1. Top of page
  2. Introduction
  3. Older criteria used to identify the DC lineage
  4. Newer criteria used to identify the DC lineage
  5. References

Research on the molecular markers and functions of dendritic cell (DC) subsets is exceptionally active (1–3), explaining why it is difficult to fully cover the field even with the 21 superb reviews that follow in this volume. First, let us consider some of the nomenclature in the DC subset field (Table 1).

Table 1.   Some uses of the term ‘dendritic cell subsets’
Dendritic cells (monocyte-independent) and monocyte-derived dendritic cells
Langerhans cells in epidermis and other stratified squamous epithelia (vagina, anus, cervix, oral pharynx, upper esophagus)
Conventional or classical dendritic cells, and plasmacytoid dendritic cells
Resident and migratory dendritic cells, or lymphoid tissue (spleen, lymph nodes, mucosal associated tissue, thymus) and non-lymphoid tissue (lung, liver, kidney, heart, intestine, dermis of skin, eye) dendritic cells

One type of nomenclature is development oriented. Most DCs in mouse lymphoid organs in the steady state are monocyte-independent, requiring flt-3L (fms-like transcript 3 ligand) rather than M-CSF (macrophage colony-stimulating factor) for their development (4–6). In other tissues, e.g. the intestine, there are additional M-CSF dependent monocytes in the steady state (7, 8) as well as inflammatory monocytes and DCs during certain infections (9). The identification of authentic monocyte-derived DCs in vivo is an active area of research. In addition to hematopoietins like flt-3L, M-CSF, and GM-CSF (granulocyte-macrophage CSF), other pathways bear upon DC development such as the Notch/Wnt system (10).

Langerhans cells (LCs) are DCs within the epidermis and other stratified squamous epithelia. They exhibit the classical features of DCs when isolated from the skin and express a lectin, Langerin/CD207, but they differ from lymphoid tissue DCs in being flt-3L independent and M-CSF dependent (7, 11). The in vivo roles of LCs have become more mysterious as a result of the development of new genetically modified mice that allow Langerin+ cells to be depleted. In effect, many functions, previously ascribed exclusively to epidermal LCs in skin, may also involve dermal Langerin+ DCs (7, 11).

A second type of nomenclature distinguishes classical or conventional DCs from plasmacytoid DCs on the basis of distinct morphology, markers and gene expression profiles (12–15). Although popular, the adjectives ‘classical’ and ‘conventional’ are a little odd. They are not used in any other lineage when new forms are uncovered, e.g. immunologists do not speak of classical or conventional B cells to distinguish them from B1 and B2 pathways, and classical or conventional T cells to distinguish them from NKT and γδ T. The term ‘myeloid’ DC is often used as a counterpart to plasmacytoid DCs, although the latter are formally myeloid in origin.

Plasmacytoid DCs were first identified in humans; their name indicates that they share cytologic features with plasma cells. These DCs are the long mysterious, type I interferon-producing cells (12, 13). While all DCs can produce large amounts of type I interferons, plasmacytoid DCs are able to do so upon exposure to both live and inactivated viruses, as well as self nucleic acids, e.g. the DNA from inactivated herpes viruses and self-DNA, when the latter is complexed to cationic self-peptides to enhance DNA uptake. This is because plasmacytoid DCs express Toll-like receptor 7 (TLR7) and TLR9 in an endosomal location, to recognize and respond to nucleic acids.

Human DC subsets are now being distinguished with markers that are not expressed in mice, e.g. the BDCA-2 marker for plasmacytoid DCs is not found in mice, and likewise for the BDCA-1 (CD1c), BDCA-3, CD1a, CD1b markers for human DC subsets (14, 15). Although many of the markers in use are not shared between human blood and mouse lymphoid tissue DCs, there are some shared markers like DEC-205/CD205 and Langerin/CD207.

Human DC subsets are relevant to disease. In histiocytosis X, LCs expand to form granulomas or rarely a more diffuse proliferative order (16). Different DC subsets are being explored to vaccinate patients, so that distinct forms of antibody and T-cell-mediated immunity can be induced (15).

Another nomenclature distinction involves resident versus migratory DCs. ‘Resident DCs’ move directly to lymphoid tissues from a blood precursor, while ‘migratory DCs’ are also bone marrow-derived, but first enter tissues prior to migration via lymphatics to lymph nodes (4, 7, 8, 17–20). Nevertheless, mouse lymphoid and non-lymphoid organs can contain similar DC subsets, e.g. a CD103+ subset is found in both locations and is effective in cross presenting antigens on major histocompatibility complex class I (MHC I) products. The identification and properties of DCs at body surfaces continues to be a critical area of immunology and is valuably addressed by many of the articles that follow, e.g. DCs in the lung (20), eye (21), intestine (7, 8, 17, 18, 22), and skin (7, 11).

The existence of DC subsets broadens the scope of DC biology, and figuring them out is of major interest and importance. DC subsets can have distinct roles in controlling the type of immune response (23), and each subset expresses distinct pattern recognition receptors for recognition of microbes and other antigens (24). The fact that many of the markers used to define DC subsets are involved in innate immunity (antigen uptake and presentation, DC signaling) suggests that a major raison d’être for subsets is to provide distinct innate defenses, e.g. production of pathogen-relevant cytokines, chemokines, and defensins, followed by the initiation and control of adaptive immunity.

Classically, research on DCs has concentrated on their capacity to initiate immunity following encounter with different maturation stimuli. The field has expanded considerably with studies on DC receptors for antigen uptake and DC roles in immune silencing by different mechanisms. These subjects are highlighted in the chapters that follow.

Older criteria used to identify the DC lineage

  1. Top of page
  2. Introduction
  3. Older criteria used to identify the DC lineage
  4. Newer criteria used to identify the DC lineage
  5. References

All subsets of DCs fulfill longstanding criteria outlined in Table 2. However, for most, especially plasmacytoid DCs and monocyte-derived DCs, the criteria are best evident when the cells mature. ‘Maturation’ is the term used to denote DC differentiation in response to environmental stimuli, allowing DCs to link the type of immunity to the challenge at hand. Currently, the mature forms of all DC subsets show a dendritic morphology with probing processes, weak phagocytic activity, high MHC class II levels, a constellation of surface receptors some unique for antigen uptake and processing, and the capacity to initiate immunity, especially T-cell immunity.

Table 2.   Older criteria to identify white cell lineages including dendritic cells
Phagocytosis for clearance, scavenging, and microbial killing
Cell surface markers
Antigen processing and presentation
Adaptive immunity, e.g. immune initiating cells versus effector cells

The ‘probing’ morphology of DCs

DCs exhibit unusual probing movements, continually extending and retracting long processes from different points of the cell body. The distinct morphology of DCs was a key criterion used to discover DCs (25) and to develop methods to enrich these relatively infrequent cells (26–28), so that their molecular markers and functions could begin to be identified (29–34). The motility of DCs has now been studied in living lymphoid tissues, using two photon microscopy (35–39). Lindquist et al. (40) reported that steady-state DCs in intact lymph nodes did not translocate but instead formed and retracted processes, continually probing their environment, much as was observed initially in vitro (25).

A relationship of motility to another hallmark of DCs, abundant synthesis and expression of MHC II, has been found. The cytosolic domain of the MHC II-associated invariant chain (CD74) regulates the actin-based motor protein, myosin II (41). As Faure-Andréet al. write, ‘the use of common regulators for antigen processing and cell motility provides a way for DCs to coordinate these two functions in time and space’.

Phagocytosis and scavenging

Mature DCs only phagocytose particles to a limited extent, but some uptake takes place in immature DCs, e.g. DCs developing from bone marrow progenitors with GM-CSF, and in LCs freshly isolated from the epidermis (42, 43). Flow cytometry now makes it possible to quantify endocytosis of different cell types within complex cell mixtures. As illustrated in Fig. 1, when one injects fluorescent heat-killed microorganisms or soluble proteins intravenously to mice, and then prepares cell suspensions 30 min to 3 h later, only a low frequency of CD11chigh DCs cells take up particles, and then only in small amounts. By contrast, a sizeable fraction of F4/80high CD11clow macrophages show a large shift of fluorescence upon uptake of soluble ovalbumin (or bovine serum albumin, not shown), latex, and E. coli particles. Alternatively, one can inject particulates and look for phagocytosis by different cell types in tissue sections (Fig. 2). Macrophages but not DCs become heavily labeled with the injected tracer for endocytosis, even when the two cell types are juxtaposed in the marginal zone of spleen.


Figure 1.  Comparison of endocytic activity of splenic CD11chighDCs and F4/80highred pulp macrophages by flow cytometry. Balb/c mice were injected intravenously with PBS, soluble Alexa 488 labeled OVA, or fluorescent particulates (Yellow-Green Polystyrene, Alexa 488 labeled killed E.coli). Thirty minutes or 3 h later, spleens were harvested, digested with collagenase D (Roche), and the cell suspensions were depleted of B cells using CD19 MACS beads and LS MACS columns (Miltenyi Biotec, Auburn, CA, USA). The negatively selected cells were stained with different cocktails of antibodies to identify DCs (CD11c, CD11b, CD3, CD19, Ter119, DX5, and CD8) or red pulp macrophages (CD11c, CD11b, F4/80, CD3, CD19, Ter119, and DX5). As shown on the top, among single cells (gate 1), the live cells (gate 2), singlets (gate 3), and the CD19−, Ter119−, CD3− and DX5−cells (gate 4, 5, and 6) were selected for further analysis and examined for the expression of CD11c and CD11b. CD11blow CD11clow cells were further analyzed for F4/80 expression to identify red pulp macrophages (see Fig. 2A). The uptake of blood borne tracers was analyzed in DCs, both CD8+ and CD8, and macrophages.

Download figure to PowerPoint


Figure 2.  Comparison of endocytic activity of splenic DCs and macrophages in situ. Splenic sections of Balb/c mice were examined 30 min after intravenous inoculation of Yellow-Green Polystyrene. Cryosections were stained with anti-F4/80 (A), anti-MOMA-1 (B) or anti-CD8 (D) followed by Alexa 555 labeled anti-rat IgG (Red). CD11c staining was performed using anti-CD11c mAb (Red in C) followed by Biotin labeled anti-Hamster and Alexa 555-Streptavidin. Sections were further stained with Alexa 647 labeled anti-SIGNR1 (Blue, A and B) or anti-B220 (Blue, C and D). Most Yellow-Green Polystyrene accumulate in red pulp macrophages (F4/80+, Red in A), marginal zone macrophages (SIGNR1+, Blue in A and B) and marginal metallophilic macrophages (MOMA-1+, Red in B), but not in CD11c+ or CD8+ DCs (Red in C and D, respectively). (Scale bar 50 μm).

Download figure to PowerPoint

In contrast to the particles just mentioned, DCs do phagocytose dying cells (or their fragments) (44, 45). In mice, this uptake takes place primarily in the CD8+ DC subset (46) and is especially active when NK cells kill targets in their vicinity (47, 48). With respect to potential receptors involved in uptake of dying cells, CD8+ DCs selectively express αVβ5 integrin, CD36, DEC-205, and Langerin (49, 50), Tim-3 (51), Treml4 (52), and CLEC-9A (53–55). Among these, treml4 and CLEC9A bind to necrotic cells (annexin+ propidium iodide+), while Tim-3 recognizes apoptotic forms (annexin+propidium iodide). In mouse lymph nodes, DCs can take up other DCs that have migrated into the node (56, 57). Human monocyte-derived DCs in tissue culture also cross present dying cells and receptor-targeted proteins (58, 59). With respect to function, presentation of dying cells leads to cross presentation of antigen and immune tolerance in the steady state, or immunity upon maturation (45, 47, 48). A new area involves the uptake of small or nanoparticles, which seems well developed in DCs (60). An important distinction in studies of phagocytosis is the difference between scavenging or clearance functions, which involves large amounts of uptake that is readily visualized, and the small amounts of uptake that can suffice to present antigen, which can be tough to visualize but has been apparent for some time in DCs (61).

Markers, typically molecules involved in antigen uptake and presentation

The first DC-restricted monoclonal antibody was termed 33D1 (62) but only recently was the corresponding antigen identified, the DCIR2 lectin (63). DCIR2 marks a subset of DCs in the bridging zones of splenic white pulp nodules, which rapidly forms MHC II-peptide complexes when antibodies deliver antigen directly to the DCIR2 receptor in vivo (63). DCIR2 is a valuable marker for DCs in mouse spleen, but it has yet to be detected on many other forms of mouse DCs. In humans, a single DCIR homolog is expressed on many types of myeloid cells including DCs (64).

NLDC-145, the second DC-restricted monoclonal antibody, mainly stains DCs in the T-cell areas of different lymphoid tissues (65), but in other organs, the marker is abundant on epithelia, e.g. airway and intestinal epithelium (66). When cloned, the molecule recognized by NLDC-145 was found to be a lectin localizing to coated pits and endocytic compartments. It was renamed DEC-205/CD205 (67) and proved to be the first in a series of lectin-like uptake receptors on DCs in vivo (68). DEC-205/CD205 is not DC specific, because small amounts of DEC-205 are expressed by many types of leukocytes (66, 69), but the CD8+ DC subset expresses high levels. Likewise, when one injects anti-DEC-205 antibody into mice, the monoclonal antibody targets primarily to T-cell area DCs (70). In human lymph nodes, CD205 is abundant on MHC IIhigh, CD11chigh DCs in the T-cell areas but not macrophages in the lymph node medulla (71). At this point in the literature, DEC-205/CD205 remains the only lectin that has been visualized on large numbers of T-cell area DCs in mouse and human lymphoid organs.

A new but incisive way to identify lineages involves receptors for hematopoietins. Most DCs in mouse lymphoid tissues are CD135 or flt3high but CD115 or c-fms/M-CSFRlow (72). The use of these receptors to mark human DCs is just beginning.

Another DC marker is CD207 or Langerin, a lectin responsible for the formation of distinctive Birbeck granules in LCs (73). In mice, Langerin is also made by dermal DCs, which are a major source for Langerin+ DCs in skin-draining lymph nodes (74–77). Additionally in Balb/c but not C57Bl/6 mice, Langerin is detected on CD8+ DCs in all lymphoid organs, not just skin-draining nodes (78). Langerin is not found on monocytes, macrophages, granulocytes, and plasmacytoid DCs (79). Thus, Langerin is the most specific marker for CD8+ DCs including detection of CD8+ DCs in tissue sections. For example, staining with the valuable L31 anti-Langerin monoclonal antibody identifies a major reservoir of CD8+ DCs in the marginal zone of spleen, which in spite of a location amidst many phagocytic macrophages, take up very few latex and bacterial particles (79). Nevertheless, Langerin itself mediates antigen uptake and presentation, as shown with anti-Langerin antibodies engineered to deliver antigens (50).

Many markers of plasmacytoid DCs are involved in their hallmark feature, the production of abundant type I interferon (80, 81). BDCA-2/CD303 on human cells inhibits interferon production (82) while in mice, trem PDC enhances interferon production (83). A marker for mouse PDCs, PDCA-1, is identical to BST-2, which is induced in many cells after exposure to type I interferon (84). At this time, few antigen uptake receptors are known for PDCs except for siglec H in mouse PDCs (85).

Monocyte-derived DCs are found in lymphoid tissues under specific stimulating conditions, e.g. infection with Leishmania major (86), Listeria monocytogenes (87), and influenza virus (88). At the moment, the markers of these cells are not specific, i.e. high CD11b and CD11c expression, but clearly the cells derive from monocytes during infection. The cells also have yet to be tested for some critical functions, such as the stability of DC differentiation and their capacity to initiate immunity.

As discussed below, the functional roles of these DC marker molecules is under active study in vivo. Nonetheless, it is probably significant that DC subset markers are often innate receptors, especially lectins that can contribute to antigen uptake and innate signaling (89). These subsets are primarily defined by the ligands they capture and the agonists they respond to.

Antigen presentation and its regulation

High expression of MHC class II was the first molecular feature that became apparent when enriched populations of DCs were prepared (26, 27, 90, 91). High MHC II was also noted in DCs in sections of lymphoid (92, 93) and non-lymphoid (94, 95) tissues, setting the stage for what was to emerge, i.e. that DCs are major stimulators of helper T-cell responses in vivo.

CIITA is a transcriptional coactivator of MHC II genes and a small cohort of non-MHC molecules such as the invariant chain, RFX5, and Rab4 (96). The CIITA locus has four promoters. Promoters p1 and p3 are used by myeloid and lymphoid cells, respectively, but while most DCs use p1, plasmacytoid DCs uses p3 (97), like B cells. The function of p2 is not known, but p4 is used in many cell types to increase MHC II expression in response to IFN-γ. When DCs mature, CIITA transcription is silenced (98), although shortly after receiving the maturation signal, DCs increase MHC II biosynthesis for a few hours (99–101). Thus, MHC II gene transcription can be under distinct lineage-restricted controls.

Antigen processing and presentation are also regulated during DC maturation. Stimuli like lipopolysaccharide cause MHC II to redistribute from MHC II+ compartments to the cell surface (99, 100). When peptide–MHC II complexes are followed directly in DCs, they form within lysosomes during maturation and then move to the cell surface (102–104), thus setting the stage for recognition by the T-cell receptor for antigen. A key control for peptide–MHC II complex formation is at the level of acidity, as maturation stimulates the assembly of a proton pump to lower pH within the endocytic system (105). In general, however, DC lysosomes are relatively lacking in proteases, in contrast with the lysosomes of macrophages, which are rich (106). Yet another regulation entails MHC II turnover. DC maturation is associated with less ubiquitylation of the MHC II β chain and more stability of cell surface MHC II (107, 108).

Plasmacytoid DCs exhibit differences in the regulation of antigen presentation, e.g. instead of dampening the presentation of soluble antigens following maturation, these DCs sustained peptide–MHC II formation (109, 110). MHC II synthesis and peptide loading both can be maintained in activated plasmacytoid DCs, whereas these events are curtailed in other maturing DC subsets.

DCs are active in two ‘non-classical’ pathways for antigen presentation: autophagy of cytosolic components for presentation on MHC II (111), and cross presentation of endocytosed substrates on MHC I (2). Plasmacytoid DCs also cross present antigen (112–114). With respect to processing for MHC I, maturing DCs contain a mixture of both conventional and immunoproteasomes (115, 116).

A new field is to move away from isolated DCs and instead study DCs in intact organs, particularly by targeting antigens selectively to them. One way is with monoclonal antibodies to DC receptors. The monoclonal antibodies are genetically engineered or chemically coupled to specific antigens (50, 63, 70, 117, reviewed in 118). Targeting of antigens to DCs enhances presentation to both CD4+ and CD8+ T cells more than 100-fold relative to non-specific control Ig fusion antibodies and to non-targeted antigen. Targeting provides an efficient way to interrogate DC receptor and subset function in vivo.

Initiation of the immune response

All DC subsets, when mature, are potent initiators of T-cell responses in culture and in vivo. Initially this was studied with stimuli that do not require antigen processing like superantigens (119), mitogens (29, 120, 121), and the mixed leukocyte reaction (MLR) (30, 122, 123). During an MLR, numerous clones of CD4+ and CD8+ T cells recognize peptide-allogeneic MHC complexes on stimulator cells, followed by extensive proliferation and development of effector functions (cytokine production, cytolytic activity). Prior to the identification of DCs, the MLR was used to detect MHC incompatibility in donors and recipients of organ transplants or ‘antigenicity’. However, when DCs were enriched and compared with MHC+ B cells and macrophages or selectively depleted with DCIR2 antibody and complement, the DCs proved to be responsible for ‘immunogenicity’, being potent initiators of immunity in vitro (30, 122, 123) and later in vivo (124).

This brings up a nomenclature issue that is more than semantics, as it deals with two major themes of immunology: one is antigenicity, what is recognized by T cells; the other is immunogenicity, what is required for specific immune responses by T cells. Often the term antigen-presenting cell (APC) is used synonymously with DCs, but the concepts and terms are different. The term APC best refers to antigenicity, how any cell can use its MHC products to present peptides that are recognized by receptors on T and NK cells, the major function of MHC products. DCs are a lineage of ‘accessory’ or stimulator cells controlling T-cell responses, both immunogenicity and tolerogenicity, and coupling antigen presentation with a number of other features, e.g. their localization and migration in situ and ability to polarize CD4+ T cells by several pathways to the T-helper 1 (Th1) type. Together, these specializations allow DCs to convert proteins, vaccines, microbes, and cells to peptide–MHC complexes and then select, and either activate or tolerize, rare clones from within the polyclonal repertoire in animals and humans.

Newer criteria used to identify the DC lineage

  1. Top of page
  2. Introduction
  3. Older criteria used to identify the DC lineage
  4. Newer criteria used to identify the DC lineage
  5. References

Recent advances better explain the development and function of DCs and are emphasized in the reviews that follow in this volume. These newer criteria (Table 3) are relevant to all subsets of DCs.

Table 3.   Newer criteria to identify cell lineages and follow their development and function
Functions in immune tolerance, e.g. deletional and regulatory mechanisms
Hematopoietins, e.g. G-CSF, M-CSF, flt-3L, GM-CSF, and their receptors
Committed progenitors, e.g. monocyte and dendritic cell progenitor (MDP) common dendritic cell progenitor (CDP), preDC
Transcriptional programs and controls, e.g. PU1, mafs, Id2, IRFs, E2-2, NF-kB

Control of immune tolerance

For some time in the tolerance field, accessory cell requirements were relatively underemphasized, and classically, very large doses of antigen were used to tolerize or silence specific clones of lymphocytes. The situation changed when antigens were first targeted to DCs in lymphoid tissues in the steady state. Surprisingly, tolerance was the outcome even with low doses of antigen (45, 47, 70, 117), and even in NOD (non-obese diabetic) mice undergoing autoimmune diabetes (125). Plasmacytoid DCs also can induce tolerance (126–131). Selective depletion of DCs from mice leads to autoimmunity (132).

Two types of tolerance take place when antigens are captured by DCs in the steady state. One is T-cell deletion, which can involve ligation of PD-1 and CTLA-4 on T cells (133). A second is induction of foxp3+ regulatory T cells (Tregs) (134, 135), particularly via the CD103+ (136, 137) or DEC-205+ (138, 139) DC subset (reviewed in 140). As discussed below, flt-3L is a hematopoietin for expanding DC subsets and Tregs in vivo (141). This may explain the efficacy of flt-3L in treating autoimmunity in mice (141–143). A likely area of future progress will be to learn how to harness DCs or specific subsets of DCs to specifically silence the polyclonal T-cell repertoire in vivo.


Early studies with human blood distinguished DCs from monocytes, e.g. a lack of CD14; these studies showed that DCs were major inducers of human T-cell responses (123, 144). In the steady state, DCs in lymphoid tissues develop independently of monocytes (145–147). These distinctions are now being explained in terms of specific hematopoietins for monocyte-independent and -dependent DC formation in vivo.

M-CSF is a major ligand for c-fms [the other is IL-34 (148)], the tyrosine kinase growth factor receptor expressed by mononuclear phagocytes (149). M-CSF is needed for macrophage formation from monocytes. By contrast, the combination of GM-CSF and an IL-4 family member (IL-13, IL-15) drives DC development from monocytes (150, 151). Different cytokines, when combined with GM-CSF, may yield different functional forms of monocyte-derived DCs (152), although this field needs to be extended in vivo. In the steady state, M-CSF and GM-CSF dependent monocyte-derived cells are found in several peripheral tissues, e.g. the lamina propria of the intestine (153, 154) and lung (155). Epidermal LCs are M-CSF dependent (156), and they are distinct among DC subsets in their requirement for TGF-β1 (157), which in turn signals through Id2 and Runx3 transcription factors (158, 159).

Flt-3L is the ligand for another tyrosine kinase receptor, flt-3, and DCs are the major lineage responding to flt-3L (160–163). Administration of flt-3L to mice and humans greatly expands the major types of DCs in vivo except for the M-CSF dependent Langerhans cells (164–169). Expansion takes place in lymphoid organs, rather than bone marrow (170), and in non-lymphoid organs like intestine (153, 154). STAT3 is crucial for delivering Flt3 signals to developing DCs, but is dispensable for GM-CSF-mediated DC development, which is mediated by STAT5 (171, 172). The ability to drive DC expansion in vivo with flt-3L (160, 165, 173, 174), and the use of flt-3−/− mice to deplete DCs have become invaluable in studies of the DC lineage and its subsets.

Large numbers of DCs can be generated in tissue culture from bone marrow stimulated with either GM-CSF or flt-3L. It is now appreciated that GM-CSF DCs correspond to monocyte-derived DCs primarily, while flt-3L supports development of monocyte-independent DCs, i.e. CD8+, CD8, and plasmacytoid DCs (170, 175–177). It will be important to revisit these different sources of DCs and compare their functional properties following expansion from bone marrow progenitors in vitro.

DC subset development from committed precursors

In addition to hematopoietins, understanding the development and function of DC subsets is being illuminated by the identification of committed precursor cells. Several recent reports have used expression of different tyrosine kinase bearing receptors i.e. c-kit/CD117, flt-3/CD135, and c-fms/CD115, to identify precursors to DCs and monocytes in mice. Geissmann (178) has identified a common progenitor for monocytes and DCs (MDP) that expresses both c-fms and c-kit. Manz, Shortman, and their colleagues have identified a common DC progenitor (CDP) that is marked by flt-3 but lower levels of c-kit than MDP. CDP gives rise to DCs and plasmacytoid DCs but not monocytes (161, 162).

Liu et al. (72) have recently unraveled how these different precursors operate in vivo. MDPs and CDPs are found primarily in the bone marrow rather than blood and lymphoid tissues. When genetically marked MDPs and CDPs are isolated and injected back into the femurs of other mice, MDPs form CDPs and both DCs and monocytes, while CDPs no longer form monocytes but only DCs and plasmacytoid DCs. The next developmental step is a ‘preDC’ for both CD8+ and CD8 DCs, but not plasmacytoid DCs. The preDC are CD11cintMHC IIlowCD135+. Potential precursors to plasmacytoid DCs also can be found in bone marrow (179). The commitment to separate DC and monocyte pathways in the steady state thus occurs in the bone marrow, whereupon preDCs travel in the blood, enter lymphoid organs via high endothelial venules, and then undergo further proliferation and differentiation to form the characteristic DC network found in the T-cell areas of lymphoid organs.

Transcriptional programs and controls

Transcriptional arrays are now being carried out on different subsets of DCs and are revealing discrete programs for each. Gene arrays on CD8+ and CD8 mouse spleen DCs show considerable differences (63, 180). Interestingly, however, both human and mouse DC subsets are more similar to one another than to other cell types like monocytes (181). Genomic approaches are now mainstream to characterize DC subsets and follow their responses.

Likewise, another major current effort is to identify transcription factors driving the differentiation programs of DC subsets. Genetic deletion reveals key roles of several transcription factors in DC development, although in most cases to date, a deficiency in a particular transcription factor affects more than one DC subset (Table 4). Recently, transcription factors required by specific DC subsets have been identified, such as E2-2 for the development of plasmacytoid DCs (182) and Batf3 for CD8+ DCs (183). The future of the DC subset field will benefit enormously by identifying additional DC-restricted transcription factors and determining how they instruct development and function.

Table 4.   Transcriptional factors selective for development of DC subsets
Transcription factorCD8+ DCsCD8 DCsB220+ pDCsEpidermal LCsReferences
  1. Numbers indicate the approximate change relative to wildtype mice.

  2. ND, not determined.

RelB−/−PresentAbsentNDPresent(184, 185)
Ikaros DN−/−AbsentAbsentNDPresent(186, 187)
Ikaros IkL/LPresentPresentAbsentND(188)
PU.1−/−PresentAbsentNDND(189, 190)
IRF4−/−Present10%50%ND(192, 193)


  1. Top of page
  2. Introduction
  3. Older criteria used to identify the DC lineage
  4. Newer criteria used to identify the DC lineage
  5. References