The formation and function of the neutrophil phagosome

Summary Neutrophils are the most abundant circulating leukocyte and are crucial to the initial innate immune response to infection. One of their key pathogen‐eliminating mechanisms is phagocytosis, the process of particle engulfment into a vacuole‐like structure called the phagosome. The antimicrobial activity of the phagocytic process results from a collaboration of multiple systems and mechanisms within this organelle, where a complex interplay of ion fluxes, pH, reactive oxygen species, and antimicrobial proteins creates a dynamic antimicrobial environment. This complexity, combined with the difficulties of studying neutrophils ex vivo, has led to gaps in our knowledge of how the neutrophil phagosome optimizes pathogen killing. In particular, controversy has arisen regarding the relative contribution and integration of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase‐derived antimicrobial agents and granule‐delivered antimicrobial proteins. Clinical syndromes arising from dysfunction in these systems in humans allow useful insight into these mechanisms, but their redundancy and synergy add to the complexity. In this article, we review the current knowledge regarding the formation and function of the neutrophil phagosome, examine new insights into the phagosomal environment that have been permitted by technological advances in recent years, and discuss aspects of the phagocytic process that are still under debate.

The neutrophil phagosome is a distinctive organelle, formed from an invagination of the plasma membrane to completely enclose an engulfed particle. A host of complementary processes and pathways then transforms the phagosome environment from a largely inert cellular inclusion into one optimized for the degradation of ingested particles. 4 Within the phagosome, two major cytotoxic events take place: the production of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase-derived reactive oxygen species (ROS) and the delivery of microbicidal proteins from pre-formed granules. The potential of these mechanisms to wreak havoc intracellularly and extracellularly necessitates exhaustive and complex regulatory systems, as excessive or aberrant neutrophil activation has been implicated in tissue damage in multiple inflammatory and autoimmune diseases. 5 The complex mechanisms governing the phagocytic process in neutrophils have been a particular area of contention. Neutrophils are challenging to study, in part due to their short life span, abundance of degradative enzymes, and significant challenges in genetic manipulation. 6 As a result, there is a relative paucity of neutrophilspecific phagocytosis research. Many aspects of phagosome maturation have been studied in macrophages and then assumed to apply to neutrophils, despite disparities in phagocytic function (reviewed in [7]). One notable difference is that in most phagocytes (including macrophages), phagosome maturation follows an endocytic maturation pathway, whereby the phagosome fuses with lysosomes to form a phagolysosome, 7 whereas in the neutrophil, the process of maturation is more rapid due to the presence of pre-formed granules situated within the cytoplasm. This, combined with the neutrophil's streamlined killing mechanisms, leads to a very different phagosome environment. Despite these differences, study of the phagosome in other cell types has been informative to a degree as certain parallels can be drawn.
In this article, we will review the processes and mechanisms which shape the environment of the neutrophil phagosome and how these aid its pathogen-killing abilities, including the formation of ROS by NADPH oxidase, the movement of ions across the phagosomal membrane, the pH of the formed phagosome, and the delivery of granule contents. We will highlight controversies in the complex field of neutrophil phagocytosis, where further investigation is paramount to understand their role in health and disease, which is often a fine balance between harm and help ( Figure 1).

| S TUDYING THE PHAG OSOME THROUG HOUT HIS TORY
Elie Metchnikoff , termed the father of cellular innate immunity, was the first to identify phagocytosis as a central host defense mechanism. He initially observed motile cells encircling an inserted thorn in starfish larvae, following which he realized the broader significance of phagocyte recruitment in host defense. 8 At around the same time, Paul Ehrlich identified complementary humoral adaptive immunity. Ehrlich's precise staining techniques led to the modern era of leukocyte biology; his neutral dyes were able to identify "epsilon granules" in neutrophils, which he called "cells with polymorphous nuclei". 8 In 1960, it was shown that, following engulfment of bacteria, neutrophils underwent granule fusion (degranulation) with the phagocytic vacuole, and that the contents of the granules were "consumed," indicating their usage against pathogens. 9 A substance named "phagocytin" was the first microbicidal granule element to be identified, 10 which was later shown to comprise a number of different cationic antimicrobial proteins. 11 Subsequently, the advancement of staining and cell fractionation techniques, and cell imaging modalities including electron microscopy established the granule subset classification we know today, of which azurophil, specific, and gelatinase are the predominant subtypes.
Research into neutrophil phagocytosis has been full of contradictions and controversies. Shortcomings have been in part due to the inherent limitations of all in vitro and ex vivo studies. As neutrophils do not proliferate in vitro, much research has been undertaken using freshly isolated human donor cells, meaning donor-dependent variation is unavoidable. The neutrophil-like cells differentiated from HL-60 and PLB-985 cell lines have been used as an alternative; however, these cells are imperfect models for phagocytosis as intracellular killing is much less efficient, potentially due to the lack of specific granules in HL-60 cells. 12 The recent use of swimming zebrafish larvae, which are small, transparent, and only utilize innate immunity, in combination with high-resolution live imaging has been revolutionary in following neutrophil responses to microorganisms in vivo and in real time. 14 Additionally, intra-vital two-photon microscopy has enabled real-time live imaging of neutrophils phagocytosing bacteria or viral prey in murine lymph nodes, with important implications for antigen presentation. 13,14 Other difficulties in phagocytosis research have arisen due to the transience of the phagosome, and an array of confounding factors present within this organelle; these include phagosomal membrane potential, osmotic strength, pH of the phagosome and cytosol, and the complex interplay of many different ion channels, enzymes, and chemical reactions. 15 However, recent technological advances have increased our knowledge of neutrophil phagocytosis: In particular, the use of automated fluorescence microscopy to observe phagosome formation and maturation 16,17 and genetargeting technology in murine neutrophils 18 have permitted more focused experiments.
Despite information from these new experimental techniques, several controversies remain, including the role of ROS in establishing the intra-phagosomal environment. The increased oxygen consumption seen during phagocytosis in neutrophils was originally thought to be due to mitochondrial respiration, until the production of ROS within the phagosome was discovered. However, the extent to which ROS are directly antimicrobial is still actively debated.
Another area of continued uncertainty is the intra-phagosomal pH; due to the difficulties of measuring pH in such a small volume, purported phagosomal pH values have varied hugely over the years, ranging from acidification to neutral to alkalinization.

| PHAGOC Y TOS IS S IG NALING
During phagocytosis, multiple signaling cascades are activated, which result in the rearrangement of the actin cytoskeleton into a nascent phagosome (i.e., the stage of the phagosome immediately following initial sealing), after which phagosome translocation toward the center of the cell occurs, alongside phagosomal maturation and pathogen killing. [19][20][21] F I G U R E 1 Phagosome formation and maturation. (A) Overview of the major steps in neutrophil phagosome formation following pathogen detection. Events 1-4 are depicted in further detail in panel B. (B) 1: An opsonized pathogen engages Fc receptors (FcγR) or complement receptors (e.g., CR3) to initiate phagocytosis. Both FcγR and CR3 can employ immunoreceptor signaling pathways: SH2-domainbearing proteins (e.g., Syk) associate with phosphorylated ITAM, signaling downstream through phosphatidylinositol 3-kinase (PI3K) and/ or phospholipase Cγ (PLCγ). CR3 also employs independent inside-out and outside-in integrin signaling pathways. 2: Phagocytic receptor signaling induces regulation of the actin cytoskeleton via Rac and/or Rho. Myosin motor control of actin rearrangement drives extending pseudopod protrusions from the plasma membrane to form the phagocytic cup around the pathogen. 3: Cytosolic specific/gelatinase granules deliver proteins to the membrane of the forming phagosome, for example, the membrane-bound subunits of NADPH oxidase, gp91 phox (NOX2), and p22 phox . Actin polymerization at the pseudopod tips facilitates membrane sealing to complete the phagocytic vacuole around the pathogen. 4: The formed pathogen-containing phagosome translocates toward the granule-rich centriole within the neutrophil cytosol. NADPH oxidase generates antimicrobial reactive oxygen species inside the phagosome. The negative charge generated by this process is compensated by an influx of protons. Cytosolic azurophil granules, containing cytotoxic proteins, for example, elastase, fuse with the phagosome membrane to deliver their contents to the lumen of the phagosome The magnitude of the neutrophil phagocytic response to pathogens is substantial, which is unsurprising given that a primary function of neutrophils is pathogen destruction. Recent work from our group demonstrated that although there were minimal changes in protein expression between neutrophils exposed to Staphylococcus aureus bioparticles for 15 minutes compared to untreated cells, approximately one third of the phospho-proteome was altered. 22 These datasets are publicly available in the PRIDE database (https:// www.ebi.ac.uk/pride/ reference PXD017092). This dramatic change in protein phosphorylation indicates that S. aureus encounter and phagocytosis result in a significant signaling stimulus for the neutrophil. Reactome database 23 pathways enriched in our dataset of phagocytosing neutrophils include Rho GTPase signaling, neutrophil degranulation, membrane trafficking and nuclear membrane breakdown, vesicle-mediated transport and phosphatidylinositol signaling, which are discussed in detail below. The use of unbiased techniques to understand dynamic signaling in neutrophils has been hampered by the release of degradative enzymes during cell lysis, though our group 22 and others [24][25][26] have demonstrated these techniques are becoming feasible with modern technologies. It is anticipated that further insights into complex signaling networks will be gained through the ongoing application of these methods.
Phagocytosis is initiated by the engagement of various receptors on the surface of neutrophils, often by endogenous opsonic ligands such as immunoglobulins or iC3b, a product of the complement system. Opsonins coating the pathogen are recognized by the neutrophil, stimulating phagocytosis, and they play an additional role by helping to overcome the repellent forces between the neutrophil and the negatively charged cell wall of many bacteria. 27 Microbial pathogen-associated molecular patterns (PAMPs) also bind to a variety of neutrophil receptors, such as dectin-1. 28,29 Thus, neutrophils can internalize both opsonized and non-opsonized particles. Here, we focus on opsonin-mediated phagocytosis signaling pathways activated by Fc gamma receptors (FcyR) and complement receptors, which exhibit distinct mechanisms.
Following receptor activation, pseudopods form from the plasma membrane to produce a cup-shaped enclosure of the target particle that is enabled by the rearrangement of the actin cytoskeleton. 30 Actin polymerizes at the leading edge and around the phagosomal cup, and this polymerization persists until the constriction and closure of the phagosome 31 : As the phagosome matures, the actin network disassembles to complete closure of the phagosome cup and allow the fusion of granules. 32,33 Originally, pseudopod extension to surround a micro-organism prior to engulfment was thought to be exclusive to FcyR-mediated phagocytosis as macrophage studies described "sinking" of complement-opsonized targets. 34,35 More recent live-cell imaging of phagocytosis in macrophages has contested this, with evidence of membrane ruffles and protrusions encapsulating targets in complement receptor 3 (CR3)-mediated phagocytosis. [36][37][38] However, slower "sinking" phagocytosis also occurs as an alternative mechanism requiring less involvement of membrane extensions. Moreover, in macrophages, the sinking phenomenon was shown to be CR3dependent but FcγR-independent, whereas both CR3 and FcγR appear necessary for the formation and closure of FcγR-mediated phagocytic cups. 39 This is likely to be the case in neutrophils also, as neutrophils with CR3 mutations displayed decreased ability to ingest IgG-opsonized targets. 40

| FcyR receptors
Neutrophils express FcγRI (CD64), FcγRIIA (CD32), and FcγRIIIB (CD16), 41 where the predominant Fc receptor subtypes are FcγRIIA and FcγRIIIB. 42 The class I and III receptors form multimeric complexes, while class II receptors exist as monomers with a unique phosphorylation motif and inhibitory action when engaged. 43 As circulating unbound IgGs are ubiquitous, phagocytes must be able to distinguish between these and IgG-associated with particles or immune complexes. Immunoglobulin receptors are therefore activated by clustering, mediated by the simultaneous engagement of multiple ligands, as opposed to ligand-induced conformational change. 44,45 Clustering induces the activation and recruitment of Src family kinases, which results in the phosphorylation of tyrosine residues in the immunoreceptor tyrosine activation motif (ITAM) within the FcyR signaling subunit.
Phosphorylation of ITAM generates docking sites for proteins bearing SH2 domains, including the tyrosine kinase, Syk. Subsequent phosphorylation of Syk leads to the recruitment of various signaling proteins to the activated FcyR complex. Highlighting the importance of Syk in FcyR-mediated phagocytosis, Jaumouillé et al found that in macrophages, Syk regulated FcyR responsiveness by increasing lateral receptor mobility and clustering through a reduction in actin polymerization. 46 Although this mechanism was not demonstrated directly, neutrophils from Syk-deficient mice displayed a similar reduced ability to ingest IgG-opsonized particles. 47,48 Phosphorylation of Syk leads to the recruitment of adaptor proteins to the activated FcyR complex, leading to the activation of lipid-modification enzymes, including phosphatidylinositol 3-kinase (PI3K) and phospholipase Cγ (PLCy). PI3K is responsible for the accumulation of phosphatidylinositol 3,4,5-trisphosphate (PtdIns (3,4,5) P3), derived from phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5) P 2 ). PtdIns(4,5)P 2 is also the substrate of PLCγ, and hydrolyzes into inositol1,4,5-triphosphate (Ins(1,4,5)P3) and diacylglycerol (DAG), which also act as messengers in the phagocytic signaling cascade. 29 Despite the actions of PI3K and PLCγ, levels of PtdIns(4,5)P 2 increase in the early stages of macrophage phagocytosis, accumulating at the site of particle engagement and at the pseudopod tips.
Levels abruptly decrease upon internalization, as actin disassembles to allow phagosome detachment, suggesting that accumulation of PtdIns(4,5)P2 is associated with initial actin recruitment, and its hydrolysis is associated with subsequent actin degradation and remodeling. 32,33 This is evidenced by the reduced actin disassembly during phagocytosis of IgG-opsonized latex beads when PtdIns(4,5)P2 hydrolysis was inhibited in macrophages. 33 The role of PtdIns(4,5)P2 in neutrophil phagosomes is less certain. Similar phosphoinositide dynamics in the phagosomal cup have been demonstrated between macrophages and the neutrophil-like HL-60 cell line, although this study did not differentiate between PtdIns(4,5)P2 and PtdIns (3,4,5) P3 accumulation. 49 Conversely, a study of human and murine neutrophils showed that although PtdIns(4,5)P2 was abundant in the plasma membrane, unlike macrophages it decreased rapidly during phagosome formation and was undetectable after sealing. 21 The synthesis of PtdIns(3)P on phagosomal membranes appears to be universal across phagocytosis, 50

| Complement receptors
Complement receptors are categorized into CR1 and CR2, formed by short consensus repeat (SCR) elements; CR3 and CR4, which belong to the β2 integrin family; and CRIg, which belongs to the immunoglobulin Ig-superfamily. 53 CR3 (also called CD11b/CD18 or Mac-1) is the most efficient phagocytic complement receptor. 54,55 CR3 receptors are αMβ2 integrins that are activated by outside-in (binding to extracellular ligands) and inside-out (intracellular protein binding that changes integrin conformation and thus affinity state) signaling (reviewed in 56,57 ). With regard to phagocytosis, the complement fragment C3bi, a potent serum opsonin, is the most important ligand; however, CR3 is promiscuous and can bind a range of other ligands, including extracellular matrix proteins, surface receptors, blood coagulation proteins, and microbial surface molecules.
Patients with leukocyte adhesion deficiency (LAD), a condition caused by genetic mutations in β2 integrins (and thus CR3), experience severe recurrent bacterial infections. 58 Neutrophils from patients with LAD type 1 are CR3-deficient and defective in phagocytosis when stimulated in vitro but appear to have normal unstimulated IgG-dependent phagocytosis. 40 This suggests that there are CR3-dependent and CR3-independent mechanisms of phagocytosis and that the recurrent infections in LAD patients may be associated with failure to increase phagocytosis in response to inflammatory stimuli. However, neutrophils from these patients also display significant migration defects so it is difficult to tease out the relative contribution of this impairment, compared with phagocytosis defects, to infection susceptibility in vivo.
Inside-out activation appears critical for CR3-mediated phagocytosis. Inside-out activation involves the transduction of an intracellular signal to the cytoplasmic and then the extracellular domain of an integrin, with signaling often initiated by G protein coupled receptor (GPCR) or Toll-like receptor (TLR) activation, or cytokine stimulation. 53 Once inside-out signaling has facilitated conformational changes, from low to high affinity, outside-in signaling occurs. 38 This always involves linkage of integrins to the actin cytoskeleton but, depending on the effector response required, recruits and activates different adaptor proteins (reviewed in 59 ).
As well as the distinct signaling pathways described above, CR3 also employs an immunoreceptor-like signaling mechanism through phosphorylation of ITAMs on receptor-associated transmembrane adaptors: Neutrophils lacking ITAM-containing adaptor proteins are partially defective in integrin-mediated phagocytosis. 60 Similar to FcyR signaling, protein adaptors provide docking sites for the SH2 domains of Syk kinases which are then activated, followed by downstream signaling events. This study also found that Src family kinase phosphorylation of the adaptor proteins was essential for the association of the SH2 Syk domains and phagocytosis. These signaling events are important for rapid target ingestion: In macrophages, deletion of Syk or ITAM adaptors accentuated the inefficient and slow sinking method of phagocytosis. 39 An alternative mechanism, where inefficient phagocytic cups were formed via the extension of membrane ruffles, was also inhibited by the deletion of Syk or ITAM adaptors, suggesting that Syk signaling induces ruffling, a phenomenon which can also be observed during FcyR-mediated phagocytosis. 39 The majority of early work on integrin signaling was performed in macrophages, initially describing FcγR-or CR3-mediated phagocytosis as two discrete mechanisms. 61 More recently, it has been suggested that CR3 and its downstream effectors are essential for phagocytosis of both opsonized and non-opsonized targets. 62

| GTPase and cytoskeleton dynamics
Following protein and lipid kinase activity, signaling cascades induce actin polymerization and localized membrane remodeling. Rho family small GTPases play a central role in actin dynamic regulation and can switch between an active (GTP-bound) and inactive (GDPbound) state. FcyR-mediated phagocytosis is thought to involve predominantly Rac1, Rac2, and Cdc42, whereas complement-mediated phagocytosis utilizes Rho. Rac and Cdc42 direct lamellipodial and filopodial membrane protrusions, respectively, whereas Rho induces the assembly of contractile actomyosin filaments. 63,64 However, in macrophages, RhoG involvement has also been identified in FcyRmediated phagocytosis and iC3b-opsonized particle uptake is greatly reduced in Rac1-and Rac2-deficient cells, 36 suggesting some crossover of these pathways. It is likely that this is also true for neutrophils but has yet to be established.
In neutrophils, FcyR-mediated phagocytosis was markedly impaired in Rac2-but not Rac1-deficient cells, 65 indicating a non-redundant role for the Rac2 isoform. A rare inhibitory Rac2 mutation has been described in a patient who presented with recurrent and severe bacterial infection; however, phagocytic capacity was not assessed directly and Rac2 is also implicated in granule translocation (section 7.4) and NADPH oxidase function. 66  to the phagosomal cup membrane of neutrophil-like PLB-985 cells at the beginning of phagocytosis was phosphatidylserinedependent, and that inhibition of this recruitment resulted in reduced ROS production. 67 In murine neutrophils, assembly and activation of NADPH oxidase following FcyR-mediated phagocytosis was completely dependent on Rac2 whereas following complement-mediated phagocytosis, there was redundancy between Rac1 and Rac2. Further differences were identified, where the oxidase rapidly (less than 6 seconds) accumulated on sealed phagosomes formed in response to complement-opsonized prey, but during slower (more than 10 seconds) phagocytosis of IgGopsonized targets the oxidase could assemble at the base of a forming phagosome. 68 In FcyR-mediated phagocytosis, fluorescence resonance energy transfer (FRET) stoichiometry of macrophages revealed that Cdc42 was activated early and localized to actin in the extending pseudopod, Rac1 was active across the phagocytic cup and during closure, and Rac2 was active predominantly during contractile activities and closure of the phagosome. 31 Similarly, in neutrophils, Rac1 was preferentially recruited to actin-rich pseudopods due to the negative charge generated by phospholipids whereas Rac2 localized to the intermediately-charged phagosomal membrane. 69 However, neutrophil and macrophage GTPase dynamics are not identical as genetic deletion or pharmacological inhibition of murine or human neutrophil Cdc42 does not impair phagocytosis, although bacterial killing was compromised. 70 These spatio-temporal phagocytic control mechanisms for the different GTPases suggest that differential regulation of these enzymes coordinates the localized membrane remodeling.
Intriguingly, the use of pharmacological depolymerization agents showed that an intact actin cytoskeleton is required to target uptake and Rac2 translocation to the site of particle attachment in neutrophil-like PLB-985 cells, but disruption of the actin cytoskeleton once phagocytosis had been initiated does not prevent translocation of Rac2 toward the phagosome. 71 Indeed, actin depolymerization following FcγR-crosslinking was shown to enhance the oxidative burst, 65 indicating that the timing of actin assembly and disassembly is important.
The formation of membrane protrusions and closure of the phagocytic cup are facilitated by myosin "motors," which can control actin assembly, crosslinking and rearrangement, and are thus important regulators of membrane deformation, protein localization and phagosome translocation during phagocytosis (reviewed in detail in 72 ). Myosin contractility is thought to be particularly important in FcyR-mediated phagocytosis in neutrophils, where inhibitors of myosin ATPase prevented particle uptake, 73 and mechanical models suggest there is a requirement for protrusive force. 30 Rho and Rac activation are thought to be downstream events that follow the phosphorylation of the guanine exchange factor (GEF) Vav by Syk. 36,74 Neutrophils from Vav knockout mice are significantly deficient in FcyR-and integrin-mediated phagocytosis. 65,75 Downstream of Rho and Rac, the Arp2/3 complex is responsible for actin polymerization. In murine neutrophils, Rac1 and Rac2 isoforms were shown to differentially regulate actin assembly using different pathways, where Rac2 mediated the majority of its effect on actin via the Arp2/3 complex. 76 During CR3-mediated phagocytosis, actin polymerization is propagated by two distinct mechanisms via RhoA. In one pathway, RhoA can activate Rho kinase, which phosphorylates and activates myosin II, leading to recruitment of Arp2/3; inhibition of Rho kinase or myosin II activity results in reduced Arp2/3 recruitment and actin cup assembly in murine macrophages. 77 A non-redundant role for Arp2/3 in neutrophils is less certain, however, as neutrophils isolated from a patient with a rarely described ARPC1B deficiency (lacking Arp2/3) did not demonstrate any defect in phagocytosis. 78 In a separate pathway, RhoA can recruit the actin nucleator mDia1, which is recruited to the phagocytic cup via the microtubule-associated protein CLIP-170. 79 Inhibition of mDia1 reduces actin polymerization and particle uptake. 80 Highlighting the importance of RhoA signaling, the complement product, C5a (an anaphylatoxin implicated in sepsis pathogenesis), inhibits RhoA activation in a PI3Kδ-dependent manner, which prevents actin polymerization and reduces neutrophil phagocytosis. 81 Importantly, neutrophils from critically ill patients exhibit a similar phenotype, that is, reduced phagocytosis and a failure to activate RhoA or polymerise actin, which may contribute to the increased risk of infection in this patient group.
Various actin-binding proteins have been shown to be important during neutrophil phagocytosis. The mammalian actin-binding protein 1 (mAbp1), an adaptor protein phosphorylated by Syk, has been implicated in complement-mediated phagocytosis. 82 Syk was found to be necessary for the translocation of mAbp1 to the site of engulfment during phagocytosis and down-regulation of mAbp1 led to a depletion of clustered β2 integrins in high-affinity conformation.
This high-affinity conformation is essential to generate the tensile strength required for phagocytosis and the absence of mAbp1 led to severe defects in β2 integrin-mediated phagocytosis. 82 Inhibition of the actin-binding protein, coronin-1, has also been found to arrest phagocytosis. Coronin-1 is recruited to the phagosomal cup early, alongside actin. This suggestion that it may regulate actin is supported by the presence of coronin-1 at the leading edge of migrating cells. 83 Coronin-1 co-localizes with filamins, the most potent actin crosslinkers, at the phagocytic cup, where it appears to be involved in non-opsonic phagocytosis in response to PAMPs. 84 In macrophage phagosomes, phosphorylated Syk and paxillin colocalize and recruit vinculin, which facilitates target internalization by anchoring F-actin. These exact phosphorylation sites have not been detected in neutrophils, but it is likely that a similar process occurs. 38 Overall, distinct signaling mechanisms characterize FcγR and CR3-initiated phagocytosis, which results in different, though overlapping, processes for particle uptake and phagosome formation. In vitro, it is often necessary to dissect the roles of various signaling molecules individually. These experiments have highlighted important redundant and non-redundant roles for selected proteins, which have also been identified clinically through mutations, albeit rarely. In practice, however, there is likely substantial crossover and synergy between the immunoglobulin and complement pathways, which optimizes phagocytosis in vivo.

| NADPH oxidase
The formation of NADPH oxidase gives rise to some of the key changes in the intra-phagosomal environment, starting with ROS production. In most cell types, ROS are produced as a by-product from a variety of processes, such as mitochondrial respiration, 85 but in neutrophils, the majority of ROS are actively generated by NADPH oxidase (NOX enzyme complexes). NOX enzymes are "professional" generators of ROS, as reducing molecular oxygen to form ROS is their sole enzymatic function. In leukocytes, NOX2 is the main catalytic subunit, which is highly expressed in neutrophils. 86 The enzyme NOX2 is dormant in the circulating quiescent neutrophil: Activation requires an initial priming step followed by full assembly of the NADPH oxidase complex from membrane-bound and regulatory cytosolic subunits. 86 At rest, the two membrane-bound subunits, gp91 phox (which is the main redox center, NOX2) and p22 phox , are embedded in specific granules. Upon phagocytosis, and before the phagocytic cup is even sealed, these subunits are delivered to the phagosomal membrane by granule exocytosis. 87 Translocation of additional regulatory cytosolic subunits to the membrane-bound gp91 phox /p22 phox heterodimer to form the active NADPH oxidase complex is tightly controlled via the activation of a series of kinases, including protein kinase A (PKA), phosphoinositide 3-kinase (PI3K), and mitogen-activated protein kinases (MAPKs). 50,88 The small GTPase, Rac, is also important for NADPH oxidase activation. Guanine nucleotide exchange factors (GEFs) convert Rac-GDP to the active Rac-GTP, which translocates to the phagosomal membrane and there binds gp91 phox and p67 phox . [87][88][89] The obligate binding partner, p22phox, ensures stability of the heterodimer and accommodates docking of the regulatory subunits. 89 NADPH oxidase activation is a dynamic process whereby a cascade of kinase signaling, calcium release, and small GTPase (predominantly Rac2) activation lead to subsequent NOX2 activation. 90,91 Activation of voltage-gated proton channels, chloride channels, and subsequent calcium fluxes is also utilized to regulate the enzyme complex. 87 The activation process is differentially regulated by various receptors, depending on the stimulus. 92 Robust activation of NOX2 is seen following FcγR and integrin receptor ligation. Other receptors, such as formyl peptide receptors, which detect bacterial cell wall products, or GPCRs, can activate the enzyme complex directly but to a lesser extent. Neutrophil priming is an important control mechanism, which prevents inappropriate triggering of ROS generation. Here, an initial priming stimulus, for example, ligand binding to TLRs or TNF exposure, generates a pre-activated neutrophil phenotype, which results in a substantially enhanced activation of the NADPH oxidase when the cell encounters a second stimulus. 93 The range of upstream signaling pathways indicates that the threshold for NADPH oxidase activation is high, and that the ROS response is heterogenous, dictated by which signal is received. This complexity ensures tight regulation, with the aim of avoiding prolonged inflammation and oxidative damage.

| The production of ROS in the phagosome
ROS are chemically reactive radical and non-radical derivatives of oxygen, the former containing at least one unpaired electron. NOX2 (gp91 phox ) is the main redox center, transferring two electrons provided by cytosolic NADPH via internal heme moieties to the phagosome, whereupon oxygen is reduced to superoxide (O 2 · − ). ROS react with many biomolecules, including DNA, proteins, lipids, and carbohydrates, to cause damage. The study of ROS has historically been focused on their capacity to cause cellular toxicity. In the 1930s, the rapid release of ROS was found to be associated with the formation of hydrogen peroxide (H 2 O 2 ), 94 which can generate potentially cytotoxic derivatives, such as the highly reactive hydroxyl radical.
These findings were reinforced by the early discovery of bacterial ROS-eliminating enzymes, such as superoxide dismutase. 95

| Reactive oxygen intermediates
The long-held consensus has been that the ROS produced by NOX2, derived from superoxide anions, worked directly to kill bacteria, 88,96,97 although this has now been contested (section 8). The initial question was which product was key to the microbial killing? The direct product of NOX2, superoxide anion (O 2 · − ), is thought not to be highly bactericidal as it is weakly reactive and unable to migrate far from the site of production. However, a few of its derivatives are more reactive and able to diffuse through pathogen membranes. 88

| The role of ROS in the phagosome
The complex nature of ROS production illustrates the dynamic environment of the phagosome. Early findings by Klebanoff suggested that the bactericidal mechanism of neutrophils was mediated through MPO-catalyzed iodination and chlorination. 110

| ME TABOLIS M DURING PHAGOC Y TOSIS
Phagocytosis is an active process: Protein phosphorylation and the generation of ROS require an energy source. Unlike monocytes, whose oxygen consumption, and thus phagocytic capacity, is substantially reduced by mitochondrial respiratory chain inhibitors, neutrophils derive most of their energy from glycolysis, even under aerobic conditions. 116,117 Early studies of metabolism in phagocytosing neutrophils have shown varied results: Some studies indicated increased glucose utilization and lactate production while others did not. [118][119][120][121] Variations in host species, phagocytic prey, or the availability of extracellular glucose may explain these experimental discrepancies.
The accepted dogma is that neutrophils rely on glucose-fueled   138 In stimulated neutrophils, the cytosolic level of Cl − is lower than the estimated concentration of 70 mM in the phagosome, suggesting active transport mechanisms enable Cl − accumulation in the phagosome. 139 Two Cl − channels (ClC), the cystic fibrosis transmembrane conductance regulator (CFTR) and ClC-3, have been identified on neutrophil phagosomes. 140,141 Neutrophils from Clcn3(−/−) mice display reduced phagocytosis and ROS generation, demonstrating its importance in neutrophil phagocytic function. 141 It has been suggested that ClC-3 is a Cl − /H + antiporter rather than an ion channel, meaning it could extrude H + against the electrochemical gradient. 142,143 This appears counter-productive to the accumulation of H + in the phagosome; however, ClC-3 may be partly responsible for H + leak out of the phagosome (section 6).
CFTR is a cAMP-activated chloride channel, which contributes approximately 50% of the total halide transport activity in neutrophils and is responsible for maintaining phagosomal HOCl levels. 144 CFTR is defective in patients with cystic fibrosis (CF) 145 ; neutrophils from CF patients with the typical ΔF508 homozygous mutation exhibit significantly reduced CFTR content in the phagosomal membrane 146 and have impaired HOCl production. 140 Although there are additional reasons why CF patients suffer recurrent infections, an impaired antimicrobial environment in the neutrophil phagosome due to defective Cl − transport may be an important contributing factor, and killing of Pseudomonas aeruginosa (a common pathogen in CF) by neutrophils has been shown to be chloride-dependent (reviewed in 147 ) ( Table 1).
The K + / Cl − cotransporter KCC3 is also postulated to be present on the neutrophil phagosome: KCC3-deficient murine neutrophils exhibited reduced NADPH oxidase activity due to disturbances in the recruitment and phosphorylation of its subunits. 162 Furthermore, KCC activity was increased synchronously with neutrophil activation and was required for ROS generation, supporting the theory that KCC3 contributes to microbial killing and potentially to charge compensation.
Calcium ion (Ca 2+ ) transport through the phagosomal membrane is more contentious; although it has been thought to regulate phagocytosis, 163 the source of Ca 2+ is still debated and the requirement of Ca 2+ release for phagosomal oxidase activity remains controversial.
This ion does, however, play an important role in regulating signaling and cytoskeletal dynamics during particle ingestion and granule fusion.

| PH OF PHAG OSOME AF TER ING E S TION
The idea of phagosomal acidification has been in existence for as long as the concept of phagocytosis itself: Metchnikoff originally hypothesized that acidification was the cause of death for ingested organisms. 164 However, the phagosomal pH of neutrophils is inherently different from other phagocytes as it acidifies more weakly, at least at early timepoints. 126,165 Over the years, there have been many contradictions regarding neutrophil phagosomal pH. Investigators initially reported an acidification, 164 but further studies suggested a biphasic pH change, with initial alkalinization followed by a modest acidification to pH ~6.5. 165,166 Subsequently, the intra-phagosomal pH was suggested to remain neutral. 126,127 Proponents of the biphasic pH change proposed that initial alkalinization was due to the consumption of H + by the formation of H 2 O 2 , and that this alkalinity was beneficial as it optimized bacterial killing by granule proteases. 165 This theory was supported by the sustained acidification seen when NAHPH oxidase is inhibited. Similarly, neutrophils from patients with chronic granulomatous disease (CGD) (which lack a functional NOX2) undergo a rapid and extreme acidification during phagocytosis, which is associated with impaired bactericidal function. 165 In contrast to healthy neutrophils, CGD neutrophils did not display the usual phagosome swelling, which was found to be independent of hydrolytic enzyme content and degranulation. It was suggested that normal phagocytic vacuoles enlarge due to an increase in the osmotically active products of bacterial digestion but that this does not occur in CGD neutrophils as the extreme acidification impairs digestion and thus swelling.
The second step in the proposed biphasic pH change is that of modest acidification, occurring once the respiratory burst subsides, and suggested to enhance activity of proteins with acidic pH optima, for example, hydrolases. The H + pump, V-ATPase, was hypothesized to be coupled to NADPH oxidase activity and to be responsible for the moderate acidification. However, in opposition of this theory, the rate of superoxide anion production was unrelated to V-ATPase activity and inhibition of the pump did not provoke alkalization of oxidase-active phagosomes. 126 Potential limitations of the studies suggesting biphasic pH change 165 include use of the pH sensor fluorescein, non-sealed phagosomes, and dye leakage. 127,166 In addition, pH changes to optimize granule proteins may not be necessary; there are examples of prominent neutrophil enzymes that potentiate MPO-mediated killing even when heat inactivated, for example, elastase and cathepsin G. 167,168 This could be because their cationic nature enables strong, disruptive electrostatic interactions with microbial surfaces, that consequently cause the microbe to be more susceptible to attack.
The current consensus is that neutrophil phagosomes acidify more weakly (certainly compared to macrophages) or remain neutral, due to robust NADPH oxidase activity. 169 This would certainly make biological sense as NADPH oxidase functions optimally at neutral pH, so maintenance of this would allow maximal ROS production. 170 Indeed, studies of single cells 126 127 In support of these findings, Chemaly et al. confirmed an inverse correlation of intra-phagosomal ROS production and V-ATPase accumulation on phagosomes. 170 Our group has generated recent evidence for the development of a more acidic pH as phagocytosis progresses (Figure 2). Using bacterial targets conjugated to a pH-sensitive dye, we demonstrated progressive acidification of the maturing neutrophil phagosome at later time points (up to 120 minutes) that was class 3 PI3K-and V-ATPase-dependent. 22 However, calibrated probes were not used at early time points to measure the exact pH. Of clinical relevance, exposure to the pro-inflammatory complement product, C5a, reduced the proportion of neutrophils phagocytosing (as expected given its effects on RhoA signaling) but also impaired phagosomal acidification in neutrophils that had ingested prey, indicating a distinct mechanism of failed phagosomal acidification. The C5a-mediated reduction in acidification was associated with a reduced phosphorylation of the V-ATPase G subunit.
Defective phagosomal acidification was also demonstrated in neutrophils from critically ill patients vulnerable to secondary infection, suggesting that acquired phagosome dysfunction in the context of systemic inflammation may increase susceptibility to infection. 24 Another interesting observation is that when neutrophils phagocytose other apoptotic neutrophils, for example found that azurophil granule-phagosome fusion and late-stage killing in FcyR-mediated phagocytosis was Ca 2+ -independent but that a separate mechanism of early azurophil delivery in the forming phagosome was Ca 2+ -dependent. 189 This suggests that it is only a transient early phase of azurophil granule, and potentially other granule, fusion and killing that is Ca 2+ -dependent and that later phagosome maturation mechanisms, such as ROS production and ion fluxes, Actin polymerization is also independent of Ca 2+ in neutrophils. 190 Conversely, actin severing and depolymerization by gelsolin, which is crucial to dissolve the thick polymerized actin ring that surrounds the forming phagosome, is Ca 2+ -dependent. 191 Other possible targets for Ca 2+ transients during phagosome formation include the Ca 2+ -dependent protease calpain, which has been implicated in clustering of β2 integrins 192 ; the Ca 2+ -binding protein synaptotagmin, which translocates to CR3-initiated phagosomes and is involved in particle uptake in macrophages 193,194 ; or the various Ca 2+ -binding annexins, which promotes membrane fusion and may be involved in actin dynamics during phagocytosis. 195,196 197 This suggests that the actin cytoskeleton may not be involved in granule-phagosome fusion.

| Cytoskeletal reorganization
Microtubule polymerization is also promoted by phosphorylation signaling cascades in activated neutrophils (distinct from actin); late phagosomes are found in close proximity to the centriole, part of the microtubule organizing center (MTOC) where granules are rich. 198 The use of colchicine, which binds to tubulin and prevents its polymerization into microtubules, led to phagosome disorganization and no preferential association with centrioles within the neutrophil, suggesting that microtubules are required to translocate the phagosome toward the granule-rich centriole; 198 in macrophages, this was observed directly. 199 In neutrophils, colchicine also inhibited the association of MTOC and azurophil granules with the phagosome, implicating microtubule involvement in the movement of granules and their delivery to the phagosome, as well as movement of the phagosome itself. 197 The association of kinesin, a motor protein involved in microtubule mediated transport, with neutrophil granules and microtubules also supports this theory. 200

| GTPases
The cytoskeleton's involvement in neutrophil trafficking is complex and the different granules are variously associated with actin and microtubules, which may explain how granules are differentially directed and regulated. 183,197 Actin and microtubule dynamics are predominantly controlled by Rac and Rab GTPases, which facilitate granule migration to the phagosome by orchestrating cytoskeletal rearrangement. Vesicle-membrane tethering, docking, and fusion are then regulated by Munc family proteins and SNAREs (soluble NSF-attachment protein receptors).
The precise secretory machinery associated with a particular granule is linked to its classification and function. Neutrophils display a non-redundant role for Rac2 in the secretion of azurophil, but not specific or gelatinase granules: Neutrophils from mice deficient in Rac2 showed absent azurophil granule exocytosis whereas specific and gelatinase granule release was unaffected. 201 Rac2 deficiency appears to mainly affect actin function in phagocytosis. 202 Rac1 and Rac2 are also required for the formation of NADPH oxidase, depending on the microorganism phagocytosed. 203 An example of directional granule trafficking control is demonstrated by Rab27a. Subcellular fractionation and immunoelectron microscopy has shown that Rab27a is located predominantly on gelatinase and specific granules, with lesser localization to azurophil granules. 204 Rab27a directs all granules for plasma membrane fusion but granules lacking Rab27a are still able to fuse with the phagosome: Neutrophils from Rab27a-deficient mice exhibited normal phagosome maturation and azurophil granule recruitment. 205,206 Conversely, Munc13-4, an effector of Rab27a present on all three granule subtypes, was found to be essential for phagosomal maturation and delivery of azurophilic granules to the phagosome, with its absence leading to impaired intracellular bacterial killing. 205 Therefore, Munc13-4 is a protein of key significance in phagosomal killing due to the importance of MPO in formation of oxidized halides, and the delivery of toxic proteases and bactericidal peptides.
Munc13-4 may regulate granule tethering to the phagosome, which correlates with the finding that tethering of secretory lysosomes to the plasma membrane in cytotoxic T lymphocytes requires the Munc13-4-Rab27 complex. 207 Alternatively, Munc13-4 may form complexes with SNARE proteins to regulate azurophil granule fusion.
Regulation of differential granule fusion and mobilization is also mediated by Src tyrosine kinases and granule associated-SNARE complexes, whereby association with certain family members dictates whether a granule is directed to the phagosome or plasma membrane. 208,209 For example, the Src family member, Hck, was found to be localized to azurophil granules, with translocation directed toward the phagosome 210 whereas Fgr was associated with specific granules, and increased their fusion with the plasma membrane. 211 Similarly, differential granule fusion is directed by various members of SNARE complexes. 208,212 Overall, a number of secretory control mechanisms ensure that the majority of azurophil granules (unlike specific and gelatinase granules) are preferentially targeted to the phagosome rather than the cell surface membrane, thus establishing a highly toxic phagosomal environment while limiting the capacity for host tissue damage.

| OXIDATIVE VER SUS NON -OXIDATIVE K ILLING
Conventionally, the two arms of oxidative (ROS) and non-oxidative (granule protein) killing in neutrophil phagosomes have been seen as opposing mechanisms, with evidence both for and against a directly microbicidal role for ROS in the phagosome. However, current research suggests a more synergistic approach.
As outlined in section 3, it was elucidated early on by Klebanoff that the respiratory burst in neutrophils produces a large amount of hydrogen peroxide (H 2 O 2 ). 94 (Table 1). This is demonstrated by the preserved ability of CGD neutrophils to kill Escherichia coli, where bacterial clearance is thought to be mediated by Bactericidal/permeability-increasing protein, 215,216 indicating that non-oxidative mechanisms can produce sufficient microbicidal effect against certain bacteria. An opposing theory of non-oxidative killing was thus proposed whereby the microbicidal capacity of neutrophils was due to the abundant granule proteins instead of ROS. It was suggested that the purpose of the electrogenic current of NADPH oxidase was to drive compensatory ion fluxes to optimize the conditions in the phagosome for granule enzymes (rather than primarily to generate ROS), which could also explain why pathogen killing is impaired in CGD. 136 However, whether any cations other than H + significantly influx into the phagosome is debatable and depends largely on the pH of the phagosome: If the pH remains close to neutral (as currently thought), then the influx will be predominantly H + whereas if the pH rises (initial alkalinization) then other cations must be involved (section 6).
Calculations based on the small amount of osmotic swelling of the phagosome estimate that most of the charge compensation must be osmotically neutral. 128 Thus, it is likely that the only cation entering the phagosome is H + .
Evidence for the importance of neutrophil granule enzymemediated killing is provided by mice deficient in the proteases, cathepsin G, and elastase. Neutrophils from these mice exhibited normal respiratory burst, ROS production, and iodination, but the mice were unable to resist infection with S. aureus (a prominent cause of infection in CGD) or Candida albicans (which also causes severe infection in MPO-deficient mice). 136  The mechanism of oxidative killing is still controversial.
Evolutionarily, the theory that NADPH oxidase only produces ROS for the purpose of membrane potential changes is unlikely due to the toxicity of its products. However, as the lifespan of neutrophils is so short, it is questionable whether neutrophils need to be protected against ROS. Presently, the conclusion is that there is not a single predominant mechanism, but a complex synergistic relationship between oxidative and non-oxidative killing that provides alternative compensatory systems, as is often the case with the immune system ( Figure 3).

| WHEN THE PHAG OSOME C ANNOT FO RM
Neutrophils are powerful and efficient killers, but bacteria are formidable opponents and with their rapid evolutionary speed they have developed a range of evasion tactics. A common evasion tactic of bacteria is the formation of aggregates that circumvent phagocytosis. Large fungal hyphae also obviate phagocytosis. However, ligation of the phagocytic dectin-1 receptor, which detects fungal elements, can divert degradative proteins from the phagosome to the nucleus, thereby enabling the formation of neutrophil extracellular traps (NETs) in situations where a pathogen is too large to engulf. 226 NETs, composed of decondensed chromatin acting as a scaffold for cytosolic and granule proteins, act to trap pathogens which cannot be phagocytosed, and can exert antimicrobial effects via exposed antimicrobial molecules. 227 Metzler et al. found that in a subset of resting neutrophils there is a protein complex, termed the azurosome, which is localized on the azurophil membrane. 228 (Table 1) phagocytosis, NET-generation may be dependent on both ROS and granule proteins.
The ability of neutrophils to "sense" pathogen size during attempted phagocytosis is intriguing. When neutrophils were exposed to Candida albicans hyphae, over 70% of total elastase was seen in the nucleus, accompanied by NET release, compared to less than 20% when Candida albicans was in the easily phagocytosable yeast form. 226 The exact mechanism is unclear, but it is possible that the "decision" for NETosis versus phagocytosis is dependent on whether elastase is delivered to the phagosome: phagocytosis, which acts more rapidly than NETosis (seconds to minutes versus minutes to hours), sequesters elastase to the phagosome and inhibits the pro-

| PHAGOSOME RE SOLUTION
Phagosome resolution is the concluding stage in the phagocytic process, comprising disposal or recycling of ingested contents and membrane resources, as well as providing the opportunity for antigen presentation. 231,232 Resolution is an important step that allows phagocytes to return to homeostasis but, in comparison with phagosome formation and maturation, resolution remains largely unexplored. Consequently, the molecular basis of phagosome resolution is rather speculative, predominantly based upon related processes such as lysosome turnover and autophagy. Limited experiments in phagosomes have almost exclusively been performed using macrophages (reviewed in [4]), whose phagosomes mature through fusion with endosomes and lysosomes in a manner which is very different from neutrophil phagosome-granule fusion. Therefore, although knowledge acquired from these experiments is often extrapolated, findings in macrophages during this phagocytic stage may not be truly representative of the situation in neutrophils.
The objectives of phagosome resolution are manifold: safe disposal of destroyed prey; waste management (including recycling) of ingested contents, including nucleic acids, proteins, and lipids; antigen presentation to lymphoid cells; and resorption of the phagosomal membrane. Resolution was originally thought to proceed in a similar fashion to unicellular eukaryotes, which egest indigestible content and plasma membrane. 233 However, recent work in macrophages has shown evidence of shrinkage and fission, 234,235 with resolution of phagolysosomes through fragmentation, by vesicle budding, tubulation, and constriction. 236 It is unclear how this reformation of lysosome-like organelles during fragmentation would apply to neutrophils as their granules are pre-formed. In addition to mechanical contraction, exportation of organic osmolytes produced by target degradation is essential for volume loss during phagosome resolution to prevent osmotically-induced hydrostatic pressure (reviewed in 237 ).
One conundrum of the phagosome maturation and resolution process is that phagocytes must be able to target the intraphagosomal lipids of ingested prey for degradation while maintaining an intact phagosomal membrane that is protected from lipolytic enzymes, but also be able to resorb this membrane once internal degradation is completed. The mechanisms controlling phagosomal membrane recycling are largely unknown but possibilities for the handling of various lipid species are discussed in. 4 There is a particular paucity of information regarding signaling events for phagosome resolution. In macrophages, the fragmentation process is thought to involve phosphoinositides, particularly PtdIns4P, and Rab GTPases, 237 which direct tethering to the endoplasmic reticulum. 235 Whether this process occurs in neutrophils is uncertain.
Given that phagosome resolution is a pivotal step in the successful completion of phagocytosis, this stage has garnered surprisingly little attention. In neutrophils, the processes required for phagosome resolution are almost completely unknown. Macrophages must be equipped to perform multiple rounds of phagocytosis while maintaining phagocytic and degradative capacity but it is not clear how vital this step is in the relatively short-lived neutrophil, which is itself efferocytosed by macrophages during inflammation resolution.
Given the unique nature of neutrophil granules, and the involvement of proteins required for phagosome fragmentation throughout phagosome maturation, there will be considerable challenges in elucidating this process.

| CON CLUS ION
For the past few decades, progress in neutrophil biology has been extremely slow; many researchers have opted to study cell types that are more easily cultured and genetically manipulated. Over time, however, the neutrophil has been revealed as unique, and its plasticity and complexity underestimated. The view that neutrophils are single function suicide killers has been overtaken by evidence of integration of complex signals to make "decisions" and instigate a range of different signaling pathways in the face of invading pathogens. This is particularly apparent in the neutrophil phagosome, which has phenotypic and functional plasticity. We now know that the neutrophil phagosome environment changes to optimize killing, but an interesting debate still surrounds the complex interplay between various systems.
One key area of contention regards the antimicrobial role of ROS, which are small, transient, and ubiquitous, juxtaposed with that of granule proteins, which are macromolecules with specific killing mechanisms. For years, researchers have believed a single mechanism to be predominant; however, more recent research suggests synergism. This is highlighted by the different but overlapping consequences of CGD, MPO deficiency, and cathepsin G and/or elastase knockout. As may be expected, the ability of the compromised host to prevent bacterial or fungal infection also varies depending on the target organism. 136,224,238,239 It is likely that some pathogens are more susceptible to a particular mode of killing. For example, neutrophil killing of S. aureus appears to require ROS as cells experiencing a lack of molecular oxygen under hypoxia displayed impaired bacterial killing (but not ingestion). 240 This makes sense in the context of the wider immune system, which employs a combination of different effector systems to tackle infection.
Another area of phagosome research which has not yet provided a conclusive answer is that of intra-phagosomal pH. Variation in experimental techniques and theories has led to a range of different persuading arguments; however, there is consensus that neutrophils are unique phagocytes, with a weaker acidification than macrophages due to their increased ROS production, and it is likely that the phagosome acidifies more as it matures. The lack of decisive data on pH has also

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available