Abstract

   Macrophages represent a major immune cell population in atherosclerotic
   plaques and play central role in the progression of this lipid-driven
   chronic inflammatory disease. Targeting immunometabolism is proposed as
   a strategy to revert aberrant macrophage activation to improve disease
   outcome. Here, we show ATP citrate lyase (Acly) to be activated in
   inflammatory macrophages and human atherosclerotic plaques. We
   demonstrate that myeloid Acly deficiency induces a stable plaque
   phenotype characterized by increased collagen deposition and fibrous
   cap thickness, along with a smaller necrotic core. In-depth functional,
   lipidomic, and transcriptional characterization indicate deregulated
   fatty acid and cholesterol biosynthesis and reduced liver X receptor
   activation within the macrophages in vitro. This results in macrophages
   that are more prone to undergo apoptosis, whilst maintaining their
   capacity to phagocytose apoptotic cells. Together, our results indicate
   that targeting macrophage metabolism improves atherosclerosis outcome
   and we reveal Acly as a promising therapeutic target to stabilize
   atherosclerotic plaques.

   Subject terms: Monocytes and macrophages, Cardiovascular diseases
     __________________________________________________________________

   Inhibition of the metabolic enzyme ATP-citrate lyase can attenuate
   atherosclerosis by preventing dyslipidemia and potentially also by
   reducing macrophage-mediated inflammation. Here, the authors show that
   specific targeting of ACLY in macrophages results in more stable
   atherosclerotic plaques.

Introduction

   Atherosclerosis is a lipid-driven chronic inflammatory disorder of the
   arteries in which macrophages are the most prevalent immune cells and
   define disease development^[62]1,[63]2. One of the crucial initiating
   events in atherosclerosis is the infiltration and accumulation of
   cholesterol-rich lipoproteins within the intima of the arterial vessel
   wall. Subsequent modification of such lipoproteins induces an
   inflammatory response and as a result, monocytes are recruited to the
   vessel wall and differentiate into macrophages. Although macrophages
   can worsen disease progression by propagating inflammation, they can
   also stabilize atherosclerotic plaques by promoting the formation of a
   fibrous cap and by clearing apoptotic cells (ACs) to prevent necrotic
   core formation. Uptake of cholesterol gives macrophages their foamy
   appearance and this can lead to apoptosis^[64]3. Initially, ACs are
   efficiently cleared by neighboring macrophages in a protective process
   termed efferocytosis^[65]4. In advanced atherosclerosis, insufficient
   efferocytosis induces AC accumulation and necrotic core formation. In
   combination with thinning of the protective fibrous cap this can make
   plaques vulnerable to rupture and can cause cardiovascular events such
   as myocardial infarction and stroke^[66]5.

   Given the prominent role of macrophages in the progression of
   atherosclerosis, modulating their responses is considered as a
   promising strategy to treat atherosclerosis^[67]6. The last decade,
   modulation of intracellular metabolic pathways has emerged as a new
   tool to reshape deranged macrophage functions and is considered as a
   new therapeutic opportunity^[68]6–[69]9. Indeed, recent
   immunometabolism research highlights that intracellular metabolic
   reprogramming is a key controller of macrophage activation^[70]7. High
   rates of glycolysis supports the inflammatory properties op macrophages
   that are activated with lipopolysaccharide (LPS), whereas increased
   mitochondrial oxidative phosphorylation supports interleukin
   (IL)-4-induced macrophage responses. Recent literature revealed an
   important role for ATP citrate lyase (Acly) in translating metabolic
   changes into altered macrophage phenotype and function^[71]10–[72]12.

   Acly is a key metabolic enzyme that converts mitochondria-derived
   citrate into acetyl-CoA and oxaloacetate within the cytosol.
   Acly-dependent acetyl-CoA incorporation into histone promotes
   chromosome accessibility and regulates both LPS- and IL-4-induced
   macrophage activation^[73]10,[74]11. In addition, acetyl-CoA fuels the
   synthesis of fatty acids, cholesterol, and the acetylation of
   non-histone proteins^[75]13,[76]14. Serving as a link between
   carbohydrate and lipid metabolism makes Acly a promising target to
   lower low-density lipoprotein (LDL) cholesterol levels and to reduce
   cardiovascular risk^[77]15–[78]17. Bempedoic acid is a competitive Acly
   inhibitor that is specifically activated in hepatocytes where it
   upregulates LDL receptor (LDLr) expression. Although promising data
   from clinical trials highlight that hepatic Acly inhibition reduces LDL
   cholesterol levels in hypercholesterolemic patients, the tools to study
   this key metabolic hub in macrophages in vivo remained absent.

   Here, we first demonstrate that Acly is activated in inflammatory
   macrophages and in human atherosclerotic plaques. Next, we define its
   formerly unknown role in regulating both plaque and macrophage
   phenotype. Using a conditional genetic knockout mouse model, we find
   that Acly deficiency in myeloid cells induces a stable plaque phenotype
   as demonstrated by increased collagen content and fibrous cap
   thickness, along with a decreased necrotic core size. Further
   functional, lipidomic, and transcriptional characterization in vitro
   show that this in vivo plaque phenotype is linked to deregulated fatty
   acid and cholesterol biosynthesis and reduced liver X receptor (LXR)
   activation within the macrophages. This results in macrophage
   apoptosis, along with more efficient efferocytosis and increased
   clearance ACs. As such, we show that targeting macrophage Acly improve
   atherosclerosis outcome and can serve as a promising therapeutic target
   to stabilize atherosclerotic plaques.

Results

Acly is activated in inflammatory conditions

   Acly can be regulated at distinct levels, with phosphorylation at
   serine 455 serving as a key posttranslational modification to promote
   its enzymatic activity^[79]18. To study Acly expression and activity in
   inflammatory conditions, we first stimulated bone marrow-derived
   macrophages (BMDMs) with toll-like receptor 4 agonist (LPS for 24 h as
   a prototypical model of inflammatory (also called classical or M1)
   activation. Although Acly gene expression and protein levels remained
   unaffected, we observed a marked increase of phosphorylated Acly
   (p-Acly) in inflammatory macrophages (Fig. [80]1a, b). As these data
   suggested that Acly activation might support inflammatory responses we
   next examined Acly phosphorylation within human atherosclerotic
   plaques. Whereas total Acly was present in most cells in and around the
   plaque, activated (p-)Acly predominantly colocalized with CD45^+CD68^+
   macrophages in human atherosclerotic plaques (Fig. [81]1c, d,
   Supplementary Fig. [82]1a–c). In addition, unstable plaques showed
   increased abundance of activated Acly when compared to stable plaques
   (Fig. [83]1e). Rupture-prone plaque areas are known to be dominated by
   inflammatory macrophages^[84]19 and together with our observation that
   the levels of activated p-Acly were increased in inflammatory
   macrophages, this prompted us to study whether targeting macrophage
   Acly could improve atherosclerosis outcome.

Fig. 1. Acly is phosphorylated in inflammatory conditions in vitro and in
vivo.

   [85]Fig. 1
   [86]Open in a new tab

   a Relative normalized expression of Acly in unstimulated and
   LPS-stimulated macrophages. b Normalized protein levels from cell
   lysates of unstimulated and LPS-stimulated macrophages. Samples were
   immunoblotted with antibodies against Acly, phosphorylated Acly
   (p-Acly) and α-tubulin. Acly/p-Acly quantification on the blots derive
   from samples of the same experiment and gels/blots were processed in
   parallel. *P = 0.0441. c Representative immunohistochemical staining
   for macrophages (CD68) and p-ACLY in human plaques from 16
   stable/unstable plaques. Scale bar represents 100 µm. d Quantification
   of colocalization, percentage of p-Acly^+ macrophages overlapping with
   CD68^+ area. e Quantification of p-Acly^+ area in the lesion.
   *P = 0.0402. Values represent mean ± SEM (n = 3 technical replicates of
   three pooled mice a, n = 3 one representative image of three technical
   replicates of three pooled mice (b, western blot), n = 7/9
   stable/unstable plaques d, e. *P < 0.05; by two-tailed Student’s t test
   (b, e). Source data are provided as a Source Data file (a–e).

Macrophage Acly deletion stabilizes atherosclerotic plaques

   To examine the role of macrophage Acly in atherogenesis, we crossed
   Acly^fl/fl mice^[87]20 with mice expressing Cre recombinase under
   control of the Lyz2 promoter/enhancer regions (LysMcre) to generate a
   new genetic mouse model that specifically lacks Acly in the myeloid
   cell lineage (macrophages, monocytes, and neutrophils). After
   confirming the knockdown of ACLY in macrophages, myeloid dendritic
   cells, and neutrophils from those mice (hereafter referred to as
   Acly^M-KO, Supplementary Fig. [88]1d, e), we transplanted lethally
   irradiated atherosclerosis-susceptible Ldlr-deficient mice with bone
   marrow from either Acly^M-KO mice or Cre-negative Acly^fl/fl littermate
   control mice. After validating efficient engraftment in both
   experimental groups (Supplementary Fig. [89]1f), the mice were fed a
   high-fat diet (HFD) (Fig. [90]2a). All mice exhibited a similar
   increase in body weight and plasma cholesterol, and similar end-state
   hepatic triglyceride levels, whereas the increase of plasma
   triglycerides was attenuated in Acly^M-KO mice (Supplementary
   Fig. [91]1g–j). Also, both groups showed a comparable blood leukocyte
   composition with only B-cell levels being slightly lower in Acly^M-KO
   mice (Supplementary Fig. [92]1k–o). As splenic B-cell levels remained
   unaltered, there is no apparent general defect in B-cell development or
   activation. It rather suggests an indirect, and yet unknown, effect of
   myeloid Acly knockdown on circulating B cells.

Fig. 2. Acly deletion in macrophages elicits a favorable plaque phenotype.

   [93]Fig. 2
   [94]Open in a new tab

   a In order to asses atherogenesis, bone marrow cells of either control
   of Acly^M-KO were transplanted in lethally irritated Ldlr^−/− mice. Six
   weeks after transplantation, mice were put on a HFD for 10 weeks. b
   Representative toluidine blue staining of atherosclerotic plaques
   within the aortic roots of control and Acly^M-KO mice. Scale bar
   represents 500 μm. c Quantification of plaque size. d Plaque phenotypes
   as classified by an experimental pathologist. e Toluidine blue staining
   of atherosclerotic lesions with necrosis indicated in red. Scale bar
   represents 200 μm. f Quantification of the necrotic area. *P = 0.0173.
   g Sirius red staining of plaques. Scale bar represents 200 μm. h
   Quantification of the collagen deposition. *P = 0.0201. i Minimal cap
   thickness was measured at the thinnest region of the fibrotic cap
   surrounding the necrotic core as indicated by the arrows in g.
   *P = 0.0292. j Quantification of MOMA-2^+ area for macrophages. k
   Quantification of Ly6G^+area for neutrophils. l Normalized expression
   of Tgfb1 in aortic arches of control and Acly^M-KO mice. **P = 0.0038 m
   Representative immunohistochemical analysis of macrophages (MOMA-2) and
   TGF-β in mouse plaques from control or Acly^M-KO mice (n = 6/6 ctrl/KO
   from one experiment). Scale bar represents 100 µm. n Quantification of
   colocalization, percentage of TGF-β^+ area in MOMA-2^+ area. o
   Quantification of TGF-β^+ area in the lesion. Values represent
   mean ± SEM (n = 20/17 (ctrl/KO in b–l), n = 6/6 (ctrl/KO in n, o).
   *P < 0.05; **P < 0.01 by two-tailed Student’s t test (f, h, i, l).
   Source data are provided as a Source Data file (c, d, f, h–l, m, o).

   After 10 weeks of HFD, immunohistochemical and gene expression analysis
   on lesions confirmed Acly knockdown in myeloid cells within the
   atherosclerotic plaques of Acly^M-KO mice (Supplementary Fig. [95]2a,
   b). Although the atherosclerotic plaque size was similar, pathological
   scoring of plaques indicated an increase in thick fibrous cap atheromas
   in the Acly^M-KO group (Fig. [96]2b–d). The observation that the
   necrotic area was significantly smaller in Acly^M-KO-transplanted mice
   (Fig. [97]2e, f) indicated that the lesions of those mice are more
   stable as necrotic core formation is associated with plaque instability
   and rupture^[98]5. Moreover, analysis of Sirius red staining
   highlighted increased collagen content and fibrous cap thickness in
   plaques of Acly^M-KO-transplanted mice (Fig. [99]2g–i), another sign of
   plaque stability. Immunohistochemical analysis revealed that the
   increase in collagen deposition was not accompanied by differences in
   macrophage and neutrophil content in the lesions (Fig. [100]2j, k,
   Supplementary Fig. [101]2c). Plasma triglyceride levels did not
   correlate with the plaque characteristics, indicating that lower
   triglyceride levels were not the cause of the more stable plaque
   phenotype (supplementary Fig. [102]2d). Transforming growth factor beta
   (TGF-β) is a key anti-atherogenic cytokine that can suppress immune
   effector functions and stabilizes atherosclerotic lesions by inducing
   fibrotic activities of smooth muscle cells and fibroblasts^[103]21. We
   therefore assessed its expression in aortic arches and detected
   increased Tgfb1 expression in Acly^M-KO-transplanted mice
   (Fig. [104]2l). As labeling of TGF-β and the macrophage marker MOMA-2
   suggested an increase of macrophage TGF-β but not total TGF-β
   (Fig. [105]2m–o), the increased Tgfb1 levels were probably caused by
   macrophages within the lesion. Collectively, these in vivo data
   demonstrate that Acly^M-KO-transplanted mice had a favorable plaque
   phenotype, reflected by decreased necrosis and increased collagen
   deposition and fibrous cap thickness.

Acly regulates in vitro macrophage polarization

   A possible mechanism to explain the more fibrotic plaque phenotype in
   the Acly^M-KO mice is a change in macrophage polarization. Although
   inflammatory (M1) macrophages can promote atherogenesis, IL-4-induced
   macrophages (M2) are regarded as atheroprotective because of their
   anti-inflammatory and pro-fibrotic homeostatic properties^[106]1.
   Previous in vitro studies with inhibitors suggested that Acly is
   involved in both inflammatory and IL-4-induced macrophage
   polarization^[107]10,[108]22,[109]23. Therefore, we next sought to
   reassess these findings with the aim to study whether altered
   macrophage polarization could clarify the differences in plaque
   phenotypes (Fig. [110]3a).

Fig. 3. Altered macrophage polarization in vitro does not explain the plaque
phenotype in vivo.

   [111]Fig. 3
   [112]Open in a new tab

   a Differentiated control and Acly^M-KO macrophages (BMDMs) were
   stimulated with LPS. b Expression of LPS-induced surface markers as
   measured by flow cytometry. Representative histograms and quantified
   surface expression (ΔMFI = [median fluorescence intensity][positive
   staining] – [MFI][pooled control]) are shown. CD80 *P = 0.0160. c
   Relative normalized gene expression of indicated LPS-induced genes. Il6
   ***P < 0.0001, Nos2 ***P < 0.0001, Ccl2 ***P = 0.0002. d Production of
   IL-6 ***P < 0.0001, TNF *P = 0.0281, NO ***P < 0.0001 and ROS
   ***P < 0.0001. e Normalized expression of M1 genes in aortic arches of
   control and Acly^M-KO-transplanted mice. Il1b **P = 0.0030, Tnf
   *P = 0.0183, Ccl2 *P = 0.0321. Values represent mean±SEM of n = 3 (b–d)
   technical replicates of one out of three representative experiments or
   n = 20/17 (ctrl/KO in e). *P < 0.05; **P < 0.01, ***P < 0.001 by
   ordinary one-way ANOVA with Bonferroni post hoc test for multiple
   comparisons (b–d) or two-tailed Student’s t test (e). Source data are
   provided as a Source Data file (b–e).

   Distinct macrophage subsets differentially affect plaque phenotype. As
   previous inhibitor studies reported an involvement of Acly in both M1
   and M2 macrophage polarization^[113]10,[114]11, we next aimed to
   investigate whether altered macrophage polarization in the absence of
   Acly could clarify the differences in plaque phenotypes (Fig. [115]3a).
   Hereto, we first stimulated macrophages from control and Acly^M-KO mice
   with LPS to elicit an inflammatory macrophage response. Surprisingly,
   our genetic Acly^M-KO model indicated that Acly is not needed for
   inflammatory macrophage responses. The LPS-induced surface markers
   CD40, CD86, and major histocompatibility complex (MHC)-II were
   upregulated to the same extend in both groups and CD80 expression was
   higher in Acly^M-KO macrophages (Fig. [116]3b). Moreover, the
   LPS-induced levels of pro-inflammatory genes and factors such as IL-6,
   tumor necrosis factor (TNF), nitric oxide (NO), and reactive oxygen
   species (ROS) were even increased in the absence of Acly (Fig. [117]3c,
   d). Likewise, we detected increased inflammatory gene expression in
   vivo in the atherosclerotic lesions and peritoneal macrophages of the
   Acly^M-KO group (Fig. [118]3e, Supplementary Fig. [119]2e).

   Next, we assessed the effect of Acly deficiency on IL-4-induced
   macrophage polarization (Fig. [120]4a). Hereto, we measured the
   IL-4-elicited expression of commonly used surface markers and detected
   reduced levels of CD206, CD273, and CD301 on the surface of
   Acly-deficient macrophages, along with blunted M2-associated gene
   expression (Fig. [121]4b, c). This decreased IL-4 response may be
   explained by lower histone 3 lysine 27 (H3K27) acetylation in the
   absence of Acly and confirms a previous study that applied Acly
   inhibitors during IL-4-responses in macrophages to link Acly-mediated
   production of acetyl-CoA to histone acetylation^[122]10. Conversely,
   H3K27 levels were similar in naive and LPS-treated control and
   Acly^M-KO macrophages and thus histone acetylation is not the mechanism
   explaining the macrophage phenotype in this setting (Supplementary
   Fig. [123]2f). Together, the enhanced inflammatory activation and
   reduced IL-4 responses in the absence of Acly do not explain the
   advantageous plaque phenotype of Acly^M-KO-transplanted mice. Our data
   highlight that M1/M2 polarization in vitro not necessarily reflects the
   in vivo context and underscore that caution is necessary when
   interpreting inhibitor effects.

Fig. 4. IL-4 response is decreased after Acly deletion in macrophages.

   [124]Fig. 4
   [125]Open in a new tab

   a Differentiated control and Acly^M-KO macrophages (BMDMs) were
   stimulated with IL-4. b Expression of IL-4-induced surface markers as
   measured by flow cytometry. Representative histograms and quantified
   surface expression (ΔMFI = [median fluorescence intensity][positive
   staining] – [MFI][pooled control]) are shown. CD206 ***P < 0.0001,
   CD273 ***P = 0.0003, CD301 ***P = 0.0008. c Relative normalized gene
   expression of IL-4-induced genes. Arg1 ***P < 0.0001, Retnla
   ***P = 0.0004, Ym1 ***P < 0.0001. Values represent mean ± SEM of n = 3
   b or n = 4 c technical replicates of one out of three representative
   experiments. ***P < 0.001 by ordinary one-way ANOVA with Bonferroni
   post hoc test for multiple comparisons (b, c). Source data are provided
   as a Source Data file (b, c).

Acly does not regulate core metabolic processes

   To identify the mechanism by which myeloid deletion of Acly influences
   macrophage and plaque phenotype, we explored changes in core metabolic
   pathways as a potential explanation of the increased inflammatory
   response in Acly-deficient macrophages.

   Metabolomics revealed that the expected LPS-induced accumulation of
   citrate was more pronounced in the absence of Acly, signifying
   decreased Acly-mediated conversion of citrate into
   acetyl-CoA^[126]24,[127]25. Yet, loss of Acly led to no further changes
   in the abundance of other tricarboxylic acid cycle (TCA) cycle
   intermediates (Fig. [128]5b). Extracellular flux analysis, and glucose,
   lactate and ATP measurements showed no differences in mitochondrial
   function, glycolysis and ATP production, indicating that other
   mechanisms should explain the observed inflammatory phenotypes of
   Acly-deficient macrophages (Fig. [129]5c–g).

Fig. 5. Mitochondrial function remains unaltered in Acly^M-KO BMDMs.

   [130]Fig. 5
   [131]Open in a new tab

   a Differentiated control and AclyM-KO macrophages (BMDMs) were left
   untreated or stimulated with LPS (b) relative abundance of TCA cycle
   intermediates ***P  <  0.0001. c Lactate secretion after 24 h LPS
   treatment. d Metabolized glucose in 24 h. e Cellular energy charge as
   calculated by [ATP] + .05[ADP]/([AMP] + [ADP] + [ATP]). f Cellular ATP
   concentration. g Seahorse analyses and extracted parameters. Values
   represent mean ± SEM from n  = 12 technical replicates from two
   combined experiments b, n =  3 technical replicates from three pooled
   mice from one of three representative experiments (c, d, g), n = 6
   technical replicates from three pooled mice from one of two
   representative experiments (e, f). ***P < 0.001 by two-way ANOVA with
   Bonferroni post hoc test for multiple comparisons. Source data
   are deposited in  MTBLS2159 (b) or provided as a Source Data file
   (c–g).

Deregulated cholesterol metabolism in the absence of Acly

   To reveal such mechanism, we next performed RNA-sequencing on BMDMs
   from control and Acly^M-KO mice. Plotting genes in a volcano plot and
   performing pathway analysis on the differentially expressed genes
   identified cholesterol biosynthesis as the most deregulated pathway,
   with most targets related to this pathway being upregulated in
   Acly^M-KO macrophages (Fig. [132]6a, b). In addition to genes involved
   in cholesterol metabolism,^[133]10 fatty-acid metabolism, and cell
   cycle genes were differentially expressed. Distinct genes involved in
   cholesterol and fatty-acid import and transport (Ldlr, Cd36, Scarb1,
   Fabp3, Slc27a4) were increased in the absence of Acly. In contrast,
   Acly^M-KO macrophages showed decreased expression of LXR-regulated
   genes that promote cholesterol efflux (Abca1, Abcg1, Apoe) or limit
   cholesterol uptake (Mylip) (Fig. [134]6c). The most significantly
   increased gene in Acly^M-KO macrophages was Dhcr24, encoding the enzyme
   that converts desmosterol into cholesterol as the terminal step in
   cholesterol synthesis (Fig. [135]6d). In line with increased Dhcr24
   expression, we measured lower levels of desmosterol and
   desmosterol-induced genes in Acly^M-KO macrophages (Fig. [136]6e, f).
   By increasing cholesterol synthesis and import, and via limiting its
   export, Acly-deficient macrophages possibly try to cope with the
   reduced supply of Acly-mediated acetyl-CoA, and by doing so they manage
   to secure total cholesterol levels (Fig. [137]6g). Moreover,
   Acly-deficient macrophages might rescue acetyl-CoA production by
   upregulating Acss2 (Fig. [138]6h). This gene encodes acyl-coenzyme A
   synthetase short-chain family member 2, an enzyme that converts acetate
   into acetyl-CoA. This is in line with earlier reports that describe
   induction of Acss2 in adipocytes and mouse embryonic fibroblasts upon
   Acly deletion^[139]20,[140]26.

Fig. 6. Acly^M-KO macrophages exhibit deregulated cholesterol biosynthesis.

   [141]Fig. 6
   [142]Open in a new tab

   a Volcano plot with up- or downregulated genes in Acly-deficient
   macrophages (BMDMs) highlighted in red or blue, respectively. b
   Ridgeline plot of significantly enriched Reactome pathways. Log[2]FC
   distributions of member genes are plotted on the x axis. Color of the
   distributions represents enriched pathway P value. c Expression of
   genes involved in import/transport of cholesterol and fatty acids, and
   genes that promote cholesterol efflux. d Schematic view of cholesterol
   metabolism and the effect of Acly deficiency (red indicating
   upregulation; blue downregulation). e Desmosterol levels in control and
   Acly-deficient macrophages. f Genes previously found to be induced by
   desmosterol, plotted in a heatmap for control and Acly^M-KO
   macrophages. g Cholesterol levels in control and Acly-deficient
   macrophages. h Acss2 expression in control and Acly-deficient
   macrophages. Values represent mean from n = 2 technical replicates from
   three pooled mice e, g, or mean ± SEM n = 3 technical replicates from
   three pooled mice (h). Source data are provided as a Source Data file
   e, g or deposited in [143]GSE126690 (a–d, f, h).

   The resulting deregulated cholesterol metabolism and suppressed
   desmosterol levels could explain the lower expression of LXR-target
   genes (Fig. [144]6f) and possibly also altered polarization
   (Figs. [145]3, [146]4) in Acly^M-KO macrophages. Indeed, LXR activation
   is involved in reprogramming fatty-acid metabolism and inhibiting
   LPS-induced inflammatory-response genes in macrophages^[147]27.
   Supporting this hypothesis, exposing macrophages to LXR agonist GW3965
   decreased inflammatory responses in both control and Acly^M-KO
   macrophages (Supplementary Fig. [148]3a). As such, reduced LXR
   activation could provide one potential mechanistic explanation for the
   elevated inflammatory responses of Acly-deficient macrophages.

Fatty-acid synthesis is affected in the absence of Acly

   Apart from cholesterol metabolism, reactome pathway analysis also
   indicated “fatty acyl-CoA biosynthesis” and “fatty-acid metabolism” as
   top deregulated pathways (Fig. [149]6b). As Acly provides acetyl-CoA as
   a precursor for fatty acid synthesis^[150]28,[151]29, we measured fatty
   acid levels in control and Acly^M-KO macrophages by performing
   lipidomics measurements. Whereas RNA-sequencing analysis suggested a
   lipid-laden phenotype of Acly^M-KO macrophages, total lipid pools were
   similar in control and Acly^M-KO macrophages (Fig. [152]7a, b).
   Conversely, levels of omega 3 and 6 fatty acids, including arachidonic
   acid (AA) tended to be increased in Acly^M-KO macrophages
   (Fig. [153]7c, d). Cyclooxygenase (COX)-1 and COX-2 produce
   prostaglandins from AA and these lipid mediators are known to affect
   inflammatory responses^[154]30. The genes encoding COX-1 and COX-2
   (Ptgs1 and Ptgs2) and downstream prostaglandins (PGD[2], PGF[2a], and
   in particular PGE[2]) were deregulated in Acly^M-KO macrophages
   (Fig. [155]7d). Moreover, Acly^M-KO macrophages showed a decreased
   expression of prostaglandin E 4 receptor (EP4, Ptger4) (Fig. [156]7e).
   Loss of EP4 has previously been described to enhance inflammatory
   responses^[157]31,[158]32. Subsequently, PGE[2]- and LPS-treated
   Acly^M-KO macrophages display a reduced responsiveness to PGE[2] when
   compared to control macrophages (Fig. [159]7f, g).

Fig. 7. Deletion of Acly leads to deregulation of lipid mediators and
decreased responsiveness to prostaglandin E2.

   [160]Fig. 7
   [161]Open in a new tab

   a Total sum of measured lipids by metabololipidomics. b Neutral lipids
   analyzed by Bodipy 493/503. c Sum of AA, DHA, EPA, AdA, LA, ALA/GLA,
   DPAn-3, and DGLA omega 3 and 6 fatty acids (FA’s) measured by
   metabololipidomics. d Lipid abundances and genes involved in
   prostaglandin synthesis. e Ptger4 expression in control and
   Acly-deficient macrophages. f Effect of PGE2 treatment on cytokine
   secretion in control and Acly^M-KO LPS-treated BMDMs. g Effect of PGE2
   treatment on cytokine secretion in control and Acly^M-KO LPS-treated
   BMDMs. Values represent mean ± SEM (n = 3 technical replicates from
   three pooled mice a–e), or mean (n = 2 technical replicates from three
   pooled mice (f, g)) Source data are provided as a Source Data file
   (a–d, f) or deposited in [162]GSE126690 (d, e).

   Together, these findings indicate that Acly^M-KO macrophages rewire
   cholesterol and fatty-acid metabolism to deal with the absence of Acly,
   and these metabolic changes are accompanied by altered inflammatory
   responses.

Acly^M-KO macrophages show increased apoptosis and efferocytosis

   In addition to cholesterol and fatty acid metabolism, pathway analysis
   identified “cell cycle checkpoints” and “TP53 regulates transcription
   of cell cycle genes” among the most-regulated pathways (Fig. [163]6b),
   with most genes related to these pathways being increased in Acly^M-KO
   macrophages (Fig. [164]8a). As the transcription factor p53 is a key
   modulator of both apoptosis and senescence, we assessed those cell
   cycle-related processes more detailed. We stained DNA of control and
   Acly^M-KO macrophages with propidium iodide and measured increased
   fractional DNA content (Sub-G1), a hallmark of ACs, in Acly^M-KO
   macrophages (Fig. [165]8b). Moreover, β-galactosidase activity
   measurements indicated increased senescence in Acly^M-KO macrophages
   (Fig. [166]8c). Supporting those data, the expression of p53 target
   genes related to cell cycle mostly increased in the absence of Acly
   (Fig. [167]8a). Noteworthy, the expression of the LXR-dependent
   anti-apoptosis genes Cd5l (AIM), Lbp, and Mafb (all known to promote
   macrophage survival^[168]33–[169]35) was reduced in the absence of
   Acly, providing a potential link between reduced LXR activation
   (Fig. [170]6f) and increased apoptosis in these cells. Increased
   apoptosis within atherosclerotic lesions needs to be paralleled by
   increased clearance of those cells to prevent the growth of necrotic
   cores and plaque destabilization. Therefore, we questioned whether
   improved efferocytosis of apoptotic macrophages could explain the
   smaller necrotic cores in atherosclerotic lesions of Acly^M-KO mice.
   Hereto, we co-cultured control or Acly^M-KO BMDMs with apoptotic
   macrophages that were labeled with the pH-sensitive dye pHrodo and
   measured efferocytosis. Culturing the cells at 4 °C served as a
   negative control to set the pHrodo^+ gate and effective internalization
   of ACs was also validated by imaging flow cytometry (Fig. [171]8d). To
   validate these results in vivo, we analyzed apoptosis and efferocytosis
   in the atherosclerotic plaques from control and Acly^M-KO mice by
   terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)
   double staining with MAC3. In vivo apoptosis, but not efferocytosis,
   appeared to be increased in plaques from Acly^M-KO mice (Fig. [172]8e,
   f). These assays highlighted that Acly-deficient macrophages
   demonstrate more-efficient efferocytosis in vitro, an atheroprotective
   macrophage function associated with stabilized atherosclerotic plaques
   in vivo (Fig. [173]8d).

Fig. 8. Deletion of Acly results in increased apoptosis, growth arrest, and
more-efficient efferocytosis.

   [174]Fig. 8
   [175]Open in a new tab

   a Expression of genes associated with indicated reactome pathways. b
   Flow cytometric analysis of control and Acly^M-KO macrophages stained
   with propidium iodide to asses cell cycle distribution. ***P = 0.0009,
   *P = 0.0385. c Senescence in control and Acly^M-KO BMDMs by
   b-galactosidase activity measurement. Scale bar represents 100 µm.
   **P = 0.0034. d In vitro phagocytosis of pHrodo-stained apoptotic cells
   by DiD-labeled control or Acly^M-KO BMDMs measured by flow cytometry
   after incubation at 37 °C (or 4 °C as negative control) and validated
   by imaging flow cytometry (Imagestream, scale bar represents 7 µm).
   Quantification of efferocytosis at 37 °C as a percentage of
   DiD-positive cells that incorporated an apoptotic cell. ***P = 0.0002.
   e Quantification of TUNEL^+ cells in lesions of Ldlr^−/− mice
   transplanted with control or Acly^M-KO bone marrow. *P = 0.0203 (f).
   Ratio of TUNEL^+ cells bound to MAC3^+ macrophages and free TUNEL^+
   cells. Values represent mean ± SEM (n = 3 technical replicates of one
   out of three representative independent experiments in b–d) and
   n = 20/19 (ctrl/KO in e, f)). *P < 0.05; **P < 0.01, ***P < 0.001 by
   two-way ANOVA with Bonferroni post hoc test for multiple comparisons b
   or two-tailed Student’s t test (c, d). Source data are provided as a
   Source Data file (b–f) or deposited in [176]GSE126690 (a).

Discussion

   Targeting ACLY in liver cells has emerged as an attractive aid to lower
   serum LDL-C levels in hypercholesteremic patients^[177]16,[178]17. In
   the present study, we demonstrate that myeloid deficiency for Acly
   deregulated fatty-acid and cholesterol metabolism and leads to enhanced
   apoptosis, combined with improved efferocytosis. In vivo, this results
   in more stable atherosclerotic plaques, reflected by smaller necrotic
   cores and more collagen deposition. As such we identify targeting
   macrophage Acly as a potential new therapeutic avenue in cardiovascular
   disease.

   Noteworthy, Acly targeting with bempedoic acid (ETC-1002, a prodrug
   activated by acyl-CoA synthetase solute carrier family 27 member 2
   (SLC27A2) in the liver) is already applied in the clinic as a new aid
   to lower serum LDL-C in hypercholesteremic patients^[179]36,[180]37.
   Acly inhibition reduces cholesterol synthesis in hepatocytes, leading
   to the upregulation of LDLR expression and clearance of LDL-C particles
   from the blood^[181]38. Similar as in liver cells, Acly deficiency in
   macrophages leads to the upregulation of Ldlr, probably to compensate
   for deregulated cholesterol biosynthesis. In addition to elevated
   expression of the LDL receptor, loss of Acly enhances the expression of
   CD36 and SR-B1. Besides their role in cholesterol uptake, both class B
   scavenger receptors are key mediators of efferocytosis^[182]39,[183]40.
   Impaired efferocytosis drives the accumulation of necrotic debris in
   atherosclerotic plaques and can lead to plaque rupture and
   cardiovascular events. In accordance with increased CD36 and SR-B1
   expression in Acly^M-KO macrophages, the capacity of these cells to
   clear ACs was improved. Silencing of ACLY was recently found to
   suppress DNA damage repair, resulting in genomic instability and cell
   death^[184]41. In agreement, Acly-deficient macrophages were more
   senescent and prone to apoptosis. Together, increased apoptosis and
   more effective clearance of such apoptotic macrophages may explain the
   reduced necrotic core formation in plaques of Acly^M-KO mice. Along
   with reduced necrotic core formation, collagen deposition was increased
   in the plaques of Acly^M-KO-transplanted mice. In line with the
   increase in collagen content and fibrous cap thickness, the plaques of
   those mice showed increased mRNA levels of TGF-β1, a key pro-fibrotic
   cytokine which is secreted by macrophages upon efferocytosis^[185]42.
   Thus, efferocytosis might trigger the synthesis of TGF-β1 that drives
   the increased collagen deposition in plaques of myeloid Acly-deficient
   mice. Interestingly, although efferocytosis is generally believed to be
   an anti-inflammatory process, we observed augmented expression of
   inflammatory genes in plaques of Acly^M-KO mice. Moreover, cultured
   Acly-deficient macrophages were found to be hyperinflammatory. We found
   that the expression of LXR-target genes and abundance of main LXR
   agonist desmosterol were diminished in both unstimulated and
   LPS-stimulated Acly-deficient macrophages. In addition to their role in
   cholesterol efflux, LXRs are established negative regulators of
   inflammatory macrophage responses^[186]43. Therefore, blunted LXR
   activation provides a potential mechanistic explanation for the
   elevated inflammatory responses of Acly-deficient macrophages. In
   addition, LXR activation is believed to promote macrophage
   survival^[187]33,[188]35. LXR-target genes Cd5l (AIM), Lbp and Mafb
   have been found to inhibit apoptosis and promote macrophage survival
   and were all downregulated in Acly-deficient macrophages. Therefore,
   diminished LXR activation might contribute to the increase in apoptosis
   upon Acly deletion. Interestingly, Lauterbach et al.^[189]11 recently
   applied Acly inhibitors and siRNA-mediated knockdown to demonstrate the
   role of Acly in regulating acute LPS responses via histone acetylation.
   The difference between our observations and the ones of Lauterbach et
   al. might be explained by the knockdown method (i.e., short-term siRNA
   vs long-term LysMCre-driven) and other experimental variables. Clearly,
   the role of Acly in macrophages is stimulus- and time-dependent and
   also appears to depend on the experimental setting as Namgaladze et al.
   observed no effect of Acly silencing^[190]44. As such, this is an area
   that warrants future research.

   Collectively, we demonstrate that macrophage-specific targeting of Acly
   elicits a favorable atherosclerotic plaque phenotype. With
   liver-specific inhibition of ACLY already being applied as a
   cholesterol-lowering aid, we anticipate our observations in macrophages
   to spike the development of new cell-specific ACLY inhibitors that
   target different cell types to tackle the disease from distinct angles.

Methods

Human tissue collection

   Anonymised human plaque tissues for the determination of protein
   expression were obtained from the Maastricht Pathology Tissue
   Collection (MPTC) under approval of the Medical Ethical Committees
   AZm/UM from the Maastricht University Medical Centre. The
   formalin-fixed paraffin-embedded human carotid samples used for
   immunohistochemistry were collected at autopsy (n = 16, mean age 72.6
   years, all men). Collection, storage in the MPTC and use of tissues
   were performed in agreement with the “Dutch Code for Proper Secondary
   use of Human Tissue”. None of the co-authors was involved in tissue
   collection during autopsy. No prior informed consent was directly
   obtained from the patients, but an opt-out arrangement was in place and
   hence tissues were not used in case of objection by relatives in line
   with the guidelines provided by the Dutch Code for Proper Secondary use
   of Human Tissue. Samples represented the following stages of
   atherosclerosis (n = 7/9 per. respectively, stable/unstable group):
   pathological intimal thickening and plaque with intraplaque
   hemorrhage^[191]45.

Mice experiments

   C57Bl/6J mice with loxP sites flanking exon 9 of the Acly gene
   (Acly^fl/fl)^[192]20 were crossed with Lyz2-Cre transgenic mice to
   generate mice with a myeloid-specific deletion of Acly (Acly^M-KO).
   Ldlr^−/− mice on a C57BL/6 background were purchased from Jackson
   Laboratories. All mice were maintained in an animal facility at an
   ambient temperature of 20–24 °C with 40–70% relative humidity under a
   12:12 h light–dark cycle. All mouse experiments were conducted after
   approval by the Committee for Animal Welfare (University of Amsterdam
   and VU university Amsterdam).

   For bone marrow transplantation, female Lldr^−/− mice were lethally
   irradiated with six Gray on 2 successive days. After randomization
   based on weight, mice were divided in two groups of 20 and injected
   intravenously with 5 × 10^6 bone marrow cells from either control or
   Acly^M-KO mice. From one week before until 5 weeks after
   transplantation, recipient Lldr^−/− mice received autoclaved tap water
   containing antibiotics (100 mg/L neomycin and 60,000 U/l polymyxin B
   sulfate). Bone marrow transplantation efficiency was evaluated by qPCR
   for Ldlr on DNA extracted from blood. Six weeks after the
   transplantation, mice received a HFD) (15% fat and 0.15% cholesterol,
   Ssniff Spezialdiäten). After 10 weeks HFD, mice were sacrificed by
   CO[2]asphyxiation. Hearts, livers, and spleens from killed mice were
   collected and hearts and livers were frozen in tissue-tek (DAKO) for
   histological analysis. Blood was collected before the start of the diet
   and five and nine weeks after the start of the diet for lipid profiling
   and to determine leukocyte composition. Peritoneal macrophages were
   isolated after thioglycolate injection from 6/6 (control/Acly^M-KO)
   mice as previously described^[193]46.

Histology

   Sections of 7 μm were cut on a Leica 3050 cryostat and stained with
   toluidine blue (0.2% in phosphate-buffered saline (PBS)) to quantify
   plaque area and Sirius red (0.05% in saturated picric acid solution) to
   measure collagen deposition using Adobe Photoshop software (Version. In
   addition, toluidine blue-stained slides were used to measure necrosis
   and to categorize plaques by an experienced pathologist. Minimal cap
   thickness was measured on the thinness part of the cap surrounding the
   necrotic core. Sections were fixed in acetone and incubated with
   antibodies against MOMA‐2 (1:4000, AbD Serotec), Ly6G (1:200, BD
   Pharmingen), total Acly (1:100, Abcam), or MOMA-2 (1:4000, AbD Serotec)
   together with TGF-β (1:100, Abcam) to determine macrophage, neutrophil,
   total Acly and macrophage-related TGF-β content, respectively,
   (antibodies detailed in the Supplementary Table [194]1). MOMA-2, Ly6G,
   and Acly were captured using biotin-labeled rabbit anti-rat antibody
   (Vector Laboratories) as secondary antibody. Hereafter, the signal was
   amplified using the ABC kit (Vector Laboratories) and subsequently
   visualized using the ImmPACT AMEC Red substrate kit (Vector
   Laboratories). Combined MOMA-2 and TGF-β staining was visualized using
   Alexa-555-labeled goat anti-rat antibody (Molecular probes, 1:400) for
   MOMA-2 and Alexa647-labeled goat anti-rabbit antibody (Molecular
   probes, 1:400) for TGF-β. Hereafter, the signal was measured by
   multispectral imaging using a Vectra Polaris slide imager (Perkin
   Elmer), and subsequently cleaned from autofluorescent spectra by
   measuring an unstained lesion. ImageJ was used for analysis of positive
   area and for colocalization with the ImageJ plugin JaCOP.

   To analyze apoptosis and efferocytosis with TUNEL, sections were
   stained with MAC3 following paraformaldehyde fixation, and
   permeabilization with 0.1% triton X-100. MAC3 signal was detected using
   biotinylated rabbit anti-rat antibody, streptavidin-ABC-AP, and
   visualized using vector blue (all Vector Laboratories). Next, TUNEL
   enzyme incorporated fluorescein-labeled dNTPs into ACs which were then
   visualized with anti-FITC–HRP, and diamino benzidine (DAB, DAKO)
   according to manufacturer’s instruction (In situ cell death detection
   kit, Sigma). The efferocytosis index was calculated as the ratio of
   TUNEL + AC bound to MAC3 + cells over free ACs (bound AC/free AC), as
   previously described^[195]47,[196]48.

   For the analysis of Acly in human plaques, slides were deparaffinised
   and antigens were retrieved by cooking in a Tris-EDTA buffer with pH 9.
   Plaque slides were incubated with anti-phosphorylated ATP citrate lyase
   (pSer^454) (Sigma, 1:100) or anti-ATP citrate lyase (Abcam, 1:100),
   anti-CD68 (Abcam, 1:75) or anti-CD45 (DAKO, 1:100), and DAPI
   (Invitrogen, 1:1000) to visualize macrophages and phosphorylation of
   ATP citrate lyase in the plaque tissue. Subsequently, slides were
   incubated with anti-rabbit Alexa488 (Molecular probes, 1:400) and
   anti-mouse Alexa647 (Invitrogen, 1:400). Fluorescent signal was
   captured with multispectral imaging using a Vectra Polaris slide imager
   (Perkin Elmer). ImageJ was used for analysis of positive area and for
   colocalization with the ImageJ plugin JaCOP. A representative image of
   7 unstable plaques was shown.

Blood, spleen and liver parameters

   Levels of cholesterol and triglycerides in plasma were enzymatically
   determined according to the manufacturer’s protocol (Roche). Hepatic
   triglyceride levels were enzymatically determined on snap frozen liver
   samples according to the manufacturer’s protocol (Cayman chemical).
   Spleens were crushed and digested by enzymatic digestion. Blood and
   spleen leukocyte composition was assessed by flow cytometry after
   antibody staining with CD45-APC-Cy7 (1:100) for total leukocytes,
   CD11b-PE-Cy7 (1:400), MHC-II-PerCP Cy5.5 (1:100) and Ly6C-APC (1:100)
   for monocytes, Ly6G-FITC (1:100) for neutrophils, NK1.1-PE (1:100) for
   natural killer cells, CD3-PerCP Cy5.5 (1:100), CD4-APC (1:200) and
   CD8-FITC (1:200) for T cells and CD19-PE (1:100) for B cells.
   Fluorescence was measured on a Beckman Coulter CytoFLEX and analyzed
   with CytExpert software.

Bone marrow isolation/neutrophil isolation and BMDM/BMDC culture

   Bone marrow cells were extracted by flushing femurs and tibias of
   control and Acly^M-KO mice. Cells were seeded and differentiated to
   macrophages in complete Roswell Park Memorial Institute (RPMI)-1640
   containing 25 mM HEPES, 2 mM l-glutamine, 10% FCS, 100 U/ml penicillin,
   100 μg/ml streptomycin and 15% L929-conditioned (LCM) medium for 7
   days. BMDMs were left unstimulated or stimulated with 100 ng/ml LPS
   (Sigma) or 20 ng/ml IL-4 (Peprotech) for 24 h.

   Bone marrow cells were differentiated to dendritic cells in IMDM
   containing 10% FCS, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM
   Glutamax, 50 µm β-mercaptoethanol and 30 ng/ml X63 for 7 days.

   Neutrophils were isolated from bone marrow cells by density gradient as
   also detailed in the reference^[197]49. Bone marrow cells were
   resuspended as max 100 × 10^6 per ml ice-cold PBS, after which 1 ml
   cell suspension was added on top of a double density gradient. The
   lower layer contained 3 ml of 150 µM NaCl-Percoll mix with a density of
   1.119 g/ml topped up with a layer of Lymphoprep with a density of
   1.077 g/ml. After centrifugation (30 min, 2000 rpm, 25 °C, no brake)
   neutrophils were present at the interphase. After washing with RPMI,
   viability was determined and cells were frozen until further analysis.

Cell lines

   RAW264.7 cells were cultured in complete RPMI-1640 containing 2 mM
   l-glutamine, 10% FCS, 100 U/ml penicillin and 100 µg/ml streptomycin.
   Cells were passed when 80–90% confluency was reached by scraping in
   fresh medium. Cells with passage number 6–12 were used for experiments.

Immunoblotting

   BMDMs were washed with cold PBS and lysed on ice in NP40 cell lysis
   buffer (ThermoFisher) supplemented with fresh protease inhibitor
   cocktail (Sigma-Aldrich) and fresh PhosSTOP (Sigma-Aldrich). Lysates
   were scraped, collected and centrifuged for 10 min at 2000 × g and
   4 °C. Supernatants were equalized on protein concentration after
   quantification with the BCA assay (ThermoFisher), heated for 10 min at
   95 °C, loaded on Bolt 4–12% Bis-Tris gels (ThermoFisher) and
   transferred onto nitrocellulose membranes (Bio-Rad). Membranes were
   blocked for 1 h with 5% milk powder (Campina) or 5% BSA (Sigma) in
   Tris-buffered saline with TWEEN 20 (TBS-T) and incubated overnight with
   primary antibodies against ACLY (1:1000, Abcam), phosphorylated ACLY
   (1:1000, Cell Signaling Technology), and α-Tubulin (1:2000,
   Sigma-Aldrich, #T5168). Next day, membranes were incubated with
   horseradish peroxidase (HRP)-conjugated secondary antibodies (1:2000,
   Thermo Fisher Scientific) in milk or BSA (5% in TBS-T) and developed
   using SuperSignal West Pico Chemiluminescent PLUS Substrate (Thermo
   Fisher Scientific).

   For histone acetylation analysis, BMDMs were washed with cold PBS and
   lysed on ice in Triton Extraction Buffer (0.5% Triton X-100, 0.02%
   (w/v) NaN[3] in PBS) supplemented with fresh protease inhibitor
   cocktail (Sigma-Aldrich). Lysates were scraped, collected and
   centrifuged for 10 min at 2000 × g and 4 °C. Subsequently, pellets were
   resuspended in 0.2 M HCl and rotated overnight at 4 °C for acid histone
   extraction. Next day, samples were centrifuged for 10 min at 2000 × g
   and 4 °C, and supernatants (containing the extracted histones) were
   collected. Immunoblotting of histone extracts was performed as detailed
   above, with primary antibodies directed against H3 (1:5000, Cell
   Signaling Technology) and H3K27ac (1:1000, Diagenode).

Gene expression analysis

   Aortic roots were disrupted with glass beads in a MagnaLyser system and
   total RNA was extracted from homogenates using TRIzol reagent (Life
   Technologies) and the RNeasy Mini Kit with DNase treatment (QIAGEN).
   For BMDMs, total RNA was isolated from macrophages in 24-well plates
   (5 × 10^5 cells/well) using the RNeasy Mini Kit with DNase treatment
   (QIAGEN) or GeneJET RNA Purification Kit (Thermo Fisher Scientific).
   Isolated RNA from aortic arches and BMDMs was reverse-transcribed into
   cDNA using the High-Capacity cDNA Reverse Transcription Kit
   (ThermoFisher) and gene expression analysis was performed with SYBR
   Green Fast mix (Applied Biosystems) on a ViiA7 system (Applied
   Biosystems). Levels of expression were normalized to those of
   housekeeping genes Rplp0 and Ppia. Primer sequences used for gene
   expression analysis are provided in Supplementary Table [198]2.

Cytokine, NO, and ROS production

   Production of IL-6, TNF, and NO was quantified in supernatants using
   ELISA (Life Technologies) or the Griess test (Sigma) respectively,
   following the supplier’s protocols. In order to determine ROS
   production, macrophages were incubated with 20 μM CM-H[2]DCFDA
   (ThermoFisher) for 30 min at 37 °C (5% CO[2]). Hereafter, cells were
   washed with PBS and fluorescence was measured on a Beckman Coulter
   CytoFLEX.

Flow cytometry

   To examine surface marker expression, BMDMs were detached using citrate
   buffer (17 mM tri-Sodium citrate dehydrate and 135 mM potassium
   chloride in water), transferred to V-bottom well plates, incubated with
   1:50 anti-CD16/CD32 Fc-block and stained for 20 min at room temperature
   in the dark with antibodies. Antibodies were directed to CD40-APC
   (Biolegend, 1:100), MHC-II-PerCP/Cy5.5 (Biolegend, 1:200), CD80-BV650
   (Biolegend, 1:100) and CD86-BV510 (Biolegend, 1:100) for LPS-stimulated
   macrophages or directed to CD71-PE (BD PharMingen, 1:250), CD206-APC
   (Biolegend, 1:250), CD273-PE (BD PharMingen, 1:250), CD301-AF647
   (Serotec, 1:250) or isotype controls (Biolegend) for IL-4 stimulated
   macrophages (IGG2A-PE, IGG2A-APC), diluted in staining buffer (PBS with
   0.5% BSA and 2.5 mM ethylenediaminetetraacetic acid; EDTA). After
   staining, cells were washed and resuspended in staining buffer,
   measured on a Beckman Coulter CytoFLEX or BD LSR Fortessa and analyzed
   using FlowJo (Tree Star). Surface expression was calculated as
   ΔMFI = (median fluorescence intensity)[positive staining] – (median
   fluorescence intensity)[isotype or FMO control]. MFI controls (isotype
   or FMO) were measured on a pool of all samples.

Metabolomics

   Macrophages were quenched by ice-cold methanol. D[3]-aspartic acid,
   D[3]-serine, D[5]-glutamine, D[3]-glutamate, ^13C[3]-pyruvate,
   ^13C[6]-isoleucine, ^13C[6]-glucose, ^13C[6]-fructose-1,6-biphosphate,
   ^13C[6]-glucose-6-phosphate, adenosine-^15N[5]-monophosphate and
   guanosine-^15N[5]-monophosphate (5 µM) served as internal standards.
   After addition of chloroform, samples were centrifuged and the “polar”
   top layer was dried in a vacuum concentrator and dissolved in
   methanol/water (6/4, v/v). Analysis was done using a Thermo Scientific
   (U)HPLC system coupled to a Thermo Q Exactive (Plus) Orbitrap mass
   spectrometer (Waltman) with a SeQuand ZIC-cHILIC column. MS data were
   acquired in full scan, negative ionization mode with a mass resolution
   of 140,000^[199]46. Interpretation of the data was performed on the
   Xcalibur software (ThermoFisher). Further analyses were performed in R
   using ggplot2, ropls, and mixOmics packages.

Metabolic determinants and extracellular flux analysis

   Glucose (GOD-PAP method, Biolabo) and ATP (ThermoFisher) levels were
   enzymatically determined according to manufacturer’s protocol. Lactate
   was enzymatically determined on 3% metaphosphoric acid (MPA)
   deproteinized cell culture supernatants. After centrifugation
   (20,000 × g, 10 min) supernatants and l-lactic acid standards were
   measured for background fluorescence in a 0.5 M Glycine-0.4 M Hydrazine
   buffer, pH 9.0 with 3.6 mM NAD with a Mithras LB-940 reader at
   λex/λem = 340–10/450–10 nm five times every 2 min. Reactions were
   started by adding 0.5 M Glycine-0.4 M Hydrazine buffer, pH 9.0 with
   50 µg/mL l-lactate dehydrogenase from rabbit muscle (Roche) and again
   measured at λex/λem = 340–10/450–10 nm every 2 min until a stable read
   was achieved.

   For extracellular flux analysis, the XF-96 Flux Analyzer (Agilent) was
   used to assess OCR and ECR on macrophages from both phenotypes. BMDMs
   (8 × 10^4 cells/well) were plated on XF-96-cell culture plates
   (Agilent), treated with LPS or left untreated. In all, 1 hr before the
   assay, cells were washed and replaced with Dulbecco’s Modified Eagle’s
   Medium (DMEM) (Sigma-Aldrich) without glucose, phenol red, and sodium
   bicarbonate. The run consisted of 2 min mixing, 3 min measuring and
   subsequent four injections: glucose (final assay concentration 25 mM),
   oligomycin A (OM, final assay concentration 1.5 µM), FCCP (final assay
   concentration 1.5 µM) with sodium pyruvate (final assay concentration
   1 mM), with last injection of antimycin A (AA, final assay
   concentration 2.5 µM) with rotenone (final assay concentration
   1.25 µM). Data were analyzed using Wave software as detailed
   before^[200]50.

Transcriptomics

   Total RNA was isolated from BMDMs using a RNeasy Mini Kit with DNase
   treatment (QIAGEN) and strand-specific libraries were constructed using
   the KAPA mRNA HyperPrep kit (KAPA Bioystems). Subsequently, samples
   were sequenced on an HiSeq 4000 instrument (Illumina)^[201]46.
   Afterwards, reads were aligned to the mouse genome mm10 by STAR 2.5.2b.
   BAM files were indexed and filtered with SAMtools and raw tag counts
   and RPKM values were summed using HOMER2’s analyzeRepeats.pl script.
   Differential expression was assessed using the DESeq2 package in an R
   environment. Reactome pathway enrichment analysis was performed using
   the ReactomePA package for significantly differentially expressed genes
   (adjusted p value ≤ 0.05). Significantly enriched pathway list was
   filtered on redundant pathways that represent parent functionality.
   Summarized pathway list was visualized using ridge plots, by extracting
   gene ids and plotting fold change distributions using in-house scripts.
   Volcano plot and heatmaps were generated using the ggplot2, ggrepel,
   and pheatmap package.

Analysis of sterols in macrophages

   Sterols were analyzed in 2 × 10^6 macrophages, homogenized in 100 μl
   PBS followed by sonication in a sonicator bath for 2 min. One hundred
   microliters of internal standard solution (65 nmol of epicoprostanol)
   was added and mixed followed by addition of 1.0 ml of saponification
   reagent (0.3 M KOH in ethanol). This was incubated for 2 h at 80 °C.
   After cooling, 2 ml of hexane and 0.5 ml of water was added and mixed.
   Three hundred microliters of the hexane layer was taken to dryness
   under a flow of nitrogen at 40 °C and 120 μl of BSTFA (+1% TMCS) was
   added and incubated for 30 min at 80 °C. One microliter of the
   resulting trimethylsilyl (TMS)-ethers of the sterols was injected on a
   GC system (CPSil5 column, Agilent GC 7890B) with FID detection for
   quantification using the added internal standard. The identity of the
   sterols was confirmed by GC-MS using separation on a Agilent GC 7890B
   followed by selected ion monitoring of the TMS‐derivatives on an
   MSD5977A MS detector in the EI + -mode.

Metabololipidomics

   Lipids were extracted from macrophages (5 × 10^5 cells/condition) in
   24-well plates, cultured in serum-free phenol red-free DMEM medium
   supplemented with 25 mM HEPES, 2 mM l-glutamine, 100 U/ml penicillin,
   and 100 μg/ml streptomycin. Cell lysates were quenched by adding 1.5 mL
   methanol and 4 µl internal standard solution [2H4]LTB[4], [2H8]15-HETE,
   [2H4]PGE[2], [2H5]DHA, [2H11]14(15)EET and [2H4]PGF[2α] (50 ng/mL in
   MeOH) was added. Samples were then placed in −20 °C to equilibrate and
   spun down for 10 min at 16.200 g at 4 °C. Supernatants were diluted in
   7.5 mL H[2]O, pH was corrected to 3.5 using formic acid (99%) and
   subsequent solid-phase extraction was used to clean the samples
   further. Samples were loaded on C18 Cartridges (Sep-Pak, Vac3 3cc
   (200 mg)) prewashed with MeOH and H[2]O and subsequently washed with
   H[2]O and n-hexane. Extracted lipids were eluted using methylformate
   and collected in glass tubes. The elutes were dried at 40 °C using a N2
   flow, reconstructed in 200 µL MeOH (40%) and transferred to deactivated
   glass inserts.

   Oxylipid content of the samples was measured using a targeted
   HPLC-MS/MS method according to the protocol of Jónasdóttir et al.
   ^[202]51. Separation of lipids was accomplished using a Shimadzu
   LC-system, which consisted of two LC-30AD pumps, a SIL-30AC autosampler
   (at 4 °C) and a CTO-20AC column oven (50 °C). Samples were injected
   (40 µL) and compounds were separated on a Kinetex C18 column
   (50 × 2.1 mm, 1.7 µm column), which was protected with a C8 pre-column
   (Phenomenex). A gradient of H[2]O (eluent A) and MeOH (eluent B), both
   with 0.01% acetic acid, was used to elute the samples at a constant
   flowrate of 400 µL/min. Other machine settings were in line with the
   protocol of Jónasdóttir^[203]51 and scheduled MRM mode was used to
   detect compound peaks. Compounds were detected using their relative
   retention times together with characteristic mass transitions and these
   and other individually optimized parameters can be found in the
   Supplementary table [204]3.

Cell cycle analysis

   After 48 h of starvation in medium without LCM, macrophages (1 × 10^6
   cells/well) in 12-well plates were washed with PBS and subsequently
   detached with citrate buffer (17 mM tri-Sodium citrate dehydrate and
   135 mM potassium chloride in water) and transferred to tubes. Cells
   were centrifuged and ice-cold 70% ethanol was drop-wise added to the
   pellet while vortexing. After fixation at 4 °C for at least 24 h, cells
   were washed two times with PBS and permeabilized with 0.1% Nonidet P-40
   for 10 min. Permeabilized cells were transferred to V-bottom 96-well
   plates and washed with PBS. Subsequently, cells were resuspended in PBS
   with 50 µg/ml propidium iodide and 25 µg/ml RNAse and incubated for
   30 min at room temperature. After staining, cell cycle status was
   determined with flow cytometry.

In situ staining of senescence-associated β-galactosidase activity

   β-Galactosidase activity was assessed in situ after culturing cells on
   glass coverslips. Cultured cells were fixed with 4% paraformaldehyde
   for 5 min and incubated overnight at 37 °C in senescence-associated
   b-galactosidase staining solution (0.1%
   5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside, 5 mM ferrocyanide,
   5 mM ferricyanide, 150 mM NaCl, 2 mM MgCl, 40 mM citric acid, 40 mM
   sodium phosphate, pH 6.0). Coverslips were mounted with Mowiol and
   captured using a bright field microscope using a ×10 objective. Cells
   were quantified using ImageJ cell counting software.

In vitro efferocytosis

   A day before the assay, macrophages were stained for 20 min with
   2.5 µl/ml DiD-stain, washed with PBS and seeded (10^5 cells/well) in
   96-well plates. On the day of the assay, RAW264.7 cells were killed by
   a heat shock of 70 °C for 5 min. Subsequently, killed cells were
   stained with pHrodo at a concentration of 2 mg/ml for 30 min at RT and
   washed with PBS. Dead RAW264.7 cells were resuspended in medium and
   incubated as 5:1 with the DiD-stained macrophages for 1.5 h. After
   washing, uptake of pHrodo-stained dead cells by DiD-stained macrophages
   was determined with flow cytometry and validated by imaging flow
   cytometry (ImageStream). A 4 °C incubation served as control to set the
   positive pHrodo gate.

Statistical analysis

   Data are presented as mean ± standard error of the mean (SEM).
   Statistical significance of data was tested using a two-tailed
   Student’s t test (when comparing two groups), ordinary one-way analysis
   of variance (ANOVA) followed by Bonferroni’s post hoc comparison with
   test multiple groups or two-way ANOVA followed by Bonferroni’s post hoc
   comparison for grouped analyses in GraphPad Prism software. P
   values < 0.05 were considered significant, with levels of significance
   being indicated as follows: *p < 0.05; **p < 0.01; ***p < 0.001.

Reporting summary

   Further information on research design is available in the [205]Nature
   Research Reporting Summary linked to this article.

Supplementary information

   [206]Supplementary Information^ (1.2MB, pdf)
   [207]Peer Review File^ (3MB, pdf)
   [208]41467_2020_20141_MOESM3_ESM.docx^ (11.4KB, docx)

   Description of Additional Supplementary Files
   [209]Supplementary Data 1^ (47.1KB, xlsx)
   [210]Reporting Summary^ (434.7KB, pdf)

Acknowledgements