Oxysterol binding protein regulates the resolution of TLR-induced cytokine production in macrophages

Significance

Our study has revealed an unexpected role for the lipid transport molecule Osbp in modulating inflammation. Altering Osbp activity by binding to 25-hydroxycholesterol, an oxysterol which is produced by macrophages in response to microbial components, delays the resolution of inflammation and the restoration of immune homeostasis by increasing the production of triglyceride species. The innate immune response is a double-edged sword: It is absolutely required for host defense but unregulated and causes inflammatory disease. While the mechanisms that initiate inflammation are relatively well studied, the resolution phase has received much less attention. A more complete understanding of the mechanisms that restrain inflammation will aid efforts to develop interventions that blunt immune mediated damage without disabling host defense.

Abstract

Toll-like receptors (TLRs) on macrophages sense microbial components and trigger the production of numerous cytokines and chemokines that mediate the inflammatory response to infection. Although many of the components required for the activation of the TLR pathway have been identified, the mechanisms that appropriately regulate the magnitude and duration of the response and ultimately restore homeostasis are less well understood. Furthermore, a growing body of work indicates that TLR signaling reciprocally interacts with other fundamental cellular processes, including lipid metabolism but only a few specific molecular links between immune signaling and the macrophage lipidome have been studied in detail. Oxysterol-binding protein (Osbp) is the founding member of a family of lipid-binding proteins with diverse functions in lipid sensing, lipid transport, and cell signaling but its role in TLR responses is not well defined. Here, we demonstrate that altering the state of Osbp with its natural ligand, 25-hydroxycholesterol (25HC), or pharmacologically, sustains and thereby amplifies Tlr4-induced cytokine production in vitro and in vivo. CRISPR-induced knockdown of Osbp abrogates the ability of these ligands to sustain TLR responses. Lipidomic analysis suggested that the effect of Osbp on TLR signaling may be mediated by alterations in triglyceride production and treating cells with a Dgat1 inhibitor, which blocks triglyceride production and completely abrogates the effect of Osbp on TLR signaling. Thus, Osbp is a sterol sensor that transduces perturbations of the lipidome to modulate the resolution of macrophage inflammatory responses.

Toll-like receptors (TLRs) on macrophages sense conserved pathogen-associated molecular patterns and initiate innate immune responses. It is increasingly appreciated that the metabolic state of the macrophage when the pathogen is encountered controls the amplitude and duration of the response. Lipid metabolic pathways intersect with TLR signaling (1–5); however, only a limited number of specific mechanisms have been identified and thoroughly examined. The, 25 hydroxycholesterol (25HC) is formed from cholesterol by the enzyme cholesterol 25 hydroxylase (Ch25h), whose expression is strongly upregulated by type I interferons (6–8). This oxysterol, whose endogenous production is increased by TLR activation, has diverse effects on immune responses (6, 7, 9–13). In previous work, we found that 25HC amplifies Tlr3 responses in macrophages and exacerbates the pathology of influenza infection, (6) however, the mechanism for this effect remained unclear. 25HC is known to bind or directly regulate several molecules including the liver X receptor (Lxr) and the sterol regulatory element binding protein (Srebp) transcription factors and the lipid transport molecule oxysterol binding protein (Osbp) (14–16). While both Lxr and Srebp have previously been shown to regulate TLR signaling pathways in macrophages (1–3, 17), to the best of our knowledge, Osbp has not previously been described to have a direct role in TLR signaling.

Osbp is a multifunctional sterol sensing molecule whose activity and localization are regulated by the relative concentration of cholesterol and oxysterols, both of which are strongly affected by inflammation. Osbp is recruited to ER-trans Golgi network membraned contact sites by the binding of oxysterols such as 25HC or in conditions of cholesterol depletion. This results in cholesterol transport into the Golgi which alters downstream lipid metabolic processes and modulates the cellular lipidome (15, 18–23). Osbp has been implicated in several aspects of host–pathogen interactions. For instance, Osbp is recruited to replication organelles by several viruses which utilize its cholesterol exchange function to facilitate their life cycle, and drugs targeting Osbp have been shown to be effective antivirals (24). Additionally, Osbp exchanges cholesterol and PI4P at several membrane contact sites including contact sites between the ER and endosomes and excessive PI4P at these sites has been shown to drive assembly of the Nlrp3 inflammasome (25). In cells lacking Osbp, PI4P builds up at ER-endosome contact sites (EECS) leading to spontaneous assembly of Nlrp3 inflammasomes which enables LPS-induced Il1b production in the absence of a second cytoplasmic signal (25). In this work, we demonstrate that Osbp also plays a critical role in controlling the resolution phase of the TLR-induced inflammatory response in macrophages.

Results

25HC Sustains Inflammatory Cytokine Production in Macrophages.

25HC has been shown to affect signaling of Tlr3 and the adapter Trif (6) but its full role in TLR signaling more broadly is unknown. The expression of Ch25h is also induced by TLR agonists that signal via the adapter Myd88 (SI Appendix, Fig. S1A) and 25HC broadly affects the induction of multiple critical cytokines and chemokines downstream of both Tlr1/2 [Pam3CSK4 (PAM3)] and Tlr4 [Lipopolysaccharide (LPS)] (Fig. 1A). A detailed temporal analysis by RT-PCR shows that the predominant effect of 25HC is to sustain the expression of TLR-responsive genes (Fig. 1B and SI Appendix, Fig. S1B). ChIP-seq analysis for acetylated histone H4 (AcH4) indicates that this effect is mediated epigenetically (Fig. 1C and SI Appendix, Fig. S1C). This delayed resolution of the inflammatory response leads to dramatic increases in the secretion of critical cytokines and chemokines including Il6 (Fig. 1D and SI Appendix, Fig. S1D) and processed Il1b (SI Appendix, Fig. S2). The serum concentration of 25HC following LPS stimulation in mice is approximately 0.3 μM (9), and its effect on macrophage cytokine production is maintained down to at least 0.25 μM (Fig. 1E).

Fig. 1.

25HC sustains inflammatory cytokine production in macrophages. (A) Expression measured by RNA-seq analysis of selected cytokines and chemokines following 18 h of stimulation with LPS or PAM3 in BMDMs from WT mice left untreated (black) or treated with 25HC prior to stimulation (red arrows) or from Ch25h−/− mice (purple arrows). The expression of the included genes at 18 h was altered by 25HC treatment by more than four-fold (FDR < 0.05) for both stimuli. (B) Expression of Il6 measured by real-time PCR in BMDMs left untreated or treated with 25HC and then stimulated with LPS or PAM3. (C) Acetylated histone H4 signal at the Il6 promoter region in BMDMs left unstimulated, treated with 25HC, stimulated with PAM3 for 18 h, or treated with 25HC and then stimulated with PAM3 for 18 h. Plots show the cumulative signal from two biological replicates presented as fragments per million, normalized to the mean value across samples. The transcriptional start site is indicated with an arrow. (D) Secretion of Il6 measured by ELISA in BMDMs left untreated or treated with 25HC and then stimulated with LPS or PAM3 for 18 h. (E) Secretion of Il6 measured by ELISA in BMDMs left untreated or treated with the indicated concentrations of 25HC and then stimulated with LPS for 18 h. Bars/points indicate the mean of three biological replicates; error bars indicate SEM. (A, D, and E). Similar results were obtained in more than three independent experiments (B, D, and E).

Osbp Mediates Macrophage Inflammatory Responses to 25HC.

25HC targets several intracellular receptors including Srebp, Lxra/b, and Osbp and we sought to determine which of these might be responsible for its effect in sustaining TLR responses. Although both Srebp and Lxra/b have previously been demonstrated to play roles in macrophage immune responses (1–3, 17), the sustained inflammatory response induced by 25HC is retained in Lxra/b^−/− and Srebf1^−/− macrophages indicating that these molecules do not mediate the effect of 25HC on cytokine production (SI Appendix, Fig. S3A). Analysis of the global transcriptional response to 25HC treatment revealed that, consistent with Osbp’s role in transport across the ER–Golgi interface and with prior reports (26), 25HC-responsive genes were enriched in gene sets consisting of ER and Golgi-related genes (SI Appendix, Fig. S3B). Taken together, these data suggested a role for Osbp in regulating TLR signaling.

Mice lacking Osbp are not viable (22) and a conditional knockout does not yet exist. Therefore, to study the effect of Osbp on Tlr4 signaling, we employed several drugs, including itraconazole and OSW-1, that, like 25HC, are ligands for Osbp (15, 22, 24, 27, 28). We treated macrophages with either 25HC, itraconazole, or OSW-1, stimulated them with LPS or PAM3, and assayed cytokine production by ELISA. Both drugs phenocopied the effect of 25HC on Il6 production (Fig. 2A). Furthermore, the effect of OSW-1 and itraconazole on the macrophage transcriptome, both prior to and following LPS stimulation, is strikingly similar to that of 25HC (Fig. 2B). These results were consistent with the hypothesis that Osbp plays a critical role in sustaining Tlr4-induced cytokine production.

Fig. 2.

Osbp mediates macrophage inflammatory responses to 25HC. (A) Secretion of Il6 measured by ELISA in BMDMs left untreated or treated with the indicated Osbp ligands and then stimulated with LPS or PAM3 for 18 h. (B) BMDMs were left untreated or treated with 25HC, itraconazole, or OSW-1 then were left unstimulated or stimulated with LPS and their transcriptomes analyzed by RNA-seq. Top panels: Comparison of the effect of the indicated Osbp ligands on the expression of genes whose expression is altered by at least two-fold (FDR < 0.01; avg. log₂(counts-per-million reads) > 2 across all conditions) by treatment with at least one of the indicated Osbp ligands. Bottom panels: The core LPS response was defined as the 637 genes whose expression at 6 h was altered by at least 10-fold (FDR < 0.01; avg. log₂(CPM reads) > 2 across all conditions). The plots compare the effect of itraconazole or OSW-1 to that of 25HC on the expression at 18 h following stimulation of core response genes whose expression at 18 h was altered by at least two-fold by either of the indicated treatments on each plot. (C) Percentage of unedited sequences remaining in the indicated cell pools as measured by TIDE sequencing (29). (D) Western blot for Osbp in macrophages differentiated from Osbp g1, g5, and g6 pools. (E) Il6 protein level 18 h following LPS stimulation of macrophages derived from the indicated CIM lines and treated with the indicated compounds. Bars/points indicate the mean of three biological replicates; error bars indicate SEM (A and E). The blue line indicates best fit and r = Pearson correlation (B). Similar results were obtained in 3 (A) or 2 (E) independent experiments.

25HC, itraconazole, and OSW-1 target both Osbp and its closely related family member Osbp2 (commonly referred to as Orp4) (24). To determine the contribution of each protein to resolving TLR responses, we used a line of ER-Hoxb8 conditionally immortalized bone-marrow progenitors derived from mice that constitutively express Cas9 (30) to generate CRISPR-edited macrophages. In the presence of β-estradiol, these cells can be propagated indefinitely and then, by removal of β-estradiol and addition of M-CSF, differentiated into macrophages that are phenotypically very similar to BMDMs (30). We infected these “CIM (conditionally immortalized macrophage)” cells with lentiviruses carrying CRISPR guide RNAs targeting three different regions of Osbp or Orp4 (one guide per culture) and expanded them under antibiotic selection prior to differentiating them into macrophages. We confirmed the efficiency of editing (>70%) by “Tracking of Indels by Decomposition (TIDE)” sequencing (29) (Fig. 2C) and measured protein levels by western blot where possible (No antibody for murine Orp4 is commercially available.) (Fig. 2D). Knockdown of Osbp, but not Orp4, completely abrogated the ability of 25HC (Fig. 2E), itraconazole, and OSW-1 (SI Appendix, Fig. S4) to augment LPS-induced production of Il6, establishing that Osbp mediates the effect of 25HC on TLR signaling.

Treatment of macrophages with 25HC sustains expression of Il1b mRNA (SI Appendix, Fig. S2A) and dramatically increases the amount of Il1b produced over 18 h (SI Appendix, Fig. S2 B and D), although we do not detect any significant change in Il1b production in Ch25h−/− macrophages (SI Appendix, Fig. S2C). A recent study found that maintaining low levels of phosphatidylinositol 4-phosphate (PI4P) at EECS stabilizes these sites and restrains Nlrp3 inflammasome activity and that in cells lacking Osbp the inflammasome assembles spontaneously such that LPS alone is sufficient to release biologically active cleaved Il1b without the requirement for a second cytoplasmic signal (25). 25HC drives Osbp to ER-trans Golgi network membrane contact sites and thus is likely to deplete it from EECS which, based on the report from Zhang et al. (25), would be expected to enhance inflammasome activity. Consistent with this, we observe production of low levels of Il1b in LPS-stimulated macrophages treated with Osbp ligands (SI Appendix, Fig. S5A) without the requirement for the second signal (ATP). To disentangle the effect of increased inflammasome activation from other effects of Osbp on TLR signaling we treated macrophages with MCC950, a potent inhibitor of Nlrp3 activation, following treatment with the Osbp ligands and prior to stimulation with LPS. Inhibition of Nlrp3 significantly attenuated the production of Il1b but had no effect on the ability of Osbp ligands to enhance Il6 production suggesting that increased inflammasome activation is not driving this effect (SI Appendix, Fig. S5 B and C).

Osbp Mediates the Inflammatory Response to Sepsis.

To determine whether manipulating Osbp affected inflammatory responses in vivo in a similar manner to what we observed in vitro, we examined the effect of modulating Osbp activity on the outcome of LPS-induced sepsis, a condition that is largely driven by dysregulated cytokine production. We treated WT mice with OSW-1 or itraconazole prior to LPS injection and observed that the treated mice had substantially higher serum levels of Il6, Cxcl1, and Cxcl2 following LPS challenge (Fig. 3 A and B). Although the potency of these two compounds for modulating Osbp differs, it is important to note that neither compound alone led to elevated serum cytokine levels. Furthermore, consistent with this observation, Ch25h−/− mice, which cannot make 25HC, produce lower levels of circulating cytokines in response to LPS-induced sepsis (Fig. 3C). At the dose of LPS used in these experiments and the time points examined, we were not able to detect Il1b in the serum.

Fig. 3.

Osbp mediates the inflammatory response to sepsis. (A and B) Serum concentrations of the indicated cytokines at 24 h following i.p. challenge with LPS in untreated mice or mice treated with itraconazole (A) or OSW-1 (B) for 24 h. (n = 7 to 13 mice/condition) (C) Serum concentrations of the indicated cytokines at 24 h following i.p. challenge with LPS of WT or Ch25h−/− mice. Plot shows cumulative results of 3 independent experiments with a total of 14 mice/condition. Bars/points indicate the mean of three biological replicates; error bars indicate SEM (A–C). Similar results were obtained in 2 independent experiments (A and B).

Osbp Controls the Resolution of Inflammation by Altering Macrophage Lipid Metabolic Pathways.

Osbp’s best-described role is as an intracellular lipid transporter (15, 19, 22) and we hypothesized that it affects TLR responses by regulating cellular lipid content. We used mass-spectrometry-based lipidomics to examine the effect of altering Osbp function on the global lipid profile in LPS-stimulated macrophages. Similar to their effect on the transcriptional response, 25HC, OSW-1, and itraconazole-induced changes in lipid composition, both prior to and following LPS stimulation, that are strikingly similar (Fig. 4A). 25HC is well described to induce cholesterol ester (CE) formation (31), and indeed, we observed massive upregulation of CEs in macrophages treated with 25HC and LPS and moderate upregulation with OSW-1. However, itraconazole had no effect on CE production (Fig. 4A). 25HC and OSW-1 bind the conserved sterol binding domain of Osbp while itraconazole binds a different, unidentified site (32). To test the hypothesis that binding the sterol domain is required for CE synthesis, we treated macrophages with the Osbp ligand TTP-8307 (33) which, like itraconazole, binds to a site distinct from the sterol binding domain. Like itraconazole, TTP-8307 sustained cytokine secretion without having a significant effect on CE production (Fig. 4 B and C). These results suggest that the effect of Osbp on sustaining TLR responses is not dependent on elevated production of CEs. Furthermore, the fact that four distinct Osbp ligands have the same effect on sustaining TLR-induced cytokine production argues against these findings being attributable to the off-target effect of any single ligand.

Fig. 4.

Osbp controls the resolution of inflammation by altering macrophage lipid metabolic pathways. (A) BMDMs were left untreated or treated with 25HC, itraconazole, or OSW-1 then were left unstimulated or stimulated with LPS and their lipid content analyzed by mass-spectrometry. The plots depict the abundance ratios between treated and untreated cells for the indicated number of species (n of 385 that were robustly detected, see Materials and Methods) whose abundance was affected by at least one of the indicated Osbp ligands by at least 50% (FDR < 0.05) in either unstimulated macrophages (Top panels) or macrophages stimulated with LPS for 18 h (Bottom panels). Colors indicate lipid class: CE = Cholesterol Ester, CER = Ceramides, DAG = Diacylglycerol, FFA = Free Fatty Acids, HCER = Hexosylceramides, LPC = Lysophosphatidylcholine, LPE = Lysophosphatidylethanolamine, PC = Phosphatidylcholine, PE = Phosphatidylethanolamine, SM = Sphingomyelin, TAG = Triacylglycerol. (B) BMDMs from WT mice were treated with the indicated Osbp ligand followed by stimulation with LPS for 18 h and Il6 measured by ELISA. (C) Data from (A) summarized by total abundance of the indicated lipid classes. (D) BMDMs from WT mice were treated with the indicated Osbp ligand followed by stimulation with LPS or LPS + T863 for 18 h and Il6 measured by ELISA. Data from untreated samples are repeated across the top and bottom panels, which depict separate experiments, to enable appropriate scaling. Bars indicate the mean of the biological replicates; error bars indicate SEM (B–D). Similar results were obtained in 3 (B) or 2 (D) independent experiments.

All four of the Osbp ligands strongly increased macrophage triglyceride (TG) content (Fig. 4C) and triglyceride production in macrophages has previously been demonstrated to play a role in TLR signaling (34). We therefore hypothesized that Osbp mediated accumulation of triglycerides is, at least in part, responsible for the ability of Osbp to amplify TLR responses. To test this hypothesis, we treated BMDMs with the Osbp ligands followed by treatment with LPS and T863, a selective and potent antagonist of Dgat1 that inhibits triglyceride production in cells, and assayed cytokine production by ELISA. While T863 only modestly inhibited LPS-induced Il6 production (in line with a previous report (34)), it completely abrogated the ability of Osbp ligands to enhance Il6 production, which is consistent with the hypothesis that these ligands amplify macrophage cytokine responses via enhanced triglyceride content (Fig. 4D). A previous study suggested that elevated triglyceride levels in macrophages lead to enhanced Il6 secretion by increasing production of PGE2 (34). However, we found that the levels of PGE2 induced by each of the four Osbp ligands tested here did not correlate with the level of Il6 secreted, and therefore, we consider it unlikely that this is the mechanism by which the 25HC–Osbp interaction sustains cytokine production (SI Appendix, Fig. S6).

Discussion

In this work, we have demonstrated a previously unappreciated role for Osbp in the LPS-induced TLR signaling pathway in macrophages. TLR4 activation leads to increased production of the enzyme Ch25h which in turn increases production of 25HC. Our data suggest that 25HC, acting by binding to Osbp, sustains and therefore amplifies macrophage cytokine responses. Experiments using BMDMs from Lxra/b−/− and Srebf1−/− mice suggest that neither of these proteins plays a significant role in this effect. Deletion of Srebpf2 is embryonic lethal but a floxed Srebpf2 mouse has been generated (35). Although in previous studies breeding these mice to mice expressing Cre under the control of the albumin promoter was sufficient to significantly reduce expression in hepatocytes (35), we were unable to achieve knockdown in BMDMs by breeding to LysM-Cre mice. Therefore, we cannot formally exclude a partial role for Srebf2 in 25HC-mediated amplification of cytokine responses.

While we have not determined the precise mechanism by which Osbp ligands increase TG levels in macrophages, our transcriptomic data strongly suggest that they induce the ER stress response, likely by altering cholesterol levels in the ER (SI Appendix, Fig. S3B) and several studies have demonstrated that activation of the ER stress response leads to TG accumulation (36–39). Our results are consistent with prior report (40) which showed that treatment of HUVECs treated with the Osbp ligands Schweinfurthin G and OSW-1 resulted in elevated TG content. While we cannot rigorously determine the relative contributions of TG production and uptake, the fact that inhibiting Dgat1, an enzyme involved in TG synthesis, abrogates the phenotype strongly suggests that production is the predominant mechanism.

Our data suggest a model in which TLR-induced (or other) changes in the relative concentrations of the natural Osbp ligand 25HC are transduced through Osbp to both transcriptional and lipid metabolic networks. 25HC binds to Osbp with extremely high affinity and high concentrations can effectively “lock” it to the Golgi membrane (18). This disrupts its lipid transport function, which drives changes in the global lipidome, and prevents Osbp from performing its other functions in the cytosol or at other MCS (e.g., EECS) and leads to increased triglyceride production which in turn sustains the secretion of proinflammatory cytokines such as Il6.

Osbp controls the duration and therefore the magnitude of macrophage immune responses and is therefore a critical intersection point linking lipid metabolic and innate immune pathways. Multiple agents that target selected functions of Osbp have been described, one of which is an FDA-approved drug in common clinical use. Thus, Osbp represents a “druggable target” that could allow more precise tuning of the immune response.

Materials and Methods

Mice.

WT (C57BL6/J; JAX: 000664), Ch25h−/− (B6.129S6-Ch25h^tm1Rus/J; JAX: 016263), Srebpf2^fl/fl (Srebf2^tm1Jdh/J JAX: 030826), and LysM-Cre (B6.129P2-Lyz2^{tm1(cre)Ifo/J}) strains of Mus musculus were obtained from The Jackson Laboratory. Srebf1−/− mice were obtained from Dr. Hitoshi Shimano, University of Tsukuba, Tsukuba, Ibaraki, Japan. Lxra−/−; Lxrb−/− double-knockout mice were bred in house by crossing Lxra−/− (B6.129S6-Nr1h3^tm1Djm/J JAX: 013761) and Lxrb−/− (B6.129S6-Nr1h2^tm1Djm/J JAX: 014633) mice that were obtained from The Jackson Laboratory. All KO mouse experiments used only homozygous animals. All mice were housed and maintained in specific pathogen-free conditions at the Seattle Children’s Research Institute (SCRI). All experiments were approved by the Institutional Animal Care and Use Committee and then performed in compliance with the relevant protocols. Healthy 8- to 20-wk-old female mice without any previous procedure history were used for all experiments and were age-matched within each experiment.

Cells.

Bone marrow–derived macrophages (BMDMs) were prepared as described previously (41, 42).

Conditionally immortalized macrophages (CIM) were grown as described in(30). Briefly, progenitor CIM cells were maintained in RPMI with 10% FCS, Pen/Strep, 4 mM L-glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 43 μM β-mercaptoethanol, 2 μM β-estradiol (Sigma #E2758), and murine GM-CSF 20 ng/mL (Peprotech 315-03). Progenitor CIMs were maintained in suspension in nontreated tissue culture dishes at densities below 500,000 cells/mL before the removal of β-estradiol and differentiation. Macrophages were differentiated for 6 d on nontreated tissue culture dishes in RPMI with 10% FCS, Pen/Strep, 4 mM L-glutamine, and 50 ng/mL human M-CSF (Peprotech 300-25), then lifted in 1 mM PBS/EDTA, counted, and replated on tissue culture dishes prior to stimulation.

CRISPR editing of CIM cells.

Target	Guide	Sequence	Algorithm
Osbp	1	GGATGTTATCCTAAAAAGCG	CRISPick (43, 44)
Osbp	5	AAGAGACCAGCGAATCCGAC	GeneScript (45)
Osbp	6	GGCAAAGCAGCACAATCACC	GeneScript (45)
Orp4	1	GTGTCAGGGATTGGTGTCCA	CRISPick (43, 44)
Orp4	2	GCGTCGAGATCCGAACCTGG	CRISPick (43, 44)
Orp4	3	TAGTTACCTTTCGGGCTGCC	ChopChop (46)

LentiCRISPR sgRNAs (single guide RNAs) were designed using several different algorithms (see table), selecting for the highest-scoring guides in target specificity and efficiency for Osbp and Orp4. Oligos spanning the sgRNA sequence were annealed and ligated into the lentiGuide-Puro backbone (Addgene plasmid #52963, gift from Feng Zhang).

All lentiviral constructs were transformed into Stbl3 bacteria (ThermoFisher Cat: C737303) for propagation of plasmid DNA. All plasmids were prepared using a Nucleobond Xtra Maxi Kit (Takara Cat:740414.100). Coding sequences of sgRNAs were confirmed by automated sequencing (Genewiz). As described in (47), lentivirus stocks were produced by PEI-mediated transfection into 293FT cells. For lentiviral vectors, plasmid amounts were 0.3 μg CMV-VSV-G (Addgene plasmid #8454), 0.75 μg psPax2 (Addgene plasmid #12260), and 0.95 μg transgene (sgRNA construct). Supernatants were passed through 0.45-μm syringe filters (Corning). CIM cells were modified by lentiviral sgRNA constructs similar to previously described protocols (30). Cells were resuspended in medium with protamine sulfate (Sigma, 1 μg/mL), aliquoted to 6-well plates with 0.5 × 10⁶ cells in 0.5 mL of medium per well, and 1.5 mL of lentiviral stocks were added. Cells were placed under selection with puromycin (12 μg/mL, Invivogen) two days after transduction for 1 wk and puromycin-resistant populations were allowed to expand. Only cell lines transduced with virus-carrying guide sequences that exhibited at least 70% editing efficiency as measured by TIDE sequencing (29) were used in experiments.

Stimulations.

Frozen stocks of LPS (List Biological #R595) or PAM3 (Invivogen #vac-pms) were thawed and sonicated for 5 min before dilution in BMDM medium. The diluted stimuli were again sonicated for 5 min before addition to BMDM cultures for a final concentration of 10 ng/mL for LPS and 300 ng/mL for PAM3. CIM cells were stimulated with 100 ng/mL LPS. Unless otherwise indicated, cells were incubated with 25HC (5 μM) (Sigma #H1015), OSW-1 (0.1 nM) (Cayman Chemicals #30310), itraconazole (1 μM) (Sigma #419825), or TTP-8307 (10 μM) (MedChemExpress #HY-124806) for 18 h prior to stimulation. To activate the Nlrp3 inflammasome, cells were stimulated with 5 mM ATP (Sigma #A2383) for 45 min prior to collection of supernatants. To block activation of the Nlrp3 inflammasome, cells were treated with MCC950 (10 μM) (MedChemExpress # HY-12815). To inhibit Dgat1, cells were treated with T863 (50 μM) (MedChemExpress #HY-32219).

Western Blot.

After 6 d of differentiation, BMDM were replated at 10⁶ cells/well of a TC-treated 6-well dish. At each time point, culture medium was removed, and the cells were washed once with PBS before protein collection with RIPA buffer and with protease inhibitors. Western blotting analyses were performed using standard techniques and transblotted onto nitrocellulose membranes. After blocking with 5% milk, membranes were probed with the relevant primary antibodies (Cleaved Il1b – E7V2A (Cell Signaling #63124), Osbp Polyclonal (Proteintech #11096-1-AP)) followed by detection with an HRP-conjugated secondary antibody (Anti-mouse IgG-HRP (Cell Signaling #7076), Anti-rabbit IgG HRP (Cell Signaling #7074)). Where applicable, membranes were stripped, reblocked, and probed with HRP-conjugated anti-β-actin (bActin-HRP (Abcam #ab20272)). Images were quantified using ImageJ densitometry analysis.

Enzyme-Linked Immunosorbent Assay (ELISA).

BMDM cell culture supernatants were collected and concentrations of specific cytokines were determined using commercial ELISAs according to the manufacturer’s specifications. All ELISA Duo Set kits were purchased from R&D Systems [Il6 (DY406), Cxcl1 (DY453), Cxcl2 (DY452), Il1b (DY401), and Csf2 (DY415)] or Cayman Chemical [PGE2 (514010)]. Absorbance was recorded using a SpectraMax M2 plate reader (Molecular Devices).

RNA Preparation.

After 6 d of differentiation, BMDM were replated at 400,000 cells/well of a TC-treated 24-well dish. Cells were harvested for total RNA extraction using the SV 96 total isolation RNA kit (Promega). RNA integrity was checked using an Agilent 2,100 Bioanalyzer.

Quantitative Reverse Transcription PCR.

RNA (10 μL) was converted to cDNA using Super Script II and Oligo dT18-20 from Invitrogen according to protocol (20 μL reaction). Then, 20 μL cDNA was diluted to 100 μL and 2 μL was taken for RT-qPCR using Taq Man primer probe and Taq Man Fast Universal PCR master mix. All primer probe sets were purchased from ThermoFisher (Il6 (Mm00446190_m1), Cxcl1 (Mm00433859_m1), Cxcl2 (Mm00436450_m1), Il1b (Mm01336189_m1), Csf2 (Mm01290062_m1), and Ch25h (Mm00515486_s1)). Data were normalized to Eef1a1 expression in individual samples (Integrated DNA technologies–Eef1a1 forward primer for custom TaqMan assay: 5’ GCAAAAACGACCCACCAATG 3’, Eef1a1 reverse primer for custom TaqMan assay: 5’ GGCCTGGATGGTTCAGGATA 3’, Eef1a1 probe for custom TaqMan assay: 5’/56-FAM/CACCTGAGCAGTGAAGCCAG/36-TAMSp/3’).

RNA-seq.

Library preparation (Illumina TruSeq stranded kit) and sequencing (Illumina NovaSeq6000 S4, 2 × 150 bp paired-end reads) were performed by a commercial vendor (Psomagen).

For all samples, initial quality assessment was performed with FastQC version 0.11.5 (48) before read ends consisting of 33% or more of the same nucleotide were removed. Reads were aligned to the mouse genome (NCBI GRCm38 mm10) using GSNAP version 2018-07-04 (49). Feature counts were extracted using the featureCounts software (50), distributed as part of the Subread package v1.5.2. Normalization (to find counts-per-million (CPM) for each gene) and differential analysis of gene expression was calculated using edgeR version 3.26.8 (51). Data are deposited in GEO: GSE201132.

Gene Set Enrichment Analysis.

The set of 541 genes whose expression was altered by at least 1.5-fold (FDR < 0.05; avg. log₂(CPM reads) >4 across all conditions) by 25HC treatment was searched against the GO Biological Process Database (2018) for enriched gene sets using the enrichR R package (52). Significance was assessed using the “Adjusted p-value” of the enrichment.

ChIP-seq.

Immunoprecipitation (IP) for acetyl-H4 was performed as described in (53) using a rabbit polyclonal antibody (Millipore catalog number 06-866). Libraries were prepared from immunoprecipitated DNA using the KAPA Hyper Prep kit (KR0961-v3.15) and sequenced as 2 × 75 bp paired-end reads. Reads were aligned to the mouse genome (NCBI GRCm38 mm10) using GSNAP (49) with splice detection disabled. Duplicate reads were removed and, using the start and end locations of the pairs to define fragment lengths, only reads with lengths <500 nucleotides were retained. Broad peaks were called using MACS2 (2.1.0) (54). ChIP-signal is presented as fragments per million, normalized to the mean value across samples. Data are deposited in GEO: GSE201132.

Lipidomics.

After 6 d of differentiation, 5 × 10⁶ BMDM cells were replated and treated as indicated. Culture medium was then removed, and the cells were washed twice with PBS, then lysed in 200 μL 0.1 × PBS, scraped and lifted, and flash-frozen on dry ice. Lipidomic analysis was performed by the Mitochondria and Metabolism Center at the University of Washington as described previously (55). Briefly, lipids were extracted using dichloromethane/methanol and 13 classes of lipids were analyzed using the Lipidyzer platform consisting of an AB Sciex 5500 MS/MS QTraps system equipped with a SelexION for differential mobility spectrometry (DMS). Multiple reaction monitoring was used to target and quantify 1,070 lipids in positive and negative ionization modes with and without DMS. Lipid abundances were quantified using a representative set of isotopically labeled standards and are expressed as nmol/mg of total protein in the cell lysate. The analysis presented in Fig. 4 was performed on the 385 species that were detected at a concentration 1.5× the level in a blank sample in at least 24 of 36 samples. Statistical analysis of differential lipid abundance was performed using the limma package (56) in R on the log₂-transformed concentration measurements. The complete dataset is presented in Dataset S1.

LPS Sepsis Model.

Mice were injected i.p. with 20 mg/kg LPS (day 0). For experiments with Osbp ligands, mice were injected with 0.01 mg/kg OSW-1 i.p. or 40 mg/kg itraconazole i.p. (or PBS for controls) at day-1 and day 0. Mice were killed 24 h after injection of LPS and blood was harvested by cardiac puncture. Serum was isolated by centrifugation and cytokine/chemokine levels were measured by ELISA.

Quantification and Statistical Analysis.

Data were analyzed using R. Data are either presented as the average of individual technical or biological replicates, as indicated, with a bar or point representing the mean with SEM error bars. Comparisons between two groups were assessed with a two-sided Student’s t test. P-values are indicated on graphs as appropriate: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Precise values of n and numbers of independent experiments are indicated in figure legends.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

RNA-seq, ChIP-seq data are publicly available in GEO (GSE201132) (57).