Unravelling Mononuclear Phagocyte Heterogeneity

PeRSPeCtiveS

phenotypical modulation7. Even in the viewpoint absence of inflammation and infection, macrophage populations8 display bewildering Unravelling mononuclear phagocyte heterogeneity. For example, numerous mono nuclear phagocyte populations are found in heterogeneity the mouse spleen, including red and white pulp macrophages, ‘reservoirs’ of mobilizable monocytes9, marginal zone metallophilic Frédéric Geissmann, Siamon Gordon, David A. Hume, Allan M. Mowat and macrophages and outer zone macrophages, Gwendalyn J. Randolph and further complexity arises from the presence of resident, motile and migratory Abstract | When Ralph Steinman and Zanvil Cohn first described dendritic cells activated DCs in the same organ10. (DCs) in 1973 it took many years to convince the immunology community that Another problem is the marked hetero these cells were truly distinct from macrophages. Almost four decades later, the geneity in the turnover and variable lifespan DC is regarded as the key initiator of adaptive immune responses; however, of macrophages — hours to weeks — distinguishing DCs from macrophages still leads to confusion and debate in the depending on stimulation, programmed cell death and/or injury. In the mouse, field. Here, Nature Reviews Immunology asks five experts to discuss the issue of genetic manipulation and intravital imaging heterogeneity in the mononuclear phagocyte system and to give their opinion on methods make it possible to study direct the importance of defining these cells for future research. precursor–product relationships. However, the dynamics of tissue populations in humans remain largely unexplored. Myeloid cells are infamously Recent in vivo experiments in mice have heterogeneous. Do you think that this increased our understanding of the develop Gwendalyn J. Randolph. Yes, there is has led to confusion in the field when ment and functions of DC and macrophage confusion. However, I think the confusion discriminating between macrophages and subsets1,12,15,64,65 and should reduce confu can be resolved when the right considera dendritic cells (DCs)? sion. However, despite this progress in mice, tions are given. First, confusion should be corresponding human subsets are yet to minimal when spleen and lymph nodes are Frédéric Geissmann. Most cells types be characterized. Numerous studies have considered. In these locations, where his in the body are identified by their origin, attempted to recapitulate some of the hetero torically the pioneering work on DCs and anatomical location, function, and pheno geneity of human DCs and macrophages to some extent macrophages has centred, it type. Myeloid cells are arguably no more in vitro. However, in vitro differentiation of is clear that CD11chi mononuclear phago heterogeneous than other cell types (such blood monocytes or bone marrow cells does cytes are DCs rather than macrophages. An as lymphoid cells or epithelial cells), and not recapitulate in vivo differentiation. For abundance of recent evidence confirms that therefore heterogeneity may not be solely example, Langerhans cells and microglial cells spleen and lymph node DCs are function responsible for the confusion between may not renew from bone marrow2–4. ally distinct from macrophages, do not arise macrophages and DCs. from monocytes, and share fewer properties Confusion has arisen in part from the Siamon Gordon. As far as mononuclear with monocytes than macrophages do11–17. use of nonspecific cell surface markers, myeloid cells are concerned, I recognize a So here is where I see the contemporary such as CD11c (also known as αX integrin), family of closely related cells with a common confusion arising. DCs and macrophages in as a surrogate for function, anatomical origin, which branches into irreversibly dif nonlymphoid organs are poorly character location, and lineage. This has led to the ferentiated sublineages, such as macrophages, ized compared with their lymphoid tissue name ‘DC’ being used for several cell DCs and osteoclasts5. counterparts. At last, work in vivo on macro types, including bona fide classical DCs, as Part of the confusion in the field arises phage and DC biology in nonlymphoid well as activated monocytes, tissue macro from the adaptability and plasticity of organs is intense. Many of us have assumed phages, interferon (IFN)producing cells macrophages. As a result of their extensive that the same set of markers that defini and even some natural killer (NK) cells and receptor repertoire and versatile biosynthetic tively distinguish DCs from macrophages eosinophils. capacity, they are highly responsive to the in lymphoid organs will also apply in non microheterogeneity present in different lymphoid organs, but it is turning out that Confusion has arisen in part tissue environments6. Furthermore, follow this premise is not true. So, I think the main ing constitutive or inflammationinduced confusion arises when scientists assume that from the use of nonspecific cell migration between different compart what is true in lymphoid organs is also true in surface markers ments of the body, they undergo marked nonlymphoid organs and what is true under

the contributors mononuclear phagocytes in the gut wall, for example, are monocytederived, F4/80+ and Frédéric Geissmann gained a medical degree in 1996 and a Ph.D. in immunology in 1999, both from express the macrophage colonystimulating the University of Paris, France. After a postdoctoral fellowship with D. Littman in New York, USA, he factor 1 receptor (CSF1R; also known as started his own lab in Paris (the Unit 838 of Inserm) to study the in vivo development and functions CD115) and all the known endocytic recep of cells of the ‘mononuclear phagocyte system’ and their roles in the pathogenesis of inflammatory 26–28 disorders. He moved his lab to London, UK, in 2008 and is currently the Arc Chair of inflammation tors ; they have even been said to ‘resemble 26 Biology at King’s College, London, and the head of the newly established Centre for Molecular and macrophages’ . But they have been termed Cellular Biology of Inflammation. DCs solely because they express CD11c. And that really does confuse the field. Siamon Gordon is an Emeritus Professor of Cellular Pathology at the University of Oxford, UK. He obtained his medical degree at the University of Cape Town, South Africa, in 1961 and a Ph.D. at Allan m. mowat. In my own field of Rockefeller University, New York, USA, in 1971, under the tutorage of Z. A. Cohn. Since 1976, he has mucosal immunology, myeloid cell hetero been at the Sir William Dunn School of Pathology at the University of Oxford, retiring from the geneity has led to considerable confusion, Glaxo–Wellcome Professorship of Cellular Pathology in 2008. He spent a sabbatical at the US National and many accepted truths probably need Institute of Allergy and Infectious Diseases (NIAID) and the US National Cancer Institute (NCI). to be reinterpreted. In the gut, there are David A. Hume (F.R.S.E.) is Director of The Roslin Institute and Research Director of the Royal (Dick) extraordinary numbers of cells of myeloid School of Veterinary Studies at the University of Edinburgh, UK. He is an international authority on origin present in the absence of inflamma macrophage differentiation and the function of macrophages in infection, inflammatory disease tion. There is a consensus that both macro and cancer. He is especially interested in transcriptional regulation, transcriptional networks and phages and DCs in the gut are generally the evolution of macrophage biology across vertebrate species. He did his Ph.D. at the Australian hyporesponsive29,30, and this overall conclu National University, Canberra, Australia, and his postdoctoral work with S. Gordon at the University sion holds irrespective of what one calls of Oxford in the early 1980s. Before coming to Edinburgh, he was at the Institute for Molecular the cells, but recent years have introduced Bioscience at the University of Queensland, Australia, where he was Director of the Australian increasing complexity, particularly in terms Research Council Special Research Centre for Functional and Applied Genomics. of DC biology31. More and more individual DC subsets Allan M. Mowat is Professor of Mucosal Immunology at the University of Glasgow, UK, where his laboratory is engaged in studies of macrophages and dendritic cells in the mouse intestinal mucosa. have been defined based, for example, on He is also involved in a large amount of undergraduate teaching and contributes to textbooks in their ability to induce the differentiation of general and mucosal immunology. T helper 17 (TH17) cells or forkhead box p3 + (FoXP3) regulatory T (TReg) cells, promote Gwendalyn J. Randolph completed her Ph.D. at the State University of New York in Stony Brook, IgA classswitching in B cells or produce New York, USA, in pathology and immunology. She then carried out postdoctoral research at the proinflammatory factors, such as tumour Rockefeller University and Weill Medical College of Cornell University, New York. She is now necrosis factor (TNF), interleukin23 (Il23) Professor in the Department of Gene and Cell Medicine and a member of the Immunology Institute and nitric oxide (No)27,29,32–34. Almost with at Mount Sinai School of Medicine, New York. Her laboratory focuses on the trafficking of out exception, it is claimed that specific DC monocytes, dendritic cells, and macrophages, including in the context of inflammatory diseases or macrophage populations may be respon such as atherosclerosis. sible for these diverse properties, and it is implied that these could be manipulated spe homeostatic conditions is also relevant in receptor) mice to ‘selectively’ ablate them19,21. cifically for practical purposes, such as vac inflammation. None of these lines should be Subsequently, cells that were morphologi cine development and disease therapy. Many blurred until we know more about these won cally and functionally active phagocytes of these socalled DC subsets are not clearly derful populations in different anatomical were found to ‘convert’ into DCs22. This led defined and it is essential to understand sites and tissue states. to the concept of the immature DC and then that different groups use different methods of inflammatory or tumour necrosis factor to obtain and characterize DCs. often the David A. Hume. The mononuclear phago and inducible nitric oxide synthase-producing starting populations are preselected on the cyte system is a family of cells derived from DC (TIP DC)23. As noted also by G.j.R, basis of arbitrarily defined levels of expres a committed progenitor in the bone mar there is no obvious distinction between sion of markers that have been assumed to row18. The number of subpopulations of these cells and resident or classically be specific for either DCs or macrophages, mono nuclear phagocytes one can define is an activated macrophages, respectively. but are actually expressed by both35. exponential function of the number of mark If we consider the DC, as originally 19 ers examined ; the heterogeneity is essen defined by Ralph Steinman and Zanvil Cohn, it is essential to understand tially infinite because each gene has its own it was defined functionally by its ability to intrinsic probability of being expressed20. stimulate in an allogeneic mixed leukocyte that different groups use It is clear that the definition of DCs has reaction (MlR)24, and the distinction from different methods to obtain changed with time. The original description a macrophage is clear. These DCs do not and characterize DCs of DCs described them as not actively endo express the widely studied macrophage cytic and not part of the mono nuclear phago marker F4/80 (also known as EMR1), CD11b cyte system18,19. CD11c, the favoured DC (also known as αM integrin) or Fc recep Similar reservations cast doubt on the marker, was already known to be expressed tors, they are not phagocytic and they do not widely accepted view that a specific popu by most tissue macrophages before the use adhere to fibronectin; we used these criteria lation of DCs can take up antigens from of CD11creporter transgenes as markers to isolate this form of DC from the intes the intestinal lumen by sending processes of DCs, and of CD11c–DTR (diphtheria toxin tinal lamina propria25. As for the rest, most through the epithelium36,37. Supported by

elegant imaging data, these findings pro Perhaps the best functional attribute I In the long term, functional properties voked great interest throughout the field can think of to distinguish DCs and macro such as antigen processing and presentation and in the wider immunological commu phages is the property of DCs to localize in or migration are too fluid to serve as strict nity. Textbooks describe the phenomenon, the T cell zone of lymphoid organs where barriers for giving a cell its name. ongoing and many experimental studies on vac they can stimulate T cells. This attribute and future developments will make it possible cines, inflammation or infection rely on it combines several features that DCs seem to to define DCs and macrophages by lineage. as a starting paradigm. However, these DCs specialize in — potent, though not exclu Classification according to lineage would have been characterized entirely by their sive, ability to activate T cells and the ability somewhat disrupt the current functional and expression of CX3Cchemokine receptor 1 to migrate into the T cell zone itself. Indeed, markerbased classification. For example,

(CX3CR1) and MHC class II, rather than if I can paraphrase a comment I heard from we know that monocytederived cells can on any functional criteria. In fact, these Marc jenkins (personal communication): develop properties of DCs, including potent cells seemingly cannot migrate to draining one of the landmark contributions brought antigen presentation and migratory abilities, lymph nodes, or present antigen to specific about by Ralph Steinman’s discovery of and we currently call some monocytederived T cells35. DCs was how it highlighted the specialized cells DCs. Nonetheless, I think that the best I think that these ‘sampling’ cells may be nature of the T cell zone. As for migration, solution to integrating DC and macrophage tissueresident macrophages, contributing my favourite research subject, I am unceas biology in the future is to define DCs as those little or nothing to shaping primary T cell ingly impressed by experiments that com cells that arise from dedicated DC precursors responses and being responsible more for pare DCs and macrophages in their ability that we know exist17, but which need more scavenging functions in the mucosa itself. to migrate — DCs are far superior. Indeed, characterization and study, and to define hi hi These are distinct properties, and it is now evidence that CX3CR1 CD11b MHC macrophages as those cells that descend from essential that all previous work in the field class II+CD11c+ cells that sample the intesti either embryonic macrophages or blood is reexamined taking into account the nal lumen (socalled lamina propria DCs)36,41 monocytes. Communicating science is obvi most recent advances in phenotyping and do not obviously migrate to mesenteric ously easier when cells have a fitting name, functional characterization. Similar over lymph nodes efficiently26,39 is one reason but we shouldn’t be afraid to let the names laps are emerging from studies of myeloid why these cells may be better classified as evolve to fit contemporary advances in our subsets previously thought to be distinct macrophages instead of their present knowledge and to ultimately make further populations in lung, skin and other organs. classification as DCs. classification easier as the field advances.

Are there any good phenotypical Glossary markers for DCs and macrophages, or should the distinction be based solely on CD11c–DTR (diphtheria toxin receptor) mice Microglial cell Mice genetically engineered to express the diphtheria A phagocytic cell of myeloid origin that is involved in the function? If so, what does a DC do that a toxin receptor under the control of the CD11c promoter. innate immune response in the central nervous system. macrophage can’t and vice versa? Following administration of diphtheria toxin, cells Microglial cells are thought to be the brain-resident expressing CD11c are transiently depleted in these macrophages. G.J.R. Generally, cells with high surface animals. These mice have typically been used to study DC + levels of F4/80 are macrophages, but lower function, but certain CD11c macrophage populations are Mixed leukocyte reaction also depleted by diphtheria toxin treatment. (MLR). A tissue-culture technique for testing T cell levels of F4/80 are commonly observed on reactivity and APC activity. A population of T cells is DCs. Furthermore, much has been made Cross-presentation cultured with MHC-mismatched APCs, and proliferation of of CD11c as a marker for mouse DCs but, The mechanism by which certain APCs take up, process the T cells is determined by measuring the incorporation as mentioned above, the accuracy of this and present extracellular antigens on MHC class I of 3H-thymidine into the DNA of dividing cells. molecules to stimulate CD8+ T cells. This property is marker as a means to track DCs depends atypical, as most cells exclusively present peptides derived Myeloid-derived suppressor cells on the anatomical site in question. from endogenous proteins on MHC class I molecules. (MDSCs). A group of immature CD11b+GR1+ cells (which Myeloid cells with high levels of CD11c, include precursors of macrophages, granulocytes, DCs and though not those with low to inter mediate Kupffer cell myeloid cells) that are produced in response to various CD11c, are DCs in lymphoid organs. A specialized macrophage that lines the sinusoidal vessels of tumour-derived cytokines. These cells have been shown to the liver. These cells regulate local immune responses, and induce tumour-associated antigen-specific CD8+ T cell However, in the lung, high levels of CD11c remove microbial particles, endotoxin and other noxious tolerance. 38,39 are found on macrophages , and there substances that penetrate the portal venous system. are plenty of other anatomical locations Osteoclast outside of lymphoid organs where macro Langerhans cell Multinucleated giant cells, of myeloid origin, that are + A type of DC that is resident in the epidermal layer of the skin. responsible for bone resorption. Osteoclasts degrade bone phages are CD11c . Moreover, CD11b matrix and solubilize calcium from bone. was once thought to be a universal, albeit M1 macrophages not specific, marker for macrophages. A macrophage subtype that produces pro-inflammatory Plasmacytoid DCs However, it turns out that splenic macro cytokines and has cytotoxic functions. An immature DC with a morphology that resembles that of phages are CD11b low or negative40, as are a plasma cell. Plasmacytoid DCs produce type I IFNs in M2 macrophages response to viral infection. 38,39 pulmonary macrophages . There are a A macrophage subtype that acts to dampen inflammatory handful of Ctype lectins that seem to be responses and scavenge debris, as well as to promote Tumour necrosis factor and inducible nitric oxide specifically expressed by DCs, but there angiogenesis and tissue remodelling and repair. synthase-producing DC is currently not enough knowledge to (TIP DC). Monocyte-derived DCs that produce high Marginal zone metallophilic macrophage quantities of tumour necrosis factor and nitric oxide. make definitive claims about any marker A type of macrophage that surrounds the splenic white These cells develop in mice from GR1+ monocytes during combination, especially in nonlymphoid pulp, adjacent to the marginal sinus, and is involved in infection with certain bacteria, such as Listeria compartments. trapping particulate antigens. monocytogenes, or following myocardial damage.

A.m.m. The big problem we all encounter characterized by their location in defined Although flow cytometry has been of is that the conventional markers used to areas of lymphoid organs and are uniquely great help in defining lymphocyte hetero identify macrophages and DCs in mice, such specialized in processing exogenous anti geneity, embedded tissue macrophages are as F4/80, CD11c, CD11b and MHC class II, gen for presentation to T cells and T cell fragile and difficult to isolate from solid have turned out not to be specific. This priming. Plasmacytoid DCs are poor antigen organs, even after enzymatic tissue diges seems to be even more of an issue in tissues presenting cells, but produce and secrete tion. In situ immunochemistry is an art and such as the respiratory tract and intestine, large amounts of type 1 IFN following a qualitative tool, sensitive to fixation condi where the immunological environment is Tolllike receptor 9 (TlR9)signalling. tions, even with amplification and retrieval different to lymphoid tissue and there is The question implies a dichotomy of DCs methods, especially when monoclonal continual antigen exposure. and macrophages that may not be relevant antibodies are used. Functional analysis is Recent work shows that CD103 (also to biology. Both DCs and macrophages a sine qua non for defining DCs, but it does known as αE integrin) and CX3CR1 expres consist of distinct subsets, and have many not lend itself to singlecell studies. sion define distinct populations of myeloid functions. The paradigmatic DCs of text Not enough attention has been paid to cells in the intestinal mucosa, although books live in the periphery, uptake antigen, the expression of antimicrobial effector mol the exact nature of some of the subsets migrate from the periphery to the lymph ecules, such as lysozyme45, which is abun identified is contentious26,35,42. There is a node and initiate the immune response. dant in neutrophils and also readily secreted population of CD11c+MHC class II+ cells This cell resembles the langerhans cell, by monocytes and macrophages, but is only expressing CD103, which have become and the ability to migrate and carry antigen poorly expressed, if at all, by DCs. It is hoped the prototypic ‘mucosal’ DCs, inducing has been proposed as a criterion to define that microarray and proteomic analysis will gut homing molecules and FoXP3 expres DCs. Macrophages are best defined by yield new markers to detect proteins that sion in naive T cells by producing retinoic their phagocytic activity and are generally distinguish macrophages from DCs, both in 29 acid . The cells expressing CX3CR1 are less thought to be tissueresident cells. mice and in humans. However, no single set well defined. on the basis of the expres However, recent evidence indicates of markers can be expected to apply to all sion of MHC class II and CD11c, some that macrophages can migrate, and that stages of cell differentiation and activation. groups have separated this population into langerhans cells are not important for T cell In characterizing macrophages and DCs, conventional macrophages and a subset of priming44. Also, although some macrophages I like to use the analogy of a surviving text DCs with functions and origins distinct (for example, Kupffer cells) are actively fragment attributed to Archilocus, a Greek from the CD103+ DCs26,42,43. However phagocytic, other macrophages (for example, philosopher, and made famous by Isaiah closer examination shows that many, if not microglial cells) show only weak phagocytic Berlin46 among others: “The fox knows many + all, CX3CR1 cells in gut mucosa express ability. So we do not really have any good things, the hedgehog knows one big thing”. MHC class II and some CD11c, making functional criteria for defining macrophages To my mind, the macrophage is the fox, able these rather inappropriate distinguishing and DCs as they represent not just two to carry out a several functions, whereas + markers. In most respects, the CX3CR1 distinct populations, but various cell types. the DC is specialized for one big function, population actually seems to be generally namely to activate naive T lymphocytes10. homo geneous and is indistinguishable from S.G. There is no single morphologic or pro tissue macrophages35. tein marker of macrophages or DCs which is D.A.H. The DC doctrine is that DCs can Most importantly, recent work shows unambiguous. CD11c expression, for exam uniquely present antigen to naive T cells, so + that CX3CR1 cells in the mucosa do not ple, is also abundant on alveolar macro a cell that can do this is, by definition, a DC. migrate from the mucosal tissue to the phages, yet these cells have poor expression I feel this viewpoint is untenable. Myeloid draining mesenteric lymph nodes or present of F4/80, a widely used mouse macrophage antigenpresenting cells (APCs) have various antigen to naive T cells — only the CD103+ antigen marker. Additionally, the F4/80 anti APC activities that can stimulate a range of subset of DC can do this35. I think DCs and gen is readily detected on immature DCs, effector T cell outputs. The F4/80+CD11b+ macrophages should be distinguished func although it is downregulated on DCs after macrophages that we isolate from the intesti tionally in this way, as I am not persuaded maturation. CD11b can be present or absent nal lamina propria do not activate T cells in that macrophages can migrate and activate on tissue macrophages in situ. MlRs; they actually suppress the response. naive T cells to any great extent in normal But it has been argued that intestinal conditions. Rather, their functions are more F4/80+CD11b+ myeloid APCs can stimulate Not enough attention has + 27 related to proteolytic and catabolic activities, the development of FoXP3 TReg cells . such as destruction of pathogens, scavenging been paid to the expression Another study showed that the cells we of cellular debris and tissue remodelling. of antimicrobial effector call macrophages (the authors called them F.G. Phenotypic markers or functions molecules CD11b+CD103– DCs) can present peptide alone are usually not sufficient. As for any and activate naive oTII transgenic CD4+ other cells, macrophages and DCs should T cells26. It is not clear that this assay has the be characterized by a combination of cell Commonly, high constitutive expression same APC requirements as the MlR. ular origin, anatomical location, function of MHC class II is used to define mature The problem really arises with in vitro and phenotype. For example, alveolar DCs in the mouse. Resident macrophages assays21. In my opinion, macrophages as macrophages express CD11c, but are in the mouse do not express MHC class II some define them are actually mixtures of best defined by their anatomical location antigens until they are activated by IFNγ. stimulatory and inhibitory cells from dif and their morphology. Anatomical loca However, MHC class II is constitutively ferent areas of the primary organs, whereas tion is also crucial to identify microglial present on human tissue macrophages, DCs are a minor population of cells exclu and Kupffer cells. Classical DCs are best even in the absence of overt infection. sively from the stimulatory end of the

myeloid APC continuum21. The unique antigen presentation and the cellular mecha optimal proliferation to differentiate into function is actually at the ‘macrophage’ end nisms that allow DCs to initiate immunity macrophages19. Although DCs were origi of this continuum; I think that these are the or promote tolerance has been made using nally thought to be CSF1independent, this cells that generate T cell tolerance. The these cells49. is clearly not the case54 and antibody against DC believers term macrophages that can Macrophages can also be obtained from CSF1R is now used for the purification of suppress T cell responses ‘tolerogenic’ DCs47! monocytes or bone marrow precursors fol DC progenitors15. So, the problem is one of nomenclature. lowing culture with CSF1. This experimental FlT3, the receptor for FlT3l, is a marker The term DC should be used exclusively for model has allowed investigators to identify for haematopoietic stem cells and com the APCs that reside in T cell areas of spleen cytokines and bacterial products that control mon myeloid progenitors, and it is retained and lymph node and the small numbers of their effector functions. M1 macrophages, the on classical DCs in the spleen. FlT3l can similar cells at mucosal surfaces. All the con prototypical activating stimuli of which are expand myeloid APC populations in vitro and tributors agree that these cells are not phago IFNγ and lipopolysaccharide (lPS), exhibit in vivo15 but, like CSF1, probably acts on the cytic, and probably not descended from cells potent microbicidal properties, whereas HPPCFC in combination with other factors. M2 macrophages that ever were phagocytic. we simply need support TH2type effector to accept that some macrophages, defined by functions and may play a role in resolving G.J.R. DCs and macrophages derived in origin, CSF1dependence in vivo, markers inflammation through endocytic clearance culture show functional properties generally — such as F4/80, CD11b and Fc receptors — and trophic factor synthesis50. consistent with their counterparts in vivo. and phagocytic activity, can in fact present However, the heterogeneity of the mono However, they are only approximations of antigen to naive T cells and the controversy nuclear phagocyte system is poorly recapitu cells that exist in vivo. They are useful for evaporates. lated by these in vitro models. In other words, cell biological studies, for migration stud in vitrogenerated macrophages and DCs, ies to some extent and in some cases they What culture conditions should a albeit useful to study the cell biology of phago hold valuable promise as agents of cellbased researcher use to generate macrophages cytosis and antigen presentation, for example, immune therapies. Cultured DCs fuelled the and DCs? How do the in vitro-generated and do not represent a model to study the special growth of the DC field in the mid1990s, so in vivo cells compare? ized functions of the diverse cell types that are they are historically important. present in vivo, or the regulation of their devel Now, the DC field has moved beyond A.m.m. like others, we use the simple and opment and functions by the tissue micro this phase and makes use of cultured DCs in conventional methods of growing bone mar environement. As an example, many tissue combination with in vivo models. Generally, row cells in either granulocyte–macrophage macrophages and DC subsets do not derive DCs cultured in GMCSF are thought to be colonystimulating factor (GMCSF) or CSF1 from blood monocytes, and some may not counterparts of inflammationderived DCs, (also known as MCSF) to obtain mouse DCs even derive from the bone marrow1,4. In vivo but this needs to be formally shown. and macrophages, respectively. These meth studies are required to analyse the functions I have less experience with cultured ods have the advantage of generating reason of the mononuclear phagocyte system. macrophages. Perhaps they are overused ably homogeneous and reliable numbers of in studies of M1 and M2 polarization states, cells which can be manipulated in vitro. D.A.H. Individual colonies produced in vitro and peritoneal macrophages isolated by peri However, the cells obtained are probably from bone marrow in GMCSF contained toneal lavage may too often be thought to not representative of their counterparts in tis classical DCs as well as granulocytes and model macrophages from any anatomic site. sues. Their early stage of differentiation and macrophages51. Bone marrow contains a However, overall, I think the field appro homogeneity are unlikely to be replicated common progenitor (the macrophage and priately uses cultured DCs and macrophages in vivo. More importantly, they have not been DC progenitor (MDP)52), and a socalled without overextending the interpretation of exposed to any of the environmental factors common DC progenitor (CDP)15. CDPs the data generated from them. known to regulate myeloid cell differentia are probably identical to highproliferative tion and function in tissues. For example, potential colonyforming cells (HPPCFC) S.G. The ability to generate large numbers DCs and macrophages in the intestine show that require more than one factor to pro of DCs from mouse bone marrow and from many unusual properties compared with duce colonies19. There are at least seven human blood monocytes with the aid of their equivalents in other tissues, blood or cytokines that one should consider in growthfactor cocktails has proved irresist bone marrow and these are known to be studying myeloid APC differentiation; CSF1, ible to investigators for obvious reasons; for conditioned by local factors29,30,48. However, GMCSF, granulocyte colonystimulating example, the direct isolation of these cells the mediators involved are still being identi factor (GCSF), Il3, FMSrelated tyrosine from blood and tissues is tedious and often fied and no in vitro conditions have yet been kinase 3 ligand (FlT3l), Il4 and IFNγ. results in poor yields and the introduction found that replicate the full characteristics of A dogma has emerged that cells grown in of artefacts. the populations in situ. The same issues apply GMCSF are DCs, and those grown in CSF1 less well appreciated is the rapid ‘accul to other tissues, and so real advances will are macrophages19, but cells grown in both turation’ of macrophages that occurs ex vivo. only be made using cells isolated from the of these conditions are phagocytes and have Kupffer cells, for example, do not express organs themselves. been studied as functionally distinct macro CD11b in situ, but rapidly acquire this com phages53. Both cell types are in vitro artefacts; plement receptor in cell culture. Even modi F.G. DCs obtained in culture from mono in vivo, progenitors do not encounter any of fying culture vessels with media and growth cytes or bone marrow precursor cells exist in these factors in isolation, and there are no factors cannot prevent this artefact of isola two functionally and phenotypically distinct obvious in vivo counterparts to the in vitro tion. Alveolar macrophages are round and states, immature and mature. Fundamental derived cells. HPPCFCs require a combina loosely adherent in vivo, but are profoundly progress in understanding the cell biology of tion of Il3, GMCSF, CSF1 and IFNγ for altered in morphology by adherence to

tissue culture plastic, as are all macrophages. between expression of MHC class II and the human macrophages and DCs will be dif The use of undefined, nonphysiological macrophage marker F4/80 in mononuclear ficult to predict from the in vitro study of supplements such as fetal bovine serum is phagocytes found in mucosal sites55. More human cells or from mouse studies alone. particularly egregious, and defined media progress would be made by focusing on the should be used wherever possible. functions on genes and gene products; the A.m.m. I’m old fashioned and reductionist In situ analysis is therefore mandatory, infinite subdivision of myeloid cells using enough to think that selective therapeutic but DCs and macrophages in the tissue can multicolour flow cytometry with arbitrary immunomodulation will only be possible if be difficult to access, quantify or test func gates is a futile and unproductive avenue for the relevant cells and molecules are identi tionally. Therefore an integrated approach, future research. fied accurately. In the case of macrophages with appropriate awareness of artefact, is and DCs, there is ample evidence that dis required. My own fantasy is to be able to rec tinct functions may correlate with different reate the full panoply of macrophage pheno it is not possible to define phenotypic markers such as CD11b, F4/80, macrophages and DCs as types observed in vivo, from embryonic stem CX3CR1, CD103 and CD8α. without cells and other haematopoietic stem cells, separate entities; they are doubt, this needs to be taken into account entirely in vitro. This will require knowing a continuum of progeny of a if their properties are to be exploited in their in situ phenotype in detail and elucida practice. For instance, if DCs are the only tion of the cellular and extracellular environ common precursor cells that can prime and shape naive T cell ment in different organs, such as the liver, functions, targeting them will be essential; gut, lung, brain and uterus. F.G. Because macrophages and DCs con for example, for developing novel vaccines, sist of several cell types with discrete and enhancing antitumour responses and How important is the issue of defining essential functions, defining the develop inhibiting autoimmune responses. macrophages and DCs for future ment and functions of these cells in their A number of therapies are already being research and potential immune therapies? natural in vivo context (particularly in developed that target DC surface molecules, humans) is essential for our understanding and it could be equally useful to apply simi D.A.H. It is my view, discussed in detail of innate and adaptive immune response, lar strategies to the phenotypic subsets of elsewhere18–21, that all cells of the mono and is therefore necessary to improve treat DC which seem to have distinct functions nuclear phagocyte system can interact with ments for inflammatory and infectious or which operate in different tissues. The T cells. It does not matter whether they are diseases, as well as for cancer. The success CD103+ population of intestinal DCs is an called macrophages or DCs, what matters is of modern vaccines and immune therapies excellent example of this, as it has a crucial the outcome of their interaction in a patho has already benefited from an improved role in shaping mucosal immune responses 29 logical or therapeutic context. Antigen knowledge of macrophage and DC biol by the generation of gut homing TReg cells . presentation is a regulated function. we ogy56. Increasing our understanding of the In the skin, an apparently similar subset can manipulate APC activity empirically functions of human macrophages and DCs can potently cross-present antigens to CD8+ using adjuvants such as microbial products in vivo is likely to be crucial for the design T cells59. These properties are being explored or various cytokines. There are no mark of new therapeutic approaches, but this task as targets for vaccination or induction of ers that predict in vivo APC function other is challenging. tolerance. than high expression of MHC class II and one of the key challenges is to translate Conversely, macrophages may be more costimulatory molecules and cytokine knowledge from mouse studies to the human important to target for their proinflamma expression. By purifying out the suppressive system. For example, we, and others, have tory and catabolic functions, without neces ‘macrophages’, we may succeed in getting a proposed that mouse GR1+ monocyte sarily interfering with the initiation phase population of DCs that is more effective in derived TIP DCs23 that arise following of immune responses. For all these reasons, a particular T cell activation assay in vitro, infection with Listeria monocytogenes or establishing celltype specific phenotypic or but classical DCs are a small subset of the during myocardial damage, resemble human functional markers will be essential. potential APCs. Most of the active APCs M1 macrophages, whereas the mouse cells for adoptive immunotherapy are in the derived from GR1– monocytes resemble S.G. Macrophages and DCs provide attrac macrophage fraction21. human M2 macrophages1,52,57. However, tive targets for therapy. Selective targeting of I think that it is not possible to define although TIP DCs and M1 macrophages recruited or resident DCs or macrophages, or macrophages and DCs as separate entities; share several functions and express com of effector cytokines, should not incapacitate they are a continuum of progeny of a com mon genes, one must keep in mind that M1 the entire innate defence system against path mon precursor. The suggestion that DCs macrophages are human cells obtained from ogens or disrupt the homeostatic functions are a separate lineage is clearly wrong20. The culture in vitro, but TIP DCs are mouse cells of macrophages. The ability to harness DC DCs defined by Steinman and Cohn are that differentiate in vivo. of note, splenic antigenpresentation through enhanced vac cells of the mononuclear phagocyte system GR1+ monocytes can differentiate into cination strategies could transform tumour that occupy a particular niche in lymphoid myeloid-derived suppressor cells (MDSCs) immunization and overcome lymphoid and tissues, no more or less unique than the spe that promote the development of tumour myeloid suppressor mechanisms. cialized cells of the mononuclear phagocyte induced regulatory T cells and T cell anergy The increased understanding of pattern system found in other organs (for example, in tumourbearing mice58. recognition receptor signalling will aid the Kupffer cells and microglia). At mucosal Therefore, it may be naive to postulate development of better adjuvants to boost surfaces, mononuclear phagocytes are that human M1 macrophages will always adaptive immune responses. Targeting adapted to sample, process and present anti behave like TIP DCs in vivo. Indeed, it is immunogens to particular DC subsets (for gen. There is an almost complete overlap likely that many of the in vivo functions of example, through the use of monoclonal

antibodies specific for molecules such Frédéric Geissmann is at King’s College London, 23. Serbina, N. V., Salazar-Mather, T. P., Biron, C. A., Kuziel, W. A. & Pamer, E. G. TNF/iNOS-producing 60 Guy’s Campus, London SE1 1UL, UK. as DEC205 (also known as ly75) ) has dendritic cells mediate innate immune defense against e-mail: [email protected] shown promise in mouse models, although bacterial infection. Immunity 19, 59–70 (2003). Siamon Gordon is at Sir William Dunn School of 24. Steinman, R. M. & Inaba, K. Stimulation of the much more work is required to confirm primary mixed leukocyte reaction. Crit. Rev. Immunol. Pathology, South Parks Road, Oxford OX1 3RE, UK. 5, 331–348 (1985). these studies in primates and humans. e-mail: [email protected] 25. Pavli, P., Woodhams, C. E., Doe, W. F. & Hume, D. A. longacting inducers of DC typeI IFN Isolation and characterization of antigen-presenting David A. Hume is at the Roslin Institute and Royal dendritic cells from the mouse intestinal lamina production, such as the TlR3 ligand (Dick) School of Veterinary Studies, University of propria. Immunology 70, 40–47 (1990). polyinosinic–polycytidylic–polyllysine Edinburgh, Roslin, Midlothian EH25 9PS, UK. 26. Bogunovic, M. et al. Origin of the lamina propria dendritic cell network. Immunity 31, 513–525 61 e-mail: [email protected] (poly I:C:lC), need further evaluation . (2009). To be practical and effective, I think that Allan M. Mowat is at the Glasgow Biomedical Research 27. Denning, T. L., Wang, Y. C., Patel, S. R., Williams, I. R. & Pulendran, B. Lamina propria macrophages and it is necessary to target specific subsets of Centre, University of Glasgow, 120 University Place, dendritic cells differentially induce regulatory and DCs or macrophages in vivo, rather than Glasgow G12 8TA, UK. interleukin 17-producing T cell responses. Nature e-mail: [email protected] Immunol. 8, 1086–1094 (2007). by ex vivo manipulation and subsequent 28. Pavli, P., Hume, D. A., Van De Pol, E. & Doe, W. F. adoptive transfer of cells. The ability to do Gwendalyn J. Randolph is at the Mount Sinai School of Dendritic cells, the major antigen-presenting cells of Medicine, New York, New York 10029, USA. the human colonic lamina propria. Immunology 78, so safely will depend on an improved knowl 132–141 (1993). e-mail: [email protected] edge of the molecular basis of mononuclear 29. Coombes, J. L. & Powrie, F. Dendritic cells in intestinal doi:10.1038/nri2784 immune regulation. Nature Rev. Immunol. 8, myeloid cell heterogeneity in vivo. 435–446 (2008). Published online 14 May 2010 30. Platt, A. M. & Mowat, A. M. Mucosal macrophages and the regulation of immune responses in the G.J.R. Both macrophages and DCs have cru 1. Geissmann, F. et al. Development of monocytes, intestine. Immunol. Lett. 119, 22–31 (2008). macrophages, and dendritic cells. Science 327, 31. Laffont, S. & Powrie, F. Immunology: dendritic-cell cial roles in events that influence immunity, 656–661 (2010). genealogy. Nature 462, 732–733 (2009). health of tissues, and recovery from disease 2. Ajami, B., Bennett, J. L., Krieger, C., Tetzlaff, W. & 32. Uematsu, S. et al. Regulation of humoral and cellular Rossi, F. M. Local self-renewal can sustain CNS gut immunity by lamina propria dendritic cells states. I don’t see how research on any cell microglia maintenance and function throughout adult expressing Toll-like receptor 5. Nature Immunol. type can occur optimally without further life. Nature Neurosci. 10, 1538–1543 (2007). 9, 769–776 (2008). 3. Merad, M. & Manz, M. G. Dendritic cell homeostasis. 33. Mizoguchi, A. et al. Dependence of intestinal defining the cell of interest, and all cells have Blood 113, 3418–3427 (2009). granuloma formation on unique myeloid DC-like cells. to be called something. 4. Chorro, L. et al. LC proliferation mediates neonatal J. Clin. Invest. 117, 605–615 (2007). development, homeostasis, and inflammation- 34. Cerutti, A. & Rescigno, M. The biology of intestinal associated expansion of the epidermal LC network. immunoglobulin A responses. Immunity 28, 740–750 J. Exp. Med. 206, 3089–3100 (2009). (2008). 5. Gordon, S. in Fundamental Immunology (ed Paul, W.) 35. Schulz, O. et al. Intestinal CD103+, but not CX CR1+, DC and macrophage 3 (Lippincott Williams and Wilkins, Philadelphia, 2008). antigen sampling cells migrate in lymph and serve biologists don’t talk with each 6. Pluddemann, A. & Gordon, S. in Phagocyte–pathogen classical dendritic cell functions. J. Exp. Med. 206, interactions (eds Russell, D. G. & Gordon, S.) (American 3101–3114 (2009). Society for Microbiology Press, Washington DC, 2009). 36. Niess, J. H. et al. CX CR1-mediated dendritic cell other enough. Indeed, if a 3 7. Gordon, S. & Taylor, P. R. Monocyte and macrophage access to the intestinal lumen and bacterial clearance. cell is called a DC in a paper, I heterogeneity. Nature Rev. Immunol. 5, 953–964 Science 307, 254–258 (2005). (2005). 37. Rescigno, M., Lopatin, U. & Chieppa, M. Interactions think that it is less likely that a 8. Mebius, R. E. & Kraal, G. Structure and function of the among dendritic cells, macrophages, and epithelial spleen. Nature Rev. Immunol. 5, 606–616 (2005). cells in the gut: implications for immune tolerance. macrophage biologist will read it 9. Swirski, F. K. et al. Identification of splenic reservoir Curr. Opin. Immunol. 20, 669–675 (2008). monocytes and their deployment to inflammatory 38. Gonzalez-Juarrero, M., Shim, T. S., Kipnis, A., and vice versa sites. Science 325, 612–616 (2009). Junqueira-Kipnis, A. P. & Orme, I. M. Dynamics of 10. Steinman, R. M. Dendritic cells: versatile controllers of macrophage cell populations during murine pulmonary the immune system. Nature Med. 13, 1–5 (2007). tuberculosis. J. Immunol. 171, 3128–3135 (2003). 11. Naik, S. H. et al. Intrasplenic steady-state dendritic 39. Jakubzick, C. et al. Blood monocyte subsets Right now, a limitation to progress in cell precursors that are distinct from monocytes. differentially give rise to CD103+ and CD103– Nature Immunol. 7, 663–671 (2006). pulmonary dendritic cell populations. J. Immunol. immunology — and by extension, immuno 12. Fogg, D. K. et al. A clonogenic bone marrow 180, 3019–3027 (2008). logical therapies — is that DC and macrophage progenitor specific for macrophages and dendritic 40. Idoyaga, J., Suda, N., Suda, K., Park, C. G. & cells. Science 311, 83–87 (2006). Steinman, R. M. Antibody to Langerin/CD207 localizes biologists don’t talk with each other enough. 13. Liu, K. et al. Origin of dendritic cells in peripheral large numbers of CD8α+ dendritic cells to the marginal Indeed, if a cell is called a DC in a paper, lymphoid organs of mice. Nature Immunol. 8, zone of mouse spleen. Proc. Natl Acad. Sci. USA 106, 578–583 (2007). 1524–1529 (2009). I think that it is less likely that a macro phage 14. Varol, C. et al. Monocytes give rise to mucosal, but not 41. Chieppa, M., Rescigno, M., Huang, A. Y. & biologist will read it and vice versa. If a neu splenic, conventional dendritic cells. J. Exp. Med. 204, Germain, R. N. Dynamic imaging of dendritic cell + 171–180 (2007). extension into the small bowel lumen in response to tral term is used in a paper, such as ‘CD11c 15. Onai, N. et al. Identification of clonogenic common epithelial cell TLR engagement. J. Exp. Med. 203, cell’62 or ‘antigenpresenting cell’, this may Flt3+M-CSFR+ plasmacytoid and conventional 2841–2852 (2006). dendritic cell progenitors in mouse bone marrow. 42. Varol, C. et al. Intestinal lamina propria dendritic cell also limit or alter its impact by limiting the Nature Immunol. 8, 1207–1216 (2007). subsets have different origin and functions. Immunity ability of the right scientists to find it. 16. Jakubzick, C. et al. Lymph-migrating, tissue-derived 31, 502–512 (2009). dendritic cells are minor constituents within steady- 43. Niess, J. H. & Adler, G. Enteric flora expands gut + Matt Collin, whose work focuses on state lymph nodes. J. Exp. Med. 205, 2839–2850 lamina propria CX3CR1 dendritic cells supporting human DCs and macrophages63, has sug (2008). inflammatory immune responses under normal and 17. Liu, K. et al. In vivo analysis of dendritic cell development inflammatory conditions. J. Immunol. 184, gested to me that M1 or classically activated and homeostasis. Science 324, 392–397 (2009). 2026–2037 (2010). 18. Hume, D. A. The mononuclear phagocyte system. 44. Kissenpfennig, A. et al. Dynamics and function of macrophages are probably the same cells as Curr. Opin. Immunol. 18, 49–53 (2006). Langerhans cells in vivo: dermal dendritic cells TIP DCs (M. Collin, personal communica 19. Hume, D. A. Differentiation and heterogeneity in the colonize lymph node areas distinct from slower mononuclear phagocyte system. Mucosal Immunol. migrating Langerhans cells. Immunity 22, 643–654 tion). Indeed, I cannot find any instance 1, 432–441 (2008). (2005). where the same study shows both cell types 20. Hume, D. A. Probability in transcriptional regulation 45. Keshav, S., Chung, P., Milon, G. & Gordon, S. and its implications for leukocyte differentiation and Lysozyme is an inducible marker of macrophage in the same place. This is just one example inducible gene expression. Blood 96, 2323–2328 activation in murine tissues as demonstrated by in situ of how further work and open dialogue in (2000). hybridization J. Exp. Med. 174, 1049–1058 (1991). 21. Hume, D. A. Macrophages as APC and the dendritic 46. Berlin, I. The hedgehog and the fox: an essay on defining macrophages and DCs could lead cell myth. J. Immunol. 181, 5829–5835 (2008). Tolstoy’s view of history (Weidenfeld and Nicolson, to a clearer and more comprehensive body of 22. Randolph, G. J., Inaba, K., Robbiani, D. F., Steinman, London, 1953). R. M. & Muller, W. A. Differentiation of phagocytic 47. Steinman, R. M., Hawiger, D. & Nussenzweig, M. C. literature for all scientists. That, in turn, can monocytes into lymph node dendritic cells in vivo. Tolerogenic dendritic cells. Annu. Rev. Immunol. 21, only help to advance the field. Immunity 11, 753–761 (1999). 685–711 (2003).

48. Iliev, I. D., Mileti, E., Matteoli, G., Chieppa, M. & of the presence of Langerhans cells. J. Exp. Med. 207, Rescigno, M. Intestinal epithelial cells promote colitis- 189–206 (2010). protective regulatory T-cell differentiation through 60. Bonifaz. et al. In vivo targeting of antigen to maturing dendritic cell conditioning. Mucosal Immunol. 2, dendritic cells via the DEC-205 receptor improves 340–350 (2009). T cell vaccination. J. Exp. Med. 199, 815–824 49. Mellman, I. & Steinman, R. M. Dendritic cells: (2004). specialized and regulated antigen processing 61. Longhi, M. P. et al. Dendritic cells require a systemic machines. Cell 106, 255–258 (2001). type I interferon response to mature and induce 50. Martinez, F. O., Sica, A., Mantovani, A. & Locati, M. CD4+Th1 immunity with poly IC as adjuvant. J. Exp. Macrophage activation and polarization. Front. Biosci. Med. 206, 1589–1602 (2009). 13, 453–461 (2008). 62. Patsouris, D. et al. Ablation of CD11c-positive cells 51. Inaba, K. et al. Granulocytes, macrophages, and normalizes insulin sensitivity in obese insulin dendritic cells arise from a common major resistant animals. Cell Metab. 8, 301–309 (2008). histocompatibility complex class II-negative 63. Haniffa, M. et al. Differential rates of replacement of progenitor in mouse bone marrow. Proc. Natl Acad. human dermal dendritic cells and macrophages during Sci. USA 90, 3038–3042 (1993). hematopoietic stem cell transplantation. J. Exp. Med. 52. Auffray, C., Sieweke, M. H. & Geissmann, F. 206, 371–385 (2009). Blood monocytes: development, heterogeneity, and 64. Liu, K. et al. In vivo analysis of dendritic cell relationship with dendritic cells. Annu. Rev. Immunol. development and homeostasis. Science 324, 27, 669–692 (2009). 392–397 (2009). 53. Fleetwood, A. J., Lawrence, T., Hamilton, J. A. & 65. Naik, S. H. et al. Development of plasmacytoid and Cook, A. D. Granulocyte-macrophage colony- conventional dendritic cell subtypes from single stimulating factor (CSF) and macrophage CSF- precursor cells derived in vitro and in vivo. dependent macrophage phenotypes display Nature Immunol. 8, 1217–1226 (2007). differences in cytokine profiles and transcription factor activities: implications for CSF blockade in Competing interests statement J. Immunol. 178 inflammation. , 5245–5252 (2007). The authors declare no competing financial interests. 54. MacDonald, K. P. et al. The colony-stimulating factor 1 receptor is expressed on dendritic cells during differentiation and regulates their expansion. DAtABASeS J. Immunol. 175, 1399–1405 (2005). UniProtKB: http://www.uniprot.org 55. Hume, D. A. Immunohistochemical analysis of murine CD8 | CD11b | CD11c | CD103 | CSF1 | CSF1R | CX CR1 | mononuclear phagocytes that express class II major α 3 DeC205 | F4/80 | FLt3 | FLt3L | G-CSF | GM-CSF | iFN | iL-3 | iL-4 histocompatibility antigens. Immunobiology 170, γ 381–389 (1985). FURtHeR inFoRMAtion 56. Steinman, R. M. & Banchereau, J. Taking dendritic Frédéric Geissmann’s homepage: http://kcl.ac.uk/schools/ Nature 449 cells into medicine. , 419–426 (2007). siamon Gordon’s homepage: http://users.path.ox.ac. 57. Nahrendorf, M. et al. The healing myocardium uk/~cholt/ sequentially mobilizes two monocyte subsets with David A. Hume’s homepage: http://www.roslin.ed.ac.uk/ divergent and complementary functions. J. Exp. Med. Allan M. Mowat’s homepage : http://www.gla.ac.uk/ 204, 3037–3047 (2007). departments/immunology/researchgroups/allanmowat/ 58. Movahedi, K., et al. Identiﬁcation of discrete tumor- Gwendalyn J. randolph’s homepage: induced myeloid-derived suppressor cell http://www.mountsinai.org/profiles/gwendalyn-j- subpopulations with distinct T cell-suppressive activity. Blood 111, 4233–4244 (2008). randolph?id=0000072500001497156102 59. Henri, S. et al. CD207+ CD103+ dermal dendritic cells All lInkS ARe ActIve In tHe onlIne pDF cross-present keratinocyte-derived antigens irrespective