Macrophage polarization is linked to Ca2+-independent phospholipase A2β-derived lipids and cross-cell signaling in mice[S]

Phospholipases A2 (PLA2s) catalyze hydrolysis of the sn-2 substituent from glycerophospholipids to yield a free fatty acid (i.e., arachidonic acid), which can be metabolized to pro- or anti-inflammatory eicosanoids. Macrophages modulate inflammatory responses and are affected by Ca2+-independent phospholipase A2 (PLA2)β (iPLA2β). Here, we assessed the link between iPLA2β-derived lipids (iDLs) and macrophage polarization. Macrophages from WT and KO (iPLA2β−/−) mice were classically M1 pro-inflammatory phenotype activated or alternatively M2 anti-inflammatory phenotype activated, and eicosanoid production was determined by ultra-performance LC ESI-MS/MS. As a genotypic control, we performed similar analyses on macrophages from RIP.iPLA2β.Tg mice with selective iPLA2β overexpression in β-cells. Compared with WT, generation of select pro-inflammatory prostaglandins (PGs) was lower in iPLA2β−/−, and that of a specialized pro-resolving lipid mediator (SPM), resolvin D2, was higher; both changes are consistent with the M2 phenotype. Conversely, macrophages from RIP.iPLA2β.Tg mice exhibited an opposite landscape, one associated with the M1 phenotype: namely, increased production of pro-inflammatory eicosanoids (6-keto PGF1α, PGE2, leukotriene B4) and decreased ability to generate resolvin D2. These changes were not linked with secretory PLA2 or cytosolic PLA2α or with leakage of the transgene. Thus, we report previously unidentified links between select iPLA2β-derived eicosanoids, an SPM, and macrophage polarization. Importantly, our findings reveal for the first time that β-cell iPLA2β-derived signaling can predispose macrophage responses. These findings suggest that iDLs play critical roles in macrophage polarization, and we posit that they could be targeted therapeutically to counter inflammation-based disorders.

T1D is a consequence of autoimmune destruction of cells, involving activation of cellular immunity and inflammation that lead to infiltration of islets by leukocytes (1). Lipid signaling is recognized as a key modulator of inflammation and immune responses (2). For instance, arachidonic acid (AA) and oxidized lipids [i.e., prostaglandins (PGs) and leukotrienes (LTs)] (3) generated via its metabolism can trigger immune responses leading to -cell death (4,5). They are generated via cyclooxygenases (COXs) and lipoxygenases (LOs), subsequent to phospholipase A 2 (PLA 2 )-mediated hydrolysis of AA (6-12) from membrane glycerophospholipids.
Given the importance of macrophages to T1D development and evidence for a role of iPLA 2  in modulating macrophage polarization and function, we sought to identify iDLs generated by activated macrophages, which could be targeted to counter autoimmune/inflammatory-based disorders. To facilitate our assessments, we utilized targeted/ quantitative MS analyses of eicosanoids produced by macrophages isolated from WT (M WT ) and iPLA 2  / (KO, M KO ) mice. As a control for the genotype, we performed similar analyses with macrophages isolated from RIP. iPLA 2 .Tg [transgenic (Tg), M Tg ] mice, which selectively overexpress iPLA 2  in -cells (43,44).
We found that iPLA 2  deficiency led to an attenuated pro-inflammatory and amplified anti-inflammatory lipid profile that was consistent with a macrophage M2 phenotype. In contrast, M Tg were found to exhibit an opposite landscape, encompassing an exaggerated pro-inflammatory and attenuated anti-inflammatory lipid profile that was associated with a macrophage M1 phenotype. These findings suggest that iPLA 2  modulates the inflammatory lipid profile and raises the intriguing possibility that increased expression of iPLA 2  in the -cells can confer increased susceptibility of macrophages to activation.

Animals
Mouse breeders obtained from Dr. John Turk (Washington University School of Medicine, St. Louis, MO) were used to generate colonies of WT, RIP.iPLA 2 .Tg (selectively overexpressing iPLA 2  in -cells only), and global iPLA 2 -KO mice at the University of Alabama at Birmingham. Tg founders (TG1 line) were mated with WT C57BL/6J mice (Jackson Laboratory) to generate RIP.iPLA 2 .Tg (Tg) mice and WT mice, and male and female iPLA 2  +/ pairings were used to generate iPLA 2  / (KO) and iPLA 2  +/+ (WT) mice, as previously described (43,45). Prior to experimentation, the mice were genotyped, as described (44), utilizing primers described in Table 1. Animal experiments were conducted according to approved IACUC guidelines at the University of Alabama at Birmingham. As our earlier studies suggested differences between the WT generated by the two schemes, data from the KO and Tg genotypes were compared against their corresponding WT littermates.

Isolation and culture of peritoneal macrophages
Mice (7-8 weeks of age) were euthanized by CO 2 inhalation and cervical dislocation. Peritoneal macrophages were obtained by filling the peritoneal cavity with 5 ml of cold PBS containing 2% FBS, massaging gently, and withdrawing the cell-containing solution. Cells were pelleted at 300 g for 5 min and resuspended in growth medium [Eagle's minimum essential medium (Sigma-Aldrich; M0894), 2.0 mg/ml sodium bicarbonate (Fisher Scientific; BP328-500), 2 mM l-glutamine (Life Technologies; 25030-081), 100 U/ml penicillin-100 g/ml streptomycin [Life Technologies; 15140-122), and 10% heat-inactivated FBS (Life Technologies; 16000044)] supplemented with 10% L929 cell-conditioned medium (source of M-CSF). Macrophages from a single collection were sufficient to seed six 60 mm nontreated culture dishes. Adherent macrophages appeared after 16 h of culture. All experiments were performed with expanded freshly isolated peritoneal macrophages under classical and alternate activation conditions, as described below. Macrophages isolated from WT, iPLA 2  / (KO), and RIP.iPLA 2 .Tg (Tg) mice are designated as M WT , M KO , and M Tg , respectively.

Macrophage activation
Macrophage activation was accomplished according to previously published methods (42). For classical activation, macrophages were treated with 15 ng/ml recombinant IFN (R&D Systems; 485-MI-100) for 8 h in growth medium followed by addition of 10 ng/ml ultrapure LPS (InvivoGen; tlrl-3pelps) and incubated for 16 h at 37°C. For alternative activation, macrophages were treated with 8 ng/ml recombinant IL-4 (R&D Systems; 404-ML-010) in growth medium for 16 h. Naïve macrophages, which received no activation stimuli, were maintained in growth medium with no additional treatment. In some experiments, the macrophages were pretreated for 1 h with either a secretory phospholipase A 2 (sPLA 2 ) (LY315920, 10 M; Cayman Chemical) or cytosolic (c)PLA 2  (CAY 10502, 50 nM; Cayman Chemical) inhibitor prior to activation. The inhibitors were present during the entire activation period.

Macrophage mRNA target analyses
Macrophages cultured in 60 mm non-tissue culture-treated dishes were lysed in 1 ml of TRIzol (Life Technologies; 15596-026). Total RNA was prepared and purified using RNeasy mini kits (QIAGEN; 74104), and 1 g RNA was converted to cDNA using the Superscript III first strand synthesis system (Life Technologies; 18080-051), according to manufacturer's instructions. The cDNA was diluted 10-fold and used as template in conventional or real-time quantitative PCR (qPCR). cDNA transcripts were amplified with primers ( Table 2) designed using NCBI Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Real-time qPCR was carried out using SYBR Select Mastermix (Life Technologies; 4472908) according to the manufacturer's instructions. Relative gene expression levels were determined using the 2 Ct method.

Eicosanoid preparation
Eicosanoids were extracted using a modified extraction process, as previously described (46,47). Medium from cells (2 ml) was combined with an internal standard mixture comprised of 10% total volume methanol (200 l) and glacial acetic acid (10 l) before spiking with internal standard (20 l) containing the following deuterated eicosanoids (2 pmol/l, 40 pmol total) (all standards purchased from Cayman Chemicals): (d 4 ) 6-keto-

Analysis of eicosanoids by UPLC ESI-MS/MS
Eicosanoids were separated using a Shimadzu Nexera X2 LC-30AD coupled to a SIL-30AC auto injector, coupled to a DGU-20A5R degassing unit in the following way [as previously described by us (47)]. A 14 min reversed phase LC method utilizing an Acentis Express C18 column (150 × 2.1 mm, 2.7 m) was used to separate the eicosanoids at a 0.5 ml/min flow rate at 40°C as we previously described (47). The column was equilibrated with 100% solvent A [acetonitrile:water:formic acid (20:80:0.02, v/v/v)] for 5 min and then 10 l of sample were injected. Solvent A (100%) was used for the first 2 min of elution. Solvent B [acetonitrile:isop ropanol:formic acid (20:80:0.02, v/v/v)] was increased in a linear gradient to 25% solvent B at 3 min, to 30% at 6 min, to 55% at 6.1 min, to 70% at 10 min, and to 100% at 10.10 min. Solvent B (100%) was held constant until 13.0 min, where it was decreased to 0% solvent B and 100% solvent A from 13.0 min to 13.1 min. From 13.1 min to 14.0 min, solvent A was held constant at 100%. All solvents were purchased from Fischer Scientific. Eicosanoids were analyzed via MS using an AB Sciex Triple Quad 5500 mass spectrometer, as previously described by us (47). Q1 and Q3 were set to detect distinctive precursor and product ion pairs. Ions were fragmented in Q2 using N 2 gas for collisionally induced dissociation. Analysis used multiple-reaction monitoring in negative-ion mode. Eicosanoids were monitored using precursor → product MRM pairs. The mass spectrometer parameters used were: curtain gas, 20 psi; CAD, medium; ion spray voltage, 4,500 V; temperature, 300°C; gas 1, 40 psi; gas 2, 60 psi (declustering potential, entrance potential, collision energy, and cell exit potential varied per transition). MRM transitions with corresponding declustering potentials, collision energies, entrance potentials, and collision cell exit potentials are shown in supplemental Table S1.

Selectivity of transgene function
As described (43), to generate Tg mice with selective overexpression of iPLA 2  in -cells only, rat iPLA 2  cDNA was inserted downstream of the rat insulin promoter (RIP) at a site within the rabbit globin gene sequence. As such, transcription of the sequence encoding iPLA 2  is under control of RIP, and transgenic overexpression of iPLA 2  is expected only in cells that express insulin, i.e., pancreatic islet -cells, but not in other cells (i.e., macrophages). To verify that induction of the transgene was specific to -cells, PCR analyses were performed using islet and macrophage cDNA as a template with two pairs of primers. One pair amplified the sequence in the internal control fatty acid-binding protein, intestinal (Fabpi) gene, and the primer sequences were (Fabpi 5′) cctccggagagcagcgattaaaagtgtcag and (Fabpi 3′) tagagctttgccacatcacaggtcattcag (expected product size, 450 bp). The other primer pair amplified a sequence that spanned the junction of iPLA 2  and globin cDNA in the TG construct. The primer sequences were (TG 5′) ctaggctcagacatcatgctggacgaggt and (TG 3′) aagatctcagtggtatttgtgagccaggg (expected product size, 200 bp). Subsequently, the islets and macrophages were processed for insulin content and iPLA 2  immunoblotting (1° antibodies: iPLA 2 , E-8, sc-166616; tubulin, TU-02, sc-8035) analyses, as described (48,49).

Statistical analysis
Lipidomics samples were analyzed via a multivariate approach using SPSS. In experiments with more than two conditions, samples were analyzed by ANOVA on the R platform with a Tukey's post hoc test. In experiments using two conditions, samples were analyzed using a Student's t-test. Presented data are mean ± SEM and, where appropriate, as fold-change relative to control (vehicle). Values of P < 0.05 were considered significant.

iPLA 2  / model verification
As previously reported (43,45), DNA was generated from tail clips and progeny were genotyped by PCR analyses. The primers used for the WT or for the disrupted iPLA 2  / (KO) sequence were expected to yield product sizes of 1,400 or 400 bp, respectively. Consistently, only a 1,400 bp product in WT (lanes 1 and 2) and a 400 bp product in KO (lanes 3 and 4) were evident (Fig. 1).
The iPLA 2  / group is referred to as KO and macrophages isolated from these mice as M KO , and macrophages from WT, M WT .

Comparison of basal lipid species in WT and KO
The MS protocol identified several lipid species (Table 3), including pro-inflammatory and anti-inflammatory PGs, LTs, HETEs, DHETs, EETs, and specialized pro-resolving lipid mediators (SPMs). Analyses of lipid production under basal conditions revealed no significant differences in the abundances of detected lipids between WT and KO (supplemental Table S2). These findings suggest an overall similar profile in M WT and M KO production of lipids. KO AA, subsequent to its release from membrane glycerophospholipids, can be metabolized by COXs and LOs to generate oxidized bioactive products, designated eicosanoids. Some of these lipids are considered to be proinflammatory. They include several PGs, LTs, and HETEs (Table 3). Further, epoxidation of AA by cytochrome P450 epoxygenases leads to the generation of EETs, which are then hydrolyzed by soluble epoxide hydrolases to proinflammatory DHETs. Comparison of macrophage proinflammatory lipid production in response to classical activation revealed an increase in pro-inflammatory PGs generated by both M WT and M KO in response to IFN + LPS ( Fig. 2A). However, the increase in PGE 2 in the KO group was of lower magnitude relative to the WT group. Overall, a nearly 40% lower production of pro-inflammatory PGs (pool of PGs presented in Fig. 2A) by M KO relative to M WT was evident (Fig. 2B). However, production of other pro-inflammatory lipid species, including LTs ( Fig. 3), HETEs (supplemental Fig. S1A), or DHETs (±14,15 and ±11,12; supplemental Fig. S2A), by M KO was similar to M WT , and the production of (±)8,9-DHET by M KO was higher relative to M WT . These observations suggest that iPLA 2  promotes production of select pro-inflammatory PGs in macrophages.

Comparison of anti-inflammatory lipid production from activated WT and M KO
In addition to generation of pro-inflammatory lipids, metabolism of AA can generate anti-inflammatory lipids (Table 3), such as EETs, which have been reported to manifest anti-inflammatory properties (50). Further, dihomo--linolenic acid (DHGLA) can be converted to PGE 1 , and SPMs are generated via metabolism of AA (lipoxin A 4 ) or DHA (resolvins D 1 and D 2 ). With classical activation, antiinflammatory PGs increased similarly in the WT and KO (Fig. 4A). In contrast, production of resolvin D 2 by M KO was significantly higher (Fig. 4B) relative to M WT ; whereas, production of EETs was similar between WT and KO (supplemental Fig. S2B). Further, alternative activation did not promote differential changes in the production of SPMs (data not shown). These observations suggest that iPLA 2  can modulate macrophage production of select SPMs.
As expected, real-time qPCR analyses (Fig. 5A) confirmed a relative absence of iPLA 2  mRNA in M KO under both basal and activated conditions. iPLA 2  mRNA in M WT was modestly, but not significantly, affected by classical activation (approximately 25% increased) and alternative activation (approximately 25% decreased) relative to basal expression. We previously demonstrated that classical activation of macrophages promotes an M1 inflammatory phenotype, whereas alternative activation favors an M2 anti-inflammatory phenotype (42). To confirm that macrophages used in the present lipid analyses exhibited the expected phenotype, they were stimulated with IFN + LPS or IL-4 and the macrophages were assessed for expression of markers for M1 [arginase 2 (Arg2)] and M2 [mannose receptor C type 1 (MRC1)], respectively, by real-time qPCR. Activation induced the corresponding markers in M WT and M KO ; however, relative expression of Arg2 was reduced (Fig. 5B) and MRC1 increased (Fig. 5C) in M KO in comparison with M WT . These findings are consistent with the reduced pro-inflammatory lipid profile in the KO. However, induction of Arg2 and MRC1 in M WT and M KO by alternative and classical activation, respectively, was similar (supplemental Fig. S3A, B).

RIP.iPLA 2 .Tg (Tg) model verification
In view of the observed effects on macrophage lipid production associated with iPLA 2 -deficiency and literature evidence for intercellular communication between inflamed sites and immune cells, we explored the possibility that signals generated by -cells could manifest effects in macrophages. To address this, we compared lipid production from M Tg and M WT littermates. We previously demonstrated that the Tg mice overexpress iPLA 2  selectively in -cells (44). To assess the genotype of the RIP.iPLA 2 .Tg (Tg), PCR analyses were performed in the presence of two sets of primers expected to yield product sizes of 450 and 200 bp for the WT (Fig. 6, lanes  1-3) and the RIP.iPLA 2 .Tg (Fig. 6, lanes 4-6), respectively. Such analyses (Fig. 6) revealed only a 450 bp band in the WT and a 200 bp in the Tg. The RIP.iPLA 2 .Tg group is referred to as Tg and macrophages isolated from these mice as M Tg .

Comparison of basal lipid species in WT and Tg
Because the KO and Tg mice were derived through different breeding schemes, the Tg mice were compared with their own littermate WT mice. Lipidomics analyses of basal production of lipids revealed no significant differences in the abundances of lipids between M WT and M Tg (supplemental Table S3). These findings suggest that similar to M KO , inherent production of lipids by M Tg is unaffected.

Comparison of pro-inflammatory lipid production from activated WT and M Tg
Classical activation of the macrophages promoted higher production of select pro-inflammatory PG lipids (6-ketoPGF 1  and PGE 2 ) by M Tg relative to M WT (Fig. 7A). Though PGD 2 production by M Tg was decreased, there was an overall 2-fold higher production of pro-inflammatory PGs (pool of PGs presented in Fig. 7A) by M Tg relative to M WT (Fig. 7B). Further, while LTD 4 , LTC 4 , and LTE 4 production was similar in M WT and M Tg (Fig. 8), LTB 4 was significantly elevated in M Tg relative to M WT . However, production of HETEs (supplemental Fig. S1B) or DHETs (supplemental Fig. S4A) was not significantly different between M WT and M Tg . These observations suggest that M Tg are primed to respond to classical activation.

Comparison of anti-inflammatory lipid production from activated WT and M Tg
Classical activation increased PGE 1 production by M Tg relative to M WT (Fig. 9A), but the production of SPMs by M WT and M Tg was similar (data not shown). Alternative activation promoted a 3-fold increase in resolvin D2 production by M WT ; however, its production by M Tg was unchanged from basal production (Fig. 9B). EETs were not significantly different between M WT and M Tg (supplemental Fig. S4B). These observations suggest that M Tg are predisposed to differential generation of PGE 1 , derived from DHGLA, and select SPMs.

Impact of sPLA 2 and cPLA 2  on macrophage lipid profiles
Noting that other PLA 2 s may also contribute to the altered lipid profile, we sought to determine whether sPLA 2 s or cPLA 2 s contribute to the observed changes in eicosanoid production by M Tg . The C57BL/J6 macrophages are reported to express GV and GX sPLA 2 (51, 52) and cPLA 2  (16), but not GIIA sPLA 2 (51,(53)(54)(55)(56). mRNA analyses revealed that sPLA 2 mRNA in M Tg was not altered under basal or activated conditions, whereas cPLA 2 mRNA was higher under basal but similar under activated conditions relative to M WT (Fig. 10). We next determine whether the lipid profiles of M WT and M Tg were modulated by inhibitors of sPLA 2 [LY315920 (57), which inhibits several sPLA 2 s including GIIA, GV, and GX] or cPLA 2  [CAY 10502 (58)]. Lipidomics analyses revealed that neither basal ( Table 4) nor activated (Table 5) production of lipids by M WT and M Tg was affected by LY315920 or CAY 10502. Comparisons between M WT and M Tg exposed to LY315920 revealed a higher basal production of several pro-inflammatory PGs and LTE 4 by M Tg relative to M WT (Table 4). In contrast, basal production of lipids by M WT and M Tg exposed to CAY 10502 was unaffected. Classical activation resulted in increases in lipid production by both M WT and M Tg (Table 5), with production by M Tg significantly higher relative to M WT . Such increases were maintained in the presence of LY315920 or CAY 10502. These findings suggest that GV and GX sPLA 2 s and cPLA 2  are not likely contributors to the inflammatory lipid profile in activated M Tg .

Phenotype comparison of M WT and M Tg
In view of the observations that M Tg exhibit a higher inflammatory state, we assessed macrophage phenotype with activation. We found that classical activation of M Tg promoted a greater induction of the M1 marker Arg2 (Fig.  11A), while the M2 marker MRC1 was reduced with alternative activation (Fig. 11B) relative to M WT . Whereas, induction of Arg2 in M WT and M Tg by alternative activation was not different, induction of MRC1 by classical activation was reduced in the M Tg (supplemental Fig. S3C, D). These findings are consistent with a heightened proinflammatory and reduced anti-inflammatory lipid profile in M Tg . No differences were observed in the expression of ALOX-12 or PTGS2 mRNA, which encode 12-LO and COX2, respectively, in M WT and M Tg under basal (Fig.  12A) or activated (Fig. 12B, C) conditions. Under both basal and activated conditions, iPLA 2  mRNA was modestly and similarly increased in M Tg (Fig. 12D) relative to M WT . To determine whether the transgene is induced in M Tg , PCR, insulin content, and iPLA 2  immunoblotting analyses were performed in islets and macrophages isolated from WT and Tg mice. As expected, transgene expression was evident in both islets and macrophages from Tg mice but not from WT mice (Fig. 13A). However, as islet -cells but not macrophages produce insulin (Fig. 13B), overexpression of iPLA 2  protein was only detected in the islets and not in the M Tg relative to M WT (Fig. 13C). These findings suggest that signals generated by iPLA 2 overexpressing -cells, upon appropriate stimulation, prime macrophages to adopt an M1 phenotype, independent of transgene leakage.
Our collection of earlier findings suggests that iDLs are key contributors to inflammation, -cell death, and subsequent development of T1D. While eicosanoids have been linked with inflammation development (46,83), to date, very little is known about the eicosanoid profile of activated macrophages and whether iDLs contribute to an inflammatory state. Therefore, in view of our findings of a link between iPLA 2  and macrophage polarization and responses, we utilized a targeted lipidomics approach with  commercially available eicosanoid standards and WT, iPLA 2  / (KO), and RIP.iPLA 2 .Tg (Tg) mouse models to identify iDLs generated by activated macrophages. Our studies revealed two salient and previously undescribed features: 1) the predominant anti-inflammatory M2 phenotype due to iPLA 2 -deficiency is associated with reduced generation of select pro-inflammatory eicosanoids and increased production of the SPM, resolvin D2; and 2) greater production of select pro-inflammatory eicosanoids and reduced resolvin D2 from M Tg is associated with macrophage polarization toward the inflammatory M1 phenotype. These findings, for the first time, reveal an association between selective changes in eicosanoids and SPMs with macrophage polarization and, further, that the relevant lipid species are modulated by iPLA 2  activity. Most importantly, our findings unveil the possibility that events triggered in -cells can modulate macrophage responses. This is supported by observations in the RIP.iPLA 2 .Tg model with a select overexpression of iPLA 2  in -cells; yet, macrophages from these mice exhibit an opposite spectrum relative to M KO . To our knowledge, this is the first demonstration of intercellular signaling, in particular iPLA 2 -derived intercellular signaling, originating from -cells affecting macrophage function.
iPLA 2  deficiency, verified by genotype and real-time qPCR analyses, was associated with a significant reduction in the pool of pro-inflammatory PGs with a selective decrease in the abundance of PGE 2 . Other recognized inflammatory lipids, including LTs and DHETs, were not significantly affected by iPLA 2  deficiency. While there were no significant differences in the induction of antiinflammatory PGs by M WT and M KO , production of resolvin D2 by M KO in response to classical activation was significantly higher relative to M WT . These lipid changes are associated with decreased Arg2 (M1 marker) and increased MRC1 (M2 marker) in M KO . Resolvin D 2 arises from the metabolism of DHA, and lipids containing DHA are substrates for iPLA 2  (84). However, sPLA 2 s exhibit greater activity toward lipids containing DHA than does iPLA 2  (84), and the group V and IIa sPLA 2 s are activated during inflammation (85,86). Consistent with this premise, it has been suggested that at the onset of inflammation, iPLA 2  activity predominates and leads to increased production of inflammatory lipids (2). However, a subsequent resolution phase is unmasked during which groups IIA and V sPLA 2 s are activated. Taken together with our present findings, it is likely that iPLA 2  deficiency is permissive for sPLA 2 activation to affect resolution through generation of resolvin D 2 . This latter phase appears to occur in the absence of increased expression of sPLA 2 s, at least at the time point studied. Acknowledging that in vitro systems are artificial and limited, we speculate that the in vivo inflammatory landscape gives rise to a more dramatic lipid profile.
Intriguingly, we found that the production of proinflammatory lipids by M Tg was increased. These included 6-keto PGF 1 , PGE 2 , and LTB 4 . Concurrently, there was increased production of anti-inflammatory PGE 1 but decreased production of PGD 2 , suggesting differential regulation of pathways leading to generation of these PGs during inflammation. The increase in PGE 1 most likely reflects an attempt at resolution (87). Interestingly, in contrast to M KO , resolvin D2 production by M Tg was not altered with classical activation, but was significantly reduced under alternative activation. This suggests a decrease in the metabolism of DHA, which has been reported to increase susceptibility to inflammation (88,89). In our earlier study, we observed that the polarization toward M2 macrophage phenotype with iPLA 2  deficiency was not a direct effect on iDL production, but was most likely due to the overall decrease in an inflammatory state, which is permissive for the evolution of anti-inflammatory processes (42). The differential production of resolvin D2 by M KO and M Tg supports this possibility; in the absence of iPLA 2 , there is a reduced inflammatory state, which unmasks pathways leading to production of resolvin D2; whereas a heightened inflammatory status in M Tg mitigates such processes. Overall, our lipidomics analyses suggest a  Fig. 9. Production of anti-inflammatory PGE 1 and SPMs M WT and M Tg . Peritoneal macrophages isolated from 8-week-old WT and Tg mice were treated with vehicle [control (C)], IFN + LPS, or IL-4 and the media collected for lipidomics analyses. The data represent activated (A) fold-changes in lipids relative to C. Data are mean ± SEM determined from four independent experiments. A: PGE 1 ± IFN + LPS. Control (pmol lipid/1e +06 ) M WT and M Tg : 0.033 ± 0.011 and 0.017 ± 0.003. B: SPMs ± IL-4. Control (pmol lipid/1e +06 ) M WT and M Tg : resolvin D2, 0.561 ± 0.073 and 1.074 ± 0.524; resolvin D1, 0.134 ± 0.042 and 0.136 ± 0.039; lipoxin A 4 , 0.050 ± 0.008 and 0.050 ± 0.006. *M Tg significantly different from M WT , P < 0.05. pro-inflammatory status in M Tg and, consistently, polarization of M Tg favored an M1 over M2 phenotype.
Examination of potential involvement of other PLA 2 s expressed in macrophages revealed an absence in induction of sPLA 2 GV or cPLA 2  in activated M Tg . Because M Tg exhibited more profound lipid changes with classical activation, we tested the impact of inhibiting sPLA 2 s or cPLA 2  on the lipid profile under this condition. Production of eicosanoids by M WT or M Tg under basal or activated conditions were found to be similar following exposure to CAY 10502, which inhibits cPLA 2 , or LY315920, which inhibits mGIB, mGIIA, mGIID, mGIIE, mGV, and mGX sPLA 2 (personal communication). In the presence of LY315920, basal production of pro-inflammatory PGs by M Tg was found to be higher relative to M WT , most likely a reflection of modest decreases in their abundances in the WT group. Nevertheless, the inability of either CAY 10502 or LY315920 to mitigate the select lipid changes in M Tg relative to M WT suggests that neither sPLA 2 s nor cPLA 2  contribute to the pro-inflammatory lipid profile in M Tg during the activated state during the time course of our studies. Additionally, expression of ALOX-12 and PTGS2 mRNA, which encode 12-LO and COX2 proteins, respectively, were similar in the M WT and M Tg , suggesting that they likely are not responsible for the differences between the two groups.
These studies add to a growing body of literature indicating a role for iPLA 2  in inflammatory disease, particularly in T1D. We previously reported that glucose-stimulated insulin secretion (GSIS) is accompanied by iPLA 2 -mediated hydrolysis of AA from -cell membrane glycerophospholipids and that overexpression of iPLA 2  leads to higher GSIS (6,7,9). As reported in those studies, the RIP.iPLA 2 .Tg mice exhibited lower circulating glucose levels and amplified GSIS insulin (43). Further, whereas survival of -cells (INS-1 or in RIP.iPLA 2 .Tg islets) was unaffected by iPLA 2  overexpression under basal conditions, ER stress and proinflammatory cytokines amplified apoptosis of the -cells (44,48,61,(90)(91)(92). It is important to note that the impact of iPLA 2  on insulin secretion is evident within 1 h of glucose stimulation and is not associated with an increase in iPLA 2  expression. In contrast, its effect on -cell survival is manifested after several hours of exposure to ER stressors or pro-inflammatory cytokines and is associated with induction of iPLA 2  (44,48,61,62,90,92,93).
In this regard, it was intriguing to find that iPLA 2  mRNA was modestly higher in M Tg relative to M WT . However, this did not translate to greater protein expression or increased production of lipids by resting M Tg . In contrast, classical activation led to significantly higher production of the select lipids by M Tg , and this may be a consequence of increased availability of substrates derived through iPLA 2 -mediated hydrolysis of sn-2 fatty acyl substituents for metabolism by downstream enzymes. This is analogous to previous reports of minimal effects of higher iPLA 2  expression in -cells under basal conditions, but amplified in response to stress (43,44), suggesting that the iPLA 2  overexpressing -cells are sensitized to respond to stimulation. Our current observations raise the possibility that there may be cross-talk between -cells and macrophages in vivo, to the extent that M Tg are predisposed to exhibiting enhanced responses to classical activation leading to an inflammatory landscape. While transfer of stress-induced signals between cells of different types has been suggested (60,94,95), to our knowledge, this is the first demonstration of lipid signaling generated by -cells having an impact on an immune cell that elicits inflammatory consequences.
While the present study did not discern the intercellular signaling molecules in RIP.iPLA 2 .Tg mice, it is likely that they are derived directly or indirectly through an increase in iDLs, as a consequence of increased iPLA 2  expression. In support of this, recent studies reveal that inhibition of iPLA 2  activity mitigates inflammatory processes in different cell systems (18,78,(96)(97)(98), raising the possibility that bioactive lipids generated through -cell-iPLA 2  activity or factors arising from the effects of such lipids may serve as potential candidate intercellular signals. For instance, insulin secretion from -cells is accompanied by a parallel iPLA 2 -mediated hydrolysis of AA from -cell membranes and generation of PGE 2 , which is mitigated by selective inhibition of iPLA 2  (7, 99). Mice with selective overexpression of iPLA 2  in - cells exhibit lower basal blood glucose (43). Further -cells in these mice express higher pPERK and generate more ceramides under basal conditions (44). This leads to the possibility that circulating insulin, PGE 2 or other iDLs, ER stress factors, or ceramides originating from -cells could prime macrophages to respond more robustly to classical activation. This is supported by the greater in-duction of Arg2 and lower induction of MRC1 in M Tg , which sets the table for the onset and progression of an inflammatory state. Further detailed studies are needed to identify the specific signals from -cells that affect macrophages and they are currently underway. The importance of intercellular signaling originating from -cells is realized in view of findings that -cells in the diabetes-prone NOD mice express higher iPLA 2  during the prediabetic phase relative to -cells in spontaneous diabetes-resistant models (20). This early phase encompasses spontaneous evolution of inflammatory signaling leading to the onset of insulitis. Our findings offer the possibility that higher activation of iPLA 2  in the -cells of  NOD mice during this phase contributes signals that can be transmitted to infiltrating leukocytes to initiate immune responses that work in concert with events in the -cells to amplify the onset and progression of inflammation that subsequently induces -cell death and T1D. As discussed in a recent review (100), inflammation and -cell dysfunction are also associated with the development of insulin resistance, and iPLA 2  has been implicated in this process as well, suggesting that signaling between -cells and immune cells must also be considered as a contributor to the development of T2D.
In summary, our findings reveal that iPLA 2  modulates macrophage production of select lipids, which collectively are recognized to be associated with an enhanced inflammatory state. The selective nature of the lipid changes suggests that the impact of iPLA 2  is not broad but is specific to certain pathways that lead to the generation of such lipids. We find that induction of an M1 inflammatory macrophage phenotype is associated with increased production of iPLA 2 -derived pro-inflammatory lipids and that responses in macrophages can be modulated by iPLA 2 derived signals of -cell origin. These findings raise the importance of assessing more carefully the mechanisms governing modulation of the function of immune cells constituting an inflammatory landscape. Fig. 13. Transgene verification. Pancreatic islets and peritoneal macrophages were isolated from 8-week-old WT and Tg mice for the following analyses: A: Transgene PCR. Analyses were performed using cDNA from WT or RIP.iPLA 2 .TG (Tg) mice, and primers that amplify sequence that either spans junction between iPLA 2  and globin. cDNA (200 bp product, lower band) or is within internal control Fabpi (450 bp product, upper band). Lane L, molecular weight ladder; lane N, control reaction without template. B: Insulin content was determined by ELISA using acid-extracted samples. C: iPLA 2  and tubulin (loading control) immunoblotting were performed using 30 g protein. The data were obtained from two independent replicates.