Macrophages are innate immune cells that populate all tissues and have a number of homeostatic roles (e.g., removing dead cells and cellular debris, recycling iron). One of their main immune functions is to recognize and phagocytose pathogens (1). Following recognition, they also produce cytokines that can recruit and induce differentiation of monocytes and T cells. Additionally, they play an important role in tissue repair once the immune response has been terminated (1).

Macrophages can respond to a variety of molecular cues through activation of Toll-Like Receptors (TLRs), and will secrete cytokines accordingly. Following activation, they will transition from a quiescent state to an activated one. Depending on the cues they receive, they can adopt different activated phenotypes. Based on in vitro experiments, activated macrophages have been historically divided into M1 pro-inflammatory or M2 anti-inflammatory macrophages (2). Stimulating macrophages with lipopolysaccharide (LPS) and interferon-γ (IFN-γ) triggers their differentiation into M1 macrophages whereas interleukin 4 (IL4) and IL13 promote an M2 phenotype (3, 4).

M1 and M2 macrophages display different metabolic properties. M1 macrophages have increased glucose uptake through the Hypoxia Inducible Factor 1α (HIF1α)-dependent upregulation of glucose transporters. This increased glucose uptake is essential to fuel glycolysis and thus produce ATP. M1 macrophages scarcely use mitochondrial respiration as they have targeted breaks in the tricarboxylic acid (TCA) cycle ( Figure 1A ) (5–7).

M2 macrophages, on the other hand, produce ATP mostly through mitochondrial respiration and have only limited glycolytic activity. Fueling the TCA cycle, M2 macrophages enhance their uptake of fatty acids. Fatty acid oxidation (FAO) provides Acetyl-CoA to a fully functional TCA cycle (8). TCA cycle-derived NADH is then used by the electron transport chain (ETC) to generate ATP through ATP synthase ( Figure 1A ) (7, 9).

In vivo, environmental factors that vary between different tissues or between homeostasis and disease are increasingly recognized as powerful dictators of macrophage metabolism. Macrophages in inflammatory conditions and inflammatory diseases are predominantly glycolytic. For example, macrophages in rheumatoid arthritis (RA) up-regulate the enolase enzyme, which stimulates production of pro-inflammatory cytokines, and show elevated succinate levels (10, 11). In diet-induced obesity, adipose tissue macrophages are hypermetabolic, up-regulating both glycolytic activity and mitochondrial respiration (12). Yet, these macrophages were long considered M1-like and are known to promote insulin resistance and metabolic syndrome (9). Such environmentally adapted profiles raise the question of precise directionality between macrophage metabolism and inflammatory profile.

Here, we use bone marrow-derived macrophages (BMDMs), with inhibitors of specific metabolic pathways to shed light on the direction of control between cellular metabolism and capacity to mount an inflammatory response. Using cytokine production as a functional read-out for macrophage phenotype, we show that impairing metabolic pathways drastically impacts macrophage phenotype. Some metabolic pathways have differential effects on cytokine secretion profile, suggesting that additional factors may directly influence macrophage phenotype in synergy with metabolic parameters.

Metabolism is tightly linked to macrophage polarization as M1 macrophages are predominantly glycolytic while M2 macrophages rely heavily on mitochondria. Here, we studied the impact of macrophage cellular metabolism on macrophage phenotype. We show that macrophage polarization depends on cellular metabolic activities and more particularly on some pivotal metabolic pathways like glycolysis. Interestingly, this dependency is functionally specific with respect to the cytokines affected and alters secretory capacity to a greater extent than transcription.

Additionally, we show that transcription of some cytokines was unexpectedly boosted by metabolic inhibition. It is possible that, following inhibition of a given pathway, metabolites used by this pathway are redirected towards other pathways in which activity is increased. This could be a mechanism of adaptation to the metabolically challenging environment. Such compensatory mechanisms have already been described in monocytes, in which IL6 production is quickly restored after depriving monocytes of ATP and glucose (13).

The impact of cellular metabolism on quiescent macrophage polarization differs between M1 and M2 macrophages. The impact of metabolic inhibition on M2 differentiation seems to be more limited than on M1 differentiation. This is of particular interest in some contexts like metabolic syndrome. In this disease, macrophages face changes in metabolic substrate availability, such as lipid overload. Moreover, in treated individuals, macrophages are also exposed to metabolically active therapeutic agents. Macrophages can accumulate in metabolic organs, adipose tissue and the liver, and exhibit an M1-like phenotype which is central to metabolic decline and the development of insulin resistance (9, 14, 15). Thus, our data suggest that changes in the metabolic environment of the cells could lead to changes in macrophage metabolic status and subsequently in macrophage phenotype. Alternatively, in therapeutic perspectives, altering availability of substrates or functioning of metabolic pathways can profoundly impact macrophage secretory profile. The question remains in defining at-risk groups and controlling cell-specific delivery of metabolically acting drugs. Future preclinical work will shed light on the feasibility of such approaches.

The M1/M2 classification has been challenged since it was first proposed. It poorly recapitulates the diversity of macrophage subpopulations found in vivo. Macrophages display great phenotypic and functional diversity depending on the tissues they populate, their ontogeny, or their states in health and disease (16). Such specificity could also arise in macrophage reliance on metabolic pathways or on adaptability of cellular metabolism. Additionally, recent studies have uncovered diversity in macrophage populations of a given tissue. For instance, 17 showed the existence of two subpopulations of macrophages that are conserved across several tissues and are localized in distinct tissular niches (17). Whether such subpopulations share common metabolic features or whether their different localizations also translate into different metabolic activities remains largely unknown. Differences in metabolic status between tissue-resident and monocyte-derived macrophages are still under investigation.

Future work could decipher the metabolic diversity of macrophages in vivo, adding mechanistic insight and rationale for redeployment of metabolically active therapeutics. Until recently, such investigations were considered technically challenging. However, techniques allowing single-cell resolution in analysis, e.g., SCENITH (18), are gaining accessibility and bring us closer to overcoming the technical challenges encountered when studying cellular heterogeneity across tissues, both in health and disease.

Macrophage polarization is known to play an active role in a number of diseases such as metabolic syndrome, as mentioned above, auto-immune diseases or cancer. While M1-like macrophages are important drivers of the progression of metabolic diseases, a phenotypic shift from M1-like to M2-like macrophages is known to favor cancer progression and eventually metastasis by dampening the anti-tumor immune response (19). Macrophage metabolism could therefore become an interesting therapeutic target in such diseases. Indeed, targeting relevant metabolic pathways could allow reprogramming of macrophage polarization and thus improve the disease. In addition, macrophages themselves are more and more considered as potential therapeutic agents, for example as chimeric antigen receptor (CAR)-macrophages targeting and phagocytosing tumor cells (20, 21). Combinations of such therapeutic macrophages with metabolic drugs aimed at modulating their phenotype could be powerful tools in a wide range of diseases.

Source data will be made available by the authors upon reasonable request.

RT, LO and FA designed the study. RT, LO and TE produced and analyzed experimental data. RT and FA wrote the manuscript. LO, TE and NV contributed in manuscript writing. All authors contributed to the article and approved the submitted version.

This work was supported by the French National Research Agency (Agence Nationale de la Recherche; ANR) ANR-JCJC grant for the MitoFLAME Project ANR-19-CE14-0005 to FA. RT was supported by a grant from the European Foundation for the Study of Diabetes (EFSD). LO was supported by Fondation de la Recherche Médicale (FDT202106013230).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.