Soluble epoxide hydrolase maintains steady-state lipid turnover linked with autocrine signaling in peritoneal macrophages

Summary Soluble epoxide hydrolase is a widely distributed bifunctional enzyme that contains N-terminal phosphatase (N-phos) and C-terminal epoxide hydrolase (C-EH) domains. C-EH hydrolyzes anti-inflammatory epoxy-fatty acids to corresponding diols and contributes to various inflammatory conditions. However, N-phos has been poorly examined. In peritoneal macrophages, the N-phos inhibitor amino-hydroxybenzoic acid (AHBA) seemed to primarily interrupt the dephosphorylation of lysophosphatidates and broadly attenuated inflammation-related functions. AHBA activated intrinsic lysophosphatidate and thromboxane A2 receptors by altering lipid-metabolite distribution; downstream the signaling, phospholipase C was facilitated to dampen intracellular Ca2+ stores and AKT kinase (protein kinase B) was activated to presumably inhibit inflammatory gene expression. Our data suggest that N-phos maintains steady-state phospholipid turnover connecting autocrine signaling and is a prospective target for controlling inflammatory responses in macrophages.


INTRODUCTION
Polyunsaturated fatty acids (PUFAs), such as arachidonic and eicosapentaenoic acids, are converted into lipid mediators by cyclooxygenase (COX), lipoxygenase (LOX), and cytochrome P450. 1 COX and LOX produce prostanoid and eicosanoid precursors, while P450 monooxygenase generates anti-inflammatory epoxy-fatty acids, which are further metabolized into pro-inflammatory acyl diols by soluble epoxide hydrolase (sEH). 2 Therefore, sEH inhibition is thought to exert anti-inflammatory effects via the accumulation of epoxidated fatty acids such as epoxy-eicosatrienoic and epoxy-docosapentaenoic acids. 2 Indeed, sEH inhibitors and epoxy-fatty acid analogs develop beneficial effects on various inflammation-related disease models including tissue flares, hypertension, cardiac hypertrophy, hyperglycemia, and hyperlipidemia. 2,3 In mammals, sEH is broadly expressed in various cell types of body tissues and forms a homo-dimer in the cytoplasm and peroxisome. 4 Importantly, sEH is a bifunctional enzyme with a molecular mass of 62 kDa; the lipid epoxide hydrolase activity is located in its 36-kDa C-terminal domain, while the 25-kDa N-terminal domain possesses lipid phosphatase activity. 4 As described previously, the role of the C-terminal epoxide hydrolase (C-EH) has been extensively characterized by means of several inhibitors and mutant mice in various disease models. 2,3 However, the N-terminal phosphatase (N-phos) has been poorly examined. In in vitro enzyme assays, N-phos catalyzes the dephosphorylation of several phosphorylated lipids including isoprenoid pyrophosphates, lysophosphatidic acids (LPAs), and sphingosine-1-phosphate. [5][6][7] A recent study using the knock-in rats bearing an N-phos-defective mutation indicates that N-phos mainly converts LPAs into monoacylglycerols (MAGs) in the liver. 8 Moreover, the N-phos-defective mutation likely confers resistance to fat mass gain during high-fat feeding and heart injury induced by ischemia-reperfusion, suggesting that N-phos takes part in pathological complications. 8 It is therefore important to examine the patho-and physiological role of N-phos using the chemical tools that selectively modulate N-phos activity. Kihara et al. currently showed that 3-amino-4-hydroxy benzoic acid (AHBA) specifically inhibits N-phos without affecting C-EH activity. 9 In this study, we focus on AHBA-mediated N-phos inhibition in thioglycolate-elicited peritoneal macrophages (TGPMs), which are widely used as a primary-culture model for immunological studies. Our results suggest that N-phos catalyzes a key dephosphorylation to maintain steady-state phospholipid turnover, and also that AHBA-induced N-phos inhibition stimulates autocrine signaling to attenuate inflammation-related functions in TGPMs. . The Ca 2+ ionophore ionomycin (IM) was used to trigger Ca 2+ release and deplete intracellular stores. The inhibitors used were the N-phos-specific inhibitor AHBA, the C-EH-specific inhibitor AUDA, and the dual inhibitor ebselen. In the dot plots, resting Fura-2 ratio, ionomycin-induced response, and SOCE amplitude are statistically analyzed. The data represent means G SEM., and the numbers of cells and mice examined are shown in parentheses. Significant differences from the vehicle-treated group are marked with asterisks (**p < 0.01 in one-way ANOVA and Dunnett's test). (B) AHBA reduces Ca 2+ release in response to P2Y receptor activation. The imaging traces show averaged time courses with shaded areas indicating standard errors (n = 10 TGPMs in each group). In the dot plot, ATP-induced responses are statistically analyzed. A significant difference is marked with an asterisk (*p < 0.05 in t-test). iScience Article trisphosphate (IP 3 ) receptors in response to ATP-induced P2Y receptor activation ( Figure 1B). Ebselen reduced ionomycin-induced Ca 2+ release as similar to AHBA, while AUDA exerted no effects on intracellular Ca 2+ stores ( Figure 1A). Therefore, independently of C-EH blockade, N-phos inhibition likely reduced the Ca 2+ contents of IP 3 -sensitive stores in TGPMs. This consequence was also supported by imaging data using another set of inhibitors, i.e., the N-phos inhibitor oxaprozin, 12 the C-EH inhibitor TPPU, 13 and the dual inhibitor N-acetyl-S-farnesyl-L-cysteine 12 ( Figure S1A).
In our preliminary search, AHBA treatment also reduced Ca 2+ stores in the Raw 264.7 macrophage cell line, A20 lymphoma cell line, primary-cultured mouse embryonic fibroblasts and hepatocytes dissociated from adult mice ( Figure S1B). Therefore, acute N-phos inhibition seemed to shrink intracellular Ca 2+ stores in various cell types. We then attempted to clarify the molecular mechanism underlying the AHBA-induced shrinkage of Ca 2+ stores in TGPMs.
AHBA PLC-dependently facilitated store Ca 2+ leakage In contrast to ATP-induced Ca 2+ release through IP 3 receptor channels ( Figure 1B), TGPMs exhibited no Ca 2+ response to caffeine, an effective activator for ryanodine receptor channels ( Figure S2A). This observation was consistent with the negligible expression of the Ryr subtype genes in TGPMs, 14 indicating that IP 3 receptors function as predominant Ca 2+ release channels in TGPMs. Although IP 3 receptor gating is dependent on both IP 3 and Ca 2+ , 15 short-term AHBA treatment reduced Ca 2+ stores without affecting resting [Ca 2+ ] i in TGPMs ( Figure 1A). To examine if AHBA dampened Ca 2+ stores by facilitating steadystate IP 3 generation, we used the phospholipase C (PLC) inhibitors U73122 and manoalide. Co-application of either the inhibitor diminished AHBA-induced reduction in store Ca 2+ contents ( Figure 2A). Moreover, the AHBA-induced reduction was clearly time-dependent; longer AHBA exposure resulted in more severe reduction in Ca 2+ stores ( Figure 2B). After a prolonged AHBA exposure (10 mM for > 20 min), intracellular stores were largely depleted, and the resting [Ca 2+ ] i became elevated in a Ca 2+ -containing bath solution. The elevated [Ca 2+ ] i was thought to be due to facilitated SOCE driven by the STIM-Orai activation. 16 The previous observations suggested that AHBA facilitated IP 3 generation by moderately enhancing ''idling'' PLC activity under resting conditions. The resulting enhanced IP 3 generation might gradually stimulate Ca 2+ leakage mediated by IP 3 receptors, but did not trigger Ca 2+ transients or waves, either of which requires the coincident activation of large numbers of IP 3 receptors. In addition to IP 3 liberation, PLC also generates diacylglyceroles (DAGs), which are endogenous activators of protein kinase C (PKC). However, the PKC inhibitor bisindolylmaleimide IV exerted no effect on AHBA-induced reduction in Ca 2+ stores, suggesting that PKC did not affect store Ca 2+ -handling in TGPMs ( Figure S2B).

AHBA altered phospholipid metabolism
We next examined AHBA-induced effects on lipid metabolism. Total lipid extracts prepared from TGPMs treated with or without AHBA (10 mM for 5 min) were subjected to outsourced lipid metabolomic analysis. As notable observations in the multi-phospholipid analysis, lysophosphatidylcholine (LPC), LPA and phosphatidylinositol triphosphate (PIP 3 ) contents were tended to increase in AHBA-treated TGPMs when compared with control cells ( Figure 3A; Table S1). The elevated LPC content probably reflected the specific hydrolysis of PCs, because AHBA treatment did not affect other lyso-phospholipid species. The eicosa/docosanoid analysis likely suggested that AHBA treatment increased contents of some PUFAs such as arachidonic and eicosapentaenoic acids without affecting non-PUFA contents ( Figure 3B and Table S2). It was also predicted that the generation of thromboxane B2 (TXB2), a non-enzymatic metabolite of thromboxane A2 (TXA2), was stimulated in AHBA-treated TGPMs (Table S2). This observation, together with immunochemical TXB2 measurements (Figure 7), suggested that AHBA treatment facilitated TXA2 synthesis.
Of the N-phos substrates reported thus far, [5][6][7] LPAs significantly accumulated in AHBA-treated cells. Therefore, the conversion of LPAs to MAGs may be mediated predominantly by N-phos in TGPMs. On the other hand, several possibilities could be drawn from the AHBA-induced shifts in lipid metabolite; however, one reasonable explanation is as follows. The increased LPC and PUFA contents suggested that AHBA may stimulate phospholipase A2 (PLA2), while the elevated PIP 3 content suggested that phosphoinositide 3-kinase (PI3K) may be activated by AHBA. Furthermore, we focused on the LPA accumulation and TXA2 induction in AHBA-treated TGPMs, because both of the metabolites act as bioactive autacoids. Gene expression profiles in a public database and our observations in reverse transcription polymerase chain reaction (RT-PCR) and western blot analyses indicated that the LPA receptor subtype genes Lpar1 and Lpar5 were weakly expressed in TGPMs ( Figure S3). LPAR1 and LPAR5 subtypes are known to potentially couple to the guanine nucleotide-binding proteins Gq, Gi/o, and G 12/13 . 17 The application of oleoyl-LPA at a submaximal dose induced no Ca 2+ transient in TGPMs ( Figure 4A), in contrast to definitive Ca 2+ transients triggered by ATP-induced P2Y receptor activation ( Figure 1B). This discrepancy was likely due to contrasting receptor densities and G-protein-coupling orientations; Lpar mRNA content was roughly estimated to be at least an order magnitude less than P2ry mRNA content ( Figure S3), and LPA receptors did not efficiently couple to Gq in TGPMs as explained further.
Similar to AHBA treatment, oleoyl-LPA clearly reduced ionomycin-induced Ca 2+ release within a few minutes in TGPMs ( Figure 4A). However, unexpectedly, the oleoyl-LPA-induced reduction in Ca 2+ stores was largely attenuated by the PLA2 inhibitor pyrrophenone and the TXA2 receptor antagonist daltroban ( Figure 4A). Therefore, TXA2 production and signaling may essentially contribute to the reduction induced by LPA receptor activation, and it was unlikely that LPA receptor activation directly induced Gq-mediated PLC stimulation.
Of the LPA receptor antagonists that have been variedly developed, Ki16425 preferentially inhibits LPAR1/ 3, and H2L5765834 selectively blocks LPAR1/3/5. When either antagonist was co-treated, AHBA did not reduce ionomycin-induced Ca 2+ release ( Figure 4B), suggesting that AHBA inhibited the conversion of iScience Article LPA to MAG to increase intracellular LPA content and thus induced autocrine activation of LPA receptors toward Ca 2+ store reduction. Taken together with the effects of the TXA2 signaling inhibitors, co-activation of LPA and TXA2 receptors likely underlay AHBA-induced reduction in store Ca 2+ content in TGPMs.

AHBA facilitated kinase signaling
Of the divergent PLA2 subtypes, cytosolic phospholipase A2a (cPLA2, also known as PLA2G4A) often contributes to prostanoid and eicosanoid production in response to various stimuli. 18 Extracellular signal-regulated kinase (ERK) and Rho-associated protein kinase (ROCK) are stimulated downstream of Gi/o and G 12/13 , respectively, 19,20 and both kinases can activate cPLA2 by phosphorylating Ser505. 21,22 In western blot analysis of total cell lysates prepared from TGPMs, the signal intensity of phospho-cPLA2 was markedly increased in response to AHBA treatment (10 mM for 20 min). This increase was inhibited by the co-application of H2L5765834 ( Figure 5A). In addition, AHBA-induced cPLA2 phosphorylation was disrupted by the ERK inhibitor FR180204, while the ROCK inhibitor Y-27632 had no effect (Figure 5B). Therefore, AHBA-induced LPA receptor activation probably stimulated ERK-mediated cPLA2 phosphorylation, but was unlikely to efficiently facilitate G 12/13 -ROCK signaling in TGPMs.
ERK is activated by phosphorylation at Thr183 and Tyr185 by mitogen-activated protein kinase kinase (MEK) downstream of Gi/o activation. 19 AHBA treatment elevated phospho-ERK content, and this elevation was inhibited by H2L5765834 ( Figure 5C). Therefore, cPLA2 was probably activated downstream of Gi/o-ERK signaling resulting from LPA receptor activation in AHBA-treated TGPMs. The activated cPLA2 was thought to selectively react with PCs to generate lyso-PCs and PUFAs, because PCs are preferable substrates for cPLA2. 23 Furthermore, it has been also reported that PUFAs generated by cPLA2 activation are predominantly converted into TXA2 in macrophages. 24 Downstream of Gi/o signaling, in parallel with the ERK activation, PI3K is often stimulated to generate PIP 3rich membrane compartments, where AKT kinase (protein kinase B) is assembled and activated. 25 The active form of AKT contains phospho-Thr308 and phospho-Ser473, which are separately phosphorylated by phosphoinositide-dependent kinase 1 (PDK1) and mammalian target of rapamycin complex 2 (mTORC2), respectively. PDK1 activation likely requires ERK activity; ERK phosphorylates and activates ribosomal S6 kinase-2 (RSK2), and the activated RSK2 then recruits and activates PDK1 in a COS7 expression system. 26 In AHBA-treated TGPMs, AKT phosphorylation was elevated at both sites, and this elevation was abolished by H2L5765834 ( Figure 5D) and the PI3K inhibitor TG100-115 ( Figure S4A). Therefore, LPA receptor activation seemed to sequentially induce PI3K and AKT activation in AHBA-treated TGPMs. Moreover, FR180204 significantly inhibited PDK1-dependent Thr308 phosphorylation but exerted no obvious iScience Article effect on mTORC2-dependent Ser473 phosphorylation ( Figure S4A), suggesting that ERK also contributed to PDK1-mediated AKT activation in TGPMs. On the other hand, as expected, the mTORC1/2 inhibitor torin 1 disrupted AHBA-induced Ser473 phosphorylation without affecting T308 phosphorylation ( Figure S4B).

AHBA autocrine-activated TXA2 receptors
TXA2 primarily functions as an auto/paracrine mediator to activate TXA2 receptors coupling with Gq. 18 In gene expression and western blot analyses, the TXA2 receptor gene Tbx2r was marginally active in TGPMs, and the estimated mRNA content was roughly similar to that of Lpar mRNAs ( Figure S3). In Ca 2+ imaging, the TXA2 receptor agonist I-BOP did not induce Ca 2+ transients but clearly reduced store Ca 2+ content in TGPMs ( Figure 6A). The PLA2 inhibitor pyrrophenone and the COX/LOX inhibitors licofelone and thymoquinone disrupted AHBA-induced reduction in Ca 2+ stores ( Figures 6B and S2C). Furthermore, the TXA2 receptor antagonists daltroban and picotamide also abolished the AHBA-induced reduction in Ca 2+ store ( Figure 6C). Therefore, TXA2 generation and TXA2 receptor activation seemed to contribute to the AHBA-induced effect on TGPMs. TXA2 receptor activation is supposed to stimulate Gq and its downstream effector PLCb. 23

AHBA facilitated TXA2 and IP 3 generation
To confirm the predicted induction of TXA2 generation in AHBA-treated TGPMs (Table S2), we used an enzyme-linked immunosorbent assay (ELISA) kit to quantify TXB2 released into cultured supernatants. iScience Article AHBA treatment (10 mM for 20 min) markedly elevated TXB2 content, indicating that TXA2 production was enhanced in AHBA-treated TGPMs ( Figure 7A). TXB2 contents were also increased in TGPMs treated with the N-phos inhibitor oxaprozin and the dual inhibitor ebselen ( Figure S5A). AHBA-induced TXB2 accumulation was inhibited by co-treated H2L5765834, thymoquinone, pyrrophenone, or FR180204, supporting the contribution of LPA receptors, ERK, cPLA2, and COX to the facilitated TXA2 production. Although (D) AKT phosphorylation and AHBA-induced LPA receptor activation. Total AKT and its Thr308-and Ser473phosphorylated forms were examined in TGPMs treated with the indicated inhibitors. The inhibitors used are the LPA receptor blocker H2L5765834, the TXA2 receptor blocker daltroban, the ERK inhibitor FR 180204, the ROCK inhibitor Y-27632, and the PI3K inhibitor TG100-116. Total cell lysates were prepared from TGPMs treated with the combination of inhibitors for 20 min and subjected to immune-blot analysis. The resulting immune-signals were captured as shown in the representative digital images and quantitatively analyzed as presented in the plot graphs. The relative abundance is related to the average level of the right-side group in each graph. The data represent means G SEM., and the mice examined are shown in parentheses. Significant differences between the groups indicated by bars are statistically examined (ns: not significant, *p < 0.05 and **p < 0.01 in ANOVA and Sidak's test). iScience Article picotamide and daltroban are both TXA2 receptor antagonists, picotamide is known to also act as a thromboxane synthase inhibitor and thus seemed to inhibit the AHBA-induced enhancement.
We next examined AHBA-induced effects on IP 3 generation using a commercial kit for quantifying intracellular IP 1 as a surrogate for IP 3 generated during incubation period. AHBA treatment (10 mM for 20 min) significantly elevated IP 1 content in TGPMs, indicating AHBA-induced facilitation of IP 3 generation (Figure 7B). IP 1 contents were also increased in TGPMs treated with oxaprozin and ebselen ( Figure S5B). The AHBA-induced IP 1 accumulation was inhibited by co-treated Ki16425, H2L5765834, and daltroban  (Figures 4 and 6). Of the kinase inhibitors tested, FR180204 significantly suppressed the AHBA-induced IP 1 accumulation likely due to disruption of ERK-mediated cPLA2 activation for TXA2 synthesis. Therefore, AHBA likely facilitated steady-state IP 3 production in a manner that was dependent on sequential activations of LPA receptors, ERK, cPLA2, TXA2 receptors, and PLC in TGPMs.

AHBA attenuated macrophage functions
AHBA had no severe toxic effect on TGPMs at less than 100 mM during 24-h incubation ( Figure S5C). We next examined the effects of AHBA (30 mM) on chemotaxis, phagocytosis, and H 2 O 2 generation, all of which are fundamental macrophage functions. In migration across a membrane equipped with micropores during 4 h, TGPM chemotaxis toward ATP was inhibited by AHBA ( Figure 8A). This inhibitory effect was almost eliminated by co-treated H2L5765834 and daltroban. AHBA also inhibited migration to C-C motif chemokine ligand 2 (CCL2), but did not affect chemotaxis to the bacterial outer membrane component lipopolysaccharide (LPS) ( Figure S5D). The results suggested that AHBA can broadly attenuate migration signaling initiated by activation of G-protein coupled receptors as the sensors of chemical gradients.
In the phagocytosis assay monitoring the uptake of fluorescence-labeled microbeads for 3 h, AHBA significantly inhibited both phagocytic cell rates and capacities, thus leading to marked reduction in phagocytic index ( Figure 8B). Co-treated H2L5765834 and daltroban considerably restored the phagocytic index reduced by AHBA. In the imaging assay using an H 2 O 2 -sensitive indicator, AHBA inhibited the accumulation of fluorescence-positive dye which was produced by reacting with H 2 O 2 during 3-h incubation, iScience Article indicating that NADPH oxidase 2 (NOX2) activity was attenuated in AHBA-treated TGPMs ( Figure 8C). This AHBA-induced attenuation was almost abrogated by co-treated H2L5765834 and daltroban.

AHBA attenuated macrophage M1 polarization
By established mediator treatments for 24 h, primary-cultured marrow-derived macrophages (BMDMs) can be polarized to M1-and M2-like macrophages. 27 BMDMs largely became positive for the M1 marker CD86 in response to the M1 inducers LPS and interferon-b (IFN-b), and co-treated AHBA (30 mM) remarkably reduced the appearance of CD86-positive cells ( Figure 9A). RT-PCR analysis detected that the representative M1-marker genes Cd86, Tnf for tumor necrosis factor alpha (TNF-a) and Nos2 for inducible nitric oxide synthase 2 were highly activated after the M1-polarizing treatment, and the upregulations were largely attenuated by co-treated AHBA. On the other hand, BMDMs turned positive for the M2 marker CD206 in response to the M2 inducers interleukin-4 (IL-4) and IL-13, and co-treated AHBA did not affect iScience Article CD206-positive cell ratios ( Figure 9B). The M2-marker genes Cd206, Ym1 for chitinase 3-like 3 and Arg1 for arginase 1 were transcriptionally upregulated after the M2-inducing treatment, and co-treated AHBA did not alter the M2-maker gene expression. Therefore, AHBA seemed to attenuate M1 polarization without affecting M2 polarization in BMDMs. Both IL-4 and IL-13 react to the receptor complexes containing the IL-4 receptor a, and activate overlapping signaling systems so-called JAK-STAT pathways, in which JAK kinase phosphorylates and activates the signal transducer and activator of transcription (STAT) proteins. 28 Our observation suggested that AHBA exerted no obvious effects on JAK-STAT signaling.
The powerful M1 inducer LPS stimulates cell-surface toll-like receptors (TLRs) and activates several signaling pathways, leading to induction of inflammation-related gene products in macrophages. Of the LPS-inducible M1 polarization-related genes previously reported, 26 10 typical inflammation-related genes were selected for our RT-PCR analysis ( Figure S8). For example, mRNAs derived from the Tnf and Il1b for IL-1b were obviously induced by the LPS treatment for 4 h, and these LPS-induced effects were largely inhibited by co-treated AHBA (30 mM). The AHBA-induced effects were occasionally interrupted by iScience Article co-treated H2L5765834 and daltroban; the interruption efficiencies of the receptor blockers were highly varied among the genes. Consistent with the RT-PCR data, ELISA quantification showed that LPS-induced TNF-a secretion was moderately inhibited by AHBA, and that the receptor blockers attenuated the AHBAinduced inhibition ( Figure S5E). The observations likely suggested that AHBA broadly inhibited M1-related gene expression and thus attenuated M1 polarization.

DISCUSSION
In this study, we examined the mechanism underlying the AHBA-induced reduction in store Ca 2+ content in TGPMs. Based on the data obtained, we propose an N-phos-mediated functional link between steadystate lipid metabolism and autocrine signaling, and the proposed scheme, together with a simplified PC turnover cycle, is illustrated in Figure S7. Leuiller et al. reported that N-phos likely converts LPAs to MAGs in hepatocytes based on lipidomic data from N-phos-deficient knock-in rats. 8 In accordance with their conclusion, AHBA-induced alterations in lipid profiles suggest that N-phos is also mainly responsible for metabolizing LPAs to MAGs in TGPMs (Figure 3). Our proposed scheme has the following sequential events; (1) AHBA inhibits the conversion of LPAs to MAGs, (2) LPAs thus accumulate and weakly stimulate intrinsic LPA receptors coupling to Gi/o, (3) downstream of Gi/o activation, cPLA2 is stimulated by ERKmediated phosphorylation, and AKT is also stimulated in the PI3K-dependent manner, (4) Activated cPLA2 then generates excessive LPCs and PUFAs, (5) accumulated PUFAs are predominantly converted to TXA2, which is further transformed to inactive TXB2, (6) TXA2 stimulates intrinsic TXA2 receptors, (7) TXA2 receptor activation potentiates PLCb activity through Gq stimulation to stimulate steady-state IP 3 generation, (8) finally, increased IP 3 production gradually facilitates IP 3 receptor-mediated Ca 2+ leakage to reduce store Ca 2+ content. Our observations in AHBA-treated TGPMs suggest that TXA2 receptor activation requires LPA receptor activation (Figure 6), and both TXA2 and LPA receptor antagonists had similar inhibitory effects on the AHBA-induced PLC activation and reduction in Ca 2+ stores (Figures 4  and 7). Therefore, TXA2 receptors seem to be mainly responsible for AHBA-induced PLC activation, while the direct contribution of LPA receptors may be negligible. AHBA primarily inhibits dephosphorylation of LPAs, but may fine-readjust a functional link between phospholipid turnover, mediator synthesis, and autocrine signaling in TGPMs. Since LPA receptor subtypes are differentially expressed in various cell types, 17 N-phos inhibition may generally increase intracellular LPA contents for autocrine-activation of endogenous LPA receptors, leading to the modulation of various intracellular signaling in different cell types. However, the predicted modulations may be commonly accompanied by PLC activation because AHBA reduced store Ca 2+ contents in several non-excitable cell types tested ( Figure S1B).
We demonstrated that AHBA attenuated the macrophage-related functions in TGPMs ( Figure 8). Downstream of AHBA-induced LPA receptor activation, ERK and AKT pathways and TXA2 receptor-mediated store Ca 2+ leakage are probably stimulated as mentioned previously. All of the proposed effects would be blocked by H2L5765834, while daltroban would interrupt PLC activation and following Ca 2+ leakage from stores. Both PI turnover and store Ca 2+ handling influence a wide variety of cellular responses. For example, migration and phagocytosis require PLC activity and Ca 2+ -dependent processes for controlling membrane dynamics, 29,30 and H 2 O 2 -producing NOX2 activity is stimulated by Ca 2+ in cardiac myocytes. 31 The AHBA-induced inhibitions of chemotaxis, phagocytosis, and H 2 O 2 generation were largely attenuated by both the LPA and TXA2 receptor blockers at similar potencies (Figure 8), suggesting that the inhibitions are mainly due to PLC activation and/or deranged Ca 2+ stores in AHBA-treated TGPMs.
AHBA inhibited LPS-induced gene expression, likely leading to attenuation of in vitro M1 polarization ( Figures S6 and 9). LPS-induced TLR activation stimulates several Ser/Thr-kinase pathways to activate the transcription factors NF-kB (NF-kB), activator protein 1 (AP-1) and interferon regulatory factors 3 and 5 (IRF3/5), facilitating transcription of inflammation-related genes. 32,33 Activation of NF-kB and AP-1 is regulated by intracellular Ca 2+ (ref. 34 and 35). On the other hand, the PI3K/AKT pathway converges the control of inflammatory responses by functioning for the negative regulation of NF-kB-mediated gene expression in macrophages. 36 In fact, AKT overexpression impairs inflammation-related responses in macrophages, while inflammatory lesions are exacerbated in AKT-deficient mice. 37 Furthermore, in response to LPS-evoked TLR activation, the Tyr-kinase Syk and its downstream effector PLCg are also activated and act as the important signaling regulators. 38 Store Ca 2+ release predicted in the Sky-PLCg pathway is likely interfered by AHBA-induced activation of store Ca 2+ leakage. Therefore, activated AKT signaling and store Ca 2+ leakage may mainly contribute to the inhibition of LPS-induced inflammation-related gene activation in AHBA-treated TGPMs. NF-kB, AP-1, and IRF3/5 differentially boost transcription depending on the ll OPEN ACCESS iScience 26, 107465, August 18, 2023 iScience Article gene, and it may be reflected by the observations that AHBA and receptor blockers exerted divergent effects on the modulation of the LPS-inducible gene expression.
Macrophages are closely associated with tissue inflammation, and our observations suggest that N-phos inhibition broadly attenuates inflammatory functions in macrophages. Therefore, we propose that N-phos is a surefire target for developing anti-inflammatory drugs. In previous studies on sEH, the dual inhibitors that block both N-phos and C-EH were often used, and the role of C-EH in the regulation of inflammation was emphasized. Our present observations suggest the possibility that the pathophysiological role of C-EH might have been overestimated. After all, it is thus important to re-examine the anti-inflammatory role of sEH inhibitors by clearly dissecting the distinct effects induced by N-phos and C-EH inhibitions.

Limitation of the study
The major limitation of this study is only based on assessments using chemical tools. To further confirm our conclusion, it is necessary to prepare N-phos-deficient knock-in macrophage models for comparing pharmacological effects between the mutant and wild-type cells.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

Lead contact
Further information and any related requests should be directed to and will be fulfilled by the lead contact, Hiroshi Takeshima (takeshim@pharm.kyoto-u.ac.jp).

Materials availability
This paper did not generate new unique reagents.
Antibodies were commercial sources described in the STAR methods key resources table.
Data and code availability d All data for evaluating the contributions in the paper are presented in the paper and/or the supplemental information.
d Microarray data based on the ImmGen (Immunological Genome Project) reference set (see key resources  table).
d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Chemicals, primers and mice
The reagents and antibodies used in this study are described in the STAR methods key resources table. Synthetic primers used for PCR analysis are listed in Table S3. C57BL/6J mice were purchased from Shimizu Laboratory Supplies (Kyoto, Japan). All experiments in this study were conducted with the approval of the Animal Research Committee according to the regulations on animal experimentation at Kyoto University.

Preparation of TGPMs
C57BL/6J male mice (9-11 weeks old) were injected i.p. with 1 mL of 30 mg/mL Brewer thioglycolate (Defco) 3 days before cell preparation. The mice were then injected with phosphate-buffered saline (PBS), and the suspended cells were collected from the peritoneal cavity, recovered by low-speed centrifugation, re-suspended in culture medium (RPMI 1640 supplemented with 10% fetal calf serum and penicillin/streptomycin cocktail) and pre-seeded on plastic dishes for 2 h. After non-adherent cells were removed, adherent TGPMs were scraped off with Cell Spatula (Techno Plastic Products Inc., Switzerland) and then seeded on polylysine-coated glass-bottom dishes (MatTek Co.) for imaging analysis after 2026-h culture or on plastic dishes for biochemical assessments after 3-h culture.

Isolation and differentiation of BMDMs
From the femoral bones dissected from male C57BL mice (8-10 weeks old), bone marrow cells were flushed out with cold phosphate-buffered saline (1 mL/femur). After passing through mesh filters, isolated bone marrow cells were cultured in DMEM medium supplemented with glutamine, pyruvate, 10% fetal calf serum, 0.5% penicillin/streptomycin cocktail and 50 ng/mL macrophage colony stimulating factor (PeproTech) for 5 days to differentiate into BMDMs. On culture day 5, M1 polarization was achieved by supplementation with 100 ng/mL LPS (Fujifilm Co., Japan) and 20 ng/mL IFN-g (Fujifilm Co., Japan) for 24 hrs, while M2 polarization was achieved by supplementation with 20 ng/mL interleukin-4 (PeproTech) and 20 ng/mL interleukin-13 (PeproTech) for 24 hrs. M1-and M2-polarizing cells thus generated were subjected to immune-staining and RT-PCR analyses.