.m. *P<0.05, two-tailed unpaired t-test except h,j.

Figure 10. Atheroprotective effects of trehalose are dependent on macrophage autophagy and p62.

(a,e) mφATG5-KO (a) or p62-KO mice (e; all on ApoE-null background) were fed a western diet for 2 months with concurrent vehicle or trehalose administration (2 g kg−1 given three times per week i.p. and 3% ad libitum in drinking water). Graphs represent quantification of Oil Red O-stained atherosclerotic plaques at the level of aortic root (statistical significance of differences was calculated using Mann–Whitney U-test). (b,f) Immunofluorescence images of ATG5-KO (b) or p62-KO (f) macrophages using DAPI and antibodies against polyubiquitinated proteins (FK-1) and p62 (scale bar, 5 μm). Average p62 intensity (b) or ubiquitin intensity (f) per cell and number of cells is shown. (c,g) ATG5-KO (c) or p62-KO (g) macrophages were co-incubated with cholesterol crystals and trehalose (or vehicle). Per cent of caspase 3/7+ cells were quantified in three independent experiments (number of cells shown under each bar). (d,h) ATG5-KO (d) or p62-KO (h) macrophages were treated as indicated and cell culture media assayed for IL-1β by ELISA (n=3 independent wells for each treatment). Data are presented as mean±s.e.m. *P<0.05, NS: not significant, two-tailed unpaired t-test except a,e. (i) Graphical summary of the benefits of harnessing macrophage autophagy–lysosomal biogenesis in atherosclerosis: TFEB overexpression in macrophages initiates its nuclear localization and autophagy–lysosomal biogenesis (1 and 2). This activation modulates several downstream mechanisms in macrophages that contribute to a reduction in atherosclerosis (3 and 4), such as decreased IL-1β secretion (3), p62-dependent polyubiquitinated protein sequestration and autophagic clearance and decreased apoptosis (4). Our data largely implicate the atheroprotective effects of TFEB to be dependent on the aggrephagy of p62-enriched inclusion bodies and associated reductions in apoptotic cell death (4). The disaccharide trehalose is an inducer of macrophage autophagy and autophagy–lysosomal biogenesis and reduces atherosclerosis by recapitulating these TFEB-induced pathways (5).