All chemicals and reagents were reagent grade or better. HPLC-grade solvents, DMEM, and FBS were and obtained from Millipore-Sigma or Fisher unless otherwise indicated. [1-¹⁴C]oleic acid (18:1; 59 mCi/mmol, 0.1 mCi/ml) and obtained from PerkinElmer (Waltham, MA), 10-hydroxyoctadecanoic acid (h18:0) was obtained from AA BLOCKS, INC (San Diego, CA), palmitic acid (16:0), stearic acid (18:0), and 18:1 were obtained from Matreya, LLC (State College, PA), 9-octadecadienoic acid (9-HODE) and 13-octadecadienoic acid (13-HODE) were obtained from Cayman Chemical (Ann Arbor, MI). [1-¹⁴C]h18:0 was prepared using purified oleate hydratase as described previously (Radka et al., 2021b). GW7647, GW0742, and rosiglitazone were provided in the Peroxisome Proliferator-Activated Receptors (PPAR) Panel assay from INDIGO Biosciences, Inc. (State College, PA).

C57BL/6 (Strain 000664) and Ppara ^−/− knockout (Strain 008154) mice were obtained from Jackson Laboratory (Farmington, CT). Bone marrow was isolated from 8-10 week old C57BL/6 and Ppara ^−/− knockout mice by the St. Jude Children’s Research Hospital Animal Resources Center, and macrophages were derived from the bone marrow using the method established by Toda et al. (2021). Cells (4.2 x 10⁷) were seeded in 6-well tissue culture plates and grown in DMEM and 10% fetal bovine serum with 1 ng/ml CSF-1 (ThermoFisher) at 37°C 5% CO₂. On day 6, bone marrow-derived macrophages (BMDM) were washed with phosphate buffered saline and then changed to media containing 1% dimethyl sulfoxide (DMSO) ± fatty acid treatment and incubated 20-24 h. RAW 264.7 mouse macrophage cells (ATCC number TIB-71) were grown in DMEM plus 10% fetal bovine serum containing 1% DMSO ± fatty acid at 37°C 5% CO₂.

The RNEasy kit (Qiagen) was used to isolate RNA from BMDM cells incubated 20 h in DMEM 1% DMSO ± 20 μM h18:0, and the TRIzol reagent (ThermoFisher) was used to isolate RNA from RAW 264.7 cells incubated 20 h in DMEM 1% DMSO ± 100 μM h18:0. RNA was quantified using the Quant-iT RiboGreen RNA assay (ThermoFisher) and quality checked by the 2100 Bioanalyzer RNA 6000 Nano assay (Agilent) or 4200 TapeStation High Sensitivity RNA ScreenTape assay (Agilent) prior to library generation. Libraries were prepared from total RNA with the TruSeq Stranded Total RNA Library Prep Kit according to the manufacturer’s instructions (Illumina, PN 20020599). Libraries were analyzed for insert size distribution using the 2100 BioAnalyzer High Sensitivity kit (Agilent), 4200 TapeStation D1000 ScreenTape assay (Agilent), or 5300 Fragment Analyzer NGS fragment kit (Agilent). Libraries were quantified using the Quant-iT PicoGreen ds DNA assay (ThermoFisher) or by low pass sequencing with a MiSeq nano kit (Illumina). Paired end 100 cycle sequencing was performed on a NovaSeq 6000 (Illumina).

Read alignment to the M22 release of the GRCm38 mm10 Mus musculus reference genome was performed with STAR version 2.7 (Dobin et al., 2013). Gene level quantification was determined using RSEM version 1.3.1 (Li and Dewey, 2011). The top 3000 most variable (and informative) genes from each data set were used to perform principal component analysis using the prcomp function, available as part of the standard R distribution. Two filtration steps were taken to prepare data for analysis: only protein coding genes were used/non-coding genes were excluded, and genes with very low expression values were removed. Only genes that passed raw read count > median.library.size/10⁶ (i.e., CPM>0.1) [default cutoff] in the smallest sample-size group were included in pairwise comparisons. Differential gene expression was modeled using the voom method, which is available in the limma R software package (Law et al., 2014). Differentially expressed genes (DEGs) were defined by P value < 0.05 or FDR < 0.05. The top DEGs are defined by the cutoff of FDR < 0.05 and ranked by the change in counts per million in descending order. Gene set enrichment analysis was done using the Reactome signatures from the Molecular Signatures Database (https://www.gsea-msigdb.org/gsea/msigdb) release 7.5.

Fatty acids were investigated with commercially available assays for human PPARα, PPARβ/δ, and PPARγ (INDIGO Biosciences, Inc., State College, PA). PPAR assay kits contain proprietary cells, media, control compounds, luminescence detection reagents, and 96-well plate. Assays were performed according to the manufacturer protocols. Briefly, PPAR reporter cells were seeded and incubated 4 h at 37°C 5% CO₂ in cell recovery medium. The medium was exchanged for compound screening medium containing 0.4% DMSO only or 10 μM fatty acid (16:0, 18:0, 18:1, h18:0, 9-HODE, or 13-HODE) in 0.4% DMSO. 10 nM GW7647, 3.3 nM GW0742, or 300 nM rosiglitazone in 0.4% DMSO were used as control agonists for PPARα, PPARβ/δ, and PPARγ respectively. A fatty acid titration using PPARα cells was performed using 0, 0.08, 0.16, 0.31, 0.63, 1.25, 2.5, 5, and 10 μM 16:0, 18:0, 18:1, or h18:0. Cells incubated 20 h at 37°C 5% CO₂, then the medium was removed, the luminescence detection reagents were added, and luminescence was quantified using a BioTek microplate reader (Agilent Technologies, Inc; Santa Clara, CA).

BMDMs isolated from wildtype C57BL/6 mice were incubated with 1 μM [1-¹⁴C]18:1 with 9 μM 18:1, or 1 μM [1-¹⁴C]h18:0 with 9 μM h18:0 in DMEM 1% DMSO. After overnight incubation at 37°C 5% CO₂, the cells were harvested and washed twice in 1 ml of 10 mg/ml fatty acid free bovine serum albumin in phosphate-buffered saline, then the lipids were extracted using the Bligh and Dyer method (Bligh and Dyer, 1959) from each experiment. Equal amounts of radioactivity (50,000 cpm/sample) of BMDM lipids labeled with [1-¹⁴C]18:1 or [1-¹⁴C]h18:0 were spotted on a Silica Gel H plate (Analtech, Inc., Newark, DE) and polar lipids were developed in chloroform:methanol:ammonium hydroxide (60/35/8, v/v/v). Fatty acid methyl esters were prepared from 50,000 cpm of each sample as described previously (Radka et al., 2020), and the entire preparation was spotted on a Silica Gel H plate (Analtech, Inc., Newark, DE) and developed in hexane:ether:acetic acid (80/20/1, v/v/v). The labeled lipids were visualized with a PhosphorImager system.

BMDMs isolated from C57BL/6 or Ppara ^−/− knockout mice or RAW 264.7 cells were incubated for 24 h ± 10 μM h18:0. The cell culture media was collected and the lipids were extracted using the Bligh and Dyer method (Bligh and Dyer, 1959). The lipid extracts were dried under N₂ and resuspended in chloroform:methanol (1:1). The fatty acid fraction was isolated using a Discovery DSC-NH₂ solid-phase extraction column (Supelco, Bellefonte, PA) as described previously (Radka et al., 2020). The free fatty acids were converted to their 3-picolylamide derivatives (Radka et al., 2021a; Li and Franke, 2011) for LC-MS/MS analysis. The amount of each hFA was calculated based on the known amount of [U-¹³C]18:1 spiked into the sample at the beginning of the sample preparation (Cambridge Isotopes Labs). Picolylamide-hFA were analyzed using a Shimadzu Prominence UFLC attached to a QTrap 4500 equipped with a Turbo V ion source (Sciex). Samples (5 µL) were injected onto an XSelect HSS C18, 2.5 μm, 3.0 x 150 mm column (Waters) at 45°C with a flow rate of 0.4 ml/min. Solvent A was 0.1% formic acid in water, and solvent B was acetonitrile with 0.1% formic acid. The HPLC program was the following: starting solvent mixture of 70% A/30% B; 0 to 5 min, isocratic with 30% B; 5 to 15 min, linear gradient to 100% B; 15 to 23 min, isocratic with 100% B; 23 to 25 min, linear gradient to 30% B; 25 to 30 min, isocratic with 30% B. The QTrap 4500 was operated in the positive mode, and the ion source parameters for the picolylamide-hFA multiple reaction monitoring (MRM) parameters were: ion spray voltage, 5,500 V; curtain gas, 15 psi; temperature, 300°C; ion source gas 1, 15 psi; ion source gas 2, 20 psi; declustering potential, 25 V, and a collision energy, 40 V. MRM masses (Q1/Q3) were: picolylamide-h12:0, 307.1/109.0; picolylamide-h14:0, 335.1/109.0; picolylamide-h16:0, 363.1/109.0; picolylamide-h18:0, 391.1/109.0, and picolylamide-[U-¹³C]18:1, 391.1/109.0. The system was controlled by the Analyst software (Sciex) and analyzed with MultiQuant™ 3.0.2 software (Sciex).

The SSTI thigh infection model used 8-12-week-old C57BL/6 (Jackson Laboratory strain 000664) or C57BL/6 Ppara ^-/- (Jackson Laboratory strain 008154) male and female mice. The S. aureus clinical MRSA strain AH1263 was grown in Luria broth overnight with aeration at 37°C and back diluted in fresh medium to grow to an OD_{600 =} 0.4. AH1263 cells were washed twice in phosphate buffered saline and then resuspended in fresh phosphate buffered saline to obtain an inoculum of ~2.5 – 3.5 x 10⁶ CFU/50 μl. AH1263 cells were introduced by 50 μl intramuscular injections, and 11 infections (C57BL/6 or C57BL/6 Ppara ^-/-) were allowed to proceed for 24 hours. The infected tissue were excised and homogenized in 1 ml of ice-cold phosphate buffered saline and the bacterial burden was determined by serial diluting and selective plating using mannitol salt. For further analysis, the homogenized samples were spun down and the supernatant was removed. The cytokines in the supernatant were analyzed by uncoated TNF-α and uncoated IL-6 ELISA kits. The supernatant was diluted 1:4 in phosphate buffered saline and the IL-6 abundance was measured according to the manufacturer’s instructions for the IL-6 ELISA kit (Invitrogen Cat. No. 88-7064). TNF-α was measured in undiluted supernatant according to the manufacturer’s instructions for the TNF-α ELISA kit (Invitrogen Cat. No. 88-7324).

Data analyses were performed using Prism version 9 (GraphPad Software, Boston, MA).

All animal experiments were performed with prior review and approval by the St. Jude Institutional Animal Care and Use Committee. All mice were maintained in biosafety level 2 facilities in accordance with IACUC protocol number 566-100442-10/16.

All RNA-seq reads generated and used in this study were deposited in the NCBI Sequence Read Archive under BioProject PRJNA1002616.

We used RNAseq to determine if the signaling function of h18:0 elicited changes in macrophage gene expression. Both the RAW 264.7 macrophage cell line and primary bone marrow-derived macrophages (BMDM) isolated from C57BL/6 mice were incubated either with or without h18:0 for 20 h, the cellular RNA isolated, and RNAseq performed using Illumina NovaSeq. Approximately 90% of the reads mapped to the Mus musculus genome in all cases. The reactome classification in the Molecular Signatures Database (MSigDB; https://www.gsea-msigdb.org/gsea/msigdb) for gene set enrichment analysis was used to classify the differentially expressed genes into functional pathways. This analysis showed that genes related to the metabolism of lipids were the most highly up regulated group of genes in the dataset in both RAW cells ( Figure 1A ; q = 0.007) and BMDM ( Figure 1B ; q = 0.002). Genes associated with the proliferation of peroxisomes were significantly up regulated in both the RAW cell line ( Figure 1C ) and BMDM ( Figure 1D ). A heat plot of the salient lipid metabolic genes associated with PPAR regulation in the RAW cell line and BMDM highlight the similarities in the genetic response to h18:0 ( Figure 1E ). These data suggest that h18:0 activates one of the PPAR transcriptional regulators, although it is not possible to definitively ascribe the changes elicited by h18:0 to a specific PPAR isoform because they regulate a highly overlapping set of genes (Bugge and Mandrup, 2010; Zhang et al., 2015).

We employed a set of three cell lines each with an engineered luminescence reporter gene to detect the ability of ligands to specifically activate each PPAR isoform (Velazquez et al., 2022). Ligands were delivered in DMSO and DMSO alone or with saturating amounts of PPAR-specific ligands that were used as controls to define the assay baseline and the maximal response ( Figure 2 ). PPARα was potently activated by h18:0 with a response comparable to the maximum signal in the assay ( Figure 2A ). Both 9-HODE and 13-HODE also potently activated PPARα. PPARβ/δ was not activated by any of the hydroxy fatty acids ( Figure 2B ). We measured a modest, insignificant PPARγ response to h18:0, 9-HODE and 13-HODE compared to non-hydroxy fatty acids ( Figure 2C ). Although previous reports indicate that PPARγ interacts with 9- and 13-HODE (Itoh et al., 2008; Nagy et al., 1998; Yokoi et al., 2009), in our comparative experiments, the hydroxy fatty acids did not activate PPARγ like they did PPARα. Dose response experiments confirmed the ability of h18:0 to activate PPARα-driven transcription ( Figure 2D ). These data point to PPARα as a target for OhyA-derived metabolites.

The metabolism of h18:0 by macrophages was investigated by tracing the fate of radiolabeled [1-¹⁴C]h18:0 prepared enzymatically using S. aureus OhyA to hydroxylate [1-¹⁴C]18:1 (Radka et al., 2021b). BMDM isolated from wildtype C57BL/6 mice were incubated for 24 h with [1-¹⁴C]18:1 or [1-¹⁴C]h18:0, and the cellular and media lipids were extracted and quantified ( Figure 3A ). Virtually all of the [1-¹⁴C]18:1 was incorporated into the cells with very little radiolabel remaining in the media. In contrast, the total amount of [1-¹⁴C]h18:0 recovered was less than half of the [1-¹⁴C]18:1 recovered, and the label was split between the cells and the media indicating incomplete BMDM catabolism of [1-¹⁴C]h18:0 ( Figure 3A ). The lower overall recovery of label derived from [1-¹⁴C]h18:0 indicated that the carbon-14 from h18:0 was being converted to non-lipid metabolites.

We followed our bulk analysis of total fatty acid mass balance with an analytical analysis of fatty acid utilization by examining which lipids were labeled using equal counts of radioactivity recovered from the cellular lipids in both sample sets. The distribution of the two labeled fatty acids in the cellular lipids was determined by thin-layer chromatography ( Figure 3B ). Both [1-¹⁴C]18:1 and [1-¹⁴C]h18:0 labeled the same major cellular phospholipids: phosphatidylcholine, phosphatidylethanolamine and sphingomyelin ( Figure 3B ). Because hydroxy fatty acids are not normally found in mammalian phospholipids, we took the additional step of converting the acyl chains of the cellular lipids to fatty acid methyl esters to determine if [1-¹⁴C]h18:0 was directly incorporated into lipids ( Figure 3C ). The [1-¹⁴C] atoms from [1-¹⁴C]18:1 remained as fatty acid as expected. However, the methyl esters derived from [1-¹⁴C]h18:0 were not hydroxylated indicating that these fatty acids were derived from the degradation of [1-¹⁴C]h18:0 to [1-¹⁴C]acetyl-CoA, which was subsequently incorporated into fatty acids by the macrophage fatty acid biosynthetic pathway. The mass balance of [1-¹⁴C]h18:0 indicates half of the fatty acid disappears from the lipophilic fraction, and the analytical analysis demonstrates h18:0 was not used directly for lipid synthesis, but rather degraded and the acetyl-CoA products were recycled to build new acyl chains as shown by the conversion of [1-¹⁴C]h18:0 to [¹⁴C]18:1.

RAW 264.7 macrophages were incubated for 24 h with 10 μM h18:0, the media was extracted and the fatty acids derivatized to determine the composition of hydroxy fatty acids in the media (Radka et al., 2021a). Neither h18:0 nor its shortened products were detected in RAW macrophages confirming that hydroxylated fatty acids were below detection in untreated cultured cells ( Figure 4A ). However, in the presence of h18:0, hydroxyhexadecanoic acid (h16:0), hydroxytetradecanoic acid (h14:0), and hydroxydodecanoic acid (h12:0) were detected ( Figure 4A ). These shorter hydroxy fatty acids were interpreted as intermediates in the β-oxidation of h18:0 each shortened by one round of β-oxidation.

We determined if the appearance of these truncated hydroxy fatty acids was PPARα dependent by comparing their abundance in BMDM isolated from C57BL/6 or Ppara ^−/− mice exposed to 10 μM h18:0 for 24 h ( Figure 4B ). Media samples were analyzed using LC-MS/MS with [U-¹³C]18:1 as the internal standard to quantify the amounts of hydroxy metabolites present ( Figure 4C ). There was low recovery of h18:0 and its metabolites from wild-type BMDM consistent with the robust oxidation of h18:0 by primary macrophages. Any modest differences in truncated hydroxy fatty acid distributions are likely related to differences in h18:0 catabolism efficiencies between the RAW 264.7 macrophage cell line and primary BMDMs. In contrast, the recovery of h18:0 was over an order of magnitude higher in BMDM from Ppara ^−/− mice indicating that these cells are deficient in the degradation of h18:0. The absence of PPARα exacerbates the inefficient catabolism of h18:0 ( Figure 3A ) as shown by the accumulation of truncated hydroxy fatty acids ( Figure 4C ). These data are consistent with the activation of PPARα by h18:0 activating a PPARα-dependent transcriptional program to accelerate the oxidation and degradation of the h18:0 signal.

The impact of PPARα on S. aureus virulence was assessed in an immunocompetent skin/soft tissue (SSTI) infection model using C57BL/6 and C57BL/6 Ppara^-/- mice. The SSTI infection model showed S. aureus established an infection in wildtype C57BL/6 mice, but the bacterial burden in the Ppara ^−/− mice was 74-fold lower ( Figure 5A ). The same inoculum was used to infect C57BL/6 and Ppara^-/- mice, and the differences in the titers observed could be attributed to both pathogen outgrowth (or lack thereof) as well as innate immune clearance. We measured the abundance of TNF-α and IL-6 in the thigh muscle tissue and found similar levels of both cytokines in C57BL/6 and Ppara^-/- infections ( Figures 5B, C ). C57BL/6 mice have a genetic predisposition that confers greater resistance to S. aureus infection in terms of control of bacterial growth and survival compared to other mouse strains (von Kockritz-Blickwede et al., 2008). Enhanced neutrophil recruitment to the infection site is proposed to be a major component of the superior C57BL/6 innate immune response and resistance to staphylococcal infection (von Kockritz-Blickwede et al., 2008). These data show that PPARα signaling is also an important component, playing a key role in promoting S. aureus survival in the host and linking bacterial unsaturated fatty acid hydroxylation to the initiation of a specific mammalian signal transduction pathway. The staphylococcal infection phenotype shows the robust impact of PPARα on the innate immune response, even in a resistant mouse strain.