The guardians of pulmonary harmony: alveolar macrophages orchestrating the symphony of lung inflammation and tissue homeostasis

The origins of tissue-resident alveolar macrophages

The origins of alveolar macrophages have been traced to different macrophage precursors that are produced in independent waves during development. The initial phase consists of yolk sac-derived macrophages, which originate from embryonic age 8.5–9.0 (E8.5–9.0) in mice. The subsequent phase encompasses fetal liver monocytes that are formed in the fetal liver starting from E12.5 onwards [8–10]. It is now recognised that they are established during embryonic development, achieving full occupancy of the alveolar space by day 3 post-natally. They exhibit a stable phenotype persisting throughout the lifespan, governed by instructive cytokines [8] and metabolites. These regulatory elements prompt the activation of specific transcription factors, thus imprinting a tissue-specific function on macrophages [11, 12]. They continue to exist in adulthood, while the involvement of bone marrow precursors (blood monocytes) is not substantial during homeostasis [10, 13]. Following birth, they mature into long-lived cluster of differentiation (CD)11c^hi, sialic acid binding Ig-like lectin F (SiglecF)^hi tissue-resident alveolar macrophages (TR-AMs) in response to granulocyte–macrophage colony-stimulating factor (GM-CSF) [14, 15] released by the alveolar epithelium and providing lung-specific tissue identity [15, 16]. Once situated there, they continue to reside and self-renew locally throughout an individual's life [8, 10, 13, 17]. Regarding the distinct types of macrophages in the lung, those present in the alveoli under homeostatic conditions (at least partially recovered by bronchoalveolar lavage (BAL)) [18] are referred to as TR-AMs. Their distinct transcriptional signature in mice is characterised by the expression of Ear1, Pparg, Ear1, Chil3, Fabp4/5 and the surface markers F4/80, tyrosine-protein kinase Mer (MerTK), CD64, CD206, CD11c and SiglecF, while lacking CD11b expression [19–23]. Whereas interstitial macrophages reside in the lung parenchyma and express CD64, CD11b and CD11c, they do not display SiglecF expression [24, 25]. While exhibiting a comparable abundance [26] to alveolar macrophages, interstitial macrophages do not constitute the primary investigative focus of the subsequent sections in this review.

In humans, the profile of TR-AM-expressed genes includes Fabp4, Serping1 and Apoc1, while markers at the cell surface are human leukocyte antigen-D-related, CD206, CD169, CD64 and macrophage receptor with collagenous structure (MARCO) [23, 27, 28]. TR-AMs locate to and closely interact with both alveolar type I and type II cells [29], forming a functional unit characterised by an intricate cellular cross-talk relying on GM-CSF [16], CD200R–CD200 interaction and autocrine macrophage transforming growth factor-β (TGF-β) signalling in homeostatic conditions. In this condition, TR-AMs downregulate the expression of CD11b, thereby fine-tuning their phagocytic activity, which is an important function together with the efficient removal of pulmonary surfactant, for optimal lung gas exchange [12, 30–34]. The development of TR-AMs and acquisition of their lung phenotype depends on co-stimulation by GM-CSF and TGF-β, leading to the activation of the transcription factor PPARγ (peroxisome proliferator-activated receptor gamma), a key regulator of lipid metabolism [15, 35–37]. PPARγ-knockout TR-AMs fail to develop into mature alveolar macrophages [35, 37].

Disruption of biological harmony: loss of TR-AM homeostasis and repopulation dynamics

Following an insult, infiltrating monocytes undergo programming to adopt transcriptional programmes characteristic of TR-AMs via the intermediate state of bone marrow-derived macrophages (BMDMs), which specifically denotes alveolar macrophages generated in vivo from bone marrow progenitors [19]. TR-AMs are depleted to varying extents after injury, with BMDMs repopulating the empty niche by receiving tissue signals with respect to location and frequency through processes that are not well understood.

TR-AMs serve as a frontline defence against invading pathogens, detecting danger signals such as PAMPs (pathogen-associated molecular patterns) commonly found in various microorganisms. This recognition, coupled with the release of pro-inflammatory mediators in response to injury, triggers the activation of the macrophage response [38]. Following activation, they immediately produce important mediators that contribute to inflammation, such as type I interferon (IFN), interleukin (IL)-1, IL-6 and tumour necrosis factor-α (TNF-α) [39]. In the initial stages of inflammation, leukocytes, including neutrophils, infiltrate lung tissue [38, 40]. This recruitment process is guided by chemotactic signals locally generated by TR-AMs [41], incoming cells and cytokine-stimulated lung parenchyma. These signals play a crucial role in directing leukocyte migration towards the site of inflammation, where they participate in immune responses, i.e. combating infection within the lung parenchyma.

Intense inflammation or a severe infection frequently results in the depletion of TR-AMs residing in tissue [25], termed the “macrophage disappearance reaction” [20] (figure 1c). When sterile inflammation is induced in mice by introducing lipopolysaccharide (LPS), there is a temporary decrease in TR-AMs within the tissue and their replenishment during the resolution of inflammation appears to primarily depend on local proliferation [42]. However, when a more injurious, highly inflammatory challenge occurs, as observed after administration of bleomycin to model lung fibrosis or after infection with influenza virus, this leads to TR-AM depletion with C–C chemokine receptor (CCR) 2-dependent replenishment from lymphocyte antigen 6 family member C (Ly6C)^hi circulating monocytes via the BMDM intermediate [43, 44]. Upon severe viral infection, the TR-AM niche becomes predominantly re-populated by monocytes [45–47]. Accordingly, significant TR-AM depletion occurs during human infections, as documented recently for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection [45, 48–50]. The precise mechanism behind TR-AM depletion remains elusive, prompting the question of whether they undergo direct infection-related cell death, are challenged by pro-apoptotic ligands released from other leukocytes or if their disappearance results from the loss of a survival signal within the alveolar epithelial–TR-AM functional unit upon epithelial cell injury. So far, the reason for this depletion reaction, as well as the signals indicating an empty niche, has not been clarified. It was speculated that the efferocytic role of TR-AMs, fostering their pro-resolution, anti-inflammatory phenotype, may prove harmful in the initial states of infections where pathogen containment is their main task. Therefore, their temporary disappearance could be advantageous in facilitating effective host defence through recruitment of more pro-inflammatory BMDMs [51]. Given that direct refilling of the alveolar niche is crucial to maintaining lung tissue integrity and gas exchange, immediate repopulation relies on different programmes: either on local proliferation of the remaining terminally differentiated TR-AMs [52–54] or on the differentiation of recruited Ly6C^hi monocyte precursors via BMDM intermediates (figure 1b) [16, 44, 55–57]. It is speculated that the extent of the contribution of circulating precursors versus local TR-AM proliferation likely depends on the extent of initial TR-AM depletion alongside the severity of inflammation; however, the signals that fine-tune this balance have not been elucidated. Despite a similar surface marker expression, BMDMs exhibit very distinct transcriptional, epigenetic, metabolic and functional phenotypes as compared to fetal monocyte-derived TR-AMs and remain distinct for months after engraftment into the TR-AM pool [58–61].

Functional phenotypes in the context of macrophage origin

Recent technological advances, such as single-cell transcriptomics, genetic models and fate-mapping, have unveiled significant variations and differences among macrophages. These variations exist even within the same tissue when considering the alveolar macrophage phenotype, their (epi)genetic activity and metabolic profiles [62]. Alveolar macrophages exhibit a high degree of adaptability, displaying various characteristics and performing different functional roles. These traits are influenced by both their source of origin [58] and the specific microenvironment [63] of the tissue in which they reside. As to the role of origin, it is well known that TR-AMs typically exhibit restricted plasticity compared to BMDMs [64, 65]. Cell-transfer experiments using different macrophage precursors showed that tissue imprinting prevails over origin to determine their cellular identity [66]. Conversely, cell ontology rather than tissue imprinting seems to determine the overall plasticity of the cell, with BMDMs typically exhibiting increased plasticity compared to TR-AMs [64, 65]. This is manifested as a higher immunoreactivity towards respiratory insults in BMDMs. Concomitantly, it was suggested that long-lasting steady-state tissue imprinting is what restricts TR-AM plasticity and, over time, BMDM-derived TR-AMs gradually lose plasticity [66]. Recent data revealed that BMDM-derived TR-AMs outcompete remaining TR-AMs of fetal monocyte origin after influenza virus infection due to a higher proliferative potential, thereby inducing a state of increased immunoreactivity [58]. Furthermore, Hou et al. [59], showed that BMDMs exhibit heightened susceptibility to environmental cues compared to TR-AMs, contributing to cytokine storms and immune dysregulation in severe infections. These data suggest that origin rather than factors such as epigenetic memory determines their responses to respiratory viral infections in the long term [58, 67]. Prolonged tissue residency, however, maintains a state of relative hyporesponsiveness [68] and limits macrophage plasticity, favouring evolutionary advantages in maintaining tissue homeostasis [66].

As to their pro-inflammatory capacity, BMDMs exhibit a notable distinction from fetal monocyte-derived TR-AMs [69], evidenced by a tendency to produce elevated levels of pro-inflammatory cytokines such as TNF-α, inducible nitric oxide synthase-derived NO and IFN-β playing a role in promoting alveolar injury associated with influenza A virus infection [69–71], LPS-induced injury [72] or bacterial infections [73]. BMDMs may then gradually transition to an anti-inflammatory phenotype within days after the initial inflammation in a sterile injury model [74], promoting cell proliferation and tissue repair to establish tissue homeostasis [69, 74]. Similarly, pro-inflammatory/tissue-damaging BMDMs gradually acquire features of pro-regenerative cells along their differentiation trajectory towards TR-AMs in influenza infection (figure 1b) [55]. This has also been observed in other inflammatory contexts, including bleomycin-induced lung injury [56] or mouse house dust mite-induced asthma [56, 75].

The question arises as to the mechanisms linking the ontogeny of macrophages to their functional properties and the capacity for functional plasticity, pivotal for elucidating macrophage biology and its dynamic response in various physiological contexts. Examination of the gene-expression patterns in monocytes recruited to inflamed areas indicates that these cells instigate a macrophage differentiation process early on [19]. As monocytes differentiate into BMDMs during acute or chronic inflammation, they retain open chromatin regions relative to their monocyte origin for an extended period, providing a distinct epigenetic legacy resulting from the differentiation programme that might be associated with increased responsiveness to microenvironmental stimuli [44, 76]. This is distinct from epigenetic imprinting called “trained immunity” that defines a process where an initial insult leaves a lasting memory on innate leukocytes, leading to faster and enhanced cellular responses upon repeated exposure to further stimuli [77, 78].

Within the alveolar niche, macrophages of different origin are further influenced by the tissue environment, whether in homeostasis or in inflammation. These multifaceted signals arise from the alveolar epithelium and are further shaped by acute or chronic inflammatory stimuli such as damage-associated molecular patterns, PAMPs and cytokines, toxic agents, microbiome composition, and nutrient availability and composition [79, 80]. Particularly noteworthy is the pivotal role of alveolar macrophage efferocytosis within an inflamed milieu or under homeostasis, a process vital for efficiently clearing dying or dead cells within the lung [15, 69], promoting alveolar macrophage anti-inflammatory and resolution or tissue homeostasis phenotypes [81, 82]. To comprehend the complexity of macrophage function, an understanding of the integration of these diverse signals at different layers of phenotype regulation is imperative [25].

The intricate interplay between macrophage origin [58] and the dynamic lung microenvironment [63] represents a current area of scientific inquiry characterised by unresolved complexities. While recent advancements have revealed significant heterogeneity among macrophage populations, encompassing variations in phenotype, genetic activity and metabolic profiles, the precise mechanisms linking origin to function and functional plasticity remain incompletely elucidated. Key questions persist regarding the nuanced interplay between monocyte differentiation trajectories, epigenetic programming and the diverse signalling milieu within the alveolar niche. Moreover, the intricate balance between tissue-specific imprinting and environmental stimuli in shaping macrophage functionality underscores the need for further investigation. As such, addressing the unresolved aspects of macrophage origin and function represents a critical frontier in pulmonary immunology and tissue homeostasis research.

Navigating functionality through metabolic mastery

Accumulating evidence underscores the importance of metabolic pathways and intermediates in functional polarisation of macrophages [83]. To date, most of the data available was generated in BMDMs, where glycolysis clearly supports inflammatory responses, while oxidative phosphorylation (OXPHOS) is associated with an anti-inflammatory phenotype [84, 85]. In this respect, data for TR-AMs is comparatively limited. TR-AMs inhabit an environment characterised by high oxygen and low glucose concentrations. Inflammation and hyperglycaemia increase glucose concentrations in the airspace, which promotes bacterial infections [86]. OXPHOS seems to be vital for TR-AMs in health and disease. During homeostasis, alveolar macrophages require OXPHOS to handle lipid contents and a defective OXPHOS reduces alveolar macrophage numbers and leads to pulmonary alveolar proteinosis [34]. Woods et al. [87] demonstrated the reliance of TR-AMs on OXPHOS and not glycolysis for inflammatory responses. In contrast, infiltrating monocytes during influenza virus infection clearly favour glycolysis to promote inflammation [87]. Similarly, human alveolar macrophages require OXPHOS and not glycolysis to generate inflammatory cytokines, while monocyte-derived macrophages rely on glycolysis at least in response to LPS [88]. Interestingly, impaired glycolysis in alveolar macrophages was shown to limit type-2 responses induced by IL-4, which depends on the lung environment [89]. Remarkably, hypoxia activates hypoxia-inducible factor 1-α (HIF-1α) to induce glycolysis, which promotes inflammatory responses and correlates with improved survival of TR-AMs during lung injury [90]. Alveolar macrophages from patients with COPD exhibit impaired OXPHOS and compared to “healthy smokers”, a dysfunctional ability to mount glycolysis, which might affect antimicrobial responses [75]. These findings stress the importance of studying the immunometabolic functions of TR-AMs during pathologic conditions, as feeding pathways and functions are highly context-dependent. The observation that human alveolar macrophages from the upper (harvested from sputum) or lower airways (harvested by BAL) are metabolically distinct, adds another layer of complexity [91]. During tissue regeneration, genes associated with glycolysis are downregulated, while genes participating in mitochondrial OXPHOS metabolism are upregulated [74, 92, 93]. Most of the data in this context was generated using BMDMs. Future research on the immunometabolic regulation of TR-AMs during resolution of inflammation will shed light on potential tissue and disease-specific differences. The therapeutic potential of targeting the mitochondrial metabolism was shown by Xian et al. [94], who demonstrated that metformin improves outcomes in lung injury. Mechanistically, metformin inhibits complex I of the electron transport chain, ultimately limiting NLR family pyrin domain containing 3 inflammasome activation by reducing the generation of oxidised mitochondrial DNA [94, 95]. Overall, we are just beginning to understand the immunometabolic properties of TR-AMs. Already known and yet-to-be-identified key metabolic regulations might pave the way for macrophage-targeted therapies in lung inflammation [96].

Alveolar macrophages: architects of tissue repair

During acute and chronic inflammation, the alveolar epithelium is damaged and needs to receive stimuli for regeneration and repair of the epithelial barrier. Such signals drive immediate repair mechanisms to restore the alveolar barrier or drive endogenous stem-cell proliferation and differentiation processes. For example, upon severe viral infections, but not limited to it, IFN-dependent expression of TRAIL (TNF-related apoptosis-inducing ligand) causes apoptosis of infected and noninfected alveolar epithelial cells (AECs) [70, 71]. In addition, IFN/TRAIL pathway activation impairs the resolution of alveolar oedema-limiting lung gas exchange function [97]. Concomitant to inflammation, alveolar macrophages foster wound healing by stimulating cell recruitment, the proliferation of fibroblasts and angiogenesis to limit tissue damage. In the best of cases, following resolution of inflammation, a successful repair process reaches full restoration of the tissue architecture and function, with no remaining fibrosis [98]. Besides the recognised and well-documented protagonism of lung macrophages driving the initiation and resolution of the inflammatory process, in recent years, the macrophage field experienced an emergence of data supporting that these cells are contributors and orchestrators of tissue regeneration.

A precise time-resolved message delivered by the tissue can switch macrophages towards a pro-repair phenotype in sterile acute lung injury models. Among those signals, epithelial expression of IL-33 drives an “alternatively activated” phenotype in ST2⁺ alveolar BMDMs that actively promote tissue regeneration upon lung injury [99]. The pro-repair phenotype of recruited BMDMs also depends on the production of IL-4 and IL-13 mainly by group 2 innate lymphoid cells, that signals through IL4rα in macrophages [100]. In a model of lung injury triggered by helminth infection, the IL-4/IL-13/IL-4rαa pathway needed to be supported by the concomitant recognition of apoptotic cells through the receptors Axl and MerTK to effectively induce a repair phenotype in lung macrophages that contribute to wound healing and reducing haemorrhage in BAL fluid [101]. In addition, metabolites such as ATP secreted from apoptotic epithelial cells via caspase-mediated opening of pannexin1 channels can mediate epithelial–macrophage crosstalk to induce lung regeneration by promoting epithelial cell proliferation after injury [102].

It has been long known that, during acute lung injury, alveolar macrophages express growth factors that can contribute to epithelial proliferation in lungs [103–105]. Indeed, human alveolar macrophages can promote AEC growth by expression of platelet-derived growth factor (PDGF), insulin-like growth factor 1 and fibroblast-derived growth factor [106]. After pneumonectomy in mice, TR-AMs rapidly expand and stimulate the enlargement of the remaining lung by enhancing angiogenesis through the expression of vascular endothelial grow factor, PDGF and epiregulin, which belongs to the epidermal growth factor family [107]. Later, it was shown that this effect depends on the recruitment of CCR2⁺ monocytes and the activation of an IL-13/IL-4r signalling pathway within the hematopoietic compartment to induce alternative activation of macrophages and a pro-repair phenotype [100]. Although it is commonly observed that the pro-repair profile of macrophages correlates with a shift in phenotype, we demonstrated that TNF-α derived from early-recruited inflammatory macrophages can promote AEC production of GM-CSF to induce epithelial proliferation in an autocrine fashion [108]. These data suggest that early pro-inflammatory signals may be immediately translated to evoke epithelial responses that support the macrophage–epithelial unit important for maintaining or regaining alveolar homeostasis.

Production of epiregulin and amphiregulin (figure 1b) is characteristic of a pro-repair macrophage phenotype [107, 109]. Particularly, amphiregulin secretion from lung macrophages, either tissue-resident or bone marrow-derived, assists lung tissue repair by inducing TGF-β activation in pericytes that promote vascular network reconstitution, improving lung function and wound healing [109]. It was additionally proposed that macrophage derived-amphiregulin protects epithelial integrity by acting in an epithelial growth factor receptor dependent way [110]. Alternatively, lung macrophages can mediate lung tissue regeneration in an IL-4/IL-13/IL-4rα-independent fashion by the expression of low molecular weight cytokines, named trefoil factor family (TFF) molecules (TFF1, TFF2, TFF3), which induce a prompt movement of epithelia over a denuded basement membrane. The TFF effect relies on the induction of Wnt4 and Wnt16 gene expression and activation of the noncanonical Wnt pathway [111]. We have reported a novel regeneration mechanism in which placenta-expressed transcript 1 (Plet1) stimulates the proliferation of alveolar epithelial type II and of further epithelial progenitor cells, reconstituting a functional macrophage–epithelial alveolar unit, as well as re-establishing a tight lung barrier, after severe viral pneumonia. We observed that Plet1 expression in BMDMs indicates a transition towards a pro-repair and homeostatic TR-AM phenotype, and alveolar Plet1 deposition could represent a novel therapeutic strategy to protect and repair the epithelial barrier [55].

Besides the repair of acutely injured alveolar tissue, ameliorating declining lung function in ageing and progressive or end-stage lung disease represents an unmet clinical need. Studying macrophage-driven pro-regenerative pathways in chronic lung disease and ageing has therefore come into focus. However, pro-repair macrophage phenotypes share activation patterns often linked to fibrotic processes [112].

Repair gone awry: macrophages in fibrosis

Fibrosis is an aberrant fibrogenic response resulting from a tissue repairing defect together with an unresolved inflammatory process often due to repetitive injuries, with acquisition of mesenchymal cell features in parenchymal cells, maintenance of reprogrammed mesenchymal cells that escape apoptosis, extensive matrix production and gradual loss of organ function. Generally, a timely and spatially limited process of matrix deposition, mesenchymal cell activation and proliferation is characteristic of and indispensable for a successful tissue repair response, and self-limiting. Given that macrophages are key effectors in such responses, aberrant fibrogenic processes with extensive scarring have been linked with aberrant macrophage responses, for example in idiopathic pulmonary fibrosis (IPF) [43, 113–115]. The proliferation of TR-AMs and the recruitment of CCR2⁺ BMDMs (figure 1d) are responsible for the augmented macrophage numbers in pulmonary fibrosis [116]. Still, most reports underline the participation of BMDMs in fibrotic tissue damage by adding inflammation, but also evolve towards a macrophage-intrinsic programme that directly drives the fibrotic response [43]. The first evidence resulted from the observation that CCR2^−/− mice, which fail to recruit monocytes/BMDMs to inflamed tissues, presented less fibrosis in response to bleomycin [117]. This was further supported by studies where depletion of monocytes in several mouse models of fibrosis resulted in its attenuation [113]. More recently, using a lineage tracing system, specific depletion of BMDMs, but not of TR-AMs, dampened fibrogenesis in response to bleomycin [43], Macrophages and fibroblasts are capable of forming cell–cell circuits that enhance survival through the cross-expression of growth factors. This symbiotic interaction yields a remarkably stable relationship that is resilient to perturbations, potentially facilitating the progression of fibrosis and conferring resistance to treatment [118]. Such dynamics may elucidate the limited efficacy of corticosteroids in halting the advancement of pulmonary fibrosis, particularly among IPF patients. Nevertheless, the prevailing understanding posits that pro-inflammatory cells might ultimately amplify the activities driven by pro-fibrogenic macrophages, in conjunction with fibrogenic growth factors localised within the microenvironment [119]. In support of this hypothesis, it was reported that a subtype of pro-fibrotic monocytes specifically develops from granulocyte/macrophage progenitors, under the control of CCAAT/enhancer binding protein β (C/EBPβ). Although deletion of the C/ebpβ gene in mice prevented bleomycin-induced fibrosis but not inflammation, it remains unclear whether the antifibrotic effects of gene deletion relate to a direct fibrotic effect or a determinant role of C/ebpβ in monocyte-to-macrophage differentiation [120]. Using cell-specific lineage tracing systems in gene-deletion experiments will help to clarify how macrophage-expressed genes contribute to fibrosis. Ultimately, the inhibition of monocyte recruitment decreases inflammation but, most importantly, prevents accumulation of pro-fibrotic macrophages.

Pro-fibrotic alveolar macrophages were identified in several pre-clinical models and in patients [22, 115, 121, 122]. The pro-fibrotic phenotype of alveolar macrophages in pulmonary fibrosis is characterised by the expression of TGF-β and PDGFs, chitinase-3-like protein 1–chemokine ligand (CCL) 18, osteopontin (spp1) and IL-1β, all of which promote myofibroblast activation [115, 123, 124]. Deep phenotyping by transcriptomics of disease-associated pro-fibrotic alveolar macrophages that localise in the fibrotic niche suggests a transitional gene expression profile between monocyte-derived BMDMs and TR-AMs [121, 124]. In this sense, McCowan et al. [125] recently showed that the resolution of bleomycin-induced fibrosis entails the repopulation of the alveolar niche by BMDMs and further transition to resident alveolar macrophages, mediated by the transcription factor early growth response factor 2.

Enriched during pulmonary fibrosis in both patients and lungs of mice, pro-fibrotic alveolar macrophages promote collagen I production by interaction with mesenchymal cells. Their differentiation in vivo was driven by IL-17A, TGF-β and growth factors in chronic lung fibrosis [121]. Pro-fibrotic alveolar macrophages can also express inhibitors of collagenolytic enzymes such as tissue inhibitor of metalloproteinases 1 (TIMP1) and TIMP2, affecting collagen synthesis and degradation [116]. Moreover, Ucero et al. [126] reported Fra-2-dependent collagen IV production by alveolar macrophages. This Fra-2/collagen IV pathway is dispensable for BMDM recruitment but can activate myofibroblast in vitro, denoting an exclusively pro-fibrotic pathway contributing specifically to the fibrotic process [126].

HIF-1α is a stress-inducible transcription factor that functions as a master regulator of oxygen homeostasis and may play an important role in TGF-β1-driven fibrogenesis since silencing of HIF-1α expression markedly decreases TGF-β1 production in a TR-AM derived cell line (MH-S) and attenuates the development of bleomycin-induced fibrosis [127]. The metabolic pathways acting in fibrotic alveolar macrophages accordingly contribute to this phenotype. IPF alveolar macrophages present higher glycolytic activity, which supports the production of the reactive oxygen species associated with fibrosis. This is accompanied by a decreased expression of cis-aconitate decarboxylase and the consequent defect in itaconate production that was reported to have antifibrotic properties [128]. In addition, itaconate exerts a modulatory function over the production of another tricarboxylic acid cycle metabolite, succinate, that instead acts in a pro-inflammatory fashion via the IL-1β–HIF-1α axis [129].

A novel mechanism, described by Wang et al. [130], links impaired epithelial regeneration to the recruitment of pro-fibrotic macrophages. IPF patients present an accumulation of transitional AECs (alveolar differentiation intermediates (ADIs)) that are in between the differentiation trajectory from AEC II to AEC I. Single-cell RNA sequencing analysis showed that ADIs stuck in this transitional stage exhibit increased expression of defined keratins, particularly krt8 (keratin 8), as well as claudin 4, stratifin and TGF-β pathway genes such as integrin β6. Using a bleomycin lung fibrosis model, the authors showed that krt8 expression correlates with CCL2 production by ADIs, increasing the recruitment of BMDMs that acquire a pro-fibrotic chemokine expression phenotype (TGF-β and IL-1β) and, in turn, promote the accumulation of ADIs, configuring a positive feedback loop driving extensive fibroblast activation and fibrogenesis. Interestingly, when fibrosis resolves, this transitional AEC population ultimately differentiates into AEC I [130, 131]. In addition, macrophage colony-stimulating factor (M-CSF)/M-CSF receptor (M-CSFR) signalling could also be a pathway involved in the maintenance of pro-fibrotic BMDM in both mice models and human pulmonary fibrosis [116].

A pro-fibrotic BMDM subset was also described in BAL fluid of patients with coronavirus disease 2019 (COVID-19) acute respiratory distress syndrome and lung fibrosis. These cells are characterised by expression of CD163, CD206, lymphatic vessel endothelial hyaluronan receptor 1 and complement component 1q, and the fibrosis-promoting genes legumain, secreted phosphoprotein 1 and TGF-β. Gene set enrichment analysis showed significant similarities between the pro-fibrotic alveolar macrophages observed in COVID-19 and IPF. Contrary to conventional beliefs within the context of IPF, the mechanism responsible for the pro-fibrotic phenotype of macrophages in COVID-19 appears to be directly driven by SARS-CoV-2 macrophage infection; however, the intracellular viral gene products mediating this very specific response remain to be defined [49].

Since alveolar macrophages clearly contribute to fibrogenesis, macrophage-centred strategies aiming to dampen lung fibrosis are being continuously explored [132–134]. Pro-fibrotic alveolar macrophages are frequently characterised by increased mannose receptor expression (CD206) and suppression of CD206⁺ macrophage differentiation reduced fibrosis in mice [22, 135]. These findings prompted the development of mannosylated albumin nanoparticles incorporating TGF-β small-interfering RNA (TGF-β1-siRNA MANPs), which, when administered intravenously, were preferentially internalised by SiglecF⁺ CD11b⁺ monocyte-derived alveolar macrophages and reduced disease severity by promoting the regression of lung fibrosis induced by bleomycin in mice [136]. It was also reported in a bleomycin lung fibrosis model that pro-fibrotic BMDMs exhibit high expression of the folate receptor FRβ. This enabled the development of another novel therapeutic candidate by the linkage of folate to a Toll-like receptor (TLR) 7 agonist that demonstrated anti-fibrotic capacity on macrophages (FA-TLR7-54). The therapeutic administration of FA-TLR7-54 in a pulmonary fibrosis mouse model reprogrammed pro-fibrotic macrophages into fibrosis-suppressing macrophages, reducing pro-fibrotic cytokine release, hydroxyl proline biosynthesis and collagen deposition, and improving disease outcome without evidence of toxicity [137].

The aged macrophage

Ageing entails the inherent deterioration of cellular function throughout various tissues and organs. This is characterised by a diminished capacity to respond to different stimuli, a decline in tissue-regenerative capabilities and an elevated susceptibility to conditions such as cancer and cardiovascular and pulmonary diseases [138]. With the progression of ageing, several intrinsic cellular phenomena manifest, including epigenetic alterations, reduction of stem cell capacity and the induction of cellular senescence. The latter denotes a stress-induced state of cell-cycle arrest, concomitant with the secretion of inflammatory factors, recognised as the senescence-associated secretory phenotype (SASP). Cellular senescence ensues through replicative and stress-related pathways (figure 1d) involving the activation of p53 and p16^INK4a (p16). This activation subsequently triggers the induction of p21^CIP1 (p21), culminating in cell-cycle arrest [139]. Senescent cells, despite being viable, exhibit phenotypic and metabolic alterations [140], chromatin reorganisation, changes in gene expression, and stable arrest in proliferation that is unresponsive to mitogenic stimuli [141, 142]. Senescence prompts the growth arrest of damaged cells and, concurrently, the SASP attracts various immune cells that facilitate their clearance [143]. The generation of the SASP plays a role in establishing an enduring condition of prolonged and low-grade inflammation, commonly referred to as inflammaging. In this context, chemokine (C-X-C motif) ligand 9 has been identified as a biomarker [144]. Senescence also manifests in the immune system [138, 145]. Immunosenescence can be characterised as the cumulative effect of various age-related defects leading to immune disbalance. The outcomes involve compromised reactions to pathogens or vaccines, increased mortality and elevated susceptibilities in the elderly. Viral infections have the potential to worsen immune ageing through both direct and indirect mechanisms, resulting in heightened dysfunction and inflammation, especially in older individuals [146].

The lung is a complex organ whose manifold cell populations [147] are continuously exposed to environmental stimuli over their lifespan, additionally driving cellular senescence and ageing signatures [138]. Monocytes and alveolar macrophages exhibit the most significant transcriptional alterations [63, 138]. In this context, the ageing process coincides with a decline in the precision and effectiveness of both the innate and adaptive immune systems [148], leading to a diverse array of variations in both the phenotypic features and functional properties of BMDMs and TR-AMs. BMDM profiles demonstrate an increase in IL-10 expression. The accumulation of IL-10 during ageing is associated with immune suppression and the onset of immunosenescence [149]. The lung epithelium appears to be influenced by the ageing process, albeit to a lesser extent than BMDMs and TR-AMs. Alterations in transcriptional factors such as ETS variant transcription factor 5 and E1A binding protein P300, which play roles in promoting epithelial repair in bleomycin-induced lung injury [150], indicate a diminished reparative capacity in elderly individuals.

In the murine lung, the ageing process induces substantial alterations in the TR-AM population. These modifications manifest as a reduction in the macrophage quantity and investigations have unveiled numerous genes that display differential expression when comparing young and aged mouse lungs [151]. Boe et al. [152] reported that ageing did not result in significant alterations in the overall counts of interstitial or alveolar macrophages within the lung interstitium and parenchyma in mice. In fact, the composition of myeloid cells in the murine lung remained relatively stable as mice aged. Regarding the functional consequences of ageing, however, a recent report elegantly described that the ageing pulmonary microenvironment, particularly matrix components, drives macrophage ageing signatures, characterised by reduced proliferation responses and impaired responsiveness to GM-CSF, by using genetic lineage tracing with sequential injury, heterochronic adoptive transfer and parabiosis experiments [63].

Age-related alterations have far-reaching implications in macrophages, particularly affecting phagocytosis and the clearance of apoptotic neutrophils, which in turn leads to detrimental consequences for inflammatory responses [153]. Research involving aged mice has underscored a distinct decline in their wound-healing capabilities compared to their younger counterparts, primarily attributed to reduced macrophage phagocytic capacity. This age-related phagocytosis impairment is further exemplified by the diminished clearance of apoptotic cells in aged mice, resulting in unresolved and chronic inflammatory responses [154]. Importantly, this deficiency in efferocytosis, which perpetuates inflammation, has also been observed in human subjects [155]. As an example, in the elderly, macrophages display indications of decreased responsiveness to IFN-γ and an elevated baseline production of cytokines, specifically TNF-α and IL-6 [152, 156]. Noteworthy key molecules, including major histocompatibility complex-II, TLR 4, MARCO and CD206, exhibited increased expression. Increased expression of TNF-α and IL-6 exhibits a positive correlation with the increased expression of cyclooxygenase 2 (COX-2) and subsequent synthesis of prostaglandin 2 in aged macrophages when contrasted with their young counterparts [157]. Although the involvement of pathways such as COX-2 and p38 mitogen-activated protein kinase has been implicated in the modification of cytokine production in old mice and elderly individuals, a comprehensive investigation into the exact mechanisms driving these age-related changes is lacking. It is conceivable that, given the diverse functions of macrophages, the tissue microenvironment plays a significant role in shaping cytokine production. Alternatively, variations in macrophage metabolism and changes in phagocytic capabilities could equally contribute to these observed differences [158]. In the murine lung, the ageing process diminishes the capacity of alveolar macrophages to mitigate lung damage in the context of influenza virus infection. Additionally, ageing results in a reduction of alveolar macrophage phagocytosis of apoptotic neutrophils, a downregulation of the scavenging receptor CD206 and the retention of neutrophils during influenza infection [151]. The transcriptional profiles of TR-AMs from young and aged mice were notably distinct, with significant alterations observed in 3545 genes, including the marked downregulation of cell-cycle pathways [151]. Therefore, ageing triggers deficient phagocytosis by TR-AMs, contributing to an exacerbation of lung damage [146, 151]. Regarding immunometabolism in the ageing macrophage phenotype, NAD⁺ (nicotinamide adenine dinucleotide) has substantial age-related fluctuations [159]. It is observed that NAD⁺ synthesis diminishes with age, akin to changes occurring during immune responses. This age-related alteration in NAD⁺ levels exerts a discernible influence on macrophage responses, accentuating their pro-inflammatory functions [160, 161].

It is noteworthy to highlight an age-associated phenomenon known as clonal haematopoiesis of indeterminate potential (CHIP) [162]. The aggregation of sets of somatic mutations in haematopoietic stem cells over the course of a human lifespan represents a distinctly delineated phenomenon that becomes increasingly evident with advancing age [163, 164]. This phenomenon has been identified as a contributor to inflammation and, given the continuous exchange of TR-AMs by BMDMs, may significantly affect lung health and disease in the aged. CHIP contributes to the altered inflammatory response seen in chronic diseases related to inflammatory macrophage phenotypes, including cardiovascular disease [165], COPD, IPF [162, 166, 167] and atherosclerosis [168]. The findings indicate an augmented risk of both the incidence and severity of COPD in patients exhibiting CHIP, thereby emphasising the potential clinical implications. In these COPD patients, CHIP was predominantly characterised by DNA (cytosine-5)-methyltransferase 3A (DNMT3A) CHIP-mediated hypomethylation of phospholipase D family member 5, a condition associated with increased levels of pro-inflammatory cytokines and a decline in lung function [162]. Additionally, carriers of CHIP with mutations in DNMT3A, tet methylcytosine dioxygenase 2, additional sex combs-like 1 and Janus kinase 2 exhibited an almost twofold increase in the risk of coronary heart disease in humans and accelerated atherosclerosis in mice [165]. First approaches, aimed at replacing CHIP mutation-carrying resident macrophages in different tissues including the lung [169] with “healthy” bone marrow precursors to restore organ homeostasis, indicate potential for therapeutic targeting. However, research on the role of CHIP mutations in lung macrophages of different ontogeny driving chronic lung diseases is still in its infancy. Looking ahead, the acknowledgment of CHIP as a risk factor may pave the way for the development of targeted therapeutic interventions aimed at mitigating the impact of CHIP mutations, including macrophage-related pathomechanisms and altering the course of pulmonary and cardiovascular health in ageing populations.

Future directions

Despite accumulating knowledge on alveolar macrophage biology, many open questions remain. This includes, for example, how different levels of macrophage programming such as origin, epigenetic memory and local microenvironmental cues, together with metabolic constraints, are integrated and shape macrophage phenotype trajectories in human disease. Particularly, the potential role of processes associated with inflammaging and age-related somatic mutations and the role of macrophage immunosenescence in such processes need to be addressed. Identifying biomarkers and mechanisms of alveolar macrophage phenotype acquisition would allow us to enter new avenues for trait or outcome prediction and for the treatment of acute and chronic lung disease in humans. Putative targeted interventions involve actively reprogramming lung macrophages in a disease-, compartment- and disease course-specific manner to generate macrophage-based cell therapies and to use or target macrophage effector molecules to fine-tune host defence, inflammation resolution or lung repair processes [50, 96, 170].