Acinetobacter baumannii phenylacetic acid metabolism influences infection outcome through a direct effect on neutrophil chemotaxis

Systemic Infection with Live A. baumannii Causes Lethal Disease in Zebrafish.

One of the most life-threatening forms of A. baumannii disease is bloodstream infection (1). As A. baumannii is not known to be a natural zebrafish pathogen, we assessed its capacity to cause disease in zebrafish embryos. A reference hospital-acquired A. baumannii strain [American Type Culture Collection (ATCC) 17978] was inoculated into the circulation of embryos 48 h postfertilization (hpf). As few as 300 cfu of bacteria caused lethal disease, with mortality dependent on the infecting inoculum (Fig. 1A). Similar findings were observed for other clinical A. baumannii strains (AB307, A9844, AB0059, ATCC 19606) (Fig. S1 A–D). Neither injection of washed, heat-killed bacteria (Fig. 1A) nor culture filtrates from exponential or stationary-phase growth caused disease. These data show that live bacterial cells are required for A. baumannii pathogenesis in zebrafish.

To determine the ability of zebrafish to mount an effective response to A. baumannii bloodstream infection, we performed in vivo bacterial growth kinetic studies after infection with 1 × 10³ cfu of A. baumannii per embryo (Fig. S1E). The bacterial burden increased to 20 h postinfection (hpi) but then subsided by 44 h to undetectable levels (Fig. S1E), suggesting that zebrafish can mount an innate immune response that clears an initially expanding bacterial population. Conversely, after inoculation of A. baumannii into the zebrafish yolk sac, a site devoid of innate immune cells (14), zebrafish were hypersusceptible to lethal disease (Fig. S1F), further supporting a role of innate immune responses in the defense against A. baumannii.

Protective Phagocyte Responses to A. baumannii Infection in Zebrafish.

In mammals, neutrophils play an important role in response and clearance of A. baumannii infection (4, 5). To examine phagocyte–A. baumannii interactions in zebrafish, we used transgenic embryos with fluorescently labeled leukocytes. Localized infection with GFP-expressing bacteria (A. baumannii) was established within somatic muscle. Within minutes of inoculation, neutrophils migrated to the site of A. baumannii infection, where they accumulated, phagocytosed bacteria, and then migrated away, clearing the bulk of infection by 4 h (Fig. 1B and Movie S1). Histopathology of the infection site confirmed neutrophilic phagocytosis of bacteria (Fig. S1G). Macrophages, which have also been shown to be important first responders to A. baumannii infection in mammals (15), were also recruited to the infection site but appeared to have less phagocytic activity (Movie S2), and similar numbers were seen with a tissue trauma control (PBS solution injection) (Fig. 1C). These data highlight the specificity of neutrophils in responding to pathogen-associated factors and macrophages in responding to tissue damage.

Within neutrophils, pathogen degradation typically occurs by hydrolytic enzymes within acidic phagolysosomes, whereas the remainder of the cytoplasm remains at neutral pH (16). To characterize the intracellular handling of A. baumannii following phagocytosis by neutrophils in vivo, we labeled A. baumannii with pHrodo-dextran, which emits red fluorescence only under low pH conditions. In Tg(mpx:GFP) zebrafish with green fluorescent neutrophils, red pHrodo-dependent fluorescence emerged within cytoplasmic vacuoles at 50 min postinfection (Fig. 2A, Fig. S2A, and Movie S3), confirming that intracellular handling and degradation of A. baumannii following neutrophil phagocytosis is within acidic phagolysosomes.

To further characterize the functional role of leukocytes in A. baumannii infection, we depleted zebrafish of neutrophils by knocking down the gene encoding the zebrafish ortholog of the granulocyte colony stimulating factor receptor [colony stimulating factor 3 receptor (csf3r)] (17). This led to greater susceptibility of zebrafish embryos to lethal A. baumannii infection (Fig. 2B). In contrast, irf8 knockdown, which depletes macrophages but expands the neutrophil population (18), did not impact on A. baumannii infection survival (Fig. 2C and Fig. S2 B and C). These data suggest a greater functional requirement for neutrophils than macrophages in controlling A. baumannii infection.

Bacterial Determinants of Virulence in Mammals Are Conserved in Zebrafish.

Although a number of bacterial virulence mechanisms are host-specific (19), some are likely conserved across a range of higher-order species, including insects, fish, and mammals (20). To assess the predictive potential of zebrafish for A. baumannii virulence, we first assessed A. baumannii mutants with deletions in genes that encode a sensor histidine kinase (ΔgacS) and a response regulator (ΔgacA), both of which have displayed marked attenuation for virulence in a murine septicemia model (21). A. baumannii ΔgacS and ΔgacA were significantly attenuated for virulence compared with wild-type A. baumannii in a zebrafish bloodstream infection model (Fig. 3A). Complementation by expressing the full-length copies of the deleted genes restored virulence to wild-type A. baumannii levels (Fig. 3A). Moreover, neutrophil depletion using the csf3r morpholino restored the virulence capabilities of A. baumannii ΔgacS (Fig. 3B), suggesting that the virulence attenuation of this mutant was neutrophil-dependent.

Quorum-sensing is an evolutionary conserved signaling mechanism between cells within a population that facilitates the control of growth, nutrient acquisition, virulence, and other biological processes in bacteria (22, 23). Quorum sensing is yet to be established as a virulence factor in Acinetobacter. We tested an Acinetobacter M2 mutant with a deletion of the autoinducer synthase gene (abaI), which is necessary for 3-hydroxy-C₁₂-homoserine lactone production (24). Compared with wild-type Acinetobacter M2 infection, zebrafish infected with Acinetobacter ΔabaI had significantly greater survival (Fig. S2D). Virulence was partly restored with complementation. These findings were then corroborated in a mammalian septicemia model (Fig. S2E). Together, these data show that diverse intrinsic Acinetobacter virulence determinants are shared for infections in zebrafish and mammals, and support the use of zebrafish to study established, as well as new, mechanisms of Acinetobacter disease.

An A. baumannii Global Virulence Regulator Drives Extravascular Neutrophil Migration Patterns During Infection.

We have shown that neutrophils orchestrate a coordinated response to a localized A. baumannii infection including rapid migration to the site of infection, phagocytosis, and bacterial processing within phagolysosomes, followed by migration away. Thus far, infection model studies have concentrated on host-derived factors that determine neutrophil chemotaxis such as IL-8, leukotriene B4, C5a, and hydrogen peroxide (2, 10, 25, 26). Less is known about bacterial factors that directly drive neutrophil migratory patterns and trafficking. During localized tissue infections into the zebrafish somatic muscle, we observed a difference in neutrophil behavior with the A. baumannii ΔgacS mutant. Although the neutrophils migrated to the infection site and phagocytosed bacteria, they failed to migrate away as occurred in wild-type A. baumannii infection (Fig. 4A and Movie S4). This resulted in persisting clusters of neutrophils at the infection site until 48 hpi (Fig. 4B and Fig. S3A). Importantly, complementation of ΔgacS restored neutrophil migratory patterns to those of wild-type (Fig. 4B). As a control, other virulence-attenuating A. baumannii deletion mutants (ΔabaI and ΔargH) were tested but showed wild-type neutrophil migratory patterns (Fig. 4B), highlighting the specificity of this response to A. baumannii ΔgacS. To control for site of infection, we also infected zebrafish in the otic vesicle, which, under normal conditions, is devoid of neutrophils, but is another established site for studying a localized infection. Similar findings were observed (Fig. 4C and Fig. S3B).

To characterize neutrophil dynamics in more detail, we used time-lapse confocal microscopy and 3D cell tracking software to quantify neutrophil migratory parameters (Fig. 4D and Fig. S3C). Following infection with wild-type A. baumannii, neutrophils migrated to and from the site of infection consistently during the 5-h observation period (Fig. 4D). However, following infection with A. baumannii ΔgacS, there was intense neutrophil migration to the site of infection during the first 75 min, followed by minimal further neutrophil ingress or egress for the remainder of the observation period (Fig. 4D). To determine if this neutrophil response coincided with increased host-derived chemotactic stimuli, we quantified at 1 h and 12 h after tail muscle infection the expression of mRNA encoding the chemokine IL-8, which is an established host-derived chemotaxin in tissue infection (2), and leukotriene A4 (LTA4) hydrolase that converts LTA4 to LTB4 (27). The expression of both mRNAs increased with infection, but no difference was seen between wild-type and A. baumannii ΔgacS (Fig. S4A). Collectively, these data indicate the presence of a strong and persistent chemoattractant in A. baumannii ΔgacS infection driving vigorous neutrophil influx followed by migratory arrest, and opens the possibility that this is bacterially derived.

Neutrophil Dwelling in A. baumannii Infected Tissue Is Mediated by an Impaired Phenylacetic Acid Catabolic Pathway.

A. baumannii GacS is a global virulence regulator, controlling 674 genes across diverse functions (21). Interestingly, after deletion of A. baumannii gacS, the genes with greatest repression (as much as 200-fold) belonged to a single operon made up of 15 coding sequences known as the paa operon (21). The function of this operon is to degrade aromatic compounds such as PA, which is derived from phenylalanine, to form acetyl- and succinyl-CoA used in the TCA cycle (28). We therefore hypothesized that A. baumannii ΔgacS is altering neutrophil responses through repression of the paa operon, leading to accumulation of aromatic compounds and subsequent bacterial-driven chemoattraction.

To test whether the ΔgacS-mediated neutrophil response is caused by loss of the PA catabolic pathway, we induced a localized infection in the zebrafish otic vesicle with an early PA pathway A. baumannii mutant (ΔpaaA). As shown in Fig. 5A, neutrophil responses were the same as that seen for A. baumannii ΔgacS (Fig. 4C), with prolonged neutrophil dwelling at the site of infection compared with wild-type A. baumannii. The phenotype was restored to wild-type levels with complementation (ppaaA) (Fig. 5A). Consistent with there being neutrophil relocation and clustering in the otic vesicle and not local neutrophil proliferation, we documented that neutrophil aggregation in the otic vesicle during A. baumannii ΔpaaA infection was concurrent with reduction of neutrophils normally resident in the caudal hematopoietic tissue (Fig. 5B). Highlighting the specificity of this neutrophilic response to A. baumannii ΔpaaA, we infected zebrafish with another isogenic A. baumannii mutant with a deletion in a GacS-regulated gene csuD (21). No difference in neutrophil trafficking compared with wild-type was seen (Fig. S4B).

To determine the functional impact of this neutrophil dwelling response on bacterial survival, we assessed bacterial densities at 6 hpi for wild-type A. baumannii, ΔpaaA, and complement strains (all grow similarly in vitro). Significantly fewer bacteria were found for ΔpaaA (Fig. 5C), and this was associated with attenuated virulence for the ΔpaaA mutant compared with wild-type and complement strains (Fig. 5D). Together, these data indicate that the A. baumannii PA catabolic pathway is important in mediating neutrophil chemotaxis and tissue responses to acute infection, and may be a mechanism by which A. baumannii uses for immune evasion and disease progression.

PA Mediates Neutrophil Chemotaxis.

To confirm whether a metabolic by-product was accumulating in the tissues after infection with A. baumannii ΔpaaA, we first injected the otic vesicle of zebrafish with A. baumannii culture filtrate taken from stationary-phase in vitro growth (Fig. 5E). Neutrophil chemotaxis and dwelling occurred with supernatant taken from A. baumannii ΔpaaA, whereas neutrophil chemotaxis was less with supernatant from wild-type and complemented A. baumannii strains (Fig. 5E). Interestingly, injection of washed, heat-killed A. baumannii ΔpaaA caused neutrophil chemotaxis, but neutrophil dwelling was not observed (Fig. S4C), indicating that metabolism from live bacteria is required for the production of the causative factor.

To identify the molecule responsible for the observed neutrophil responses, we performed untargeted global metabolomics by MS on culture filtrate taken from wild-type A. baumannii, ΔpaaA and the complemented strain. In total, more than 1,600 metabolite features were detected, and the only metabolite that showed a significant accumulation (more than twofold; P < 0.05) in A. baumannii ΔpaaA compared with the wild-type and complemented strains was putatively identified as PA (Fig. S5 and Dataset S1). Subsequent identification and accurate quantification of PA by using targeted liquid chromatography (LC)/MS confirmed fourfold accumulation of PA in ΔpaaA supernatant compared with wild-type (Fig. 5F). PA is an intermediate product of the bacterial phenylalanine metabolic pathway, and a direct substrate of paaA. To confirm that PA was the causative driver of the observed neutrophil tissue responses, we injected purified PA (Sigma) into the otic vesicle of zebrafish at varying concentrations. This showed a bell-curve dose–response relationship characteristic of chemoattractants (29), with a concentration as low as 3.4 ng per embryo having an effect, and peaking at 37 ng per embryo (Fig. S6A). Neutrophil migration was seen within 2 hpi (Fig. 5G). When purified PA was added to wild-type culture filtrate and injected into the otic vesicle of zebrafish, we observed the same neutrophil trafficking characteristics as observed for infection with A. baumannii ΔpaaA (Fig. S6B). Furthermore, the addition of purified PA to wild-type A. baumannii cells before infection led to greater embryo survival compared with infection with wild-type A. baumannii alone (Fig. S6C). Finally, to determine if PA was acting directly on neutrophils rather than through stimulation of endogenous cytokines, and to assess whether this effect was observed with mammalian neutrophils, we performed ex vivo neutrophil migration assays by using murine bone marrow-derived neutrophils. Purified PA caused significantly greater neutrophil transmigration compared with the control (Fig. 5H). Together, these data confirm that, in the absence of a functional PA catabolic pathway, PA accumulates with live bacterial cells and then acts as a bacterial-driven neutrophil chemoattractant.

Bacterial Strains and Growth Conditions.

A. baumannii strains (Table S1) were grown in LB liquid or solid media at 37 °C supplemented with carbenicillin (150 μg/mL) or kanamycin (50 μg/mL) as appropriate. GFP expression and complementation of Acinetobacter M2 ΔabaI and ATCC 17978ΔpaaA were performed by using pWH1266 as described elsewhere (38). Primers used for this study are shown in Table S2.

Zebrafish Strains and Maintenance.

Wild-type Tübingen, Tg(lyz:DsRed)^nz50 (39), Tg(mpeg1:mCherry)^gl23 (40), Tg(mpx:GFP)^il14 (41), and Tg(mpeg1:Gal4FF)^gl25 × Tg(UAS-E1b:Eco.NsfB-mCherry)^c264 zebrafish embryos were maintained in the Monash University FishCore facility according to standard protocols (SI Materials and Methods).

Microinjection of A. baumannii into Zebrafish Embryos.

Overnight bacterial cultures were adjusted to OD₆₀₀ of 1, washed, and resuspended in 2% (wt/vol) polyvinylpyrrolidone in PBS solution. Microinjections were performed at 48 hpf, and all inocula were confirmed by colony counts (42, 43). Embryos were monitored daily for survival to 96 hpi, and survival analyses were performed by using the Kaplan–Meier method. The Mantel–Cox test was used for statistical significance (P < 0.05). A range of 500–800 cfu of bacteria were injected into the muscle or otic vesicle for neutrophil and macrophage migration assays. Methods for in vivo A. baumannii growth kinetics and quantitative real-time PCR are shown in SI Materials and Methods.

Morpholino Depletion of cfs3r and irf8.

Antisense morpholino oligomers (MOs) directed at csf3r (30) and irf8 (18) were purchased from Gene Tools. Volumes of 1 nL were microinjected into the yolk of one-cell embryos, delivering MOs at the following final concentration in distilled water: 25 μM of csf3r^ATG MO and 25 μM of irf8^ATG MO. A standard control MO was used as a negative control.

Leukocyte Enumeration.

Leukocyte units (LUs), a surrogate parameter proportional to leukocyte numbers determined by analysis of digital images, were computed as previously described and validated (43). When appropriate, LUs are referred to as “neutrophil units” or “macrophage units.” Leukocyte numbers were referred to when the total number of neutrophils was counted. A Mann–Whitney test was used for statistical significance (P < 0.05).

Microscopy and Image Processing.

For microscopy, zebrafish were immobilized in 1% low melting temperature agar. Routine bright-field and fluorescence imaging was performed by using a Zeiss Lumar V12 stereo dissecting microscope with an AxioCam MRm camera running AxioVision 4.8 software. Confocal microscopy was performed with a Zeiss LSM 710 device with a Plan-Apochromat 20×, 0.8 N.A. objective, and Zen software (v4.0) (SI Materials and Methods). Images were processed in Imaris (Bitplane) for cell-tracking mode, and data were analyzed in the R program using ggplot2 (26). Vibratome sectioning, histopathology, and pHrodo staining are described in SI Materials and Methods.

Murine Disseminated Infection Model.

The model was performed as described previously (21) using female BALB/c mice aged 6–8 wk (n = 15 per group) infected via intraperitoneal injection (SI Materials and Methods). Animal protocols were approved by the Monash University Animal Ethics Committee.

Metabolomics.

Metabolite samples were prepared from cell culture supernatants by solvent extraction and analyzed by LC/MS with hydrophilic interaction liquid chromotography (HILIC) and high-resolution accurate MS (44) (SI Materials and Methods). Metabolomics data were analyzed with the IDEOM software (45), and PA abundance confirmed with Tracefinder (Thermo). Quantification of PA was confirmed by using a complementary analytical method, reversed-phase chromatography coupled to high-resolution MS (Q-Exactive; Thermo; SI Materials and Methods).

Ex Vivo Neutrophil Chemotaxis Assay.

Neutrophils were isolated from the bone marrow of C57BL/6 mice by Histopaque density gradient centrifugation, and chemotaxis was performed as described previously (46) (SI Materials and Methods).

Zebrafish Strains and Maintenance.

Embryos were grown in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgSO₄, equilibrated to pH 7.0) at 28 °C. For microinjection and imaging, embryos were anesthetized in E3 medium containing 0.02% Tricaine (ethyl 3-aminobenzoate). Embryos for microscopy were also maintained in E3 medium containing 0.003% (wt/vol) 1-phenyl 2-thiourea from 12 hpf to prevent pigmentation.

In Vivo Acinetobacter baumannii Growth Kinetics.

Zebrafish embryos (48 hpf) were infected with A. baumannii via the circulation with a lower inoculum (1 × 10³ cfu per embryo) and at 0, 2, 8, 20, 32, 44, 56, and 78 hpi. Five embryos were randomly selected for cfu enumeration. Embryos were pooled and homogenized, serially diluted, and plated onto solid LB media with carbenicillin (150 μg/mL) antibiotics for calculation of cfu per embryo. Three independent experiments were performed.

Quantitative RT-PCR.

RNA was isolated from 4 hpi embryos by using TRIzol Reagent (Life Technologies) following the manufacturer’s instructions. Purified RNA was subjected to on-column treatment with RNase-free DNase I (Qiagen) according to the manufacturer’s instructions. A total of 2 μg of RNA was used in a reverse transcription reaction by using SuperScript III (Life Technologies) and random decamers per the manufacturer’s instructions. The resulting cDNA was diluted three times for use in quantitative real-time PCR. Each 20-μL reaction mixture contained 2.5 μL of cDNA, 50-nM concentrations of each gene-specific primer (listed in Table S2), and 10 μL of SYBR Green PCR Master Mix (ABI). Real-time PCR was performed on a BioRad thermocycler. Known concentrations of A. baumannii ATCC 17978 genomic DNA were used to construct a gene-specific standard curve so that the concentration of template in each reaction mixture could be determined. Quantitation was performed on triplicate wells. All reactions were normalized against β-actin, and melting curve analysis confirmed the presence of only one PCR product. Significance was determined by using a t test with a significance threshold of P ≤ 0.05.

Microscopy and Image Processing.

Confocal microscopy was performed with a Zeiss LSM 710 with a Plan-Apochromat 20×, 0.8 N.A. objective running Zen software (v4.0). Images were 12-bit, 512 × 512 pixels. Z-depth ranged from 48 to 58 slices with Z-intervals optimized for 1:1:1 X:Y:Z reconstruction. Time intervals were optimized for each experiment and ranged from 10 to 30 s. Excitation laser wavelengths were 489 nm for EGFP and 561 nm for mCherry and DsRed. Emission detection used a BP495-555 filter for EGFP emission and a LP575 filter for mCherry and DsRed emission.

Vibratome Sectioning and Histopathology.

Embryos were fixed in 2% paraformaldehyde overnight at 4 °C, washed in PBS solution, and embedded in 4% low-melting-point agarose/E3 medium. When it had become solid, the agarose block was mounted onto a Vibratome chuck with a cyanoacrylate adhesive. The holding chamber was filled with PBS solution, and 250-μm transverse sections were cut by using a Leica V1200S Vibratome. For visualization of the nuclei, the sections were stained with NucBlue Live Ready Probes reagent (Life Technologies) per the manufacturer’s instructions. Sections were imaged by using a Leica Sp5 confocal laser scanning microscope.

pHrodo Staining.

Overnight cultures of A. baumannii were washed twice in PBS solution, suspended in 0.08% paraformaldehyde/PBS solution, and incubated at 37 °C for 30 min. Bacteria were then transferred to 2–8 °C for 15–20 min. The bacteria were washed twice in PBS solution and resuspended in carbonate buffer (pH 8.4). pHrodo (Invitrogen) was added at final concentration of 0.1 mM and incubated at room temperature for 45 min in the dark. The bacteria were washed once in 750 μL of HBSS before being resuspended in HBSS. The stained cells were stored at 4 °C in the dark until required.

Murine Disseminated Infection Model.

Overnight cultures of M2 W/T, M2ΔabaI, and the complemented strain (M2 pabaI) were prepared with an equal volume of 10% porcine mucin. Female BALB/c mice aged 6–8 wk (n = 15 per group) were infected via intraperitoneal injection with 5 × 10⁴ cfu of wild-type or mutant bacteria. Mice were monitored for clinical signs of illness and euthanized when necessary. Time to euthanasia was assessed by Kaplan–Meier method, and the Mantel–Cox test was used to determine statistical significance (P < 0.05).

Metabolomics.

Culture supernatants from overnight bacterial cultures in LB media from four independent biological replicates were prepared by membrane filtration (0.22 µm) and stored at −80 °C. Before analysis, samples were thawed, and 10-µL sample was extracted with 250 µL extraction solvent (chloroform/methanol/water; 1:3:1). Samples were mixed thoroughly by vortex and centrifuged at 15,000 × g for 10 min to remove particulates before analysis by LC/MS.

Metabolomics samples were analyzed by HILIC chromatography coupled to high-resolution MS based on the pHILIC-AC method (44). The chromatography used a Dionex RSLC U3000 HPLC system (Thermo) with a 150 × 4.6 mm × 5 µm ZIC-pHILIC column (Merck). The mobile phase consisted of 20 mM ammonium carbonate (A) and acetonitrile (B), with a gradient from 80% B to 50% B (15 min), wash at 5% B (18–21 min) and reequilibration at 80% B (24–32 min). Mass spectra were collected on a Q-Exactive Orbitrap mass spectrometer operating at high resolution (35,000) with positive and negative mode electrospray (rapid switching). Source conditions were sheath gas 25, auxiliary gas 12, spray voltage 3.5 kV, probe temperature 150 °C, capillary temperature 320 °C, and S-lens RF level 55 eV. All samples were analyzed in a single batch with pooled quality control samples throughout the batch, and four external standard compounds were included in all samples. Standards mixtures containing more than 200 authentic metabolites were analyzed with the batch to aid metabolite identification.

Metabolomics data were analyzed with the open-source IDEOM software. Briefly, raw data were converted to mzXML format and peaks extracted with XCMS. Samples were aligned and filtered with mzMatch and saved as peakML format, and peak data imported to the IDEOM Excel interface. Data were filtered with a relative standard deviation (RSD) threshold of 30%, intensity threshold of 100,000, and the default IDEOM (version 20) settings for other filters. Metabolites were putatively identified based on exact mass and retention time by using a 2-ppm mass window after data-dependent recalibration and retention times from authentic standards where available, or predicted retention times for other metabolites. Annotations from the EcoCyc database were used to prioritize known bacterial metabolites among isomers. Metabolite identifications are based on this automated process and can be considered level-2 identifications according to the Metabolomics Standards Initiative. Semiquantitative analysis based on LC/MS peak height was performed between the mutant strain and each of the wild-type and complemented strains by using Welch’s t test (α = 0.05) for all putative metabolites. Relative quantification of PA was confirmed by manual integration of peak areas with TraceFinder software (Thermo Fisher Scientific).

Accurate quantification and identification of PA in culture supernatants was confirmed by comparison with authentic standards by using a complementary analytical method, reversed-phase chromatography (Kinetex EVO C18 2.6 µm, 50 × 3 mm; Phenomenex) coupled to high-resolution MS (Q-Exactive; Thermo Fisher Scientific). The mobile phase consisted of 0.01% formic acid (A) and methanol (B), starting at 30% B from 0 to 0.5 min, to 70% (0.6 min), 90% (2 min), 90% (3.1 min), 30% (3.2 min), and 30% (5 min). The flow rates were 0.4 mL/min from 0 to 2.3 min and 0.6 mL/min from 2.3 to 5 min. Accurate quantification of PA (m/z 135.0452, retention time = 1.97 min) was achieved by using an external calibration curve with ¹³C₂-PA (m/z 137.0519) as the internal standard.

Ex Vivo Neutrophil Chemotaxis Assay.

Briefly, bone marrow was collected and resuspended in chemotaxis buffer (in 50 mL of buffer: 23.7 mL of RPMI medium, 23.7 mL HBSS, 2% FBS, 2% glutamine, and 1% penicillin/streptomycin), and red blood cells were lysed with lysis buffer. Migration assays were performed in 96-well plates (Corning) with a known concentration of PA (Sigma, catalog no. P16621) in the lower well, and neutrophils were added to the insert, which goes on top of the chemotaxis buffer. Neutrophils migrate to the bottom of the well in response to a direct chemotactic stimulation. A chemotaxis index was calculated based on the number of neutrophils that migrate to the test well divided by the number of neutrophils that migrate to the control well, which contains media only. At least five independent experiments were performed.