Nonclassical Patrolling Monocyte Function in the Vasculature

Abstract

Nonclassical patrolling monocytes are characterized by their unique ability to actively patrol the vascular endothelium under homeostatic and inflammatory conditions. Patrolling monocyte subsets (CX3CR1^highLy6C⁻ in mouse, and CX3CR1^highCD14^dimCD16⁺ in humans) are distinct from the classical monocyte subsets (CCR2^highLy6C⁺ in mouse, and CCR2^highCD14⁺CD16⁻ in humans) and exhibit unique functions in the vasculature and inflammatory disease. Patrolling monocytes function in a number of disease settings to remove damaged cells and debris from the vasculature, and have been associated with wound healing and the resolution of inflammation in damaged tissues. This review highlights the unique functions of these patrolling monocytes in the vasculature and during inflammation.

Introduction

Monocytes are the primary inflammatory cell type that infiltrates early atherosclerotic plaques ¹, and considerable evidence implicates monocytes as critical to the development of atherosclerosis ^2–7. Atherosclerotic lesion size correlate strongly with the number of circulating monocytes, and combined inhibition of the monocyte recruitment factors CCL2, CX3CR1 and CCR5 nearly abolishes atherosclerotic development in hypercholesterolemic mice ⁵. In mouse models of cardiac injury the depletion of monocytes leads to higher mortality and reduced ventricular function, suggesting that monocytes are necessary for proper wound healing and rebuilding of heart tissue ⁸. Thus, understanding the role of monocyte activities in the vasculature, cardiac injury and inflammatory disease is important to understanding the etiology of cardiovascular disease and potential therapeutic targeting of monocyte activities.

Currently three types of monocytes have been defined in humans and in mice ^9–14. Monocytes are characterized in humans by their positive expression of HLA-DR, CD11b, and differential expression of CD14 and CD16, and in mice by their positive expression of CD115 and CD11b, and differential expression of Ly6C and CD43. “Classical” monocytes (CD14⁺CD16⁻ in humans, and Ly6C⁺CD43⁻ in mouse) express high levels of the chemokine receptor CCR2 and can migrate to sites of injury and infection where they differentiate into inflammatory macrophages ^{9, 15}. In contrast, “nonclassical” monocytes (CD14^dimCD16⁺ in humans, and Ly6C⁻CD43⁺ in mouse) have high levels of the adhesion related receptor CX3CR1 and exhibit a unique ability to patrol the resting vasculature and remove debris ^{16, 17}. In addition, a 3^rd subset of “intermediate” monocytes (CD14⁺CD16⁺ in humans, and Ly6C⁺CD43⁺ in mouse) have been identified which have high expression of CX3CR1, but generally possess inflammatory characteristics. Some earlier studies in humans grouped CD16⁺ intermediate monocytes with CD14^dimCD16⁺ nonclassical patrolling monocytes, though current evidence suggests that intermediate monocytes are distinct and do not actively patrol the vasculature ¹⁰.

Nonclassical monocytes actively and continuously patrol the luminal side of the vascular endothelium both at steady state and during inflammation. Monocyte patrolling has been observed in the small blood veins and arteries of the dermis, mesentery, lung, and brain ^{10, 17, 18}. However, these cells are not restricted to the vasculature as nonclassical monocytes also undergo diadepesis and are found within atherosclerotic plaques, as well as within the parenchyma of multiple other tissues ^{15, 19–21}. This review focuses on the differentiation, recruitment, regulation, and physiological functions of “nonclassical” patrolling monocytes in the vasculature and in inflammatory diseases.

Patrolling monocyte development

Monocytes are short-lived mononuclear phagocytes that are continuously replaced throughout life from a common committed progenitor. In mice, monocytes mostly live only two days in circulation in the absence of inflammation, but a small fraction can live seven days or more ²². Nonclassical Ly6C⁻ monocytes are generally longer lived than classical Ly6C⁺ monocytes ²³. In the absence of classical Ly6C⁺ monocytes, the half-life of nonclassical Ly6C⁻ monocytes is increased from around 2.5 days to 11 days suggesting an increased surveillance function of nonclassical patrolling monocytes in the vasculature when the classical monocyte population has been disrupted or recruited to inflammatory sites ²³.

Monocytes develop predominantly in the bone marrow from a CD117⁺CD115⁺CX3CR1⁺ monocyte dendritic cell precursor (MDP) that also gives rise to splenic dendritic cells ²⁴. Recently a common monocyte progenitor (cMoP) subset of the MDP was described that has potential to produce both monocyte subsets but not dendritic cells, and are defined as CD117⁺CD115⁺Ly6C⁺Flt3^{− 25}. A number of studies support the notion that Ly6C⁺ monocytes give rise to Ly6C⁻ monocytes. BrdU pulse chase studies have shown rapid incorporation of the thymidine analogue into the DNA of Ly6C⁺ monocytes followed by a gradual displacement of the Ly6C⁺ population by Ly6C⁻ monocytes in the BrdU labeled fraction ²³. Adoptive transfer studies have shown that congenically labeled Ly6C⁺ monocytes give rise to Ly6C⁻ monocytes 1–3 days post transfer ^{21, 23, 25, 26}. Corresponding studies of human monocyte subset differentiation and lifespan have yet to be conducted.

Ly6C⁺ monocytes can give rise to Ly6C⁻ monocytes in vivo, however this does not exclude the existence of an alternative route for Ly6C⁻ monocytes development independent of the Ly6C⁺ population. Indeed, genetic evidence for this proposal exists. Two myeloid determining transcription factors, IRF8 or KLF4, both specifically regulate Ly6C⁺ monocyte production without effecting Ly6C⁻ monocyte numbers. Studies of either global IRF8^−/− mice, or fetal liver transplant of KLF4^−/− cells into irradiated wild type recipients, both report dramatically reduced numbers of Ly6C⁺ monocytes in the BM while retaining relatively normal Ly6C⁻ monocyte numbers^{27, 28}. These findings imply a pathway for Ly6C⁻ monocyte development that is independent of Ly6C⁺ monocytes, perhaps originating directly from the cMoP precursor. However, other explanations for this phenotype include the enhanced survival of Ly6C⁻ monocytes in these models. Therefore it remains possible that Ly6C⁻ monocytes derive from either blood or bone marrow Ly6C⁺ monocytes, from an independent bone marrow monocyte progenitor, or from a combination of all three of these scenarios. In order to resolve these issues a detailed understanding of the factors and pathways regulating the development and survival of both Ly6C⁺ and Ly6C⁻ monocyte populations will be required.

A major advance in our understanding of patrolling nonclassical monocyte differentiation was made with the discovery that the transcription factor Nur77, encoded for by the gene NR4A1, is absolutely required for Ly6C⁻ monocyte development^{16, 29}. Nur77 is highly expressed in patrolling monocytes, and patrolling monocytes are specifically missing in the blood, spleen, and bone marrow of Nur77 knockout mice ²⁹. The few patrolling monocytes remaining in the bone marrow of Nur77 knockout mice are arrested in S phase of the cell cycle and undergo apoptosis, implying that Nur77 functions as a master regulator of the differentiation and survival of nonclassical patrolling monocytes from myeloid progenitors in the bone marrow. Unlike the chemokine receptor-deficient models used to study monocyte function (discussed below), Nur77-deficient mice lack Ly6C⁻ monocytes both in the bone marrow and periphery. In the homologous human CD14^dimCD16⁺ population of patrolling monocytes there is also high Nur77 expression, and likely common function ^{10, 30}.

Mechanisms of monocyte recruitment, adherence, patrolling and survival in the vasculature

Mature classical and nonclassical monocytes must exit the bone marrow into the peripheral circulation in order to perform their duties. The migratory properties of monocytes are facilitated by the actions of chemokines, and both classical and nonclassical monocytes express a different repertoire of chemokine receptors enabling their differential mobilization from the BM and into tissues. Classical monocytes express high levels of CCR2 and migrate from the BM into the vascular circulation in response to its ligands CCL2, CCL7 and CCL12 ^31–33. Consequently CCR2^−/− mice have reduced Ly6C⁺ monocytes in the periphery, but an increased frequency of Ly6C⁺ monocytes in bone marrow, whereas the distribution of Ly6C⁻ monocytes in these mice is relatively normal.

Less is known about the molecular pathways governing the emigration of nonclassical monocytes from the bone marrow and their recruitment to sites of vascular damage or inflammation (See figure 1). Ly6C⁻ monocytes are recruited to sites of atherosclerosis, which was reduced ~40% with CCR5 blockade, suggesting that CCR5 expression is at least partially responsible for nonclassical monocyte recruitment ¹⁹. However, CCR5 expression is found at low levels on patrolling monocytes putting into question if this recruitment is directly acting on patrolling monocytes. Recent insight on patrolling monocyte recruitment has been gained from functional studies of sphingosine-1-phosphate (S1P) receptors, which are major regulators of leukocyte activation and trafficking. Administration of the S1P agonist FTY720 facilitates the internalization and degradation of S1P receptors and leads to a reduction in monocyte egress from the BM and spleen into the vasculature ^{34, 35}. Localized FTY720 application can selectively enhance recruitment of patrolling monocytes to ischemic vessels in a S1PR3 dependent manner, suggesting S1P can regulate recruitment of patrolling monocytes ³⁶. Furthermore, S1PR5 is highly expressed on patrolling monocytes and is important for their homeostasis. S1PR5^−/− mice have greatly reduced blood-circulating and splenic Ly6C⁻ monocytes, but have normal numbers in the bone marrow ³⁷. This effect was attributed to defective cell intrinsic emigration of S1PR5-deficient monocytes from bone marrow, as the survival of intravenously transferred S1PR5-deficient Ly6C⁻ monocytes were comparable to wild type controls.

Figure 1. Patrolling and Recruitment of Nonclassical Monocytes.

Resting nonclassical monocytes actively patrol the vasculature at a speed of ~12μm/min in a manner independent of blood flow. Nonclassical monocyte patrolling requires LFA1 binding with ICAM1 on vascular endothelial cells. Nonclassical monocyte patrolling and activation is also partially CD11b-dependent. In response to inflammation, vascular damage or infection, there is a release of chemoattractant factors from either endothelial cells, damaged tissues or other recruited immune cells that attract patrolling monocytes. These chemoattractant factors include CX3CL1, TLR7 agonists, CCL5 and S1P, which monocytes can respond to via intrinsic expression of cognate receptors CX3CR1, TLR7 and possibly CCR5 or S1PR. Patrolling monocytes are then either recruited locally to sites of vascular injury or can enter areas of inflammation such as atherosclerotic plaques, nephritic kidneys or arthritic joints.

Patrolling monocytes interact with the vascular endothelium via the β2 integrin, lymphocyte function-associated antigen-1 (LFA-1), and the chemokine receptor CX3CR1 ¹⁷. A pioneering study by Auffrey et al described the characteristic “patrolling” behavior that nonclassical monocytes display on the luminal side of the vascular endothelium ¹⁷. Human CD14^dim monocytes have also demonstrated patrolling behavior when infused into mice ¹⁰. This patrolling is a much slower process (~12μm/min) than the rolling process of classical monocytes and is independent of the direction of blood flow. Since in vivo analysis of patrolling function of monocytes is difficult to assess in humans, the majority of studies examining nonclassical monocyte patrolling function and activities have used mouse models. Patrolling is in part dependent on the expression of LFA-1, which is highly expressed by nonclassical monocytes. Antibody blockade of either subunit of the LFA-1 complex, CD11a or CD18, leads to the rapid and prolonged disassociation of non-classical monocytes from the endothelium in vivo and abolishes patrolling behavior. The LFA-1 ligands ICAM1 and ICAM2 act in a redundant manner to facilitate LFA-1-dependent adhesion. ICAM1/2^−/− mice have a greatly reduced frequency of nonclassical monocyte adhesion to the vasculature, which is a phenocopy of the CD11a^−/− (ITGAL) mice¹⁶. The capacity for other integrin complexes to mediate patrolling has not been shown, however studies to date cannot rule this out. Interestingly neutrophil arrest in the mouse is entirely dependent on LFA-1, whereas human neutrophils utilize both LFA-1 and MAC-1 (ITGAM/CD11b) ^{38, 39}. Whether or not this is also the case in monocyte patrolling is not known. Further, neutrophil arrest requires integrin conformational changes that allow for high affinity binding. This process involves cytoskeletal adaptor proteins Talin-1 and Kindlin-3 ⁴⁰. The requirement for integrin high affinity binding conformational changes in monocyte patrolling as well as the requirement for integrin activating adaptor proteins is not known and will be important to understanding how the process of patrolling takes place.

Both classical and non-classical monocyte subsets express the chemokine receptor CX3CR1, although patrolling monocytes express it at much higher levels ⁹. CX3CR1^high patrolling monocytes respond to CX3CL1 (also known as fractalkine), a chemokine that is expressed in tissues and can be used as an apoptotic cell recognition receptor. CX3CR1 deletion reduces the patrolling of Ly6C⁻ monocytes ¹⁷, and recruitment to the spleen during bacterial infection⁴¹. CX3CR1 also provides a survival signal to Ly6C⁻ monocytes. CX3CR1^−/− mice have fewer circulating Ly6C⁻ monocytes, which has been attributed to reduced monocyte survival resulting from CX3CR1-dependent expression of the anti-apoptotic protein BCL2 ⁴². In this context it is interesting to note that global IL17RA-deficiency leads to a similar reduction in Ly6C⁻ monocyte survival ⁴³, and that BCL2 is an IL17 target gene in fibroblast-like synoviocytes ⁴⁴. However a causal link between IL17RA-dependent signaling and BCL2 expression in patrolling monocytes has not been established. Although CX3CR1 is dispensable for the initiation of monocyte patrolling, CX3CR1-CX3CL1 interactions stabilize patrolling in a CD11b-dependent manner ¹⁶. Specifically, TLR7 agonists mediate an increase in epithelial CX3CL1 expression that led to a direct increase in the retention time of patrolling monocytes at the epithelial surface.

Infection and tissue damage are not the only environmental cues that result in monocyte mobilization. Other physical factors such as exercise and stress have been demonstrated to recruit monocytes into circulation. Exercise leads to a more than a four-fold increase of nonclassical monocytes in the blood ^{45, 46}. This exercise induced increase can be inhibited by β-adrenergic receptors blockade suggesting a catecholamine-dependent mechanism of recruitment ⁴⁶. Studies have shown that monocytes express β-adrenergic receptors, and rapidly appear in circulation from vascular marginal pools, a process called demargination, in response to catecholamine release ⁴⁷. Chronic stress also induces a β3-adrenergic receptor dependent increase in monocyte progenitors in bone marrow, and classical monocytes in circulation, promoting atherosclerotic development ⁴⁸. In addition to their progenitors, monocytes themselves also express β-adrenergic receptors, thus future studies to learn how stress and exercise induced β-adrenergic receptor signaling directly affects monocyte function in the vasculature and atherosclerotic development should be of interest.

Activation activities of patrolling monocytes

Classical monocytes circulate in the bloodstream and respond to inflammatory cues such as microbial products, cytokines and chemokines produced by infected and/or damaged tissue. In the presence of inflammatory stimuli, circulating inflammatory monocytes can quickly extravasate into affected nonlymphoid tissues in a CCR2–CCL2-dependent manner, where they differentiate into macrophage and dendritic cell subsets³². Upon stimulation these monocytes produce inflammatory cytokines such as IL12, NO, TNF-α and IL-1 and can differentiate into TNF-α producing inflammatory macrophages and DCs (Tip-DCs) ⁴⁹. These responses result in increased immune activation and bacterial clearance, but also significant tissue damage.

Patrolling monocytes, on the other hand, produce only low levels of proinflammatory cytokines in response to bacteria-derived stimuli such as LPS and PAM3CK4 but high levels of anti-inflammatory, and pro-wound healing factors such as IL1 receptor antagonist, IL10R, apolipoproteins ApoA and ApoE, and CXCL16 in studies using CD14^dimCD16⁺ human monocytes ¹⁰. Patrolling monocytes express lower levels of CD14 (a TLR4 co-receptor), suggesting a weaker LPS response ⁵⁰. However, little is known about the ability of patrolling monocytes to extravasate and differentiate into effector macrophage populations in non-homeostatic conditions. Interestingly, murine patrolling monocytes are the first immune cell detected to extravasate into the peritoneum following listeria monocytogenes infection, and produce high levels of TNF α showing that these cells are not always anti-inflammatory counterparts to the proinflammatory classical monocytes ¹⁷. In addition, CD16⁺ human monocyte stimulated with tumor cells showed enhanced production of TNFα, IL12 and NO compared to classical monocytes, suggesting that nonclassical monocytes may be an important subpopulation of monocytes involved in antitumor response ⁵¹. Further, both human and murine patrolling monocytes strongly respond to viruses and nucleic acids via a TLR7/8-MEK pathway by producing proinflammatory cytokines like TNFα and IL1β as well as chemokines CCL3, CCL5, and CXCL10 ^{10, 16}. Nonclassical monocytes can recruit neutrophils that function to kill damaged or infected cells in response to nucleic acid-derived TLR7-mediated “danger signals” within the vasculature¹⁶. TLR7 activation results in intravascular retention of Ly6C⁻ monocytes, and subsequent removal of cellular debris by the retained monocyte. It is not known if other cell types are directly recruited by patrolling monocytes in response to CCL3, CCL5, and CCL10 production, though this is of high interest to the field. In the absence of inflammatory cues, patrolling monocytes scan the vasculature and uptake microparticles along the endothelium. Thus these cells function as “intravascular housekeepers” that scavenge microparticles, remove cellular debris from the vasculature, and can vigorously respond to cell damage and infection. Therefore, it is possible that nonclassical monocytes function as terminally differentiated cells.

In addition to uptake of cellular debris, patrolling monocytes engulf apoptotic cells and can then cross present antigen from the engulfed cell to T cells in the spleen. Interestingly, these antigen presenting Ly6C⁻ monocytes express high levels of PDL1 following apoptotic cell engulfment and suppress endogenous antigen-specific T cell responses ⁵². Thus Ly6C⁻ monocytes may contribute to self-antigen tolerance. Studies in human monocytes show that CD16⁺ nonclassical monocytes express ILT4, a receptor that binds HLA-G. The ILT4/HLA-G interaction is associated with fetal tolerance during pregnancy ⁵³. Taken together these studies in mouse and man suggest that patrolling monocytes function in some capacity in antigen presentation to T cells, and may aid in eliciting a tolerogenic response.

Comparisons of human and mouse patrolling monocytes

Much of our knowledge about the functions of human non-classical monocytes is derived from comparative transcriptomic analyses. A number of studies have compared gene and protein expression between monocyte subsets in humans and mouse ^{10, 50, 54, 55}. These expression studies identify a number of unique genes and activities in the patrolling monocyte population (See table 1). Two independent studies have provided transcriptional array data of all three human monocyte subsets and show that the ‘intermediate’ (CD14⁺CD16⁺) monocytes share a transcriptional repertoire that is in between that of both the nonclassical (CD14^dimCD16⁺) and classical (CD14⁺CD16⁻) subsets ^{10, 56}. While there is disagreement as to which subset the intermediate monocyte is more closely related, this might simply reflect differences in placement of the gating used to sort this population by flow cytometry. Nevertheless, the intermediate monocyte population only represents a minor frequency (5–10%) of the total nonclassical subset ⁵⁰, thus the inclusion of this population should have little confounding influence on the interpretation of data comparing the mouse and human classical and non-classical subsets.

Table 1.

Genes up-regulated in nonclassical monocytes in both humans and mice grouped by function

Cellular function	Gene Name	Reference
Lineage-Determining Receptor	CSFR1	50
Survival/Differentiation	Nur77 (NR4A1), BCL2A1A, DUSP5, HES1, OCT2 (POU2F2), TCF7L2	29, 50, 58
Adhesion and Patrolling	CX3CR1, CD43 (SPN), CD11a (ITGAL/LFA-1), CD11c (ITGAX), CD31 (PECAM1), RRAS	16, 50
Immune Regulation	CD16 (Fcgr3), TNF, IL10R, IL1B*, IL6*, IL1RA*, TGFBR3, LAIR1, LTB, KLRD1, GZMA	50, 10, 16, 58
Viral Immunity	TLR7, IFITM1, IFITM2, IFITM3	10, 54, 58

^*In response to TLR7 stimulation

Ingersoll et al published a detailed comparative analysis between the human classical CD14⁺CD16⁻ and nonclassical CD14⁻CD16⁺ subsets, and the mouse Ly6C⁺ and Ly6C⁻ populations ⁵⁰. This analysis demonstrated statistically significant similarities in the global gene expression profiles of the mouse and human classical and non-classical subsets alike ⁵⁰. Cros et al independently confirmed these findings and went on to show that human CD14^dimCD16⁺ monocytes patrol the vasculature of lymphoid deficient RAG1^−/−IL2RG^−/−CX3CR1^GFP mice ¹⁰. Genomic studies also verify that nonclassical human monocytes have increased metabolic activities and efficient cytoskeletal dynamics that are indicative of actively patrolling cells ^{54, 55}. These data provide compelling evidence that the human CD14^dimCD16⁺ population represent a functionally orthologous cell subset to the mouse Ly6C⁻ patrolling monocyte.

Conserved up-regulated genes encode many of the proteins that impart the key mechanistic features of the nonclassical subset, these include the patrolling-associated integrins LFA-1 and MAC-1 (ITGAM/CD11b), the chemokine receptor CX3CR1 as well as the inflammatory cytokine TNF. In total 63 genes were more abundant in both human and mouse nonclassical monocytes relative to their classical counterparts ⁵⁰. The roles of most of these genes in monocyte biology are yet to be determined, but future studies will likely provide significant new insight into the functions of these cells. Other functional similarities between the mouse and human non-classical monocytes include an attenuated response to stimulation with the TLR4 agonist LPS (due to relatively low LPS/TLR4 co-receptor CD14 expression), and a heightened response to the viral DNA sensor TLR7 (due to relatively high TLR7 expression)^{10, 16}.

Although global similarities have been drawn between the human CD14^dimCD16⁺ monocytes and Ly6C⁻ monocytes there are notable differences. For example the transcription factor PPARγ and known target genes including CD36 and FABP4 are higher in murine Ly6C⁻ monocytes, but not observed in human CD16⁺ monocytes. Importantly, mouse nonclassical monocytes also expressed higher levels of genes involved in apoptotic cell uptake including Tgm2, Treml4, CD36, and CD51 whereas these genes were more highly expressed by human CD14⁺ classical monocytes, suggesting potential differences in apoptotic cell recognition ⁵⁰. MHCII expression is also low/absent on mouse Ly6C⁻ patrolling monocytes, but high on human patrolling monocytes, which may suggest better presentation of certain antigens by the human subset ^{16, 50}. In total Ingersoll et al. found 33 genes to be reciprocally regulated between the human and mouse subsets ⁵⁰. Thus, some caution must be made in making direct comparisons in patrolling monocyte function between species.

A recent ChIP-Seq analysis of primary human classical and non-classical monocytes has provided mechanistic insight into transcriptional differences between these cell types ⁵⁴. Transcription factor binding motifs for the CEBP and ETS factors are enriched in the enhancers of all monocytes. This results is expected as these motifs are bound by the myeloid pioneer transcription factors CEBP and PU.1 which collaborate to drive the myeloid gene expression program ⁵⁷. Transcription factor binding sites for KLF, IRF and NR4A transcription factors were more abundant in the non-classical subset. Currently no obligatory roles for KLF or IRF transcription factors have been implicated in non-classical monocyte function or development although KLF2 is more abundant in CD16⁺ monocytes ⁵⁸, however the nuclear receptor transcription factor NR4A1 is crucial for Ly6C⁻ monocyte development ²⁹. Studies of mouse monocyte enhancer activity will inform whether or not the same enrichment for IRF and KLF factors is conserved between species.

Roles in disease

The unique ability of nonclassical monocytes to actively patrol the vasculature and potential capacity to resolve inflammation make them attractive targets for disease therapy. Generally, nonclassical monocytes are thought to be involved in the resolution of inflammation and differentiate into resident macrophage populations that work to heal wounding and resolve inflammation¹⁷. Studies have suggested that patrolling monocytes can preferentially differentiate into CD11c⁺ resident lung macrophages, implying a specialized role for these effector cells in the lung ^{20, 42}. There is also some evidence that patrolling monocytes can differentiate into resident or anti-inflammatory “M2-like” macrophages, but the evidence for this is far from definitive and may be specific for the tissue and/or stimulus. Others have found that recruitment of patrolling monocytes and subsequent monocyte-mediated inflammatory responses (including TNF and ROS production) in certain tissues may potentially aggravate autoimmune diseases such as lupus nephritis and arthritis induced joint inflammation. These findings are summarized below and in Table 2.

Table 2.

Observed activities of nonclassical monocytes in disease.

Disease	Observed Activity	Reference
Atherosclerosis	CCR5 mediated recruitment to atherosclerotic plaques and reduced accumulation with statin treatment	19, 59
	Number of circulating cells positively correlated with atherosclerotic lesion size	5
	Regulated survival and recruitment to atherosclerotic plaques by CX3CR1	17, 42
	Nur77 hematopoietic knockout mice with absence of patrolling monocytes develop increased atherosclerosis and inflammatory macrophage content	30
	Remove lipids including oxLDL from circulation, and increased numbers in circulation in response to oxLDL treatment	60
	Preferential uptake of oxLDL by CD16+ monocytes in hypercholesterolemic patients	61
	Inversely correlated with high-density lipoprotein (HDL) cholesterol levels and associated with increased APOE4 expression in hypercholesterolemic patients	72
	Percentage of cells in circualtion correlated with small HDL levels	73
	Circulating CD16+ monocyte levels positively correlated with vulnerable plaques in coronary heart disease patients	68, 69, 70
Myocardial infarction (MI) and Vascular Wounding	Recruitment during the reparative phase of MI	74, 77
	Higher numbers in circulation of MI patients that did not develop ventricular thrombus formations	78
	Nur77 hematopoetic knockout mice have adverse cardiac remodeling post MI	79
	Recuitment associated with the recovery of vascular flow and a regenerative phenotype in a hind-limb ischemia model	75
	S1PR3 mediated recruitment and correlation with arteriogenesis in ischemic microvessels	36
	Redundant role in progression and recovery of ischemic stroke	81
Neurological Disease and Damage	Attracted to and actively remove amyloid beta peptides from brain vasculature	82
	Differentiate into perivascular macrophages and important role in maintaining blood brain barrier	18
	Prevent excito-toxicity and neuronal cell death	83
	Beneficial recruitment to the injured spinal cord	84
Lupus and Kidney Disease	Accumulation of nonclassical monocytes to glomerular vessels	10, 87
	Accumulation Associated with elevated levels of CX3CL1, proliferative glomerular lupus nephritis lesions, and disease activity	87
	Accumulate, recruit neutrophils, and remove damaged endothelial cell in the kidney vasculature in response to a TLR7-induced “danger signal”	16
Arthritis	Critical for the initiation and progression of sterile joint inflammation, but derived macrophages may also be important for arthritic resolution	92
	Increased circulating levels in rheumatoid arthritis patients associated with elevated levels of C-reactive protein, rheumatoid factor and more active disease	93, 94

Atherosclerosis

Currently, little is known about the importance of nonclassical monocytes and their role in atherosclerotic development. Tracking studies have demonstrated that both classical and nonclassical subsets can enter atherosclerotic plaques in mice ¹⁹. Atherosclerotic lesion size has been shown to positively correlate with the number of both Ly6C⁻ and Ly6C⁺ circulating monocyte subsets ⁵. Quantitative intravital imaging has also demonstrated a substantial accumulation of nonclassical mouse monocytes in lesions that can be reduced with statin treatment ⁵⁹. In blood draws from APOE deficient mice on high fat diet there is an increase in the percentage of classical monocytes and decrease in the nonclassical population¹⁹. However, low numbers of patrolling monocytes analyzed in blood draws could be due to increased adherence and patrolling of nonclassical monocytes in the vasculature, and not due to a true reduction in cell number, though this has yet to be properly examined.

Experiments suggest that the Ly6C⁺ and Ly6C⁻ monocyte subsets use different chemokines for recruitment to atherosclerotic lesions ¹⁹. Ly6C⁻ monocyte recruitment to sites of atherosclerosis was reduced ~40% with CCR5 blockade, whereas Ly6C⁺ monocyte recruitment was reduced ~50% with either CCR5, CCR2, or CX3CR1 blockade ¹⁹. These findings suggest that CCR5 expression is at least partially responsible for nonclassical monocyte recruitment, but there are likely other recruitment factors yet to be identified.

Nonclassical monocytes actively patrol the vasculature and have been shown to phagocytize microparticles and mediate the removal of damaged endothelial cells in the vasculature ^{10, 16}. Nonclassical Ly6C⁻ monocytes can preferentially scavenge and accumulate lipids including oxLDL from the vasculature, and increase in numbers in response to oxLDL treatment ⁶⁰. Likewise in patients with familial hypercholesteremia, CD16⁺ monocytes preferentially uptake oxLDL ⁶¹. Therefore it is easy to speculate that nonclassical monocytes may work to repair the vasculature or remove lipids from circulation under damaging inflammatory atherosclerotic conditions.

It has also been suggested that nonclassical monocytes may differentiate into resident anti-inflammatory “M2-like” macrophage populations that could suppress inflammatory conditions in atherosclerotic plaques ^{49, 62}. However, a recent study suggests that while monocyte recruitment is important for early plaque development, it is local macrophage proliferation and not continual monocyte recruitment that dominate atherosclerotic plaque growth ⁶³. This provocative possibility, in its early stage of investigation, merits closer study.

Work in our laboratory has identified that in the absence of Nur77 and patrolling monocytes, the remaining macrophage populations are polarized to a more inflammatory phenotype leading to increased atherosclerotic development ³⁰. Macrophages from Nur77^−/− mice express relatively high levels of tumor necrosis factor-α (TNFα) and nitric oxide, and low expression of Arginase-I upon inflammatory stimuli. Three independent studies have confirmed a direct role of Nur77 in resolving myeloid cell-mediated inflammation and suppressing atherosclerotic development ^{30, 64, 65}. A conflicting study found no role of Nur77 in atherosclerotic development and may represent an outlier in the literature ⁶⁶. In the majority of studies linking Nur77 deficiency to atherosclerotic development, Nur77 deficient macrophages are polarized toward a pro-inflammatory phenotype, exhibiting increased IL-12 and nitric oxide synthesis in response to activation by toll-like receptor (TLR) agonists. This increase in macrophage activation is driven by enhanced NF-κB activity and TLR expression in the absence of Nur77 ^{30, 67}. These results suggest that Nur77-dependent monocytes may be atheroprotective, and that the nonclassical monocytes themselves, or their derived effector cells, may suppress inflammatory macrophage activity. We are currently investigating if these differences in macrophage activity are related to the absence of a patrolling monocyte derived anti-inflammatory macrophage subset or if these effects are related to intrinsic Nur77 activity in macrophages.

In humans, increased numbers of circulating CD16⁺ monocytes are associated with increased coronary heart disease ⁶⁸. CD16⁺ monocyte levels are also positively correlated with vulnerable plaques in coronary heart disease patients, and levels of CD16⁺ monocytes are significantly decreased in patients receiving statin treatment ^{69, 70}. Unfortunately, these human studies did not distinguish between CD16⁺CD14^dim nonclassical and CD16⁺CD14⁺ generally more inflammatory intermediate monocyte populations. A recent study of over 900 patients has suggested that it is mainly the CD16⁺CD14⁺ intermediate monocytes that are positive predictors for cardiovascular events, while the CD16⁺CD14^dim nonclassical monocyte subset showed no correlation⁷¹.

In hypercholesterolemic patients, the number of nonclassical monocytes is inversely correlated with high-density lipoprotein (HDL) cholesterol levels and is associated with increased APOE4 expression, a factor related to higher plasma cholesterol ⁷². Percentage of nonclassical monocytes is also positively associated with small HDL levels ⁷³. Nevertheless, the association between nonclassical monocytes and atherosclerosis does not necessarily imply causality. Research into the functional relationships between patrolling monocytes, production of inflammatory or anti-inflammatory factors, and ability to remove lipids or repair the vasculature in cardiovascular disease is needed to establish positive effector cell associations.

Myocardial infarction and wounding

With respect to myocardial infarction (MI) and wound healing, patrolling monocytes have been associated with reparative and proangiogenic effects ^74–76. In mouse models of MI the depletion of monocytes leads to higher mortality and reduced ventricular function suggesting that monocytes facilitate wound healing and cardiac remodeling ⁸. Classical monocytes migrate into infarcted heart tissue during the early inflammatory phase of injury, and nonclassical monocytes are sequentially recruited at day 5 during the reparative phase ⁷⁴. Similar findings were observed in patients where nonclassical monocytes peaked at 5 days post MI in the reparative phase, suggesting a possible reparative role of nonclassical monocytes ⁷⁷. MI patients that did not develop ventricular thrombus formations had higher circulating nonclassical monocytes further suggesting a protective function of the subset ⁷⁸. Mice in which bone marrow derived cells lack Nur77 expression serve as a model for the absence of patrolling monocytes. These mice display adverse cardiac remodeling post MI, again highlighting a role for patrolling monocyte in post MI wound healing ⁷⁹. However, other anti-inflammatory effects of Nur77 on classical monocyte derived macrophages may also affect the MI outcome in this model. Further research is needed to examine the mechanisms by which patrolling monocytes may respond and repair damaged cardiac tissue. For example, it will be interesting to learn whether patrolling monocytes contribute to the local cardiac macrophage pool or if they function as terminally differentiated cells within the myocardium.

In a hind-limb ischemia model, proangiogenic nonclassical monocytes were associated with recovery of vascular flow and a more regenerative phenotype ⁷⁵. Similarly, S1PR3 recruited nonclassical monocytes were correlated with arteriogenesis in ischemic microvessels ³⁶. With laser-induced focal tissue damage of the ear dermis in mice, CX3CR1^high nonclassical monocytes quickly align along collagen fibers at the outer edges of the wounding, in a prime location for repairing tissue ⁸⁰. However, another study concluded that in a cerebral hypoxia-ischemia model patrolling monocytes were redundant in the progression and recovery of ischemic stroke ⁸¹. Additional research is needed to determine if these conflicting findings can be attributed to patrolling monocyte function in specific tissues or general differences in ischemic models. Combined these data generally demonstrate a beneficial effect of nonclassical monocytes in vascular repair and restoring organ function.

Neurological diseases and damage

Nonclassical Ly6C⁻ monocytes patrol the brain and nervous system vasculature and have been associated with a variety of generally protective activities ^{18, 82}. Nonclassical patrolling monocytes are attracted to amyloid beta peptides and have been observed actively removing amyloid beta from the brain vasculature ⁸². This suggests that patrolling monocytes may be potential therapeutic targets for reducing amyloid beta deposits associated with Alzheimer’s disease. Live imaging of the brain has also revealed Ly6C⁻ CX3CR1^high monocyte differentiation into perivascular macrophages, a cell that is important for maintaining the blood-brain barrier and preventing damaging inflammatory cell influx into nervous tissue ¹⁸. Nonclassical monocytes were attracted during endotoxemia by TNFα, IL1β and angiopoietin-2.

Other possible roles of patrolling monocytes in the nervous system include preventing excito-toxicity and beneficial recruitment to the injured spinal cord. Absence of patrolling monocytes using hematopoietic deletions of Nur77 or CX3CR1 exacerbated excito-toxicicity and neuronal cell death ⁸³. Another study demonstrated that Ly6C⁻ CX3CR1^high monocytes are actively recruited via the choroid plexus to help recovery of spinal cord injuries. ⁸⁴. Taken together, these finding reveal multiple protective roles of patrolling monocytes in the vasculature of the brain and nervous system.

Lupus and kidney disease

Glomerulonephritis is the leading cause of kidney failure and death in Lupus patients. Kidney glomerulus inflammation results from an increased duration and retention of migratory leukocytes and patrolling monocytes in a CD11b dependent manner ^{16, 85, 86}. Accumulation of nonclassical monocytes to glomerular vessels in the kidneys of Lupus patients has been documented ^{10, 87}. Recruitment to the kidney vasculature is at least partially mediated by CX3CL1 (fractalkine) expression in the damaged kidney. Elevated levels of CX3CL1 in the glomerular vessels of lupus patients is associated with recruitment of CD16⁺ monocytes, proliferative glomerular lupus nephritis lesions, and disease activity ⁸⁷. Interestingly, glucocorticoid therapy had a tendency to decrease both glomerular fractalkine expression and CD16⁺ monocyte numbers. Nonclassical human monocytes can preferentially make TNF and CCL3 in response to serum from Lupus patients, suggesting an inflammatory and active state of these cells under Lupus-like conditions ¹⁰. The pattern recognition receptor TLR7 plays important roles in autoimmune responses directed against DNA- and RNA-containing nuclear antigens and the pathogenesis of Lupus in susceptible hosts ^88–91. TLR7 agonists can specifically induce nonclassical monocytes to produce proinflammatory cytokine including TNF, CCL3, IL6, IL1beta, and CXCL1 ¹⁶. In response to TLR7-induced “danger signals” patrolling monocytes accumulate in the kidney microvasculature which recruit neutrophils and work to remove damaged endothelial cells ¹⁶. These findings suggest that though patrolling monocytes may initially be recruited to protect the kidney vascular endothelium, they could also contribute to tissue damage in susceptible individuals.

Arthritis

Nonclassical monocytes have also been associated with both the initiation and resolution of autoimmune joint inflammation ⁹². An elegant study by Misharin et al. demonstrated that Ly6C⁻ monocytes are actively recruited to injured joints from the vasculature, can differentiate to inflammatory macrophages in the joint, and are critical for the initiation and progression of sterile joint inflammation in mouse models ⁹². However, with the development of arthritis Ly6C⁻ monocyte-derived macrophages shift to an alternatively activated phenotype, which promote the resolution of joint inflammation.

In humans nonclassical monocyte recruitment is likewise associated with rheumatoid arthritis. Nonclassical monocytes in blood were significantly increased in patients with rheumatoid arthritis compared to healthy controls ^{93, 94}. Patients with elevated nonclassical monocyte were associated with more active disease, and elevated levels of erythrocyte sedimentation rates, C-reactive protein and rheumatoid factor ⁹⁴. Patients responding to therapy developed lower nonclassical monocyte frequencies, but nonresponders increased their frequency. Additionally, the expression of the chemokine receptors CCR5, CCR1 and ICAM-1 were higher on nonclassical monocytes from patient with active arthritis ⁹⁴. These finding show that nonclassical monocytes infiltrate inflamed joints and may contribute to both the induction and resolution of joint inflammation.

Potential targeting and future research

Though many interesting functions have been associated with patrolling monocytes, the kinetics of their recruitment and capacity to differentiate into effector macrophage populations in inflammatory disease are still uncertain. Furthermore, the extent to which patrolling monocytes directly participate in the resolution of inflammation and the specific populations of macrophages/myeloid effector cells (if any) that derive from patrolling monocytes is currently unknown. Better in situ techniques and models to distinguish monocyte and derived macrophage populations in tissues are needed to verify the extent of which nonclassical patrolling monocytes contribution to the myeloid compartment in inflammatory disease. Other functional characteristics of both monocyte subsets including differences in their respective capacities to proliferate, particularly in response to insult, are still unknown, and critical to our understanding of these cells.

Specifically targeting nonclassical monocyte production and activities via Nur77 mediated regulation or liposomes may be of therapeutic benefit. Directed recruitment of nonclassical monocytes by S1P, TLR7 agonists, CX3CL1 or CCL5 may also be of interest to enhancing repair of damaged vascular or tissues. Interestingly, for possible targeting of aberrant nonclassical monocyte activities, it has been shown that glucocorticoid treatment can deplete non-classical monocyte populations within a few days of treatment, and may contribute to the mechanism of glucocorticoid mediated immune suppression ^{95, 96}. Additional research is needed to delineate the inherent function of patrolling monocyte subsets in inflammatory diseases.

Significance.

Recent studies have identified at least two distinct monocyte sub-populations within the circulation—classical and nonoclassical patrolling monocytes. While many studies have examined the role of classical monocytes within the vasculature and in response to inflammatory diseases, comparatively few have assessed the unique activities of patrolling monocytes. Patrolling monocytes function to remove damaged cells and debris from the vasculature, and have been associated with wound healing and the resolution of inflammation in damaged tissues. This review summarizes the current knowledge and unique functions of nonclassical patrolling monocyte in the vasculature and during inflammatory disease.