Macrophage polarization in pathology

Abstract

Macrophages are cells of the innate immunity constituting the mononuclear phagocyte system and endowed with remarkable different roles essential for defense mechanisms, development of tissues, and homeostasis. They derive from hematopoietic precursors and since the early steps of fetal life populate peripheral tissues, a process continuing throughout adult life. Although present essentially in every organ/tissue, macrophages are more abundant in the gastro-intestinal tract, liver, spleen, upper airways, and brain. They have phagocytic and bactericidal activity and produce inflammatory cytokines that are important to drive adaptive immune responses. Macrophage functions are settled in response to microenvironmental signals, which drive the acquisition of polarized programs, whose extremes are simplified in the M1 and M2 dichotomy. Functional skewing of monocyte/macrophage polarization occurs in physiological conditions (e.g., ontogenesis and pregnancy), as well as in pathology (allergic and chronic inflammation, tissue repair, infection, and cancer) and is now considered a key determinant of disease development and/or regression. Here, we will review evidence supporting a dynamic skewing of macrophage functions in disease, which may provide a basis for macrophage-centered therapeutic strategies.

Introduction

Unlike other myeloid cells, such as neutrophils, characterized by a very short life span of few days, macrophages live longer, up to some months in tissues, hence representing key orchestrators of chronic inflammatory disorders. Heterogeneity and plasticity are hallmarks of mononuclear phagocytes, not paralleled by any other cell type in the body [1]. Indeed, macrophages can either protect tissue integrity and heal damaged tissues or under different contexts (e.g., chronic inflammation), be major destroyers of tissues, thanks to their production of inflammatory cytokines and proteolytic enzymes; they are not only essential cells in host defense against pathogens and primers of adaptive immune responses, but can also promote cancer cell proliferation and dissemination [2, 3]. This functional heterogeneity is only partially reflected by different phenotypes and morphological appearance. Rather, it is mostly due to different transcriptional programmes that are specifically activated by locally expressed signals [4–7].

Most tissue resident macrophages derive from the yolk sac or embryonic hematopoietic stem cells and self-maintain locally; however, at certain sites such as the gut and dermis, macrophages content depends on the continuous supply from bone marrow progenitors [8]. Blood circulating monocytes can be divided into two main subsets: patrolling monocytes (CX3CR1^high Ly6C⁻ in mouse; CX3CR1^high CD14^dim CD16⁺ in human), and inflammatory monocytes (CCR2^high Ly6C⁺ in mouse; CCR2^high CD14^high CD16⁻ in human). This latter population is recruited at inflammatory sites by specific chemokines, among which CCL2 [9, 10].

Under steady state conditions, it is noteworthy that an almost universal growth factor dictates their differentiation: Colony Stimulating Factor 1 (CSF-1), which recognizes a unique tyrosine kinase receptor (CSF-1R), although a second ligand (IL-34) to this receptor has been described in the last years [11].

Monocytes and macrophages also respond to the Granulocyte Macrophage-Colony Stimulating Factor 1 (GM-CSF) [12]. GM-CSF enhances monocyte migration across human blood–brain barrier endothelial cells and acts as an activator of macrophages and microglia [13]. In the lung, GM-CSF and CSF-1 control the local proliferation of alveolar macrophages [14], as well as their differentiation [15] and survival [16].

Signals present in the different microenvironments derived from pathogens, injured tissues, or activated effectors of the adaptive immunity, trigger distinct genetic programmes in the differentiating macrophages inducing different functional polarization states. The current paradigm describes two functional subsets: the M1 or classically activated and the M2 macrophages or alternatively activated, although it must be considered that these functional phenotypes represent the two extremes of a broad spectrum of differentiation states [17, 18].

Microbial stimuli (e.g., LPS) alone or in concert with cytokines (e.g., IFNγ, TNFα and GM-CSF), induce classically activated M1 macrophages. These are key effectors of the host response against intracellular bacteria and produce immunostimulatory cytokines (IL-12, IL-1β, TNFα, IL-6, and IL-23) and other effector molecules of the inflammatory response (e.g., reactive oxygen and nitrogen intermediates). In addition to M1 signals (e.g., IFNγ and LPS), engagement of additional receptors (such as phagocytic Fc receptors) is required to support a robust and prolonged production of ROS [19]. Moreover, differences exist between the human and murine systems in that NO production is peculiar of mouse macrophages and basically absent in the human counterpart. Hence, when appropriately activated, M1 macrophages have cytotoxic activity toward cancer cells and mediate resistance against tumors [18].

In contrast, IL-4 and IL-13, as well as IL-10 and TGFβ, induce alternatively activated M2 macrophages (IL-12^low/IL-1β^low/IL-10^high), which are poor antigen-presenting cells and suppressors of Th1 responses [18, 20]. M2 macrophages are key effectors for the resistance to parasites and in Th2 responses. Furthermore, while the persistence of M1-activated macrophages and their reactive products can induce tissue damage (e.g., chronic inflammatory conditions), the presence of M2-polarized macrophages is associated with the dampening of inflammation, scavenging of debris, angiogenesis, and tissue healing [21, 22]. However, in the context of tumors, these tissue trophic functions and the lack of cytotoxic ability promote cancer progression [23, 24]. Furthermore, M1 and M2 macrophages differently orchestrate the immune response by secreting distinct sets of chemokines: M1 cells recruit Th1 lymphocytes by producing CXCL9 and CXCL10, while M2 macrophages recruit Th2 lymphocytes, Tregs, as well as eosinophils and basophils with CCL17, CCL22, and CCL24 [25]. Thus, as we will discuss in this review, M1 and M2 macrophages are involved in distinct pathological settings.

As mentioned above, the binary distinction into M1 and M2 macrophages, albeit operationally useful, is an oversimplification of the complex biology of mononuclear phagocytes. Indeed, recent evidence has been provided that human macrophages exposed to a wide range of stimuli, such as TLR agonists, lipids, pro- and anti-inflammatory cytokines, activate distinct transcriptional programmes, suggesting a much wider spectrum of activation for macrophages than previously perceived [7]. Furthermore, exposure of M2 macrophages to M1 signals, or viceversa, can induce “re-polarization” of already differentiated macrophages, another evidence of their high functional plasticity. Of note, this re-education of macrophages is currently pursued for therapeutic purposes [26].

Macrophages in tissue homeostasis and repair

Macrophages are key players in the maintenance of tissue homeostasis and restoration of tissue integrity after injury [2]. Already, during embryonic development, macrophages participate in tissue and organ remodeling. Organs for which there is evidence of a macrophage role are bones, pancreas, mammary gland, and the nervous system [27]. Consistent with their function, fetal macrophages display an M2-phenotype [28] and contribute to the extracellular matrix (ECM) remodeling, proliferation of epithelial cells, angiogenesis, and tissue architecture. At the interface between fetus and mother in decidua and placenta, macrophages show an immunosuppressive phenotype, maintaining a tolerogenic milieu [29]. Due to this central role, macrophage tissue densities are tightly controlled. In the uterus of pregnant mouse, macrophage densities are locally controlled by M-CSF (CSF-1), whose in situ level of activity varied widely between uterine tissue layers. CSF-1 acts in part by inducing macrophage proliferation and in part by stimulating the extravasation of Ly6C^high monocytes that served as macrophage precursors [30].

Upon injury, mononuclear phagocytes that accumulate in damaged tissues mainly derive from circulating monocytes, although there is evidence in skin and muscle that resident macrophages contribute to the inflammatory infiltrate [31, 32]. In mice, Ly6C⁺CCR2⁺CX3CR1^low monocytes migrate to damaged tissues, where they contribute to the establishment of a pro-inflammatory response through the production of TNFα, IL-1, and nitric oxide. Monocytes expressing Ly6C⁻CCR2⁻CX3CR1^high are found in both inflamed and healthy tissues, and their function is usually associated with an anti-inflammatory and reparative process [9, 10]. In many tissues, the healing process progresses by overlapping phases of acute inflammation, resolving inflammation, proliferation, and remodeling [2, 22]. During these different phases of repair, macrophages undergo dynamic changes: M1-polarized macrophages are essential to initiate the inflammatory response after injury, while switching to an M2 phenotype seems to be fundamental for the resolution of inflammation and the regeneration of injured tissues [2, 22, 33] (Fig. 1). Dynamic changes from M1 to M2 phenotype of recruited monocytes have been observed in different model of ischemic heart and kidney diseases [34]. In humans, chronic venous ulcers represent a failure in resolving chronic inflammation and infiltrating macrophages fail to switch from an M1 to an M2 phenotype [35].

Fig. 1.

Role of macrophages in tissue repair. During the different phases of tissue repair, macrophages undergo dynamic changes, switching their phenotype from an M1- to an M2-phenotype. In the early inflammatory phase, M1 macrophages produce inflammatory cytokines and mediators, such as IL6, TNFα, IL1, and NO, stimulating innate immune cells (e.g., neutrophils) and defending host from pathogen colonization. During the resolution phase, macrophages initiate an M1 to M2 phenotype switch, gradually acquiring an anti-inflammatory phenotype, which includes down-regulation of inflammatory mediators, increased production of anti-inflammatory cytokines (e.g., TGFβ and IL10), phagocytosis of apoptotic neutrophils, and removal of damaged cells. In the proliferation phase, M2 macrophages produce a variety of growth factors, such as EGF, FGF, and VEGF, inducing the proliferation of various cell types involved in the healing process. Finally, in the remodeling phase macrophages contribute to the maturation of the regenerated tissue, reorganizing the extracellular matrix, the vasculature, and the scar tissue

The major role of macrophages in the inflammatory phase is to produce inflammatory cytokines such as IL-1, IL-6, and TNF-α which trigger innate immunity and defend the host from pathogen colonization. During the resolution phase, macrophages down-regulate the production of inflammatory mediators, increase the release of anti-inflammatory cytokines, such as TGF-β and IL-10 and actively remove damaged cells [36–38]. In a model of skeletal muscle injury, macrophages switched from Ly6C⁺CCR2⁺CX3CR1^low to Ly6C⁻CCR2⁻CX3CR1^high-resolving phenotype, as a consequence of phagocytosis of muscle cell debris [39]. In addition to clearing necrotic debris, macrophages phagocytose apoptotic neutrophils, a mechanism that contributes to the resolution of inflammation and progression toward the proliferative phase of healing [40].

During the proliferating phase, macrophages produce a variety of mediators such as epidermal, fibroblast, and vascular growth factors, able to induce the proliferation of different cell types involved in the healing process. The essential role of macrophages in tissue regeneration has been investigated in several mouse model of tissue injury. In injured liver, macrophages promote the expansion of liver progenitor cells and limit the number of myofibroblasts, a process associated with increased levels of MMPs and IL-10. Macrophage depletion results in reduced liver progenitor differentiation and maturation [41]. Similarly, in injured kidney, IL-10-expressing macrophages limit inflammation, protect tubular cells from apoptosis, and induce regenerative proliferation [42].

In the remodeling phase, macrophages support the maturation of the regenerated tissues through the reorganization of the extracellular matrix, the vasculature, and scar tissue formation. This phase can take very long time: human magnetic resonance imaging indicated that scar tissue can persist at the repair site up to months or years after injury [43]. In mice, skin wound healing usually resolves in 1–2 weeks, but scar maturation, including collagen cross-linking and replacement of collagen III by collagen I, may last up to 6 months [44]. In the remodeling phase, macrophages undergo phenotypic changes, increasing the expression of CD206 and CD163, producing TGF-β and several matrix proteases, and decreasing the expression of VEGF, arginase I, and insulin-like growth factor [37, 45].

Although monocytes/macrophages are essential for the correct healing process, long-term accumulation and/or dysregulation of macrophage phenotype and functions result in failure to heal or in fibrosis. Mouse models of diabetes exhibit impaired healing and over-accumulation of macrophages, whose depletion accelerates wound closure [46, 47]. In chronic venous ulcers, prolonged accumulation of pro-inflammatory macrophages lead to a ROS-mediated DNA damage, fibroblast cellular senescence, and defective tissue repair [35]. The same observation was found in impaired healing outcome in mouse models of myocardial infarction and kidney injury [48, 49]. Fibrosis is a common feature of inflamed lung and other parenchymal organs. It has been demonstrated that macrophages play a pivotal role in lung fibrosis, especially in TGF-β-mediated fibrotic response [50, 51]. TGF-β expression and M2-polarized macrophages have also been involved in the fibrosis associated to Duchenne muscular dystrophy [52]. In the injured liver, macrophage depletion during the fibrinogenic phase leads to a reduced scar formation and decreased number of myofibroblasts. On the contrary, during the recovery phase, the absence of macrophages results in a failure of matrix degradation and fibrosis [53].

In summary, resolution of inflammation and tissue repair processes involve mononuclear phagocytes, whose phenotype, usually set in an M2 or M2-like mode, may vary depending on the phase of healing progression and tissue type. Perturbation of this dynamic reprogramming of macrophage functions may lead to a failure of resolution of inflammation and tissue repair, eventually resulting in the formation of dysfunctional fibrotic tissue.

Macrophages and cancer

Tumor-associated macrophages (TAMs) are the major component of tumor infiltrating leukocytes and are key players in the link between inflammation and cancer [54]. In established tumors, TAMs generally display an M2-like phenotype: they are devoid of cytotoxic activity, produce growth factors for cancer and endothelial cells, and have immune-suppressive functions [2, 55–57]. However, considerable TAM heterogeneity emerged in recent years, with distinct features in different types of tumors and even at different sites of the same microenvironment [58].

Macrophage phenotype can be modulated during the transition from benign lesions to cancer: in the cancer-initiating phase TAMs may have immunostimulatory functions [59], but at later stages the microenvironment enriches in growth factors and inflammatory mediators, such as CSF-1, IL-4, IL-10, and TGF-β, which cause a shift in macrophage polarization so that they acquire an M2 phenotype with tumor-promoting functions [60, 61].

In the tumor microenvironment, TAMs may influence different aspects of tumor progression. In particular, they promote tumor growth and dissemination, sustain angiogenesis, contribute to matrix degradation, and suppress anti-tumor adaptive immune responses. TAM are a major source of reactive mediators such as cytokines, chemokines, growth factors, reactive oxygen and nitrogen species, and proteolytic enzymes. Among the soluble factors produced by TAM, IL-6 and TNFα are of prime importance. IL-6 activates the transcription factor STAT3 which regulates proliferation and survival pathways in tumor cells [62]. TNFα is a primary inflammatory cytokine that at high concentrations elicits tissue destructive reactions, but at low concentrations, as those usually produced by TAMs, stimulates tumor growth and angiogenesis. Furthermore, TNFα induces the production of chemokines which further recruit myeloid cells (e.g., CCL2) [63, 64].

TAMs are principal inducers of the “angiogenic switch,” the acquisition of a new vasculature that supports tumor growth providing oxygenation and nutrients [65]. Genetic or pharmacological depletion of macrophages in experimental mouse tumors results in the inhibition of angiogenesis in tumors [66, 67]. The role of hypoxia in guiding macrophage functions has been investigated. New evidence indicate that while hypoxia is not a major driver of TAM subset differentiation, it specifically fine-tunes the phenotype of M2-like MHC-II^low TAM [68].

TAMs preferentially localize in the hypoxic areas of tumors, where they express the transcription factor HIF 1α that induces the transcription of different genes associated to angiogenesis, such as VEGF, bFGF, PDGF, and prostaglandin E2 [69]. Overall, in the hypoxic areas of tumors TAMs activate a transcription program in which mitogenic, pro-invasive, pro-angiogenic, and pro-metastatic genes are up-regulated[70]. We previously showed that hypoxic induction of HIF-1α in TAMs influences the positioning and function of tumor cells, stromal cells, and TAMs, by selectively up-regulating the expression of the chemokine receptor CXCR4[71]. Furthermore, HIF-1 activation mediates expression of the CXCR4 ligand CXCL12, a chemokine involved in angiogenesis and cancer metastasis [72, 73]. Recently, Casazza et al. showed that cancer cell-derived Sema3A promotes accumulation of tumor-associated macrophages (TAMs) into avascular areas of tumors cancer, through a Neuropilin-1 (NRP1)-dependent signaling cascade [74].

Pro-angiogenic TAMs include a sub-population of macrophages characterized by the expression of the angiopoietin (ANG) receptor Tie2 (Tie2 expressing monocytes, TEMs). TEMs often align along the blood vessels in a Tie2-ANG-interacting manner, and targeting this interaction results in a reduction of angiogenesis [75].

TAMs play a critical role in the process of tumor cell invasion and metastasis. TAMs produce enzymes and proteases that regulate the degradation of the extracellular matrix (ECM), such as matrix metalloproteases (MMPs), plasmin, urokinase-type plasminogen activator (uPA) and its receptor, osteonectin, and cathepsins [76–78]. ECM disruption by TAMs facilitates tumor cell migration and spreading [54].

TAMs also contribute to an immunosuppressive environment within tumors: they are poor antigen-presenting cells unable to secrete IL-12, while in turn produce IL-10 and TGF-β to block T cell proliferation, suppress cytotoxic T cell responses, and activate regulatory T lymphocytes [57, 79–81].

Based on the tumor-promoting functions of TAMs, it is not surprising that TAM infiltration usually correlates with reduced patients survival, as observed in different tumor types: ovarian and breast cancer, follicular B lymphoma, soft tissue sarcoma, and classic Hodgkin’s Lymphoma [82–85]. New strategies targeting TAMs represent an active area of research to improve anti-tumor therapies (see below).

Macrophage polarization in infectious diseases

A remarkable example of macrophage plasticity is displayed in response to microbial agents, where monocytes/macrophages acquire distinct polarized programs to establish specific protective conditions. However, several evidence also indicate that microbes might derange macrophage functions to avoid cytotoxic functions and evade the immune response. Finally, activation of programs supporting tissue homeostasis and healing require a functional reprogramming of macrophages.

Sepsis

M1-polarized inflammation in response to bacteria like Listeria monocytogenes [86], Salmonella typhi and typhimurium [87], acts as a primary defense mechanism. However, to avoid tissue damages an M1/M2 shift of macrophage polarization occurs during the transition from early to late stages of infection, hence restraining inflammation and favoring tissue repair. Coherently, a typical M1/M2 signature transition was observed during treatment or convalescence of typhoid fever patients, in peripheral blood cells [88]. Similarly, the systemic inflammation occurring during the initial phase of “Systemic Inflammatory Response Syndrome” (SIRS) associated with severe sepsis, is followed by a state of tolerance, that supports the compensatory anti-inflammatory response syndrome (CARS) [89–91]. Using an in vitro model of LPS-induced tolerance, we demonstrated that the SIRS to CARS transition underlies an M1-M2 switch of polarized monocyte/macrophage activation [92]; according a distinct M2 gene expression profile was confirmed in endotoxin tolerant human mononuclear phagocytes [93]. Although this functional reprogramming represents a protective mechanism to counteract overwhelming inflammation, it also associates with increased risk of relapse and susceptibility to secondary infections and mortality [94, 95].

Mycobacterium tuberculosis infection

Despite the progress in prevention, diagnosis, and treatment, Tuberculosis (TB) remains as one of the world’s deadliest infective diseases (Global Tuberculosis Control, World Health Organization report 2014) and several lines of evidences underline the role of alternative activated/M2 macrophages in the promotion of TB pathogenesis [96–100], suggesting new potential therapeutic interventions. Mycobacterium tuberculosis (MT) induces human macrophages to produce IL-10, which in turn favors bacterial survival and growth inside macrophages, by blocking the phagosome maturation and the host immune response [101]. In agreement, IL-10 polymorphisms were associated with increased TB susceptibility and increased expression of IL-10 correlated with higher levels of MT antigen CFP32 in the sputum of active TB patients [102]. In contrast, neutralization of IL-10 in ex vivo cultured PBMC from active TB patients resulted in increased IL-12 secretion by monocytes, as well as enhanced T cell proliferation and IFNγ production [103, 104]. Alveolar macrophages (AM), which are committed to control local inflammation, show a uniquely immunoregulatory M2-oriented phenotype, which encompasses low antigen presentation capacity, limited production of oxidants, and enhanced production of anti-inflammatory cytokines [102]. Hence, AM avoid excessive lung injury, but limit defense against invading pathogens. For these reasons, during the early phase of MT infection the extent of the M1- and Th1-polarized immune response is not sufficient to radically eliminate mycobacteria [96]. Interestingly, MT can bias macrophage polarization toward an M2 polarization state through different mechanisms. In this regard, engagement of TLR2 by MT heat shock protein 60 (Mtbhsp60) results in clathrin-dependent endocytosis, associated with increased production of IL-10 [105]. Further, binding of MT mannose-capped lipoarabinomannan to the macrophage mannose receptor (CD206) enhances the expression of Peroxisome proliferator-activated receptor γ (PPARγ) [97], that skews macrophage activation toward an M2-phenotype [106]. In agreement, PPARγ is higher in PBMCs derived from tuberculosis patients and knockdown of PPARγ results in decreased expression of M2 genes in MT-infected human macrophages (i.e., IL-10, Arginase I and II, Dectin 1, Mannose Receptor), paralleled by the increased expression of anti-microbial M1 genes (e.g., iNOS, TNF and IL-6) [97, 107]. PPARγ is also engaged by both MT and host macrophage lipids and, in synergism with other lipid-sensing nuclear receptors (e.g., Testicular receptor 4), supports intracellular MT survival through distinct mechanisms, including inhibition of phagolysosome maturation and ROS/NO production, and increased foam cell formation [107].

Furthermore, upon granuloma formation, ATP released by either dying cells or activated T lymphocytes may influence macrophage activation. In this regard, exposure of human MT-infected macrophages to high levels of ATP (1-3 mM) induces bacterial clearance by enhancing phagosome-lysosome fusion and autophagy in a P2 × 7-dependent manner [108, 109]. However, this high concentration of ATP appears unlikely reached in vivo [110] and gene expression profile of human MT-infected macrophages exposed to low levels of ATP (100 μM) showed an M2-skewed phenotype [111]. This latter event is likely dependent on the CD39-mediated ATP to AMP hydrolysis, with the subsequent engagement of the adenosine A_2A receptors [111] and triggered expression of an M2-like phenotype [112]. Finally, similarly to other pathogens (e.g., Leishmania parasites, Legionella pneumophila, Trypanosoma cruzi, and Toxoplasma gondii), MT can promote AMP generation by production of ATPase [113].

Helicobacter pylori infection

Helicobacter pylori is the causative agent of chronic gastritis, peptic ulcer disease, gastric MALT lymphoma, and gastric cancer, that is the third leading cause of cancer-related death worldwide (http://globocan.iarc.fr). H. pylori has developed several mechanisms to escape the immune response and survive in the stomach. This bacterium triggers a mixed Th1/Th17/Treg immune response that results in ineffective eradication of the infection, paralleled by sustained epithelial damage evolving to gastritis and gastric cancer [114]. Mouse and human studies highlighted the pivotal role played by macrophages in driving this complex inflammatory response. In patients with H. pylori infection and chronic gastritis, mucosal macrophages promote Th17 expansion by producing BAFF that directly induces CD4⁺ T cell differentiation and production of Th17-inducing cytokines by macrophages [115]. Gobert and colleagues demonstrated, in both mice and men, that the Cag⁺ H. pylori strain induces gastric macrophages to express hemeoxygenase 1 (HO-1), which restrains transcription of pro-inflammatory M1 genes and upregulates selected regulatory genes (i.e., IL-10) [116]. Additional studies showed either in vitro or in vivo (i.e., gastric mucosa of mice and humans) that H. pylori induces macrophages to express the M2 markers arginase II (Arg2) [117–119], Ornithine decarboxylase (ODC) [120, 121], and Spermine oxidase (SMO) [122, 123]. Similar to other pathogens, including MT [124], H. pylori induced up-regulation of these enzymes contributes to immune evasion by diverting the metabolism of l-Arginine toward the production of ornithine and polyamines, rather than anti-microbial NO. These findings agree with the observation that polyamines promote the M2 phenotype by stimulating IL-4-induced gene expression, while inhibiting LPS-induced gene expression [125]. Further, polyamines may exert anti-inflammatory effects by inhibiting pro-inflammatory gene expression and promoting macrophage apoptosis [122, 123]. In gastric epithelial cells, H. pylori-induced SMO expression generates oxidative stress and DNA damage, hence supporting neoplastic transformation [126]. Coherently, gastric tissues from H. pylori-infected patients and gastric tumors showed high levels of polyamines and increased ODC transcript levels correlate with an enhanced risk of gastric cancer in humans [126].

In both human gastric epithelial and monocytic cells, H. pylori inhibits IFNγ signaling through activation of Src homology-2 domain–containing phosphatase (SHP)2, in CagA- and ROS-dependent manners [127], thus favoring gastric cancer development [128–130].

HIV infection

Despite the advent of potent, highly active antiretroviral therapy (HAART) has greatly improved the life of immunodeficiency virus (HIV) infected people, HIV infection remains a major clinical challenge. The major hurdles of HAART are inaccessibility to some anatomical sites, like brain and testes, and resistance of some infected cells, including dendritic cells, latently infected CD4⁺ T cells, and monocytes/macrophages, that act as viral reservoirs. As a consequence, HIV infection is associated with chronic comorbidities and aging-related complications, including cancer [131], neurocognitive [132], and cardiovascular disorders [133]. Contradictory in vitro evidence indicated that both M1- and M2-polarized macrophages can restrain HIV infection. However, M1 polarization appears the most powerful antiviral program, inhibiting virus entry and replication at pre- and post-integration levels, limiting virus spreading and enhancing activation of adaptive immunity [131, 134]. An M1 toward M2 polarization switch of circulating monocytes occurs during the transition from early to late phases of HIV infection, which might support the establishment of chronic latent infection [135]. Macrophages represent the most important HIV reservoir, due to their ubiquitous distribution, long half-life, and resistance against the viral cytopathic activity. Increasing evidence indicate that infected monocytes/macrophages are crucial promoters of HIV-related comorbidities. HIV-infected monocytes cross blood–brain barrier allowing virus entry in the central nervous system [136]. Besides direct viral neurotoxicity, macrophages driven-neuroinflammation plays a major role in neuronal dysfunction and death. Indeed, HIV-infected macrophages produce pro-inflammatory cytokines like TNFα and IL-1β that activate astrocytes and microglia to secrete neurotoxic factors, such as excitatory amino acids and inflammatory mediators [132]. Similarly, along with HIV-mediated lipid dysfunctions, HIV-dependent activation of both endothelial cells and monocytes crucially promotes the pathogenesis of atherosclerosis [133].

Parasites infection

Parasite infections represent a major health problem in developing countries, where both protozoans and helminths are endemic and based on the World Health Organization (WHO) about 200 millions of cases of malaria were reported in 2013. Despite malaria could be treated, therapy is extremely expensive and more than half million of infected people, mostly African children, still die every year [137]. Moreover, helminth infections (e.g., Schistosoma haematobium, Clonorchis sinensis, and Opisthorchis viverrini) are causally associated with higher incidence of cancer in the infected organs [138]. Depending on the causative agent of infection, human parasites promote either M1- or M2-polarized activation programs. Protozoans (e.g., Leishmania, Toxoplasma, Trypanosoma, Plasmodium) generally elicit an early M1 macrophage polarization that restrains parasitemia and controls the disease, followed by a partial M1 to M2 shift that limits inflammation-dependent tissue damages, but supports chronic infection [139]. Hence, a delicate M1 vs M2 inflammatory balance tightly controls the disease outcomes. Fitting this paradigm, at early stages of Plasmodium infection activated innate immune cells produce type I Interferons that in turn induce NK and NKT cells to secrete IFNγ [140]. An IFNγ-dependent priming of monocytes/macrophages is crucial to promote resistance against infectious insults, but favors an exaggerated release of pro-inflammatory cytokines in response to Plasmodium-derived pathogen-associated molecular patterns (PAMP) and malaria-associated damage-associated molecular patterns (DAMP), leading to the clinical manifestations of malaria and associated syndromes (systemic inflammation, anemia, metabolic acidosis, cerebral and placental malaria)[140]. To counteract malaria immunopathology, monocytes/macrophages upregulate M2-related immunoregulatory molecules, such as HO-1 [141] and IL-10 [142], which hamper both oxidative burst and inflammation while increasing the susceptibility to bacterial superinfections, especially non-typhoidal Salmonella [143]. Macrophage plasticity may account for several other vicious synergisms between pathogens, worsening disease outcome in co-infected individuals. For example a synergistic increase of arginase activity in circulating PBMC is associated with poor immune response and prognosis of Leishmania donovani–HIV co-infected patients [144].

In contrast, during helminth infections macrophages acquire an M2 phenotype capable to control different clinical parameters, including parasites clearance, confinement in granulomas, repair of tissues damaged by parasites transit, and early cytotoxic immune responses [145–147]. However, even during helminth infections M2-polarized activation could be a double-edged sword. In fact, the induction of oxidative stress along with immunosuppressive M2-skewed inflammation are recognized as the key mechanisms linking Schistosoma haematobium, Clonorchis sinensis, and Opisthorchis viverrini infections with the development of bladder cancer and cholangiocarcinoma [145]. In a small proportion of Schistostoma mansoni and japonicum infected individuals, the type 2 switch that occurs during disease progression avoids parasites clearance and results in a life-threatening hepatosplenic schistosomiasis associated with chronic hepatic fibrosis, portal hypertension, and hemorrhages [146]. While Th2 expansion is recognized as a key promoter of fibrosis, the role of M2 macrophages is still controversial. Mouse studies indicate that an enhanced arginase-1 activity in M2 macrophages limits fibrosis by inhibiting the proliferation of Th2 cells [148]. In line with the crucial recruitment of bone marrow-derived CCR2^HiLy6C^Hi CX3CR1^Lo monocytes in fibrotic areas of damaged tissues, CCR2^−/− mice showed decreased fibrosis in response to S. Mansoni eggs [149].

Macrophage polarization in allergy

Allergic diseases are the prototypical disorders associated with Th2/M2-polarized inflammation. Allergic asthma is the most common chronic inflammatory disease and is often found in conjunction with other allergic disorders, like atopic eczema and allergic rhinitis, supporting the idea that these different allergic responses represent disparate outcomes of common immune disorders. A number of studies in mice indicate macrophages as key orchestrators of allergic asthma and promoters of inflammatory responses associated with lung injury, fibrosis, and goblet cells hyperplasia [150, 151]. In line, analysis of bronchial biopsy specimens showed increased number of CD206⁺ and stabilin-1⁺ macrophages in asthmatic patients, supporting a correlation between the percentage of M2 macrophages and disease severity [152]. An increased number of circulating CD14⁺CD16⁺ intermediate monocytes, expressing an M2-like phenotype has been found in patients with allergies and bronchial asthma [153]. Furthermore, in vivo studies demonstrated that in response to endobronchial-allergen challenge macrophages of asthmatic patients undergo M2 polarization and consequently support Th2 inflammation [154].

Along with the M2 polarizing cytokines IL-4 and IL-13, epithelial-derived IL-33 [155] and thymic stromal lymphopoietin (TSLP) [156] were pointed as additional amplifiers of M2-polarized macrophage activation. Accordingly, asthma patients show high levels of IL-33 in both serum and lung epithelial cells [155] and increased concentrations of TSLP have been found in both skin biopsies and in the airways of patients with atopic dermatitis [157] and asthma [158], which correlated with disease severity.

Despite the large body of evidence supporting the link between airways disease and Th2/M2 inflammation, accumulating studies indicate asthma as an heterogeneous disease with many endotypes [159]. Within this puzzling scenario M1-polarized macrophages may contribute to the pathogenesis of asthma by releasing inflammatory cytokines and nitric oxide, which support exacerbation of lung injury and airway remodeling, particularly in patients with steroid-resistant or severe forms of asthma [160, 161].

Obesity and metabolism

Obesity-associated inflammation contributes to the development of metabolic diseases including cardiovascular, fatty liver disease, diabetes, and cancer. This scenario is advocated by accumulation and activation of adipose tissue (AT) immune cells, including B cells, T cells, neutrophils, and a larger macrophage population [162], which create an active immunological tissue that establishes systemic chronic metabolic inflammation (metaflammation). Hence, sensing of adipose tissues by immune cells, macrophages, in particular, represents as key event of obesity-associated comorbidities [163].

AT macrophage accumulation correlates with the degree of obesity and adipocytes from obese subjects promote their accumulation and activation through the release of inflammatory mediators, such as chemokines, TNF, or free fatty acids (FFA) [162, 164, 165].

Selected chemokines and chemokine receptors (e.g., CCL2/CCR2 or CCL5/CCR5) [162, 166, 167] and the macrophage-derived apoptosis inhibitor of macrophages (AIM) [168] guide AT macrophage accumulation. AT macrophages organize in crown-like structures around dead adipocytes [169, 170], where they acquire an inflammatory M1-like phenotype and contribute to metabolic dysfunctions via secretion of inflammatory mediators [165, 171–174]. The appearance of crown-like structures follow the onset of high-fat feeding (Strissel KJ, Stancheva Z et al. Diabetes 2007). In response to excess of energy, adipocytes undergo hypertrophy and/or hyperplasia and smaller adipocytes have lower markers of inflammation [175, 176].

Inflammatory cytokines produced by AT macrophage (e.g., TNF, IL-6, IL-1β) counteract the insulin-sensitizing action of adiponectin and leptin, thus leading to insulin resistance [162, 165, 166]. In particular, TNF-α stimulates adipocyte lipolysis contributing to elevated serum FFA concentrations, which can lead to decreased insulin sensitivity [164, 177].

Despite evidence indicate that the AT macrophage population is a mix of M1 and M2 cells, weight loss is associated with a shift-back toward M2-like phenotype [171, 178] and increased number of lipolytic F4/80 + CD11c-CD301 + M2-like macrophages [179].

AT macrophages sense adipose tissue damage-associated molecular patterns (DAMPs) through pattern recognition receptors. Fetuin A, released by adipocytes in response to FAA exposure, represents an endogenous ligand of TLR4 that allows FFA-TLR4 cross-talk and leads to an M1-like phenotype of macrophages, as well as to inflammatory adipocytes [180]. In concert, the NLRP3 inflammasome promotes adipose tissue inflammation, by sensing FFA, ATP, ceramides, glucotoxicity, cholesterol, amyloid-β, and urate crystals [163]. This pathway is active in type 2 diabetes and is inhibited by metformin in diabetic patients [181]. Of relevance, expression of the inhibitor of κB kinase epsilon (IKBKE) in hematopoietic cells limited the NLRP3 priming and the associated metaflammation [182]. AT macrophages also act as antigen-presenting cells promoting T cell proliferation and Th1 skewing of CD4 + T cells with increased IFNγ production, eventually supporting M1-like functions of AT macrophages [183].

Hypercholesterolemia supports the expansion of circulating Ly6C^high inflammatory monocytes through the increased expression of the common β-subunit of the IL-13 and the GM-CSF receptor, that promotes proliferation of haematopoietic stem and progenitors cells (HSPC) [184, 185]. Ly6C^high monocytes are recruited to progressing atherosclerotic lesions, as progenitors of the M1 macrophages found in the plaques [184–186]. Evidence in mouse models of atherosclerotic regression have shown the M1/M2 balance in plaques to be dynamic, with M1 predominating in disease progression and M2 in regression [187]. Furthermore, oxidized LDL induce a macrophage population termed Mox, found in the progressing plaques and distinct from M1 or M2 macrophages. MOx macrophages are characterized by increased expression of nuclear factor erythroid 2-related factor 2 (NRF2)-dependent genes and in reactive oxygen species [188].

Iron homeostasis is emerging as a new activity of AT macrophages and elevated body iron levels have been linked with metabolic syndrome and obesity [189]. Recent studies indicated that M2 macrophage polarization induces an increased capacity for iron uptake and release (iron-recycling phenotype), characterized by elevated gene expression of Tfrc, Cd163, Hmox1, and Fpn, while in contrast M1 polarization elicits an iron sequestration phenotype [190]. More recently, a population of M2-polarized AT macrophages, with elevated cellular iron content (MFe^hi) and an iron-recycling gene expression profile, has been described [191]. Selective decrease in MFe^hi iron content was reported in obesity, which lead to adipocyte iron overload, associated with increased lipolysis and lipid peroxidation, decreased adiponectin, and increased ROS production [189]. The enzyme heme oxygenase 1 (HO-1) metabolizes heme into iron, bilirubin, and CO and is part of the cellular defense against oxidative stress [192]. In apparent discrepancy, in obese human subjects HO-1 was demonstrated to be a positive predictor of metabolic disease and macrophage-specific depletion of HO-1 in mice resulted in resistance to diet-induced insulin resistance, reducing steatosis, and liver toxicity [193].

In line with the established association between M1 inflammation and insulin resistance, genetic ablation of key transcription factors guiding M2 macrophage polarization [21] demonstrated that myeloid-specific disruption of PPARγ, PPARδ, and Krüppel-like factor 4 (KLF4) predisposes to development of diet-induced obesity, insulin resistance, and glucose intolerance [106, 194].

Beige fat provides a thermogenic defense against cold and obesity. In response to cold, IL-4–driven M2 macrophages release noradrenaline in brown adipose tissue (BAT) and white adipose tissue (WAT), which coordinates fatty acid mobilization and energy expenditure [195]. More recently, Qiu Y et al. identified eosinophils, type 2 cytokines interleukin (IL)-4/13, and alternatively activated macrophages as a key circuit of the thermogenic defense, showing that macrophages recruited to cold-stressed white adipose tissue via the chemokine receptor CCR2 undergo alternative activation and produce catecholamines, required for browning of WAT [196]. Furthermore, Perino et al. recently demonstrated that pharmacological and genetic inhibition of PI3 Kβ and PI3 Kγ in mice resulted in increased norepinephrine release from the sympathetic nervous system, with increased lipolysis and browning of the white adipocytes [197].

Overall, the current knowledge over AT macrophages indicate that M2-like AT macrophages orchestrate adipose tissue homeostasis and promote insulin sensitivity, whereas M1-like AT macrophages promote obesity-associated inflammation and insulin resistance.

Concluding remarks

Progresses have been made in defining the role of macrophages and their polarization state in pathology. Molecular determinants of M1 versus M2 polarization include members of the PPAR, KLF, IRF, STAT, NF-κB, and HIF families [21]. In addition, epigenetic modifications with involvement of miRNAs, histone methylation and acetylation have emerged as regulators of phagocyte activation and function [5, 198–201]. Functional macrophage polarization has reported in vivo, under physiological (embryogenesis and pregnancy and normal maintenance of selected tissues, such as testis and adipose) and pathological conditions (chronic inflammation and tissue repair, metabolic and vascular disorders, infection, and cancer). Several evidence indicate polarized macrophage activation as a key component in pathology (Fig. 2). However, in a number of pathological conditions, including infection (e.g., H. pylori) and neurodegenerative disorders [202] macrophage populations express mixed or unique phenotypes. As mentioned above, the binary distinction into M1 and M2 macrophages, remains a useful oversimplification of the biology of mononuclear phagocytes and new evidence indicate that a wider spectrum of macrophage activation exists. This vast functional plasticity is dependent on the type and combination of stimuli encountered by macrophages [7] and integrates the memory effects elicited at transcriptional level by selected stimuli [5]. Understanding the role of coexisting phenotypes of macrophages and the mechanisms driving their dynamic changes during disease progression will be a key challenge in the coming years. This effort will likely provide the basis for macrophage-centered diagnostic and therapeutic strategies.

Fig. 2.

Mechanisms and markers of polarized macrophages in pathology. a Key signals driving M1 vs M2 polarization of macrophages. Inflammatory cytokines (IFNγ, TNFα), pathogen-associated molecular patterns (e.g., LPS), and damage-associated molecular patterns (e.g., HMGB1, HSP, Fetuin A, ATP) induce M1-polarized activation, whereas Th2 cytokines (IL-4, IL-13), anti-inflammatory molecules (IL-10, GC, AMP), and immunocomplexes (Ic) induce M2-polarized activation. IL-33 and thymic stromal lymphopoietin (TSLP) act as M2 amplifiers (asteriks). M1 and M2 signals engage signaling pathways involving different kinases (e.g., Akt1 and 2) and family of transcription factors (e.g., STATs, IRFs, NF-κB, KLFs, HIFs). Further, PPARγ and PPARδ control distinct aspects of M2 macrophage activation and the oxidative metabolism. c-Myc and c-Maf, respectively, regulate a subset of IL-4- and IL-10- inducible genes. Macrophage-polarized activation is also controlled at epigenetic levels by different miRNAs (miR-155, miR-124) and proteins involved in histone methylation and acetylation (bromodomain-containing BET proteins, JMJD3). Inflammatory cytokines (TNFα, IL-1β, IL-6, IL-12, IL-23, IL-27), Th1-recruiting chemokines (CXCL9, CXCL10, CXCL11), and the co-stimulatory receptor CD40, represent distinct markers of M1 polarization. M2 macrophages express higher levels of genes encoding anti-inflammatory cytokines (IL-10, TGFβ), Th2-recruiting chemokines (CCL17, CCL18, CCL22), c-type lectin (CD206, CD301, dectin-1), and scavenger receptors (CD163, Stabilin-1). Further, M1- and M2-polarized macrophages express distinct enzymes involved in iron and amino acid metabolism. Hence M1 macrophages exert anti-microbial activities by iron retention (ferritin, CP, DMT-1, Nramp-1) and ROS/NOS production (iNOS, gp91phox, p22phox), whereas M2 macrophages promote wound healing through iron recycling and release (Tfr, HO-1, Fpn) and polyamine production (Arg 1, Arg 2, ODC, SMO). b M1 and M2 macrophage polarization in disease. Association of M1 vs M2 macrophage polarization in distinct diseases. Dynamic reprogramming of macrophage polarization occurs during the progression of both infections (sepsis, protozoans, HIV) and cancer and mixed macrophage phenotypes can coexist (e.g., H. pylori infections). Systemic Inflammatory Response Syndrome (SIRS); Compensatory Anti-inflammatory Response Syndrome (CARS); Interferon γ (IFNγ) lipopolysaccharide (LPS) High Mobility Group Box 1 (HMGB1); Heat Shock Proteins (HSPs); glucocorticoids (GC) Adenosine monophosphate (AMP); Nuclear factor κB (NF-κB); Signal Transducer and Activator of Transcription (STAT); Interferon Regulatory Factor (IRF); Hypoxia Inducible Factor (HIF); Krüppel-like factor (KLF); peroxisome proliferator-activated receptors (PPAR); CCAAT/enhancer binding protein (C/EBP); Inducible nitric oxide synthase (iNOS); ceruloplasmin (CP); natural resistance-associated macrophage protein 1 (Nramp-1); divalent metal transporter-1 (DMT-1); Arginase (Arg); ornithine decarboxylase (ODC); spermidine oxidase (SMO); hemeoxygenase-1 (HO-1); ferroportin (Fpn); transferrin receptor (TfR)