Abstract

Introduction

Macrophages are highly plastic immune cells involved in tissue homeostasis, immune response and inflammation. In recent years, numerous studies have demonstrated that macrophage plasticity is closely linked to a metabolic adaptation that supports the energy requirements and cellular growth associated with these functions (1). In the inflammatory milieu, stimulated macrophages face challenging metabolic conditions in which they must compete for nutrients and oxygen with other immune cells or pathogens invading the tissue. In this environment, macrophages must generate sufficient energy output to sustain growth, phagocytosis, antimicrobial functions, and also to fuel the highly demanding synthesis of inflammatory mediators that contribute to the propagation or resolution of the inflammatory response. In a simplified view, macrophage polarization states have been defined as proinflammatory M1 (classically activated) or anti-inflammatory M2 (alternatively activated), a classification that mirrors the Th1/Th2 polarization of T cells (2). The definition of these activation states contributed to the understanding and categorization of the opposing functions of macrophages: killing and repairing (3). However, in reality, the dichotomous M1/M2 distinction is not representative of what takes place in an in vivo setting, as M1 and M2 stimuli do not exist alone in tissues. Instead, the macrophage population represents a continuum of phenotypes that stands between these two extremes, implying that discrete populations should not be so crudely depicted (4). Indeed, transcriptome analysis of stimulated macrophages with different activation signals show that these cells experience a transcriptional reprogramming that extends the M1/M2 paradigm (5). However, since most of the preceding literature has used the nomenclature based on the M1/M2 classification as a tool for dissecting the complex macrophage phenotypes, this terminology has been maintained throughout some parts of this review.

Pathological situations in which nutrient availability is compromised, such as infection, chronic inflammation, diseases associated with metabolic/nutrient imbalance (diabetes, obesity, atherosclerosis) or ischemia/reperfusion events associated with organ transplantation or surgery, generate metabolic stress that potentially subverts macrophage functions to induce maladaptive polarization states (6–8). Macrophages can perceive these signals in the tissue microenvironment via metabolic sensors that coordinate metabolic and transcriptomic rewiring and are therefore very responsive to any abnormal imbalance associated with pathology. For instance, hypoxic (oxygen-limiting) environments associated with inflammation or ischemia activate cellular sensors for oxygen and the hypoxia-inducible factor (HIF), which induce a metabolic switch from oxidative to glycolytic metabolism and proinflammatory polarization that further exacerbates the inflammatory response (9, 10). This hypoxic environment is also closely linked to an endoplasmic reticulum (ER) stress response, which is critical for the integration of the metabolic and inflammatory responses in macrophages. The ER organelle plays a central role in cellular nutrient sensing, activating the signaling pathway called the unfolded protein response (UPR) under metabolic stress conditions such as hypoxia or nutrient imbalance (amino acid or glucose deprivation, infectious process, etc.). This response is partially mediated by the mTORC1 pathway, which is a positive regulator of protein synthesis, and cell growth that coordinates the cellular balance between anabolic pathways and energy consumption in macrophages (11).

Considering all this evidence, it is clear that cellular sensors for oxygen and ER stress pathways contribute critically to the signal integration and metabolic adaptation associated with various pathological conditions. In this context, macrophage polarization lies at the intersection between metabolic imbalance and inflammation, and understanding the molecular pathways connecting these processes will be critical for the development of new therapeutic strategies. Here, we review how ER stress and hypoxic responses are organized and connected with macrophage function, focusing particularly on the maladaptive polarization states associated with the pathological contexts in which the metabolic balance in macrophages is compromised.

Molecular Mechanisms in ER Stress: Unfolded Protein Response

The ER has a crucial role in maintaining cellular functions, such as protein folding, maturation and assembly of proteins that are trafficked along the secretory pathway, as well as preserving cellular calcium homeostasis. Several physiological and pathological conditions involving imbalance in ER folding capacity, accumulation of misfolded proteins, hypoxia, amino acid or glucose deprivation, oxidative stress, viral infection or disruption of ER calcium balance can trigger ER stress and activate the UPR that maintains cellular homeostasis and cell survival (12). This mechanism rescues the cells from the damage caused by ER stress, and in the event of unresolvable stress, induces apoptosis. The UPR comprises three major signaling pathways, which are initiated by the activation of three protein sensors: activating transcription factor 6 (ATF6), pancreatic eukaryotic translation initiation factor 2α (eIF2α) kinase (PERK), and inositol-requiring enzyme 1α (IRE1α). Under normal conditions, these sensors are bound to glucose-regulated protein 78 (GRP78), an ER chaperone, also known as BiP (binding immunoglobulin protein), that maintains them in an inactive state. Under ER stress conditions, GRP78 dissociates from the sensors and binds to unfolded proteins (13), allowing activation by dimerization or translocation (Figure 1). Accordingly, activated IRE1α performs two enzymatic functions upon dimerization: serine/threonine kinase and endoribonuclease (RNase) activity (14). This RNase domain in IRE1α initiates the non-conventional splicing of XBP1, which produces a translational frameshift and creates XBP1s (spliced), a potent transcriptional activator (15). XBP1s is translocated to the nucleus and induces the transcription of an extensive variety of chaperones and enzymes that together increase ER size and function. During prolonged ER stress, the beneficial effects of IRE1α activation through the activity of XBP1s may be reduced and could activate the properties of IRE1α to promote inflammation and cell death (16). Similar to IRE1α, dissociation of GRP78 from the ER luminal domain of PERK induces activation through homodimerization and autophosphorylation. Activated PERK phosphorylates the eukaryotic translation initiation factor (eIF2α), which reduces the degree of recognition and translation of the AUG initiation codon, resulting in a lower protein load on the damaged ER (17). Remarkably, eIF2α can be phosphorylated independently of ER stress by the general control non-derepressible 2 (GCN2) kinase, which is induced by amino acid deprivation (18). This results in the activation of ATF4, Nrf2 (nuclear erythroid 2 p45-related factor 2), and NF-κB (nuclear factor kappa β), a master transcription factor with numerous functions, including regulation of the inflammatory response (19). ATF4 also activates the transcription of C/EBP homologous protein (CHOP), which is involved in ER-stress-mediated apoptosis and cell death (20). The third arm of the pathway relies on ATF6, a type II transmembrane protein. There are two isoforms of ATF6 (ATF6α and ATF6β), but only ATF6α is required to activate UPR. When ATF6 senses ER stress, it is transported to the Golgi, where it is cleaved by site 1 and site 2 proteases (S1P and S2P) (21). The cleaved soluble cytosolic fragment can now enter the nucleus and trigger the expression of genes encoding proteins such as GRP78, GRP94, p58IPK, and XBP1 that function to increase ER capacity and folding as well as the expression of genes of the ERAD pathway involved in protein degradation (22, 23). Of note, cleaved ATF6 can act as an enhancer to increase the CREBH-mediated (cAMP response element-binding protein H) acute inflammatory response (24), representing a link between ATF6 and inflammation.

Open in a separate window

Figure 1

Crosstalk between hypoxia and the unfolded protein response (UPR) pathways. Hypoxia induces protein misfolding due to the lack of oxygen required for the formation of disulphide linkages, which leads to endoplasmic reticulum (ER) stress and the activation of UPR. Under basal conditions, BIP binds to PERK, IRE1α and ATF6 and prevents their activation. In response to ER stress during hypoxia, BIP dissociates from the three sensors, allowing their activation. Activated IRE1α mediates XBP1 mRNA splicing (XBP1s) that it is translocated to the nucleus and regulates the expression of genes involved in cellular maintenance pathways. Activated PERK phosphorylates eIF2α resulting in a general slowdown of protein synthesis that allows the activation of ATF4, a master transcription factor which induces the expression of target genes such as CHOP. Moreover, there is an inhibition of mTOR-dependent protein translation. Besides, HIF-1α avoids proteasome degradation promoted by prolyl-hydroxylases (PHDs) and von Hippel-Lindau (VHL) protein, which allows its nuclear translocation and the induction of genes required for adaptation to a hypoxic environment. In this context, the inhibition of PHD3 activity enables ATF4 stabilization. Furthermore, ATF4 and XBP1 upregulate the expression of the ubiquitin ligase Siah2, which in turn provokes the degradation of PHDs favoring HIF-1α expression, thus establishing another link between hypoxia and UPR pathways. Moreover, XBP1 and HIF-1α trigger a transcriptional complex that regulates the expression of genes associated with HIF-1α through the recruitment of RNA polymerase II. In addition, hypoxia-induced cell damage is counteracted by the UPR pathway through the upregulation of autophagy by ATF4 and IRE1α. Direct induction of autophagy may be also achieved by BIP-dependent activation of AMPK, which attenuates AKT-mTOR signaling.

UPR in Macrophage Physiology and Function

It is generally accepted that ER stress and inflammation are two mechanisms that allow tissue and cells to resolve infections or stress-induced situations. Various studies have demonstrated the crosstalk between these pathways and show that the UPR route is essential to ensure the normal physiology of dendritic cells and macrophages. Toll-like receptor (TLR) signaling induces ER stress in macrophages, which in turn amplifies the responses after the binding of the TLRs to their ligands (25). TLR2 and TLR4 induce the activation of IRE1α, and consequently of XBP1s, through a mechanism that needs the NADPH oxidase NOX2 and TNF receptor-associated 6 (TRAF6). The knockdown of IRE1α in J774 macrophages cells, prevents the production of IL-6, TNF-α, and interferon β (IFN β). This result shows the crucial effect of the IRE1α-XBP1s axis on the TLR-mediated response (26). Moreover, several studies have reported that ER stress activates glycogen synthase kinase (GSK)-3β, accelerating inflammation and lipid accumulation (27, 28). Experiments performed on RAW264 macrophages revealed that the IRE1α pathway is involved in the increase in IL-1β production through GSK-3β activation. In addition, pharmacological inhibition of GSK-3β shows that GSK-3β inhibits XBP1 splicing, entailing a decrease in TNF-α expression and possibly in TLR signaling (29).

On the other hand, it is known that ATF4 directly binds the IL-6 promoter linking the PERK branch with inflammation (30). The involvement of this UPR branch in inflammation has also been noted in another study in human macrophages, in which the induction of the ATF4-CHOP branch was attenuated by TLR signaling in a TRIF (TIR-domain containing adaptor inducing interferon-β) dependent way (31), whereas the other UPR pathways were not suppressed. Interestingly, TLR engagement did not impede phosphorylation of PERK or eIF-2α, but phospho-eIF-2α failed to promote translation of the CHOP activator ATF4 because TLR-TRIF signaling enhances eIF2B-GEF activity to counteract the effects of p-eIF2α (32). In this way, the beneficial aspects of prolonged physiological ER stress can be achieved, avoiding the negative effects of prolonged CHOP expression because deficiencies in this pathway allow the macrophages to survive, potentiating the immune response. The ATF6 branch synergizes with TLR stimulation in liver macrophages, enhancing NF-κB signaling in the context of liver ischemia reperfusion (33). ER stress in murine bone marrow-derived macrophages (BMMs) induced by chemicals and followed by TLR stimulation yields significantly higher levels of TNF-α and IL-6, but lower levels of IL-10. Moreover, ATF6 KO liver macrophages showed that TNF-α and IL-6 depend on ATF6 in response to TLR4 activation. In addition, the AKT anti-inflammatory signaling pathway is inhibited by the ER stress response in BMMs, and the absence of ATF6 is enough to restore its activation and enhance NF-κB signaling (33).

Finally, in an experimental model of lung injury and fibrosis, GRP78 heterozygous mice have fewer lung macrophages due to their higher caspase-3 and CHOP levels compared with those of wild type (WT) mice. It should be noted that an increase in the frequency of T cells induced by bleomycin was associated with an increment in M2 macrophages in this lung damage mouse model (34). CHOP deficiency in mice allows macrophages to accumulate by inhibiting ER stress-mediated cell death, as assessed in experimental models of liver and pulmonary fibrosis, suggesting that GRP78 protects from lung fibrosis while up-regulation of CHOP promotes it (35). The IRE1-XBP1 arm of the UPR is involved in the synergy of IL-6 with IL-4 and IL-13 in profibrotic “M2-like” hyperpolarization in human and mouse cells. On the basis of all these findings, the authors proposed that an inhibitor of the IRE1-XBP1 axis might be beneficial in preventing the profibrotic capacity of macrophages (36).

The influence of UPR inhibition was analyzed in RAW 264.7 macrophages in which specific knockdown was performed in each UPR signaling arm individually (GRP78, PERK, and IRE1α). These experiments illustrated the differential effects on macrophage polarity due to differential activity in an adipose triglyceride lipase (ATGL), which cleaves triacylglycerol to generate non-esterified fatty acids. GRP78 knockdown in macrophages stimulated with LPS exhibits elevated ATGL levels relative to PERK knockdown macrophages. The same study demonstrated that PERK inhibition prompts an elevated glucose uptake in macrophages independently of LPS stimulation. Metabolic flux analysis demonstrated that IRE1α and PERK inhibition increases mitochondrial metabolism and basal glycolysis in contrast to GRP78 inhibition, which reduces the values of these two parameters (37). All these studies demonstrate the importance of UPR pathways in innate immune functions (Figure 2).

Open in a separate window

Figure 2

Schematic depiction of the regulation of macrophage metabolic phenotype, polarization and function by hypoxia (A) and UPR-dependent (B) pathways as well as their association with the corresponding pathological scenario. (A) In hypoxia, HIF-1α avoids degradation and translocate to the nucleus to induce the expression of glycolytic enzymes such as hexokinase (HKII), phosphofructokinase (PFKFB3) and pyruvate dehydrogenase kinase (PDK), and glucose transporters such as GLUT1. On the other hand, the inactivity of prolyl-hydroxylase 2 (PHD2) during hypoxia entails the increase of PDK1 which further contributes to the glycolytic metabolic switch. Moreover, OXPHOS shutdown provokes the accumulation of the TCA cycle intermediates citrate and succinate which are employed by the macrophage to produce microbicide and proinflammatory molecules. Succinate in turn increases the activity of HIF-1α. These metabolic adaptations potentiate proinflammatory macrophage polarization, characterized by their increased capacity to mount Th1 responses, increased phagocytic and antigen presentation capacities and enhanced inflammasome activation. In contrast, anti-inflammatory macrophages are involved in processes related to the resolution of the inflammatory response and homeostasis reestablishment. Lastly, hypoxia can be found in several pathological situations, where the balance between pro- and anti-inflammatory macrophages determines disease outcome. Proinflammatory macrophages are key mediators in atherosclerosis and autoimmune diseases such as rheumatoid arthritis. Their accumulation in non-healing wounds has been associated with the chronification of inflammation in the wound. However, these macrophages are required to mount proper anti-tumor and anti-bacterial responses, where a shift toward an anti-inflammatory phenotype has been shown to be detrimental. (B) Increased lipid accumulation and ATGL activity promotes “M2-like” macrophage polarization in the absence of GRP78 (also known as BIP), whereas decreased ATGL activity and elevated glucose uptake, leading to enhanced glycolysis to favor “M1-like” macrophage polarization, have been detected in the absence of PERK. Conversely, PERK knockdown induces an “M2-like” polarization state, although the explanation of this discrepancy requires further experimentation (indicated by a question mark in the picture). The same happens with respect to the IRE1α arm of the UPR where, on one hand, its activation is necessary for IL6 synergy with IL4 and IL13 to enhance anti-inflammatory macrophage polarization and, on the other hand, its increased activation in adipose tissue macrophages (ATMs) and in bone marrow-derived macrophages (BMMs) leads to a shift to a proinflammatory state. Typical features of proinflammatory macrophages, such as TLR signaling, inflammasome activation and proinflammatory cytokines expression, are enhanced by the activation of the three arms of the UPR: IRE1α and ATF6 synergize with TLR signaling, IRE1α is involved in the increase in IL1β through GSK3β activation, and ATF4 directly binds to the IL6 promoter. Finally, metabolic disorders such as obesity, fatty liver and atherosclerosis induce an increase of the IRE1α and PERK pathways of the UPR in macrophages, highlighting the link between a metabolic imbalance and maladaptive macrophage polarization states.

Molecular Mechanisms in Hypoxia

Aerobic organisms need oxygen for mitochondrial respiration and a wide range of other metabolic processes, for which reason sophisticated mechanisms for sensing changes in oxygen tension have evolved to modulate wide-ranging aspects of cell behavior. Hypoxia is a condition in which the oxygen supply is reduced relative to demand, causing oxygen levels to drop below 1%. Although it is mainly associated with pathological situations such as cancer, cardiovascular disease, dementia and diabetes, it can also be part of normal physiology, for example, during fetal development (38). During hypoxia, cells experience a reduction in mitochondrial ATP production due to lack of oxygen, which is the terminal electron acceptor in the mitochondrial electron transport chain (ETC) required to generate energy by oxidative phosphorylation (OXPHOS). Under these circumstances, cells switch their metabolism to glycolysis to obtain energy in an oxygen-independent manner, although this is at the cost of a poorer ATP yield (39). In addition, hypoxic cells are threatened by the accumulation of excessive reactive oxygen species (ROS) arising from the mitochondria (40). Thus, to survive at low levels of oxygen, cells must reduce energy consumption, temporarily arrest the cell cycle, upregulate antioxidant mechanisms to decrease ROS excess, and secrete survival and proangiogenic factors (41–44). Cells manage to develop all these adaptations in ways that are essentially mediated by hypoxia-inducible factors (HIFs) (41) and by inducing pathways that result in the general slowdown of processes involving oxygen and energy consumption, which mainly involve the UPR (45) and mTOR (46) pathways. These pathways synergize to activate common downstream targets and may also influence one another, such as in the case of mTOR and HIF-1α mutual regulation (47).

The Impact of Hypoxia on Macrophage Phenotype-Shaping

Macrophages are confronted by differential oxygenation levels under several pathological conditions. Hypoxic areas can be found in inflamed, infected and ischemic tissues, as well as in regions where local vessel growth cannot maintain an adequate blood supply due to cellular overgrowth, as occurs in adipose and malignant tissues (67). In these varied settings, macrophages must adjust their metabolic requirements to generate energy in an oxygen-independent fashion. Numerous studies have demonstrated that proinflammatory macrophages are highly dependent on glycolysis and that anti-inflammatory and pro-homeostatic macrophages show a stronger preference for mitochondrial OXPHOS to generate ATP (68). Hypoxia is a master driver of glycolysis, since oxygen deficit results in limited OXPHOS and cells rely on glycolysis to generate ATP. HIF-1α is fundamental to this process, which induces the expression of glycolytic enzymes such as hexokinase II (HKII) (69), phosphofructokinase (PFKFB3) (70) and glucose transporters such as GLUT1 (71). HIFs can also upregulate the expression of the enzyme pyruvate dehydrogenase kinase (PDK), which in turn provokes a drop in the conversion rate of pyruvate to acetyl CoA by the action of the enzyme pyruvate dehydrogenase (PDH), which prevents the entry of glucose into the TCA cycle (72). In this way, hypoxia-mediated activation of HIFs increases the glucose/carbon flux to the glycolytic route, but minimizes the flux through the TCA cycle and OXPHOS. Other oxygen sensors, such as prolyl-hydroxylases (PHDs) upstream of HIF, also directly contribute to the glycolytic reprogramming of macrophages. To take a specific example, the PHD2 isoform has been found to be important for regulating the glycolytic switch by altering PDK1 activity, as demonstrated by specific PHD2 deletion in the macrophage lineage in vivo (73). The shutdown of the TCA cycle under hypoxic conditions due to the lack of oxygen and to the effects of NO on the mitochondrial ETC causes intermediates, such as citrate and succinate, to accumulate. Citrate is then exported from the mitochondria into the cytosol, where it is converted to acetyl-CoA for use in several biosynthetic pathways and in the formation of microbicide and proinflammatory molecules. In the case of the metabolite succinate, this can be further oxidized to generate ROS and to increase the activity of HIF-1α (74–77). All these metabolic adaptations allow proinflammatory macrophages to develop their functions in the inflamed hostile microenvironment (Figure 2). Glycolysis can also take place in the presence of oxygen, a process known as the Warburg effect. This phenomenon was first detected in tumor cells, but was soon also observed in macrophages, since glycolysis serves not only as an adaptation to the hypoxic microenvironment, but also as a rapid means of obtaining energy, macromolecules for the catabolic cellular pathways, and of reducing equivalents such as ROS and NO, which are required for bactericidal activity (78).

Remarkably, HIF-1α activation can occur under normoxia, which allows the initiation of an inflammatory response before the tissue becomes hypoxic. Stimulation of macrophages with LPS increased HIF-1α protein levels and the formation of the functional HIF-1 complex (79). The induction of HIF-1α in the context of inflammation was soon found to depend on the presence of NF-κB (80). A relationship between HIF-1α glycolytic reprogramming and macrophage migratory activity through the HIF-1α/PDK1 axis was also discovered (81). HIF-1α is also recruited to the CXCR4 promoter, which mediates the chemotactic responses to its ligand CXCL12, which is expressed in hypoxic environments (82). Conversely, M2 macrophages do not need this rapid switch to glycolysis and obtain much of their energy from fatty acid oxidation and oxidative metabolism, which can be sustained for longer periods. This is consistent with their functional roles, as they appear later in the inflammatory response during the resolution phase and fulfill longer-term functions such as angiogenesis and extracellular matrix remodeling. HIF-1α and HIF-2α isoforms are differentially expressed in pro- and anti-inflammatory macrophages, respectively, according to their induction by Th1 and Th2 cytokines and by their regulation of inducible NOS2 (“M1-like”) and arginase 1 (“M2-like”) (83). Collectively, these findings demonstrate that metabolic adaptation is central to the polarization and functional activity of macrophages during hypoxia (Figure 2).

In conjunction with these metabolic changes, hypoxia is also involved in several other macrophage functions such as inflammasome activation, phagocytosis, antigen presentation and trained immunity. In hypoxic microenvironments IL-1β activity is enhanced through an increase in the synthesis of pro-IL1β and its processing enzyme, caspase-1, in LPS-stimulated human macrophages (84). Moreover, hypoxia boosts phagocytosis in macrophages via the activation of a p38 MAP kinase-HIF-1α link (85) and IFNγ-dependent upregulated expression of phagocytic receptors (86). Related with this process, hypoxic macrophages display an enhanced antigen-presentation capacity through the upregulation of costimulatory and antigen-presenting receptor expression (86, 87) (Figure 2). Innate immune cells such as macrophages have recently been found to be able to mount a memory response by a process known as trained immunity. This mechanism relies on major epigenetic changes upon initial stimulation that make the cells more responsive to subsequent stimuli (88). Induction of trained immunity is accompanied by significant changes in cellular metabolism. Indeed, the accumulation of several TCA metabolites, such as succinate and fumarate, could modulate the epigenetic landscape required for priming (89, 90). Elevated glycolysis is a central feature of this phenomenon, which, in turn, is dependent on the activation of mTOR through the dectin-1–Akt–mTOR–HIF-1α pathway (10). Oxygen may also directly regulate epigenetic enzymes such as histone demethylases (91), which are important in macrophage physiology (92). However, the exact link between these enzymes and oxygen deprivation in the context of trained immunity or macrophage polarization has not yet been assessed. On the other hand, TCA-intermediate metabolites could in turn influence HIF-1α stabilization in macrophages through their influence on the activity of hydroxylases such as succinate, which inhibits PHDs (77). It is of particular note that glycolysis-derived metabolites like lactate are known HIF-1α stabilizers (93). Specifically, lactate is an important mediator of VEGF expression in macrophages (94), where it induces Arg1 and other anti-inflammatory genes in tumor associated macrophages (TAMs) in an HIF-1α dependent way. This contributes to the development of a pro-homeostatic programme from which cancer cells may benefit to promote their growth (95).

Metabolic and Signaling Crosstalk Between UPR and Hypoxia

It is important to note that although the aforementioned oxygen-sensitive signaling pathways are activated independently, there is growing evidence that they may also act in an integrated way, influencing each other and common downstream targets that affect gene expression, metabolism and cell survival.

Involvement of the UPR and Hypoxia Pathways in the Maladaptive Polarization of Macrophages in Pathology and Chronic Inflammation

Concluding Remarks and Perspectives

Although much work is still required to fully understand the metabolic molecular cues that drive macrophage polarization, compelling evidence summarized in this review suggests that the roles of hypoxia and the UPR are much more important than previously thought. An imbalance in the UPR components occurs in several pathological and inflammatory disorders, placing macrophages in maladaptive polarization states in which their metabolic balance is compromised. The contribution of hypoxia to the fine balance of the UPR regulatory arms is being explored, and it has been demonstrated that UPR activation influences hypoxia tolerance. It is known that the hypoxic pathological microenvironment can direct macrophage metabolism toward proinflammatory polarization that worsens the inflammatory response. Hence, as this hypoxic environment is closely linked to the UPR, which, in turn, is critical to the integration of the metabolic and inflammatory signals in macrophages, exploring this relationship may yield tools for designing more effective immunotherapies and for developing new strategies for reshaping macrophage phenotypes in several clinical settings. Many studies discussed in this review used very general UPR inhibitors, so before developing new clinical tools, it would be necessary to investigate animal models with specific deletions of the UPR individual components in macrophages, or develop macrophage-specific drug delivery systems, such as liposomes and nanoparticles. There are no clinical trials currently underway using specific inhibitors of the UPR pathways as treatment (just pan inhibitors such as the chemical chaperones TUDCA and PBA). In the case of HIF-1α, several clinical studies in cancer have been carried out, but the lack of selectivity of the inhibitors used increases the difficulty of correlating molecular and clinical responses in patients (165). Moreover, myeloid HIF-1α and HIF-2α may have non-redundant or even opposing functions in myeloid cell physiology (9). As such, discerning the roles of each isoform individually and developing selective isoform inhibitors should be a priority. Nevertheless, phenotypic conversion of macrophages offers attractive therapeutic targets and stress and oxygen-sensing pathways are clearly key determinants of their physiological and pathological functions. Moreover, outstanding questions regarding macrophage phenotypic conversion still need to be addressed and a close characterization of macrophage diversity within specific microenvironments must be further explored to elucidate the specificity of cellular metabolism of macrophages in each disease, for instance by using emerging sophisticated technologies such as single-cell analysis, multiplex immunohistochemistry and mass cytometry. Many questions arise concerning the clinical contexts in which UPR and/or HIF inhibitors are most likely to be therapeutically effective. Clinical translation will be challenging: how will it be possible to target these molecules in a myeloid-specific way, and how efficient will those therapies be? Future studies on these aspects will allow new treatment regimens to be developed for the broad spectrum of the aforementioned pathologies.

Author Contributions

PD-B, MS, and RR designed and wrote the review. RR and CL-L coordinated and edited the manuscript. All authors contributed to the manuscript revision, and read and approved the submitted version.

Footnotes

References