Neutrophil heterogeneity and fate in inflamed tissues: implications for the resolution of inflammation

Abstract

Neutrophils are polymorphonuclear leukocytes that play a central role in host defense against infection and tissue injury. They are rapidly recruited to the inflamed site and execute a variety of functions to clear invading pathogens and damaged cells. However, many of their defense mechanisms are capable of inflicting collateral tissue damage. Neutrophil-driven inflammation is a unifying mechanism underlying many common diseases. Efficient removal of neutrophils from inflammatory loci is critical for timely resolution of inflammation and return to homeostasis. Accumulating evidence challenges the classical view that neutrophils represent a homogeneous population and that halting neutrophil influx is sufficient to explain their rapid decline within inflamed loci during the resolution of protective inflammation. Hence, understanding the mechanisms that govern neutrophil functions and their removal from the inflammatory locus is critical for minimizing damage to the surrounding tissue and for return to homeostasis. In this review, we briefly address recent advances in characterizing neutrophil phenotypic and functional heterogeneity and the molecular mechanisms that determine the fate of neutrophils within inflammatory loci and the outcome of the inflammatory response. We also discuss how these mechanisms may be harnessed as potential therapeutic targets to facilitate resolution of inflammation.

INTRODUCTION

Acute inflammation is a localized, self-limited innate host defense mechanism against invading pathogens and tissue injury. Typically, polymorphonuclear neutrophil granulocytes (PMN) are among the first cells to be recruited from the blood to the site of infection or injury. Neutrophils possess an impressive array of mechanisms that contribute to elimination of offending agents and necrotic tissues (93, 199). Excessive or dysregulated neutrophil recruitment and activation are also capable of inflicting tissue damage and prolong inflammation (93, 134, 197). Preclinical data indicate that impaired neutrophil removal from inflamed tissues results in aggravation and prolongation of the inflammatory responses (197). Neutrophil-driven inflammation is a common mechanism underlying many pathological conditions, including cardiovascular, acute respiratory, neurodegenerative, metabolic, and autoimmune diseases, sepsis, asthma, and cancer (136, 159). Hence, efficient inactivation and removal of emigrated neutrophils is critical for timely resolution of inflammation and return to homeostasis.

The molecular events governing the multiple-step process of neutrophil trafficking into tissues have been well defined (136, 157, 205). However, the fates of emigrated neutrophils have received considerably less attention. This is likely based on the long-held view that halting influx of additional neutrophils and clearance of neutrophils that fulfilled their immediate mission in the tissues can explain their rapid decline in inflamed areas and return to homeostasis. This review will focus on recent advances in molecular events that determine the fate of neutrophils that have emigrated into the inflamed and injured tissue and discuss how these mechanisms may be harnessed as potential therapeutic targets of the future.

NEUTROPHIL HETEROGENEITY

Phenotypical and functional heterogeneity is well known for lymphocyte and macrophage populations and assures diverse functions in homeostasis, inflammation, or immunity (90, 311). In contrast, neutrophils are traditionally considered as a homogeneous population of terminally differentiated cells with highly conserved function. Human blood neutrophils have banded or segmented nuclei, contain heterogeneous cytoplasmic granules, and express Mac-1(CD11b), CD15, CD16/Fcγ receptor III, CD33, and CD66b, which along with lack of CD14 expression are widely used to distinguish PMN from eosinophils and monocytes. However, emerging evidence challenges the classical view of PMN homogeneity. In healthy subjects, subsets of neutrophils expressing the glycoprotein CD177 (18, 112), olfactomedin-4 (45, 162), or variable T cell receptor-like immune receptors (84, 230) have been identified. CD177 is required for surface presentation of proteinase 3 (18), a serine protease normally located in the primary granules, which mediates transendothelial migration of CD177⁺neutrophils (144). Proteinase 3 is a major antigen in anti-neutrophil cytoplasmic antibody (ANCA)-associated systemic vasculitis (275). Patients with ANCA-associated vasculitis have higher number of CD177/proteinase 3-positive neutrophils (113) that may also predict relapse (232). The functional properties of this neutrophil subset are controversial (113, 260), and genetic deletion of CD177 did not affect neutrophil migration in mice (300). Hence, defining the roles of CD177⁺ and CD177⁻ subsets will require further investigations. Olfactomedin-4, expressed in only 20–25% of human peripheral blood neutrophils (45), does not affect PMN influx into tissues (294), whereas it enhances bactericidal capacity against Staphylococcus aureus in mice (162). A small subset (3–5%) of human neutrophils expresses variable T cell receptor-like immune receptors (230), with declining repertoire diversity in the elderly (84). The function of these receptors is still unknown.

Bone marrow PMNs are derived from the granulocyte-macrophage progenitor, leading to generation of a committed neutrophil precursor (preNeu), which differentiates into nonproliferating immature and mature PMNs (72). These subsets possess distinct proliferative capacity and genetic and functional signatures (72). PreNeus expand in the bone marrow and spleen during inflammation, indicating increased intra- and extramedullary granulopoiesis (72). Unlike preNeus, immature neutrophils could enter the bloodstream and traffic into the inflamed site as efficiently as mature PMNs (72). The implications of mobilization of immature neutrophils to host defense or contribution to PMN functional heterogeneity are currently unclear.

During the acute-phase response, bone marrow neutrophils are mobilized to the blood and consequently to highly vascularized organs (136), which are governed by the cytokine G-CSF, the chemokines CXCL1 and CXCL2 (63, 252), and lactate released from inflammatory bone marrow neutrophils (132). Of note, CXCL1 and CXCL2 are also involved in enhancing different modes of neutrophil transendothelial migration (91). Chemokines acting via CXCR4 may control neutrophil return to bone marrow following senescence (174). CXCR4 expression is almost undetectable on circulating human neutrophils, whereas it is upregulated following ex vivo culture parallel with enhanced migratory capacity toward CXCL12 (194), implying homing to the bone marrow. In mice, upregulation of CXCR4 and CD11b expression and reduction of CD62L expression on circulating PMN defines a neutrophil subset named “senescent” or “aged” neutrophils, which are controlled by circadian oscillations in the number of hematopoietic stem progenitor cells (35), the microbiome (308), and a cell-autologous molecular program consisting of clock-related genes, such as Bmal1, and the CXCR2 signaling pathway (3). The population of “aged” neutrophils (CD62L^low) peaks during daylight, whereas nonaged PMNs (CD62L^high) are most prevalent in the circulation at night (3, 35). “Aging” has been suggested to favor PMN clearance (3), whereas other studies reported rapid recruitment of “aged” PMN to sites of inflammation (282). Circadian changes in CD62L and CXCR2 expression were also detected in human PMN, suggesting conservation of this phenomenon across species (3) and endorsing time-based therapeutic interventions for optimizing clinical benefit (236).

Several studies have identified intrinsic functional heterogeneity in human circulating neutrophil pool under physiological conditions (226). For example, preference of some neutrophils over others to phagocytose bacteria, termed as competitive phagocytosis, has been reported (109). However, analysis of common PMN markers failed to identify the phenotype associated with phagocytic function, highlighting the challenges in linking functional responses to phenotype.

In the inflammatory milieu, neutrophils sense and respond to cues from both pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) (22). Neutrophils integrate these signals, which in turn alter their lifespan and capacity to undergo phenotypic and functional changes, and metabolic and transcriptional reprogramming, hence, setting the foundation for PMN heterogeneity under pathological conditions (21, 143, 260). Not surprisingly, a number of PMN subsets have been reported. For instance, LPS administration to healthy volunteers led to increases in the number of circulating banded CD16^dim neutrophils (as opposed to mature fully competent segmented CD16^bright neutrophils) with impaired antimicrobial function (225) and mature hypersegmented neutrophils that can suppress T cell responses and proliferation, resembling the actions of myeloid-derived suppressor cells (224). A CD63⁺ neutrophil subset was isolated from the sputum of patients with cystic fibrosis (115, 277). These neutrophils exhibit increased neutrophil elastase and arginase 1 activity and suppress T cell function, thereby likely contributing to perpetuation and progression of the disease. Other studies have shown that IL-13⁺ neutrophils polarize macrophages toward M2 phenotype that accelerate clearance of nematodes in mice (37). A portion of circulating neutrophils express IL-17 upon stimulation with IL-6 and IL-23 both in humans and mice, parallel with increased capacity for ROS production and fungal killing (273).

Human neutrophils exposed to PGE₂ or PGD₂ induce switching from LTB₄ synthesis to production of lipoxin A₄, a potent stop signal for neutrophil trafficking (155). This lipid class-switching may define a neutrophil phenotype in the resolution phase of inflammation, for lipoxin A₄ also facilitates macrophage efferocytosis (92, 253).

Another PMN subpopulation commonly referred to as low-density neutrophils may exert immunosuppressive function (also known as granulocytic myeloid-derived suppressor cells or G-MDSC) or display proinflammatory properties (34). Low-density neutrophils have been detected in healthy pregnant women (137, 265) as well as in patients with cancer (193, 240), sepsis (118), diabetes (298), and autoimmune diseases, such as systemic lupus erythematosus and psoriasis (34, 99). Proinflammatory subsets of low-density neutrophils produce IFNα upon stimulation, display reduced phagocytic activity, and are susceptible to release neutrophil extracellular traps (NET), thereby causing injury to endothelial cells (53, 285). These actions are consistent with aggravation and progression of the clinical symptoms in lupus. Recently, CD10 was suggested as a marker to discriminating immune-suppressive mature (defined as CD66b⁺, CD10⁺) and immune-stimulatory immature (CD66b⁺, CD10⁻) neutrophils present within heterogeneous PMN populations in patients with acute or chronic inflammatory conditions (173). Elevated CD10 and CD66b expression on oral PMN from patients with chronic periodontitis identifies a proinflammatory neutrophil subset distinct from naïve circulatory neutrophils and parainflammatory PMNs, which are present in the healthy oral cavity and interact with commensal biofilms without inducing clinically evident inflammation (75).

Studies on tumor-infiltrating PMNs in mice revealed distinct subsets of neutrophils that play opposing immune-modulator roles (80, 226). A PMN population, referred to as N1, is characterized by hypersegmented nuclei, production of proinflammatory cytokines, and potent tumor killing capacity (80). By contrast, the N2 subset displays an immature phenotype and increased arginase activity and favors tumor growth (80). Immature PMNs, defined as CD101⁻ cells (as opposed to CD101⁺ mature PMNs), were found to accumulate at the tumor site and correlate with cancer progression (72). It is still unclear whether the N1 and N2 populations share a common origin or are originated from naïve mature neutrophils and G-MDSC, respectively (81, 240). Interestingly, the anti-inflammatory cytokine transforming growth factor-β (298) and IFNβ (6, 116) have been implied in polarizing tumor-associated neutrophils toward an anti-angiogenic and anti-tumorigenic phenotype. Other studies have shown that expression of the proto-oncogene Met is essential for recruitment of anti-tumor neutrophils (76). A unique subset of tumor-associated PMNs has been identified in early-stage (small size) human lung cancer (261). This subset possesses characteristics of both neutrophils and antigen-presenting cells and could trigger an anti-tumor T cell response. Intriguingly, these cells were completely absent in larger tumors (261), indicating the phenotypical and functional divergence of tumor-associated PMNs during human lung cancer progression (241).

MULTIDIRECTIONAL MIGRATION

Neutrophils are usually the first cells to accumulate in injured and microbe-infected tissues (93, 199). Neutrophils egress from the blood to tissues is a multistep process involving PMN activation, rolling on, and adhesion to the activated endothelium, followed by transendothelial migration and chemotactic migration toward the inflammatory locus. The molecular mechanisms underlying and governing this process have been described in detail (136, 157, 205). The acute inflammatory response is protective and self-limited and resolves on its own under healthy conditions. Proteins such as annexin A1 (218, 262) and galectin-1 (150, 301) and proresolving lipids, including lipoxins, resolvins, protectins, and maresins, are produced during the resolution phase of self-limited inflammation (155, 253, 254) and function as potent stop signals for PMN activation and trafficking into tissues (218, 253, 257). Blocking PMN recruitment is an important event in terminating the inflammatory reaction for uncontrolled, excessive neutrophil accumulation and can perpetuate tissue damage that aggravates the initial protective inflammatory response and may evolve into chronic inflammation. Figure 1 summarizes neutrophil mobilization into sites of infection or tissue injury and PMN functions in host defense.

Fig. 1.

Neutrophil kinetics and fate of neutrophils in inflamed tissues. Neutrophils are rapidly recruited from the blood to infected or injured tissues. Neutrophil trafficking is a tightly regulated multistep process. Mobilization of annexin A1 to the neutrophil surface results in detachment of adherent polymorphonuclear neutrophil granulocytes (PMNs), thereby limiting their recruitment. Following extravasation, PMNs swarm toward and form clusters around the affected site. Neutrophils may neutralize and clear bacteria and damaged cells through phagocytosis, release of net extracellular traps (NETs), or necroptosis. Phagocytosis induces apoptosis in PMNs, followed by removal of apoptotic cells by macrophages through efferocytosis. Efferocytosis induces polarization of macrophages toward proresolution phenotype. NETs immobilize and kill bacteria in the extracellular milieu and may also present autoantigens to trigger adaptive immunity. Under certain conditions, PMNs can undergo necroptosis (programmed necrosis) parallel with or independent from NET formation. Necrotic debris may then be cleared by macrophages. Although several molecular switches that govern neutrophil responses to bacteria have been identified, the signals that direct the choice between phagocytosis, NET formation, and necroptosis are still poorly characterized. Emigrated PMN may reenter the circulation through reverse transendothelial migration (TEM) or exit from inflamed tissues through the lymphatic vessels. These “recirculated” PMNs may carry and present antigens to T and B cells in the lymph nodes or return to bone marrow for degradation. CRAMP, cathelin-related antimicrobial peptide; DAMPs, damage-associated molecular patterns; GSDM D, gasdermin D; JAM-C, junctional adhesion molecule-C; LTB₄, leukotriene B₄; LXA₄, lipoxin A₄; NE, neutrophil elastase; PAMPs, pathogen-associated molecular patterns; ROS, reactive oxygen species; RvD1, resolvin D1; RvE1, resolvin E1.

Neutrophil Detachment

According to the classical concept of neutrophil trafficking, adhesion of PMNs to endothelial cells is an essential step to direct them to sites of inflammation. Advances in imaging techniques have revealed that adherent neutrophils can detach and move away from sites of inflammation. Detachment of adherent PMNs is mediated by increased production and secretion of annexin A1 in response to glucocorticoids (171, 218), mast cell-stabilizing anti-allergic drugs (302), natural serine protease inhibitors (284), and the gaseous mediator hydrogen sulfide (27). Annexin A1 is mobilized from the cytoplasm pool to cell surface and signals through the lipoxin A₄/formyl-peptide receptor 2 (ALX/FPR2) (217, 218). Lipoxin A₄ evokes mobilization of annexin A1 (26) and acts in concert with annexin A1, thus forming an endogenous anti-inflammation loop to limit PMN trafficking into inflammatory loci (26). Interestingly, a small portion of neutrophils continued to pass the endothelial barrier following treatment with glucocorticoids (171), likely representing a functionally distinct subpopulation. The potential importance of annexin A1 controlling neutrophil adhesion is supported by the findings of enhanced neutrophil transmigration in annexin A1-deficient mice (36). However, the phenotype, function, and fate of detached neutrophils remain to be investigated.

Neutrophil Swarming

Following extravasation from the blood, PMN behavior can switch from mere chemotaxis toward the infected or injured site to a swarm-like migration pattern, referred to as neutrophil swarming (148, 202). Neutrophil swarming is characterized by sequential phases of highly coordinated interstitial chemotaxis and progressively accelerated exponential PMN accumulation, culminating in the formation of neutrophil clusters around the damage core to insulate the affected site from the surrounding healthy tissues (43, 133, 147, 148). Neutrophil swarms require LTB₄ released from PMNs (4, 148). The signal(s) for enhanced LTB₄ formation has not been elucidated, although contact of PMNs with necrotic cells is critical for initiation of swarming (281). It is possible that upon sensing necrotic tissue, pioneer neutrophils release ATP through connexin 43 channels, which in turn opens P2X1 channels, leading to elevations in intracellular calcium and subsequent activation of 5-lipoxygnease and LTB₄ production (228). Coordinated activation of LTB₄ synthesis would then build a stable LTB₄ gradient to drive concerted waves of migration. Consequent microscale arrays confirmed the dominant role of LTB₄ during human neutrophil swarming and also identified a large number of protein mediators that function to enhance neutrophil accumulation during swarming (235). Among these proteins are CXCL8, galectin-3, CXCL7, lipocalin-2, and pentraxin-3. The integrins Mac-1and LFA-1 are critical for PMN accumulation in the collagen-free injury center (148). Signaling through CXCR1, CXCR2, ALX/FPR2, and LTB₄ receptor BLT1 coordinates neutrophil congregation at damaged sites and maintains tight neutrophil association in the aggregation stage (49, 148). These findings identify a feedforward loop centered on LTB₄ in both the recruitment and the clustering phases of swarming. By cloaking the injured area (removing debris and sequestering DAMPs), tissue-resident macrophages can prevent swarming and consequently maintain tissue functionality (281). However, because cloaking did not interfere with swarm progression after its initiation and PMNs still appeared around cloaked areas, it is unlikely that macrophage inhibition of swarming was mediated by secretion of chemorepulsive signals (281). Moreover, development of human neutrophil swarms was associated with generation of the potent proresolving lipid mediators lipoxin A₄ and resolvin E3 parallel with slower growth and reduced size of swarms (235).

Neutrophil swarming has been observed in vitro around dying cells (169) and in animal models of sterile inflammation (148, 281), infection with a wide range of pathogens, including Staphylococcus aureus (124), Escherichia coli (140), Toxoplasma gondii (43), and Aspergillus fumigatus (140), and following influenza vaccination (227). Under these conditions, swarming was found to limit the extent of tissue damage and infection, implying an important role in host defense. Consistently, neutrophils isolated from trauma patients (235) or from patients receiving immunosuppressive therapy following transplantation (119) were shown to have a limited ability to swarm. Thus, impaired neutrophil swarming may contribute to increased susceptibility and inability to clear Staphylococcus infections following major trauma (154) and fungal infections with Aspergillus fumigatus conidia in patients during immunosuppressive therapy (119). Excessive or uncontrolled swarms may contribute to aggravation and propagation of tissue damage, as was shown in bacteria and ischemia-reperfusion-induced pulmonary inflammation in mice (140) and inflammation flares around uric acid crystals in gout disease (170).

Reverse Transendothelial Migration

Studies using intravital microscopy documented the ability of neutrophils to exhibit motility away from injured tissues back into the vascular lumen, known as reverse transendothelial migration (TEM) (46, 146, 299). This phenomenon was first demonstrated using in vitro models of neutrophil transmigration (30) and subsequently in zebrafish embryos (176) and in the mouse cremaster circulation (299). Neutrophil reverse TEM is most prevalent in tissues subjected to ischemia-reperfusion injury that is associated with reduced expression of junctional adhesion molecule-C (JAM-C) at endothelial cell junctions (246, 299). Pharmacological blockade or genetic deletion of JAM-C enhances the frequency of neutrophil reverse TEM in mouse cremaster venules (299). Under ischemia-reperfusion, excessive production of LTB₄ leads to activation of the LTB₄ receptor BLT1 on neutrophils and release of neutrophil elastase (46). Neutrophil elastase binds to Mac-1 (32), which is also a ligand for JAM-C and cleaves JAM-C (46). In vitro, reversely migrated neutrophils display a phenotype (ICAM-1^high, CXCR1^low) distinct from tissue-resident (ICAM-1^low, CXCR1^low) or circulating (ICAM-1^low, CXCR1^high) neutrophils and increased capacity to produce superoxide (30, 299). ICAM-1^high neutrophils were also detected in both the systemic and pulmonary circulation of mice subjected to either cremaster muscle or lower-limb ischemia-reperfusion injury (299). In the cremaster model, an association was found between percentage of ICAM-1^high neutrophils within the pulmonary vasculature and the severity of lung inflammation (299).

At present, the functional implication of this phenomenon is unclear. One possibility is that reverse TEM serves as a protective mechanism to enhance clearance of PMNs from infected or injured tissues, thereby contributing to dampening the inflammatory response (176, 206). Alternatively, neutrophil reverse TEM may be associated with dissemination of systemic inflammation and distant organ injury (46, 299).

Lymphatic Transit

Neutrophils may also exit from the inflamed tissue through the lymphatic vessels. Neutrophil accumulation in draining lymph nodes in mice has been detected following injection of a variety of bacteria, protozoan parasites, and viruses (1, 43, 102, 103, 148). Neutrophils have been observed in the subcapsular sinus, medullary region, interfollicular zone, and T cell zone, depending on the pathogen and the mode of PMN recruitment (103). Neutrophil localization within the lymph node affects its function. For instance, rapid localization of neutrophils to the subcapsular sinus can kill S. aureus and destroy infected cells (124), thereby limiting further pathogen spread. In the interfollicular zone, neutrophils are positioned in close proximity to T and B cells and innate-like lymphocytes such as NK cells (124, 127) and are required for optimal NK cell activity (117). Some studies have reported that PMNs can present antigen to T cells in vitro (271). However, it is not known whether lymph node neutrophils can present antigen to T cells in vivo.

Neutrophils may enter lymph nodes from the bloodstream through high endothelial venules or lymphatic vessels (25, 124). The surface molecules governing PMN transmigration across high endothelial venules are relatively well defined and include CD62L, CD62P, CD162, LFA1 (CD11a), Mac-1 (CD11b), ICAM-1, and CXCR4 (103). Studies with CXCR2-deficient mice showed reduced neutrophil influx to lymph nodes in a tumor lysis model (25), whereas CXCR2 was dispensable for Staphylococcus aureus-induced PMN accumulation in lymph nodes (102). The mechanisms governing neutrophil migration via lymphatic vessels are still being investigated. Neutrophil migration from inflamed skin to draining lymph nodes is guided by CXCR4, and its ligand CXCL12 expressed by lymphatic endothelial cells (102). These findings would imply CXCR4 as a potential therapeutic target to reduce neutrophil entry into lymphatic vessels. Other studies suggest that Mac-1 may also be required for neutrophil entry into lymphatics following intradermal injection of Mycobacterium bovis in mice (237). The counter-ligand(s) for Mac-1-dependent migration has not yet been identified.

Staphylococcus aureus-pulsed neutrophils that migrated to the sentinel lymph nodes exhibit a distinct phenotype (characterized as CD11b^high, CD62L^low, and CXCR2^low) and upregulation of major histocompatibility complex (MHC) II and the costimulatory molecules CD80 and CD86, implying their ability of antigen presentation and initiation of adaptive immunity (102). Moreover, PMNs from vaccinia virus-infected skin were reported to migrate to the bone marrow and enhance activity of virus-specific memory CD8⁺ (cytotoxic) T cells (60). Collectively, these observations would indicate that neutrophil-associated antigen transport could contribute to trigger adaptive immunity. By contrast, other studies have shown that neutrophils carrying Leishmania major (221) or Toxoplasma gondii (48) contribute to immune evasion and pathogen spread throughout the mouse intestine.

NEUTROPHIL LIFESPAN

Human circulating neutrophils have a short half-life (145, 270), although one study estimated their lifespan to be 5.4 days (223). In the inflammatory microenvironment, neutrophils receive opposing cues that modulate their lifespan and their mode of death. Cell death is a major mechanism to eliminate neutrophils from inflammatory loci for assuring timely resolution of inflammation. Neutrophils are classically thought to die via apoptosis (244). Apoptosis is a form of programmed cell death that involves cell shrinkage and chromatin condensation, followed by fragmentation of the entire cell into sealed apoptotic bodies. Execution of this constitutively expressed cell death program renders neutrophils unresponsive to inflammatory stimuli (135) and promotes their removal via efferocytosis (89, 138, 243) with minimal or no damage to the surrounding tissue. Although PMN apoptosis has been studied most extensively, neutrophil subsets undergo other modalities of caspase-independent regulated cell death, such as NETosis (82) and necroptosis (213). Necrotic cell death can be accidental or programmed (i.e., necroptosis) and is characterized by loss of plasma membrane integrity, release of intracellular contents into the extracellular space, and proinflammatory actions. Moreover, if apoptotic neutrophils are not cleared in a timely fashion, they may undergo secondary necrosis. Release of neutrophil constituents may contribute to killing bacteria (28) or to sustain collateral tissue damage and trigger autoimmunity (98, 263). Thus, the mechanisms activated by emigrated PMNs to contain pathogens or tissue damage through phagocytosis, necrosis, or NET formation would ultimately determine their own fate, the means of clearance of debris generated by apoptosis versus necrosis and ultimately the outcome of the inflammatory response.

Apoptosis and Phagocytosis-Induced Cell Death

Neutrophil lifespan is thought to be extended within inflammatory loci. PMN adhesion to endothelial cells and most inflammatory mediators and microbial constituents found in the inflammatory microenvironment generate prosurvival cues to prolong neutrophil longevity by delaying the constitutive apoptotic program (47, 122, 152). Activated neutrophils secrete protons that contribute to interstitial acidification, yielding pH between 5.5 and 7.0, in inflammatory loci (64, 88, 195). Extracellular acidosis is a danger signal (231), which through activation of multiple signaling pathways, including ERK 1/2, phosphoinositide 3-kinase (PI3K)/Akt, NF-κB, and cAMP/PKA leads to preventing degradation of Mcl-1, a key regulator of neutrophil survival (62), and subsequently suppresses PMN apoptosis and enhances the apoptosis-delaying actions of inflammatory mediators (65). Extracellular acidosis affects the outcome of inflammation in a context-dependent manner. For example, genetic deletion of the pH sensor TDAG8 aggravates LPS-induced acute lung injury in mice (279), whereas proton pump inhibitors protect mice from the deleterious actions of persisting acidosis in acute systemic inflammation (14). Acidosis is a characteristic feature of patients with sepsis, and more severe acidosis is associated with poor prognosis (203). Hypoxia in the inflammatory microenvironment and bacterial infections even under normoxia can induce release of HIF-1α (222) and HIF-2α (276), which activate prosurvival mechanisms in neutrophils, control their bactericidal activity, and restrict systemic spread of infection. Moreover, activated neutrophils release myeloperoxidase, which has been implicated in killing bacteria and tissue destruction (134, 198), binds to Mac-1 on human neutrophils (149), induces PMN activation and degranulation (68, 149), and delays intrinsic apoptosis (68). Thus, myeloperoxidase triggers a feedforward loop to amplify and perpetuate the inflammatory response (68). Accordingly, myeloperoxidase prolongs neutrophil-mediated acute lung injury in mice (68). Genetic deletion of myeloperoxidase in mice attenuates PMN accumulation and organ damage after renal ischemia-reperfusion (177) and reduces E. coli septicemia-induced lung injury and mortality (29). Conversely, aspirin-triggered 15-epi-LXA₄ interrupts the myeloperoxidase-centered circuit and redirects neutrophils to apoptosis (69).

Neutrophil apoptosis has emerged as one of the critical determinants of the outcome of the inflammatory response. Extended neutrophil lifespan through delayed apoptosis is observed in patients with sepsis (128), acute respiratory distress syndrome (178), severe asthma (280), or acute coronary artery disease (87) and is associated with disease progression and poor prognosis. Moreover, neutrophil heterogeneity is a common finding in the bronchial lavage of patients with chronic obstructive airway diseases, where the number of hypersegmented PMNs inversely correlates with lung function (163). In experimental models, delaying neutrophil apoptosis adversely affects the duration of inflammation (68, 120), whereas treatment with cyclin-dependent kinase inhibitors, 15-epi-LXA₄, or IFN-β counters prosurvival cues, facilitates neutrophil apoptosis and efferocytosis, and consequently enhances resolution of inflammation (69, 238, 242). Genetic deletion of the apoptosis-related effector protein in the TGFβ signaling pathway (ARTS/Sept4_i2) prolongs inflammation in mice (168), underscoring the importance of the intrinsic (mitochondrial) pathway of apoptosis in the resolution of inflammation.

Phagocytosis-Induced Cell Death

Typically, phagocytosis of opsonized bacteria or necrotic cells accelerates neutrophil apoptosis and facilitates the removal of apoptotic neutrophils via efferocytosis. Lateral clustering of Mac-1 [or complement receptor 3 (CR3)] initiates complement-mediated phagocytosis (79), which is governed by a delicate balance between Mac-1 and the complement C5a receptor (C5aR or CD88) (97, 187). Diminished expression of CXCR1 or Mac-1 on neutrophils results in reduced phagocytosis of bacteria (20, 105). Pharmacological blockade or genetic deletion of C5aR reduces phagocytosis of E. coli and P. aeruginosa by human and mouse neutrophils, respectively (187). The PAMP bacterial DNA and the danger-associated signal mitochondrial DNA, sharing the bacterial ancestry, reduces phagocytosis and phagocytosis-induced apoptosis in human neutrophils through neutrophil elastase and proteinase 3-mediated shedding of C5aR (Fig. 2) (250). Consistently, bacterial DNA delays clearance of E. coli, neutrophil apoptosis, and efferocytosis and prolongs E. coli-evoked lung injury in mice (250). TLR9-mediated cleavage of C5aR may represent a potential strategy evolved by bacteria to avoid complement-mediated destruction (146) and may also explain defects in neutrophil function, including impaired phagocytosis under pathological conditions characterized by release of bacterial DNA (233, 249) or mitochondrial DNA (295, 309). Aspirin triggered 15-epi-LXA₄ and 17-epi-RvD1, acting through the ALX/FPR2, effectively counter TLR9 signaling, prevent release of neutrophil elastase and proteinase 3, and restore the balance between Mac-1 and C5aR expression, phagocytosis, and phagocytosis-induced apoptosis in human PMNs (250). The therapeutic potential of these lipids is illustrated by the observations of accelerated bacterial clearance, enhanced neutrophil apoptosis, and efferocytosis in a mouse model of acute lung injury (250). Of note is that other proresolving lipids, such as resolvin E1, acting through the LTB₄ receptor BLT1 (66), and resolvin D5, which signals through GPR32 (39), facilitate phagocytosis by naïve PMNs via distinct receptors and molecular mechanisms.

Fig. 2.

Modulation of neutrophil phagocytosis and phagocytosis-induced cell death. Typically, Mac-1-mediated phagocytosis of opsonized bacteria or cell debris leads to polymorphonuclear neutrophil granulocytes (PMN) apoptosis by reactive oxygen species (ROS)-dependent activation of caspase-8. Ligation of Toll-like receptor 9 (TLR9) by bacterial DNA (CpG DNA) or mitochondrial DNA (mtDNA) released from necrotic cells evokes release of contents of azurophil granules. Neutrophil elastase (NE) and proteinase 3 (PR3) downregulate complement 5a receptor (C5aR) and by altering the Mac-1/C5aR ratio reduce phagocytosis, bacterial clearance, and phagocytosis-induced cell death (PICD). Myeloperoxidase (MPO) binds to Mac-1 and generates survival signals for PMNs, consistent with prolongation of inflammation. Specialized proresolving lipid mediators efficiently reverse these actions (right). Resolvin E1 (RvE1), acting through the leukotriene B₄ receptor (BLT1), enhances phagocytosis, whereas aspirin-triggered 15-epi-LXA₄ (lipoxin A₄) or 17-epi-RvD1 (resolvin D1), through lipoxin A₄/formyl-peptide receptor 2 (ALX/FPR2), inhibits TLR9-stimulated degranulation, restores Mac-1/C5aR ratio, phagocytosis, and bacterial killing, and counters MPO-generated survival signals, thereby facilitating resolution. Hck, hematopoietic cell kinase.

Receptors Integrating Life and Death Decisions

Outside-in signaling through β2-integrins.

The β2-integrin Mac-1 is best known for mediating neutrophil adherence to endothelial cells and the extracellular matrix (136, 157) and phagocytosis of complement C3b-opsonized bacteria (61) as well as mediating phosphatidylserine- and C3b-dependent efferocytosis by human macrophages (186). The β2-integrins function as bidirectional “signaling machines” (114) that also regulate the neutrophil lifespan. Engagement of Mac-1 with its endothelial counterreceptor ICAM-1, fibrinogen, immune complexes, platelets, or myeloperoxidase induces activation of the PI3K/Akt and MAPK/ERK signaling pathways, leading to Mcl-1 accumulation and suppression of the intrinsic apoptotic program (68, 182). Mac-1 ligation in conjunction with TNF-α or FAS ligand directs neutrophils toward apoptosis though induction of NADP oxidase/ROS-mediated activation of Lyn and SHIP [Scr-homology 2 (SH2)-containing inositol 5-phosphatase], which inactivates phosphoinositide 3-kinase (182). This mechanism may be relevant for limiting tissue accumulation of neutrophils in vivo. Phagocytosis of C3b-opsonized bacteria induces ROS-dependent activation of caspase-8, which overrides the ERK-mediated survival signal, resulting in apoptosis (66, 182). Caspase-8 forms a heterodimer with FLIP (FLICE-inhibitory protein), which suppresses RIPK3-dependent necroptosis (207, 208), thereby preventing release of neutrophil content.

Downregulation of Mac-1 expression and subsequent modulation of outside-in signaling by proresolving lipid and protein mediators (218, 253, 257) are critical for limiting further PMN recruitment, the lifespan of emigrated neutrophils and neutrophil-mediated tissue damage.

Opposing signaling through formyl peptide receptors.

Formyl peptide receptor (FPR) ligands are ubiquitous in inflammatory loci, are highly diverse functionally, and govern neutrophil trafficking, lifespan, and clearance (303). FPR1, the high-affinity receptor for N-formyl peptides released from bacteria and mitochondria of necrotic cells, is critical for neutrophil killing of bacteria (86, 161, 209), host-commensal interaction during dysbiosis (188), and guiding neutrophils to the necrotic center of focal injury in mice (40, 184).

The lipoxin A₄ receptor/formyl-peptide receptor 2 (ALX/FPR2) binds an unusually large number of structurally distinct host- and bacteria-derived agonists, which often evoke opposing biological activities (303). For instance, the acute-phase protein serum amyloid A (SAA) induces neutrophil activation and promotes inflammation (108), whereas LXA₄, RvD1 and annexin A1 inhibit neutrophil activation and adhesion to the endothelium, thereby limiting PMN trafficking into tissues (217, 218, 253, 257).

Within inflammatory loci, neutrophils receive and integrate opposing cues from various ALX/FPR2 ligands, and their fate would ultimately depend on the relative abundance of these mediators. For instance, SAA (67) and the antimicrobial peptide LL-37 (287) delay whereas annexin A1 enhances neutrophil apoptosis (220). LXA₄ and 15-epi-LXA₄, which do not affect PMN apoptosis (123), override the prosurvival signal from SAA, thereby redirecting neutrophils to apoptosis (67). By contrast, SAA produced within the lung opposes LXA₄ and contributes to persisting inflammation in patients with chronic obstructive pulmonary disease (COPD) (24). Annexin A1 and RvE1 enhance synthesis of LXA₄ (107, 217), forming a feedforward resolution circuit. Ligand selectivity may directly be related to the conformational landscape of FPRs that alters the activation of downstream signaling pathways (50, 74, 142). Thus, LXA₄ or annexin A1 stimulate ALX/FPR2 homodimerization and activation of the p38 MAPK-MAPKAPK-Hsp27 pathway to augment the production of the anti-inflammatory cytokine IL-10 (50). Annexin A1-derived peptide Ac2-26 facilitates FPR1-ALX/FPR2 heterodimer formation and activation of JNK and caspase-3, thereby propagating proapoptotic signaling (50, 74). SAA and LL-37 do not induce dimerization; rather, they signal through ALX/FPR2 isomer to activate the Akt and ERK pathways to prevent degradation of Mcl-1 and delay apoptosis (50, 67). Conceivably, abundance of proinflammatory ALX/FPR2 ligands would contribute to the initial phase of the inflammatory response, whereas increased production of proresolving ligands would counter the actions of proinflammatory mediators during the resolution phase.

Efferocytosis

Removal of apoptotic cells is an essential process for development and maintenance of the organism’s health. Efferocytosis is orchestrated by “find me” signals that direct macrophages and dendritic cells toward dying neutrophils and “eat me” signals expressed on the surface of the apoptotic cell, which facilitate their recognition and engulfment (57, 138, 190, 267). Uptake of apoptotic cells is a silent nonphlogistic process that leads to reprogramming of macrophages from the inflammatory phenotype to an anti-inflammatory phenotype (9, 258) and subsequently to a proresolving CD11b^low subset with minimal phagocytic activity (31, 242, 247). This latter subset is characterized by expression of 12/15-lipoxygenase and IFN-β-related gene signature. Inhibition of cyclin-dependent kinases 5 and 9 enhances neutrophil apoptosis and accelerates PMN clearance and macrophage reprogramming (111, 238). Phagocytosis of apoptotic neutrophils triggers metabolic switch in engulfing macrophages, leading to glycolysis, lactate release through SLC16A1, and generation of an anti-inflammatory microenvironment (191). Loss of phagocytic capacity in macrophages (when macrophages become satiated) is associated with suppression of glycolysis and increased oxidative phosphorylation (31). Mitochondrial ROS production through the uncoupling protein UCP2 (211), the fission-promoting protein dynamin-related protein (DRP1) (291), and apoptotic corpse-derived l-arginine converted by arginase 1 and ornithine decarboxylase to ornithine and putrescin within the macrophage has been implicated in maintaining efficient efferocytosis. Thus, controlling ROS burden from apoptotic corpses following neutrophil efferocytosis is likely maintained by mitochondrial fission, whereas satiation is associated with arginase 1 shutdown (247).

Apoptotic neutrophils express CCR5, CXCR4, and the atypical chemokine receptor ACKR2/D6 on their surface, which affect the outcome of inflammation (8, 104, 212). CCR5 binds inflammatory chemokines and accelerates their clearance (8), whereas ACKR2/D6 binds CCL5 and enhances reprogramming of resolution phase macrophages (13, 212). Satiated macrophages are a likely source of SPMs such as LXA₄, RvD1, and maresin 1 (164, 247). These lipids enhance efferocytosis, macrophage reprogramming, and/or departure from resolving sites (41, 51, 92, 247). Similar properties were reported for protein effectors of resolution, including protein S (164), annexin A1 (51), IL-10 (33), DEL-1 (139), and IFN-β (242). IFN-β seems to play a unique role in the resolution of inflammation, as it is produced by satiated macrophages and possesses multiple proresolving activities, including facilitation of bacterial clearance (125, 242), acceleration of neutrophil apoptosis (7, 242), enhancing efferocytosis, and macrophage reprogramming to the CD11b^low phenotype (242) (Fig. 3). Thus, efferocytosis of neutrophils and macrophage reprogramming are tightly controlled by proresolving effectors that are produced following the efferocytic trigger.

Fig. 3.

IFN-β regulation of macrophage-neutrophil interactions during resolution. Phagocytosis of bacteria usually induces apoptosis in neutrophils. Polymorphonuclear neutrophil granulocytes (PMN) and tissue resident cells secrete inflammatory mediators that attract blood-borne monocytes to the inflammation site, where they mature to phagocytic inflammatory macrophages that contribute to clearing bacteria and apoptotic PMN. Apoptotic PMN express “find me” signals to attract monocytes and macrophages and “eat me” signals that allow their recognition and engulfment via efferocytosis. Efferocytosis polarizes macrophages from an M1 inflammatory phenotype to an M2 reparative phenotype. Upon reaching efferocytic satiation, M2 macrophages convert to the proresolving Mres phenotype that possesses low phagocytic capacity. Satiated macrophages, and to lesser degree tissue resident macrophages, produce IFN-β, which in turn stimulate bacterial clearance by PMN, redirect neutrophils to apoptosis, and accelerate efferocytosis by resolution phase macrophages. IFN-β reprograms macrophages to the CD11b^low phenotype and facilitates their migration to lymphoid organs. In addition to its local actions, IFN-β likely limits excessive tissue repair, fibrosis, bone resorption, and myelopoiesis.

PMN efferocytosis not only governs the responses of the engulfing phagocyte but could also lead to new use of proteins stored in PMN granules. For example, uptake of apoptotic PMN was shown to assist macrophages in limiting the growth of Mycobacterium tuberculosis (272). Lactoferrin, an antibacterial protein stored in the specific granules, was found to accumulate in resolution phase macrophages and undergo fragmentation (165). The lactoferrin fragments released to the extracellular milieu stimulate macrophage reprogramming in vivo, and the formation of resolution-promoting aggregated extracellular traps by human PMN (165).

NET Formation

In addition to phagocytosis and degranulation (i.e., release of preformed antimicrobial peptides, such as myeloperoxidase, neutrophil elastase, or cathelicidin from specific granules), neutrophils can release extracellular traps (NET) to immobilize and kill invading pathogens in the extracellular milieu (16, 28) and to degrade cytokines and chemokines (245). NETs target many pathogens, including Staphylococcus aureus, Mycobacterium tuberculosis, and Aspergillus fumigatus (98), activated platelets and platelet-derived microparticles (172, 180, 286), monosodium urate crystals (245), coagulation factors (175), or microparticles (56, 239). Recent reviews detail the molecular pathways leading to NET formation as well as differences in its composition (121, 210, 263). In general, NETs compromise a DNA scaffold, citrullinated histones, and an array of granule proteins, which absorb pentraxin 3 and complement (70, 307). NET formation, commonly referred to as NETosis, which is considered a form of necrotic cell death (185). Reports also exist of complete release of nuclear DNA, resulting in formation of anuclear but still viable cells (304) and selective extrusion of mitochondrial DNA to form extracellular traps without affecting viability (305). Although differences in the mechanisms responsible for the release of nuclear versus mitochondrial DNA have been proposed (306), the implications of these processes in innate immunity and disease remain to be investigated. Pathways leading to NET formation differ in terms of dependence on ROS (58, 129, 210). Activation of the Raf-MEK-ERK and p38 MAPK pathways triggers NADPH oxidase-dependent release of NETs (100, 130), whereas HMGB1 released from activated platelets, leukocytes, or necrotic cells induces ROS-independent NET formation (269). Because HMGB1 effectively inhibits phagocytosis (15) and acidic environment, generated by H⁺ ion release during phagocytosis, impairs NET formation (179), these mediators may function as molecular switches to direct the most effective neutrophil response to an insult and represent alternative outcomes of PMN activation. Phagocytosis of bulky phosphatidylserine-exposing particles, such as apoptotic bodies or platelets, renders neutrophils unable to release NETs (172, 181). By contrast, smaller particles or tampering with phosphatidylserine-associated moieties or recognition triggers NETosis and leads to accumulation of apoptotic debris (172). Regardless of the origin of DNA, NETs are degraded directly by DNase 1 (101) released predominantly from dendritic cells (151) or by the cytosolic exonuclease TREX1 (DNase III) following engulfment by macrophages (73, 151). The antimicrobial protein LL-37 protects NET against degradation by bacterial nucleases (246) and is required for the uptake of NET by macrophages (151). Because many of the molecules externalized through NETosis are potent autoantigens, aberrant NET formation and/or impaired NET degradation is recognized as trigger of autoimmune diseases, such as lupus erythematosus and rheumatoid arthritis in susceptible individuals (98).

NET formation is an important mechanism to trap and contain invading pathogens and demarcate the infected areas (44, 83). However, NET formation is increasingly being recognized to contribute to a range of pathologies. For example, clinical studies have reported an association between augmented NET generation and the severity and mortality of pneumonia- or sepsis-associated ARDS (153, 158) and COVID-19-associated ARDS (17, 312). Reducing NET by intratracheal instillation of DNase I or partial deletion of protein deiminase 4 (PAD4^+/−) reduced the severity of lung injury and increased survival in a mouse model of severe bacterial pneumonia (153). Intriguingly, complete PAD4 deficiency and attenuated NET formation and lung injury increased bacterial burden, suggesting that bacterial infections may shift the balance of the beneficial and deleterious effects of NET in ARDS. NET formation is also controlled by signaling through ALX/FPR2, for mice with genetic deletion of Fpr2 (the mouse equivalent of human ALX/FPR2) produced excess NETs, and this was associated with more severe lung injury and increased mortality (153).

Necroptosis

Necrotic cell death leads to the release of large amounts of DAMPs and necrotic debris, which are potent inducers of inflammation (213). Recent studies on signaling pathways that are activated in dying neutrophils indicate that under certain conditions PMNs can undergo necroptosis or programmed necrosis, as opposed to necrosis, which occurs in a disorderly manner following cell injury (213). Ligation of TNFR1 in mouse neutrophils leads to activation of the receptor-interacting protein kinase 1 (RIPK1)-RIPK3-mixed lineage kinase domain-like protein (MLKL) signaling and translocation of MLKL1 to the inner leaflet of plasma membrane, leading to membrane permeabilization (297). By ubuquitinilating RIPK1 (293), X-linked IAP (XIAP) functions as a switch to direct neutrophils to either necroptosis or apoptosis (297). It should be noted that high concentrations of TNF-α were required to induce necroptosis in mouse neutrophils, and this phenomenon has not been demonstrated in human neutrophils.

Ligation of adhesion receptors, including CD11b, CD18, CD15, and CD44 in human neutrophils primed with GM-CSF, induces activation of the RIPK3-MLKL-p38 MAPK-PI3K pathway, leading to NADPH oxidase-mediated generation of ROS and subsequent necroptosis (19, 289). In line with an effector role for ROS, neutrophils from patients with chronic granulomatous disease (caused by a genetic defect in NADPH oxidase) do not undergo necroptosis following ligation of adhesion receptors (289). ROS has also been suggested to act upstream of RIPK1-RIPK3-MLKL, thereby triggering necroptosis upon exposure of neutrophils to monosodium urate crystals (54, 55). Consistently, pharmacological blockade of RIPK1 with necrostatin-1 or genetic deletion of RIPK3 reduced neutrophil necroptosis in a mouse model of gouty arthritis (54). In the course of necroptosis, NET formation was also detected in this model (54, 55). However, this could be explained by necrosis-related passive release of chromatin (290), for NET release from live neutrophils occurs independent of RIPK3 and MLKL (5).

Phagocytosis of methicillin-resistant Staphylococcus aureus can also induce necroptosis in human neutrophils, as some ingested bacteria may survive within the phagosome and prevent phagocytosis-induced apoptosis (94, 95). This requires RIPK3 but is independent of RIPK1 and MLKL (94) and is associated with neutrophil serine protease-mediated IL-1β secretion (141, 278). The mechanism of S. aureus-evoked IL-1β production in neutrophils differs from that in macrophages, in which it is mediated by the canonical NLRP3 inflammasome and caspase-1 (278). Necroptosis may allow the escape of viable S. aureus from dead PMNs, liberation of intracellular neutrophil contents, and generation of IL-1β, which aggravate local tissue injury and lead to persistent infection (95). Activation of the RIPK3-MLKL pathway was detected in neutrophils in inflamed tissue samples from patients with neutrophilic diseases (289), indicating that neutrophil necroptosis occurs under in vivo inflammatory conditions, although the relevance of necroptosis to these pathologies remains elusive.

Dying cells release a myriad of DAMPs, including mitochondrial formyl peptides, purines, LTB₄, cytokines, and chemokines, and necrotic debris activates the complement cascade to generate the chemoattractant C3a and C5a, which collectively serve as necrotic “find me” signals (38, 296). These signals guide recruitment of phagocytes to the vicinity of necrotic targets (184, 296). Necrotic cells also express “eat me” ligands that facilitate their uptake. Some of these ligands partially overlap with equivalent apoptotic signals (e.g., exposure of phosphatidylserine and LTB₄) (296), whereas others, such as complement C1q deposited on necrotic cell membrane (85) and externalization of annexin A1 on the surface of necrotic cells (23), are unique to necrotic debris. Little is known about the ligands and receptors mediating the clearance of necrotic neutrophils. The importance of removing necrotic PMNs is illustrated by the detrimental effect of neutrophil antigens such as proteinase-3 to trigger autoimmunity, as reported in vasculitis (113, 232).

Therapeutic Approaches to Facilitate Neutrophil Clearance from Inflamed Tissues

Resolution of inflammation is an active process enabling the affected tissues to restore function (253, 257). The resolution phase of self-limiting inflammation is skewed toward proresolving mediators, as illustrated by changes in lipid profile with increases in generation of LXA₄ and resolvin D5 in a mouse model of bacterial peritonitis (39). Conversely, insufficient activation of resolution mechanisms contributes to excessive or prolonged inflammatory response to the initial insult (197, 253). For example, stable atherosclerotic plaques exhibit a controlled chronic inflammation phenotype, whereas defect in proresolution mechanisms precedes plaque rupture (77, 78). Ligation of TLR9 is associated with markedly reduced production of proresolving lipid mediators, such as 15-epi-LXA₄ and RvD1, which likely contributes to propagation of bacterial infections in mice (250). Thus, restoring dysregulated or defective endogenous resolution circuits is anticipated to promote resolution by correcting the imbalance between proinflammatory and proresolving mediators. Indeed, recent research identified many resolution mediators as promising agents to target neutrophils to achieve this goal. Table 1 summarizes results obtained in animal disease models and patients.

Table 1.

Potential therapeutic approaches to target neutrophils for promoting resolution of inflammation

Molecule	Receptor	Species	Disease/Model	Key Mechanisms, Actions on Neutrophils	Effects on Disease	Reference
Annexin A1	ALX/FPR2	Mouse	Mesenteric circulation	Detachment of adherent PMNs, ↓emigration		215
		Mouse	Dermatitis		↓Inflammation	160
		Mouse	Zymosan-induced dermatitis	↑monocyte recruitment, ↑efferocytosis	↓Inflammation	183
		Mouse	Blood-neutrophil senescence	↓PMN senescence, ↑migratory capacity	↑PMN clearance in bone marrow	52, 167
Peptide Ac 2-26	FPR1, ALX/FPR2	Rat	Peritonitis	↑macrophage accumulation, ↓PMN accumulation	↓Inflammation ↓Systemic inflammation	266
		Mouse	Intestinal ischemia-reperfusion-induced lung injury	↑IL-10, ↓PMN recruitment into the lung	↓Lung injury	96
		Mouse	Pleurisy	↓Mcl-1 expression, ↓NF-κB, ↓ERK 1/2, ↑caspase-3, ↑PMN apoptosis	↑Resolution	283
		Mouse	Skin allograft	↓Transmigration, ↑PMN apoptosis	↓Tissue damage, ↑graft survival	274
LXA4, 15-epi-LXA4	ALX/FPR2	Mouse	Asthma	↓Prostanoids, ↓IL-5, ↓IL-13, ↓cysteinyl LTs	↓Airway inflammation, ↓hypersensitivity	156
		Mouse	Ischemia-reperfusion injury-induced lung injury	↓BLT signaling, ↓5-LO signaling, ↓PMN accumulation	↓Lung injury	40
		Mouse	Peritonitis	↓PMN accumulation, ↑phagocytosis, ↑efferocytosis, ↑PMN migration to lymph nodes, ↑IL-10, ↑efferocytosis	↑Resolution	92, 248
		Mouse	Cystic fibrosis	↓PMN accumulation	↓Bacterial burden, ↓disease severity	126
		Mouse	ARDS, acute lung injury	Restores impaired phagocytosis, ↓NE and PR3 release, ↑PMN apoptosis, ↑efferocytosis	↑Bacterial clearance, ↑resolution	250
		Mouse	ARDS, acute lung injury	↓MPO signaling, ↓Mcl-1, ↑PMN apoptosis, ↑efferocytosis	↑Resolution, ↑survival	69
		Rabbit	Periodontitis	↓LTB4 signaling, ↓PMN accumulation	↓Inflammation, ↓bone loss	256
Resolvin D1	GPR32	Mouse	Peritonitis	↓PMN transendothelial migration, ↓PMN accumulation, ↑bacterial clearance, ↑IL-10, ↑phagocytosis, ↑efferocytosis	↓Inflammation	247, 268
		Mouse	Renal ischemia-reperfusion injury	↓TLR-mediated activation of macrophages, ↓PMN infiltration	↑Renal function, ↓fibrosis	59
17-epi- resolvin D1	ALX/FPR2	Mouse	ARDS, acute lung injury	Restores impaired phagocytosis, ↓NE and PR3 release, ↑bacterial killing, ↑PMN apoptosis, ↑efferocytosis	↑Bacterial clearance ↑Resolution	250
Resolvin D2		Mouse	Microbial sepsis, peritonitis	↓PMN accumulation, ↑bacterial clearance, ↑peritoneal macrophage	↓Inflammation, ↑survival	264
Resolvin D5	GPR32	Mouse	Bacterial peritonitis	↓PMN infiltration, ↑PMN phagocytosis, ↑bacterial clearance	↓Antibiotic need, ↑resolution, ↑survival	39
Resolvin E1	ChemR23	Mouse	Peritonitis	↓PMN trafficking, ↑efferocytosis, ↑lymphatic PMN, ↓IL-12, ↓NF-κB	↓Inflammation	10, 247, 248
		Mouse	Colitis	↓PMN recruitment	↓Inflammation, ↑survival	12
	BLT1, ChemR23	Mouse	Peritonitis	Partial BLT1 agonist/antagonist, ↓LTB4 signaling, ↓PMN trafficking, ↓IL-12, ↓TNF-α, ↓iNOS	↓Inflammation, ↓body weight loss, ↑survival	10, 11, 248
		Mouse	Bacterial pneumonia	↓PMN recruitment, ↑bacterial clearance, ↓IL-1β, ↓HMGB1	↓Inflammation, ↑survival	251
	BLT2	Mouse	Acute lung injury, peritonitis-associated lung injury	↓PMN accumulation, ↑phagocytosis, ↑PMN apoptosis, ↑efferocytosis	↓Tissue injury, ↑resolution ↑survival	66
		Rabbit	Periodontitis	↓PMN infiltration	↓Inflammation, ↓bone loss,	106
Maresin 1	?	Mouse	Peritonitis	↓PMN accumulation, ↑efferocytosis	↓inflammation	259
		Mouse	ARDS	↑Platelet-PMN interactions, ↓PMN trafficking, ↑efferocytosis,	Organ protection, ↓tissue hypoxia, ↓edema	2
Protectin D1	?	Mouse	Renal ischemia-reperfusion injury	↓PMN infiltration, ↓TLR-mediated macrophage activation	↑Renal function, ↓fibrosis	59
		Mouse	Peritonitis	↓PMN infiltration, ↑phagocytosis, ↑efferocytosis, ↑lymphatic PMN	↑Resolution	248, 255
Galectin 1		Mouse	Peritonitis	↓PMN adherence, ↑PMN clearance, ↑IFN-β, ↓PR3	↑Resolution	150, 301
Protein S/GAS6	TAM receptors	Mouse	Periodontitis	↓PMN accumulation,↑ bacterial clearance, ↑IL-10	↑Resolution	196
		Mouse	Acute lung injury	↑PMN clearance, ↑efferocytosis	↑Resolution	200
DEL-1	LFA1 (CD11a)	Mouse	Acute lung injury	↓PMN accumulation	↑Resolution	42
		Mouse	Periodontitis	↓PMN accumulation,↓TNF-α, ↓IL-17	↑Resolution ↓Bone loss	71
	Phosphatidylserine, αVβ3 integrin	Mouse	Peritonitis	↓PMN infiltration, ↑RvD1, ↑RvE1, ↑efferocytosis		139
Interferon-β	IFNα/βR1	Mouse	Bacterial pneumonia, peritonitis	↑PMN apoptosis, ↑bacterial clearance, ↑efferocytosis, ↓Mcl-1, ↑IL-10,↓IL-6, ↓TNF-α	↑Resolution	242
Elafin (PI3)	NE inhibitor	Human	Early phase ARDS	↓NE activity	↓Plasma NE level biomarker?	292
		Mouse	Pleurisy	Mobilization of annexin A1, ↓NE activity, ↑PMN apoptosis,↑M2 macrophages, ↑efferocytosis	↑Resolution	284
		Mouse	Gout	Similar actions as in pleurisy	↓Inflammation	284
PF-1355	MPO inhibitor	Mouse	Immune complex vasculitis	↓MPO activity, ↓PMN recruitment, ↓NET formation	↓Pulmonary vascular injury	310
		Mouse	Anti-GBM disease	↓MPO activity, ↓PMN recruitment, ↓NET formation	↑Renal function	310
Anti-TNF antibody	TNF neutralization	Human	Rheumatoid arthritis	↓NET formation ex vivo, ↓ expression of PMN auto-antigens	?	131
Agonistic FAS antibody	FAS	Human	Severe trauma	Stimulation of FAS, ↓Mcl-1, ↑PMN apoptosis ex vivo	?	214
R-roscovitine, NG 75	Inhibition of cyclin-dependent kinases	Mouse	Pleurisy, Aacute lung injury	↓PMN accumulation, ↑apoptosis, ↑efferocytosis	↑Resolution	238
		Human	Cystic fibrosis	Restores suppressed PMN apoptosis ex vivo	?	189
Hydrogen sulfide		Mouse	Endotoxemia in the mesenteric circulation	↑annexin A1 mobilization, ↓PMN trafficking	↓Inflammation	27
Nedocromil	Mast cell stabilization	Mouse	Mesenteric ischemia-reperfusion injury	↑annexin A1 mobilization, ↓PMN accumulation	↓Tissue injury	302
DNase I		Mouse	ARDS	↑NET degradation	↓Lung injury	153, 158
		Mouse	Systemic lupus erythematosus	↑NET degradation	↓Autoantibodies, ↓mortality	166
Dornase-α (synthetic DNase I)		Human	Respiratory failure after severe trauma	↑NET degradation	↓Respiratory failure	229

ARDS, acute respiratory distress syndrome; BLT1, leukotriene B₄ receptor 1; BLT2, leukotriene B₄ receptor 2; DEL-1, developmental endothelial locus-1; iNOS, inducible NO synthase; 5-LO, 5-lipoxygenase; LT, leukotriene; MPO, myeloperoxidase; NE, neutrophil elastase; NET, neutrophil extracellular trap; PI3, phosphoinositide 3; PMN, polymorphonuclear leukocytes; PR3, proteinase-3; TLR, toll-like receptor; ? indicates yet undefined receptor/effect.

Stopping PMN recruitment into inflamed sites and stimulating efferocytosis are common features of most proresolving lipid (253, 257) and protein mediators (150, 218). These mediators possess multipronged, somewhat overlapping functions but also express special cell-specific functions mediated through distinct receptors and mechanisms in the resolution phase. ALX/FPR2 binds multiple proresolving ligands, including annexin A1, annexin A1-derived peptide Ac2-26 (217), LXA₄ and its aspirin-triggered 15-epimeric form 15-epi-LXA₄ (253, 257), and 17-epi-RvD1 (257), whereas RvD1 and RvD5 signals through GPR32 (39) and RvE1 signals through ChemR23 (268) and BLT1 (66, 253). Ligation of these receptors was found to reduce tissue injury (10, 12, 35, 69, 96, 126, 250, 274), decrease bacterial burden (39, 126, 250), and enhance phagocytosis (39, 66, 92, 247, 248, 250, 251, 255, 256, 264), PMN apoptosis (69, 70, 250, 274, 283), and efferocytosis (2, 10, 69, 70, 92, 167, 183, 247, 248, 250, 255, 259, 268), contributing to preservation of organ functions (2, 59, 106, 156), acceleration of resolution (2, 10–12, 39, 40, 42, 52, 59, 69, 71, 92, 96, 106, 126, 131, 156, 160, 167, 183, 189, 196, 200, 214, 247, 248, 250, 251, 255, 256, 259, 264, 266, 268, 283, 292, 300, 310), and increased survival (39, 69, 248, 251, 264) under a variety of pathological conditions (Table 1). There is evidence that some of these mediators form feedforward amplifying loops, for example, LXA₄-annexin A1 (26) and RvE1-LXA₄ (107), to counter PMN recruitment and activation, thereby driving resolution. ALX/FPR2 agonists can be harnessed in conjunction with antibiotics (39) or antiviral drugs (192) to fortify the host response to infections (204, 219).

Pharmacological induction of PMN apoptosis with cyclin-dependent kinase inhibitors (238) or synthetic annexin A1-derived peptides such as peptide Ac 2-26 (274, 283) mimic the resolution enhancing actions of SPMs. The clinical potential of this approach is illustrated by the efficacy of an agonistic FAS antibody to counter suppressed apoptosis in PMN from patients with severe trauma ex vivo (214). The gaseous mediator H₂S (27) and the mast cell-stabilizing drug nedocromil (302) were shown to counter the deleterious actions of PMN through mobilization of annexin A1. Pharmacological blockade of the activity of PMN-derived enzymes, neutrophil elastase, or myeloperoxidase attenuated tissue injury and promoted resolution in murine pleurisy and immune complex vasculitis (284, 310). The neutrophil elastase inhibitor elafin efficiently reduced plasma elastase activity in patients with ARDS (292), suggesting the potential use of elafin as a biomarker of the disease. Reducing NET formation or accelerating its degradation with DNase I (153, 158), MPO inhibitor (310), and anti-TNF antibody (214) is under investigations to prevent propagation of organ damage and autoimmune responses (98). Of note is that the synthetic DNase I analog dornase-α is currently being tested in phase III clinical trials in patients with severe trauma-associated respiratory failure (229). Because presentation of severe COVID-19 resembles known NET-associated pathologies ARDS and microthrombosis (17, 312), targeting neutrophils or NET formation may represent promising approaches in the treatment for COVID-19.

Sex differences in the resolution of inflammation have been observed in mice (216, 288) and humans (234). Studies on skin blisters in healthy volunteers suggest accelerated resolution in females compared with males, presumably due to reduced PMN accumulation and elevated levels of D-series resolvins (234). Future studies are warranted to further explore differences in resolution mechanisms between sexes.

CONCLUDING REMARKS

Consideration of neutrophil functional heterogeneity and novel paradigms in neutrophil biology provides a different vantage point of their roles in host defense and disease. Clearance of neutrophils from infected or injured tissues is critical for efficient resolution and return to homeostasis. Impaired PMN removal has harmful consequences. Therefore, it is essential to further explore the mechanisms that determine the fate of emigrated neutrophils. These mechanisms govern phenotypical and functional heterogeneity of neutrophils, although linking phenotype to specific functions is challenging. Understanding the diversity in PMN functions is essential to develop neutrophil-specific therapies. The current anti-inflammatory drugs alleviate symptoms of inflammation (swelling, pain) but do not promote and may even delay resolution and tissue repair. An attractive alternative approach would be to stimulate egress of neutrophils from inflammatory loci. Results from preclinical models indicate that this can be accomplished by the application of SPMs such as lipoxins, resolvins, annexin A1, galectin 1, protein S, and IFN-β, which can prevent the deleterious effects of excessive or prolonged PMN-mediated inflammation. Although large-scale clinical trials with these compounds seem distant, initial results with targeting neutrophil-specific functions (e.g., inhibition of neutrophil elastase or degradation of NET) hold promise as disease-modifying intervention. Further studies are needed to investigate whether therapeutic interventions to facilitate removal of pathogenic neutrophils from inflamed tissues will lead to resolution of inflammation underlying many chronic diseases.

GRANTS

This work was supported by Canadian Institutes of Health Research Grants MOP-97742 and MOP-102619 (to J.G.F.), Israel Science Foundation Grant No. 678/13, the Rosetrees Trust, and the Wolfson Family Charitable Trust (to A.A.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.G.F. and A.A. prepared figures; J.G.F. and A.A. drafted manuscript; J.G.F. and A.A. edited and revised manuscript; J.G.F. and A.A. approved final version of manuscript.