Monocyte ontogeny

In mice, monocytes consist of at least two functionally and phenotypically distinct subsets. Classic monocytes, or inflammatory monocytes, are Ly6C ^high (GR-1 ⁺ ) with proinflammatory properties, migrating to injured ^, or infected tissues from the blood or bone marrow via signaling through chemokine receptor CCR2. Nonclassical Ly6C ^neg (GR-1 ^neg ), or patrolling, monocytes reside in the vasculature and crawl along the luminal vessel surfaces of endothelial cells to surveil tissues during the resolution of inflammation. Ly6C ^neg monocytes enter into tissues during the latter phase of infection, via CX3CR1 signaling, exerting a tissue repair immune regulatory program. Ly6C ^hi monocytes have a short-lived half-life of 20 hours but constitute the steady-state precursors to the blood-residing Ly6C ^neg migratory pool such that the abundance of Ly6C ^hi monocytes dictates the pool of Ly6C ^neg progeny. Monocytes are thought to be derived from the common monocyte progenitor (cMoP), a myeloid precursor population found in the bone marrow in mice, and phenotypically defined as CLEC12A ^hi CD64 ^hi subsets in humans. Progenitor populations were originally thought to derive in a hierarchical progression from common myeloid progenitor (CMP)- granulocyte-monocyte progenitor (GMP)- monocyte-macrophage dendritic cell progenitor (MDP)-cMoP-monocyte. However, it has been demonstrated that under specific inflammatory contexts a bifurcation in the pathway occurred whereby CMP arose independently from MDP into GMP-derived neutrophil-like monocytes whereas MDP differentiated into monocyte-derived DCs. Thus, distinct ontogenies likely underlie observed functional heterogeneity in monocyte populations. Besides originating from the bone marrow, large pools of monocytes reside in the spleen and lungs, where they are maintained and are recruited en masse in response to inflammation, injury, or other systemic insults.

Most studies investigating monocyte origins and deployment throughout the body have used murine models. Human counterparts exist and have both similar and distinct functional attributes compared with their murine counterparts, which are outlined in Table 4.1 . In humans, monocytes represent 10% of the nucleated cells in the blood, whereas in mice monocytes represent 4% of nucleated blood cells. In humans, monocytes are defined based on CD14 and CD16 expression. CD14 ⁺⁺ CD16 ⁻ monocytes are more similar in function and phenotype to Ly6C ^hi monocytes in mice, including having high expression of CCR2 and CD115 while also having lower CD16 expression levels. On the other hand, CD14 ⁺ CD16 ⁺ monocytes also appear to have similar attributes to Ly6C ^neg monocytes, including having elevated CX3CR1 expression. Despite these similarities, important differences between murine and human monocytes have been identified. For example, comprehensive transcriptional and flow cytometric profiling of murine monocytes and their human equivalents revealed that murine monocytes had high PPARγ expression signatures, whereas human monocytes had opposed patterns of surface receptors responsible for recognition and uptake of apoptotic cells and phagocytosis.

Monocyte function: Differentiation during homeostasis and inflammation

Monocytes are recruited in waves, where short-lived Ly6C ^hi monocytes contribute to early inflammation and Ly6C ^neg monocytes specialize in tissue repair and immune regulation. Ly6C ^neg monocytes are longer-lived, with a half-life of up to 2 weeks. Upon entering tissues, and depending on the signals present within a given tissue environment, monocytes will either act as terminally differentiated effectors, as outlined previously, or differentiate into tissue macrophage and dendritic cell populations.

In steady-state conditions, little evidence exists to indicate that classic Ly6C ^hi monocytes contribute significantly to the pool of tissue-resident macrophages. ^{, ,} In the context of pathologic inflammation, Ly6C ^hi monocytes repopulate both tissue-resident macrophage and dendritic cell pools (for in-depth review, see Wynn et al. ). One interesting exception to this paradigm is the lamina propria myeloid cells residing in the intestines. Using macrophage depleted conditions, monocyte fate mapping, adoptive transfer of monocytes, and parabiosis experiments, Varol et al. and Bogunovic et al. both demonstrated that circulating inflammatory Ly6C ^hi monocytes replenish the pool of immune-regulatory CX3CR1 ⁺ macrophages but not of CD103 ⁺ lamina propria migratory dendritic cell subsets. ^, This process occurs during homeostasis and is thought to differ from other steady-state macrophage pools because of the low-level tonic signaling by the commensal microbiota. Follow-up studies demonstrated that signals from the commensal microbiota were essential for monocyte-driven repopulation of luminal macrophages, because antibiotic treatment significantly depleted monocyte recruitment into the intestines, whereas germ-free mice exhibited a significant reduction in luminal macrophage levels. Thus, the chronic signals derived by the gut microbiota were central for continued recruitment of Ly6C ^hi monocytes into the lamina propria. During acute inflammation, such as during colitis, Ly6C ^hi monocytes respond to bacterial products and give rise to proinflammatory effectors that eventually differentiate into migratory antigen-presenting dendritic cell subsets. On the other hand, noninflammatory patrolling monocytes in both mice and humans directly give rise to migratory dendritic cell subsets in homeostatic settings. This occurs in the absence of acute inflammation, where Ly6C ^hi monocytes recirculate back to the bone marrow and are converted into Ly6C ^neg populations that replenish depleted dendritic cell populations in mucosal, but not splenic, tissues.

Monocytes are versatile cell types and, similar to other myeloid-derived counterparts, exhibit sensitivity to the environment in which they reside. Functionally, monocytes exhibit plasticity and have a broad functional repertoire and differentiation potential that is dictated by the signals present in the tissues that they are recruited into. These properties endow monocytes with the potential to elicit robust inflammatory responses or participate in tissue remodeling and repair. Understanding the basis for monocyte fate and function in specific inflammatory contexts will be important for targeting these functionally diverse subsets for the treatment of various diseases, including cancer.

Macrophage ontogeny

Macrophages populate tissues throughout the body in distinct waves during embryogenesis, with inputs from the yolk sac, fetal liver hematopoietic stem cells, or bone marrow. The first wave originates from the yolk sac, where erythroid–myeloid precursors either populate tissues directly or establish a niche of embryonic hematopoietic stem cells in the fetal liver. Monocytes originating from the fetal liver constitute the second wave of macrophage seeding into embryonic tissues. With the exception of the intestinal and cardiac macrophages, the latter of which are renewed by adult monocytes during aging, tissue macrophages are able to self-renew into adulthood. Brain microglia are the only population of tissue macrophages seeded exclusively by yolk sac embryonic precursors, whereas heart macrophages, liver Kupfer cells, and skin Langerhans cells are seeded by both the yolk sac and fetal liver. ^, Alveolar macrophages, kidney macrophages, and red-pulp macrophages are seeded only by fetal liver monocytes. ^,

Macrophage cellular functions: Tissue-specific functions and the spectrum model of activation

Macrophages are considered professional phagocytes, a cell that specializes in phagocytosis, deriving from the Greek words phago (to devour) and cytos (cell). Phagocytosis is a cellular process in which innate immune cells engulf and ingest large materials (greater than 0.5 μm in size) such as dead and dying cells, pathogens, or other particulate materials. Phagocytosis by macrophages and other professional phagocytes is essential for immune homeostasis and host defense, a process that was first described by Elie Metchnikoff, who received a Nobel Prize for his work in 1908 (for in-depth review, see Gordon ). Phagocytosis occurs in multiple phases beginning with sensing of the particle to be ingested through receptors on the cell surface, ^, internalization of the particle through the plasma membrane and into a distinct cellular compartment termed the phagosome , and fusion of the phagosome with lysosomes, an organelle responsible for degradation of phagocytosed macromolecules.

Tissue-resident macrophages exhibit multiple specialized functions for distinct tissue compartments and are identified by specific transcription factors that are induced in response to the tissue microenvironment. Several functionally nonredundant macrophage populations reside in the spleen. Macrophages residing in the red pulp are regulated by the transcription factor PU.1, enabling localization to red pulp for phagocytosis of damaged and dying red blood cells scavenging released iron to maintain iron homeostasis. On the other hand, marginal zone splenic macrophages are dependent on LXRa signaling, where they specialize in the clearance of apoptotic cells while also regulating selective uptake of apoptotic cells by CD8a dendritic cells, a subset in the spleen responsible for regulating autoreactivity to self-antigens. Alveolar macrophages express the transcription factor B lymphoid transcription repressor BTB and CNC homology 2 (Bach2), which enables lipid handling and clearance of surfactant in the alveolar spaces of the lungs. They also facilitate gas exchange and prevent against hypoxia during infection.

Although there are multiple additional examples of tissue specification for macrophages, these studies underscore that macrophage function and phenotype are dictated by the signals present within the tissues that they reside. The advent of single-cell technologies to measure differences in gene expression, protein production, and chromatin regulation have enabled discovery of specific signaling pathways contributing to macrophage heterogeneity during health and disease. This concept is central to the heterogeneity of macrophages observed in disease settings such as cancer. Through transcriptional and epigenetic profiling of tissue-resident macrophages, Lavin et al. identified that aside from lineage-specific transcription factors, tissue-specific microenvironmental signals regulated macrophage identity and function through the induction of core transcription factors and specific chromatin modifications. Tissue signals were so robust that transplanted macrophage precursors or mature macrophages were effectively epigenetically “reprogrammed” when transferred into new tissue environments, exemplifying macrophage functional and transcriptional plasticity. Macrophage plasticity and the ability to be reprogrammed in response to changes in the microenvironment are key considerations when developing anticancer therapies, as will be outlined in the following.

The conceptual framework of macrophage activation has long maintained that macrophages can be polarized into two opposing states: M1, or classically activated inflammatory macrophages, and M2, or alternatively activated immune regulatory macrophages. Studies evaluating macrophage response to acute infection, allergies, obesity, and asthma supported the paradigm that macrophages existed in one of two polarized states. However, this model did not account for the specialized functions of tissue macrophages, macrophages during sterile inflammation and tissue repair, and resolution of inflammation, nor did it encompass macrophage functional changes during chronic and complex disease states such as autoimmunity, chronic infection, and cancer. This paradigm has evolved coincident with the work related to macrophage ontogeny, with elegant studies demonstrating that macrophages have a broad functional repertoire with multiple, distinct, activation programs. Using ex vivo–derived human macrophages, Xue et al. differentiated macrophages in 29 distinct conditions ranging from stimuli associated with M1 and M2 activation axis to stimuli associated with free fatty acids, high-density lipoprotein, or combinations associated with chronic inflammation. Using transcriptomics and network analysis, it was found that in conditions diverging from the classical M1 and M2 paradigm, macrophages had a broad spectrum of activation signatures. These studies support the framework that macrophage activation is multidimensional, where macrophages receive input from tissue environments and external stimuli such as sterile insult, pathogens, or cancer, culminating in a broad range of activation phenotypes. The multispectral model is highlighted in Fig. 4.1 .

Natural killer cells

Natural killer (NK) cells are central for host defense against virally infected or transformed cells. NK cells are widely distributed throughout the body, present within secondary lymphoid tissues such as lymph nodes, spleen, and mucosal associated lymphoid tissue in addition to tissues such as the lung, uterus, and liver. In the blood and lymph nodes, NK cells represent approximately 5% to 15% of total lymphocytes, whereas in the liver, lung, and uterus they are represented in higher frequencies. NK cells derive from common lymphocyte progenitors in the bone marrow. However, unlike B- and T-cell lymphocytes, NK cells do not require sensitization to kill target cells because of the lack of expression of antigen-specific receptors. Instead, NK cells express a series of germline-encoded activating and inhibitory receptors that enable distinction between self and nonself, a process referred to as education or licensing . Although NK cells are generally considered an innate lymphocyte subset, NK cells are also distinct from innate myeloid cells because they do not have the ability to phagocytose cells. Instead, NK cells possess effector capabilities resembling that of T cells, including the secretion of perforin, granzymes, and interferon gamma (IFN-γ). NK cells also express FasL and TRAIL on the cell surface, both of which bind to death receptors on recipient cells, leading to the induction of apoptosis. NK cells also produce a multitude of effector cytokines and chemokines, including tumor necrosis factor alpha (TNF-α), interleukin (IL)-10, GM-CSF, G-CSF, IL-3, CCL2, CCL3, CCL4, CCL5, XCL1, and CXCL8.

NK cells are derived from common lymphoid precursors in the bone marrow prior to populating peripheral lymphoid and nonlymphoid tissues. Commitment to the NK lineage coincides with expression of CD122, a receptor for IL-15, followed by the induction of transcription factors such as STAT5, EOMES, Nfil3, Ets, T-bet, Id2, Tox, and Tcf1. As NK cells mature, they upregulate expression of receptors such as NKp46 and CD49b, a point at which they are considered to be mature because of their potential to exert cytotoxic functions. NK cells continue to progress through a four-stage process of maturation, consisting of transcriptional and functional changes and defined in mice by upregulation of CD11b and gradual loss of CD27, with terminally differentiated CD11b ⁺ CD27 ⁻ effectors having potent cytolytic capabilities and reduced potential for expansion. In humans, NK-cell maturation is thought to associate with diminishing levels of surface CD56 expression. Using single-cell RNA sequencing, CD56 ^bright populations were found to have gene expression similar to that of the less differentiated CD11b ⁻ CD27 ⁺ subset, whereas CD56 ^dim populations resembled CD11b ⁺ CD27 ⁻ populations. However, it remains unknown whether CD56 expression truly identifies human NK subsets from the same lineage, progressing through distinct maturation states, or whether its expression defines subpopulations with unique developmental trajectories.

Central to NK cell function is the expression of a broad array of membrane proteins that calibrate NK cell interactions between other cells, resulting in inhibition or induction of cell cytotoxicity. NK receptors are broadly categorized into activating, inhibitory, and costimulatory receptors. Licensed NK cells express inhibitory receptors that contain immunoreceptor tyrosine-based inhibitory motifs (ITIMs) in their cytoplasmic domains. Upon recognition of self-MHCI, ITIM domains are phosphorylated by SH2 domain–containing protein tyrosine phosphatase 1 (SHP1), effectively inhibiting intracellular signaling pathways controlling NK cell activation. Many NK-activating receptors contain immunoreceptor tyrosine-based activation motif (ITAM)-signaling molecules that, upon receptor crosslinking, lead to activation of the ITAM motif and downstream signaling events associated with NK cytotoxicity and effector function (for in-depth review, see Yokoyama and Plougastel ).

Mast cells

Mast cells are evolutionarily conserved innate immune cells found in connective tissues throughout the body, such as perivascular sites, brain, smooth muscle, peritoneal cavity, gastrointestinal tract, and respiratory tract. Mast cells are known for initiating inflammatory responses during allergic diseases. Although mast cells are often neglected in cancer research because of their low abundance in tissue and tumor microenvironments, it has become clear that in cancer and noncancer settings, mast cells are pivotal regulators of immune function.

Non-tissue-resident mast cells develop from hematopoietic progenitor populations in the bone marrow where precursor populations are recruited to distal sites via chemokine gradients into the tissue and undergo division, ^, followed by maturation and activation by tissue microenvironmental factors. On the other hand, tissue-resident mast cells are derived from yolk sac progenitors. Tissue mast cells are long-lived and can be replenished from precursors residing in the tissue. ^, However, very little is known how progenitor tissue-resident mast cells are matured. For bone marrow–derived mast cells, progenitors differentiate from common myeloid progenitors (CMPs), a process that is tightly controlled by CCAAT/enhancer binding protein α (C/EBPα), GATA-binding factor 2 (GATA-2), GATA-binding factor 3 (GATA-3), Hes-1, and melanocyte-inducing transcription factor (MITF). The downregulation of C/EBPα in GMP along with upregulation of GATA-2, GATA-3, and Hes-1 allows GMP to differentiate into bipotent progenitors, which have the potential to become mast cells and basophils. The subsequent upregulation of MITF and further downregulation of C/EBPα lead to the commitment of progenitors into mast cells.

A tremendous amount of work has been done to identify subtypes of mast cells in animals and humans, where the classification of mast cell subsets is based on the anatomic location and/or the protein contents in the secretory granules. In mice, the most well-characterized subtypes of mast cells are connective tissue mast cells (CTMCs) and mucosal mast cells (MMCs). CTMCs are found in skin, the peritoneal cavity, and the gastrointestinal tract, whereas MMCs usually populate at mucosal sites. CTMCs have granules that contain heparin, histamine, tryptase, and chymase, whereas the granule contents of MMCs mainly comprise chymase and have low levels of histamine and undetectable levels of heparin. In humans, mast cells are classified as MC _T (mast cells containing mainly tryptase) and MC _TC (mast cells containing both tryptase and chymase). MC _T reside in the external mucosa of the gastrointestinal and respiratory tracts, whereas MC _TC are usually found in the submucosa and perivascular tissues. ^, Distinct tissue tropisms give rise to distinct functional attributes, with MC _T critically regulating the immune responses and MC _TC being involved in tissue repair. ^, Although human and murine mast cells share multiple protein signatures, mast cells from each species also produce a distinct repertoire of proteases and respond differently in the presence of the mast cell stabilizer cromolyn. Host-intrinsic and/or tissue microenvironment–driven variations can also contribute to phenotypical differences in mast cells. ^,

Neutrophils

Neutrophils, or polymorphonuclear leukocytes, are the most abundant cell type in the peripheral blood, approximating 50% to 70% of cells in human blood and 10% to 25% of cells in murine blood. Neutrophils are phagocytic and granulocytic cells considered one of the first lines of defense against pathogens. On the other hand, neutrophils also have a central role during immune homeostasis and tissue repair. The neutrophil life cycle is still an intense area of debate, owing to a lack of reliable techniques to accurately trace the path of neutrophil differentiation, circulation, tissue seeding, and eventual death in vivo. Compared with other innate immune cell populations, it is generally accepted that neutrophils are short-lived, surviving only hours to days once in the periphery. Given their critical role for innate immune defense, the numbers of neutrophils circulating in the periphery, and their short half-life, there is a significant requirement for de novo production of progenitors from the bone marrow daily. It is estimated that approximately 60% of all leukocytes produced daily in the bone marrow are granulocytic precursors, with 10 neutrophils produced and released from the bone marrow in humans and 10 neutrophils produced in mice daily. During inflammation, as many as 10 billion neutrophils can be produced in humans.

Neutrophils are derived from the multipotent progenitors in the bone marrow, undergoing commitment prior to entrance into the blood. Prior to the work by Evrard et al., it was unknown what factors contributed to the phenotypic and functional heterogeneity of peripheral neutrophils. Evrard et al. demonstrated that neutrophils arise from CMPs, transitioning into GMPs and eventually into a previously unrecognized precursor population they termed preneutrophils . In homeostatic conditions, commitment of GMP into preneutrophils is driven by expression of the transcription factor C/EBPε and expression of cKit and CXCR4. Preneutrophils are highly proliferative and reside in the bone marrow and spleen. During inflammatory stress, such as sepsis or cancer, preneutrophils rapidly expand to meet the increased demand, giving rise to two distinct neutrophil subsets: nonproliferative immature Ly6G ^lo/+ CXCR2 ⁻ CD101 ⁻ neutrophils and Ly6G ⁺ CXCR2 ⁺ CD101 ⁺ mature neutrophils within the blood. Other studies have identified that multipotent progenitors could act as a source of neutrophils in response to hematopoietic demands, whereas Cugurra et al. identified that the skull and vertebral bone marrow can supply the meninges with neutrophils under pathologic conditions such as spinal cord injury and neuroinflammation. Altogether, these studies have begun to establish the underlying mechanisms of neutrophil heterogeneity, underscoring that neutrophil phenotype and function are dependent on systemic inflammatory factors in addition to the tissue of origin.

The production, trafficking, and egress of mature neutrophils from the blood was demonstrated to occur in a diurnal pattern, where the exit of CD62L ^low CXCR4 ^high “aged” neutrophils from the blood back into the bone marrow and other tissues peaked during resting phases and was followed by emergence of “fresh” neutrophils during awake hours. Casanova-Acebes et al. went on to demonstrate that the circadian clock gene Artnl1 (BMAL1) was responsible for the diurnal regulation of neutrophil migration. BMAL1 was found to be a cell-intrinsic regulator of CXCL2 production and subsequent signaling through the ligand CXCR2, initiating transcriptional and migratory changes in neutrophils that trigger the aging cascade. On the other hand, the chemokine receptor CXCR4 antagonized aging via CXCR2 signaling. Although immature neutrophils express high levels of CXCR4 while being retained in the bone marrow, CXCR4 is paradoxically expressed at high levels in aged neutrophils, a process that is hypothesized to signify a return into the bone marrow and other sterile tissues to bolster to tissue immunity.

Functionally, neutrophils exhibit a broad range of capabilities that are dictated by the tissue environments into which they migrate, by the signals in the tissue or systemic environment they are exposed to, and by a culmination of physiologic factors that differently prime neutrophils in their functional response. For example, neutrophils in an obese setting can metabolize lipids and affect energy expenditure, whereas neutrophils in the spleen provide help to B cells. During inflammation, neutrophil longevity is increased upon activation, which can occur via recognition of tissue insults through expression of multiple innate signaling receptors. Neutrophils are highly phagocytotic cells; they can release potent amounts of reactive oxygen species and proteases from their granule contents and can release neutrophil extracellular traps, or NETs. NETs are an extracellular release of neutrophil DNA into a net-like configuration, effectively trapping microorganisms. NETs contain a high concentration of antimicrobial molecules and other molecules that are toxic to the captured microbes. As discussed below, the heterogeneity of neutrophil effector function and sensitivity to systemic signals result in significant heterogeneity in neutrophil phenotypes and functions in cancer-bearing individuals.

Monocytes

One of the major bottlenecks in defining how monocytes affect cancer is the lack of markers that differentiate monocyte-derived macrophages and dendritic cells from monocytes. This is because monocytes upregulate macrophage and dendritic cell markers such as F4/80, MHCII, and CD11c once they enter the tissue. These attributes, along with progenitor pathways that are distinct from those of other myeloid subsets during conditions of inflammation, make tracking the trajectory that monocytes take from the bone marrow into the tumor environment more challenging. Despite these challenges, lineage tracing, specialized transgenic mice, and single-cell analysis technology have provided insight into the pro- and antitumor activity of monocytes during cancer.

Tumor- and stroma-associated expression of the chemokine CCL2 is one of the main signaling pathways involved in the recruitment of inflammatory monocytes into the tumor microenvironment. Once monocytes enter the tumor microenvironment, they are coopted to enhance tumor growth, enable immune evasion, or promote metastatic dissemination. Qian et al. demonstrated CCL2 signaling recruited Ly6C ^hi CCR2-expressing monocytes into metastatic sites enhancing extravasation, tumor seeding, and growth in the lungs. Once in the tumor microenvironment, inflammatory CCR2-expressing monocytes contribute to the pool of tumor-associated macrophages. Other relevant signaling pathways associated with polarizing monocytes to acquire tumor-associated macrophage functions include CSF1 signaling. In a tumor setting, CSF1 enhances recruitment and polarization of monocytes into macrophages. However, other pathways have been identified, indicating that there are multiple tissue and microenvironmental contexts in which monocytes can become polarized to become part of the tumor-associated macrophage pool. For example, expression of the receptor for IL-4 and IL-13 (IL-4R) results in polarization and enhanced tumor-promoting functions of monocytes recruited in the bone marrow, culminating in enhanced recruitment to and outgrowth of breast tumor cells in the bone marrow. Another study identified that IL-6 and GM-CSF produced by cancer-associated fibroblasts induce polarization of monocytes into tumor-associated macrophages. Inflammatory monocytes can also directly promote metastasis. In certain tumor models, inflammatory monocytes recruited into the tumor microenvironment via CCL2 express factor XIIIA, a transglutaminase that covalently crosslinks fibrin, creating a scaffold for tumor cells to invade and metastasize. Monocytes also modulate the tumor microenvironment via production of IL-1β, IL-6, TNF-α, and CCL3, resulting in immune suppression and enhanced angiogenesis (see review by Olingy et al. ).

Nonclassical or patrolling monocytes are longer-lived than classic or inflammatory monocytes, constantly surveilling tissues and transporting antigen to draining lymph nodes. These attributes make patrolling monocytes well-suited for antitumor immune functions. Nonclassic monocytes patrol the lung environment for tumor cell debris and can activate NK cells to prevent metastatic seeding of the tissue. Mechanistically, it was subsequently shown that patrolling monocytes are expanded by tumor-produced exosomes and activated to produce IL-15, leading to subsequent activation of IFN-γ-producing NK cells that prevent metastatic colonization of the lungs. Despite the tumor protective functions of patrolling monocytes, they are still susceptible to the signals within the tumor microenvironment and can be modulated or polarized to support tumor growth. For example, tumors treated with anti-VEGFR2 therapy acquire resistance. One of the mechanisms of resistance involves upregulation of CX3CR1 and recruitment of patrolling monocytes into the tumor microenvironment. Once recruited, the monocytes enhance angiogenesis and immune suppression through the production of chemokines that recruit suppressive neutrophils into the tumor microenvironment.

Dendritic cells

Dendritic cells are critical for the initiation of adaptive tumor immune responses. However, as tumors progress, conventional dendritic cells are coopted by tumors acquiring tolerogenic functions. These dendritic cells are termed tumor-induced regulatory dendritic cells . Although distinguishing dendritic cells from macrophages and other myeloid cells is more challenging under conditions of emergency myelopoiesis, bona fide dendritic cells can be identified by expression of the transcription factor Zbtb46 ^, and the presence of DNGR1. Using an approach to selectively deplete dendritic cells in an autochthonous model of ovarian cancer, Scarlett et al. demonstrated that during early stages of tumor progression, dendritic cells were instrumental for orchestrating antitumor CD8 T-cell immune responses. However, as tumors continued to evolve and the tumor microenvironment became more suppressive, tumor-produced transforming growth factor beta (TGF-β) and prostaglandin E2 (PGE2) polarized dendritic cells to suppress T-cell responses. Mechanistically, dendritic cells became suppressive in response to sustained and elevated expression of the chromatin regulator Satb1. Although required for differentiation of conventional dendritic cells, unremitting Satb1 expression induces a phenotypic switch in conventional dendritic cells, leading to enhanced secretion of immune suppressive galectin 1 and tumor-promoting IL-6.

Conventional dendritic cells can be polarized via additional suppressive pathways, including interactions with MDSCs, regulatory T cells, and metabolic stress. For an extensive review of mechanisms associated with dendritic cell polarization and suppression, see Veglia and Gabilovich, Wculek et al., and Conejo-Garcia et al. Herber et al. observed that dendritic cells in patients and in murine models of cancer were abnormally laden with oxidized lipids, a state termed dyslipidemia , resulting in endoplasmic reticulum (ER) stress, which is discussed later, and a reduced ability of dendritic cells to activate T cells. Mechanistically, lipid accumulation was driven by expression of scavenger receptor A whereby oxidatively truncated lipids were bound to chaperone HSP70, effectively preventing transport of exogenous peptide-MHC class I complexes to the cell surface, limiting the ability of dendritic cells to stimulate CD8 T cells.

Plasmacytoid dendritic cells represent a rare population within the tumor microenvironment. As such, evaluating the effects of pDCs on cancer progression has been challenging, and it is still unclear whether pDCs conclusively have pro- or antitumor functions. However, given the plasticity in the dendritic cell lineage, it is likely that pDCs have varied function depending on the signals that are present in the tumor microenvironment. Within the tumor microenvironment, pDCs were shown to enhance immune suppression by expanding regulatory T cells in an inducible T-cell costimulatory ligand (ICOSL)-dependent manner while also inducing IL-10 in regulatory T cells or initiating type 2 immune responses in CD4 T cells, both of which are immune suppressive and protumorigenic. However, evidence also exists that plasmacytoid dendritic cells induce cytotoxic T-cell responses via production of IFN-α. Additionally, when activated with ligands for TLR7, pDCs initiate substantial antitumor immune responses.

Abnormal differentiation of myeloid cells and infiltration of tumors with immature dendritic cells are considered primary mediators of myeloid dysfunction within the tumor microenvironment. Myeloid-derived dendritic cells are often referred to as plastic or nonclassical monocytes , rather than bona fide dendritic cells, and are thought to acquire dendritic cell–like attributes once they enter the tumor environment. Myeloid-derived dendritic cells are associated with a negative prognostic outcome. Monocyte-derived dendritic cells arise from the differentiation of CCR2-expressing monocytes in response to inflammation, such as what would be found within the tumor microenvironment. Myeloid-derived dendritic cells promote immune suppression within the ovarian tumor microenvironment and are associated with poor outcome in other models of cancer.

cDC1 dendritic cells, although a rare subpopulation within the tumor microenvironment, express CD103 and are dependent on IRF8 and Batf3 signaling. Importantly, cDC1 are one of the most capable cell subsets that stimulate CD8 T cells for tumor control. In support of this, the abundance of cDC1 in tumors corresponds with positive cancer outcomes. cDC1 can also produce IL-12 to potentiate adaptive antitumor immune responses. In situ expansion of cDC1 within tumors using systemic administration of FLT3L and intratumoral poly I:C resulted in activation and enhanced response to BRAF inhibition in combination with programmed death ligand 1 (PD-L1) blockade. The response was so robust that mice were subsequently protected from tumor challenge, demonstrating the potency of cDC1 in achieving protective antitumor immunity. NK cells recruit cDC1 into the tumor microenvironment via CCL5 and XCL1 expression, identifying a key NK–dendritic cell–chemokine regulatory axis that is currently being explored as a potential cancer therapeutic. Mechanisms of tumor-induced dendritic cell dysfunction are summarized in Fig. 4.4 .

Dendritic cell polarization in the tumor microenvironment.

During early tumor progression, dendritic cells exert pressure against growing tumors through activation of cytotoxic CD8 T cells. As tumors progress, tumor-induced cytokines such as PGE2 and TGF-β upregulate the master regulator SATB1, downregulating MHC II and enhancing galectin 1 and IL-6 production from transformed dendritic cells. Dendritic cells undergo ER stress in response to accumulation of truncated oxidized lipids, diminishing processing and cross-presentation of antigen to CD8 T cells. Tumor-induced emergency myelopoiesis enhances accumulation of monocytes from the bone marrow, where some are differentiated into protumorigenic monocytic dendritic cells (MoDCs).

(Source: Figure generated in Biorender.)

Macrophages

In a tumor setting, macrophages are hijacked by tumors to promote malignant growth, invasion, metastasis, and outgrowth of tumors in distal organs. In general, the presence of macrophages associates with poor clinical outcomes for most cancer types. ^, Phenotypically, tumor-associated macrophages resemble M2-like macrophages, upregulating CD206 and acquiring a range of effector functions that enable the promotion of metastasis, angiogenesis, immune suppression, and tumor escape from therapy.

Aside from the surface expression of CD206, tumor-associated macrophages are functionally diverse and heterogenous in nature. Multiple signaling pathways have been identified that control macrophage polarization within the tumor microenvironment. One major mechanism involved in the polarization of macrophages is tumor hypoxia. Tumor hypoxia occurs because of aberrant growth of tumor cells and abnormal tumor vasculature, outpacing the limited oxygen and nutrient supply. Hypoxia forces tumors to adapt to new metabolic demands, initiating prosurvival, invasion, and inflammatory programs via upregulation of phosphoinositide 3-kinase (PI3K)/protein kinase b (Akt)/the mamalian target of rapamycin (mTOR), mitogen activated protein kinase (MAPK), and NF-κB signaling. Heterogeneity of monocyte-derived macrophages has been demonstrated using murine models and single-cell phenotypic analysis. For example, in a murine model of breast cancer, seven functionally and molecularly distinct macrophage subsets were identified based upon Ly6C and MHC II expression levels. The most suppressive macrophage populations resided adjacent to hypoxic tumor areas and had increased angiogenic and suppressive activities. Tumor-derived lactic acid was also demonstrated to have similar effects on the polarization of macrophages. Another contributor to macrophage heterogeneity is the nature of myeloid precursors that macrophages arise from. Macrophages that differentiate from M-MDSC are transcriptionally distinct and more immune suppressive compared with monocyte-derived tumor macrophages. M-MDSC-derived tumor macrophages are dependent on S100A9 activity within the tumor microenvironment, a signaling pathway that induces C/EBPβ, a transcription factor associated with potent immune suppressive activity of M-MDSCs. On the other hand, prostaglandin E2 ( PGE2), a product of COX-1 or 2 activity, enhances the suppressive function of committed M2 macrophages, ^, and inhibition of PGE2 using COX-2 inhibitors restores efficacy of PD-L1 blockade in certain tumor models.

Tumor-associated macrophages originating from tissue-resident macrophage pools complicate efforts to target tumor-associated macrophages. Tissue-resident macrophages self-renew locally, in the absence of input from the bone marrow. This establishes a scenario where tissue-resident macrophages can provide early and supportive signals to developing tumors to enhance growth, invasion, and avoidance of immune destruction. In non-small cell lung cancer, tissue-resident macrophages established a niche for developing tumors, facilitating epithelial to mesenchymal transition while also shielding developing tumors from adaptive immune pressure. For pancreatic cancer, CSF1-polarized monocyte-derived macrophages are more immune suppressive, and polarized tissue-resident macrophages have a profibrogenic phenotype.

Depending on the signals present within the tumor microenvironment, a multitude of additional transcriptional and regulatory pathways influence the polarization of tumor-associated macrophages. Signaling through the aryl hydrocarbon receptor (AhR) can polarize multiple immune cell subsets in the tumor microenvironment, including dendritic cells and macrophages. AhR resides in the cytoplasm, and when AhR ligands enter the cytoplasm via passive diffusion, the AhR receptor complex translocates into the nucleus, binding target DNA sequences located in the promoter regions of multiple genes relevant to inflammation and procancer activities. AhR ligands include kynurenine pathway metabolites and arachidonic acid metabolites, both of which have tolerogenic effects on immune cells within the tumor microenvironment. AhR signaling in macrophages and dendritic cells within the tumor microenvironment increases expression of IL-1β, IL-6, TNF-α, IL-10, and NF-κB. Hezaveh et al. added to this body of literature by demonstrating that indoles produced by certain Lactobacilli species from diet-derived tryptophan polarized tumor-associated macrophages overexpressing the AhR ligand in pancreatic cancer. The effects of the gut microbiome on cancer progression and therapy response are covered in Chapter 11 . These studies highlight the systemic effect of cancer and a cancer-affected gut microbiome on innate immune cells. Thus, potential therapies could involve dietary modifications such as consuming low-tryptophan diets. Indeed, for pancreatic cancer, provision of a diet low in tryptophan reversed macrophage-mediated immune suppression.

Transcription factors associated with macrophage polarization into suppressive proangiogenic phenotypes include, but are not limited to, jumonji domain-containing protein 3 (JMJD3) and IRF4 ^, ; STAT1, 3, and 6 signaling ; and C/EBPβ. Due to the underlying phenotypic plasticity of macrophages, protumor transcriptional changes can be triggered in response to multiple inflammatory cytokines and chemokines, PAMPs and DAMPs, metabolic changes, and changes in the vasculature and in the surrounding stroma. Other immune cells present within the tumor microenvironment can also polarize macrophages. For example, B-cell production of IL-10 leads to subsequent activation of TRIF/STAT1 signaling in macrophages, resulting in decreased proinflammatory functions; production of IL-10, IFN-β, CCL2; and upregulation of M2-like markers Ym1 and Fizz1. Other studies have found that autoantibodies polarized macrophages by binding Fc receptors on macrophages. Regulatory T cells and Th2 CD4 T cells are also able to polarize macrophages, through the production of IL-4, IL-13, and IL-10, as detailed in Gabrilovich et al. Several reviews cover signaling pathways associated with macrophage polarization extensively. ^,

Mast cells

Mast cells are potent innate immune cells whose role in inflammation extends beyond allergy and hypersensitivity reactions. As described previously, mast cells possess multiple effector functions that enable modulation of the tumor microenvironment. The effect of mast cells on cancer is variable and dependent on the types of cancer. For ovarian cancer and esophageal cancer, mast cells have been shown to be associated with a positive prognosis, whereas for melanoma, breast cancer, lung cancer, and gastric cancer, associations are either mixed or correspond with poor outcomes. Using a humanized model of melanoma, Somasundaram et al. demonstrated that mast cells are associated with immune suppression and failure of PD-1 blockade.

Mast cells are potent immune effectors and are also highly sensitive to microenvironmental cues present in the tissues in which they are residing. As such, mast cells are functionally and phenotypically modulated by tumor-associated signals while having a reciprocal effect upon the inflammatory milieu of the tumor microenvironment. Mast cells can rearrange the tumor environment through release of metalloproteinases, serine proteases, and other factors that enhance angiogenesis and remodeling of the basement membrane. Cytokines such as IL-6, CCL2, and VEGF enhance angiogenesis and recruit cells into the tumor environment, whereas profibrogenic mediators such as PDGF-β activate fibroblasts. Mast cell–produced IL-6 and CCL2 can also enhance cell proliferation and polarization of other myeloid subsets. The chemokine IL-8 (CXCL1, a proposed homologue in mice) and histamine can chemoattract and induce proliferation of tumor cells. Mast cells can also produce IL-13 and IL-4. Typically associated with an allergic response, these two cytokines have been associated with a fibrotic response, immune suppression, and tumor growth. The alarmin IL-33, which is also produced by gastric tumor cells, has been shown to induce IL-13, CSF-2, CCL3 and IL-6 from mast cells.

Mast cells can accumulate in tumors and the tumor microenvironment through multiple chemotactic factors such as SCF, CCL2, VEGF, angiopoietin 1, IL-8, CXCL1, osteopontin, and CXCL10. Activation and degranulation are also affected by signals in the tumor environment. For example, PGE2 produced by macrophages, other myeloid cells, and tumor cells can activate mast cells to degranulate via signaling through the PGE2 receptor EP3. ^, On the other hand, stimulation of EP2 and EP4 by PGE2 reduces activation of mast cells. Mast cells are also activated and phenotypically modulated by recognition of PAMPs and DAMPs. TLR4-mediated recognition of LPS can induce GM-CSF, IL-8, and IL-10 release without degranulation, whereas recognition of peptidoglycan via TLR2 signaling can trigger degranulation.

Neutrophils

Neutrophils are often synonymous with antimicrobial defense; however, their recruitment into sterile tissues suggests that neutrophils also have a role in tissue defense or damage, a process with high relevance to cancer. In patients with cancer, cancer-associated neutrophils sediment in the low-density fraction of blood gradients, along with myeloid-derived suppressor cells, which are described later. Low-density neutrophils are derived from TGF-β signaling and have been shown to be highly immune suppressive, protumorigenic, and prometastatic. GM-CSF is also a potent chemokine produced by tumors that triggers emergency myelopoiesis and pathologic expansion of immature neutrophils from the bone marrow and other extramedullary sites. CXCR1 and CXCR2 signaling on neutrophils, the ligands of which are induced by TGF-β and produced by tumor cells, are also involved in neutrophil exit from the bone marrow and into tumor-affected tissue. The ligands for these two pathways are produced in abundance by transformed cells, initiating the release of neutrophils from the bone marrow and into the tumor microenvironment.

Signaling through CXCR2 can culminate in the release of NETs into the tumor microenvironment. Unlike their host-protective role during infectious disease, NETs in the tumor microenvironment protected tumors from T cell– and NK cell–mediated cytotoxicity. NETs also have tumor metastasis effects, with studies demonstrating that NETs enhance extravasation from the primary tumor site ^, and recruit tumors into the premetastatic niche to facilitate metastatic colonization. Thus, inhibitors limiting NET formation or neutrophil egress from the bone marrow have the potential to improve adaptive immune responses and reduce tumor growth and invasion. Several studies have identified multiple heterogenous neutrophil populations in cancer-bearing individuals, suggesting that host-intrinsic factors also affect neutrophil phenotype and abundance. Indeed, in a comprehensive outlook of neutrophils in cancer biology, Quail et al. outlined multiple factors, including but not limited to age, sex, circadian clocks, obesity, the microbiome, and smoking.