Dendritic Cell Heterogeneity Tailored to Their Function

Dendritic cells (DC) represent the key link between innate and adaptive immune systems, combining direct recognition of pathogens through pattern recognition receptors such as Toll-like receptors (TLR) and the ensuing orchestration of immune responses through cytokine secretion and antigen presentation to T cells. The foundational view of DC lineage considered it composed of several discrete entities (subsets): conventional or classical DC (cDC) that represent specialized antigen-presenting cells (APC), and plasmacytoid DC (pDC) that excel at the production of type I interferon (IFN-I) and other cytokines. In turn, cDC comprise two distinct subsets: type 1 cDC (cDC1) that are capable of IL-12 production and antigen-cross presentation, and type 2 cDC (cDC2) that efficiently present MHC II-restricted antigens to T helper (TH) cells. The advent of modern technologies such as high-dimensional immunophenotyping, single-cell transcriptomics and lineage tracing added remarkable depth to the above-mentioned view of DC. The three main subsets turned out to be local peaks of a broad spectrum of DC subsets and states, with heterogeneity observed between species (e.g., unique DC-like cells in teleost fish); between tissues (e.g., distinct DC types in the intestine); between subsets (e.g., transitional DC or tDC with features of both pDC and cDC2); and within subsets (e.g., Notch-dependent vs. Notch-independent cDC2). The heterogeneity within populations appears to have multiple sources: some of it may arise from different developmental origins, while some may be generated by tissue-derived signals and environmental cues. In parallel to these new insights into DC populations, studies of the last two decades have revealed a greatly expanded horizon of DC functions. These include insights into DC contribution to TH differentiation during anti-microbial responses, to the maintenance of natural regulatory T cells (Treg) in the steady state, and to the induction of Treg by commensals and food antigens. Moreover, the essential role of DC in nearly all pathological conditions has been rapidly coming into focus: examples include the aberrant activation of DC in autoimmune diseases, the induction of anti-tumor responses by DC and its subversion by tumors, and the hypofunctional state of DC in persistent infections and inflammation. Thus, the field is facing exciting yet difficult challenges: to match the emerging heterogeneity of DC populations to their apparently diverse functions, to establish a clear genetic basis of the resulting relationships, and to reconcile functional DC subsets between experimental animals and humans. This volume attempts to highlight the current status of these efforts by featuring 19 reviews from leading laboratories. They are grouped as described below, according to the “stage” of the DC life cycle: from development and subset specification, to migration and maturation, to functions in immune responses and in tolerance, to their involvement in pathologies such as autoimmunity and cancer. Given the recent progress in the field, they provide important perspectives rather than compendia of facts, and should be considered in context of the ongoing studies (of which only a few most recent ones are cited herein). One unfortunate side effect of the rapid progress is the lack of a unified nomenclature. Indeed, every recently described DC subset is referred to by multiple names while the exact definitions of some populations are being hotly debated. Most of these issues are comprehensively addressed in the corresponding reviews; the readers are encouraged to stay tuned as the field moves towards a consensus. The immune system of the type found in mammals had arrived as a “complete package” in jawed vertebrates, implying the deep evolutionary conservation of the DC lineage. Wen and colleagues 1 review recent progress in the identification of major DC subsets in teleost fish, highlighting the orthologs of cDC and pDC and of key growth factors (e.g., Flt3) and transcription factors (e.g., Id2 and Batf3) guiding their development. They also describe specialized DC populations in zebrafish including brain-specific cDC and a unique evolutionary adaptation: metaphocytes, non-hematopoietic DC-like cells with epithelial characteristics and sensory functions. In adult mammals, DC subsets continuously develop and undergo initial specification in the bone marrow, followed by exit into the periphery and (for cDC) terminal differentiation in tissues. Among transcription factors that control DC development, IRF8 stands out for its involvement at multiple stages and in multiple subsets, consistent with its recently established role in 3D chromatin organization in DC 2, 3. The cDC1 subset represents a bona fide cell lineage that emerges from a single type of committed bone marrow progenitor, and thus displays minimal developmental or functional heterogeneity. Murphy and Murphy 4 review the specification of cDC1, focusing on the key role of IRF8 in the process. They highlight the complex regulation of IRF8 expression by stage- and lineage-specific enhancers, its dose-dependent activity on target genes, and cross-talk with other cDC1-specifying factors such as BATF3. In contrast to cDC1, cDC2 can be derived from multiple types of progenitors and manifest extensive developmental, transcriptional and functional heterogeneity. Minutti and colleagues 5 discuss the two well-established murine cDC2 subsets: the ESAM+ cDC2 that are shaped by tissue-derived Notch, lymphotoxin and retinoic acid signaling, and the developmentally distinct CLEC12A+ subset (termed cDC2A and cDC2B, respectively). The authors emphasize the functional relevance of this heterogeneity as the two cDC2 subsets promote distinct types of T cell responses, including the key role of ESAM+ cDC2 in T follicular helper cell-mediated antibody responses. Apart from the canonical pDC, cDC1, and cDC2 subsets, recent years witnessed the emergence of a distinct subset with shared phenotype and transcriptome of pDC and cDC2, the so-called “transitional” DC or tDC. Although initially considered as by-products of pDC development or as a type of cDC2 precursor, their evolutionary conservation and unique functional properties argue that tDC represent a bona fide specialized DC subset. Idoyaga and colleagues 6 describe the shared origin of tDC with the pDC lineage, their heterogeneity (yes, even they are comprised of at least two subsets!) and the distinct ability to produce IL-1β in response to viruses. They highlight the natural differentiation trajectory of tDC towards cDC2 (specifically the above-mentioned ESAM+ cDC2 subset), emphasizing that pDC, tDC, and cDC represent a continuum of differentiation states and corresponding functionalities. Further adding to the notion of developmental heterogeneity, cDC (particularly cDC2) can be derived not only from DC progenitors in the bone marrow but also from monocytes in the periphery. The contribution of this pathway in the steady state (e.g., to the ESAM− cDC2) remains to be fully elucidated, but its important role in inflammation has been clearly established in both humans and mice. Segura 7 reviews the origin of monocyte-derived DC (mo-DC) from at least two types of myeloid progenitors, transcriptional control of their differentiation and choice between mo-DC and macrophages, and developmental cues including cytokines and aryl hydrocarbon receptor agonists. She notes the emerging role of mo-DC as tissue-resident APC that can shape context-dependent T cell differentiation and re-stimulate effector and memory T cells in situ. A key part of the DC life cycle is their mobilization and migration into lymphoid organs such as regional lymph nodes (LN), followed by localization in the appropriate anatomical niche within these organs. All cDC subsets that reside in tissues or lymphoid organs undergo this process, which is driven primarily by the chemokine receptor CCR7 but was also recently shown to utilize additional signal transduction pathways 8. The migration is hard-wired into the DC lineage: it can be precipitated by pathogen sensing, but otherwise it occurs spontaneously in the absence of pathogens or even of commensal microbes. The resulting localization of migratory DC (migDC) in specific LN niches enables their interaction with T cells and other immune cells, and thus is critical for all DC functions. Lennon-Dumenil and colleagues 9 discuss the mechanisms that initiate the switch from the sensing behavior of cDC in tissues towards the migratory behavior that allows DC exit from the tissue and passage through the lymph. They emphasize the role of the mechanosensing of extracellular matrix, particularly the mechanical confinement of the nuclear envelope, and the resulting cytoskeleton remodeling as important drivers of the process. Eisenbarth and colleagues 10 give an organism-wide overview of DC migration guided by CCR7 ligands CCL19/21. They describe DC migration within the spleen, or between the skin, lung, or intestine and the respective LN, which results in the precise subset-specific localization of migDC in these lymphoid organs. Huang and Gerner 11 focus on DC positioning within the LN, including the unique localization of resident and migratory DC subsets, and their repositioning during type 1 inflammation. They emphasize the role of DC repositioning in the trafficking of other innate cells and in the remodeling of the LN architecture, in addition to its role in T cell priming. DC migration comprises a part of the broader program of maturation, which is accompanied by the profound remodeling of DC physiology and gene expression and culminates in DCs' arrival in LN and interaction with T cells. Paradoxically, this conserved program (driven largely by NF-κB signaling) can lead to two opposite outcomes: pathogen-induced immunogenic maturation leads to the priming of productive T cell responses, whereas homeostatic maturation in the steady state leads to preferential DC interactions with Treg that facilitate Treg function and prevent T cell activation. Although originally described in mice, homeostatic maturation of cDC2 has been recently documented in the human spleen 12. Akyol and Dalod 13 review key aspects of DC maturation, focusing primarily on cDC1 as a conserved DC subset with unique functions. They discuss the convergent nature of immunogenic and homeostatic maturation programs, as well as molecular differences between them. They also emphasize the emerging role of cholesterol metabolism in the process, and illustrate the precise spatiotemporal control of cDC1 maturation during the murine cytomegalovirus infection. Kumamoto and colleagues 14 detail several waves of DC interaction with T cells during T cell priming in LN, and emphasize the distinct and cooperative roles of migratory DC subsets in the process. They consider T helper type 2 (TH2) differentiation as an informative example, and highlight a specific role of CD301+ cDC2 located near high endothelial venules in TH2 priming. Unlike cDC, pDC are not known to undergo homeostatic maturation in the steady state, except for the IFN-I-driven diversification that confers antiviral resistance. On the other hand, pathogen sensing by pDC induces rapid IFN-I production followed by bona fide immunogenic maturation into cDC2-like cells. Zuniga and colleagues 15 review this activation-induced pDC differentiation, which is facilitated by IFN-I and transcriptional reprogramming. They also discuss the effect of persistent viral infections on pDC including their impaired generation and reduced functionality (exhaustion), the latter mediated by metabolic changes such as lactate accumulation. Barrat and Guery 16 discuss a unique aspect of human pDC function, i.e., their genetic mosaicism in females due to the reactivation of an inactive X-chromosome. The resulting increased dosage of key functional genes such as TLR7 enhances pDC function via a feed-forward loop, which may contribute to female bias of several autoimmune diseases. Although the term “tolerogenic DC” is used frequently, there is no clear consensus on its meaning. Indeed, it may imply (i) a particular tolerance-promoting stage in the general DC life cycle; (ii) a tolerogenic role of a common DC subset in a specific tissue, or (iii) a developmentally distinct DC subset dedicated to the induction or maintenance of tolerance. Recent advances suggest that all these scenarios contribute to the tolerogenic function of DC in the steady state. Regarding the first scenario, homeostatic maturation of all cDC subsets promotes Treg maintenance and peripheral tolerance as discussed above. This notion has been highlighted by the recent identification of transcription factor ETV3 as a specific regulator of homeostatic DC maturation whose deletion impairs Treg function and induces autoimmunity 17. Regarding the second scenario, the obvious organ to consider is the thymus, where DC may directly affect T cell selection and central tolerance. Thymic DC are reviewed by Hogquist and colleagues 18, who highlight the heterogeneity of thymic DC subsets and their specific roles in T cell selection and tolerance: cDC1 in acquiring antigens from AIRE-expressing medullary thymic epithelial cells (mTEC) and facilitating thymic Treg induction, and cDC2 in the clonal deletion of autoreactive T cells. The authors also describe the immigration of peripheral cDC2 into thymus, which may import peripheral antigens and contribute to T cell selection “in real time”. Another relevant tissue is the gut, where DC facilitate induction of tolerance to food- and commensal-derived antigens. Esterhazy and colleagues 19 discuss intestinal DC subsets and their localization that reflects specialized functions: while migratory CD103+ DCs bring antigens to mesenteric LN and facilitate Treg induction, CX3CR1+ DC locally produce IL-10. Further, they discuss how antigens of different types (soluble vs. particular) and abundance may tolerize via different DC subsets, and emphasize the imprinting of tolerogenic function by tissue-derived factors such as retinoic acid and fatty acids. Regarding the third scenario, a distinct DC-like cell with tolerogenic properties has been described in the last few years: the RORγt-expressing APC. Although these unique cells were recently proposed to branch off lymphoid development 20, 21, their shared transcriptional signature with cDC and their tolerogenic function in the intestine warrant their consideration as DC-like cells. Fu and Littman 22 emphasize that these “tolerogenic DC” or tolDC mediate the induction of Treg to intestinal antigens, while the above-mentioned CD103+ DCs may be necessary but not always sufficient for this outcome. The authors review the transcriptional regulation of tolDC (including the combined roles of RORγt and PRDM16), their important contributions to the tolerance of commensal microbes and food antigens, and their likely collaboration with other DC subsets. Consistent with their important role in tolerance, the aberrant activity of DC contributes to the tolerance breakdown in autoimmunity. For instance, aberrant production by pDC of IFN-I and other mediators such as CXCL4 has been shown to exacerbate systemic autoimmune diseases including systemic lupus erythematosus (SLE) and scleroderma, as discussed by Barrat and Guery 16. Recent genetic evidence also suggests the role of cDC maturation in SLE and other autoimmune diseases: for example, several genes uniquely expressed in mature DC, such as CCL22 and the above-mentioned ETV3, have been linked to SLE by genome-wide association studies. Two reviews in this volume describe the role of aberrant DC maturation in autoimmunity, focusing particularly on its metabolic aspects. Quintana and colleagues 23 emphasize how the reprogramming of lipid and amino acid metabolism affects DC migration and chromatin organization, facilitating the switch from homeostatic to immunogenic DC maturation in neurological autoimmunity. Cao and colleagues 24 analyze this process in the context of SLE and related autoimmune diseases. They focus on the switch from fatty acid oxidation and mitochondrial respiration in the steady state to glycolytic reprogramming during inflammation as one mechanism that facilitates aberrant maturation of DC to acquire immunogenic properties and participate in autoimmunity. Conversely, impaired immunogenic maturation and function of DC contribute to defective immune responses in conditions such as cancer and chronic inflammation. In one of the most recent examples, rescuing the impaired DC migration from the tumor site to the draining LN was shown to improve anti-tumor responses 8. Pass and Krummel 25 review the key role of cDC1 as regulators of anti-tumor CD8+ T cell immunity, their ability to form spatial hubs with NK and CD8+ T cells, and their relation to responsiveness to immunotherapy. They also describe the role of cDC2 in anti-tumor CD4+ T cell responses and of migDC in the transport of tumor antigens to LN, and highlight therapeutic opportunities such as DC activation by cytokines and/or TLR ligands and DC-based vaccines. Villadangos and colleagues 26 discuss the function of DC in systemic inflammatory response syndrome (SIRS) and related conditions, describing the early mechanisms such as the loss and systemic activation of cDC, and late mechanisms such as cytokine-mediated functional impairment. They draw parallels with impaired DC function in tumors, and discuss therapeutic approaches that may restore DC function in systemic inflammation. In summary, this volume provides a snapshot of the current exciting state of the DC field, in which several new populations are emerging, the genetic basis of their development and function is being elucidated, and major new aspects of DC function are becoming apparent. The next few years should hopefully witness the “great unification”, wherein all “shades” on the spectrum of DC types will be defined by a broad consensus and matched to their unique functions in the steady state, during normal immune responses and in pathological conditions. National Institute of Health Grant Awards AI072571, AI164728, and AI128949. The author declares no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.