Hormone-sensitive lipase drives pro-resolving macrophage polarization and enhances efferocytosis

Abstract

Background

Lipid droplets (LDs) are vital organelles in cellular energy regulation, controlling lipid storage and mobilization. However, their specific roles in macrophage function, particularly in the context of immune regulation, remain elusive. This study aimed to elucidate how LD manipulation affects macrophage phenotypes and functions, with a focus on their role in efferocytosis and inflammatory modulation. We also explored the novel function of hormone-sensitive lipase (HSL) in regulating LD metabolism and macrophage polarization.

Methods

Human monocytes were differentiated into M1 (IFN-γ/LPS) or M2 (IL-4) macrophages. To determine the functional role of HSL, we employed pharmacological inhibition, siRNA-mediated knockdown, and plasmid-driven overexpression. Phenotypic and functional effects on efferocytosis, gene expression, and cytokine secretion were evaluated using flow cytometry, qRT-PCR, and ELISA. The underlying mechanisms were explored through unbiased quantitative lipidomics, real-time metabolic flux analysis to assess mitochondrial respiration, and high-resolution imaging via confocal and electron microscopy, which confirmed alterations in the LD dynamics.

Results

M2 macrophages exhibited increased efferocytosis and contained smaller, more numerous lipid droplets with greater lipase activity compared to M1 macrophages. Inhibition or siRNA-mediated knockdown of HSL reduced efferocytosis and downregulated the expression of M2 markers, including CD206, CD163, CD200R1, and Annexin A1, while promoting a pro-inflammatory state. Conversely, HSL overexpression in naive macrophages induced a strong M2-like phenotype, characterized by increased efferocytosis and elevated mitochondrial respiration. We highlighted the key role of the cAMP-protein kinase A pathway in promoting lipolysis of LD via HSL activation. Lipidomic analysis revealed distinct lipid profiles for M1 and M2 macrophages, as well as for M2 macrophages with HSL inhibition. Lipidomic results indicated that M2 macrophages contain specific monounsaturated fatty acids, such as Nervonic Acid (FA 24:1), which can enhance efferocytosis by M1 macrophages.

Conclusions

This study identifies a novel mechanism by which LD lipolysis, mediated by HSL, promotes M2 macrophage polarization and enhances efferocytosis. These findings highlight the importance of LD metabolism in macrophage function and suggest HSL as a potential therapeutic target for immune-mediated conditions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-025-02631-z.

Introduction

Macrophages have remarkable heterogeneity and play crucial roles as innate myeloid cells in health and disease [1–4]. This functional diversity is often studied using the M1/M2 polarization model, which provides a valuable in vitro framework for distinguishing between pro-inflammatory and pro-resolving states. The polarization of macrophages into M1 (pro-inflammatory) and M2 (anti-inflammatory or tissue repairing) subtypes is a fundamental aspect of their role in host defense, inflammation, and tissue homeostasis. M1 macrophages, typically induced by cytokines such as IFN-γ and microbial products like lipopolysaccharide (LPS), produce inflammatory cytokines and reactive oxygen species that are essential for antimicrobial defense and tumor suppression. Conversely, M2 macrophages, which can be induced by cytokines such as IL-4 and IL-13, are associated with wound healing, tissue repair, and the resolution of inflammation through the production of anti-inflammatory cytokines and growth factors [5, 6]. Generally, M1 macrophages depend on glycolysis, while M2 macrophages are associated with oxidative phosphorylation and a metabolic state that involves fatty acid oxidation (FAO) [7, 8]. M2 macrophages highly express lipid-scavenging receptors and have high phagocytosis capacity [9, 10]. The importance of peroxisome proliferator-activated receptor gamma (PPARγ) in regulating the M1/M2 macrophage switch was reconfirmed in a study from Bouhlel et al.. showing that activation of PPARγ potentiates polarization of circulating monocytes into M2-like macrophages [11]. The balance between M1 and M2 macrophages is crucial for maintaining health, and dysregulation of this balance is implicated in various clinical conditions [12]. Targeting macrophage polarization to modulate the immune response is an appealing therapeutic strategy for autoimmune diseases, cancer, chronic inflammation, and obesity. However, significant challenges remain in defining macrophage polarization states [13].

Lipid droplets (LDs) are intracellular organelles that function as energy reservoirs and substrates for membrane synthesis [14–16]. In macrophages and other immune cells, LDs are dynamic structures involved in lipid metabolism, inflammatory responses, and metabolic adaptation [17]. LDs primarily consist of triacylglycerols (TAG), diacylglycerols (DAG), and cholesterol esters (CE), which are hydrolyzed by specific lipases. Adipose triglyceride lipase (ATGL) initiates TAG breakdown, hormone-sensitive lipase (HSL) processes DAGs into monoacylglycerols (MAGs), and monoacylglycerol lipase (MAGL) completes the lipolysis pathway by releasing free fatty acids (FFA) [18–20]. HSL also hydrolyzes CE, generating cholesterol and FFA [21]. These lipolytic activities provide a rapid energy supply during immune responses. LD accumulation in macrophages can be triggered by Toll-like receptor (TLR) signaling [22]. In dendritic cells, TLR ligand stimulation with LPS induces a metabolic shift toward glycolysis, leading to increased pyruvate flux into the TCA cycle and subsequent FA synthesis. These FAs are predominantly converted into TAG, contributing to the growth of LDs [23].

Efferocytosis, the process of clearing apoptotic cells, is essential for immune homeostasis, inflammation control, and tissue remodeling [24]. This physiological function is vital for maintaining cellular and tissue integrity and preventing autoimmune responses by ensuring the controlled removal of dying cells before they release inflammatory mediators [25, 26]. Dysregulation of efferocytosis induces substantial inflammatory responses and is linked with chronic inflammatory diseases, including autoimmune diseases [27, 28]. Apoptotic cells expose phospholipid phosphatidylserine (PS) on their plasma membrane, which acts as an “eat-me” signal for phagocytes such as macrophages [29]. Macrophages detect these apoptotic cells using specific efferocytic receptors, such as Mer tyrosine kinase (MERTK) and T-cell immunoglobulin and mucin domain containing-4 (TIM-4), which recognize PS directly or indirectly through bridging molecules like Growth Arrest-specific gene 6 (Gas6) and Protein S [24, 30]. Other critical scavenger receptors, such as CD36, are also involved in the process of efferocytosis. CD36 facilitates the uptake of apoptotic cells by recognizing oxidized lipids on cell surfaces [31]. The expression of these receptors and their associated bridging molecules is regulated by PPARγ, which influences the metabolic states of macrophages [32, 33]. Importantly, lipids play a significant role in the formation and maturation of phagosomes during efferocytosis [34]. However, the role of LDs in efferocytosis remains unclear and requires further elucidation.

FA 24:1 (nervonic acid, NA) is a monounsaturated, very long-chain fatty acid primarily responsible for the development and maintenance of the myelin sheath [35]. NA is predominantly found in the subcortical white matter as a component of sphingolipids and myelin and is utilized by oligodendrocytes during myelination. As such, it has been extensively studied in neurological diseases, including Alzheimer’s and demyelinating disorders [36]. While NA has traditionally been studied in the context of nervous system function, recent research suggests it may also exert broader immunomodulatory effects, particularly through the regulation of macrophage-mediated inflammation. NA has been shown to downregulate important pro-inflammatory signaling pathways such as NF-κB and MAPK in radiation-induced pulmonary injury-exposed macrophages, thereby reducing cytokine production and tissue damage [37]. Improving the efferocytotic capacity of macrophages in SLE patients can thus serve to reduce the chronic inflammation; however, utilizing NA to promote efferocytotic activity by promoting an M2 phenotype in this disease context has yet to be explored.

In this study, we investigate how lipid metabolism influences macrophage polarization and efferocytosis. We first demonstrate that in vitro-polarized M1 and M2 macrophages exhibit distinct LD characteristics and that enhanced efferocytosis in M2 cells depends on lipolytic activity. We then further identify a cAMP-PKA-HSL signaling pathway as a hallmark of the M2 state. Importantly, using a combination of pharmacological inhibition, genetic knockdown, and overexpression, we demonstrate that HSL is both necessary and sufficient to promote efferocytosis and a pro-resolving gene program. Overall, these findings reveal a comprehensive metabolic pathway where HSL serves as a central regulator, linking lipid droplet breakdown to the transcriptional and energetic needs of the pro-resolving macrophage.

Materials and methods

Materials

Orlistat (10–100 µM), T-863 (50 µM), Lalistat 1 (5 µM), Atglistatin (20 µM), and MJN110 (5 µM) were purchased from Cayman Chemical (Ann Arbor, MI). HSL-IN-1 (5–15 µM) was obtained from MedChemExpress (Monmouth Junction, NJ). Dibutyryl-cAMP (50 µM) was sourced from Stem Cell Technology (Vancouver, Canada). Phorbol 12-myristate 13-acetate (PMA) from Abcam (Cambridge, UK). FA 24:1 (Nervonic Acid, ≥ 99% capillary GC), as well as human BSA and mouse BSA, were purchased from Sigma-Aldrich (St. Louis, MO).

Human subjects and ethical approval

All studies involving human cells were performed in accordance with institutional guidelines. The research was granted an exemption by the Institutional Review Board (IRB) of the Feinstein Institutes for Medical Research. This exemption was based on the secondary use of commercially obtained, fully de-identified biological material. Leukopaks from healthy, anonymized donors were procured from the New York Blood Center (New York, NY).

Monocyte isolation and cell culture

Human peripheral blood mononuclear cells (PBMCs) were isolated from the obtained leukopaks using density gradient centrifugation with Ficoll-Paque Plus (Cytiva, Marlborough, MA). Monocytes were enriched from PBMCs through negative selection, utilizing human monocyte enrichment kits (Stem Cell Technology). The purity of the monocytes (≥ 90%) was assessed using a flow cytometer (LSRFortessa; BD, Franklin Lakes, NJ) by staining for CD14 (Supplementary Fig. 1A). The purified monocytes (1.5 × 10⁶ cells per mL) were cultured in RPMI 1640 with 10% FBS or ImmunoCult™-SF Macrophage Medium (Stem Cell Technologies) with 20–40 ng/mL of M-CSF (PeproTech, Cranbury, NJ) for 4–5 days in a 5% CO₂ incubator at 37 °C to differentiate into macrophages (M0). The M0 macrophages were then stimulated in 24-well plates (1 × 10⁵ cells per well) with either IFN-γ (20 ng/mL, PeproTech) and LPS (100 ng/mL, Merck KGaA, Darmstadt, Germany) to generate M1-like macrophages (M1) or with IL-4 (20 ng/mL, PeproTech) to generate M2-like macrophages (M2), with or without inhibitors, for 24 h.

Preparation of bone marrow-derived macrophages (BMDMs)

Bone marrow cells were extracted from the femurs and tibias of C57BL/6 mice purchased from Jackson Laboratory (Bar Harbor, ME). The animals were maintained in a conventional animal housing facility and euthanized according to approved protocols by the Feinstein Institutes IACUC committee (2022-022). The marrow was flushed with RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, or ImmunoCult™-SF Macrophage Medium (Stem Cell Technologies) with 20 ng/mL murine M-CSF (PeproTech) to promote differentiation into macrophages. Cells were cultured in a humidified incubator at 37 °C with 5% CO2. After three days, non-adherent cells were removed, and fresh medium was added with M-CSF. On day five, M0 cells were differentiated into M1 (murine IFN-γ, 20 ng/mL, PeproTech) and LPS (100 ng/mL, Merck) or M2 (murine IL-4, 20 ng/mL, PeproTech) macrophages. Cells were harvested for experiments after 24 h.

qRT-PCR

Total RNA was extracted from cells using Direct-zol RNA Miniprep Kits or Direct-zol-96 RNA Kits (Zymo Research, Irvine, CA). Complementary DNA (cDNA) was synthesized with the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Real-time PCR reactions were conducted on a LightCycler 480 II (Roche, Basel, Switzerland) using LightCycler 480 Probes Master (Roche). Primers for IL1B (Hs01555410_m1), HIF1A (Hs00153153_m1), MERTK (Hs1031973_m1), CD206/MRC1 (Hs00267207_m1), PPARG (Hs01115513_m1), ATGL/PNPLA2 (Hs00982042_m1), LIPE (HSL, Hs00943410_m1), TNF (Hs00174128_m1), CD36 (Hs00354519_m1), CD200R1 (Hs00793597_m1), ASAH1 (Hs00924966_m1), ACADVL (Hs00825606_g1), CPT1A (Hs00912671_m1), CPT2 (Hs00988962_m1), SLC25A20/CACT (Hs00386383_m1), MFN2 (Hs00208382_m1), FABP3 (Hs07287863_m1), Annexin A1 (ANXA1, Hs00167549_m1) and HPRT1 (Hs99999909_m1) were obtained from Thermo Fisher Scientific (Waltham, MA). The genes of interest were normalized to the expression of housekeeping genes, and relative expression (R.E.) was calculated using the 2^−ΔCt method.

Flow cytometry

Cells were suspended in HBSS with 3% BSA and incubated with Fc Receptor Blocking Solution (BioLegend, San Diego, CA) for 5 min at 4 °C. The cells were then incubated for 30 min at 4 °C with a panel of antibodies: Pacific Blue-anti-human CD14 (BD), Allophycocyanin (APC)-CD14, phycoerythrin (PE)-Cy7-anti-CD36, APC-CD206, and Brilliant Violet 711-anti-CD163 (BioLegend). Fixable Viability Dye (FVD), eFluor 780 or 506 (Thermo Fisher Scientific). After incubation, the cells were washed and fixed in 4% paraformaldehyde (PFA, Sigma) and stored in the dark at 4 °C until analysis. Sample acquisition was performed using either the LSR Fortessa or Symphony flow cytometers (BD), and data was analyzed using FlowJo software (BD).

Efferocytosis assay

Efferocytosis was quantified using the Efferocytosis Assay Kit (Cayman Chemical), following the manufacturer’s guidelines. Initially, M0 macrophages were stained with CytoTell™ Blue and polarized as previously described. Jurkat cells, clone E6-1 (ATCC, Manassas, VA), were labeled with CFSE and treated with Staurosporine (1 µM) overnight to induce apoptosis. To validate apoptosis induction in Jurkat cells, cells treated with staurosporine for 3–18 h were harvested, washed, and stained with the FITC Annexin V Apoptosis Detection Kit with 7-AAD (BioLegend) according to the manufacturer’s protocol. Cells were analyzed immediately by flow cytometry using an LSR Fortessa (BD) (Supplementary Fig. 1B). These apoptotic Jurkat cells were washed twice and then co-cultured with the stained macrophages at a 1:2 ratio (effector cells: Jurkat cells) for 2 h in a humidified incubator at 37 °C with 5% CO2. After incubation, cells were washed and fixed in 4% buffered formaldehyde (Sigma) for 20 min. The samples were then analyzed using the Amnis ImageStreamX Mk II Imaging Flow Cytometer (Cytek Biosciences, Fremont, CA) and the LSR Fortessa (BD). Data were processed using IDEAS software (Cytek Biosciences) or FlowJo software (BD) to evaluate CFSE positivity in the macrophages.

In a parallel assay using mouse BMDMs, CFSE-labeled apoptotic thymocytes from C57BL/6 mice, treated overnight with Staurosporine, were co-cultured with M1 or M2 macrophages at 1:5 for 90 min. Following this, harvested cells were stained with Pacific Blue™-anti-mouse F4/80 antibody, washed, and fixed as described above.

Neutral lipid staining

HCS LipidTOX™ Green Neutral Lipid Stain was used to evaluate the intensity of neutral lipids according to the manufacturer’s instructions. Briefly, harvested macrophages were fixed in 4% PFA and then stained with HCS LipidTOX™ Green Neutral Lipid Stain. After washing, cells were analyzed using the Amnis ImageStreamX Mk II Imaging Flow Cytometer (Cytek Biosciences). The geometric mean fluorescence intensity among single cells was calculated using IDEAS software (Cytek Biosciences).

Confocal microscopy assay immunofluorescence microscopy

M0-macrophages were stimulated in 8-well chambers (2 × 10⁵ cells per well) for 24 h. The cells were washed, fixed with 4% PFA, and permeabilized using BD Perm/Wash™ Buffer (BD). Immunofluorescence staining was performed using primary antibodies against Perilipin-2 (PLIN2, Thermo Fisher Scientific) or HSL (Cell Signaling Technology, Danvers, MA), followed by Alexa Fluor 594- or 488-conjugated anti-rabbit IgG antibodies. After washing, LipidTOX™ Deep Red Neutral Lipid Stain and DAPI were used for staining neutral lipids and nuclei, respectively. Imaging was performed using a confocal microscope (60x with oil, LSM900, Zeiss, Oberkochen, Germany), and the images were analyzed using Imaris software (Oxford Instruments). To quantify the relationship between HSL and LDs, the Imaris “Coloc” module was used. LD number and volume per cell were measured from 3D reconstructed images. For per-cell analysis, individual cells were segmented by creating a 3D “Surface” object for each cell boundary. Then, a second “Surface” object identified LDs within the anti-PLIN2 channel, using Gaussian filtering to reduce noise and high-intensity thresholding. The software automatically calculated the total number of LDs, and the total LD volume (µm³) based on the voxel count of all LD surfaces within each cell.

Lipase and protein kinase A (PKA) activity assays

Total lipase and PKA activities were measured in macrophage lysates prepared according to the manufacturer’s instructions. Protein concentration in each lysate was determined by a BCA assay (Thermo Fisher Scientific) for normalization.

Lipase activity was measured using the Lipase Assay Kit (Abcam, Cambridge, UK). The assay quantifies glycerol produced from the hydrolysis of a triglyceride substrate. A standard curve was generated using the provided Glycerol Standard to allow for absolute quantification. Lipase activity was calculated based on the rate of glycerol production and normalized to the total protein content of the lysate, with final units expressed as pmol/min/mg. For experiments involving pathway activation, macrophages were pre-treated with the Db-cAMP (50 µM) or the PKC activator PMA (50 nM, positive control) for 3 h prior to lysis.

PKA activity was assessed using the PKA Colorimetric Activity Kit (Thermo Fisher Scientific). This assay measures the phosphorylation of a specific peptide substrate for PKA. The raw absorbance values for each sample were first normalized to their total protein concentration. To enable comparison across independent experiments, the activity for all samples was expressed as Relative Units (RU) by setting the average normalized activity of the M1 group to 1.

Western blot

Total protein extracts from macrophages (1 × 10⁶ cells) were prepared using RIPA lysis buffer containing protease and phosphatase inhibitors (Thermo Fisher Scientific). The samples were combined with LDS Sample Buffer, boiled, and subjected to immunoblot analysis utilizing 4–12% Bis-Tris gel with MOPS running buffer. The proteins were then transferred onto a PVDF membrane (Thermo Fisher Scientific). For detection, primary antibodies against phospho-HSL (Ser660) and HSL from the Lipolysis Activation Antibody Sampler Kit (#8334, Cell Signaling Technology) were used. ECL Plus Substrate (Thermo Fisher Scientific) was used to visualize secondary antibodies conjugated with HRP (Anti-rabbit IgG, Cell Signaling Technology). The membrane was scanned using the Sapphire Biomolecular Imager (Azure Biosystems, Dublin, CA).

Plasmids and transient transfection

M0 macrophages in 24-well plates were transfected using Lipofectamine LTX (Thermo Fisher Scientific) according to the manufacturer’s protocol. The cells were transfected with either a PNPLA2-encoding plasmid (ATGL, GeneCopoeia, Rockville, MD), a LIPE-encoding plasmid (HSL, GeneCopoeia), or a control plasmid (GeneCopoeia). All plasmids carry mCherry tags. qRT-PCR confirmed the efficiency of ATGL and HSL overexpression.

For siRNA-mediated gene silencing, M0-human macrophages or BMDMs were transfected with 25 nM of Silencer pre-designed siRNA targeting human LIPE/HSL (143993 and 143994, Life Technologies), Lipe mouse pre-designed siRNA Set A (HY-RS18621, MCE), or a non-targeting control pool (Dharmacon, Lafayette, CO) using Lipofectamine LTX Transfection Reagent (Thermo Fisher Scientific). Cells were cultured 24 to 48 h after transfection before being harvested for efferocytosis, ELISA, or qRT-PCR assays.

For the efferocytosis assay, transfected macrophages were co-cultured with apoptotic Jurkat cells without prior staining with CytoTell™. After co-culture, the cells were stained with APC-Cy7-anti-CD14 (BioLegend) prior to fixation and analyzed using the LSRFortessa (BD). Data were evaluated using FlowJo software (BD), assessing CFSE positivity in mCherry and CD14-positive macrophages.

ELISA

The levels of IL-1β and TNFα in the culture supernatants from human and mouse sources were quantified using the IL-1 beta/IL-1F2 DuoSet ELISA and TNF-alpha DuoSet ELISA kits (R&D Systems, Minneapolis, MN). Human Annexin A1 levels were quantified using the Human Annexin A1 ELISA Kit (ThermoFisher). TNF measurement was also used with Meso Scale Discovery human cytokines (MSD, Rockville, MD). All assays were performed strictly in accordance with the manufacturer’s instructions.

Transmission electron microscopy

M0 macrophages were stimulated in 6-well plates (5 × 10⁵ cells per well) as described above and then fixed for 30 min with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (Thermo Fisher Scientific). The cells were subsequently post-fixed with 1% osmium tetroxide, followed by 2% uranyl acetate. They were then dehydrated through a graded series of ethanol, lifted from the monolayer using propylene oxide, and embedded as a loose pellet in LX112 resin (LADD Research Industries, Burlington, VT) in Eppendorf tubes. Ultrathin sections were cut on a Leica Ultracut UC7 (Wetzlar, Germany), stained with uranyl acetate followed by lead citrate, and examined using a JEOL 1400Plus transmission electron microscope (Tokyo, Japan) at 120 kV (1500x or 5000x) at the Analytical Imaging Facility at the Albert Einstein College of Medicine. The size and density of LD (5000x) were analyzed using ImageJ [38]. LD density was defined as the mean grayscale intensity (0 = black, 255 = white) within the boundary of each LD, and the analysis was performed by an investigator blinded to the experimental conditions.

Mass spectrometry lipidomics analysis

M0-macrophages were stimulated in 6-well plates (5 × 10⁵ cells per well) as described above, collected in Precellys^® 2 mL Tissue Homogenizing Mixed Beads Kit, snap-frozen and kept in -80 °C until analysis. The cell pellets were extracted with lipidomics extraction solvent (480 µL, methylene chloride: methanolchloride: methanol: isopropanol = 25:10:65, v/v/v + 0.1% BHT), vortexed 30 s, centrifuged for 10 min at 13k r.p.m at 4 °C. The aliquot (98 µL) was transferred into mass spectrometry analysis tubes and SPLASH LipidoMix (Avanti Lipids) as internal standard mixture (2 µL) was spiked. Ultra-high performance liquid chromatography (UPLC) was performed using a 1.7 μm particle, 2.1 mm ×100 mm, CSH C18 Column (Waters, Milford, MA, United States) coupled to a quadrupole TOF mass spectrometer (AB SCIEX, TripleTOF 5600) operated in information-dependent MS/MS acquisition mode. LC-MS conditions followed the protocol by Choi et al. with modifications for both positive and negative ion modes [39]. Raw MS files were imported and processed by the program PeakView (Sciex, ver 1.2.1, Framingham, MA). PeakView detects spectral features using extract ion chromatogram (XIC) lists from our in-house library of lipids (each defined by a unique chromatographic retention time and accurate mass, MS/MS fragmentation, and isotopic pattern). Each peak was integrated by MulitQuant (Sciex ver 3.0.3) software. The Area under the curve (AUC) of each peak was normalized by an internal standard, and then the concentration of each lipid was calculated based on the previously reported NISTIR 8185 and LIPID MAPS references [40]. These measured concentrations of each lipid species (Supplementary Table 1) were utilized for additional analysis with MetaboAnalyst (v6.0).

Preparation of nervonic Acid–BSA conjugate solution

Nervonic acid conjugated to human BSA was prepared at a 3:1 molar ratio of lipid to protein. The dissolved nervonic acid was gradually added to the BSA solution while continuously vortexing to facilitate lipid-protein binding to the carrier protein. The mixture was incubated at 37 °C for 15 min with intermittent gentle vortexing and sonication to ensure uniform dispersion and association of nervonic acid with BSA. Following incubation, the volume was adjusted to 200 µl with sterile PBS. BSA controls were prepared in the same way, but without nervonic acid.

Seahorse XF Mito stress test assay

Real-time oxygen consumption rate (OCR) was measured using a Seahorse XFe96 Analyzer (Agilent Technologies, Santa Clara, CA). For FA studies, M1-polarized macrophages were seeded (50,000 cells per well) and treated with either FA 24:1 (50 µM, complexed with BSA) or a BSA vehicle control for 18 h prior to the assay. For gain-of-function studies, M0 macrophages were seeded in a Seahorse XF96 cell culture microplate (50,000 cells per well), transfected with either an empty vector or an HSL-encoding plasmid, and cultured for 24 h. Before the assay, cells were washed and incubated in XF Base Medium supplemented with 10 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate, then equilibrated in a non-CO2 incubator at 37 °C for 1 h. OCR was measured under basal conditions and after sequential injections of oligomycin (1.5 µM), FCCP (1.0 µM), and a mixture of rotenone and antimycin A (0.5 µM each). After the assay, cells were fixed, stained with DAPI, and imaged to normalize OCR data to cell count per well. Key respiratory parameters, including Basal Respiration, Maximal Respiration, Spare Respiratory Capacity, and ATP production, were calculated using Wave Desktop software (Agilent).

Statistical analysis

Statistical analysis was conducted using GraphPad Prism 10 (GraphPad Inc., La Jolla, CA). The analyses included paired t-test for two groups, and one-way ANOVA for multiple comparisons. Other methods are described in the text or figure legends. A p-value of < 0.05 was considered statistically significant.

Results

Human M1 and M2 macrophage polarization and lipid droplet characteristics

To investigate the relationship between macrophage polarization, efferocytosis, and lipid metabolism, we first characterized the phenotype of human monocyte-derived macrophages polarized in vitro. We differentiated human CD14⁺ monocytes, purified to > 90% purity (Supplementary Fig. 1A), into naive (M0) macrophages and subsequently polarized them into pro-inflammatory M1 or anti-inflammatory M2 phenotypes (Fig. 1A). We included the M0 population as a baseline for comparisons. Consistent with successful M2 polarization, M2 macrophages showed significantly upregulated surface expression of the mannose receptor CD206, the hemoglobin scavenger receptor CD163, and CD36 compared to both M0 and M1 macrophages (Fig. 1B). Expression of the pro-resolving mediator Annexin A1 was also significantly enhanced in M2 macrophages. In contrast, M1 polarization resulted in a significant increase in the secretion of the pro-inflammatory cytokine TNFα, as well as elevated mRNA levels of TNFα, IL1β, and the hypoxia-inducible factor HIF1α, confirming a pro-inflammatory phenotype (Fig. 1C). To further validate the M2 phenotype, we assessed additional M2-associated markers. M2 macrophages displayed significantly higher mRNA expression of the master regulator PPARγ and the surface receptors CD206, CD36, CD200R1 and Mertk relative to M0 and M1 cells (Fig. 1D).

Fig. 1.

Association of Lipid Droplet Characteristics with Enhanced Efferocytosis in Macrophages. (A) Schematic representation of the experimental workflow. Human CD14⁺ monocytes were differentiated into M0 macrophages using M-CSF, then polarized into M1 (IFN-γ/LPS) or M2 (IL-4) macrophages over 24 h. After polarization, CFSE-labeled apoptotic Jurkat cells were co-cultured with these macrophages to assess efferocytosis. (B) Surface expression of the M2-associated markers CD206 and CD163, the efferocytosis receptor CD36, and the pro-resolving mediator Annexin A1 on M0, M1, and M2 macrophages, quantified by flow cytometry across all viable cells or ELISA (n = 5–7 independent donors). (C) Secreted TNFα protein levels (ELISA) and relative mRNA expression of TNFα, IL1β, and HIF-1α in M0, M1, and M2 macrophages, analyzed by qRT-PCR (n = 5–7 independent donors). (D) Relative mRNA expression of M2-associated genes PPARγ, CD206, CD36, CD200R1 and Mertk in M0, M1, and M2 macrophages, analyzed by qRT-PCR (n = 5–7 independent donors). (E) Efferocytosis efficiency of M0, M1, and M2 macrophages, quantified as the percentage of macrophages engulfing CFSE-labeled apoptotic Jurkat cells (n = 7–10 independent donors). (F) Representative cell images and histogram plots illustrating CFSE positivity in M1 and M2 macrophages. (G) Total neutral lipid content in M1 and M2 macrophages measured by HCS LipidTOX™ Green staining and quantified through ImageStream (n = 4 independent donors). (H) Representative confocal microscopy images of M1 and M2 macrophages stained for LDs with PLIN2 (red) and nuclei with DAPI (blue). Scale bar = 10 μm. (I) Quantification of LD number and total LD volume per cell from confocal images. Data points represent the average value across the four independent donors, with at least 25 cells analyzed per donor. For panels B, C, D, and E, statistical significance was assessed using a one-way ANOVA with Tukey’s multiple comparisons test. For panels G a paired t-test was applied, and for I, unpaired t-test was applied. Data are shown as mean ± SEM, except E (box-and-whisker plots) and I (mean ± SD). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Since efferocytosis is a crucial function of pro-resolving macrophages, we then compared the ability of these polarized populations to clear apoptotic cells. M2 macrophages showed significantly improved efferocytosis efficiency compared to both M1 and M0 macrophages (Fig. 1E and F; Supplementary Fig. 1B-C for apoptosis induction protocol and detailed ImageStream gating strategy). The data confirm that our in vitro polarization model accurately reflects the expected functional dichotomy, with M2 macrophages showing a greater pro-resolving ability. M-CSF cultured M0 cells are similar to M2, and M1 cytokines push M0 cells toward a pro-inflammatory state. With this confirmation, we compared M1 and M2 afterward.

Based on the hypothesis that the macrophage polarization state alters lipid metabolism, which may in turn support distinct cellular functions, we next examined the characteristics of LDs. Staining with the neutral lipid dye LipidTOX revealed that M1 macrophages contained significantly higher total lipid content compared to M2 macrophages (Fig. 1G). To further characterize these differences, we visualized LDs by staining for the LD-coating protein PLIN2 ( [41, 42], Fig. 1H), which showed higher total LD volume in M1 macrophages and smaller, more numerous LD in M2 macrophages, suggesting distinct metabolic states between the two phenotypes (Fig. 1I). Collectively, these findings show that M1 and M2 polarization states are associated with distinct efferocytosis capacities and markedly different LD numbers, sizes, and overall lipid accumulation.

Hormone-sensitive lipase (HSL) activity is critical for maintaining M2 macrophage efferocytosis and pro-resolving functions

We observed that M2 macrophages exhibited significantly higher total lipase activity compared to M1 macrophages (Fig. 2A), suggesting an increased rate of lipolysis in the M2 state. To test if this increased lipolytic activity was functionally crucial for efferocytosis, we treated M2 macrophages with Orlistat, a pan-lipase inhibitor, and observed a dose-dependent suppression of efferocytosis (Fig. 2B; Supplementary Fig. 2A). To identify the specific lipases responsible, we utilized inhibitors for the main cytosolic lipases: ATGL, HSL, and MAGL (Fig. 2C). Inhibition of HSL most potently reduced efferocytosis in M2 macrophages, with ATGL inhibition showing a more modest but still significant effect. In contrast, MAGL inhibition had no impact (Fig. 2D; Supplementary Fig. 2B). We tested these inhibitors on M1 macrophages. Interestingly, only HSL inhibition significantly reduced efferocytosis in M1 cells, while ATGL and MAGL inhibitors had no effect (Fig. 2E). We next investigated alternate lipid metabolic pathways. Inhibiting triglyceride synthesis via a DGAT1 inhibitor (DGATi) did not alter efferocytosis in either M2 or M1 macrophages (Fig. 2F; Supplementary Fig. 2C). Despite reports indicating the critical role of lysosomal lipase in M2 activation [43], lysosomal lipase inhibition with Lalistat1 did not demonstrate any noticeable effect (Fig. 2G). Together, these results pinpoint cytosolic lipolysis, primarily driven by HSL, as a key process regulating macrophage efferocytosis.

Fig. 2.

HSL is a critical regulator of efferocytosis and the pro-resolving phenotype in M2 macrophages. (A) Total lipase activity in cell lysates from M1 and M2 polarized macrophages. (n = 4 independent donors). (B) Efferocytosis efficiency in M2 macrophages with a pan-lipase inhibitor, Orlistat (50 and 100 µM). (n = 4 independent donors). (C) Schematic of intracellular triglyceride (TAG) metabolism, illustrating pathways for lipid storage via DGAT and lipolysis via cytosolic lipases (ATGL, HSL, MAGL) or lysosomal lipophagy (LAL). (D) Efferocytosis efficiency in M2 macrophages following treatment with specific inhibitors for ATGL (ATGLi/Atglistatin, 20 µM), HSL (HSLi/ HSL-IN-1, 5 µM), or MAGL (MAGLi, 5 µM). (n = 4–5 independent donors). (E) Efferocytosis efficiency in M1 macrophages following treatment with the same specific lipase inhibitors. (n = 6 independent donors). (F-G) Efferocytosis efficiency in M2 macrophages after inhibition of (F) triglyceride synthesis with a DGAT1 inhibitor (DGATi/ T863, 50 µM) or (G) lysosomal acid lipase with Lalistat 2 (LALi, 5 µM). (n = 4 independent donors). (H) M2 macrophages treated with M2 and HSLi (HSL-IN-1, 5 µM) were stained for LDs using PLIN2 and imaged with confocal microscopy. Total LD volume per cell was measured with Imaris. (n = cells from 3 independent donors). (I) Representative high-resolution confirmation of the LD phenotype by transmission electron microscopy (TEM) of M1, M2, and HSL-inhibited M2 (M2 + HSLi) macrophages. Lower panels show magnified views of the boxed areas. Red arrows indicate LDs. Scale bars = 2 μm (top), 0.5 μm (bottom). (J) Quantification of LD size (left) and density (right) in M1, M2, and M2 + HSLi macrophages was conducted using ImageJ software from a single donor with triplicate measurements per condition. Each dot represents a single LD, and they are not used for statistical comparison between groups. (K) Surface expression of CD163 and CD36 on M2 macrophages with or without HSL inhibition, measured by flow cytometry. (n = 4–6 independent donors). (L) Secreted levels of the pro-resolving mediator Annexin A1 and the pro-inflammatory cytokine TNFα from M2 macrophages with or without HSL inhibition, measured by ELISA. (n = 5–6 independent donors). (M) Efferocytosis efficiency in murine bone marrow-derived macrophages (BMDMs) polarized to M1 or M2 phenotypes, with or without HSL inhibition. (n = 6–10 biological replicates). For panels A, F, G, H, K, L, a paired t-test was used. For panels B, D, E, and M, a one-way ANOVA with Tukey’s multiple comparisons test was used. Data in A shows paired individual donors. Panels B, D-G and K-M are presented as box-and-whisker plots showing median with all points. Data in J are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Having identified HSL as a key driver of efferocytosis, we next sought to directly assess how its inhibition impacts the macrophage lipid phenotype. We performed confocal microscopy and analyzed PLIN2-stained LDs, confirming that HSL inhibition in M2 macrophages (M2 + HSLi) significantly increased the total LD volume per cell compared with untreated M2 cells (Fig. 2H). To investigate the ultrastructural basis for this accumulation, we then performed transmission electron microscopy (TEM). High-resolution micrographs showed that, while untreated M2 macrophages contained small, distinct LDs (Fig. 2I, red arrows), HSLi-treated M2 macrophages exhibited large LDs similar to those in M1 macrophages. We further quantified the characteristics of individual LDs from a representative experiment, which confirmed a trend toward larger, less electron-dense droplets in the M2 + HSLi condition compared to M2 macrophages (Fig. 2J). This altered LD morphology in HSL-inhibited M2 cells closely resembled that of pro-inflammatory M1 macrophages. Functionally, HSL inhibition led to a significant downregulation of the surface receptor CD163 and the efferocytosis co-receptor CD36 on M2 macrophages (Fig. 2K). Critically, HSL inhibition significantly reduced the secretion of the pro-resolving mediator Annexin A1 while simultaneously increasing the secretion of the pro-inflammatory cytokine TNFα (Fig. 2L).

To determine whether these phenotypic changes were associated with alterations in the macrophage transcriptional program, we performed qRT-PCR. HSL inhibition in M2 macrophages caused a significant downregulation of M2-associated genes, including the master regulator PPARγ and the functional markers CD206, CD36, and CD200R1 (Supplementary Fig. 2D). Concurrently, HSL inhibition led to an upregulation of M1-associated pro-inflammatory genes TNFα, IL1β, and HIF1α (Supplementary Fig. 2E). To confirm these findings in a different species for translational relevance, we found that HSL inhibition similarly reduced efferocytosis in murine M2 bone marrow-derived macrophages (BMDMs) (Fig. 2M). Collectively, these findings demonstrate that HSL activity is not only essential for efficient efferocytosis but also necessary to sustain the morphological, functional, and transcriptional identity of M2 macrophages, thereby preventing their reversion to a pro-inflammatory state.

HSL activation via PKA-dependent phosphorylation and macrophage function

To elucidate the mechanism underlying the elevated lipase activity observed in M2 macrophages, we next examined the activation state of HSL. Immunoblot analysis revealed that M2 macrophages exhibited significantly higher levels of HSL phosphorylation at serine 660 (pHSL S660), an established activating modification, compared to M1 macrophages (Fig. 3A). Consistent with its role in LD metabolism, confocal microscopy showed that HSL prominently co-localized with LDs in M2 macrophages, a pattern that was less evident in M1 cells (Fig. 3B, C).

Fig. 3.

The PKA-HSL signaling axis is more active in M2 macrophages and drives efferocytosis. (A) Immunoblot and densitometric quantification of phosphorylated HSL (pHSL Ser660) relative to total HSL in M1 and M2 macrophage lysates. Data points represent individual donors, showing fold change relative to the M1 mean. (n = 6 independent donors). (B) Representative confocal microscopy images of M1 and M2 macrophages stained for lipid droplets (LipidTOX, red), HSL (green), and nuclei (DAPI, blue). The merged image shows co-localization of HSL with LDs, particularly in M2 macrophages. Scale bar = 10 μm. (C) Quantification of HSL and LD co-localization from confocal images. Each data point represents the mean co-localization score from one field of view. (n = 8 fields of view from 2 independent donors). (D) Basal Protein Kinase A (PKA) activity measured in M1 and M2 macrophage lysates. (n = 7 independent donors). (E) Efferocytosis efficiency in M1 macrophages treated with or without the PKA activator dibutyryl-cAMP (Db-cAMP, 50 µM) for 3 h. (n = 7 independent donors). (F) To confirm activator efficacy, PKA activity was measured in M1 macrophages treated with or without Db-cAMP (50 µM). (n = 4 independent donors). For all panels, a paired t-test was used. Data in A shows paired individual donors. Panels C, D, and F are presented as mean ± SEM. Data in panel E are presented as box-and-whisker plots showing median with all points. *P < 0.05, **P < 0.01, ***P < 0.001

The cAMP-Protein kinase A (PKA) signaling pathway is a known upstream regulator of HSL [44], phosphorylating it at S660 to enhance its enzymatic activity [45]. We therefore hypothesized that differential PKA activity might underlie the observed differences in HSL phosphorylation. Consistent with this, direct measurements of PKA activity revealed significantly higher activity in M2 macrophages compared to M1 cells (Fig. 3D). To functionally determine whether activation of this pathway in a pro-inflammatory context could restore efferocytic capacity, M1 macrophages were treated with dibutyryl-cAMP (Db-cAMP), a cell-permeable activator of PKA. This treatment significantly enhanced the efferocytosis capacity of M1 macrophages (Fig. 3E). We confirmed that Db-cAMP treatment did, in fact, significantly increase PKA activity in the M1 macrophages (Fig. 3F). These findings establish PKA as a critical regulator of LD lipolysis via HSL activation, directly connecting this pathway to the superior efferocytic capacity of M2 macrophages.

Lipolysis enzymes remodel the M2 lipidome to generate pro-resolving fatty acids

To test the hypothesis that HSL activity alters the lipid landscape to support M2 functions, we performed comprehensive lipidomic profiling of M1, M2, and HSL-inhibited M2 (M2 + HSLi) macrophages (Supplementary Tables 1–2). The heat map and PCA analysis confirmed that M1 and M2 macrophages occupy distinct lipidomic profiles (Fig. 4A, Supplementary Fig. 3A). Relative abundances and absolute quantities of lipid species indicate distinctive lipid species between M1 and M2 macrophages; phosphatidylinositol (PI) elevates in M2, while phosphatidylethanolamine (PE), Cholesterol, TAG, and Ceramide (Cer) elevate in M1 (Fig. 4A, Supplementary Fig. 3B). Notably, the lipid profile of HSL-inhibited M2 macrophages shifted away from the M2 cluster and toward the M1 cluster, providing direct evidence that HSL activity is essential for maintaining the M2-specific lipidome (Supplementary Fig. 3A). Unbiased analysis revealed a profoundly different lipid profile between M1 and M2 macrophages, with 81 lipid species (after FDR correction) significantly altered (Fig. 4B). Specifically, FA 24:1 and several PI species, including PI (16:0/20:4), PI (18:0/18:0), and LPI 18:0, increased in M2 macrophages. Notably, among FAs, the concentration of FA 24:1 was only significantly higher in M2 compared to M1 cells, which is independent of HSL inhibition. (Fig. 4C, Supplementary Fig. 3C). We also observed a corresponding decrease in Ceramide (d18:1/24:1) in M2 compared to M1 (Supplementary Fig. 3D) and an increase in the ceramide-catabolizing enzyme ASAH1 in M2 cells, suggesting a potential pathway for FA 24:1 production (Fig. 4D).

Fig. 4.

Lipidomic Profiling of M1, M2, and HSL-Inhibited M2 Macrophages. (A) Heatmap showing the relative abundance of major lipid classes in M1, M2, and HSL-inhibited M2 (M2 + HSLi) macrophages. (n = 7 for M1, M2; n = 4 for M2 + HSLi independent donors). Cer, Ceramide; SM, Sphingomyelin; DAG, Diacylglycerol; FFA, Free Fatty Acids; LPC, Lysophosphatidylcholine; PA, Phosphatidic Acid; PC, Phosphatidylcholine; PE, Phosphatidylethanolamine; PG, Phosphatidylglycerol; PI, Phosphatidylinositol; PS, Phosphatidylserine; TAG, Triacylglycerol; CE, Cholesteryl Ester; LPI, Lysophosphatidylinositol. (B) Volcano plot of lipidomic data comparing M1 and M2 macrophages derived from quantitative analysis. Significantly altered lipid species (FDR-adjusted p-value < 0.05, fold change > 1.5) are highlighted. (C) Absolute quantification of FA 24:1 in M1, M2, and M2 + HSLi macrophages, expressed as nmol/5 × 10⁵ cells. (n = 7 for M1, M2; n = 4 for M2 + HSLi). (D) ELISA quantification of secreted Acid Ceramidase (ASAH1) in culture supernatants (n = 5, independent donors), and relative mRNA expression of Acid Ceramidase (ASAH1, n = 3, independent donors) from M1 and M2 macrophages. A paired t-test was performed. (E) Efferocytosis efficiency was measured in M1 (left) and M2 (right) macrophages treated with Nervonic Acid (FA 24:1, 1, 10 or 50 µM) for 24 h. (n = 4 independent donors). (F) Efferocytosis capacity of M1- and M2-polarized murine bone marrow-derived macrophages (BMDMs) treated with increasing concentrations (1, 10, 50 µM) of Nervonic Acid (FA 24:1) for 24 h. (n = 4–6, biological replicates). (G-H) Nervonic acid enhances mitochondrial respiration in M1 macrophages. (G) A representative Seahorse XF Mito Stress test of M1 macrophages (M1 control, red line) or M1 macrophages treated with FA 24:1 (50 µM, blue line) is shown. The trace represents the mean ± SEM of three technical replicates from a single donor. OCR: oxygen consumption rate. 1.5 µM Oligomycin, 1 µM FCCP, 0.5 µM Rotenone + Antimycin A. (H) Quantification of key mitochondrial parameters, Basal Respiration, Maximal Respiration, Spare Respiratory Capacity, and ATP Production, was calculated using Wave software. Each dot represents the mean value from one of n = 3 independent donors. Data in panels C, E, and F are presented as box-and-whisker plots showing the median with all points. Panel D shows paired individual donors. Panel H is presented as mean ± SEM. Statistical analysis was performed using a paired t-test for panels

D and H; one-way ANOVA with Tukey’s multiple comparisons test for panels C, E, and F. ns, not significant; *P < 0.05, **P < 0.01

To directly test whether specific lipid species liberated by HSL causally mediate macrophage function, we performed an “add-back” experiment. We treated pro-inflammatory M1 macrophages with exogenous FA 24:1. Remarkably, in both human and mouse macrophages, the addition of FA 24:1 was enough to significantly improve the efferocytosis ability of M1 macrophages in a dose-dependent manner, but it had no further enhancement on M2 macrophages, which already have high efferocytosis capacity (Fig. 4E, F), suggesting that the effect is most pronounced in a pro-inflammatory setting. Finally, we assessed mitochondrial metabolism using a Seahorse assay. We found that treating M1 macrophages with FA 24:1 significantly increased their OCR, boosting both basal and maximal respiration and ATP production (Fig. 4G, H). In summary, our findings suggest that lipolysis in M2 macrophages alters the cellular lipid profile, with FA 24:1 serving as a crucial metabolic substrate that enhances mitochondrial respiration and supports efferocytosis.

Genetic modulation of HSL confirms its role in driving the M2 pro-resolving program

Our pharmacological HSL inhibitor data indicated a crucial role for HSL (Fig. 2 and Supplementary Fig. 2). To confirm that these effects were specifically attributable to HSL rather than off-target activity of the inhibitor, we performed genetic loss-of-function experiments using siRNA-mediated knockdown of HSL in M2 macrophages. HSL knockdown significantly impaired efferocytosis in human M2 macrophages (Fig. 5A). Consistent with a shift away from a pro-resolving state, HSL knockdown also reduced the secretion of Annexin A1 while simultaneously increasing the release of pro-inflammatory TNFα (Fig. 5B). At the transcriptional level, HSL knockdown caused a significant downregulation of the M2-associated gene program, including PPARγ, CD163, CD36, and CD200R1, and a concurrent upregulation of the M1-associated genes TNFα, IL1β, and HIF1α (Fig. 5C, D). The necessity of HSL for efferocytosis was further validated in murine BMDMs (Fig. 5E).

Fig. 5.

Genetic knockdown of HSL replicates the effects of pharmacological inhibition, thereby confirming its essential role in sustaining the M2 phenotype. Human M0 macrophages were transfected with control non-targeting siRNA or siRNA targeting HSL. (A) Efferocytosis efficiency was measured 48 h post-transfection (n = 6 independent donors). (B) Secreted levels of the pro-resolving mediator Annexin A1 and the pro-inflammatory cytokine TNFα in culture supernatants, measured by ELISA (n = 9–10 independent donors). (C) Relative mRNA expression of M2-associated genes (PPARγ, CD163, CD36, CD200R1) analyzed by qRT-PCR. (n = 5, independent donors). (D) Relative mRNA expression of M1-associated genes (TNFα, IL1β, HIF1α) and HSL/LIPE (n = 5, independent donors). (E) Efferocytosis was measured in murine M0 bone marrow-derived macrophages (BMDMs) following siRNA-mediated knockdown of HSL (n = 6 biological replicates). For panels A, B, and E, a paired t-test was used. For panels C and D, a two-way ANOVA was used. Data in A, B, and E are presented as box-and-whisker plots. Panels C & D show paired individual donors. *P < 0.05, **P < 0.01, ***P < 0.001

Furthermore, HSL overexpression into M0 macrophages significantly enhanced efferocytosis (Fig. 6A). HSL overexpression drove the acquisition of an M2-like state, characterized by increased surface expression of CD206, CD163, and CD36, and boosted secretion of Annexin A1 (Fig. 6B). Remarkably, the upregulation of CD206 and CD36 was completely reversed upon co-treatment with an HSL inhibitor (Supplementary Fig. 4A), demonstrating that the enzymatic activity of HSL is the driver of this pro-resolving surface phenotype. Alongside, HSL overexpression suppressed the secretion of pro-inflammatory cytokines TNFα and IL-1β (Fig. 6C). We also examined downstream metabolic gene expression and mitochondrial function. HSL overexpression led to a significant upregulation of the M2-master regulator PPARγ and, critically, of genes involved in fatty acid β-oxidation (FAO), including ACADVL and CPT1a (Fig. 6D). Other FAO-related genes, such as FABP3, CPT2, CACT, or MFN2 showed no significant difference (Supplementary Fig. 4B). This demonstrated a transcriptional shift towards FA catabolism. Functionally, we found that HSL-overexpressing cells exhibited increased maximal OCR and greater spare respiratory capacity (Fig. 6E, F). Remarkably, overexpression of HSL alone, in the absence of the M2-polarizing cytokine (IL-4), was sufficient to induce a pro-resolving phenotype in naïve macrophages.

Fig. 6.

HSL overexpression is sufficient to drive a pro-resolving phenotype. Human M0 macrophages were transfected with an empty vector or an HSL-encoding plasmid. (A) Efferocytosis efficiency was measured 24 h post-transfection (n = 5 independent donors). (B) Surface expression of M2-associated markers CD206, CD163, and CD36 (gMFI by flow cytometry) and secreted levels of Annexin A1 (ELISA, n = 4–6, independent donors). (C) Secreted levels of pro-inflammatory cytokines TNFα and IL-1β were measured by ELISA (n = 6–7 independent donors). (D) Relative mRNA expression of the master regulator PPARγ and genes involved in fatty acid oxidation, ACADVL and CPT1a (n = 5–6, independent donors). (E-F) HSL overexpression enhances mitochondrial respiratory capacity in macrophages. (E) A representative Seahorse XF Mito Stress test of macrophages transfected with a control vector (Vector, red line) or an HSL-encoding plasmid (HSL, blue line) is shown. The trace represents the mean ± SEM of three technical replicates from a single donor. OCR: oxygen consumption rate. 1.5 µM Oligomycin, 1 µM FCCP, 0.5 µM Rotenone + Antimycin A. (F) Basal respiration, maximal respiration, spare respiratory capacity, and ATP production were determined using the Seahorse XF Cell Mito Stress Test report generator. Each dot indicates the average from one of n = 6 independent donors. For all panels, a paired t-test was used. Data in panels A, B (Annexin A1) and C are presented as box-and-whisker plots and in panels F as mean ± SEM. Panels B & D show paired individual donors. *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant

Discussion

The clinical impact of macrophage polarization is seen in many diseases [6, 46–48], especially in SLE. A key feature of SLE is impaired efferocytosis, where macrophages fail to clear apoptotic cells properly. This leads to necrosis and the release of autoantigens, fueling inflammation [49, 50]. In this study, we establish a causal link between lipid droplet metabolism, specifically lipolysis driven by HSL, and the functional identity of human pro-resolving M2 macrophages. We demonstrate that M2 macrophages can use a PKA-HSL signaling axis to mobilize fatty acids from intracellular LDs. This mechanism was confirmed through pharmacological inhibition, which reverted the M2 phenotype, and more definitively through genetic loss-of-function (siRNA) and gain-of-function (overexpression) studies, establishing HSL as both necessary and sufficient to drive the pro-resolving macrophage state. Our comprehensive model, summarized in Fig. 7, positions HSL as a critical metabolic checkpoint that orchestrates the pro-resolving capacity of macrophages.

Fig. 7.

Proposed model for HSL-driven metabolic and transcriptional reprogramming of pro-resolving macrophages. This schematic highlights the central role of the PKA-HSL axis in controlling the M2 pro-resolving phenotype. Upstream signals, such as those via GPCRs, increase intracellular cAMP, which activates Protein Kinase A (PKA). PKA then phosphorylates and activates HSL, leading to its translocation to lipid droplets (LDs). Elevated levels of acid ceramidase (ASAH1) in M2 macrophages may degrade ceramides (Cer) into sphingosine and FAs. These FAs, including Nervonic Acid (FA 24:1), act as important metabolic and signaling intermediates. The activation of HSL promotes a transcriptional shift by serving as natural ligands for the nuclear receptor PPARγ and by suppressing NF-κB signaling. PPARγ induces the expression of a pro-resolving gene program (e.g., CD206, CD36, Annexin A1) while reducing the transcription of pro-inflammatory genes (e.g., TNFα, IL1β). External supplementation of FA 24:1 may occur through CD36 or fatty acid transporters (FATs), supporting fatty acid oxidation (FAO) and enhancing pro-resolving activities, particularly efferocytosis. The dotted lines represent pathways that are not fully understood. Created with Biorender.com

Our initial observation that M1 macrophages accumulate large LDs, whereas M2 macrophages maintain smaller, more numerous droplets, aligns with previous findings in murine models. Previous studies have shown that IFN-γ and LPS induce the accumulation of TAG, DAG, and CE, leading to increased neutral lipid staining in murine bone marrow-derived macrophages compared to M0 or IL-4-induced M2 macrophages [51–53]. Specifically, LPS alone promotes DAG accumulation, while IFN-γ and LPS together induce CE accumulation [52]. Additionally, IFN-γ and LPS co-stimulation led to TAG accumulation enriched with polyunsaturated FA (PUFA), whereas IFN-γ alone results in TAG accumulation with fewer PUFA [52]. Our study confirmed neutral lipid accumulation in human M1 macrophages differentiated from purified human CD14⁺ monocytes. Since we could detect only a single species of CE in our lipidomic analysis, definitive conclusions regarding CE metabolism remain tentative.

HSL has been extensively studied in adipose tissue, where hormonal and nutritional signals intricately regulate its activity. Although HSL predominantly exhibits high activity levels in white and brown adipose tissue, it is also expressed in several other tissues, including the testes, adrenal glands, ovaries, cardiac muscle, and macrophages [54–56]. HSL-deficient mice exhibit DAG accumulation in white and brown adipose tissue, skeletal muscle, cardiac muscle, and testis [57, 58]. The activation of HSL is primarily regulated by hormonal signals, especially those mediated through the cAMP pathway, where PKA plays a crucial role. When activated, PKA phosphorylates HSL at specific serine residues (Ser563, Ser659, and Ser660) [59], significantly enhances its lipase activity. This phosphorylation increases the substrate affinity of HSL and facilitates its translocation from the cytosol to LD, optimizing lipolysis [60, 61].

M2 macrophages have higher basal PKA activity compared to M1 counterparts, and Db-cAMP significantly increases PKA activity in M1 macrophages (Fig. 3E, F). While the canonical IL-4/STAT6 pathway is the well-established initiator of the M2 transcriptional program, our findings suggest that direct activation of PKA in M1 macrophages (using Db-cAMP) was sufficient to significantly enhance efferocytosis. We note that this enhancement did not fully restore efferocytosis to the level observed in M2 macrophages, suggesting that while the PKA-HSL axis is a critical component, other M2-specific factors also contribute to maximal efferocytosis function. This intervention, however, did not fully recapitulate the entire M2 gene signature or surface phenotype. Moreover, an intriguing aspect of our findings is that HSL’s role in efferocytosis is not limited to M2 macrophages; its inhibition also reduces the basal efferocytosis ability of pro-inflammatory M1 cells (Fig. 2D). This indicates a dual function for HSL. We suggest that it serves a homeostatic role in all macrophages by providing the necessary fatty acids. Additionally, it has a specialized, inducible signaling function, where the PKA-HSL axis in M2 cells is highly activated to produce specific mediators that coordinate the complete pro-resolving program. Therefore, while a basic level of HSL activity seems essential for the mechanics of cell clearance, its inducible activity acts as a therapeutically relevant enhancer of inflammation resolution.

Our lipidomic analysis revealed significant alterations in membrane lipid components, particularly PI, upon following HSL inhibition. The decrease in PI species following HSL inhibition suggests their involvement in modulating macrophage functions through lipid signaling pathways that modulate macrophage function. PIs are known to participate in various signaling cascades that influence inflammation, cell survival, and vesicular trafficking processes essential for macrophage function [62, 63]. Specifically, PI serves as a precursor for phosphatidylinositol phosphates (PIPs), which are involved in signaling pathways governing essential processes such as cell migration, phagocytosis, and cytokine production [62, 64]. Additionally, plasma membrane PIPs have been shown to enhance the active conformation of β-arrestin and stabilize GPCR-β-arrestin complexes, which are critical for signaling modulation [65]. GPCR activation leads to cAMP production, which in turn activates PKA and may contribute to the polarization of M2 macrophages. Given that GCPR activation stimulated cAMP production and PKA signaling, these pathways may contribute to M2 macrophage polarization.

Lipidomic analysis identified FA 24:1 (nervonic acid, NA) enriched in M2 macrophages (Fig. 4B and C), and ceramide hydrolyzing enzyme ASAH1 levels are higher in M2 macrophages compared to M1 macrophages (Fig. 4D). Adding NA to pro-inflammatory M1 macrophages increased their efferocytosis and shifted their metabolism toward a more oxidative, M2-like state. We suggest that ASAH1 and HSL-derived fatty acids induce the PPARγ pathway, which in turn activates genes involved in efferocytosis and FAO (Fig. 7) [11]. NA may also influence membrane fluidity and lipid rafts, thereby affecting the surface expression of receptors on the plasma membrane. In addition, NA administration was reported to downregulate classic pro-inflammatory cytokines TNFα and IL6 and upregulate the anti-inflammatory mediator IL10, suggesting its ability to influence macrophage polarization toward a pro-resolving (M2-like) phenotype [66]. While future studies are needed to determine whether NA directly activates HSL, our findings firmly establish a connection between an HSL-derived lipid and the induction of a pro-resolving macrophage phenotype.

Boosting HSL activity could help improve efferocytosis and promote inflammation resolution, offering a potential therapeutic angle. On the other hand, in certain cancers, tumor-associated macrophages (TAM) create an environment that supports tumor growth and suppresses an anti-tumor immune response [67]. Blocking HSL could disrupt their lipid metabolism, possibly enhancing anti-tumor immunity, and reducing tumor-promoting effects. Beyond immune-related conditions, genetic studies have linked HSL variants to type 2 diabetes, familial partial lipodystrophy, and even Parkinson’s disease [68, 69]. These connections highlight the broader role of HSL in metabolism, making it a potential therapeutic target in macrophage-driven diseases and metabolic disorders. HSL activity is further influenced by LD-associated proteins, such as ApoL6 and PLIN1 [70]. ApoL6 interacts with PLIN1 to suppress HSL, resulting in larger LDs and increased TAG storage. Understanding these regulatory pathways could provide therapeutic avenues for modulating macrophage-driven inflammation and metabolic diseases.

Lipid metabolism is essential for phagocytosis [34]. While glucose is the primary energy source for macrophages, FAs also play a critical role in fueling immune functions [71]. Yin et al. investigated whether FA released from lipoproteins by macrophage lipoprotein lipase (LPL) could substitute glucose as an energy source for phagocytosis [72]. They demonstrated impaired Fc receptor-mediated phagocytosis in macrophages from LPL knockout mice compared to normal mice, suggesting FA are vital for macrophage energy. Similarly, ATGL-deficient macrophages in mice showed impaired phagocytosis of Escherichia coli particles [73]. These macrophages failed to efficiently hydrolyze TAG, resulting in decreased cellular FA concentrations and increased LD accumulation. Although less explored, efferocytosis in the context of LD has been investigated; van Dierendonck et al. found that HILPDA, which inhibits ATGL, did not affect efferocytosis in LPS-activated murine macrophages [74]. Our study builds on these findings by identifying the PKA-HSL axis as a key regulator of LD lipolysis in M2 macrophages, directly linking this pathway to efficient efferocytosis.

Furthermore, our study addresses the critical question of how liberated fatty acids are metabolically utilized by macrophages. Efferocytosis is an energetically demanding process that requires a constant supply of ATP [75]. By employing Seahorse metabolic analysis, we provide direct evidence that the fatty acids liberated by HSL are shuttled into the mitochondria to fuel FAO. Overexpression of HSL in naive M0 macrophages was sufficient to increase the expression of key FAO enzymes (CPT1a, ACADVL), which may enhance basal and maximal mitochondrial respiration (Fig. 6D-F). Beyond serving as a metabolic fuel, the profound alteration of the cellular lipidome may have other functional consequences. We observed that M2 macrophages have a more fluid plasma membrane compared to M1 cells, and this was reduced by HSL inhibition (Supplementary Fig. 4C). This suggests that HSL-mediated lipid remodeling may also facilitate efferocytosis by increasing membrane plasticity, potentially enhancing the ability of receptors like CD36 to cluster and engage with apoptotic cells.

Limitations and future directions

While our study provides strong mechanistic evidence using a human in vitro system, it has several limitations that open avenues for future research. First, our findings need to be validated in vivo. The use of conditional knockout mice with macrophage-specific deletion of HSL/LIPE would be the critical next step to confirm the role of this pathway in models of sterile inflammation, infection resolution, or tumorigenesis. Second, our study demonstrates that HSL activation promotes the expression of key efferocytosis machinery like Mertk and CD36. However, efferocytosis is a multi-step process involving tethering, recognition, engulfment, and degradation. Future studies should investigate whether HSL-driven metabolic reprogramming also influences the surface localization and clustering of these receptors, as well as the downstream signaling cascades initiated upon binding to apoptotic cells. Third, while we propose that HSL-derived fatty acids may serve as endogenous PPARγ ligands, the direct identity of these ligands and the mechanism of NF-κB suppression remain to be fully elucidated. Fourth, our lipidomic profiling was performed at a single time point after HSL inhibition. Future studies involving a detailed time-course analysis that tracks the kinetics of lipidome changes along with the onset of gene expression and efferocytosis would be valuable for developing a comprehensive, dynamic model of this metabolic shift. Finally, while we have identified the PKA–HSL axis as a central mediator of M2 macrophage function, the specific endogenous GPCRs that activate this pathway in M2 macrophages remain to be elucidated.

Conclusion

In conclusion, this study establishes HSL not merely as a metabolic enzyme but as a master regulator of the macrophage pro-resolving phenotype. We demonstrate that an active PKA-HSL-FAO metabolic axis is essential for maintaining the transcriptional identity and functional capacity of M2 macrophages. By coupling lipid droplet metabolism to mitochondrial energy production and nuclear receptor signaling, HSL acts as a pivotal node that enables macrophages to efficiently clear apoptotic cells and promote the resolution of inflammation. Targeting this pathway offers a promising strategy for therapeutic intervention in a host of macrophage-driven diseases. These findings also provide a foundation for exploring lipid-based approaches to restore immune homeostasis in autoimmune disorders such as SLE.

Authors’ contributions

H.W. and M.S. designed the research. H.W., B.P., M.L., and M.S. performed experiments and analyzed data. J.K. provided expert interpretation of the lipidomics data. H.W., B.P., M.L., and M.S. wrote the manuscript. H.S.S., and B.D. provided revisions. M.S. supervised all aspects of the project. All authors have read and approved the final manuscript. H.W.'s current address is: Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Faculty of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, 700-8558, Japan

Funding

This work was supported by grants from the National Institutes of Health (R01AI135063 and R21AR084084). M.S. was also supported by an Advancing Women in Science and Medicine (AWSM) award from the Feinstein Institutes for Medical Research.

Data availability

The lipidomics data generated in this study are provided in the supplementary information and source data files. All other source data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were conducted following the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of the Feinstein Institutes for Medical Research (Approval No. 2022-022). All human samples were collected from healthy, anonymized donors under protocols approved by the Institutional Review Board (IRB) of the Feinstein Institutes for Medical Research, which granted an exemption based on the secondary use of de-identified biological material.

Consent for publication

Not applicable.

Competing interests

A provisional patent application has been filed by the assignee, the Feinstein Institutes for Medical Research, based upon the subject matter contained in this manuscript. M.S. is a named inventor on this application. All other authors declare that they have no competing interests.