Monocyte trafficking in acute and chronic inflammation

Monocyte Phenotypic Heterogeneity

Monocytes are heterogeneous circulating blood cells poised to rapidly extravasate into inflamed tissues. In the bone marrow (BM), a common Lin⁻cKit^hiCD115⁺CX3CR1⁺Flt3⁺ progenitor cell, termed macrophage and DC precursor (MDP), gives rise to monocytes and numerous subsets of macrophages and DCs [1]. Two major subsets of blood monocytes have been described in mice, humans, and other species [2–4]. The two murine subsets, classical (which express Ly6C, recognized by the anti-Gr1 antibody) and nonclassical monocytes (Ly6C^lo), are distinguished by differential expression of chemokine receptors, particularly CCR1, CCR2, and CX3CR1 (fractalkine receptor) [3–6]. Classical monocytes are known to exit the BM in a CCR2-dependent fashion [7] and CCR2 ligands CCL2 (MCP-1) and CCL7 (MCP-3) help maintain homeostatic levels of monocytes in the circulation [8]. Mechanisms controlling the egress of non-classical monocytes from the BM remain elusive and, indeed, it remains to be directly demonstrated that they arise in the BM. In addition to the BM, the spleen is home to both subsets of mature monocytes which appear phenotypically similar to the two subsets found in blood (Text Box 1) [9].

Analogously, CD115⁺ human monocytes can be separated into two major subsets, typically based on their differential expression of CD14 and CD16 [6, 10]. The CD14⁺⁺CD16⁻ subset, equivalent to mouse classical monocytes, is also CCR2⁺, whereas the CD14⁺CD16⁺ (also called CD14^dimCD16⁺) subset lacks CCR2 and constitutes the non-classical population [6]. Nonclassical CD16⁺ monocytes can be further divided by their high or low levels of CD14 expression [11]. Recent work reveals a difference between the minor CD14⁺CD16⁺ and CD14^dimCD16⁺ monocyte populations in their capacity to become activated and secrete key inflammatory cytokines in response to different stimuli [12]. CD14⁺CD16⁻ and CD14⁺CD16⁺ respond to TLR2 and TLR4 ligands, whereas CD14^dimCD16⁺ monocytes respond to viral stimuli through TLR7 and TLR8. In addition, these data are in agreement with previous work [6] suggesting that CD14^dimCD16⁺ monocytes are the counterparts of nonclassical monocytes in mice. Differential gene expression profiles, including many genes involved in cell trafficking, between the monocyte subsets are well conserved between mice and humans, suggesting mouse models of trafficking may reveal clues to human monocyte behavior in disease [6]. However, when applying observations made in mouse experimental models to human disease, it is important to recognize a few key differences, such as subset ratio in the blood, that exist between humans and mice. Here, we discuss recent advances in the mechanisms controlling monocyte trafficking during acute and chronic inflammation and address controversial issues such as the role of CCR2 in directing monocytes to inflamed tissues.

Monocyte Functional Heterogeneity

As might be predicted from their differential gene expression patterns, monocyte subsets also differ functionally. Nonclassical monocytes demonstrate a patrolling behavior along blood vessel walls [13] and accumulate in non-inflamed peripheral tissues such as spleen, lung, and liver when adoptively transferred [3]. While lineage tracer studies suggest that it is unlikely that nonclassical monocytes contribute to dendritic cell (DC) populations in steady state peripheral organs, it is possible they contribute to resident macrophage populations [14]. Consensus on the homing patterns of both monocyte subsets has not been reached, but there is overwhelming evidence that recruitment of classical monocytes dominates early in inflammatory responses. For example, classical monocytes infiltrate inflamed tissues, such as in sterile peritonitis models, more robustly than or to the exclusion of their nonclassical counterparts in mice and rats, and are specifically increased in the circulation during systemic or chronic infection [3, 4, 15, 16]. Only classical monocytes migrate to the skin of mice receiving an intracutaneous injection of latex microspheres [17] and in a model of skeletal muscle injury, only classical monocytes migrate to injured tissue [18]. However, after engulfing dying cells, they differentiated into cells resembling nonclassical monocytes, which mediated tissue repair mechanisms [18]. By contrast, after myocardial infarction, both monocyte subsets appear to home to the same tissue at different stages of inflammation (Figure 1) [19]. Specifically, whereas the classical subset of monocytes first seeds the infarcted heart and exhibits inflammatory functions, the nonclassical subset is recruited at a later stage and promotes tissue healing by expressing high amounts of vascular endothelial growth factor [19]. In this report, the two subsets are under the control of distinct trafficking mechanisms, with the classical subset being recruited via CCR2 and nonclassical monocytes utilizing a CX3CR1-dependent pathway [19]. Dependence on CX3CR1 may be due to survival rather than recruitment, as this chemokine receptor is important for nonclassical monocyte homeostasis [20, 21]. To add a level of complexity, nonclassical monocytes were reportedly recruited earlier than classical monocytes, or even neutrophils, to the peritoneum following intraperitoneal injection of Listeria monocytogenes [13]. Given these disparate outcomes in different models, the relative contributions of classical monocyte recruitment and subsequent differentiation versus sequential monocyte subset recruitment to injured tissues merit further study.

The exact relationship between these subsets and stage at which they diverge during development is unclear. Although phenotypically distinct monocyte subsets exist in the BM and blood, most studies suggest that these subsets represent the same cells at different maturation stages [4, 22, 23]. Indeed, in the absence of inflammation, grafted classical monocytes can home back to the BM, differentiate into nonclassical monocytes, and return to the bloodstream in mice [23]. Monocyte subset conversion, after adoptive transfer of classical monocytes, also occurs in rats in the absence of further proliferation [15]. Further understanding of the pathways that govern monocyte subset conversion is critical to advance efforts to manipulate subset frequency to attenuate inflammation and/or enhance tissue repair.

Monocyte trafficking during acute inflammation: Listeria infection as a prototypical monocyte migration model

Monocyte migration to sites of inflammation typically occurs after neutrophil infiltration and can be sustained for days (Figure 1). Initial experiments exploring monocyte subset behavior in the context of inflammation used sterile inflammatory triggers, such as thioglycollate or lipopolysaccharide (LPS), even if the inciting agent was microbial in origin [3, 4, 15]. More recently, studies have focused on the role of classical monocytes in inflammatory models of bacterial, viral and parasitic infections and the role this subset serves in the host response.

Monocyte behavior in the context of infection is probably best characterized in Gram-positive opportunistic pathogen Listeria monocytogenes models [24]. Listeria infection provokes a robust monocytosis, an observation that led to the original name Bacterium monocytogenes for an unknown bacterium in 1926 [25]. More recent work demonstrated that Listeria infection promotes monopoiesis in the BM that is dependent on TLR signaling and is sustained by ongoing infection [26]. Intravenously administered Listeria invade splenocytes and hepatocytes and escape into the cytosol [25], which is required for the expression of the CCR2 ligands, CCL2 (MCP-1) and CCL7 (MCP-3), in target organs [27, 28]. These chemokines induce CCR2⁺ classical monocyte infiltration that controls infection directly and mediates initiation of adaptive immunity [24, 29]. Deletions in CCL2 or CCL7 result in reduced, but not ablated, monocyte recruitment, which correlates with an increase in bacterial burden [28]. This outcome is an intermediate phenotype, as compared to infection of Ccr2^−/− mice, where monocyte migration to infected spleen and liver is severely abated, and both bacterial burden and mortality are increased [24, 29, 30]. CCR2 signaling is thought to be the primary mechanism for monocyte migration to Listeria-infected tissues and mice deficient in CCR5 do not display increased susceptibility to infection [31]. Ccr2^−/− mice exhibit a specific accumulation of monocytes in the BM under both homeostatic and inflammatory conditions, suggesting that CCR2 is required for mediating monocyte egress from the BM during infection [7]. Using CCL2 reporter mice and conditional knockouts, mesenchymal stem cells and CXCL12-abundant reticular cells have been identified as the major producers of CCL2 in the BM, regulating monocyte egress in response to TLR ligands [32]. The failure to observe a role for CCR2 in tissue-specific homing in this work [7] and others indicates that CCR2 either does not play a role at this step or has a redundant role with other chemoattractants. See Text Box 2 for further discussion.

Beyond Listeria: The role of chemokine receptors in additional infection models

Listeria models demonstrate the critical role of CCR2 in mediating robust trafficking of classical monocytes, primarily by controlling BM egress. Using similar tools, such as Ccr2, Ccl2, or Ccl7^−/− mice, CCR2-dependent migration of classical monocytes appears to be a critical component of the host response to a wide variety of microbial infections (Figure 1), although not always to the benefit of the host.

Aerosol Mycobacterium tuberculosis infection induces chemokine expression, including CCL2 in lung tissue [33]. Moreover, Ccr2^−/− mice have impaired monocyte and T cell migration to the lung and increased susceptibility to high-dose infection [33]. Similarly, absence of CCL2 contributes to reduced monocyte and IFN-γ-producing T cell infiltration in the M. tuberculosis-infected lung [34]. Ccr2^−/− chimeric mice reconstituted with Ccr2^+/+ monocytes and Ccr2^−/− T cells and infected with M. tuberculosis addressed whether decreased T cell recruitment is a direct result of the absence of CCR2 or an indirect result of reduced monocyte infiltration [35]. In these mice, T cells were recruited to infected lungs in comparable numbers independently of CCR2, suggesting that M. tuberculosis susceptibility in Ccr2^−/− mice depends predominantly on monocyte infiltration [35]. Interestingly, treatment with poly-IC, a synthetic analog of dsRNA that acts as a TLR3 agonist, of mice infected with M. tuberculosis renders the mice more susceptible to infection, through the robust recruitment of classical monocytes to the lung, which support bacterial growth in this model [36]. This suggests that there is a fine balance in the magnitude of monocyte recruitment that controls infection in response to Mycobacterium infection.

While CCR2 deficiency is beneficial during Dengue virus infection, prolonging host survival and minimizing tissue damage [37], it is detrimental in a model of West Nile virus infection of the brain [38]. Similar to M. tuberculosis infection, CCL2 neutralization is beneficial for the host after West Nile virus infection [39], whereas CCR2 deficiency negatively affects host survival [38], suggesting that moderating monocyte infiltration minimizes tissue damage while still controlling viral replication.

Gram-positive uropathogens, such as Staphylococcus saprophyticus that preferentially infect the kidney, induce only modest monocyte recruitment, despite kidney-specific expression of monocytic chemokines CCL2, CCL3, and CCL5 [40]. By contrast, CCR2-mediated, CCL2-expendable, monocyte recruitment to the kidney in a model of acute ischemic-reperfusion injury is robust and harmful to the host [41]. Both of these models induce multiple monocytic chemokines but exhibit surprisingly different monocyte trafficking patterns, suggesting further investigation is needed to determine the role of monocytes in defense against pathogens and tissue injury in the kidney. Monocyte recruitment during uropathogenic E. coli (UPEC) infection may be controlled by several chemokines expressed robustly in the bladder, including CCL2, CCL3, CCL4, and CCL5 [42] but the contribution of each of these ligands is unknown. Monocyte infiltration correlates with significantly reduced UPEC, suggesting that monocytes likely play a critical role in bacterial eradication [42]. In support of this idea, recent analysis reveals that TLR4 polymorphisms and expression levels on human monocytes correlate with susceptibility to acute and chronic urinary tract infection in adult patients [43]. Seemingly contradictory findings demonstrate that UPEC infection of Ccr2^−/− mice reveals no dependency on monocyte infiltration for reduction in bacterial burden [44, 45]. However, the apparently self-limiting strain used in these studies [44, 45] is rapidly cleared from the wildtype host despite inoculums fifty times higher than those used in similar infection models where UPEC robustly colonize bladder tissue [42], thus complicating data interpretation.

Fate of Recruited Monocytes: TipDCs and immunity

Classical monocytes recruited to Listeria-infected spleen undergo differentiation into so-called TNF and iNOS-producing DCs or TipDCs [29]. Subsequent to this study, cells resembling TipDCs have been identified in bacterial infection models of Brucella melitensis [46], UPEC [44], and Salmonella [47]. Development of TipDCs have also been observed in viral infections such as influenza [48], and parasitic infection models, including Trypanosoma brucei [49], Leishmania major [50], and Toxoplasma gondii infection [51] (Figure 1).

In Salmonella infection, recruited classical monocytes differentiate further once they have reached the infected gut, and about 20–30% express TNFα or iNOS [47]. However, these cells are, in fact, poor antigen-presenting cells, despite expressing CD11c, MHC II and co-stimulatory molecules [47]. Further, while these cells are critical in controlling bacterial burden, Salmonella has been shown to block monocyte-derived DC accumulation in lymph nodes when injected into the skin [47, 52, 53], due at least in part to differences in the extracellular matrix of the host [54].

In a model of oral inoculation of Toxoplasma gondii, classical monocytes are recruited to the villi of the small intestine in a CCR2-dependent manner [51]. Mice succumb to infection more rapidly in the absence of either CCR2 or CCL2 [55]. Recruited monocytes differentiate into cells resembling TipDCs, with the ability to produce TNF-α and iNOS, which serves to eliminate local parasite burden [51]. The authors of this study stress that the TNFα and iNOS-producing cells identified in the gut of T. gondii-infected mice do not express CD11c [51, 56], differentiating them from TipDCs identified during Listeria infection [29]. However, it is possible that they are the same cells, as TipDCs are described as being CD11c^int [29], which may be difficult to discern by immunofluorescent detection methods such as those used in the T. gondii study [51]. Of note, these cells may be the same cells identified in a peritoneal parasite infection model as recruited Gr1⁺ CD68⁺ macrophages [57].

During infection, most studies have focused on initial innate immune responses, demonstrating that TipDCs play a critical role in controlling pathogen burden, but do not demonstrate, for instance, that their absence leads to impaired adaptive immunity. Indeed, TipDCs are not necessary for CD4⁺ or CD8⁺ T priming and activation in Listeria infection [29]. One exception to this is a recent study describing the accumulation in lymph nodes of two DC subsets that arise during Leishmania major skin infection [58]. One subset migrates to lymph nodes through lymphatic vessels from infected skin while the second appears to directly enter lymph nodes from the circulation [58]. Monocyte-derived DCs from the skin exhibit superior ability to activate T cells in this system [58]. As cells identified as TipDCs are the predominant cell infected in a L. major model [50], it is reasonable to conclude that here, these L. major-associated TipDCs behave as DCs, contributing to the initiation of adaptive immune responses. This is only one example and further investigation into the role of TipDCs in eradicating pathogens or in antigen presentation are needed.

One important question is how monocytes that enter the lymph node through the HEV, such as observed in L. major infection, gain access to antigen for presentation after they arrive and locally differentiate into DCs in the ensuing hours and days [58–61]. Inflamed lymph nodes (LNs) contain a population of CD11c^loCD11b⁺Gr1⁺ cells that phenotypically resemble monocytes, and appear to seed the LN through the HEV in a CCR2-dependent manner [15, 60, 62]. The recruitment of these cells may be a consequence of inflammatory signals that enter the LN from the lymph. In particular, CCL2 made in inflamed organs is rapidly transported to the HEV through the conduit system in the LN to recruit monocytes along HEVs [62]. It has been shown that these recruited monocytes undergo rapid differentiation into DC-like cells [58, 59], but they may preserve some aspects of their former monocyte phenotype for some period of time [15]. Do these monocytes have sufficient access to antigen that drains through conduits [63]? Alternatively, could antigen be passed on from DCs that emigrated to lymph nodes via the afferent lymphatic vessel from sites of inflammation [64]? It is important to address these possibilities in order to better understand the role of monocyte arrival to lymph nodes through the HEV in immune responses.

Monocyte trafficking during chronic inflammation: studies in atherosclerosis

Nearly all infection models that have studied monocyte trafficking have done so in an acute phase of infection. To gain a better picture of monocyte trafficking during chronic inflammation, one can turn to studies of the chronic inflammatory disease atherosclerosis. Atherosclerosis begins very early in life as deposition of lipid and infiltration of monocytes creates fatty streaks in the intima of arteries. Advanced atheromatous plaques are characterized by accumulation of monocyte-derived foam cells and modified lipoproteins, with lipid and dying foam cells contributing to the formation of a molten necrotic core [65]. Even before the disease begins, CD11c⁺ cells accumulate in areas of the artery that will, upon elevated cholesterol in the plasma, go on to form plaques [66]. In these nascent plaques, as in all stages of the disease, both monocyte subsets are recruited, where they rapidly accumulate cholesterol esters to become foam cells, using a combination of chemokine receptors that include CCR2, CCR5, and CX3CR1 [67].

Contribution of monocytosis and monocyte subsets to disease progression

Similarly to Listeria infection [25], monocytosis develops in apoE-deficient hypercholesterolemic mouse models and contributes significantly to the pool of monocytes available in circulation [16, 69]. Monocytosis appears to be initiated and maintained by the accumulation and retention of cholesterol in plasma membrane signaling rafts of hematopoietic cells, rendering these cells more sensitive to hematopoietic cytokines that drive monocyte precursor proliferation [70]. The number of circulating monocytes positively correlates with plaque size in mice and is thought to be an independent risk factor for cardiovascular disease in humans [71]. In mice, where monocytosis is dominated by the classical subset, this subset enters plaques that form in the aortic arch of Apoe^−/− mice at a greater frequency than the nonclassical subset [16, 69]. However, recruitment of nonclassical monocytes approaches the rate of classical monocyte entry into aortic sinus plaques (Figure 2) [72]. Thus, in contrast to acute infection or inflammation models, both monocyte subsets significantly contribute to plaque and the ratio at which they do so differs in different anatomic locations. It is uncertain if both subsets participate in disease progression from the onset or if the contribution of nonclassical monocytes increases later. In another model of chronic inflammation, chronic liver injury, only CCR2⁺ classical monocytes are recruited and appear to drive liver fibrogenesis [68]. Findings from atherosclerosis models, as well as this liver study, indicate that CCR2-dependent classical monocyte infiltration shapes disease progression during chronic inflammation.

Turning down the dial for recruitment of both monocyte subsets promotes regression

Recruited or locally differentiated nonclassical monocytes are thought to initiate repair of injured tissue [18, 19]. Thus, during reversal of atherosclerosis, one might hypothesize that the sustained recruitment of nonclassical monocytes would be desired. However, in a model of regression, in which apoE-deficient mice were treated with apoE-encoding vectors to reverse disease, both subsets of monocytes are halted from entering plaques [72]. Indeed, this cessation in recruitment accounted for the reduction in plaque macrophage burden (Figure 2) [72]. It remains possible, however, that certain aspects of “healing” during disease regression do require macrophages of a certain phenotype, but ongoing nonclassical monocyte recruitment may not be needed for remodeling and repair.

Concluding remarks

Monocyte recruitment to sites of infection, and in atherosclerosis, includes the recruitment of classical monocytes by chemokine signaling. A weakness in the current body of literature is that too few scenarios of chronic infection have been studied. Further investigation into monocyte behavior in chronic infection, such as in M. tuberculosis, may reveal a more prominent role for nonclassical monocytes, for example. Perhaps the function of nonclassical monocytes is in tissue repair, but few studies demonstrate that nonclassical monocytes are recruited to damaged tissue. Those that do are predominately in the setting of chronic atherosclerosis where there is no healing. Future studies are needed to address open questions and to bring the biology of monocyte subsets into sharper focus.