Lipid-Associated Macrophages in Obesity: Immunometabolic Programs and Emerging Therapeutic Implications—A Narrative Review

Abstract

Obesity-associated adipose tissue inflammation is accompanied by the expansion of macrophage populations enriched in lipid-handling, lysosomal, and phagocytic programs, commonly referred to as lipid-associated macrophages (LAMs). Single-cell and spatial studies support their accumulation in crown-like structures and other lipid-rich adipose tissue niches, with partially conserved LAM-like populations also identified in humans. However, direct functional evidence, particularly in human adipose tissue, remains limited. Experimental studies suggest that LAMs may participate in lipid sequestration, efferocytosis, lysosomal processing, and inflammatory regulation, although their net effects appear to vary according to species, adipose depot, metabolic stage, and experimental model. The proposed transition from an adaptive lipid-buffering state to a chronically inflammatory state should therefore be regarded as a context-dependent working model rather than an established temporal sequence. This narrative review critically examines how LAMs are defined, distinguishes direct adipose LAM evidence from findings extrapolated from broader macrophage systems, and integrates lipid metabolism, bioenergetic remodeling, inflammatory signaling, and therapeutic evidence within an explicit evidence hierarchy. Current interventions act mainly through broader macrophage, adipose tissue, or systemic metabolic pathways, and no selective LAM-targeted therapy is currently available. Further translation will require robust human LAM definitions, longitudinal functional studies, reliable biomarkers, and strategies that modify detrimental LAM-associated functions without disrupting beneficial macrophage activities.

Graphical Abstract

Introduction

Obesity has become one of the most important public health challenges worldwide. Its prevalence has continued to rise over the past decades, and it is now widely recognized as a chronic disease characterized by dysregulated energy metabolism.^1,2 In recent years, obesity has increasingly been viewed not only as a metabolic disorder, but also as a chronic inflammatory condition. A large body of evidence indicates that obesity is marked not only by excessive adipose tissue accumulation, but also by a persistent state of chronic low-grade inflammation. This inflammatory response, driven by nutrient excess and metabolic imbalance, is referred to as metaflammation and is considered a key pathological basis of obesity-associated metabolic dysfunction.³

Adipose tissue is not only an energy-storage organ, but also an immunometabolic organ with important immune regulatory functions. In addition to adipocytes, adipose tissue contains a variety of immune cells, among which macrophages represent one of the most abundant and functionally important populations. Under obese conditions, adipose tissue undergoes profound remodeling, including adipocyte hypertrophy, local hypoxia, cell death, and immune cell infiltration. Among these changes, the massive recruitment and activation of macrophages are considered central drivers of adipose tissue inflammation and metabolic disturbance. By secreting inflammatory mediators and regulating adipocyte metabolism, adipose tissue macrophages contribute to the development and progression of obesity-related metabolic abnormalities, including insulin resistance.^4,5

Traditionally, macrophages have been classified into pro-inflammatory M1 and anti-inflammatory M2 phenotypes according to their functional states. However, the conventional M1/M2 polarization model cannot fully explain the complex functional states of macrophages in obese adipose tissue or their roles in lipid metabolism.^4,6 With advances in single-cell and spatial profiling, the heterogeneity of adipose tissue macrophages has become increasingly apparent, and lipid-associated macrophages (LAMs) have emerged as a prominent obesity-associated macrophage population.^6,7 In this review, LAMs are operationally defined as adipose tissue macrophage populations identified through convergent evidence across several domains: enrichment of a lipid-handling, lysosomal, and phagocytic transcriptomic program, commonly involving genes such as TREM2, CD9, LPL, APOE, LGALS3, LIPA, CTSB, and CTSL; protein-level detection of markers including TREM2, CD9, or CD63; preferential localization to crown-like structures and other lipid-rich remodeling niches; and, where available, functional evidence related to lipid uptake, lysosomal processing, phagocytosis, or efferocytosis.^6–8 No single transcriptomic or protein marker is considered sufficient to define a LAM. Throughout this review, the term “LAM population” is used for adipose tissue macrophages supported by convergent molecular and spatial evidence. The term “LAM-like macrophages” is used for human or cross-tissue populations that share part of the LAM-associated signature but whose functional equivalence to adipose LAMs has not been fully established. Because computational trajectory analyses identify candidate developmental relationships rather than definitive ontogenetic ancestry, the term “lineage” is reserved for conclusions supported by direct fate-mapping or equivalent lineage-tracing evidence.

Persistent lipid overload can reshape macrophage metabolism through coordinated changes in lipid processing, energy metabolism, lysosomal activity, and inflammatory signaling.^4,9 These processes are closely associated with the emergence of LAM-associated programs and the remodeling of macrophage functions in obese adipose tissue.^7,8,10 However, several important questions remain unresolved. Direct evidence obtained from molecularly and spatially defined adipose LAM populations is often discussed together with findings from total adipose tissue macrophages, macrophages in other tissues, or general immunometabolic models. The developmental origin of LAMs, the functional significance of their transcriptomic programs, the proposed shift from adaptive lipid buffering to chronic inflammatory activity, and the extent to which mouse findings can be translated to human adipose tissue also remain incompletely defined. Translation is further constrained by the absence of standardized markers and clinically useful biomarkers for human LAM-like populations, uncertainty across adipose depots and sexes, and the lack of strategies that selectively deliver interventions to adipose LAMs without disturbing macrophages in other tissues.

Accordingly, this narrative review aims to provide an operational definition of adipose LAMs; critically examine the evidence supporting their origin, differentiation, and context-dependent functions; distinguish mechanisms directly demonstrated in adipose LAMs from those inferred from broader macrophage systems; compare findings from mouse models with the more limited functional evidence available in human adipose tissue; and evaluate whether proposed therapeutic strategies act directly on LAM-associated targets or influence LAM biology indirectly through macrophage recruitment, adipose tissue remodeling, weight loss, or systemic metabolic regulation. By applying this evidence-oriented framework, the review seeks to integrate lipid handling, bioenergetic remodeling, inflammatory signaling, and therapeutic implications while avoiding the interpretation of associative or cross-tissue findings as established LAM-specific mechanisms.

Review Methodology

This structured narrative review was developed using a literature-search and evidence-classification strategy. PubMed/MEDLINE, Web of Science Core Collection, and Scopus were searched from database inception to 1 April 2026. Search terms were organized into three principal domains: (1) LAM identity and adipose tissue localization, including “lipid-associated macrophage”, “TREM2-positive macrophage”, “CD9 macrophage”, “adipose tissue macrophage”, “obesity”, and “adipose tissue” (2); immunometabolic mechanisms, including “lipid metabolism”, “glycolysis”, “mitochondrial metabolism”, “lysosome”, “phagocytosis”, “efferocytosis”, and “inflammation”; and (3) therapeutic and weight-loss-related interventions, including “therapy”, “pharmacotherapy”, “weight loss”, “bariatric surgery”, “GLP-1”, “semaglutide”, “tirzepatide”, and “macrophage targeting”. The reference lists of key original studies and relevant reviews were also screened to identify additional foundational, recent, or contradictory publications.

Original studies involving human adipose tissue, mouse models of obesity, single-cell or spatial profiling, flow cytometry, tissue imaging, genetic perturbation, lineage-tracing approaches, or functional metabolic assays were prioritized. Recent studies were emphasized, whereas earlier publications were retained when they provided foundational evidence for adipose tissue macrophage biology, metaflammation, or LAM identification. Studies were considered relevant when they addressed LAM identity, origin, spatial distribution, lipid handling, lysosomal or bioenergetic remodeling, inflammatory function, metabolic consequences, or therapeutic modulation. Conference abstracts, duplicate reports, studies with insufficient methodological information, and publications without a clear connection to obesity-associated adipose macrophage biology were excluded.

Titles, abstracts, and full texts were screened for relevance according to these criteria, and the final study set and evidence classifications were subsequently verified. Any disagreements were resolved by consensus.

To avoid treating heterogeneous findings as equivalent, evidence was classified according to its directness for adipose LAM biology. Category I evidence comprised studies in which adipose LAMs were identified using convergent molecular, protein, spatial, and/or phenotypic criteria and were further examined using functional or genetic perturbation. Category II evidence comprised studies providing direct molecular or spatial identification of adipose LAMs without selective functional or causal validation. Category III evidence comprised supportive or comparative findings derived from total adipose tissue macrophages, broad myeloid populations, global genetic models, macrophages from other tissues, in vitro macrophage systems, or systemic interventions in which molecularly and spatially defined adipose LAMs were not selectively assessed. These categories describe the directness of evidence for adipose LAM biology rather than the overall methodological quality of individual studies.

Transcriptomic enrichment was interpreted as evidence of an associated molecular program rather than proof of enhanced functional capacity. Similarly, computational trajectory analyses were considered hypothesis-generating evidence of candidate developmental relationships and were not interpreted as definitive lineage evidence unless supported by longitudinal analysis, fate mapping, or equivalent lineage-tracing experiments.

Human and animal evidence was evaluated separately. Contradictory findings were retained and compared according to species, sex, adipose depot, diet composition and duration, obesity stage, genetic model, method of LAM identification, and metabolic outcome. Therapeutic studies were additionally classified according to whether the intervention modulated a LAM-associated molecular pathway, broadly altered adipose tissue macrophage biology, or acted indirectly through weight loss or systemic metabolic improvement. Because this article was designed as a structured narrative review rather than a systematic review or meta-analysis, no quantitative synthesis or formal risk-of-bias assessment was performed.

Changes in Macrophages in Obese Adipose Tissue

Remodeling of the Immune Microenvironment in Obese Adipose Tissue

During the development of obesity, adipose tissue initially undergoes a transition from adaptive lipid storage to stress-associated expansion. As adipocytes continue to enlarge beyond the local capacity for oxygen diffusion and angiogenesis, adipose tissue gradually develops focal hypoxia and insufficient perfusion.^11,12 Hypoxia is not merely an accompanying feature of obese adipose tissue; through HIF-dependent signaling, it may promote inflammatory, pro-fibrotic, and metabolically maladaptive gene programs in adipocytes and stromal cells.¹² However, whether adipose hypoxia is universally present in obesity remains controversial, likely owing to differences in detection methods, adipose tissue blood flow, depot characteristics, and obesity severity. Therefore, adipose tissue hypoxia is better understood as a context-dependent pathological alteration.

In addition to hypoxia, nutrient excess can increase the burden of protein folding and lipid processing in adipocytes, thereby inducing endoplasmic reticulum (ER) stress and amplifying inflammatory responses.^13,14 Studies have shown that proteostatic imbalance and ER stress activation in obese adipose tissue can promote the expression of pro-inflammatory adipokines and chemokines, while enhancing the infiltration of pro-inflammatory macrophages.^13,15 These findings suggest that the proteostasis disruption–ER stress–inflammation axis may represent an important mechanism underlying inflammatory remodeling in obese adipose tissue.

Adipocyte death is a key event driving the transformation of the immune microenvironment in adipose tissue.¹⁶ Lipid droplets, cellular debris, and danger-associated molecular patterns (DAMPs) released from dying adipocytes can induce the local accumulation of macrophages and the formation of crown-like structures (CLSs).^16,17 Because adipocytes are large and lipid-rich, their clearance usually requires the coordinated engulfment by multiple macrophages, resulting in the formation of CLSs, one of the most characteristic histological features of obese adipose tissue.

These stress responses and cell death processes further promote immune cell recruitment through chemokine signaling, among which the monocyte chemoattractant protein-1 (MCP-1/CCL2)–CCR2 axis is particularly important.¹⁸ This pathway not only mediates monocyte migration into adipose tissue, but is also closely associated with the development of metabolic complications such as insulin resistance and fatty liver disease. It is therefore regarded as an important link between local adipose tissue inflammation and systemic metabolic dysfunction. Together, these changes drive immune microenvironment remodeling and macrophage accumulation in obese adipose tissue (Figure 1).

Figure 1.

Remodeling of the adipose-tissue microenvironment and macrophage accumulation during obesity. In lean adipose tissue, adipocytes are relatively small and macrophages are sparsely distributed. Obesity-associated adipocyte hypertrophy, hypoxia, and endoplasmic reticulum (ER) stress promote adipocyte dysfunction and death, with the release of lipids and damage-associated molecular patterns (DAMPs). C–C motif chemokine ligand 2 (CCL2)–C-C chemokine receptor type 2 (CCR2) signaling recruits circulating monocytes, which contribute to adipose-tissue macrophage (ATM) accumulation and crown-like structure (CLS) formation around dead or dying adipocytes. Dashed red arrows denote associations of CLS-associated macrophage accumulation with TNF-α, IL-6, IL-1β, and insulin resistance; they do not imply that these outcomes are specific to molecularly defined LAMs.

Abbreviations: ATM, adipose-tissue macrophage; CCL2, C–C motif chemokine ligand 2; CCR2, C–C chemokine receptor type 2; CLS, crown-like structure; DAMP, damage-associated molecular pattern; ER, endoplasmic reticulum; IL, interleukin; LAM, lipid-associated macrophage; TNF-α, tumor necrosis factor α.

Increased Macrophage Abundance and Spatial Redistribution

The number of adipose tissue macrophages (ATMs) is markedly increased in obese adipose tissue, representing one of the most characteristic histological features of obesity-associated adipose inflammation.¹⁹ Macrophages constitute a major immune-cell population in adipose tissue, and their abundance and phenotypic diversity increase substantially during obesity, in parallel with tissue inflammation and metabolic dysfunction.^20,21 It should be noted that the reported proportion of macrophages varies across studies owing to differences in quantification methods, adipose depot location, and obesity models. Nevertheless, the overall trend of a substantial increase in macrophage abundance is highly consistent.

With the development of spatial imaging and quantitative approaches, it has become increasingly clear that obesity alters not only the number of ATMs, but also their spatial distribution. In lean adipose tissue, macrophages are generally dispersed throughout the interstitial space, whereas in obesity they progressively accumulate around dead or dying adipocytes, forming the characteristic CLSs.¹⁶ The increased abundance of ATMs is thought to arise mainly from two sources: the recruitment of circulating monocytes and the local proliferation of tissue-resident macrophages.¹⁹ These two mechanisms may coexist at different stages of obesity and are likely regulated by local microenvironmental signals.

Functional Reprogramming and Metaflammation

Obesity-associated adipose tissue inflammation is characterized by a low-grade, chronic, and metabolically coupled nature.³ During obesity progression, ATMs, together with other immune cells, continuously produce pro-inflammatory mediators and interact with adipocytes, thereby forming a self-amplifying inflammatory loop.²² Early studies established a causal link between chronic adipose tissue inflammation and insulin resistance, while more recent work has shown that this process is driven by the interplay of multiple inflammatory and metabolic pathways and is profoundly influenced by local lipid burden and tissue remodeling status.²² Macrophages can promote metabolic inflammation, participate in inflammatory resolution, and restore tissue homeostasis, indicating that obesity-associated macrophage functions should not be interpreted as uniformly pathogenic.²³

Within the traditional phenotypic framework, ATMs in obesity have often been described as shifting from an M2-like to an M1-like state, characterized by the upregulation of pro-inflammatory cytokines such as TNF-α, IL-6, and IL-1β, as well as inducible nitric oxide synthase (iNOS)-related pathways.²⁴ Through these changes, ATMs contribute to systemic insulin resistance by interfering with key nodes of insulin signaling, including IRS-1 and AKT.²⁵ However, accumulating evidence indicates that obese ATMs cannot be fully equated with classically lipopolysaccharide (LPS)-induced M1 macrophages.^26,27 Recent single-cell studies further indicate that ATM activation encompasses multiple context-dependent states and cannot be reduced to a binary M1/M2 model.^28,29

Metabolically activated macrophages (MMe) and LAMs are related but non-equivalent constructs. MMe describes a broader experimental and conceptual activation framework induced by nutrient excess, glucose, insulin, and fatty acids, and is characterized by mixed inflammatory and metabolic features. By contrast, adipose LAMs are identified through convergent molecular and spatial evidence, including a lipid-handling and lysosomal transcriptomic program, supportive protein markers, and enrichment in crown-like structures or other lipid-rich remodeling niches. An MMe phenotype alone is therefore insufficient to establish LAM identity.^7,8 Some obesity-associated adipose tissue macrophages show overlapping MMe- and LAM-associated features, including lipid uptake, lysosomal activity, lipid-droplet handling, and inflammatory signaling. This overlap is partial and context dependent. In the remainder of this review, MMe evidence is treated as supportive adipose-tissue-macrophage evidence unless a molecularly and spatially defined LAM population was directly assessed.

Discovery and Characteristics of Lipid-Associated Macrophages

Discovery of LAMs

Single-cell transcriptomic studies established the concept of adipose LAMs. Jaitin and colleagues identified a TREM2-associated macrophage population that expanded in obese mouse adipose tissue and showed a partially conserved transcriptional program in human samples. Based on coordinated enrichment of genes involved in lipid uptake, lipid catabolism, and lysosomal function—including LPL, LIPA, CD36, CTSB, CTSL, FABP4, and FABP5—this population was termed LAMs.⁷ These data defined a lipid-handling and lysosomal program and nominated TREM2 as an important LAM-associated node, but did not establish a unique lineage or a single-marker definition.

Subsequent flow-cytometric and immunofluorescence studies provided complementary protein-level and spatial support for CD9+CD63+ macrophages enriched in obese adipose tissue and crown-like-structure-rich niches. Co-expression of TREM2 and CD9 further connected transcriptomic identification with protein-marker validation and spatial localization.^8,30 Together, these findings support a reproducible LAM-associated macrophage population or state, while marker expression alone does not establish complete functional equivalence across species or studies.

Further evidence from spatial transcriptomics and tissue imaging studies demonstrated that LAMs display clear spatial organization within adipose tissue, are closely associated with CLSs, and progressively expand during obesity progression.⁸ These observations indicate that LAMs are not only a newly recognized macrophage population in obese adipose tissue, but are also spatially coupled to the local microenvironmental niche created by adipocyte death and lipid release.

Markers and Phenotypic Features of LAMs

A prominent feature of LAM-associated populations is enrichment of transcriptional programs related to lipid uptake, lipid-droplet metabolism, lysosomal degradation, phagocytosis, and tissue clearance. At the protein and phenotypic levels, CD9+CD63+ macrophages show a higher neutral-lipid burden, a finding consistent with increased lipid accumulation but not, by itself, proof of superior lipid-processing capacity.^7,30 Enrichment of LIPA, CTSB, CTSL, and other lysosome-related molecules likewise supports a lysosomal program, whereas direct measurements of lipid degradation and metabolic flux in molecularly defined adipose LAMs remain more limited. Functional studies of adipose macrophage lysosomal activity provide additional support for a role in the clearance of lipid-stressed adipocytes.³¹

TREM2 is not only a marker of LAMs, but also a context-dependent regulator of macrophage lipid handling, phagocytosis, lysosomal function, and metabolic adaptation in lipid-rich tissue environments.^10,32 In the original adipose LAM study, TREM2 deficiency impaired the LAM-associated transcriptional program and altered adipose tissue metabolic homeostasis.⁷ However, subsequent studies using different TREM2-deficient models have reported partially divergent metabolic outcomes. Sharif and colleagues reported that TREM2 deficiency aggravated diet-induced insulin resistance and hepatic steatosis, whereas Winn and colleagues found that exon 2-mediated TREM2 deletion did not worsen whole-body metabolic function in diet-induced obese mice.^10,32 These findings indicate that TREM2 signaling is important for LAM-associated lipid handling and tissue remodeling, but its systemic metabolic consequences may depend on the genetic model, adipose depot, sex, diet duration, and disease stage. Accordingly, these studies support a proposed lipid-buffering role for LAM-associated macrophages in some experimental settings, rather than a uniform function across all models of obesity.

Because LAM-associated macrophages are enriched in crown-like-structure regions, their spatial distribution overlaps with sites of adipocyte death and debris accumulation. This spatial overlap is consistent with a potential role in efferocytic clearance, a process implicated in adipose tissue homeostasis and obesity-associated inflammation.³³ More directly, lipid-droplet efferocytosis studies demonstrate TREM2- and MS4A7-dependent regulation of post-clearance inflammatory signaling, although these experiments did not uniformly isolate molecularly and spatially defined adipose LAMs.³⁴

Evidence for LAMs in Human Adipose Tissue

Although many mechanistic studies of LAMs have been performed in mouse models of diet-induced obesity, human adipose tissue studies are essential for evaluating the translational relevance of this macrophage population. Single-cell analyses of human white adipose tissue have identified macrophage populations enriched for lipid-handling, lysosomal, phagocytic, and TREM2-associated gene programs.^35–37 These LAM-like populations commonly express markers such as TREM2, CD9, LPL, APOE, and lysosome-related genes, suggesting partial conservation of lipid-handling macrophage programs between mice and humans. Recent syntheses of adipose single-cell datasets and Human Cell Atlas consensus efforts further emphasize that cell annotation, adipose depot, species, and analytical workflow must be considered when comparing LAM-like macrophage states across studies.^38,39

Nevertheless, human evidence remains more limited than mouse evidence. Most available human studies are observational and cannot directly establish whether LAM-like macrophages are protective, pathogenic, or stage-dependent in obesity-related metabolic dysfunction. Therefore, human LAM-like macrophages should be interpreted as translationally relevant but functionally incompletely defined cell states. Future studies combining single-cell transcriptomics, spatial profiling, metabolic phenotyping, and functional validation in human adipose tissue will be necessary to determine their precise roles in obesity. For this reason, the term “LAM-like macrophages” may be more appropriate when discussing human adipose tissue unless functional equivalence with mouse LAMs has been experimentally validated.^35,36

Origin and Candidate Developmental Relationships of LAMs

Current single-cell, trajectory, and spatial studies are consistent with a candidate contribution of recruited monocytes to LAM-associated macrophage states in obese adipose tissue. Intermediate pre-LAM-like states express increasing levels of TREM2, LGALS3, APOE, LPL, and lysosome-related genes and are spatially associated with lipid-rich remodeling niches. However, trajectory inference and temporal co-occurrence do not establish ontogenetic ancestry.⁸

The relative contributions of recruited monocytes, locally proliferating macrophages, and tissue-resident macrophages remain unresolved and may vary with adipose depot, obesity stage, diet duration, sex, and local signals.^40,41 LAM development should therefore be presented as a candidate state-transition model rather than a proven linear differentiation pathway. Figure 2 illustrates this inferred model and should not be interpreted as fate-mapping evidence.

Figure 2.

Candidate origin, LAM-associated clearance features, and context-dependent roles of adipose LAMs. The dashed monocyte–pre-LAM-like–LAM sequence denotes an inferred state-transition model rather than a fate-mapped lineage. Solid relationships indicate lipid-handling and lysosomal features supported in adipose-LAM studies but do not establish enhanced metabolic flux. Efferocytosis is shown as a supportive relationship because molecularly and spatially defined adipose LAMs were not uniformly isolated in the relevant functional studies. The potential lipid-buffering and inflammation/remodeling outcomes represent context-dependent alternatives rather than a confirmed temporal progression. Solid lines indicate supported adipose-LAM features; dashed lines indicate supportive, inferred, or proposed relationships.

Abbreviations: CLS, crown-like structure; LAM, lipid-associated macrophage.

Functions of LAMs and Their Relationship with Obesity

The functional significance of LAM-associated macrophages is supported most directly by their obesity-associated expansion, spatial enrichment in crown-like structures, and selected genetic perturbation studies. TREM2-deficient models alter LAM-associated transcriptional programs and the accumulation of crown-like-structure-associated macrophages. However, most manipulations are not selective for molecularly defined adipose LAMs, and different TREM2-deficient models have produced divergent systemic metabolic outcomes.^10,32 These findings support a contribution of LAM-associated programs to lipid homeostasis and tissue remodeling, but do not establish a uniform protective function.

Spatial and time-course studies show that candidate pre-LAM-like and LAM-associated states colocalize with crown-like-structure remodeling as glucose intolerance develops. This association links LAM accrual to tissue remodeling but does not demonstrate that LAMs directly cause insulin resistance.⁸ Studies of macrophage lipid-flux remodeling further support an association between lipid handling and macrophage functional plasticity. LAMs may therefore provide one cellular context in which lipid metabolism and metaflammation intersect.⁴²

A context-dependent dual-role model has been proposed for LAM-associated macrophages. Under some conditions, lipid uptake, lysosomal processing, and efferocytic clearance may buffer local lipid stress. Under prolonged or severe metabolic stress, LAM-associated inflammatory and remodeling signals may coexist with insulin resistance.^30,34 Current evidence does not establish that individual LAMs undergo a fixed transition from an early protective state to a later pathogenic state. Candidate developmental relationships and stage-dependent functions should therefore be treated as working models (Figure 2).

The principal studies used to define adipose LAMs are summarized in Table 1. Evidence directness was evaluated according to whether adipose LAMs were specifically identified using molecular evidence together with protein, spatial, and/or functional criteria and whether causal perturbation was performed. Studies based on total adipose tissue macrophages, global genetic models, or cross-tissue transcriptomic integration were classified as supportive or comparative evidence. This classification reflects directness for adipose LAM biology rather than overall study quality (Table 1).

Table 1.

Evidence Hierarchy of Major Studies Defining Adipose Lipid-Associated Macrophages in Obesity

Study	Species/Model and Tissue	LAM Identification Method	Principal Markers or Features	Main Findings	Evidence Category and Principal Limitation
Jaitin et al, 20197	Diet-induced obese mice and human adipose tissue	Single-cell transcriptomic profiling, index sorting, tissue localization, and global Trem2 perturbation	TREM2, CD9, LPL, LIPA, CD36, CTSB, CTSL, and other lipid-handling and lysosomal genes	Identified an obesity-expanded macrophage population with a conserved TREM2-associated lipid-handling program. Global Trem2 deficiency suppressed the LAM-associated program and altered adipose and systemic metabolic phenotypes.	Category I: direct identification with functional perturbation. Global Trem2 deletion was not selective for adipose LAMs; therefore, systemic metabolic effects cannot be attributed exclusively to LAMs.
Stansbury et al, 20238	Male mice; epididymal white adipose tissue during early and chronic diet-induced obesity	Single-cell RNA sequencing, spatial transcriptomics, tissue imaging, and computational trajectory analysis	TREM2, LGALS3, CTSL, APOE, LPL, LRP1, and CLS-associated spatial localization	Identified macrophages transcriptionally intermediate between monocytes and mature LAMs and described them as candidate pre-LAM states. Pre-LAM and LAM-associated signals were spatially enriched in crown-like structure niches.	Category II: direct molecular and spatial identification without definitive lineage validation. Computational trajectory and cross-sectional time-point data support a candidate developmental model but do not prove a fixed monocyte-to-LAM lineage.
Sharif et al, 202110	Diet-induced obese mice; white adipose tissue and systemic metabolic assessment	Global Trem2 deficiency, bone-marrow transplantation, adipose tissue analysis, and metabolic phenotyping	TREM2-associated macrophage and adipose tissue responses	Trem2 deficiency aggravated insulin resistance and hepatic steatosis; however, bone-marrow transplantation experiments indicated that the systemic metabolic effects were not explained solely by TREM2 expression in immune cells.	Category III: supportive but non-LAM-specific genetic evidence. The model does not selectively perturb molecularly defined adipose LAMs and indicates substantial non-immune or tissue-intrinsic effects.
Xu et al, 202535	Mouse and human macrophages from multiple tissues and inflammatory conditions	Integrated analysis of published single-cell RNA-sequencing datasets	Conserved TREM2 and LPL expression together with tissue-restricted transcriptional programs	Identified shared and tissue-specific LAM-associated gene programs across species, tissues, and inflammatory settings, supporting partial conservation of a broader LAM transcriptional state.	Category III: comparative transcriptomic evidence. Cross-tissue transcriptional similarity does not establish equivalent function or adipose LAM-specific mechanisms.
Chini et al, 202630	Diet-induced obese mice; visceral white adipose tissue	Flow cytometry, tissue imaging, transcriptomic analysis, macrophage–adipocyte interaction studies, and myeloid-specific Cd9 deletion	CD9-positive LAM-associated macrophages, crown-like structures, profibrotic and extracellular-matrix programs	CD9 promoted macrophage–adipocyte interactions, crown-like structure organization, and fibrotic remodeling. Myeloid CD9 loss reduced adipose fibrosis and improved systemic metabolic outcomes.	Category I: direct LAM-associated functional evidence. CD9 is not unique to LAMs, and myeloid-specific deletion also affects other myeloid populations; the intervention was therefore not strictly LAM-selective.
Winn et al, 202232	Diet-induced obese mice; adipose tissue and whole-body metabolic assessment	Exon 2-mediated Trem2 deletion and metabolic phenotyping	TREM2-associated macrophage responses	In contrast to some other TREM2-deficient models, exon 2-mediated Trem2 deletion did not worsen whole-body metabolic function during diet-induced obesity.	Category III: supportive and model-dependent genetic evidence. The study did not selectively manipulate an experimentally defined LAM population and highlights the influence of genetic strategy and experimental context.
Reyes-Farias et al, 202536	Humans with severe obesity; subcutaneous and visceral adipose tissue	Transcriptomic analysis of adipose tissue-infiltrating CD11b-positive myeloid cells	TREM2, CD9, GPNMB, and CD68	LAM-associated marker expression was higher in subcutaneous than in visceral adipose tissue, demonstrating depot-dependent heterogeneity in human obesity.	Category III: supportive human adipose marker evidence. The study assessed CD11b-positive myeloid populations rather than a purified, spatially validated LAM population and did not provide causal functional testing.

Notes: Category I, direct adipose LAM identification accompanied by functional or genetic perturbation; Category II, direct molecular or spatial identification of adipose LAMs without selective causal validation; Category III, supportive or comparative evidence derived from total adipose tissue macrophages, global genetic models, broad myeloid populations, or cross-tissue transcriptomic analyses. These categories describe the directness of evidence for adipose LAMs rather than the overall methodological quality of each study. Transcriptomic enrichment indicates an associated molecular program but does not, by itself, establish enhanced functional capacity. Computational trajectory analysis identifies candidate developmental relationships but does not constitute definitive lineage evidence.

Abbreviations: CLS, crown-like structure; LAM, lipid-associated macrophage; SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue.

Immunometabolic Programs Associated with LAMs in Obesity

Lipid Metabolic Programs

In obese adipose tissue, persistent lipid overload is associated with macrophage lipid-metabolic remodeling, including changes in lipid uptake, intracellular storage, lysosomal processing, and cholesterol efflux. Lipid-handling and lysosomal programs are enriched in adipose LAM-associated populations, whereas ABCG1-dependent lipid partitioning has been demonstrated mainly in broader adipose tissue macrophage populations.⁴² However, many cited studies did not directly quantify metabolic flux in isolated, molecularly defined LAMs. The following pathways are therefore discussed as coordinated lipid-handling modules rather than as a fully established LAM-specific metabolic circuit.

At the level of fatty acid uptake and transport, one of the most prominent features of LAMs is the upregulation of lipid uptake and transport programs, in which CD36 and LPL serve as important molecular components.^7,8 CD36 functions as both a fatty acid transporter and a signaling receptor that integrates lipid uptake with immune cell metabolism and functional fate.⁴³ In adipose tissue, recent studies further suggest that CD36-dependent pathways can influence LAM accumulation and tissue remodeling under lipid-rich metabolic conditions.^44,45 For example, studies of obesity-associated remodeling in brown adipose tissue have shown that CD36-associated macrophage programs participate in the processing of adipocyte-derived lipid particles and contribute to changes in adipose tissue cellular identity.⁴⁵ These findings suggest that CD36-related lipid uptake may affect not only macrophage metabolic states, but also the surrounding adipose tissue microenvironment. Likewise, LPL hydrolyzes triglycerides in lipoproteins to release fatty acids, thereby providing lipid substrates for macrophages and coupling lipid accumulation to inflammatory responses.^42,46 Relevant studies further suggest that the ABCG1/LPL axis can regulate inflammatory thresholds by altering the partitioning of fatty acids between the plasma membrane and lipid droplets, indicating that both lipid uptake and intracellular lipid distribution directly influence the intensity of macrophage inflammatory responses.⁴²

LAM-associated transcriptional programs support coordinated lipid sequestration, lysosomal processing, and phagocytic activity. Lipid-droplet formation may buffer potentially toxic free lipids, whereas enrichment of LIPA and lysosome-related proteases is consistent with post-phagocytic lipid processing. These molecular features do not alone prove enhanced metabolic capacity. Functional studies in macrophages show that lipid-droplet efferocytosis and lysosomal activity can modify inflammatory responses, providing supportive—but not uniformly LAM-specific—evidence for the functional relevance of these pathways.^31,34 More recently, ADAR1 was found to be enriched in adipose LAMs following weight-loss interventions, while functional experiments in broader macrophage systems linked ADAR1 to efferocytosis, lysosomal processing, and lipid-buffering programs; these findings are mechanistically relevant but not wholly adipose-LAM-specific.⁴⁷

In the context of cholesterol and fatty acid pool remodeling, ABCG1-dependent pathways may also influence adipose tissue macrophage plasticity and insulin resistance by regulating lipid partitioning and inflammatory thresholds.⁴² Therefore, lipid storage, lysosomal degradation, and cholesterol/lipid efflux should be viewed as interconnected modules that shape the lipid-handling and immunoregulatory phenotype of LAMs.

Overall, lipid handling is a central feature of LAM-associated macrophage programs. Enrichment of lipid-uptake, lipid-droplet, lysosomal, and efflux pathways describes an adaptive molecular and phenotypic state in lipid-rich niches; whether this state produces net homeostatic benefit depends on context and requires direct flux and perturbation studies. These lipid-handling programs may interact with the bioenergetic and inflammatory pathways discussed below.

Energy Metabolic Remodeling

In addition to lipid metabolic reprogramming, alterations in energy metabolism represent another core feature of macrophage functional reprogramming in obese adipose tissue. According to the classical immunometabolic model, pro-inflammatory macrophages primarily rely on glycolysis, whereas macrophages associated with tissue repair or homeostasis depend more heavily on mitochondrial oxidative metabolism.^48,49 However, the microenvironment of obese adipose tissue is characterized by multiple concurrent metabolic stresses, including lipid overload, hypoxia, cell death, and tissue remodeling. Under these conditions, adipose tissue macrophages may adopt metabolically flexible states that utilize multiple energy sources rather than conforming to the traditional M1/M2 metabolic dichotomy; whether this flexibility is a defining property of molecularly defined adipose LAMs remains unresolved.⁴

First, enhanced glycolysis in obesity-associated ATMs has been directly supported by experimental evidence.^50,51 In diet-induced obesity, adipose tissue macrophages exhibit increased glycolytic activity and HIF-1α activation, which can sustain local and systemic IL-1β production and thereby contribute to obesity-associated inflammation. Saturated fatty acids, such as palmitate, may further promote a pseudohypoxic macrophage state characterized by HIF-1α stabilization, increased glycolytic flux, and inflammatory cytokine production.^51,52 These findings suggest that, under conditions of lipid overload and adipose tissue stress, glycolytic remodeling may support inflammatory mediator production and contribute to impaired insulin signaling.⁵¹

Second, in terms of oxidative metabolism, obesity-associated ATMs may also engage mitochondrial oxidative pathways to support lipid processing, phagocytosis, and lysosomal activity. This interpretation is consistent with the concept of metabolically activated macrophages, in which nutrient excess and lipid-rich cues induce a macrophage state distinct from classical M1 activation.^4,26 Such macrophages can retain inflammatory activity while also activating metabolic programs related to lipid handling and oxidative metabolism.^27,50 Therefore, rather than displaying a simple glycolytic or oxidative phenotype, obesity-associated ATMs and LAM-like populations are better viewed as metabolically flexible cells whose energy programs are shaped by lipid burden, phagocytic demand, and local tissue stress.^4,48 Recent adipose-tissue studies further implicate hypoxia-linked BNIP3-dependent mitophagy and mitochondrial calcium regulation in macrophage glycolytic reprogramming and inflammatory activation.^53,54

In addition, nutrient- and energy-sensing pathways, particularly mTOR and AMPK signaling, provide a general regulatory framework for understanding macrophage immunometabolic reprogramming in obesity.^49,55 mTOR signaling integrates nutrient availability, growth signals, and inflammatory cues to regulate macrophage activation, survival, and metabolic state, whereas AMPK functions as a cellular energy sensor that links energetic stress to mitochondrial metabolism and inflammatory restraint.^49,56 In obesity-associated adipose tissue, these pathways may influence ATM and LAM-like phenotypes by modulating the balance between glycolysis, oxidative metabolism, lipid handling, and cytokine production.^49,50 However, direct evidence defining how mTOR–AMPK crosstalk specifically controls LAM formation and function remains limited, and this mechanism should therefore be interpreted as a plausible regulatory framework rather than a fully established LAM-specific pathway.^4,56

Current evidence supports metabolic flexibility primarily in total adipose tissue macrophages and metabolically activated macrophage models. LAM-like populations may combine glycolytic and oxidative programs, but direct metabolic-flux measurements in molecularly defined adipose LAMs remain scarce. Energy-metabolic remodeling should therefore be presented as a supportive framework for LAM biology rather than as an established LAM-specific phenotype.

TREM2-Associated and Inflammatory Signaling

TREM2 is a prominent LAM-associated marker and a functional regulator in selected adipose-tissue models.⁷ Direct adipose-LAM evidence supports an association between TREM2 and the maintenance of lipid-handling, lysosomal, survival, and phagocytic programs. DAP12/TYROBP–SYK signaling is well established downstream of TREM2 in myeloid cells, whereas direct demonstration of the complete TREM2–DAP12–SYK–PI3K–AKT–mTOR cascade in molecularly defined adipose LAMs remains limited.⁵⁷

Myeloid PI3K activation can promote lipid-buffering macrophage phenotypes in obesity models, but these experiments did not selectively manipulate a molecularly defined LAM population.⁵⁸ The TREM2–PI3K–AKT–mTOR relationship should therefore be presented as a proposed LAM-associated regulatory framework rather than a proven linear LAM-specific pathway.

With respect to inflammatory signaling, NF-κB and JNK are classical signaling hubs that convert lipid stress into metaflammatory responses.^3,59,60 In obese adipose tissue, adipocyte hypertrophy, hypoxia, and cell death induce macrophages to secrete inflammatory mediators such as TNF-α and IL-1β.⁶¹ These mediators further activate JNK or IKKβ/NF-κB signaling, thereby disrupting insulin signaling and sustaining chronic inflammation.⁶² FcRn-dependent IgG accumulation has also been identified as an adipose immunometabolic signal that promotes macrophage activation and insulin resistance during obesity, although this mechanism has not been demonstrated to be LAM-specific.⁶³

At the level of inflammatory execution, the NLRP3 inflammasome serves as an important effector module that translates lipid danger signals into inflammatory responses.^64,65 Under obese conditions, fatty acids, cholesterol crystals, and lipid droplets can all act as danger-associated signals that activate the NLRP3 inflammasome and promote the maturation and release of IL-1β, thereby contributing to a state of chronic low-grade inflammation.^64,66 Studies have shown that although TREM2 is not essential for lipid droplet phagocytosis itself, it is required for suppressing NLRP3 inflammasome activation and pro-inflammatory cytokine production after lipid droplet clearance, suggesting that TREM2 plays an important regulatory role during the post-clearance phase of inflammatory resolution.³⁴

In addition, TREM2 signaling is also regulated by receptor shedding. Recent studies have shown that in obese adipose tissue, TREM2 can be cleaved by ADAM10/17 to generate soluble TREM2 (sTREM2), thereby reducing cell-surface TREM2 levels and impairing macrophage phagocytic function.⁶⁷ Adipocyte pyroptosis may promote TREM2 shedding through the STING–ADAM10/17 axis, whereas inhibition of this shedding process enhances macrophage phagocytic capacity and ameliorates obesity-associated inflammation and metabolic dysfunction.⁶⁷ These findings support TREM2 shedding as a regulator of TREM2-associated macrophage function in obese adipose tissue; whether it represents a LAM-specific limiting mechanism requires direct validation.

Taken together, current evidence is insufficient to place lipid uptake, bioenergetic remodeling, TREM2 signaling, and NF-κB/NLRP3 activation into a single linear LAM-specific pathway. The strongest adipose-LAM evidence concerns TREM2/CD9-associated identity, lipid-handling and lysosomal programs, and spatial enrichment in crown-like structures.^7,8,30,34 Evidence for glycolytic, mitochondrial, mTOR/AMPK, NF-κB, JNK, and NLRP3 regulation is derived largely from total adipose tissue macrophages or broader macrophage systems. Figure 3 therefore separates direct adipose-LAM evidence from supportive and extrapolative mechanisms.

Figure 3.

Evidence-graded framework for immunometabolic programs associated with adipose LAMs. Solid relationships summarize molecular or spatial features supported in adipose-LAM studies. TREM2/CD9 expression, LPL/LIPA enrichment, lipid-handling and lysosomal programs, and CLS localization are presented as parallel associated features rather than as a causal sequence. Dashed relationships denote supportive evidence from total adipose-tissue macrophages or indirect evidence from other macrophage systems. Glycolysis/HIF-1α and OXPHOS/AMPK–mTOR are therefore shown as supportive ATM frameworks, whereas the TREM2–DAP12–SYK–PI3K–AKT–mTOR and NF-κB/JNK/NLRP3 modules are not presented as a unified LAM-specific cascade.

Abbreviations: AKT, protein kinase B; AMPK, AMP-activated protein kinase; ATM, adipose-tissue macrophage; CD9, cluster of differentiation 9; CLS, crown-like structure; DAP12, DNAX-activating protein of 12 kDa; HIF-1α, hypoxia-inducible factor 1α; JNK, c-Jun N-terminal kinase; LAM, lipid-associated macrophage; LIPA, lysosomal acid lipase; LPL, lipoprotein lipase; mTOR, mechanistic target of rapamycin; NF-κB, nuclear factor kappa B; NLRP3, NOD-like receptor family pyrin domain-containing 3; OXPHOS, oxidative phosphorylation; PI3K, phosphoinositide 3-kinase; SYK, spleen tyrosine kinase; TREM2, triggering receptor expressed on myeloid cells 2.

A context-dependent model can nevertheless be proposed: LAM-associated programs may support lipid buffering and cellular-debris clearance, whereas prolonged metabolic stress may coexist with inflammatory and tissue-remodeling signals. This should be regarded as a working model rather than a confirmed transition of individual LAMs from an early protective state to a later pathogenic state. LAMs are therefore best described as a context-dependent macrophage population or state, not as an established key cellular node that uniformly links lipid overload to insulin resistance.

Therapeutic Strategies with Direct or Indirect Relevance to LAM Biology

With the growing understanding of the mechanisms underlying obesity-related metaflammation, therapeutic research has gradually shifted from simply suppressing inflammatory responses toward modulating the metabolic states of immune cells.^3,4 LAM-associated macrophages occupy lipid-rich inflammatory niches, making their lipid-handling, lysosomal, and inflammatory programs relevant to therapeutic investigation. Because no current intervention selectively targets molecularly defined adipose LAMs, the strategies reviewed below are classified by proximity to LAM biology: monocyte-recruitment blockade, modulation of LAM-associated molecular programs, broader macrophage or inflammatory reprogramming, and adipose-niche or systemic metabolic intervention.

Because most therapeutic studies have not selectively targeted molecularly defined adipose LAMs, the strategies below are classified according to their proximity to LAM biology. Interventions involving TREM2-associated signaling, lipid handling, lysosomal activity, or phagocytosis are mechanistically proximal to LAM-associated programs but are not LAM-selective. Approaches targeting monocyte recruitment, systemic inflammation, metabolic stress, weight loss, or adipose-tissue remodeling are indirect because they primarily reshape the niche in which LAM-associated states arise.

Regulation of Macrophage Recruitment

Recruited monocytes contribute to the increased macrophage burden in obese adipose tissue and may supply candidate precursors for some LAM-associated states.^18,19 Blocking monocyte recruitment has therefore been investigated as an upstream strategy, with the CCL2/CCR2 axis representing the most studied example. Under obese conditions, adipocytes and the local inflammatory microenvironment release a variety of chemokines that promote the continuous infiltration of circulating monocytes and their subsequent differentiation into adipose tissue macrophages, thereby contributing to crown-like structure formation and the progression of metaflammation. In theory, interruption of this recruitment process could reduce inflammatory cell accumulation and cytokine release, thereby improving the inflammatory microenvironment of adipose tissue.

At present, clinical translation of this strategy has mainly been explored in the field of nonalcoholic steatohepatitis (NASH).⁶⁸ The dual CCR2/CCR5 antagonist cenicriviroc demonstrated some antifibrotic potential in the phase 2b CENTAUR study, but failed to significantly improve histological fibrosis in the subsequent Phase 3 AURORA trial. These findings suggest that although blocking immune cell recruitment is biologically plausible, this approach alone may be insufficient to reverse the established immunometabolic reprogramming network in chronic metabolic disease.⁶⁸

Because chemokine networks are redundant, inhibition of a single recruitment axis may not fully suppress monocyte influx. Because recruitment blockade acts upstream of macrophage accumulation, it may indirectly influence LAM accrual but should not be described as direct reprogramming of established LAM-associated macrophages.¹⁸ Given the unresolved ontogeny of LAMs, monocyte-derived cells should be described as candidate precursors rather than as cells that definitively give rise to all LAMs.

Modulation of LAM-Associated Immunometabolic Pathways

TREM2 is a prominent LAM-associated marker and a functional regulator in selected obesity models. DAP12/TYROBP–SYK signaling is well established downstream of TREM2 in myeloid cells, whereas PI3K–AKT–mTOR involvement in adipose macrophages provides only supportive evidence for a proposed LAM-associated cascade.^57,58 The complete cascade has not been demonstrated as a unified therapeutic pathway in molecularly defined adipose LAMs.

TM4SF19-mediated regulation of lysosomal activity provides evidence that modifying adipose-macrophage lysosomal function can influence inflammation and metabolic outcomes.³¹ Because these experiments did not selectively manipulate LAMs, lysosomal enhancement should be described as a LAM-relevant candidate strategy rather than an established method for improving LAM function.

Accordingly, TREM2-associated and lysosomal interventions remain mechanistically proximal, preclinical strategies. No study currently demonstrates selective pharmacological reprogramming of adipose LAMs. The proposed early lipid-buffering and later inflammatory states should not yet be used as a treatment rule for when LAMs should be enhanced or suppressed.

Inflammation and the Tissue Microenvironment

In contrast to approaches that directly regulate macrophage immunometabolic programs, another category of intervention strategies acts primarily by improving adipose tissue inflammation and the systemic metabolic environment, thereby indirectly influencing the formation and functional state of LAMs. By suppressing inflammatory signaling, ameliorating insulin resistance, or modulating systemic energy metabolism, these strategies reshape the tissue microenvironment in which macrophages reside and, in turn, influence their immunometabolic phenotype.

The NLRP3 inflammasome links lipid stress and damage-associated signals to IL-1β production. Preclinical NLRP3 inhibition improves selected inflammatory and metabolic outcomes in obesity models, but these studies do not selectively measure or manipulate molecularly defined adipose LAMs, and clinical translation remains limited.⁶⁹ NLRP3 inhibition should therefore be classified as an indirect microenvironmental strategy rather than a LAM-targeted therapy.

PPAR-dependent transcriptional programs regulate macrophage lipid handling and inflammatory state. Human pioglitazone studies and adipose-tissue-macrophage-directed PPAR nanomedicine demonstrate adipose or macrophage metabolic effects, but neither selectively tests LAMs.^70,71 CD36- and SREBP-related findings likewise come mainly from broader adipose-macrophage or metabolic models. These approaches should therefore be classified as supportive or indirect LAM-relevant strategies, with tissue specificity remaining a major translational challenge.

At the level of energy metabolism, AMPK functions as a cellular energy sensor that links energetic stress to fatty acid oxidation, mitochondrial function, and inflammatory regulation.⁴⁹ However, direct evidence that pharmacological AMPK modulation selectively reprograms molecularly defined adipose LAMs remains limited. Systemically acting metabolic therapies are therefore discussed separately below according to their relevance to weight loss and the LAM microenvironment.

Beyond metabolic reprogramming-based approaches, depletion strategies directly targeting macrophage abundance have also been widely explored, among which CSF1R inhibitors are the most representative. However, obesity-related studies to date indicate that simply reducing macrophage numbers has clear limitations as a therapeutic strategy in metabolic disease. In high-fat diet mouse models, Merry and colleagues found that the CSF1R inhibitor pexidartinib (PLX3397) significantly reduced the number of adipose tissue macrophages, but failed to effectively improve glucose homeostasis or visceral adipose cytokine expression, suggesting that macrophage depletion alone is insufficient to reverse established insulin resistance and metabolic dysfunction. In addition, CSF1R inhibition may produce inconsistent tissue-specific metabolic effects. Bosch and colleagues reported that although PLX5622-mediated macrophage depletion improved hepatic insulin sensitivity, it also impaired glucose tolerance and was accompanied by elevated liver enzyme levels, indicating that macrophages in different tissues may exert distinct metabolic regulatory functions.^72,73

These studies show that macrophage depletion can produce tissue-dependent metabolic outcomes and does not reliably reverse insulin resistance. Because adipose and systemic macrophage compartments contain both homeostatic and inflammatory populations, depletion should be considered a broad and indirect strategy. Functional reprogramming may be conceptually preferable to indiscriminate depletion, but direct evidence for selective pharmacological reprogramming of molecularly defined adipose LAMs remains lacking.

Weight Loss and Pharmacotherapy: Implications for LAM Biology

Weight loss provides an important clinical context for evaluating whether obesity-associated macrophage programs are reversible. Recent work examining adipose tissue during obesity and subsequent weight loss has demonstrated selective remodeling of the adipose niche rather than a simple uniform reversal of all obesity-associated cellular changes.³⁷ This observation is particularly relevant to LAM biology because the microenvironmental signals associated with LAM accumulation—including lipid release, adipocyte stress, cell death, and crown-like structure formation—may change as adipose tissue mass and metabolic stress decline. Nevertheless, current evidence does not establish whether LAM populations disappear, return to a pre-obesity state, or persist as phenotypically altered macrophages after weight loss. Mouse multi-omics studies demonstrate that obesity-associated immune-cell phenotypes may persist during weight loss and weight regain. Recent human single-cell studies indicate that lifestyle- and bariatric surgery-induced weight loss produces stepwise, cell-type-specific remodeling of adipose tissue, including reductions in myeloid-cell abundance, rather than a uniform return to the pre-obesity state.^74–76 Nevertheless, obesity-associated molecular and epigenetic memory may persist after weight loss, although current evidence does not establish whether this persistence is specifically attributable to LAMs.⁷⁷

The relationship between weight loss and LAM remodeling should therefore be interpreted cautiously. Lifestyle intervention, metabolic surgery, and pharmacotherapy can simultaneously alter adipocyte size, lipid flux, insulin sensitivity, vascular function, stromal-cell activity, and systemic inflammation.^37,78 Any accompanying change in LAM-like macrophages may consequently reflect a broad remodeling of the adipose tissue niche rather than a direct intervention on LAMs. Longitudinal studies using paired adipose tissue samples before and after weight loss are needed to determine whether changes in LAM abundance or molecular programs precede, accompany, or follow metabolic improvement.

Metformin is a systemic metabolic therapy that may indirectly influence the adipose niche associated with LAM formation. Review-level evidence supports modest effects on body weight, insulin sensitivity, and adipose metabolic pathways, but direct effects on molecularly defined LAMs have not been demonstrated.⁷⁹ Any macrophage-related effects may arise from reduced systemic nutrient stress, altered glucose and lipid availability, or broader changes in adipose tissue inflammation rather than direct reprogramming of a molecularly defined LAM population. Beyond metformin, acarbose has been reported to suppress pro-inflammatory ATM activity in diet-induced obesity through both gut-dependent and direct macrophage pathways.⁸⁰

GLP-1-based and dual-incretin therapies provide clinically relevant perturbation models because semaglutide and tirzepatide produce substantial weight loss and broad metabolic improvement in people with obesity.^81–83 In preclinical obesity models, tirzepatide reduced pro-inflammatory adipose tissue macrophage activity and improved adipose inflammation and insulin resistance; however, this study assessed broader ATM populations rather than molecularly defined adipose LAMs.⁸⁴ These effects could reduce adipocyte hypertrophy, lipid overflow, cell stress, and inflammatory signaling, thereby indirectly modifying the microenvironment in which LAM-like macrophages accumulate. At present, however, direct evidence showing that GLP-1-based therapies selectively reduce, expand, or functionally reprogram human adipose LAM-like macrophages remains insufficient. It is also unclear whether any observed immune remodeling would result from direct receptor-mediated effects on macrophages or would occur secondarily to weight loss and improved systemic metabolism.

Accordingly, lifestyle-, surgery-, metformin-, and GLP-1-associated changes should currently be interpreted as systemic or adipose niche-level modulation rather than LAM-targeted therapy. Their immediate translational value lies in providing clinically available perturbation models through which LAM behavior can be studied during metabolic improvement. Future trials should incorporate paired adipose tissue sampling, transcriptomic and protein-level LAM identification, spatial assessment of crown-like structure niches, and comparison with weight-matched controls. Such designs would help distinguish direct pharmacological effects from changes caused by weight reduction and determine whether remodeling of LAM-associated programs is related to durable metabolic benefit.

Targeted Delivery and Translational Medicine

Because systemic macrophage-pathway interventions can have tissue-dependent and off-target effects, adipose-tissue- or macrophage-directed delivery is being explored as a way to improve local exposure and reduce systemic effects.^72,85,86

In recent years, macrophage-targeted nanodelivery systems have attracted increasing attention. For example, Mohaghegh and colleagues developed simvastatin-loaded PLGA nanoparticles (Sim-NPs) for local delivery to adipose tissue. Experimental studies showed that this formulation induced anti-inflammatory polarization, promoted adipose browning, and reduced body weight in high-fat diet-fed mice.⁸⁷ Whether crown-like-structure localization can be exploited for selective delivery to LAM-associated cells remains a testable hypothesis; the current simvastatin-loaded nanoparticle study did not establish LAM-specific delivery.

In addition, certain immunometabolism-related metabolites have also been considered to possess therapeutic potential. Yu and colleagues reported that itaconate counteracts high-fat diet-induced obesity through leptin-independent mechanisms in multiple mouse models, improves glucose and lipid metabolic phenotypes, and increases energy expenditure through thermogenesis-related pathways, suggesting that immunometabolism-related metabolites may indirectly reshape the adipose metabolic and inflammatory milieu.⁸⁸ At the same time, the development of adipose tissue organoid models has provided a new platform for evaluating the effects of immunometabolic interventions in human tissues, thereby helping to establish a more reliable translational bridge between basic research and clinical application.^89,90

For translation, adipose-tissue- or macrophage-targeted delivery should be evaluated for cell-type specificity, efficacy, and systemic safety. Crown-like-structure localization provides a possible targeting rationale, but selective delivery to molecularly defined adipose LAMs has not yet been demonstrated. The therapeutic evidence discussed in this review is summarized and classified in Table 2. Importantly, none of the currently available interventions can be considered an established selective LAM-targeted therapy. The strategies were therefore classified according to whether they modulate a LAM-associated pathway, broadly alter macrophage abundance or function, preferentially target adipose tissue macrophages, or act indirectly through systemic weight loss and metabolic improvement. This classification reflects the directness of evidence for LAM modulation rather than the general clinical efficacy of each intervention.

Table 2.

Evidence Classification and Translational Relevance of Therapeutic Strategies Potentially Affecting Lipid-Associated Macrophages in Obesity

Strategy or Intervention	Primary Target or Mechanism	Experimental or Clinical Context	Evidence Related to LAMs	Main Reported Outcome	Evidence Classification	Major Limitation
Modulation of TREM2 availability and receptor shedding67	STING–ADAM10/17-mediated TREM2 cleavage and cell-surface TREM2 availability	Preclinical obesity-associated adipose tissue and macrophage models	Directly examines a LAM-associated receptor and macrophage phagocytic function in obese adipose tissue	Inhibition of TREM2 shedding preserved cell-surface TREM2, enhanced macrophage phagocytic activity, and ameliorated adipose inflammation and metabolic dysfunction	LAM-associated pathway modulation; preclinical LAM-associated pathway modulation without selective targeting of molecularly defined adipose LAMs	The intervention is not selective for molecularly defined LAMs, and the evidence remains preclinical
Cenicriviroc68	CCR2/CCR5-mediated monocyte recruitment and inflammatory cell trafficking	Phase III clinical trial in nonalcoholic steatohepatitis	No direct assessment of adipose LAM abundance, identity, or function	The intervention did not demonstrate the expected efficacy for liver fibrosis in the Phase III trial	Systemic chemokine-pathway intervention with indirect relevance to macrophage recruitment	Conducted in liver disease rather than obesity-associated adipose LAMs; clinical outcome cannot be interpreted as evidence for or against LAM targeting
NLRP3 inflammasome inhibitors69	NLRP3-dependent inflammatory activation and IL-1β-related signaling	High-fat diet-induced obesity models	No selective identification or functional assessment of LAMs	Reduced diet-induced obesity-associated inflammation and improved systemic metabolic outcomes in animal models	Indirect macrophage and systemic inflammatory-pathway intervention	NLRP3 is expressed in multiple cell types and tissues; the reported effects cannot be attributed specifically to LAMs
PPAR-targeted nanomedicine in adipose tissue macrophages71	PPAR signaling and macrophage lipid metabolism through adipose tissue-targeted delivery	Preclinical obesity models using nanomedicine delivery	Targets adipose tissue macrophages but does not selectively identify or manipulate LAMs	Improved lipid metabolism and obesity-associated metabolic dysfunction	Adipose tissue macrophage-targeted intervention with potential relevance to LAM-associated metabolic programs	The delivery system is ATM-directed rather than LAM-selective, and evidence of direct LAM reprogramming was not provided
Metformin79	AMPK-associated energy sensing, systemic glucose regulation, and metabolic improvement	Obesity-related preclinical and clinical literature summarized in a scoping review	No direct evidence of selective LAM modulation	Associated with reduced body weight, improved insulin sensitivity, and remodeling of systemic and adipose metabolic states	Systemic metabolic intervention with indirect effects on the LAM microenvironment	Changes in macrophage or LAM states may be secondary to systemic metabolic improvement; direct LAM measurements are lacking
GLP-1-based therapies81–84	Appetite regulation, weight loss, glycemic control, and systemic metabolic improvement	Established clinical therapy for diabetes and obesity	Direct evidence that GLP-1-based therapies selectively alter human adipose LAMs is currently lacking	Produce clinically meaningful weight loss and metabolic improvement and may secondarily reduce adipose tissue inflammatory stress	Systemic weight-loss and pharmacotherapy intervention with indirect potential relevance to LAMs	The effects of treatment-induced weight loss cannot be separated from direct immune-cell effects without longitudinal adipose tissue studies
Pexidartinib or PLX339772	CSF1R-dependent reduction of macrophage abundance	High-fat diet-induced obesity in mice	Reduced total adipose tissue macrophages without selective LAM assessment	Substantially reduced tissue macrophage abundance but did not improve glucose homeostasis or visceral adipose cytokine expression	Broad macrophage-depletion strategy	Macrophage depletion is not LAM-specific and may remove both harmful and beneficial macrophage populations
PLX562273	CSF1R inhibition and macrophage depletion	Preclinical metabolic assessment in mice	No selective identification or manipulation of LAMs	Produced tissue-dependent metabolic effects, including improved hepatic insulin sensitivity but impaired glucose tolerance and increased liver enzymes	Broad macrophage-depletion strategy with mixed systemic effects	The findings demonstrate tissue heterogeneity and do not support indiscriminate macrophage depletion as a LAM-directed approach
Simvastatin-loaded polymeric nanoparticles87	Local delivery to inflammatory adipose tissue macrophages and promotion of adipose tissue browning	Preclinical obesity models	Targets inflammatory macrophages in adipose tissue but does not establish selective delivery to LAMs	Promoted local adipose tissue browning and improved obesity-related metabolic parameters	Adipose macrophage-targeted delivery strategy with indirect relevance to LAMs	Macrophage phenotype definitions differ from the operational definition of LAMs, and human translational evidence is unavailable
Itaconate-based intervention88	Metabolic signaling and activation of brown adipocyte thermogenesis	Diet-induced obesity models	No direct LAM assessment	Reduced diet-induced obesity primarily through enhanced brown adipocyte thermogenesis	Systemic or adipocyte-centered metabolic intervention rather than a LAM-directed strategy	The principal mechanism is not macrophage- or LAM-specific; its relevance to LAM biology is indirect

Notes: “LAM-associated pathway modulation” refers to interventions involving molecules strongly associated with LAM biology, but not selectively delivered to or restricted to molecularly defined LAMs. “Adipose tissue macrophage-targeted intervention” refers to strategies preferentially directed toward adipose macrophages without demonstrating LAM selectivity. “Indirect systemic intervention” refers to treatments that may alter the LAM microenvironment through weight loss, improved insulin sensitivity, reduced lipid overload, or systemic inflammatory regulation without directly measuring or targeting LAMs. No intervention listed in this table currently qualifies as an established selective LAM-targeted therapy.

Abbreviations: AMPK, AMP-activated protein kinase; ATM, adipose tissue macrophage; CCR, C–C chemokine receptor; CSF1R, colony-stimulating factor 1 receptor; GLP-1, glucagon-like peptide-1; LAM, lipid-associated macrophage; NLRP3, NOD-like receptor family pyrin domain-containing 3; PPAR, peroxisome proliferator-activated receptor; TREM2, triggering receptor expressed on myeloid cells 2.

Current Experimental Evidence and Translational Gaps

Current evidence for LAM biology in obesity comes mainly from three sources: in vitro macrophage models, in vivo animal studies, and human adipose tissue analyses.^8,34,91 In vitro studies provide mechanistic insight into how lipid overload, fatty acids, dying-cell signals, and lysosomal stress reshape macrophage lipid uptake, lipid droplet handling, phagocytosis, and inflammatory signaling.⁶⁴ Mechanistic studies in broader macrophage or adipose-macrophage systems have been used to examine TREM2-associated post-clearance regulation, CD36-mediated lipid uptake, lysosomal processing, glycolytic remodeling, and NLRP3-related inflammatory activation.⁵¹ However, in vitro systems cannot fully reproduce the spatial organization, adipocyte–macrophage interaction, crown-like structure formation, and chronic metabolic stress that characterize obese adipose tissue.^89,92

In vivo animal studies, particularly diet-induced obesity models, have provided the strongest spatial and selected functional evidence for obesity-associated LAM expansion and activity in adipose tissue.^8,30 These studies have shown that LAMs expand during obesity, are spatially associated with CLS-rich remodeling niches, and express transcriptional programs related to lipid handling, lysosomal function, and phagocytosis. TREM2-related studies further support the involvement of LAM-associated pathways in adipose tissue remodeling and metabolic homeostasis, although different TREM2-deficient models have reported partially divergent systemic metabolic outcomes. Therefore, animal studies support an important role for LAMs in obesity, but also indicate that their functions may depend on obesity stage, adipose depot, sex, diet duration, and disease context.^10,32

Human adipose tissue studies are essential for evaluating the translational relevance of LAMs. Single-cell and spatial analyses have identified LAM-like macrophage populations in human adipose tissue, characterized by the expression of TREM2, CD9, LPL, APOE, and lysosome-related genes.^91,92 These findings suggest partial conservation of lipid-handling macrophage programs between mice and humans. Nevertheless, most human evidence remains observational, and direct functional validation of human LAM-like macrophages is still limited.⁹¹ It remains unclear whether these cells mainly protect against lipotoxicity, contribute to chronic inflammation, or perform different roles at different stages of obesity.

From a therapeutic perspective, strategies discussed in relation to LAM biology are largely indirect. Interventions aimed at macrophage recruitment, TREM2-associated signaling, CD36-mediated lipid uptake, lysosomal function, bioenergetic remodeling, NLRP3-related inflammation, or targeted delivery may alter LAM-associated programs or their tissue niche, but few studies directly measured molecularly defined adipose LAMs. No established selective LAM-targeted therapy is currently available. Major translational gaps include the lack of reliable human biomarkers for monitoring LAM status, the difficulty of selectively targeting adipose tissue macrophages without disturbing systemic immune function, and the need to determine when LAMs should be enhanced, restrained, or functionally reprogrammed.

Therefore, future studies should move beyond simply asking whether LAMs are beneficial or harmful. Instead, they should define stage-specific LAM functions, clarify the differences between mouse and human LAM-like populations, and develop adipose tissue- or macrophage-targeted strategies that can modulate lipid handling and inflammatory resolution with minimal systemic adverse effects. Such work will be essential for determining whether LAM-directed immunometabolic therapy can be translated into practical interventions for obesity-related metabolic dysfunction. Overall, evidence is strongest for LAM existence, obesity-associated expansion, and spatial organization; supportive but incomplete for lipid handling, lysosomal processing, and inflammatory regulation; and limited for causal validation in human obesity and for LAM-selective therapy.^35,91

Conclusion and Perspectives

The most consistent evidence supports the obesity-associated expansion of adipose macrophage populations enriched in lipid-handling, lysosomal, and phagocytic programs and their preferential localization in crown-like structures and other lipid-rich remodeling niches. Multimodal studies support TREM2- and CD9-associated LAM populations in mice and partially conserved LAM-like populations in human adipose tissue. Selected perturbation studies link LAM-associated molecules and programs to lipid processing, tissue remodeling, and metabolic outcomes, although these interventions are not strictly selective for molecularly defined adipose LAMs.

Major uncertainties remain regarding LAM ontogeny, temporal stability, functional capacity, and net effects during obesity. Computational trajectories support candidate monocyte-to-pre-LAM-like relationships but do not establish a fixed lineage, and the proposed transition from adaptive lipid buffering to chronic inflammatory activity remains a context-dependent working model. Several bioenergetic and inflammatory pathways discussed in this review are supported mainly by total adipose tissue macrophages or other macrophage systems. Direct causal evidence in human LAM-like populations is limited, and no selective LAM-targeted therapy is currently available.

The priorities for translation are standardized multimodal definitions of human LAM-like populations; longitudinal and functional studies across adipose depots, sexes, and obesity stages; stronger fate-mapping or equivalent lineage evidence; clinically useful biomarkers; and delivery strategies capable of modifying detrimental LAM-associated functions without removing beneficial macrophage activities. In the near term, paired adipose tissue studies during successful weight loss, metabolic surgery, and GLP-1-based pharmacotherapy may be more clinically actionable than immediate development of a LAM-specific drug. By separating direct adipose-LAM evidence from supportive and extrapolative findings, this review provides a framework for testing these questions without overstating the maturity of the field.

Funding Statement

The authors declare that no funding was received for this study.

Abbreviations

LAMs, lipid-associated macrophages; ATMs, adipose tissue macrophages; CLSs, crown-like structures; TREM2, triggering receptor expressed on myeloid cells 2; CD36, cluster of differentiation 36; LPL, lipoprotein lipase; LIPA, lysosomal acid lipase; APOE, apolipoprotein E; DAMPs, damage-associated molecular patterns; ER, endoplasmic reticulum; scRNA-seq, single-cell RNA sequencing; NLRP3, NOD-like receptor family pyrin domain-containing 3; AMPK, AMP-activated protein kinase; mTOR, mechanistic target of rapamycin; NF-κB, nuclear factor kappa B; JNK, c-Jun N-terminal kinase; CCL2, C-C motif chemokine ligand 2; CCR2, C-C chemokine receptor type 2; PPAR, peroxisome proliferator-activated receptor; CSF1R, colony-stimulating factor 1 receptor.

Data Sharing Statement

No new data were generated or analyzed in this narrative review.

Author Contributions

Wen Zhang: Conceptualisation, Investigation, Writing, Review & Editing.

Bo Zhu: Visualization, Writing, Review & Editing.

Jianan Xu: Validation, Writing, Review & Editing.

Xian Jin: Investigation, Writing, Review & Editing.

Junfeng Cui: Conceptualisation, Supervision, Project Administration, Writing, Review & Editing.

All authors took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no competing interests.