Complement C5a impairs phagosomal maturation in the neutrophil through phosphoproteomic remodelling

Critical illness is accompanied by the release of large amounts of the anaphylotoxin, C5a. C5a suppresses antimicrobial functions of neutrophils which is associated with adverse outcomes. The signalling pathways that mediate C5a-induced neutrophil dysfunction are incompletely understood. Healthy donor neutrophils exposed to purified C5a demonstrated a prolonged defect (7 hours) in phagocytosis of Staphylococcus aureus. Phosphoproteomic profiling of 2712 phosphoproteins identified persistent C5a signalling and selective impairment of phagosomal protein phosphorylation on exposure to S. aureus. Notable proteins included early endosomal marker ZFYVE16 and V-ATPase proton channel component ATPV1G1. A novel assay of phagosomal acidification demonstrated C5a-induced impairment of phagosomal acidification which was recapitulated in neutrophils from critically ill patients. Examination of the C5a-impaired protein phosphorylation indicated a role for the phosphatidylinositol 3-kinase VPS34 in phagosomal maturation. Inhibition of VPS34 impaired neutrophil phagosomal acidification and killing of S. aureus. This study provides a phosphoproteomic assessment of human neutrophil signalling in response to S. aureus and its disruption by C5a, identifying a defect in phagosomal maturation and new mechanisms of immune failure in critical illness.


33
Whilst several signals mediating aspects of C5a-induced neutrophil dysfunction have been established 34 (Conway Morris et al, 2011;Denk et al, 2017a;Huber-Lang et al, 2002b), a global picture of 35 signalling in neutrophils encountering common pathogens and how this process is perturbed by C5a 36 does not exist. Such studies are challenging in neutrophils owing to their high degradative enzyme 37 content and short in-vitro survival times (Luerman et al, 2010). 38

39
This study aimed to characterise the neutrophil phosphoprotein response to a common nosocomial 40 pathogen, Staphylococcus aureus, and investigate how this is perturbed by prior exposure to C5a. 41 Our differential phosphoprotein analysis implicated C5a in altered phagosomal maturation, findings 42 that we confirmed with functional neutrophil assays in C5a-treated healthy donor cells and those from 43 critically ill patients. The phosphoprotein response to S. aureus implicated the involvement of the 44 phosphatidylinositol 3-kinase VPS34, hence we continued examined the effects of this enzyme on 45 phagosomal maturation. 46

C5a induces a prolonged defect in neutrophil phagocytosis of bacteria 48
C5a induces a defect in phagocytosis of the clinically relevant bacterial species S. aureus ( Figure 1A) 49 and E. coli (1B). Pulse exposure of neutrophils to C5a revealed a persistent defect in phagocytosis 50 lasting at least seven hours (1C), with short pulses inducing a significant defect. These effects were 51 not explained by the loss of cell viability (1D). A similar prolonged defect was identified in the whole 52 blood assay (1E), representing continuous exposure of neutrophils to C5a (which cannot be washed 53 off in this assay). The ability of C5a to inhibit phagocytosis was dependent on the temporal 54 relationship between C5a and S. aureus exposure. Only pre-exposure to C5a induced the defect in 55 phagocytosis, whereas co-exposure or the addition of C5a 30 minutes after S. aureus addition failed to 56 induce a defect (1F). 57

58
To explore the potential mechanisms whereby pre-exposure to S. aureus prevents the inhibitory effect 59 of C5a, we examined whether this could be due to reduced C5aR1 expression. Although we could 60 demonstrate a reduction in C5aR1 following S. aureus exposure ( Figure S1A), this was modest and 61 similar to the reductions induced by other inflammatory mediators including lipopolysaccharide 62 (LPS) and leukotriene A (LTA), neither of which ameliorated the subsequent suppressive effect of 63 C5a ( Figure S1B). Further, C5a and not LPS, LTA, granulocyte-macrophage colony-stimulating 64 factor (GM-CSF) and tumour necrosis factor (TNF) reduced neutrophil phagocytosis ( Figure S1C). 65 To confirm the functional relevance of C5a-impaired phagocytosis, we demonstrated that C5a pre-66 treatment reduced bacterial killing of S. aureus ( Figure S1D). 67 68

S. aureus and C5a induce widespread changes in the neutrophil phosphoproteome 69
Although key signalling 'nodes' have been identified in neutrophils following C5a exposure (Conway 70 Morris et al, 2009Morris et al, , 2011, no map of global signalling networks has been produced. Given the 71 rapidity of the C5a-induced phagocytic impairment demonstrated above, and the known signalling 72 kinetics of G-protein coupled receptors (GPCRs) (Lohse et al, 2008), we examined post-translational 73 modification by phosphorylation (i.e. a phosphoproteomic approach). of phosphorylation changes within the previously reported range (Papachristou et al, 2018). Changes 79 in the human proteome were minimal (2 % of total proteome with S. aureus treatment) whereas 80 phosphoprotein expression varied markedly (31.6 % of total phosphoproteome with S. aureus 81 treatment, Table S1). Figure 2 shows the top 2.5% most variable phosphoproteins with protein 82 identification, whereas the top 25 % are shown in Figure S3 to demonstrate wider changes within the 83 phosphoproteome. The phosphoproteomic and proteomic datasets are publicly available in the PRIDE 84 database (data available to reviewers, will be made public on acceptance of manuscript). 85 86 C5a exposure induces persistent alteration in phosphoproteins across several pathways. 87 Figure 3A shows a volcano plot comparing neutrophils treated with C5a versus vehicle control. 119 88 proteins were significantly differentially phosphorylated at 1 hour, indicating persistent signalling, 89 consistent with the prolonged inhibition of phagocytosis seen in Figure 1. Notably, C5aR1 remained 90 highly phosphorylated (a modification key to its internalisation) (Braun et al, 2003) and this change 91 has been used to identify C5a-exposed, dysfunctional neutrophils (Conway Morris et al, 2009Morris et al, , 2011Morris et al, , 92 2013Schmidt et al, 2015;Unnewehr et al, 2013). Pathway enrichment using Metascape (Zhou   Figure S4). 104 105 C5a exposure prior to S. aureus reduced the phosphoprotein response to the bacterium considerably 106 ( Figure 4C). However, comparing C5a and control treated cells exposed to S. aureus, 19 proteins 107 were identified, suggesting selective pathway modulation ( Figure 4D). When mapped to known 108 pathways using Metascape (Zhou et al, 2019) and manually annotated from the Uniprot database (The 109 Uniprot Consortium, 2019), a pattern of reduced phosphorylation of phagosomal maturation proteins 110 (Tables  111   Table ) and pathways ( Figure 4E) emerged. Notably, early endosomal marker ZFYVE16 and its 112 interactor TOM1 had impaired phosphorylation following C5a exposure, as did V-type ATPase 113 subunit G1 (which is critical for phagosomal acidification). ZFYVE16 requires phosphatidylinositol-114 3-phosphate (PI3P) for recruitment to the phagosome (Sorkin & Von Zastrow, 2009). Another 115 prominent PI3P-responsive protein noted was Ras-related protein 7a (RAB7A), although this protein 116 was not differentially phosphorylated between the C5a/S. aureus and vehicle control/S. aureus 117 conditions. Figure S5 shows individual donor data for these key proteins. 118 Our dataset suggests that C5a exposure that precedes pathogen encounter prevents effective signalling 120 through the phagosomal maturation pathways, and links intracellular signalling to the prolonged 121 functional impairment noted in this context. The other major cluster of differentially phosphorylated 122 proteins were nuclear and nuclear membrane proteins, many of which are involved in mitosis and 123 nuclear envelope integrity. 124 125 C5a induces an impairment in phagosomal acidification, distinct from the impairment in 126 ingestion. 127 The phosphoproteomic signature of altered phagosomal maturation following C5a exposure, and the 128 involvement of V-ATPase suggested that C5a had effects beyond impaired ingestion of bacteria. To 129 disentangle the effects of phagocytic ingestion and phagolysosomal acidification, S. aureus 130 bioparticles co-labelled with the pH-insensitive dye AF488 and pHrodo™ red were used. Neutrophils 131 ingested particles, and then subsequently acidified the phagosome, a process which could be ablated 132 by the addition of the V-ATPase inhibitor bafilomycin (Bowman et al, 1988) (Figures 5A and B). C5a 133 pre-treatment increased the proportion of neutrophils that failed to ingest particles ( Figure 5C) and 134 increased the population that ingested particles but failed to acidify the phagosome ( Figure 5D). 135 Recent reports suggest that C5a induces Na + /H + exchanger-1 (NHE-1)-mediated cytoplasmic 136 alkalinisation (Denk et al, 2017a). An NHE-1 inhibitor did not alter the C5a-mediated effect on 137 phagosomal acidification ( Figure 5E), suggesting that the pathways mediating these two effects of 138 C5a on neutrophils are distinct. Furthermore, we confirmed previous work (Huber-Lang et al, 2002b) 139 showing C5a impaired ROS production ( Figure S6), which in combination with the current findings, 140 suggests C5a induces a generalised failure of phagosomal maturation in addition to its effect on 141 phagocytic ingestion. 142 143 144

VPS34 inhibition impairs phagosomal acidification 145
The differential phosphoprotein analysis (Tables  146   Table ) and phagosomal acidification assays ( Figure 5) demonstrated impaired phagosomal 147 maturation after exposure to C5a. As noted, several of the phosphoproteins that were differentially 148 phosphorylated are known interactors with PI3P. The phosphatidylinositol 3-kinase VPS34 is the 149 dominant source of PI3P in mammalian cells (Devereaux et al, 2013). Although VPS34 itself was 150 detected, its phosphorylation status was not significantly altered. However, the finding that C5a 151 altered the phosphorylation status of PI3P-responsive proteins led us to explore the role of VPS34 in 152 phagosomal acidification. We used the selective inhibitor, VPS34IN1 (Bago et al, 2014) to examine 153 the role of this enzyme in phagosomal acidification, and how this related to the defect induced by 154 C5a. VPS34IN1 did not alter the percentage of neutrophils that underwent phagocytosis ( Figure 6A  155 and time-course in E) but did lead to a reduction in the overall number of particles ingested ( Figure  156 6B) and a more marked reduction in pHrodo signal ( Figure 6C and time course in F), indicating 157 VPS34IN1 impairs phagosomal acidification. VPS34 inhibition also led to an impairment in the 158 killing of S. aureus ( Figure 6D), similar to that observed with C5a ( Figure S1D) without a significant 159 reduction in phagosomal ROS production ( Figure S7). 160 161 162

Neutrophils from critically ill patients exhibit defective phagosomal acidification 163
To establish the relevance of our findings to the clinical setting, we used our assay of phagosomal 164 acidification to interrogate neutrophils obtained from critically ill patients and healthy volunteers. We 165 assessed neutrophil function in critically ill patients, defining neutrophil dysfunction as phagocytosis 166 of <50% in our previously established zymosan assay ( Figure 7A), a threshold associated with a 167 markedly increased risk of nosocomial infection (Pinder et al, 2018;Conway Morris et al, 2009, 168 2011. Using our phagosomal acidification assay, we then compared patients with dysfunctional 169 neutrophils to critically ill patients with functional neutrophils and healthy controls. Dysfunctional 170 neutrophils exhibited a failure of phagosomal acidification ( Figure 7B) that was not seen in patients 171 with functional neutrophils. Furthermore, we observed a correlation between C5aR1 expression 172 (decreased after C5a exposure) and phagocytosis ( Figure 7C) and an inverse correlation between 173 C5aR1 expression and phagosomal acidification ( Figure 7D), though the latter correlation did not 174 reach statistical significance. The patients with dysfunctional and functional neutrophils could not be 175 readily identified by clinical factors such as severity of illness or precipitating insult (Table S2). These 176 data provide evidence of dysfunctional phagosomal acidification in critically ill patients and imply a 177 role for C5a in driving this dysfunction. 178

179
Our data demonstrate that C5a induces both a prolonged defect in phagocytosis of relevant pathogens 180 (S. aureus and E.coli), and persistent signalling across multiple pathways for some hours after the 181 well characterised initial signalling events such as ionised calcium flux (Blackwood et al, 1996) and 182 PIP3 generation (Houslay et al, 2016). This finding supports the proposal that persistent C5a-induced 183 signalling may mediate the neutrophil dysfunction observed in critically ill patients (Conway Morris 184 et al, 2009Morris 184 et al, , 2011. 185 To our knowledge, the data presented here (Figures 3-5) represent the deepest sequencing of the 187 human neutrophil proteome and phosphoproteome (Muschter et al, 2015;Tak et al, 2017;McLeish et 188 al, 2013). These data provide a phosphoproteomic assessment of the human neutrophil response to S. 189 aureus and C5a. Unlike transcriptomic data (Juss et al, 2016;Rorvig et al, 2013;Kobayashi et al, 190 2002), phosphoproteomics provides a direct assessment of mediators that are likely to have functional 191 implications, especially in short-lived cells such as neutrophils (Luerman et al, 2010;Fessler et al, 192 2002) and early pathogen exposure timepoints, as examined in this study. previous identification of key roles for these molecules in C5a-mediated functional deficits in 202 neutrophils (Conway Morris et al, 2009Morris et al, , 2011Scott et al, 2015). The marked suppression of the 203 phosphorylation response to S.aureus induced by C5a pre-treatment is not simply a response to 204 reduced particle ingestion. Fifteen minutes after pathogen contact there were limited differences in the 205 ingestion rates between C5a and control treatments, and these became more marked over time ( Figure  206 1). Furthermore, the differential analysis of C5a/S. aureus versus vehicle control/S. aureus conditions 207 identified defects in specific signalling pathways, most notably those involving endosomal trafficking. 208 This led us to examine the process of phagosomal maturation, and to identification of a C5a-induced 209 failure of phagosomal acidification ( Figure 5) with similar findings in critically ill patients (Figure 7). 210 Failure of phagosomal maturation and intracellular killing has been described in primary immune 211 deficiency (Buvelot et al, 2017), but has not previously been described as part of the immuno-paresis 212 of critical illness. Impaired phosphorylation in pathways involving nuclear envelope breakdown and 213 nuclear pore disassembly by C5a was unanticipated. The functional relevance of these changes 214 remains unclear, though they may be early processes in the formation of non-lethal DNA-containing 215 neutrophil extracellular traps (NETs) (Pilsczek et al, 2010). 216 217 Important signalling proteins involved in the process of phagosomal maturation (such as RAB7A, 218 TOM1 and ZFYVE16) can be recruited to the phagosomal membrane by PI3P produced 219 predominantly by VPS34 (Botelho et al, 2000;Levin et al, 2016;Sorkin & Von Zastrow, 2009). Both 220 ZFYVE16 and TOM1 phosphorylation were impaired by C5a exposure. We investigated the role of 221 VPS34 as a mediator of neutrophil bactericidal function, and found that selective VPS34 inhibition 222 produced a similar impairment in phagosomal acidification to that observed with C5a ( Figure 6). The 223 finding that a similar defect could be induced by inhibiting VPS34, the dominant source of PI3P in 224 neutrophils (Devereaux et al, 2013), adds further validation to the pathway signature identified in the 225 phosphoproteomic profile. 226 Ellson and colleagues (Ellson et al, 2001) demonstrated that PI3P plays an important role in targeting 228 neutrophil oxidase components to phagosomal membranes and its importance in phagosomal 229 maturation has also been identified in Dictyostelium discoideum (Buckley et al, 2019), murine 230 macrophages, and macrophage-like cell lines (Naufer et al, 2018). However, the role of VPS34 in 231 human neutrophils has previously been inferred indirectly (Anderson et al, 2008), owing to prior lack 232 of selective inhibitors and the difficulties of genetically manipulating human neutrophils. Anderson 233 and colleagues (Anderson et al, 2008) demonstrated a role for VPS34 in NADPH oxidase-mediated 234 reactive oxygen species generation in neutrophils. We found a non-significant reduction in ROS 235 production ( Figure S7) that was much less marked than the effect on phagosomal acidification. The 236 reasons for these divergent findings are uncertain, though may include differences in ROS 237 measurement assays, our use of a selective VPS34 inhibitor, and differences between primary human 238 neutrophils and cell lines. The mechanism by which VPS34 inhibition impairs killing of S. aureus 239 requires further investigation, as phagosomal acidification is not thought to be critical to this process 240 (Lacoma et al, 2017) and it is likely that the enzyme inhibition leads to further defects in phagosomal 241 maturation. It is intriguing to note that whilst VPS34 inhibition does not reduce the percentage of cells 242 that undergo phagocytosis ( Figure 6A), consistent with previous work (Anderson et al, 2008), it does 243 reduce the number of particles ingested ( Figure 6B). This suggests a hitherto undescribed relationship 244 between phagosomal maturation and the capacity of cells to ingest particles. 245

246
Our data also demonstrate that the timing of C5a exposure (before, alongside, or after pathogen 247 encounter) has an important effect on neutrophil function. Only pre-exposure to C5a impaired 248 subsequent neutrophil phagocytosis (Figure 2). Reduced C5aR1 availability for ligation by C5a is 249 unlikely to explain this observation, as C5aR1 downregulation is induced by multiple agents that do 250 not have the same effect on phagocytosis ( Figure S1). Given the marked phosphoproteomic response 251 to S. aureus and its distinction from the response to C5a (Figures 3 and 4), a potential explanation is 252 that signalling induced by S. aureus simply overwhelms C5a-induced phosphorylation changes unless 253 they were established prior to S. aureus exposure, though this hypothesis requires further 254 experimental validation. 255

256
This study was conducted entirely in primary human neutrophils, using C5a, an established, clinically 257 relevant modulator of neutrophil function that has been linked to a range of adverse outcomes in 258 critically ill patients. The use of clinically relevant pathogens, and the development of a whole-blood 259 bacteraemia model, increases the relevance of our study to the in-vivo situation. Impaired ingestion of 260 zymosan by patient neutrophils has been associated with adverse outcomes including development of 261 subsequent nosocomial infection (Conway Morris et al, 2011). The finding that patients with such 262 impairment also manifest impaired phagosomal acidification that correlates with markers of C5a 263 exposure (Figure 7) suggests that the identified mechanisms may be clinically relevant. 264 265 Several potential limitations should be highlighted. The phosphoproteomic response to S. aureus was 266 evoked with heat-killed bacterial particles, conjugated with fluorescent dyes, and these may not fully 267 reflect the response to live bacteria, although they do allow parallel functional assessment and 268 standardisation of the stimulus between donors and across research sites. Although whole blood is a 269 more physiologically relevant than cell-culture media, it remains an abstraction from the situation in- Assent was provided by a personal or nominated consultee.
Further details of methods and reagents described below are available in the supplementary materials.

Neutrophil isolation
Neutrophils were isolated from citrated peripheral venous blood by using a modification of the discontinuous plasma-Percoll density gradient centrifugation technique initially described by Böyum in 1968. (Boyum, 1968)

Phagocytosis of pHrodo™ S. aureus and E. coli Bioparticles by purified neutrophils
Purified human neutrophils, suspended in Iscoves Modified Dulbecco's Medium (IMDM) with 1 % autologous serum at a concentration of 5 x 10 6 /mL, were incubated in microcentrifuge tubes with purified human C5a or vehicle control. pHrodo-conjugated S. aureus or E. coli bioparticles were opsonised, in 50 % autologous serum for 30 min prior to be being added to the suspended cells.

No-wash, no-lyse whole blood assay of neutrophil phagocytosis and ROS production
Blood, collected into argatroban 150 µg/mL, was treated with inhibitors or priming agents as indicated in the respective figure legends, before being exposed to S. aureus pHrodo™/dihydrorhodamine (DHR) or E.coli pHrodo™. Aliquots were stained on ice with anti-CD16 antibody, diluted and analysed by flow cytometry (Attune NxT).
In variations on this assay, S. aureus particles labelled with the pH-insensitive dye AlexaFluor (AF)488 or dual labelled with AF488 and pHrodo red were used. pHrodo red conjugation of AF488 S. aureus was performed in-house using the pHrodo particle labelling kit (Thermofisher). Fluorescence of extracellular particles was quenched with trypan blue (0.1mg/mlL). Patient samples were analysed in a different laboratory that did not have access to an Attune Nxt flow cytometer, to fit with established workflows in this laboratory red cells were lysed using Pharmlyse (BD Bioscience, Wokingham, UK) followed by washing twice using a Facswash Assistant (BD Bioscience) prior to undertaking flow cytometry (Fortessa, BD Bioscience).

Bacterial killing assay -whole blood
Methicillin-sensitive S. aureus (MSSA) bacteria (strain ASASM6, kind gift from Prof Gordon Dougan, University of Cambridge) were grown to early log-phase. Blood was collected into argatroban and incubated with bacteria for 1 hour. Human cells were lysed by addition of pH 11 distilled water for 3 minutes before plating of serial dilutions on Colombia blood agar.

Preparation of whole human neutrophil lysates for phosphoproteomics
Neutrophils were isolated from whole blood as detailed above, and resuspended in RPMI 1640 media containing 10 mM HEPES with 1 % autologous serum (AS) at a concentration of 1x10 7 cells/mL.

Proteomic and phosphoproteomic studies
Triplicates of 1x10 7 neutrophils were treated with vehicle control or C5a (100 nM, 60 minutes) at 37 °C before addition of pHrodo™ S. aureus (15 µg/mL). Phagocytosis was allowed to occur for 15 minutes. Aliquots were withdrawn from each triplicate and pooled at the indicated timepoints. Cells were centrifuged at 400 g for 5 min at 4 °C, supernatants aspirated, and cell pellets snap frozen in liquid nitrogen. Cells were lysed by the addition of 0.5 % sodium dodecyl sulphate (SDS)/0.1 M triethylammonium bicarbonate (TEAB) buffer and sonication, before undergoing centrifugation, trypsin digestion, tandem mass tag labelling, fractionation, phosphopeptide enrichment, and liquid chromatography and tandem mass spectrometry (LC-MS/MS) analysis. The experimental schematic can be seen in Supplemental figure S8.

Statistical analysis of wet laboratory data
Data are presented as individual data points with summary statistics (median and interquartile range (IQR) or mean and standard deviation (SD) according to whether data are normally distributed.
Parametric or non-parametric statistical tests were applied as appropriate after data were tested for normality using the D'Agostino-Pearson test. Tests used for comparisons are indicated in figure legends. Two-tailed P values were computed, P < 0.05 was considered statistically significant. Nonsignificant differences have not been indicated in figures for clarity. Statistical analyses were undertaken using GraphPad Prism v8.0 (GraphPad Software; San Diego; California).

Statistical analysis of phosphoproteomics data
Spectral .raw files from data dependent acquisition were processed with the SequestHT search engine on Thermo Scientific Proteome Discoverer™ 2.1 software. Data were searched against both human and S. aureus UniProt reviewed databases at a 1 % spectrum level false discovery rate (FDR) criteria using Percolator (University of Washington). MS1 mass tolerance was constrained to 20 ppm, and the fragment ion mass tolerance was set to 0.5 Da. TMT tags on lysine residues and peptide N termini (+229.163 Da) and methylthio (+45.988 Da) of cysteine residues (+45.021 Da) were set as static modifications, while oxidation of methionine residues (+15.995 Da) and deamidation (+0.984 Da) of asparagine and glutamine residues were set as variable modifications. For TMT-based reporter ion quantitation, we extracted the signal-to-noise (S:N) ratio for each TMT channel. Parsimony principle was applied for protein grouping.
Peptide and phosphopeptide intensities were normalised across conditions using median scaling and then summed to generate protein and phosphoprotein intensities. Proteins and phosphoproteins were independently identified and quantified in all samples from all four donors; species not meeting these criteria were excluded from subsequent analysis. Log base 2 fold change (Log2FC) was calculated between conditions of interest, compared across n = 4 donors and tested for statistical significance by limma-based linear models with Bonferroni's correction for multiple testing. Hierarchical clustering using Euclidean distance was undertaken on the entire dataset. Heatmaps and volcano plots were generated as shown in Results. Statistical analyses were performed in RStudio (RStudio Team, 2016) using the qPLEXanalyzer (Papachristou et al, 2018) package, and plots were produced using the ggplot2 (Wickham, 2016) package.

Data sharing statement
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol Y et al, 2019) partner repository with the dataset identifier PXD017092 and will be made public on acceptance after peer-review   E. coli (B) bioparticles. Data are presented as the median phagocytic index for each condition for n=7 (A) or 6 (B) independent donors, *P = 0.016 (A) and 0.031 (B) by Wilcoxon's matched-pairs signed rank test. C: Neutrophils were pulsed with 100 nM C5a or PBS control for the indicated periods of time, followed by 2 washes. S. aureus bioparticles were then added and cells were incubated for the indicated time points. Data are presented as the mean and SD of the phagocytic index of C5a-treated cells relative to their paired vehicle control for n=5 independent experiments. P < 0.0001 for time and P = 0.0186 for treatment by two-way ANOVA. ***P = 0.0001 ****P < 0.0001 by Dunnett's multiple comparison test. D: Data are presented as the mean and SD of the percentage of DRAQ7 positive, dead cells for n=5 independent experiments. P = 0.378 for time and P = 0.349 for treatment by two-way ANOVA. E: Anticoagulated whole blood was pre-treated with 300 nM C5a or control for the indicated duration before phagocytosis was measured as previously indicated. Data are presented as the mean and SD of the cumulative phagocytic index for 4 independent experiments. P < 0.0001 by two-way ANOVA, ****P < 0.0001, ***P < 0.001 by Sidak's multiple comparisons test. F: S. aureus particles were incubated with isolated PMNs in the presence of 100 nM C5a or PBS added at the indicated time points, with time 0 representing the time of addition of S. aureus bioparticles. Experiments proceeded for the indicated time points and phagocytic index quantified. Data are presented as the mean and SD of the phagocytic index of C5a-treated cells relative to their paired vehicle control for n=5 independent experiments. P < 0.0001 for time and P = 0.0186 for treatment by two-way ANOVA. ****P < 0.0001 by Dunnett's multiple comparisons test.

Figure 5: C5a induces an impairment in phagosomal acidification, distinct from the impairment in ingestion
A: Exemplar flow cytometry plots of whole blood pre-treated with vehicle control or bafilomycin A (60 min; 100nM) prior to exposure to 5 µg/mL co-labelled AF488/pHrodo red S. aureus for 120 min. Both phagocytosis (x-axis) and phagosomal pH (y-axis) can be measured simultaneously in the same population of cells. pHrodo™ fluorescence increases with decreasing pH, indicating phagosomal maturity as shown. B: Conditions as in A. Data are shown as individual data points with mean for n=7 individual donors. P = 0.016 by Wilcoxon's test. C: Whole blood was pre-treated with vehicle control or C5a (300 nM; 60 minutes) prior to exposure to phagocytosis probe for 180 min. Phagocytosis without maturation (i.e. AF488 signal) is shown. Data are shown as mean and SD of n = 5 individual donors. ****P < 0.0001 by repeated-measures two-way ANOVA with Bonferroni's multiple comparisons test. D: Conditions as in C. The percentage of S. aureus particle positive (AF488+) cells with low pH (mature) and high pH (immature) phagosomes is shown for control and C5a-treated conditions. Data are shown as mean and SD of n = 5 individual donors. ***P < 0.001 by repeated-measures two-way ANOVA with Bonferroni's multiple comparisons test. E: Whole blood was pre-treated with C5a, NHE-1 inhibitor (5µM), or both, then exposed to maturation probe for 60 min. The percentage of AF488+ cells with high pH (immature) phagolysosomes is shown. Data are shown as individual data points with median from n = 7 individual donors. P = 0.0080 by Friedman's test, *P < 0.05 for Dunn's test of multiple comparisons, ns = non-significant.

Figure 6: VPS34 inhibition impairs phagosomal acidification
Whole blood was pre-treated with vehicle control or VPS34IN1 (1 µM; 60 min) prior to addition of 5 µg/mL maturation probe (A-D), or live S. aureus (E), for 120 minutes prior to analysis. A: Percentage of neutrophils that have phagocytosed bioparticles. P = 0.31 by Wilcoxon's test. n = 6 individual donors. B: MFI of ingested particles, indicating relative quantity of phagocytosis. P = 0.03. by Wilcoxon's test. n=6 individual donors. C: pHrodo™ Median Fluorescent Intensity (MFI), indicating phagosomal acidification. P = 0.03. by Wilcoxon's test. n=6 individual donors. D: After phagocytosis of live bacteria, human cells were lysed in alkaline dH2O and surviving bacteria were incubated overnight on blood agar. Bacterial survival was quantified by counting colonies. P = 0.03 by paired t-test, n=5 individual donors. E-F: Whole blood was processed as above with quantification of phagocytosis (E) and acidification (F) at the indicated time points. There was a reduction in phagosomal acidification as shown but no change in percentage of cells that underwent phagocytosis. Data are shown as mean and SD of n=5 individual donors. **P = 0.0058 for drug treatment by repeated measures two-way ANOVA with Bonferroni's multiple comparisons test.

Figure 7: Neutrophils from critically ill patients exhibit defective phagosomal acidification
A: Zymosan-based assay demonstrating differentially impaired phagocytosis in critically ill patients. Data are shown as individual patients/controls with median values indicated. n = 6 patients with dysfunctional neutrophils and 5 patients with functional neutrophils respectively. ** P=0.004 by Mann-Whitney U-test. B: Neutrophil phagosomal acidification was assessed in whole blood from critically ill patients using the maturation probe. Patients were classed as dysfunctional using the assay from A. Data are shown as individual patients/controls with mean from n = 6 patients with dysfunctional neutrophils, 5 patients with functional neutrophils and 10 healthy controls respectively. P = 0.04 by one-way ANOVA. **P < 0.01 by Holm-Sidak's test of multiple comparisons. C, D: C5aR1 expression was assessed by flow cytometry and correlated (Spearman) with phagocytosis (C) and phagosomal acidification (D) for n = 12 patients. NB: One patient's cells did not adhere to tissue culture plastic for the zymosan assay, thus they could not be assigned to dysfunctional or non-dysfunctional groups shown in A and B. C5aR1 expression and maturation probe data was available to allow inclusion in correlation analyses in C and D, hence the difference in numbers between these figures.   Tables   Table I: Differentially phosphorylated proteins between C5a and control-treated neutrophils exposed to S. aureus.
All 19 phosphoproteins with Bonferroni adjusted p-values < 0.05 for difference in phosphorylation status between the Control plus S. aureus vs C5a plus S. aureus conditions. Subcellular location and function manually annotated from Uniprot database (The Uniprot Consortium, 2019