Lipid metabolism in dendritic cell biology

SUMMARY

Dendritic cells (DCs) are innate immune cells that detect and process environmental signals and communicate them with T cells to bridge innate and adaptive immunity. Immune signals and microenvironmental cues shape the function of DC subsets in different contexts, which is associated with reprogramming of cellular metabolic pathways. In addition to integrating these extracellular cues to meet bioenergetic and biosynthetic demands, cellular metabolism interplays with immune signaling to shape DC-dependent immune responses. Emerging evidence indicates that lipid metabolism serves as a key regulator of DC responses. Here, we summarize the roles of fatty acid and cholesterol metabolism, as well as selective metabolites, in orchestrating the functions of DCs. Specifically, we highlight how different lipid metabolic programs, including de novo fatty acid synthesis, fatty acid β oxidation, lipid storage, and cholesterol efflux, influence DC function in different contexts. Further, we discuss how dysregulation of lipid metabolism shapes DC intracellular signaling and contributes to the impaired DC function in the tumor microenvironment. Finally, we conclude with a discussion on key future directions for the regulation of DC biology by lipid metabolism. Insights into the connections between lipid metabolism and DC functional specialization may facilitate development of new therapeutic strategies for human diseases.

1. INTRODUCTION

Dendritic cells (DCs), initially discovered by Ralph Steinman and Zanvil Cohn in 1973,¹ are professional antigen processing cells (APCs) that exist in all mammalian tissues and play a fundamental role in the initiation and regulation of adaptive immunity, as well as in innate immune response.^2,3 Based on their functional specialization, DCs are mainly classified into type 1 conventional DCs (cDC1s), type 2 conventional DCs (cDC2s), and plasmacytoid DCs (pDCs) in vivo.^2,3 cDCs are defined by their uniquely efficient ability to activate naïve T cells. In particular, cDC1s preferentially present antigens to CD8⁺ T lymphocytes, which activates the cytotoxic function of CD8⁺ T cells against intracellular pathogens and cancer.^2,3 In contrast, cDC2s primarily activate CD4⁺ T cells in response to extracellular pathogens, parasites, and allergens.^2,3 pDCs are well established as a key source of type I interferons during viral infections.⁴

In recent years, it has become increasingly appreciated that cellular metabolism serves as a key regulator of DC development, activation, maturation, and homeostasis.^5–9 In addition to altering immune signaling pathways that specify DC function, activation of pattern recognition receptors (PRRs) by immunogenic stimulants such as lipopolysaccharide (LPS) and polyinosinic-polycytidylic acid (poly I:C) also promotes reprogramming of cellular metabolism in DCs. Such metabolic reprogramming is necessary to provide DCs with energy (via ATP generation) to drive thermodynamically unfavorable reactions and to generate building blocks for macromolecule synthesis to meet the demands of functional activation and homeostatic maintenance. In addition, nutrients and metabolites shape signaling pathways that direct DC function. Ultimately, the dynamic and tightly regulated metabolic changes allow DCs to adapt to immunological and environmental cues, and thereby exert their function appropriately in different settings.

Lipid metabolism is a major metabolic pathway that is involved in many aspects of immune cell biology, and the relevance of this to macrophages, neutrophils, and T lymphocytes has been extensively covered in recent reviews.^10–12 How lipid metabolism orchestrates DC function is of emerging interest for tuning immune responses in infectious diseases and cancer. In this review, we describe how lipid metabolism and selective metabolites shape DC biology. Specifically, we will discuss the roles of fatty acid and cholesterol metabolism, as well as the associated metabolites, in regulating the functions of DCs and their subsets. We will also summarize how the tumor microenvironment (TME) induces the tolerization of DCs by rewiring DC intracellular lipid metabolism. Furthermore, we will discuss the challenges and open questions in understanding DC lipid metabolism and aim to provide new insights on future research in this fast-moving area.

2. OVERVIEW OF DC SUBSETS USED IN IMMUNOMETABOLIC RESARCH

DCs are a diverse group including multiple DC subsets with distinctive functional specification.^2,3 Each of the DC subsets requires unique metabolic programs to exert their specialized functions appropriately.^13–16

2.1. Primary DCs in vivo

Under steady state, primary DCs, also known as natural DCs,^8,17 are mainly composed of cDC1s, cDC2s, and pDCs (Table 1). cDCs in mice and humans express CD11C and major histocompatibility complex class II (MHC-II) molecules.^2,3 In addition, cDC1s selectively express the chemokine XC receptor (XCR1) and the C-type lectin domain family 9 member A (CLEC9A), and depend on transcription factors, including interferon regulatory factor 8 (IRF8), DNA-binding protein inhibitor 2 (ID2), and basic leucine zipper transcriptional factor ATF-like 3 (BATF3), for their development.^2,3 Functionally, cDC1s mediate antigen cross-presentation, which involves the uptake of exogenous soluble or cell-associated antigens, processing antigens into peptides, loading peptides onto the MHC-I molecules, and their presentation to CD8⁺ T cells, thereby provoking cytotoxic CD8⁺ T cell responses.^2,3 The importance of cDC1s in antitumor and antiviral immune responses are underscored by the impaired antitumor and antiviral immunity in cDC1-deficient Batf3^−/− mice and other in vivo models of cDC1 depletion.^18,19

Table 1.

Summary of lipid metabolism in dendritic cell subsets in vivo.

DC subsets	Developmental origin	Presence in vivo	Surface marker	Functional specialization	Lipid metabolism in biology
cDC1s	CDP	Lymphoid resident, peripheral tissue, and blood	Mouse: CD11C+MHC-II+XCR1+CD24+CLEC9A+CD11B−	Cross-presentation of exogenous antigens on MHC-I, and activate TH1 and CD8+ T cells; clearance of intracellular pathogens and tumors in vivo	Increased lipid droplet upon DC maturation;61 functional relevance unknown
			Human: CD11C+HLA-DR+XCR1+CLEC9A+CD141+CD1C−		Requirement of lipid droplet for maturation64
cDC2s			Mouse: CD11C+MHC-II+CD172A+XCR1−CD8α−CD24−	Presentation of antigen on MHC-II and prime both immunogenic CD4+ TH and regulatory T cells; against parasites, allergens, extracellular bacteria, and fungi in vivo	Increased lipid droplet accumulation upon maturation;61 requirement of 7α,25-HC for positioning in marginal zone bridging channel;97,98 requirement of 7α,27-HC for T cell zone migration99
			Human: CD11C+HLA-DR+ CD1C+CD172A+CD11B+CD141−		Not reported
pDCs	CDP and CLP		Mouse: SIGLECH+LY6C+B220+PDAC-1+	Production of type I interferon; against viral infections in vivo	Not reported
			Human: CD123+CD303+CD2+		Not reported

cDC1s, type 1 conventional dendritic cells; cDC2s, type 2 conventional dendritic cells; pDCs, plasmacytoid dendritic cells; CDP, common dendritic cell precursor; CLP, common lymphoid progenitor; 7α−25-HC, 7,25-dihydroxycholesterol; 7α−27-HC, 7,27-dihydroxycholesterol; T_H, CD4⁺ T helper cell.

cDC2 development requires cDC2 lineage-defining transcription factors including interferon regulatory factor 4 (IRF4), zinc finger E-box-binding homeobox 2 (ZEB2), and RELB.^2,3. Recently, additional cDC2 subtypes have been characterized, including one that selectively expresses ESAM and depends on neurogenic locus notch homolog protein 2 (NOTCH2) for development.²⁰ cDC2s are predominantly involved in presentation of MHC-II-restricted antigens to CD4⁺ T cells, and control T_H2-associated immune responses against parasites and allergens, as well as T_H17-associated immune responses to defend from extracellular bacteria and fungi.^2,3 cDC2s are commonly distinguished from cDC1s by their preferential expression of CD11B and CD172A (Table 1). However, these markers are also shared with other CD11C⁺MHC-II⁺ myeloid cells like macrophages and monocyte-derived DCs (moDCs) and cannot be used to identify cDC2s specifically in inflamed contexts (e.g., in the TME). Whereas proteins uniquely expressed by cDC2s have not yet been identified, a recent study has developed a cDC2-deficient mouse model via mutation of Zeb2 enhancer,²¹ which may help advance the study of cDC2 biology in the future.

pDCs express surface markers B220, SIGLECH, and CD317/BST2/PDCA-1, and depend on IRF8, transcription factor 4 (TCF4) and runt-related transcription factor 1 (RUNX1) for lineage commitment.⁴ pDCs have no or limited antigen presentation function compared to cDCs, and instead produce type I interferons in response to certain infections. This function is attributed to the high expression of selective PRRs such as toll-like receptor 7 (TLR7) and TLR9 that recognize single-stranded RNA and CpG dinucleotides, respectively.⁴ Studies using pDC-deficient Itgax^CreTcf4^flox/flox mice and CLEC4C-based DTR model of pDC depletion have uncovered the critical role of pDCs in controlling viral infections and contributing to autoimmunity.^22,23

2.2. DCs from in vitro-derived culture systems

In addition to these DC subsets in vivo, studies using DCs generated via in vitro culture systems have contributed to our understanding of DC biology (Table 2). In particular, bone marrow derived DCs (BMDCs) are the most widely used DCs differentiated in vitro, which can be generated by culturing bone marrow cells with different cytokines such as FMS-like tyrosine kinase 3 ligand (FLT3L) for FLT3L BMDCs,²⁴ granulocyte macrophage colony stimulating factor (GM-CSF) with or without interleukin 4 (IL-4) for GM-CSF BMDCs,^25–27 and GM-CSF plus FLT3L for iCD103⁺ BMDCs.²⁸ Furthermore, culturing human peripheral blood mononuclear cells with GM-CSF and IL-4 can induce the differentiation of moDCs from human blood monocytes.²⁹ It is noteworthy that BMDCs are heterogeneous; for example, FLT3L BMDCs include cDC1-like, cDC2-like, and pDC-like cells, while GM-CSF BMDCs comprise not only DC-like cells but also macrophage-like cells.^30,31

Table 2.

Summary of lipid metabolism in dendritic cells derived from in vitro culture systems.

	Developmental origin	Culture conditions	Composition	Lipid metabolism in biology
GM-CSF BMDCs	Bone marrow derived progenitors	GM-CSF with or without IL-4 for 5−7 days	DC-like, monocyte-like, and macrophage-like	Impairment of activation by PUFAs and SCFAs;34, 39–41,47 improvement of maturation by SFAs;34,37 upregulation of FAS but decrease in FAO upon activation;51,52,56,66 FAO blockade increases TH17 cell polarization;71 requirement of membrane cholesterol and intracellular lipid droplet for antigen presentation;61,75,79 enhancement of pro-inflammatory cytokine secretion by intracellular cholesterol85,90
FLT3L BMDCs		FLT3L for 9 days	cDC1-like, cDC2-like and pDC-like	SCFAs induce tolerogenic phenotype;44 requirement of FAO for cDC2-like cell development;65 FAS and FAO are indispensable for pDC-like cell activation57
iCD103+ BMDCs		FLT3L plus GM-CSF for 15–17 days	cDC1-like	FAS is dispensable for T cell priming60
moDCs (human)	Peripheral blood monocytes	GM-CSF plus IL-4 for 6−7 days	DC-like	PUFAs and SCFAs cause tolerogenic phenotype;36,38,45–47 FAO blockade increases maturation;67,68 membrane cholesterol is required for antigen presentation;79,80 intracellular cholesterol accumulation and 22R-HC inhibit migration;88,94 bile acids induce immunosuppression of human moDCs101

cDC1, type 1 conventional dendritic cell; cDC2, type 2 conventional dendritic cell; pDC, plasmacytoid dendritic cell; GM-CSF, granulocyte–macrophage colony-stimulating factor; FLT3L, FMS-like tyrosine kinase 3 ligand; IL-4, interleukin 4; GM-CSF BMDCs, DC from bone marrow cultures with GM-CSF or GM-CSF plus IL-4; iCD103⁺ BMDCs, DCs from bone marrow cultures with FLT3L plus GM-CSF; moDCs, monocyte derived DCs; PUFAs, polyunsaturated fatty acids; SCFAs, short chain fatty acids; SFAs, saturated fatty acids; FAS, fatty acid synthesis; FAO, fatty acid β oxidation; 22R-HC, 22R-hydroxycholesterol.

3. FATTY ACID METABOLISM

Cellular fatty acid content is regulated at multiple levels and varies based on the metabolic demands. For example, intracellular fatty acid levels can be altered when extracellular fatty acids are transported into cells by membrane transporters. Further, intracellular fatty acid composition is actively regulated via multiple pathways, including de novo fatty acid synthesis (FAS) from citrate-derived acetyl-CoA, lipolysis of triglycerides (TAGs, which form lipid droplets) by neutral hydrolases in the cytoplasm, or lipophagy of lipid droplets in lysosomes.³² These intracellular fatty acids are further mobilized to support membrane synthesis, catabolized by the fatty acid β oxidation (FAO) pathway in mitochondria, or esterified with glycerol to form TAGs and stored in lipid droplets (Figure 1a),³² highlighting the complex and dynamic regulation of fatty acid homeostasis in cells. We discuss the functional consequences of altering fatty acid content in DC biology below.

Figure 1.

The role of fatty acid metabolism in dendritic cell biology. a, Schematic of fatty acid metabolism in DCs. Sterol regulatory element-binding protein 1 (SREBP1) is the master transcription factor for fatty acid metabolism. SREBP1 promotes the expression of lipid transporters, including CD36, MSR1, and LDLR, to facilitate fatty acid uptake. Further, SREBP1 enhances transcriptional induction of de novo fatty acid synthesis (FAS) enzymes, such as acetyl-CoA carboxylase 1 (ACC1) and fatty acid synthase (FASN). FAS is initiated by ATP citrate synthase (ACLY), which converts tricarboxylic acid (TCA) cycle-derived citrate to acetyl-CoA. Then, the rate-limiting enzyme ACC1 (highlighted in red) catalyzes the conversion of acetyl-CoA to malonyl-CoA. FASN catalyzes additional steps in FAS to generate palmitate via multiple reactions. Fatty acids can enter the mitochondrial matrix for fatty acid β oxidation (FAO) through carnitine palmitoyltransferase (CPT) 1- and CPT2-dependent transport or be catalyzed by diacylglycerol O-acyltransferase (DGAT) and stored in lipid droplets, which can be broken down by lipolysis or lipophagy. b, Extracellular fatty acids regulate DC function. Short-chain fatty acids (SCFAs) and poly-unsaturated fatty acids (PUFAs) serve as negative regulators of DC activation, whereas saturated fatty acids (SFAs) facilitate DC maturation. These lipids may exert their bioactivity via the interaction with receptors (e.g., omega-3 fatty acid and its receptor GPR120) or by being transported into cells. How these lipids interplay with the signaling pathways to orchestrate DC maturation and functional activation is largely unknown and is marked by a dotted line; pMHC, peptide-MHC complex. c, Intracellular fatty acid metabolism in DC maturation. Activation of toll-like receptors (TLRs) drives reprogramming of fatty acid metabolism, in which de novo FAS and lipid droplet accumulation increase and FAO potential decreases to support DC maturation. d, DCs from the TME or treated with tumor-derived factors display increased FAO, FAS, and lipid droplet accumulation that drive DC defects in priming CD8⁺ T cell responses. Figure was generated using BioRender.

3.1. Extracellular fatty acids

Uptake of fatty acids from the environment is one of the major approaches to maintain the intracellular lipid pool in DCs. Long chain fatty acids (LCFAs), medium chain fatty acids (MCFAs), and short chain fatty acids (SCFAs) are transported into cells by LCFA transporters such as CD36, MSR-1, LDLR, MCFA transporters like GPR40 and GPR84, and SCFA transporters including SLC5A8, GPR109A, and GPR43.^32,33

The connections between extracellular fatty acids and DC functions were first established in DCs from in vitro culture systems.^34,35 Specifically, treatment with the saturated MCFA, lauric acid, increases expression of co-stimulatory ligands, including CD40, CD80, and CD86, as well as MHC-II in GM-CSF BMDCs. Lauric acid also increases LPS-induced production of inflammatory cytokines (IL-12p70 and IL-6) in a TLR4-dependent manner.³⁴ Similar as the effects of lauric acid, the common saturated LCFA, palmitic acid (PA), enhances TLR-mediated innate activation by inhibiting hexokinase, a key glycolytic enzyme, in GSF-BMDCs. PA-induced inhibition of glycolytic activity leads to disrupted mitochondrial fitness and increased generation of mitochondrial reactive oxygen species (mtROS). These alterations in turn cause an exacerbated unfolded protein response (UPR) and consequently induce DC inflammatory responses.³⁶ In contrast to the effects of saturated fatty acids, the n-3 polyunsaturated fatty acid (PUFA), docosahexaenoic acid, has been shown to inhibit LPS-induced DC maturation.³⁴ The inhibitory effects of PUFAs on DC activation are further validated by independent studies using human moDCs,^37,38 GM-CSF BMDCs,^39–41 and primary splenic DCs (Figure 1b).⁴² Although PUFAs are the natural ligands for peroxisome proliferator-activated receptors (PPARs) that are important regulators of lipid metabolism, both PPARγ-dependent and independent effects of PUFAs on DC activation have been reported.^38,39,43 Moreover, microbiota-derived SCFA butyrate enhances the capability of FLT3L-BMDCs to induce T_reg cell differentiation by inhibiting histone acetylation within DCs.⁴⁴ In line with this, butyrate and another SCFA, propionate, inhibit human moDC maturation.^45,46 Further, butyrate induces DC tolerization and alleviates the progression of models of inflammatory bowel diseases (IBDs) in vivo.⁴⁷ Additional studies are warranted to establish the mechanisms by which extracellular fatty acids alter DC functions, especially the molecular mechanisms underlying the interplay between these lipids and the immune signaling pathways.

Excessive uptake of fatty acids can result in lipid accumulation in cells, with overaccumulation of lipids being associated with DC dysfunction. For example, primary splenic DCs or GM-CSF BMDCs with higher content of TAGs are less potent activators of T cell proliferation compared to the lower lipid-bearing DCs.⁴⁸ Mechanistically, lipid accumulation in DCs dampens their ability to process and present antigens without affecting co-stimulatory ligand or pro-inflammatory cytokine expression.^48,49 Accordingly, blocking fatty acid uptake in GM-CSF BMDCs reduces lipid levels and restores DC function for activating allogeneic T cells,⁴⁸ suggesting that excessive lipid uptake leads to DC dysfunction. In agreement with these findings, other independent groups have reported that accumulation of fatty acids in GM-CSF BMDCs and in ovarian cancer-associated DCs is associated with impaired DC function in inducing T cell activation.^50–53 In contrast, hepatic DCs with high intracellular lipids more potently activate pro-inflammatory T cells, natural killer T cell (NKT) cells, and natural killer (NK) cells compared to those with low intracellular lipids, but the latter are more potent inducers of regulatory T cell (T_reg)-mediated tolerance.⁵⁴ Altogether, the contrasting effects of intracellular fatty acid accumulation on DCs from different settings suggest that environmental education may enable DCs to develop a unique pattern in lipid requirements.

Overall, these data highlight that extracellular fatty acids play an important role in the functional regulation of DC (Figure 1b). However, the discrete effects of different lipid species (e.g., fatty acids of various carbon chain lengths, and whether they are saturated or unsaturated) on the functions of DC subsets or DCs from different contexts are only emerging, and much remains to be explored. In addition, the cellular compartments where lipids exert their bioactive functions remain unclear; for instance, lipids may influence DC function by interacting with membrane lipid receptors such as free fatty acid receptor 4, also known as GPR120,⁵⁵ or entering cells via transporters (Figure 1b). Further, the mechanisms underlying how fatty acids interplay with the signaling pathways that define DC subset specialization or functions also require further investigation.

3.2. De novo fatty acid synthesis

In addition to uptake of fatty acids from the environment, DCs also synthesize fatty acids from nonlipid precursors, including glucose or glutamine upon their conversion into citrate through the process of glycolysis or glutaminolysis, respectively.³² De novo fatty acid synthesis (FAS) is an anabolic process by which citrate is transported by inner mitochondrial transmembrane protein SLC25A1 from the mitochondrial matrix to the cytosol, where it is converted to the building block acetyl-CoA by ATP-citrate synthase (ACLY).³² The rate-limiting enzyme acetyl-CoA carboxylase 1 (ACC1) then catalyzes the conversion of acetyl-CoA to malonyl-CoA. Malonyl-CoA is then committed to FAS through fatty acid synthase (FASN), which sequentially elongates the nascent fatty acid chain by adding two carbon units at each step in an NADPH-dependent manner, ultimately producing the 16-carbon fatty acid palmitate.³² The transcription factor sterol regulatory element-binding protein 1 (SREBP1) promotes transcriptional activation of a number of genes participating in fatty acid synthesis, including Acaca (encodes for ACC1) and Fasn (encodes for FASN), as well as certain lipid transporters (Figure 1a).³²

As noted above, DC maturation is associated with increased ability of antigen processing and presentation (signal 1), expression of cell surface proteins including co-stimulatory ligands (signal 2), and secretion of cytokines and chemokines (signal 3) (Figure 1b).⁸ The expansion of endoplasmic reticulum (ER) and Golgi networks is necessary for transporting the peptide MHC complex (pMHC), co-stimulatory molecules, and chemokine receptors to the plasma membrane, and for the secretion of cytokines and chemokines. De novo FAS provides the building block acyl-CoA to support ER and Golgi network remodeling. Accordingly, maturation of GM-CSF BMDCs is associated with increased FAS.⁵⁶ More importantly, inhibition of FAS via deletion of Slc25a1 (encodes for SLC25A1) or suppression of ACC1 or FASN activity using pharmacological inhibitor TOFA (ACC inhibitor) or C75 (FASN inhibitor) blocks LPS-induced maturation of GM-CSF BMDCs, as revealed by decreased expression of CD86 and CD40 and secretion of IL-6, IL-12p70, and TNFα.⁵⁶ Blockade of FAS by TOFA also impairs the immunogenicity of liver DCs, associated with the impaired CD4⁺ T cell or NK cell activation.⁵⁴ Moreover, FAS is also required for CpG-induced pDC activation, as revealed by downregulation of type I interferons and IL-6, as well as CD86 expression, upon TOFA treatment (Figure 1c).⁵⁷ Together, these studies highlight that de novo FAS can promote DC maturation to support their function in certain contexts.

Although the above observations suggest a positive role for de novo FAS in promoting DC maturation and function, inhibition of FAS can also boost DC function under certain contexts. For example, either C75 or TOFA treatment enhances antigen processing by GM-CSF BMDCs and increases their capacity to activate antigen-specific CD4⁺ and CD8⁺ T cells by elevating ER stress.⁵⁸ In addition, FAS inhibition decreases tumor explant supernatant (TES)-induced lipid accumulation in mouse GM-CSF BMDCs, and restores the ability of these cells to activate T cells.⁴⁸ Interestingly, ablation of ER stress response factor X-box-binding protein 1 (XBP1, encoded by Xbp1) inhibits FAS and subsequently downregulates lipid levels in ovarian tumor-associated DCs, thereby restoring their immunostimulatory activity in supporting antitumor T cell responses.⁵⁰ These data suggest that the TME can induce DC functional defects in activating antitumor T cell responses by increasing intracellular FAS (Figure 1d). More recently, FAS was reported to inhibit DC antigen presentation by altering DC epigenetic state. Specifically, hyperactivated de novo FAS induced by deletion of tuberous sclerosis complex subunit 1 (TSC1) deprives the availability of acetyl-CoA for histone acetylation, which in turn reduces MHC-I molecule expression and the capability of DCs in priming CD8⁺ T cells.⁵⁹ Recently, the view that de novo FAS is dispensable for DC maturation and function has also emerged. For instance, iCD103⁺ BMDCs deficient in Acaca have equivalent capability in T cell priming and pro-inflammatory cytokine expression compared to wild-type cells.⁶⁰ Further, mice bearing Acaca or Acacb (encodes for ACC2)-deficient DCs retain the capacity to clear Mycobacterium bovis infection.⁶⁰ The precise reasons for these inconsistencies remain unclear, but there may be different requirements for de novo FAS in DC subtypes or in DCs under various contexts. Further, there may be discrete effects in FAS blockade between pharmacological inhibitors, where the dose and timing of treatment can have additional effects, and genetic models. Of note, more precise analysis of primary DCs in different contexts in vivo are warranted, as the majority of studies to date have examined the impact of de novo FAS on BMDC populations in vitro.

3.3. Lipid droplets

Intracellular fatty acids can be stored as lipid droplets, also known as lipid bodies, in which fatty acids are esterified by diacylglycerol acyltransferase (DGAT) with glycerol and incorporated into lipid droplets. In contrast to DGAT-mediated lipid storage, adipose triglyceride ligase (ATGL) mobilizes lipid droplets and initiates their lipolysis in the cytosol, whereas lipophagy sequesters lipid droplets and targets them into lysosomes for turnover (Figure 1a).³²

The first study that linked lipid droplets and DC function revealed that lipid droplets are essential for DCs in mediating T cell priming.⁶¹ Specifically, GM-CSF BMDC activation is associated with increased expression of interferon gamma-induced GTPase (IGTP, encoded by Igtb) and lipid droplet accumulation. GM-CSF BMDCs, as well as primary splenic cDC1s or cDC2s, lacking Igtp fail to accumulate lipid droplets upon interferon-γ or poly I:C treatment. Functionally, Igtp-deficient GM-CSF BMDCs have selective defects in antigen cross-presentation, whereas MHC-II-restricted antigen presentation is still intact in these cells in vitro. Igtp depletion in GM-CSF BMDCs also compromises their abilities of antigen cross-presentation and activation of antigen-specific CD8⁺ T cells. Mechanistically, IGTP, which is localized to the ER and lipid droplets, interacts with the lipid droplet-associated protein ADFP, also known as Perilipin-2, and serves as a negative regulator of phagosome maturation, thereby preventing phagocytosed proteins from undergoing lysosomal degradation.⁶¹ Accordingly, Xanthohumol, a DGAT inhibitor, decreases the lipid droplet content in GM-CSF BMDCs and suppresses antigen cross-presentation.⁶¹ The immunogenic roles of lipid droplets in DCs are also validated in monocytic CD11B⁺ DCs,⁶² and human BDCA3⁺/CD141⁺ cDC1s (Figure 1c).⁶³ However, another study suggests an immune inhibitory function of lipid droplets in DCs. Specifically, lipidomics analysis of nontreated and TES-treated GM-CSF BMDCs revealed the selective accumulation of oxidatively truncated lipids in GM-CSF BMDCs upon TES treatment.⁶⁴ The combined use of imaging, biochemical approaches, and computational simulation uncovered that DCs with accumulated oxidized lipids in lipid droplets are defective in stimulating CD8⁺ T cell responses (Figure 1d),⁶⁴ suggesting that tumor-derived factors impair DC function by altering the composition and content of lipid droplets. Altogether, these data highlight important but also complex roles of lipid droplets in shaping DC functions. It is worth noting that our knowledge of lipid droplets and DC biology is largely derived from GM-CSF BMDCs treated with pharmacological inhibitors. Future studies are required to explore functional importance of lipid droplets in primary DCs in vivo, and genetic dependences of lipid droplet regulators such as DATG1/2, ATGL and lipophagy-related genes. Further, it will be important to address the mechanisms by which lipid droplets interact with the intracellular processes such as DC-derived signals 1, 2, and 3 for the modulation of adaptive immunity.

3.4. Fatty acid β oxidation

In addition to storage in lipid droplets, free fatty acids can also enter a catabolic pathway known as fatty acid β oxidation (FAO), which encompasses a cyclical series of reactions that break down two carbon units at a time in the mitochondrial matrix, leading to generation of acetyl-CoA, NADH, and FADH₂.³² These FAO-derived metabolic intermediates ultimately enter the tricarboxylic acid (TCA) cycle and electron transport chain to facilitate the generation of ATP. The PPAR family transcription factors orchestrate FAO by transcriptional activation of key FAO-associated genes including Cpt1a and Cpt2 (Figure 1a). The transmembrane proteins CPT1 and CPT2 transport LCFAs across the outer and inner mitochondrial membranes, where they are respectively localized, to facilitate FAO of LCFAs.³² In contrast, MCFAs and SCFAs can diffuse into the mitochondria passively without transporters and undergo FAO largely independently of CPT1 or CPT2 (Figure 1a).

Beyond regulating DCs through lipid anabolism, fatty acids may also affect DC biology through FAO-mediated catabolic pathways. Primary mouse DC progenitors in the bone marrow are dependent on nutrient transporters and glucose uptake for their proliferation upon FLT3L stimulation in vitro.⁶⁵ FLT3L drives the differentiation of common DC progenitors to DC subsets including cDC1s and cDC2s, as well as pDCs, allowing us to evaluate DC differentiation in vitro (Table 2).²⁴ Of note, treatment with etomoxir, a CPT1 inhibitor, favors cDC2 over cDC1 differentiation without affecting pDC differentiation,⁶⁵ suggesting that FAO may contribute to the appropriate development of DC subsets, although the mechanisms remain undefined. Several studies have examined how FAO is regulated in DC subsets, and shown that AMPK activity generally favors mitochondrial FAO over cytosolic FAS. Such a decrease in FAO in the absence of Prkaa1 (encodes for AMPK subunit alpha 1) may partly account for the decreased differentiation of cDC1s.⁶⁵ Further, TLR stimulation in cultured GM-CSF BMDCs reduces AMPK activation and FAO, and accordingly, targeting of Prkaa1 using shRNAs results in upregulation of LPS-induced expression of CD86 and IL-12p40 on GM-CSF BMDC (Figure 1c).⁶⁶ Thus, AMPK activation is associated with increased FAO and reduced BMDC maturation.

The above studies suggest that FAO likely promotes the DC tolerogenic phenotype. Consistent with this notion, inhibiting FAO attenuates the tolerogenic phenotype in primary human moDCs and an immortalized cDC1-like cell line.^67,68 Accordingly, inhibition of FAO by targeting Cpt1a using shRNAs in GM-CSF BMDCs rectifies the defects in activating CD8⁺ T cell proliferation, with such defects being attributed to elevated WNT5A induced β-catenin activity and PPARγ-dependent FAO;⁵¹ the β-catenin–PPARγ axis is also reported to induce cDC1 and cDC2 tolerization in visceral adipose tissue.⁶⁹ In addition, abrogation of FAO by etomoxir or inhibition of FAO by PPARα inhibitor GW6471 in GM-CSF BMDCs also leads to an increased capacity in priming CD8⁺ T cells (Figure 1d).⁵² Moreover, primary splenic DCs isolated from mice fed a high-fat diet have elevated FAO, and primary splenic DCs from obese mice display defects in presenting antigens to CD8⁺ T cells, which can be reversed by etomoxir treatment.⁷⁰ Mechanistically, FAO-derived ROS contributes to the reduced capacity of antigen presentation in primary splenic DCs. Antioxidants such as vitamin E or N-acetyl-l-cysteine treatment mitigate this FAO-driven accumulation of cellular ROS and consequently improve the antigen presentation capacity of primary splenic DCs.⁷⁰ Thus, some of the functional defects mediated by treatment with certain exogenous fatty acids as noted above could be attributed to increased FAO, although this requires further investigation, especially in vivo. Further, it will be important to investigate the mechanism by which FAO regulates DC functions, for example the crosstalk with the cellular energy deprivation sensor AMPK and epigenetic pathways.

Interestingly, FAO may have discrete roles in DC subsets. Indeed, contrary to the aforementioned observations in GM-CSF BMDCs, CpG-stimulated pDCs generated from FLT3L BMDCs cultures have enhanced mitochondrial pyruvate import and FAO that fuels elevated OXPHOS.⁵⁷ This effect is attributed to increased autocrine type I interferon signaling that promotes FAO via acting downstream of PPARα.⁵⁷ Importantly, suppression of FAO by etomoxir inhibits CpG-induced pro-inflammatory cytokine interferon-α, IL-6, and TNFα production, as well as co-stimulatory molecule CD86 expression, rather than affecting cell viability.⁵⁷ Additionally, reduction of Cpt1a levels by shRNA impairs pDC activation upon CpG treatment.⁵⁷ Furthermore, in a mouse model of allergic asthma, FAO has been shown to promote DC function. Specifically, depletion of mechanistic target of rapamycin (encoded by Mtor) induces FAO in lung DCs with increased production of pro-inflammatory cytokine IL-23 to shew T_H17 polarization. Etomoxir treatment mitigates the allergic airway inflammation in mice with Mtor-deficient DCs, associated with reduced activation of inflammatory lung DCs and also reduced T_H17 polarization.⁷¹ In agreement with these findings, elevated FAO is also observed in activated human moDCs.⁷² Therefore, FAO has differential effects in DC subsets and DCs from different contexts, which may reflect the metabolic heterogeneity and requirements of different DCs, and also metabolic flexibility of DCs for energy generation by switching their carbon sources to adapt to environmental changes and exert context-dependent functions. Future studies are warranted to address the functional importance and mechanistic basis of FAO in DC biology, including the use of genetic models to probe DC function in vivo.

4. CHOLESTEROL AND ITS METABOLITES

Like other sterols, cholesterol is largely a hydrophobic lipid. It is predominantly localized to cell membranes, where it interacts with adjacent lipids to regulate the rigidity, fluidity, and permeability of the lipid bilayer.⁷³ Beyond binding to lipids, cholesterol also interacts with sterol transporters and multiple transmembrane proteins, which can influence their conformational state. Moreover, cholesterol is a precursor for many biological molecules, including oxygenated derivatives of cholesterol and bile acids.⁷³ Below, we discuss how cholesterol and its derivatives impact DC functions.

4.1. Cholesterol

In DCs, cholesterol can be taken up from the blood in the form of low-density lipoprotein (LDL) particles that transport liver-synthesized cholesterol, or can be synthesized de novo.⁷³ In the cytosol, cholesterol synthesis is initiated downstream of the TCA-derived citrate and catalyzed by the rate-limiting enzymes 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR) and squalene monooxygenase (SM). Then, a series of over 20 reactions occur to promote the biosynthesis of cholesterol (Figure 2a).⁷³ SREBP2 transcriptionally regulates the expression of many genes involved in cholesterol biosynthesis, while the liver X receptors (LXRs) and PPARs oppose SREBP2-dependent cholesterol biosynthesis and instead facilitate cholesterol efflux by transcriptional induction of efflux transporters, including ATP-binding cassette subfamily A member 1 (ABCA1) and the ATP-binding cassette subfamily G member 1 (ABCG1).⁷³ This extracellular cholesterol is then packaged by apolipoprotein protein (APO) into high-density lipoprotein (HDL), which can then enter the blood circulation. Also, excess cholesterol can be esterified to cholesteryl esters by acyl coenzyme A:cholesterol acyltransferase (ACAT) and be stored in lipid droplets (Figure 2a). Thus, cellular cholesterol composition is dynamically regulated by multiple processes.

Figure 2.

Major pathways of cholesterol metabolism in dendritic cell functional regulation. a, Cholesterol metabolism in DCs. Cholesterol is synthesized from HMG-CoA that is derived from citrate, via a series of reactions using 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR) and squalene monooxygenase (SM) as the rate-limiting enzymes (marked in red). Cholesterol can be esterified by acyl coenzyme A:cholesterol acyltransferase (ACAT) to cholesteryl ester and stored in lipid droplets, be converted to 25-hydroxycholesterol (25HC) by cholesterol-25-hydroxylase (CH25H), 7,25-dihydroxycholesterol (7α,25-HC) by CH25H and oxysterol 7α-hydroxylase (CYP7B1), or to 7,27-dihydroxycholesterol (7α,27-HC) by cholesterol 7α-monooxygenase (CYP7A1) and CYP7B1, or be removed from the cell via ATP binding cassette subfamily A member 1 (ABCB1)- or ATP binding cassette subfamily G member 1 (ABCG1)-mediated efflux. Increased demand for cholesterol drives the induction of sterol regulatory element binding transcription factor 2 (SREBP2), which promotes cholesterol biosynthesis and uptake. In contrast to SREBP2, liver X receptors (LXRs) and peroxisome proliferator-activated receptors (PPARs) facilitate cholesterol efflux by transcriptionally inducing expression of ABCA1 and ABCG1. b, Cholesterol in DC functional regulation. Cholesterol in plasma membrane organizes pMHC clustering and thereby promotes antigen presentation to T cells. Intracellular accumulation of cholesterol activates NLRP3 inflammasome and consequently results in immunostimulatory functions of DCs. c, Cholesterol-derived metabolite 22R-HC decreases CCR7 expression and impairs DC migration by binding its receptor LXRα. 7α,25-HC facilitates DC positioning to the marginal zone (MZ) bridging channel by binding to EBI2. Another EBI2 ligand 7α,27-HC is essential for DC to migrate to the T cell zone. Primary and secondary bile acids dampen DC maturation through the TGR5 and FXR pathways. The mechanisms by which these oxysterols influence DC maturation and migration are unclear. Figure was generated using BioRender.

As noted above, cholesterol is a key component of the cell membrane that determines fluidity and rigidity. It is well appreciated that the cholesterol-rich lipid domain, also known as a lipid raft, is important for T-cell receptor (TCR) clustering and signaling to promote T cell activation.⁷⁴ Similarly, like TCR in the lipid raft, the location of the pMHC-II complex in the cholesterol-rich microdomain has also been observed in different APCs,^75–77 suggesting that cholesterol in the plasma membrane may contribute to TCR–pMHC-II interactions. Moreover, cholesterol regulates DC antigen presentation, as removing membrane cholesterol by methyl-β-cyclodextrin (MCD) inhibits GM-CSF BMDCs to present endocytosed MHC-II-associated antigens to OT-II CD4⁺ T cells,⁷⁸ with similar results found in B cells.^75,76 The effects of MCD in DC–T cell interaction are also observed in other groups.^79–81 Importantly, the reduction of cell surface cholesterol in DCs impairs their function against promoting the clearance of Toxoplasma gondii and hepatitis B virus in vivo,^82,83 highlighting that cholesterol in the plasma membrane serves important roles in facilitating T cell priming and adaptive immune responses. Further, APOE (encoded by Apoe) is an important cholesterol and lipid carrier, Apoe-deficient primary splenic DCs accumulate more intracellular cholesterol. This cholesterol accumulation is associated with upregulation of surface TLR4 and clustering of MHC-II in lipid rafts, which can be rectified by MCD treatment.⁸⁴ These cellular phenotypes induced by Apoe deletion in DCs likely explain the increased CD4⁺ T cell activation in Apoe^−/− mice.⁸⁴ The dysfunction of CD4⁺ T cell priming is also observed in GM-CSF BMDCs derived from Apoe^−/− mice.⁸⁵ Of note, excess cholesterol levels in DCs play a causal role in the development of autoimmune diseases in mice lacking both Apoe and oxysterol receptor LXR-beta (Nr1h2/Lxrb).⁸⁶ Specifically, cholesterol accumulation in DCs stimulates the production of factors promoting B cell proliferation and enhances antigen presentation to T cells, thereby driving the expansion of autoreactive B cells and the consequent development of autoimmune disease.⁸⁶ Thus, cholesterol modulation impacts the functional capacity of DCs to induce lymphocyte activation in vitro and in vivo. It will be interesting to determine precisely how such an accumulation of cholesterol contributes to the altered DC function, which may involve the dynamic interactions between cholesterol and pMHC complex.

Beyond regulation of signal 1, additional studies have shown that cholesterol levels can influence DC migration and cytokine production. Indeed, increasing cholesterol levels using a high cholesterol diet or ablation of Apoe in mice dampens the migration of DCs to the lymph node, lowering cholesterol levels can partially reverse these migration defects.⁸⁷ The migration of human moDCs is also impaired upon inhibition of LXR and PPARγ, which both facilitate cholesterol efflux⁸⁸. In terms of cytokine production, accumulation of cholesterol in DCs deficient for Abca1 and Abcg1 enhances IL-23 production, which subsequently elevates inflammatory responses and disrupts the quiescence of hematopoietic stem and multipotent progenitor cells.⁸⁹ Mechanistically, cholesterol promotes the activation of the NLRP3 inflammasome to provoke this inflammatory phenotype in mice with Abca1 and Abcg1 co-deficient DCs.⁹⁰

In summary, cholesterol is an immunogenic stimulant for DCs (Figure. 2b). Although cholesterol serves as a positive regulator for surface pMHC-II clustering, the precise molecular mechanism underlying the unique effects of cholesterol on MHC-II- but not MHC-I-associated antigen presentation is still largely unknown. Of note, the role of cholesterol in regulating DC migration or DC inflammasome activation needs to be further examined in different DC subtypes. Comprehensive and detailed dissection of cholesterol in diverse DC groups would advance our understanding of cholesterol metabolism in DC biology.

4.2. Cholesterol-derived metabolites

Cholesterol can also give rise to various oxygenated sterols (oxysterol) through enzymatic and non-enzymatic routes.⁷³ The oxysterol-producing enzyme cholesterol 25-hydroxylase (CH25H) catalyzes the conversion of cholesterol to 25-hydroxycholesterol (25HC), and 25HC 7α-hydroxylase (CYP7B1) can further catalyze the conversion of 25HC to 7,25-dihyroxycholesterol (7α,25-HC). Cholesterol can also be catalyzed and converted to 7,27-dihyroxycholesterol (7α,27-HC) by 7α-monooxygenase (CYP7A1) and CYP7B1 (Figure 2a). Expression of CH25H is rapidly induced by agonists for TLR3 or TLR4, as well as by interferon treatment,⁹¹ resulting in the production of 25HC, a bioactive molecule that broadly inhibits viral entry into cells.⁹² The function of DCs in antiviral immunity may partially result from the autocrine/paracrine production of interferons that increase the expression of antiviral 25HC, downstream of TLR stimulation. In addition, oxysterols such as 22R-hydroxycholesterol (22R-HC) and 25HC are natural ligands for LXRs.⁹³ Interestingly, human moDCs treated with 22R-HC show impaired migration, with a decrease in chemokine receptor CCR7 expression and defect in T cell priming.⁹⁴ Further, overexpression of sulfotransferase 2B1b (SULT2B1b) that inactivates oxysterols by sulfurization in tumor cells, or ablation of Nr1h3/Lxra in DCs, restores DC CCR7 expression and migration,⁹⁴ suggesting an inhibitory role of the oxysterols–LXRα axis for DC migration. Another oxysterol, 7α−25HC, is a potent and natural ligand for chemotactic receptor Epstein Barr virus-induced gene 2 (EBI2, also known as GPR183).^95,96 Mice deficient in Ebi2, Ch25h or Cyp7b1b show reduced frequency and number of cDC2s. Further analysis uncovered that 7α,25-HC–EBI2 pathway is required for cDC2 positioning in marginal zone bridging channels, and this positioning is important for maintaining their homeostasis and for mounting T cell-dependent antibody responses against certain bloodborne antigens.^97,98 More recently, the same group found a second EBI2 ligand, 7α,27-HC, is required for positioning cDC2s in the spleen. Upregulated EBI2 and CCR7 expression in activated cDC2s guides their position to the B-T zone interface. Under stimulation with type I interferon, the levels of EBI2 ligand 7α,27-HC are elevated, causing activated cDC2s to disperse throughout the T cell zone and induce robust T cell responses.⁹⁹ Altogether, these data suggest that oxygenated sterols play important roles in DC biology, with much remaining to be discovered.

Some oxygenated cholesterol derivatives can be further metabolized into bile acids, which are hydroxylated, amphipathic steroids that are synthesized in the liver. Bile acids are abundant in the mammalian intestine and play an important role in emulsifying dietary lipids to facilitate their absorption.¹⁰⁰ Human moDCs cultured with bile acids display a tolerogenic phenotype in response to commensal bacterial antigens, as revealed by low levels of IL-12 and TNFα.¹⁰¹ This phenotype can be reversed by deletion of bile acid receptor G-protein coupled bile acid receptor 1 (GPBAR1), also known as TGR5, in DCs.¹⁰¹ Secondary bile acids that are modified by bacteria also have bioactive effects. For example, the secondary bile acid 3β-hydroxydeoxycholic acid (isoDCA) dampens DC immunostimulatory properties and improves their ability to induce T_reg cell generation in the periphery.¹⁰² Similarly, ablation of bile acid receptor Nr1h4, also known as farnesoid X receptor (FXR), in DCs enhances the peripheral generation of T_reg cells and promotes transcriptional changes in DCs that are akin to those induced by isoDCA. These results suggest that isoDCA signals via FXR inhibition in DCs to contribute to T_reg cell induction and promote local immune tolerance in the intestines.¹⁰² Like isoDCA, gut microbiota-related secondary bile acids suppress DC activation via TGR5 signaling, which alleviates the severity of experimental autoimmune uveitis.¹⁰³ Together, these results highlight the critical roles of cholesterol derivatives in shaping DC function (Figure 2c). However, our knowledge on cholesterol-derived metabolites in DCs remains limited, and additional insights on the effects on DC subsets and in vivo function could have a significant impact on our understanding of cholesterol metabolism and DC biology.

5. LIPID METABOLISM AND DC FUNCTION IN THE TME

The immunosuppressive TME contains high levels of lipids.³² Interestingly, tumor cells exhibit considerable plasticity and flexibility in their lipid metabolism, which benefits them to adapt to the lipid-rich TME and support their survival, proliferation, metastasis, and immune evasion.³² In contrast, the functions of DCs are largely compromised by the lipid-rich microenvironment. Accordingly, primary splenic DCs from tumor-bearing mice have increased expression of the fatty acid transporter MSR1 (encoded by Msr1), which leads to the accumulation of cellular TAGs.⁴⁸ Msr1-deficient GM-CSF BMDCs do not accumulate lipids after their transfer to tumor-bearing recipients, associated with their acquisition of a more activated phenotype with improved priming of T cells.⁴⁸ Further, in the absence of Msr1, GM-CSF BMDCs are resistant to tumor-derived factors that impair their ability to promote T cell priming.⁴⁸ Interestingly, the intracellular accumulation of oxidized neutral lipids, including TAGs, cholesterol esters, and fatty acids, but not nonoxidized lipids, impair antigen cross-presentation in DCs by reducing the surface expression of pMHC-I complexes.⁴⁹ Mechanistically, oxidized lipids located within lipid droplets promote the sequestration of chaperone heat shock protein 70 (HSP70) in lipid droplets, thereby accelerating antigen degradation in lysosomes.⁶⁴ Thus, the accumulation of selective intracellular lipids appears to impair DC function and antitumor immunity, associated in part by antigen degradation that limits their capacity to provoke T cell activation.

In addition to the extracellular lipids, several studies have shown that de novo FAS also contributes to DC dysfunction in the TME. Oxidized lipids from the TME promote constitutive ER stress response via XBP1 activation, which subsequently enhances the de novo FAS pathway to increase intracellular lipid levels in DCs from orthotopic ovarian tumor model.⁵⁰ Limiting ER stress by deletion of Xbp1 decreases lipid abundance in tumor DCs and enhances their ability to prime CD8⁺ T cells.⁵⁰ Beyond the genetic perturbation of FAS, inhibiting FAS using TOFA alleviates the ability of tumor-derived factors to induce DC dysfunction and restores their ability to prime T cells.^104,105 These results together support that FAS serves as a negative regulator for DC-mediated antitumor immunity.

As described above, intracellular fatty acids can either be esterified and absorbed into the lipid droplet or enter FAO for break-down in the mitochondria. Studies show that inhibition of DGAT impedes lipid droplet synthesis and reverses the DC dysfunction elicited by tumor-conditioned medium, thereby enhancing the therapeutic efficacy of DC vaccines against tumors.¹⁰⁶ Further, tumor-derived WNT5A activates β-catenin that subsequently triggers the PPARγ pathway, which enforces FAO in DCs and induces DC tolerization.⁵¹ Mechanistically, increased FAO in DCs leads to the overproduction of indoleamine 2,3-dioxgenase-1 (IDO), which in turn suppresses production of IL-6 and IL-12.⁵¹ Further, depleting Cpt1a via siRNA or inhibition of CPT1 activity by etomoxir increases the ability of GM-CSF BMDCs to activate T cell priming, and augments DC-based therapeutic effects against tumors.⁵¹ Another independent study also shows that inhibition of FAO strengthens the antitumor function of DCs,⁵² highlighting that FAO, presumably of tumor-derived lipids, impedes the antitumor function of DCs. Additionally, cholesterol metabolism has also been reported to disrupt tumor DC function. Tumor-derived factors dampen human moDC migration by downregulation of surface CCR7 expression, and also inhibit DC immunogenic activity for T cell activation.⁹⁴ Inactivation of oxysterols in tumor cells or Nr1h3/Lxra depletion in DCs upregulates CCR7 expression, restores migration, and alleviates the immune inhibitory effects of the tumor-conditioned medium on human moDCs,⁹⁴ suggesting that cholesterol-derived metabolites from the tumor impose negative effects on DC function via the LXRα pathway. Altogether, these results suggest that tumor cells shape DC lipid metabolism, thereby contributing to DC tolerization (Figure 3).

Figure. 3.

The tumor microenvironment (TME) shifts DC lipid metabolism to induce dysfunction of DCs. Tumor-derived WNT5A activates β-catenin and peroxisome proliferator-activated receptor gamma (PPARγ) that increases FAO. Reactive oxygen species (ROS) in tumors triggers ER stress by inducing x-box-binding protein 1 (XBP1) activation, which promotes an upregulation of FAS and lipid droplet formation. Excessive lipid uptake can fuel FAO and facilitates lipid droplet accumulation in DCs, which in turn sequesters heat shock protein 70 (HSP70). In addition, tumor cells-derived oxysterols impair DC function by binding to its ligand LXR. These collective alterations decrease the immunogenicity of tumor DCs by decreasing antigen presentation, inflammatory cytokine production, and migration. TME also induces DC tolerance by an upregulation of indoleamine 2,3-dioxgenase-1 (IDO). Altogether, these alterations ultimately result in the dysfunction of DCs in stimulating CD8⁺ T cell antitumor immunity. Figure was generated using BioRender.

Much remains to be explored regarding the precise roles of FAO, de novo FAS, lipid droplets, free fatty acids, or lipid-related metabolic intermediates on DC biology, including the underlying mechanisms and their impacts on the functions of specific DC subtypes. Also, whether different sources of lipids (e.g., exogenous lipids versus de novo synthesized lipids) exert unique effects on tumor DC subtypes remains unclear. A better understanding of lipid metabolism in tumor DC biology could manifest legitimate opportunities for treating tumors through DC-based immunotherapies.

6. CONCLUDING REMARKS AND PERSPECTIVES

As we reviewed here, extensive studies have investigated the close links between lipid metabolism and DC development and their functional properties. However, we are only beginning to understand the role of lipids in DC biology, especially in vivo. Many fundamental questions remain open, including the role of lipid metabolism and lipid metabolites in regulating functional specialization of DC subsets. For example, how do lipids interplay with signaling pathways (i.e., those that induce signals 1, 2, and 3) to specify functions of DC subtypes? Do lipids from different cellular sources or compartments have distinct signaling roles? Further, the context-dependent regulation of DC biology by lipid metabolism awaits further investigation, especially in settings such as tumor progression, pathogen infections, and inflammatory diseases, including non-alcoholic fatty liver disease and IBDs. Moreover, whether and how metabolic plasticity of lipid metabolism allows DCs to adapt to environmental changes (e.g., the availability of various nutrients) and sustain their immunogenic activity in different contexts are also an interesting question to explore.

Of note, most of the current knowledge on lipid metabolism in DC biology has been obtained using GM-CSF BMDCs, which, in addition to DC-like populations, also contain significant proportions of macrophage-like cells.^30,31 Although this DC culture model has advanced our understanding of the differing metabolic profiles of immature DCs and activated DCs, these cells do not fully recapitulate the biological function of primary DC subsets in vivo. In addition, the understanding of the roles of lipid metabolism in DCs are frequently based on the application of pharmacological inhibitors. Caution needs to be taken with these results because inhibitors could have off-target effects, as was recently reported with etomoxir and C75.^107–109 Further, many studies focused on DC lipid metabolism have been performed with in vitro culture systems, and limited information is available on DC regulation in vivo functionally and mechanistically. Thus, studying the consequences of manipulating lipid metabolic regulators in DC in vivo rather than in vitro should be of high relevance, as the microenvironment is crucial for cellular metabolism and immune function. To circumvent the disadvantage of drug compounds and the limitation of in vitro studies, genetic silencing of lipid regulators in DCs using Cre-expressing mouse lines (e.g., XCR1-Cre for cDC1-specific genetic perturbation^110,111) could allow for the investigation of DC subsets with lipid metabolic impairment in vivo. Additionally, applications of systems-level approaches such as metabolomics, proteomics, ATAC-seq, single cell RNA-seq, and other single cell-based approaches (e.g., SCENITH and MetFlow^112,113) will help us to dissect the mechanisms by which lipid metabolism orchestrates DC biology. Further, CRISPR–Cas9-based genetic screening can be applied to interrogate the causative effects of lipid metabolism and DC biology on a large scale.¹¹⁴ The integrative use of combinatorial or complementary approaches (e.g., pharmacological inhibition, genetic perturbation, and systems biology approaches) will uncover key lipid metabolism-related mechanisms that regulate DC function in diverse contexts and facilitate development of new therapeutic strategies.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.