Intestinal macrophages in resolving inflammation

Abstract

Macrophages not only regulate intestinal homeostasis by recognizing pathogens to control enteric infections, but also employ negative feedback mechanisms to prevent chronic inflammation. Hence, macrophages are intriguing targets for immune-mediated therapies, especially when barrier function in the gut is compromised to trigger aberrant inflammatory responses, most notably during inflammatory bowel diseases. Recently, there has been considerable progress in our understanding of human macrophage biology in different tissues, including the intestines. In this review, we discuss some new findings on the properties of distinct populations of intestinal macrophages, how resolution of inflammation and tissue repair by macrophages could be promoted by type-2 cytokines as well as other therapeutic interventions, and highlight some challenges for translating these findings into the future for this exciting area of immunology research.

Introduction

The human gastrointestinal tract is persistently exposed to a high antigenic load of microbial and dietary nature, and as a result, necessitates a delicate coupling of immune resistance to infectious organisms and tolerance to tissue damage and inflammation. While failure to mount a robust protective response against pathogens can result in susceptibility to recurrent infection, an inadequate ability to dampen the inflammatory response and limit disease severity may contribute to excessive inflammation and irreversible tissue damage. Macrophages are a critical component of this immunological balancing act. These mononuclear phagocytes play an indispensable role in distinguishing commensal luminal antigens and potentially life-threatening pathogens, as well as in resolving inflammatory responses. Therefore, understanding the regulatory mechanisms of macrophages in states of health and intestinal disease holds promise for the optimization of therapeutics aimed at resolving inflammation and resisting pathogens. In this article, we will briefly discuss the mechanisms by which macrophages function under normal physiological conditions in the steady state (e.g. regulation of enteric neurons), and their training by the local intestinal milieu. Finally, we will discuss how macrophage-established homeostasis is perturbed and subsequently restored in the resolution of intestinal inflammation, before offering our perspective on the role of type 2 immune responses and helminth infections in this process.

Features and functions of intestinal macrophages

As the largest pool of mononuclear phagocytes in the body, macrophages are ubiquitous throughout the gastrointestinal tract, with major populations found in the stomach as well as along the length of the small and large intestines. (1). However, there are marked differences in luminal content dependent on regional specialization throughout the gastrointestinal tract and thus, the presentation of distinct challenges to the immune system. The role of the duodenum and jejunum of the upper small intestine in brush border enzyme-dependent digestion and nutrition allows for a relatively sterile and largely food antigen-derived microenvironment compared to the ileum and commensal-harboring colon which are selected for in inflammatory bowel diseases (2). In line with the role of resident macrophages found in other tissues, intestinal macrophages are highly phagocytic, expressing genes such as Mrc1, CD36, Timd4, and Mertk in the mouse (3, 4), and are important for clearing apoptotic cells through a process called efferocytosis (5, 6). Intestinal macrophages express CD64 (7) and Mertk (8), as well as CD11c, MHCII and F4/80. CD64 is also expressed on human macrophages (9) rendering it a useful marker across species. There is considerable confusion and controversy with respect to distinguishing between intestinal macrophages and dendritic cells (10), but we will not address this in further detail. The combination of CD14, HLA-DR, CD11c and CD11b is able to distinguish among four distinct macrophage populations in the human small intestine (9), which are highly phagocytic but less responsive to toll-like receptor (TLR) ligand stimulation. Importantly, while macrophages express MHCII and antigen presentation machinery, they are inferior to dendritic cells in activating naïve T cells in the lymph nodes (11, 12), but likely affect effector T cell function in the tissue (see below.)

The privileged position of macrophages in the lamina propria (LP) in close proximity to the epithelial layer and gut lumen facilitates mucosal barrier protection and the elimination of transgressing bacteria. LP macrophages play a key role in immune sampling of luminal bacteria (Figure 1). By extension of transepithelial dendrites (TEDs), macrophages expressing the fractalkine receptor CX₃CR1 can penetrate the epithelial barrier with minimal disruption to integrity (10, 13–15). The ability of LP macrophages to sample gut antigens may be altered by affecting intestinal epithelial cell differentiation as a result of disrupted crypts and reduction of Lgr5⁺ stem cells by CSF1R blockade (16). Furthermore, CSF1R-dependent macrophages affect the differentiation of additional cell types important for immune defense within the gut including Paneth cells, goblet cells, and M cells (16). Despite possessing bactericidal properties when encountering pathogens, intestinal macrophages do not elicit an overt inflammatory response to commensals (10, 17). Constitutive IL-10 signaling is essential in maintaining the appropriate regulation of intestinal macrophages and the prevention of aberrant inflammation. Deletion of the IL-10 receptor on CX₃CR1-expressing macrophages in the intestine leads to spontaneous severe colitis in mice (18) and can lead to the development of severe early onset inflammatory bowel disease (IBD) in pediatric patients (19, 20). In addition to bacterial sampling, intestinal macrophages are important influences on T cell function in the gut. The activity and function of Tregs (via production of IL-10 (21)) and Th17 cells (as a source of IL-1β (22)) in particular are likely to be influenced by the activation status of intestinal macrophages (see below.) Hence, macrophages facilitate the maintenance of homeostasis in the gut through a combination of phagocytic and anti-bacterial functions and immune regulation.

Figure 1.

Orchestration of intestinal macrophages in homeostasis, inflammation and helminth infection. Macrophages residing in the lamina propria (LP) are strategically positioned in close proximity to the epithelial layer. CX₃CR1⁺ LP macrophages participate in sampling of luminal contents via extension of transepithelial dendrites and phagocytose transgressing pathogens. Through the secretion of immunoregulatory cytokines IL-10 and TGF𝛃, LP macrophages support the expansion of regulatory T cells (Tregs). Within the deeper layers of the gut wall, the submucosa and muscularis externa, macrophages engage in a reciprocal crosstalk with a network of enteric neurons. This molecular dialog is mediated by the secretion of bone morphogenic protein 2 (BMP2) by macrophages. In response, enteric neurons secrete CSF1 which further promotes BMP2 secretion. In the setting of intestinal inflammation and breach of the epithelial layer, proinflammatory macrophages are recruited and elicit a protective immune response mediated by the secretion of inflammatory cytokines IL-1𝛃, TNFα and IL-6. Helminth infection induces a type 2 immune response mediated by an induction of Tregs and Th2 cells. Enteric neurons may sense infection by helminths and produce the neuropeptide neuromedin U (NMU) that stimulates innate lymphoid type 2 cells (ILC2s) to produce type 2 cytokines that can alternatively activate intestinal macrophages.

Tissue-resident and monocyte-derived intestinal macrophages

It is now clear that a significant proportion of tissue-resident macrophages are maintained in situ by self-proliferation from progenitors that are already seeded in the tissues during embryonic development (23, 24), whereas pro-inflammatory macrophages differentiate from infiltrating monocytes of hematopoietic origin during inflammatory responses. Depending on the particular tissue in which tissue-resident macrophages inhabit, the macrophage pool can be self-maintained throughout adulthood or rely on constant replenishment from infiltrating monocytes in the blood (25). In the case of the intestines, incoming monocytes were thought to continuously replace tissue-resident intestinal macrophages (26–28), in stark contrast to other tissue that contain macrophage populations of both embryonic and monocyte-derived origin (29). However, long lived macrophages defined by CD4 and Tim-4 expression were recently described for the mouse intestine (3). Fate-mapping experiments and bone marrow chimeric mice indicate that these macrophages can persist for 8 months in the gut (3), self-renew from local maintenance, and consist of embryonic as well as hematopoietic cells (3). This transcriptionally distinct minority of self-maintaining gut macrophages with enrichment of gene transcripts such as Nova1, Chrm2, and Efr3b involved in tissue protective and supportive processes may support the maintenance of submucosal vascular architecture and positively regulate neuronal differentiation (4). Why the gut mucosa is comprised of macrophages predominantly from monocytes, whereas other tissues have additional macrophage subsets of varying origin remains unclear. However, niche accessibility is thought to play an active role in this molecular decision (23, 30). The gut mucosa represents a vast amount of surface area available for contact with the external environment and is freely accessible throughout life. Therefore, recruitment of monocytes to the intestines could be explained by the baseline inflammatory tone of the gut due to persistent antigen exposure.

Environmental conditioning of intestinal macrophages

The molecular cues responsible for the orchestration of the differentiation of monocytes into intestinal macrophages are not entirely understood. The local gastrointestinal microenvironment must provide macrophages with a repertoire of instructive signals. CSF1 is an important growth factor for myeloid cells as well as other populations within the gut (16, 31, 32). Transcriptional profiling of human monocyte-derived macrophages grown in CSF1 and exposed to lipopolysaccharide (LPS) identified enriched promoters for CSF1-dependent monocyte-macrophage transition in IBD susceptibility loci, suggesting that maladaptation of monocytes to the gut mucosa is a key trigger for developing IBD (33).TGFβ signaling, perhaps acting through Runt-related transcription factor 3 (RUNX3), may especially be important for terminal differentiation (34). Notch signaling is another developmental signal that is likely required for intestinal differentiation (35). A vital role of the gut microbiota and microbe-derived molecules has also been suggested in the influence of macrophage population dynamics (36). Furthermore, depletion of the gut microbiota by broad spectrum antibiotics greatly reduces macrophage turnover (26). Mature intestinal macrophages upregulate genes encoding monocyte chemoattractants such as Ccl7 and Ccl8, suggesting developmental hardwiring of a self-replacement program (37). The phagocytosis of apoptotic intestinal epithelial cells by intestinal macrophages may also lead to suppression of pro-inflammatory genes (38). Hence, the combination of microbiota signals and normal apoptotic epithelial cells (which could be linked via unknown mechanisms) may lead to the conditioning of intestinal macrophages that are less responsive to stimulation by microbial antigens. Notably, as macrophages in the intestine mature and become CX₃CR1^hi, they are conditioned towards adopting this tolerogenic phenotype (39), which is a process that clearly requires the IL-10 axis (Figure 1).

Neuro-immune interactions mediated by macrophages

The immune and nervous systems are charged with the task of sensation within a landscape of fluctuating conditions typically found at barrier surfaces like that of the gastrointestinal tract. Although ultimately employing structurally distinct processes, i.e. immune recognition of pathogens versus nociception of potentially harmful stimuli, both systems rely on direct cell-cell communication mediated by series of soluble signaling molecules, including cytokines and neurotransmitters, and possess memory capacity in anticipation of re-challenge. The adult gastrointestinal tract harbors not only the largest immune cell compartment in the body, but also a prolific neuronal network which rivals that of the spinal cord (40, 41). While many research efforts have been directed towards macrophages in the lamina propria, macrophages that reside in the deeper layers of the gut wall, namely the submucosa, the muscularis externa which comprises the circular and longitudinal smooth muscle layers, and the serosa, have largely been neglected until recent advancements in imaging techniques and transcriptional profiling (42). These macrophage subsets are functionally and morphologically distinct. In contrast to proinflammatory LP macrophages that likely undergo training by luminal cues, muscularis macrophages (MM) adopt a tissue-protective phenotype, upregulating M2 genes such as Arg1 and Cd163 (42). Anatomic positioning of CX₃CR1⁺ macrophages in the muscularis externa may influence morphology (43), with the adoption of either a stellate or bipolar morphology (42). There is now evidence to suggest that gut macrophages are intimately associated with the enteric nervous system and engage in a reciprocal crosstalk (44). Gastrointestinal motility can be regulated by bone morphogenic protein 2 (BMP2) secreted by muscularis macrophages (31). In a positive feedback mechanism, BMP2 from MM acts on its receptor BMPR on the surface of enteric neurons to induce CSF1 secretion, which in turn stimulates further BMP2 expression (31, 42). However, recent findings of an expected number and distribution of phenotypically intact MM in Ret ^−/− mice and aganglionic colons of Hirschsprung disease patients undermine this signaling circuit and suggest a role for environmental influences (45). Notably, antibiotic treatment reduces BMP2 expression on MM, indicating that the gut microbiota may play a role in regulating this process (31), although the mechanism of this interaction is still not clear. These interactions could then regulate peristalsis, or contraction of smooth muscle layers, which propels an ingested bolus through the gastrointestinal tract for eventual elimination. This mechanism is further illustrated by the grossly disturbed gastrointestinal functional motility upon treatment of mice with broad spectrum antibiotics (46). Clinically, postoperative ileus (POI) may occur when this system is transiently dysfunctional and may prolong hospitalization in patients who have undergone major abdominal surgical procedures. In the presentation of POI, a subtle proinflammatory response by resident macrophages in the muscularis externa could increase neutrophil recruitment and upregulate adhesion molecules to inhibit muscle contractility, resulting in delayed intestinal transit (43, 47, 48). Anti-inflammatory molecules that counteract inflammation such as polyunsaturated fatty acid-derived pro-resolving lipid mediators may rescue gut motility (49), however the role of IL-10 in this process remains controversial (50, 51). While macrophages can influence neuronal function in this context, neuronal sensing of helminth infections may also lead to activation of immune cells in the gut (52), which may lead to the appropriate activation of intestinal macrophages (Figure 1). Hence, there are intriguing bidirectional communication mechanisms between neuronal cells and immune cells (including macrophages) in the gut that are just beginning to be elucidated.

Macrophages in the setting of gut inflammation and infection

During intestinal inflammation, there is a rapid and abundant accumulation of inflammatory Ly6C^hi monocytes in the gut. These monocytes differentiate into macrophages, but unlike their counterparts present under homeostatic conditions, they are highly responsive to stimulation by microbial antigens. They will upregulate pro-inflammatory cytokines (e.g. IL-1β, IL-6, TNF) and produce reactive oxygen species in response to TLR agonist stimulation. One possibility is that the infiltrating monocytes and differentiating macrophages have not been appropriately conditioned to the gut environment yet to become anti-inflammatory. The resident macrophages that are CX₃CR1^hi still retain their anti-inflammatory signature during inflammation. Since expression of CX₃CR1 is a final step in the maturation of mature macrophages from Ly6C^hi monocytes in the monocyte differentiation “waterfall”, it seems likely that the pro-inflammatory nature of infiltrating monocytes is slowly deconditioned in the normal gut, but under inflammatory settings they become persistently activated without becoming fully differentiated into resident immunoregulatory intestinal macrophages. The resident macrophages that already express high levels of CX₃CR1 may thus be particularly important in immune regulation during inflammation. The presence of a normal microbiota may also be important in conditioning CX₃CR1⁺ macrophages to adopt these immune regulatory properties (53), whereas in the presence of pathogens these macrophages may adopt dendritic cell-like properties by trafficking back to the lymph nodes via CCR7 to induce T cell and B cell responses (54).

While data on macrophage function in inflammation of the human gut is sparse and correlative compared to findings obtained from mouse studies, the underlying principles likely remain similar. The breakdown of mucosal immune homeostasis established and maintained by tissue resident macrophages may result in acute and chronic inflammatory states of the gut. This manifests most notably in inflammatory bowel disease, comprising Crohn’s disease and ulcerative colitis, which is a chronic inflammatory disorder of the gastrointestinal tract requiring lifelong management as a consequence of exacerbated immune responses in genetically predisposed individuals. In response to inflammatory signals from the intestinal epithelial layer, usually the initiating step of intestinal inflammation, newly recruited CD14^hiCD11c^hi monocytes and immature macrophages challenge the existing immunosuppressive population of resident macrophages (CD64⁺HLA-DR^hiCD14^lo) and differentiate into proinflammatory macrophages with a signature expression of TNFα, IL-1β, and IL-6 (55, 56). Notably, the majority of IBD susceptibility loci are expressed in monocytes and have been identified by genome-wide association studies as well as functional studies (33, 57–59). Recruited proinflammatory macrophages in this context express TREM1 (60) and could further disrupt integrity and introduce a leakage of luminal content through increased pro-inflammatory responses. CCR2⁺CX₃CR1⁺ macrophage subsets have also been demonstrated to undergo expansion in the mucosa of patients with active IBD (61). Furthermore, in the human inflamed gut, accumulated CD14^hi macrophages may provide an indication of IBD severity (62). This population is similar to the murine equivalent Ly6C^hi population that dominates under conditions of infection (63). The parallels between macrophages observed in patient endoscopic biopsy material and murine models may be key to identifying possible therapeutic targets through translational studies.

Alterations to macrophage function during resolution of inflammation

While macrophages are critical for a protective inflammatory response against pathogens, they are also of equal importance in the resolution phase of inflammation (64). Clearance of accumulated neutrophils that responded to inflammatory stimuli in the gut occurs mainly through efferocytosis by macrophages as they become apoptotic. While both tissue-resident macrophages and monocyte-derived inflammatory macrophages can clear apoptotic cells, how they respond to this process may be distinct (65). For example, expression of Tim-4 can distinguish between tissue-resident and inflammatory macrophages. Tim-4 is a PtdSer recognition receptor and macrophages lacking Tim-4 have reduced apoptotic cell engulfment (66). Probably, a combination of intrinsic signals (e.g. triggered by phagocytosis) and extrinsic signals (e.g. cytokines such as IL-4, IL-13 and IL-10) enable the intestinal macrophages to adopt a wound healing anti-inflammatory response to counteract tissue damage and restore homeostasis. Efferocytosis can induce TGF-β (67), which promotes the differentiation of Treg cells. Retinoic acid produced by IL-4 activated macrophages (68) as well as intestinal dendritic cells (69), acts synergistically with TGF-β to promote Treg differentiation. Hence, intestinal macrophages represent a key link between the combination of immune regulatory responses (e.g. Tregs and IL-10) as well as type-2 responses in the process of resolving tissue inflammation and promoting mucosal healing. ILC2s responding to the breach in barrier signals, including IL-25 and IL-33, are an important source of IL-13 alongside CD4⁺ T_H2 cells. This type-2 cytokine environment then promotes the accumulation of alternatively activated M2 phenotype macrophages.

The traditional concept of activated pro-inflammatory M1 macrophages and anti-inflammatory M2 macrophages being important for antibacterial responses and inflammation resolution respectively (70), is becoming increasingly outdated in light of the remarkable heterogeneity of macrophage populations as revealed by single-cell RNA sequencing studies and other transcriptomic analyses (71–73). While polarization of macrophages partially contributes to this heterogeneity, it nonetheless does not provide a complete explanation. However, there are clearly macrophage populations that are important for wound healing and tissue repair that are dependent on type-2 cytokines (70, 74), and are further enhanced by phagocytosis of apoptotic cells (75).

Mucosal healing is a widely recognized treatment goal in the management of complex IBD patients. However, remission is achieved in only a proportion of patients, many of whom lose response over time and require surgical management (76–78). In patients who demonstrate a clinical response to infliximab, a distinct subset of macrophages expressing CD206 is induced and expanded compared to patients who failed to respond to the biologic (79). Furthermore, tofacitinib, a small molecule JAK inhibitor developed for the treatment of IBD as well as other immune-mediated disorders, affects macrophage polarization and function (80, 81). Upon treatment with tofacitinib, both murine and human macrophages increase transcription of alternatively-activated macrophage markers and increase levels of IL-10 secretion while inhibiting IFNγ signaling, which may help ameliorate weight loss and disease activity, and further highlight its therapeutic potential (81). Hence, there is some indirect evidence that macrophage activation pathways associated with tissue repair and type 2 cytokine responses could be associated with mucosal healing in IBD patients, or with the activity of successful therapeutic agents used to treat IBD patients.

Helminth infection and intestinal macrophages

In addition to the commensal gut bacteria, the gut is home to soil-transmitted helminths, especially intestinal nematodes. These worms can cause considerable tissue damage, which requires the activation of a type 2 cytokine mediated wound healing response (82). While the type 2 response is important for protecting the gut from intestinal injury during colonization, it is also important for expulsion of the parasites and for reducing worm burden in infected individuals. Notably, infection with helminths has also been proposed to promote resolution of inflammation in IBD (83, 84). The type 2 immune response induced by helminth infections could promote goblet cell activity and restoration of the mucus layer (85), but also induce alternatively-activated M2 macrophages in the gut to promote tissue repair and resolution of inflammation (86). While for some helminths (e.g. Heligomosoides polygyrus) this macrophage response is necessary for the expulsion of the parasites (87), in other settings (e.g. Nippostrongylus brasiliensis) macrophages do not play a role in expulsion, but may instead be important for tissue repair (88, 89). In mouse studies whereby helminth infection can improve symptoms of colitis, macrophages have been credited as the cell type important for the suppression of intestinal inflammation (90, 91). Macrophages stimulated in vitro with antigens from the tapeworm Hymenolepis diminuta, or interleukin-4 can improve colitis when adoptively transferred into mice (91, 92) via a mechanism that requires intact IL-10 (93). This could be mediated via induction of regulatory T cells (94), perhaps through retinoic acid (68, 95, 96). The source of IL-4 or IL-13 for alternatively activated macrophages during helminth infections is CD4⁺ T cells as well as innate lymphoid type 2 cells (ILC2s). Recently, production of neuromedin U (NMU) by neurons was shown to stimulate ILC2s to produce IL-5 and IL-13 (52, 97, 98) which can occur in the intestine during helminth infections. Hence, neuronal sensing of helminth infection may lead to the production of cytokines by ILC2s necessary for macrophage alternative activation. Alternatively-activated M2 macrophages may also play a role in shifting the balance of the gut microbiota towards an increasingly anti-inflammatory microbial community while inhibiting pro-inflammatory bacteria (99). If type 2 cytokines could be specifically targeted to the gut, this could be a potential therapeutic strategy to promote the resolution of intestinal inflammation by alternatively-activated M2 macrophages, while concomitantly activating goblet cells to restore the mucus barrier. However, one concern may be that this approach would increase susceptibility to certain intestinal bacterial infections (100, 101) or colorectal cancer (102, 103).

Conclusions

Significant progress has been made in recent years in advancing our understanding of both mouse and human intestinal macrophage immunobiology in the context of inflammation. What are the key unanswered questions for the future? First, our understanding of the heterogeneity of macrophage subtypes in the gut is still at an early stage. With the advent of single cell sequencing technologies, the next few years will surely lead to the identification of additional subtypes and different activation states of intestinal macrophages, which may arise from different cellular lineages, and perform distinct functions under both homeostatic conditions as well as inflammatory conditions. The concept of M1 and M2 activated macrophages will become outdated, as we are already beginning to appreciate that the complexity of macrophage subsets far exceeds such simple classification. What common nomenclature will we then use to describe this heterogeneity? Because of the plasticity of macrophages and extensive conditioning by the local microenvironment, each unique tissue location within the gut will possess specific markers and differentiation cues for the macrophages that reside within. Defining a common nomenclature, presumably based on both cell surface receptors and transcriptional profiles, to describe macrophage types will become increasingly challenging for the purposes of comparing different studies. Secondly, greater effort has to be made to characterize human intestinal macrophages and to relate the function of human macrophage subsets with mouse data. While mechanistic causality is always difficult to demonstrate in human studies, ex vivo experiments that can demonstrate similar immunological properties between specific populations with a mouse macrophage counterpart (where in vivo roles can be established experimentally) is a key component missing from current studies. Finally, we have a poor understanding of between-individual variation in macrophage populations. Whereas inbred strains of mice are relatively homogenous as they are genetically identical and are exposed to similar laboratory conditions, genetic and environmental variation will affect human intestinal macrophage populations and functions. Most immunological studies are conducted only on C57BL/6 mice and different genetic backgrounds may provide new insight to some key questions. The identification of various bacterial and parasitic commensals that may vary between animal facilities can affect intestinal immunity and could explain different experimental outcomes from the intestinal system. Given the integral role of macrophages in immune-mediated and functional gastrointestinal disorders and their potential as an attractive therapeutic target, it is critical to understand monocyte/macrophage compartment control mechanisms that are responsible for tipping the balance from homeostasis to disease. This may improve the feasibility of treating various gastrointestinal pathologies including IBD with interventions designed to exploit the ability of macrophages to fine tune immunological responses and influence clinical outcomes.