Group X Secreted Phospholipase A2 Releases ω3 Polyunsaturated Fatty Acids, Suppresses Colitis, and Promotes Sperm Fertility*

Within the secreted phospholipase A2 (sPLA2) family, group X sPLA2 (sPLA2-X) has the highest capacity to hydrolyze cellular membranes and has long been thought to promote inflammation by releasing arachidonic acid, a precursor of pro-inflammatory eicosanoids. Unexpectedly, we found that transgenic mice globally overexpressing human sPLA2-X (PLA2G10-Tg) displayed striking immunosuppressive and lean phenotypes with lymphopenia and increased M2-like macrophages, accompanied by marked elevation of free ω3 polyunsaturated fatty acids (PUFAs) and their metabolites. Studies using Pla2g10-deficient mice revealed that endogenous sPLA2-X, which is highly expressed in the colon epithelium and spermatozoa, mobilized ω3 PUFAs or their metabolites to protect against dextran sulfate-induced colitis and to promote fertilization, respectively. In colitis, sPLA2-X deficiency increased colorectal expression of Th17 cytokines, and ω3 PUFAs attenuated their production by lamina propria cells partly through the fatty acid receptor GPR120. In comparison, cytosolic phospholipase A2 (cPLA2α) protects from colitis by mobilizing ω6 arachidonic acid metabolites, including prostaglandin E2. Thus, our results underscore a previously unrecognized role of sPLA2-X as an ω3 PUFA mobilizer in vivo, segregated mobilization of ω3 and ω6 PUFA metabolites by sPLA2-X and cPLA2α, respectively, in protection against colitis, and the novel role of a particular sPLA2-X-driven PUFA in fertilization.

Among the sPLA 2 isoforms, group X sPLA 2 (sPLA 2 -X) has the highest ability to hydrolyze phosphatidylcholine (PC), a major phospholipid in the outer leaflet of the plasma membrane (16 -18). Because of this property, most previous studies have postulated that sPLA 2 -X promotes inflammation by driv-ing AA metabolism. Indeed, mice deficient in sPLA 2 -X (Pla2g10 Ϫ/Ϫ ) are refractory to pulmonary and cardiovascular disorders in association with reduced eicosanoid levels (2, 19 -22). In contrast, overexpression of sPLA 2 -X in cultured macrophages elicits anti-inflammatory responses (23). Furthermore, adoptive transfer of Pla2g10 Ϫ/Ϫ bone marrow (BM) cells into LDL receptor-null (Ldlr Ϫ/Ϫ ) mice exacerbates whereas that of human PLA2G10-Tg (PLA2G10 tg/ϩ ) BM cells ameliorates atherosclerosis and associated Th1 immunity (24). These observations suggest that sPLA 2 -X also has anti-inflammatory roles. Moreover, lipidomics studies of sPLA 2 -X-treated cells or lipoproteins in vitro have demonstrated the release of 3 PUFAs in addition to 6 AA (25,26). However, the ability of sPLA 2 -X to release 3 PUFAs and the resulting physiological outcomes have not been investigated in vivo. Here, we show that sPLA 2 -X releases 3 PUFAs in vivo, thereby suppressing colitis and facilitating fertility in the respective tissues where it is highly expressed.
Histology and Immunohistochemistry-Formalin-fixed tissues were embedded in paraffin, sectioned, mounted on glass slides, deparaffinized in xylene, and rehydrated in ethanol with increasing concentrations of water. The tissue sections (4 m thick) were incubated with Target Retrieval Solution (Dako, Glostrup, Denmark) as required, incubated for 10 min with 3% (v/v) H 2 O 2 , washed three times with phosphate-buffered saline (PBS) for 5 min each, incubated with 5% (w/v) skim milk in PBS for 30 min, washed three times with PBS for 5 min each, and incubated with rabbit antiserum for mouse sPLA 2 -X at 1:500 dilution in PBS overnight at 4°C. The sections were then treated with a CSA system staining kit (Dako) with diaminobenzidine substrate, followed by counterstaining with hematoxylin and eosin.
Quantitative RT-PCR-Total RNA was extracted from tissues or cells using TRIzol reagent (Invitrogen). First-strand cDNA synthesis was performed using a high capacity cDNA reverse transcriptase kit (Applied Biosystems, Foster City, CA). PCR was carried out using Power SYBR Green or TaqMan gene expression assay (Applied Biosystems) on the ABI7700 real time PCR system (Applied Biosystems), as described previously (4 -6, 8). The probe/primer sets used are listed in Table 1. Gapdh (4352339E; Applied Biosystems) was used as an internal control.

Name
Assay no.
ware. Circulating blood cells were analyzed by the clinical blood cell analyzer Vetscan HMII (Abaxis, Union, CA). Microarray-Total RNA was purified using the RNeasy mini kit (Qiagen, Venlo, Netherlands). Microarray analysis was carried out according to the manufacturer's protocol (Agilent Technologies, Santa Clara, CA), as described previously (6,8). In brief, the quality of RNA was assessed with a 2100 Bioanalyzer. cRNA targets were synthesized with a low input Quick-Amp labeling kit. Samples were hybridized to the Whole Mouse Genome microarray kit (4x44K), washed, and then scanned using a SureScan Microarray Scanner. Microarray data were analyzed with Feature Extraction software and then imported into GeneSpring GX software. Probes were normalized by quantile normalization among all microarray data. The GEO accession numbers for the microarrays are GSE77336 and GSE77144.
Measurement of Serum Immunoglobulin (Ig) Levels-Serum titers of IgM, IgG 1 , IgG 2 , and IgE were determined by a mouse IgX ELISA quantification kit (Bethyl Laboratories, Montgomery, TX).
Dextran Sodium Sulfate (DSS)-induced Colitis-DSS of average molecular weight 36,000 -50,000 (MP Biomedicals, Solon, OH) was orally applied to 8-week-old male mice at a concentration of 1-3% (w/v) in drinking water. Changes in body weight were calculated every day. To assess the extent of colitis, body weight, stool consistency, and occult blood in the stool were monitored daily (31). Diarrhea was scored as follows: 0, normal; 2, loose stools; 4, watery diarrhea. Hemoccult was scored as follows: 0, normal; 2, hemoccult positive; 4, gross bleeding. On the last day of the experiments, blood was collected for determination of hematocrit using a Vetscan HMII; the colon was taken for histological examination, and the spleen was weighed and subjected to flow cytometry. As required for experiments, 2.5 M eicosapentaenoic acid (EPA; C20:5) and 5 M DHA (both from Cayman Chemicals, Ann Arbor, MI) in 200 l of saline were intrarectally injected into mice every day during the period of DSS treatment.
Adoptive Transfer of BM Cells-Male Pla2g10 ϩ/ϩ or Pla2g10 Ϫ/Ϫ mice (8-week-old) were used as donors and recipients. Recipients were irradiated with 10.4 gray (M-150WE; Softex, Kanagawa, Japan) and then injected with 10 7 BM cells from donors. After 12 weeks, the recipient mice were subjected to DSS-induced colitis.
Separation of Intestinal Epithelial and Non-epithelial Cells-The large intestine was removed, opened longitudinally, washed with PBS, and incubated with PBS contaning 5 mM EDTA with shaking for 30 min at 37°C. The tissue was separated into intestinal epithelial cells (IECs; leaflets of the epithelium) and non-IECs under a stereomicroscope, and the cells were washed with PBS before use.
Preparation of Colorectal Lamina Propria Lymphocytes (LPLs)-LPLs were prepared from C57BL/6 mice treated with 3% DSS for 7 days. Briefly, the colon (1 cm in length) was incu-bated with 5 mM EDTA in PBS for 20 min at 37°C, washed twice with PBS, minced, and incubated with 20 mg/ml collagenase type 4 (Worthington) plus 0.1 mg/ml DNase (Sigma) in RPMI 1640 medium (Sigma) for 50 min at 37°C. After filtration through a nylon mesh, floating cells were suspended in 40% Percoll (Sigma), applied onto 75% Percoll, and centrifuged at 1,000 ϫ g for 20 min at room temperature. The boundary cells (LPLs) were collected, adjusted at 2 ϫ 10 6 cells/ml in RPMI 1640 medium containing 10% FBS in a U-shaped 96-well plate (100 l/well), and then cultured for 48 h to assess the release of cytokines using enzyme immunoassay kits for IL-17A (eBioscience) and IL-22 (Biolegend). As required for experiments, the cells were cultured with 1-10 M lipids (Cayman Chemicals) or 10 M GSK137647, a GPR120 agonist (Tocris Bioscience, Bristol, UK).
Sperm Fertility-Analyses of spermatozoa were carried out as described previously (3). Briefly, female mice (10 weeks old) were injected intraperitoneally with 7.5 IU pregnant mare serum gonadotropin (Asuka Pharmacy, Tokyo, Japan) followed 48 h later with 7.5 IU human chorionic gonadotropin (Asuka Pharmacy). After 13 h, the oocyte cumulus complexes from the oviduct were placed in 100 l of HTF medium (ARK Resource, Kumamoto, Japan) in a 60-mm culture dish, and droplets were covered by embryo-tested mineral oil (Nakalai Tesque). Spermatozoa collected from the cauda epididymidis from male mice (8 weeks old) were allowed to swim into 50 l of HTF medium, aspirated, incubated in 200 l of HTF medium for 60 min at 37°C to permit capacitation, diluted, and added to the oocyte droplets to achieve a concentration of 200 spermatozoa/l. After incubation for 6 h at 37°C, the oocytes were washed and cultured for 24 h. Fertilization was evaluated by the presence of a second polar body and two pronuclei. As required for experiments, lipids (1 M) were added to the in vitro fertilization assay.
Electrospray Ionization-Mass Spectrometry (ESI-MS)-All procedures were performed as described previously (4,5). In brief, tissues were soaked in 10 volumes of methanol and homogenized with a Polytron homogenizer. After overnight incubation at Ϫ20°C, H 2 O was added to the mixture to give a final methanol concentration of 10% (v/v). As internal standards for determination of recovery, 1 ng of d 5 -labeled EPA, d 4 -labeled leukotriene B 4 , d 4 -labeled prostaglandin (PG) E 2 , and d 8 -labeled 15-hydroxyeicosatetraenoic acid (Cayman Chemicals) were added to the samples. The oxygenated lipids in the supernatant were extracted using Sep-Pak C18 cartridges (Waters, Milford, MA), where the samples in 10% methanol were applied to the cartridges, washed with 10 ml of hexane, eluted with 3 ml of methyl formate, dried up under N 2 gas, and dissolved in 60% methanol. The analysis of PUFAs and their metabolites was performed using a 4000Q-TRAP quadrupolelinear ion trap hybrid mass spectrometer (AB Sciex, Framingham, MA) with liquid chromatography (LC) (LC-20AP; Shimadzu, Kyoto, Japan) combined with an HTC PAL autosampler (CTC Analytics, Zwingen, Switzerland). The sample was applied to the Develosil C30-UG column (1 ϫ 150 mm inner diameter, 3 m particles) (Nomura Chemical, Aichi, Japan) coupled for ESI-MS/MS. The samples injected by the autosampler (10 l) were directly introduced and separated by a step gradient with mobile phase A (water containing 0.1% acetic acid) and mobile phase B (acetonitrile: methanol ϭ 4: 1; v/v) at a flow rate of 50 l/min and a column temperature of 45°C.
For detection of phospholipids, tissues were soaked in 10 volumes of 20 mM Tris-HCl (pH 7.4) and then homogenized with a Polytron homogenizer. Phospholipids were extracted and subjected to ESI-MS using a 4000Q-TRAP and LC-20AP with Develosil C30-UG column, as described previously (4). As an internal standard, 1 nmol of LPC(17:0) (Avanti Polar Lipids, Alabaster, AL) was added to each sample. The samples were separated by a step gradient with mobile phase A (acetonitrile/ methanol/water ϭ 1:1:1 (v/v/v) containing 5 M phosphoric acid and 1 mM ammonium formate) and mobile phase B (2-propanol containing 5 M phosphoric acid and 1 mM ammonium formate) at a flow rate of 80 l/min at 50°C.
Identification was conducted using multiple reaction monitoring transition and retention times, and quantification was performed based on peak area of the multiple reaction monitoring transition and the calibration curve obtained with an authentic standard for each lipid (Avanti Polar Lipids and Cayman Chemicals).
Statistical Analysis-Data are expressed as mean Ϯ S.E. or S.D. Statistical significance between groups was evaluated by two-tailed Student's t test or one-way analysis of variance at a significance level of p Ͻ 0.05.

Results
Immunosuppressive Phenotypes in PLA2G10 tg/ϩ Mice-Although our analysis of PLA2G10 tg/ϩ mice was underway (4), we noticed that PLA2G10 tg/ϩ mice had fewer circulating lymphocytes than did wild-type (WT) mice (Fig. 1A), contrary to our prediction that sPLA 2 -X overexpression would increase immune cells through its proposed pro-inflammatory action. Consistent with the lymphopenia, the weight of the spleen relative to that of the heart was significantly lower (Fig. 1, B and C); the number of splenocytes was ϳ50% lower (Fig. 1D), and the splenic white pulps appeared smaller (Fig. 1E) in PLA2G10 tg/ϩ mice than in WT mice. Although the proportions of splenic CD4 ϩ or CD8 ϩ T cells and CD11b ϩ monocytes/macrophages were unchanged, the proportion of CD45R ϩ B cells was lower in PLA2G10 tg/ϩ mice than in WT mice (Fig. 1, F and G). As the absolute number of splenocytes was reduced in PLA2G10 tg/ϩ mice, the total counts of splenic T cells and monocytes/macro- phages were proportionally lower in PLA2G10 tg/ϩ mice than in WT mice. Furthermore, the median fluorescence intensity of CD45R on B cells was greater in PLA2G10 tg/ϩ mice than in WT mice ( Fig. 1G), indicative of altered B cell differentiation. Despite the lower proportion of B cells, serum levels of IgG 1 and IgE, but not IgM and IgG 2a , were higher in PLA2G10 tg/ϩ mice  Total RNAs were isolated from the thymus of PLA2G10 tg/ϩ and littermate WT mice at 6 months. Equal amounts of total RNA (pooled from four mice for each genotype) were subjected to two-color gene expression microarray analysis. Data were processed using the Feature Extraction software from Agilent. Representative genes that showed decreased expression in transgenic (Tg) mice relative to WT mice are listed. than in WT mice (Fig. 1H), suggesting preferential skewing toward a Th2 response, which on the one hand promotes allergies and on the other hand suppresses Th1/Th17-based diseases such as arthritis, atherosclerosis, obesity, and colitis (32).
Resident peritoneal macrophages in PLA2G10 tg/ϩ mice showed greater expression of the M2 macrophage markers Arg1 and Cd206 than did WT mice, although expression of the M1 macrophage marker Cd68 was comparable in both genotypes ( Fig. 2A). The count of thioglycolate-induced macrophages in the peritoneal cavity was lower in PLA2G10 tg/ϩ mice than in WT mice (Fig. 2B), suggesting a reduced ability of monocytes to migrate to sites of inflammation or to differentiate into pro-inflammatory M1-like macrophages. M-CSFdriven BM-derived macrophages from PLA2G10 tg/ϩ mice showed greater expression of the M2 marker Arg1 than did WT mice, even when they were cultured with M1 polarizers (LPS ϩ IFN-␥) (Fig. 2C).
In agreement with the view that M2 macrophages and Th2 immunity counteract metabolic diseases (8,32), PLA2G10 tg/ϩ mice had lower body weight (Fig. 2D) and adiposity (Fig. 2E) than did WT mice throughout their life span. The subcutaneous fat layer, which was obviously present in WT mice, was scarcely seen in PLA2G10 tg/ϩ mice (Fig. 2F). Thus, Tg overexpression of sPLA 2 -X facilitates M2 polarization of macrophages, which may account, at least partly, for the anti-inflammatory and lean phenotypes.
We next assessed whether the anti-inflammatory phenotypes observed in PLA2G10 tg/ϩ mice might be ascribed to the capacity of sPLA 2 -X to alter lipid profiles in vivo. ESI-MS revealed that the splenic levels of AA, EPA, and DHA were significantly greater in PLA2G10 tg/ϩ mice than in WT mice (Fig. 3A). The levels of AA metabolites tended to be slightly higher in PLA2G10 tg/ϩ mice than in WT mice, but none of them reached statistical significance. Notably, the levels of 3 PUFA metabolites, such as hydroxyeicosapentaenoic acids and hydroxydocosahexaenoic acids (including protectin D1 (PD1)), were significantly increased in PLA2G10 tg/ϩ mice relative to WT mice (Fig. 3A). The increase of 3 PUFAs and their metabolites in PLA2G10 tg/ϩ mice was not limited to the spleen, because the skin levels of DHA and its metabolite PD1 were also higher in PLA2G10 tg/ϩ mice than in WT mice (Fig. 3B), although AA and its metabolite PGE 2 were also increased in the transgenic skin (4). In the colon, significant increases of EPA, rather than AA, metabolites were evident (Fig. 3C). Taken together, these results suggest a previously unappreciated capacity of sPLA 2 -X to mobilize 3 PUFAs and their metabolites in vivo. Given the well established anti-inflammatory role of 3 PUFAs and their metabolites (38,39), the lipid profiles altered thus far could explain, at least in part, the immunosuppressive phenotypes in PLA2G10 tg/ϩ mice.
Exacerbation of Colitis in Pla2g10 Ϫ/Ϫ Mice-Given these observations, we next searched for a particular pathophysiological condition under which endogenous sPLA 2 -X would play an anti-inflammatory role. To this end, we focused on inflammation in the gastrointestinal (GI) tract, where endogenous sPLA 2 -X is abundantly expressed (40,41). Inflammatory bowel disease (IBD) is a chronic, relapsing, and remitting condition of unknown origin that exhibits various features of immunological disorders, including impaired mucosal barrier function, pronounced innate and acquired immunity, and dysregulated production of cytokines, chemokines, and lipid mediators (42)(43)(44). Both 6 AA metabolites, such as PGE 2 and 12(S)-hydroxyheptadecatrienoic acid (12-HHT) (31,45,46), and 3 PUFAs or their metabolites, such as resolvins D and E (47)(48)(49), are protective against IBD. However, the PLA 2 subtypes that lie upstream of the production of these lipid mediators in this disease are currently unknown.
Among the sPLA 2 s, Pla2g10 (X) was expressed most abundantly in C57BL/6 colon, followed in order by Pla2g5 (V), Pla2g2f (IIF), Pla2g3 (III), and Pla2g12a (XIIA), whereas Pla2g1b (IB), Pla2g2d (IID), and Pla2g2e (IIE) were detected only at trace levels (Fig. 4A). Pla2g4a, and to a lesser extent Pla2g6 (which encode group IVA cytosolic PLA 2 (cPLA 2 ␣) and group VIA Ca 2ϩ -independent PLA 2 (iPLA 2 ␤), respectively), were also expressed at substantial levels in the colon. Immunohistochemistry of the colon showed that sPLA 2 -X protein was localized in IECs and goblet cells, although its staining was absent in Pla2g10 Ϫ/Ϫ mice (Fig. 4B). Consistently, Pla2g10 mRNA was enriched in Epcam-positive IECs isolated from WT colon (Fig. 4C). In DSS-induced ulcerative colitis, a well known model of IBD (50), the colorectal expression of Pla2g10 as well as Pla2g2f, Pla2g3, Pla2g12a, and Pla2g6 was decreased in mice treated for 7 days with 3% DSS (Fig. 4A), probably due to the collapse of the mucosal epithelium or in unknown ways. The expression of Pla2g4a and Pla2g5 was constant regardless of DSS challenge, suggesting that they are distributed mainly in cells other than IECs.
To assess the roles of sPLA 2 s in IBD, we applied the DSSinduced colitis model to mice lacking individual sPLA 2 s expressed in the colon. Notably, Pla2g10 Ϫ/Ϫ mice exhibited more severe colitis than did WT mice. After a lag period of several days after exposure to 1% DSS, Pla2g10 Ϫ/Ϫ mice displayed more severe body weight loss (Fig. 5A), fecal bleeding plus diarrhea (summarized as the clinical score) (Fig. 5B), and colon shortening (Fig. 5C) than did WT mice. Histologically, more advanced epithelial loss, crypt damage, ulceration, and submucosal infiltration of immune cells were evident in the colon of DSS-treated Pla2g10 Ϫ/Ϫ mice than was the case for WT mice (Fig. 5D). In comparison, mice lacking other sPLA 2 s, including Pla2g2d Ϫ/Ϫ , Pla2g2e Ϫ/Ϫ , Pla2g2f Ϫ/Ϫ , Pla2g3 Ϫ/Ϫ , and Pla2g5 Ϫ/Ϫ mice, showed no obvious phenotypes in this model (Fig. 5E).
Quantitative RT-PCR of the colon revealed that the expression levels of genes related to pro-inflammatory and Th17-related cytokines (Il1b, Il6, Il17a, Il22, and Tnf) were increased more robustly in Pla2g10 Ϫ/Ϫ mice than in Pla2g10 ϩ/ϩ mice after DSS challenge (Fig. 5F). Expression of Reg3g, which encodes an IL-22-inducible anti-bacterial protein (42,43), as well as that of CD4 ϩ and CD8 ϩ T cell markers (Cd4 and Cd8a; the latter in particular) also tended to be higher in DSS-treated Pla2g10 Ϫ/Ϫ than in Pla2g10 ϩ/ϩ mice (Fig. 5F). Expression of both M1 and M2 macrophage markers (Nos2 and Arg1, respectively) was also greater in DSS-treated Pla2g10 Ϫ/Ϫ mice than in Pla2g10 ϩ/ϩ mice, suggesting that the absence of sPLA 2 -X affected recruitment, rather than polarization, of macrophages in this setting. These results were further supported by microarray gene profiling, where colorectal expression of various cytokines, chemokines, macrophage markers, and other inflammatory genes was elevated in DSS-treated Pla2g10 Ϫ/Ϫ mice relative to Pla2g10 ϩ/ϩ mice (Table 3). Even in the control group, expression of the pro-inflammatory and anti-bacterial genes S100a8 and S100a9 was higher in Pla2g10 Ϫ/Ϫ mice than in Pla2g10 ϩ/ϩ mice, suggesting that some colorectal abnormal- ities were already present in the null mice under normal housing conditions. In contrast, expression of several epithelial markers was lower in DSS-treated Pla2g10 Ϫ/Ϫ than in Pla2g10 ϩ/ϩ mice (Table 3), consistent with the increased epithelial collapse in the former.
DSS-treated Pla2g10 Ϫ/Ϫ mice showed more profound splenomegaly (Fig. 6, A and B) and a decrease in hematocrit (Fig. 6C) relative to Pla2g10 ϩ/ϩ mice, suggesting alternation of extramedullary erythropoiesis due to colorectal bleeding. Indeed, flow cytometry of cells in the erythrocyte lineage (in terms of CD71 and TER119 expression) revealed increased accumulation of immature erythroblasts and reticulocytes, with reciprocal decreases in mature erythrocytes, in the blood (Fig. 6, D and E) and even more profoundly in the spleen (Fig. 6, F and G) of DSS-treated Pla2g10 Ϫ/Ϫ mice relative to replicate Pla2g10 ϩ/ϩ mice. This trend was already evident, albeit modestly, even in the control group (Fig. 6, D and F). These results suggest that the transition from immature to mature erythro-cytes is disturbed by Pla2g10 deficiency, particularly under the conditions of colitis.
To evaluate the relative contribution of sPLA 2 -X in the hematopoietic and non-hematopoietic compartments to DSSinduced colitis, BM cells from Pla2g10 ϩ/ϩ or Pla2g10 Ϫ/Ϫ mice were adoptively transferred into lethally irradiated Pla2g10 ϩ/ϩ or Pla2g10 Ϫ/Ϫ mice, which were then subjected to the colitis model (Fig. 7A). When the donor BM cells from Pla2g10 ϩ/ϩ or Pla2g10 Ϫ/Ϫ mice were transferred into recipient Pla2g10 ϩ/ϩ mice (WT 3 WT or KO 3 WT), there were no differences in body weight (Fig. 7B) or clinical score (Fig. 7C) between the groups. In contrast, weight loss (Fig. 7B) and an increased clinical score (Fig. 7C) were obvious in Pla2g10 Ϫ/Ϫ mice that received Pla2g10 ϩ/ϩ BM cells (WT 3 KO) in comparison with WT 3 WT or KO 3 WT chimeras. When Pla2g10 Ϫ/Ϫ mice were used as both donors and recipients (KO 3 KO), the weight loss (Fig. 7B) and increased clinical score (Fig. 7C) were similar to those in WT 3 KO chimeras. These results suggest that sPLA 2 -X in non-hematopoietic cells, most likely IECs, is mainly responsible for the protection from colitis. Of note, DSS-induced splenomegaly (Fig. 7D) and the decrease in hematocrit (Fig. 7E) were significantly more severe in KO 3 KO chimeras than in WT 3 KO chimeras, implying an additional contribution of hematopoietic sPLA 2 -X to these processes in the absence of non-hematopoietic sPLA 2 -X. sPLA 2 -X Mobilizes 3 PUFAs in Colitis-To gain insights into the mechanism underlying the anti-inflammatory action of sPLA 2 -X in colitis, lipids extracted from colon tissues of Pla2g10 ϩ/ϩ and Pla2g10 Ϫ/Ϫ mice with or without DSS treatment were subjected to ESI-MS. We found that the colon levels of EPA, docosapentaenoic acid (DPA; C22:5), and DHA were increased in WT mice following DSS treatment, whereas these changes were not evident in Pla2g10 ϩ/ϩ mice (Fig. 8A). AA also showed a similar trend but did not reach statistical significance (Fig. 8A). Pla2g10 deficiency did not alter the basal levels of these PUFAs. Strikingly, the colorectal levels of AA metabolites were not affected by Pla2g10 deficiency (Fig. 8B), whereas those of EPA or DHA metabolites, such as resolvins and 18-HEPE, were substantially lower in DSS-treated Pla2g10 Ϫ/Ϫ mice than in Pla2g10 ϩ/ϩ mice (Fig. 8C). These results, together with the results of PLA2G10 tg/ϩ mice (see above) and the reported role of 3 PUFA metabolites in the protection against colitis (47)(48)(49), raise the possibility that the mobilization of 3 PUFAs or their metabolites may underlie the anti-inflammatory role of sPLA 2 -X in colitis.

TABLE 3 Microarray gene profiling of the colon of Pla2g10 ؊/؊ mice versus Pla2g10 ؉/؉ mice in DSS-induced colitis
Total RNAs were isolated from the colons of Pla2g10 ϩ/ϩ (WT) and Pla2g10 Ϫ/Ϫ (KO) mice with or without 1% DSS treatment for 1 week. Equal amounts of total RNA (pooled from three mice for each genotype) were subjected to one-color gene expression microarray analysis. Data were processed using the Feature Extraction software from Agilent and analyzed using GeneSpring software. Fold changes (KO relative to WT) on the microarray are listed. 8D). Because these PUFAs can act on the fatty acid receptor GPR120 or GPR40 (51), we tested the effect of GSK137647, a GPR120-selective agonist, on cytokine production by LPLs. The release of IL-17A and IL-22 was suppressed by GSK137647 as efficiently as DHA (Fig. 8E), indicating that PUFAs may act, at least in part, on GPR120 on LPLs, thereby partially dampening the Th17 cytokine production. Moreover, daily intrarectal injection of 3 PUFAs (a mixture of EPA and DHA) into Pla2g10 Ϫ/Ϫ mice prevented DSS-induced body weight loss (Fig. 8F). Overall, these results further support the notion that sPLA 2 -X prevents colitis by releasing 3 PUFAs. Nonetheless, because the colorectal level of AA tended to be lower in DSStreated Pla2g10 Ϫ/Ϫ mice than in WT mice (Fig. 8A), the 3 PUFA metabolites were present at 30 times lower than the AA metabolites (Fig. 8, B and C), and AA suppressed IL-17A release by LPLs (Fig. 8D), it is possible that AA itself released by sPLA 2 -X might also contribute to the protection from colitis and that the background level of PGs might mask a pool of PGs potentially formed following mobilization of AA by sPLA 2 -X in a subset of cells.
Protective Role of the cPLA 2 ␣-PGE 2 Axis against Colitis-The above observations prompted us to ask which PLA 2 subtype(s) is linked to AA metabolism in colitis. We therefore applied the DSS-induced colitis model to mice null for Pla2g4a and Pla2g6, which are expressed in the colon (Fig. 4A). Severe weight gain, fecal bleeding, and diarrhea were seen in Pla2g4a Ϫ/Ϫ mice, but not WT mice, soon after oral application of DSS (Fig. 9, A and  B). On day 7, colorectal damage with epithelial loss and massive immune cell infiltration (Fig. 9C), splenomegaly (Fig. 9D), and decrease of hematocrit (Fig. 9E) were far more prominent in Pla2g4a Ϫ/Ϫ mice than in WT mice. In contrast, exacerbation of these parameters was not evident in Pla2g6 Ϫ/Ϫ mice (Fig. 9, A  and B). The overall phenotypes in Pla2g4a Ϫ/Ϫ mice were similar to those in Ptger4 Ϫ/Ϫ mice, which lack the PGE 2 receptor EP4 (31), or Ptges Ϫ/Ϫ mice, which lack microsomal PGE 2 synthase-1 (mPGES-1) (Fig. 9, A-E) (52), even though the overall symptoms, as revealed by the delay in body weight loss, were milder in Ptges Ϫ/Ϫ mice than in Pla2g4a Ϫ/Ϫ mice.
Lipidomics studies of the colon revealed that PGE 2 was present at a markedly lower level in DSS-treated Pla2g4a Ϫ/Ϫ mice than in WT mice (Fig. 9F). 12-HHT was also ϳ50% lower, although the changes in other prostanoids were relatively small, in DSS-treated Pla2g4a Ϫ/Ϫ mice compared with WT mice. These results indicate that cPLA 2 ␣ is preferentially coupled with PGE 2 and to a lesser extent 12-HHT in DSS-induced colitis. In contrast, the levels of EPA and DHA metabolites were unaffected by Pla2g4a deficiency. The level of PGE 2 was markedly decreased in Ptges Ϫ/Ϫ mice (to a level similar to that in Pla2g4a Ϫ/Ϫ mice, but not to zero), with reciprocal increases of other prostanoids, relative to WT mice (Fig. 9F). These results suggest the following. (i) The severe exacerbation of DSS-induced colitis in Pla2g4a Ϫ/Ϫ mice is due to the marked reduction of colon-protective prostanoids such as PGE 2 and 12-HHT (31,45,46). (ii) The milder outcome in Ptges Ϫ/Ϫ mice than in Pla2g4a Ϫ/Ϫ mice is probably because the former harbors the reduction of PGE 2 only, which may be counterbalanced by the increases in 12-HHT and/or other prostanoids through a shunting effect (53). (iii) The cPLA 2 ␣-mPGES-1 axis accounts mostly, if not entirely, for a pool of PGE 2 responsible for this disease model. (iv) There is an alternative route for the basal, cPLA 2 ␣-independent production of prostanoids such as PGF 2␣ and 6-keto-PGF 1␣ (a stable end product of PGI 2 ) in the colon. Overall, cPLA 2 ␣ and sPLA 2 -X exert a protective effect against colitis by mobilizing distinct sets of lipid metabolites, i.e. 6 AA and 3 PUFA metabolites, respectively.
Reduced Release of DHA and DPA in Pla2g10 Ϫ/Ϫ Spermatozoa-In addition to the GI tract, sPLA 2 -X is abundantly expressed in the testis, being stored in and released from the acrosomes of spermatozoa during capacitation and the acrosome reaction, and Pla2g10 Ϫ/Ϫ spermatozoa display reduced fertility, with no alteration in motility (40,54). However, the phospholipid metabolism underlying the action of sPLA 2 -X in this context remains to be determined.
To address this issue, we performed lipidomics analysis of Pla2g10 ϩ/ϩ and Pla2g10 Ϫ/Ϫ spermatozoa before and after capacitation. Consistent with the view that sPLA 2 -X is dispensable for sperm maturation (40), the PC compositions of spermatozoa before capacitation were identical between the genotypes (Fig. 10A). Notably, after capacitation, the levels of PC species with DHA or DPA, but not those with AA and other fatty acids, were significantly lower in Pla2g10 ϩ/ϩ cells than in Pla2g10 Ϫ/Ϫ cells (Fig. 10A). In accordance with this, the release of DHA and DPA but not AA and linoleic acid (Fig. 10B), as well as LPC with C18:0 (and to a lesser extent with C18:1 and C16:0) (Fig. 10C), was greater in Pla2g10 ϩ/ϩ than in Pla2g10 Ϫ/Ϫ spermatozoa. Release of EPA was very low, because EPA-bearing PC was a minor phospholipid component in mouse sperm (Fig. 10,  A and B) (3). These results suggest that sPLA 2 -X secreted from activated spermatozoa preferentially cleaves DHA-or DPAcontaining PC in the sperm membrane to release DHA, DPA, and LPC. We then evaluated the effects of these lipid products on fertilization. The fertilization ability of Pla2g10 Ϫ/Ϫ sperm with WT oocytes was lower than that of Pla2g10 ϩ/ϩ sperm, as reported previously (40,54), whereas addition of DPA and to a lesser extent LPC restored the fertilization ability of Pla2g10 Ϫ/Ϫ sperm (Fig. 10D). Thus, the lipid products released from the sperm membrane by sPLA 2 -X, particularly DPA, facilitate optimal fertilization.

Discussion
The roles of sPLA 2 s, including sPLA 2 -X, in promoting or attenuating inflammation or other pathophysiological events may be dictated by the cells from which they are secreted, the target membranes on which they act, or when and how their phospholipid-hydrolytic products are associated with the particular biological processes in cell-, tissue-, or disease-specific contexts. Given the current proposal that sPLA 2 -X is a proinflammatory enzyme (2, 16 -22), our present observation that PLA2G10 tg/ϩ mice exhibited a global immunosuppressive phe-notype was initially unexpected but appeared to be compatible with studies reporting that sPLA 2 -X overexpression in cultured macrophages elicited anti-inflammatory responses (23) and that atherosclerosis worsened in Ldlr Ϫ/Ϫ mice that had been received Pla2g10 Ϫ/Ϫ BM cells by adoptive transfer (24). In this study, by employing Pla2g10 gene-manipulated mice in combination with lipidomics, we have revealed the anti-inflammatory, rather than pro-inflammatory, features of sPLA 2 -X in vivo.
Endogenous sPLA 2 -X is constitutively expressed at a high level in the GI tract and testis (40,41,54), and this study using Pla2g10 Ϫ/Ϫ mice has shown that sPLA 2 -X mobilizes 3 PUFAs in addition to, or even in favor of, 6 AA in these tissues in the processes of colitis and fertilization, respectively. Even in PLA2G10 tg/ϩ mice, which globally overexpress human sPLA 2 -X at a super-physiological level, there are modest trends toward selective increases of 3 PUFA metabolites over 6 AA metabolites in multiple if not all tissues. These observations suggest that sPLA 2 -X has the intrinsic ability to mobilize 3 PUFA-derived metabolites in vivo. We have recently shown that sPLA 2 -IID, which is highly expressed in dendritic cells in lymphoid tissues, resolves contact hypersensitivity by mobilizing DHA-derived pro-resolving lipid mediators (5). Our results thus reveal a novel role of sPLA 2 -X as another 3 PUFA-mobilizing sPLA 2 , thereby regulating tissue-specific homeostasis.
3 PUFAs such as EPA and DHA resolve various types of inflammation, obesity, and atherosclerosis by acting on fatty acid-sensing receptors (e.g. PPARs and GPR120; see below) (51,55), by being metabolized to pro-resolving lipid mediators (e.g. resolvins and protectins) (38,39), by attenuating endoplasmic reticulum stress (56), or by increasing membrane fluidity, thus eventually altering membrane signaling or trafficking (57). It is likely that the anti-inflammatory actions of sPLA 2 -X occur through any of these mechanisms. Indeed, changes in the tissue levels of 3 PUFAs and their metabolites are correlated with the levels of sPLA 2 -X expression. 3 PUFA metabolites promote M2 macrophage polarization (58,59), prevent T cell activation or differentiation (60,61), and alter antibody production by B cells (62), a view that is relevant to the phenotypes observed in PLA2G10 tg/ϩ mice. Our results are also in accord with the aggravating role of sPLA 2 -X in asthma (2), where M2 macrophages are associated with the Th2-skewed airway inflammation (13). Therefore, we speculate that the reported roles of sPLA 2 -X in protection against atherosclerosis and obesity (24, 63) may also involve, at least in part, the mobilization of 3 PUFAs by this enzyme in a local tissue or even at a distal site (e.g. the GI tract), thus affecting the disease indirectly.
Our results do not rule out the contribution of sPLA 2 -X to AA metabolism, because this enzyme can release AA in various cultured cells (at 10 -100 ng/ml or more) (16 -18), and because several in vivo studies have shown that Pla2g10 ablation results FIGURE 9. DSS-induced colitis in Pla2g4a ؊/؊ , Pla2g6 ؊/؊ , and Ptges ؊/؊ mice. WT (n ϭ 9), Pla2g4a Ϫ/Ϫ (n ϭ 4), Pla2g6 Ϫ/Ϫ (n ϭ 5), and Ptges Ϫ/Ϫ (n ϭ 10) mice were administered 1% DSS orally. Body weight (A) and clinical score (B) were monitored at the indicated times, and colon histology (bar, 100 m) (C), spleen weight (D), hematocrit (E), and ES-MS profiles of colorectal lipids (F) were evaluated at day 7. There were no differences in the measured parameters among the genotypes under normal conditions (data not shown). Mean Ϯ S.E., *, p Ͻ 0.05, and **, p Ͻ 0.01. in reduction of eicosanoids (2, 19 -22). However, many of the previous in vivo studies did not measure 3 PUFA metabolites or discriminate whether sPLA 2 -X directly mobilizes eicosanoids or whether the observed changes in eicosanoids reflected changes in cPLA 2 ␣ expression or activation in the ongoing process of a given pathology. In fact, in the context of asthma, sPLA 2 -X secreted from the airway epithelium acts on infiltrating eosinophils in a paracrine manner to produce LPC, which in turn increases Ca 2ϩ influx leading to cPLA 2 ␣-dependent leukotriene generation (64), although it may directly mobilize AA metabolites from lung epithelial cells in an autocrine manner (20). In this study, we have shown that cPLA 2 ␣ and sPLA 2 -X are functionally segregated in the large intestine, driving non-overlapping lipid pathways (6 AA metabolism and 3 PUFA metabolism, respectively), which eventually culminates in a common outcome, i.e. protection against colitis. To the best of our knowledge, this is the first demonstration of the PLA 2 enzymes that are responsible for the release of distinct PUFAs in IBD. Moreover, the sPLA 2 -X-driven 3 PUFAs are capable of suppressing Th17 cytokine production by intestinal LPLs through GPR120, providing the first evidence for the functional linkage from a particular sPLA 2 to a fatty acid receptor. Although our study failed to show the ameliorating effect of resolvins and 18-HEPE on Th17 cytokine production by these cells, it is possible that these pro-resolving mediators could affect other steps of colitis, for instance by acting directly on epithelial cells to protect from injuries and on neutrophils to suppress their migration and to promote their clearance. In fact, resolvins block colitis when administered exogenously (47)(48)(49), and an endogenous EPA-derived epoxide attenuates allergic colitis (65).
Presumably, the mobilization of 6 AA versus 3 PUFA metabolites, or even other fatty acids and lysophospholipids, by sPLA 2 -X or other sPLA 2 s would rely not only on their intrinsic enzymatic properties but also on tissue-or disease-specific contexts such as the lipid composition of target membranes or the spatiotemporal availability of downstream enzymes, which may explain why the same enzyme often exerts pro-or antiinflammatory effects with different lipid mediator profiles in distinct settings. Indeed, sPLA 2 -IID mobilizes DHA-derived RvD1 in draining lymph nodes to suppress contact dermatitis (5) and AA-derived PGD 2 in the lung to counteract anti-viral immunity (66). sPLA 2 -V in adipose tissue releases oleic acid from lipoproteins in the process of obesity (8). Moreover, mobilization of a particular class of lysophospholipids, rather than fatty acids, is important for the function of sPLA 2 -IIF in the epidermis (67). Thus, caution should be exercised when interpreting the results of studies in which the actions of sPLA 2 are assigned only to AA metabolism. Apart from their roles in inflammation, multiple sPLA 2 s are expressed in male genital organs (68), among which two particular isoforms, sPLA 2 -III and -X, participate in sperm maturation and activation, respectively (3,54). Several lines of evidence suggest that DHA insufficiency causes asthenozoospermia with hypomotility and infertility (69,70). sPLA 2 -III is secreted from the epididymal epithelium and acts on immature spermatozoa passing through the epididymal duct to promote sperm membrane remodeling (3). As such, mature spermatozoa gain a higher proportion of DPA/DHA-containing PC species, which are crucial for sperm motility and thereby fertility. After ejaculation into the female duct, mature sperm undergo capacitation to allow hypermotility and acrosome reaction for fertilization, where the acrosome-derived sPLA 2 -X plays a promoting role (54). DPA, an intermediate in the biosynthetic conversion from EPA to DHA, has recently attracted attention as a precursor of novel 13-series resolvins with potent pro-resolving activity (71). Beyond this function, DPA is highly enriched in sperm cells, yet the biological importance of this PUFA in reproduction has been poorly understood (72). We now provide evidence that sPLA 2 -X selectively hydrolyzes DPA/DHA-containing PC species in sperm membranes to release DPA, DHA, and LPC, among which DPA has the highest ability to restore the fertilization ability of Pla2g10 Ϫ/Ϫ sperm. Although the mechanism underlying this action of DPA still awaits future studies, our results nonetheless provide new insight into the biological role of this unique PUFA in reproduction and also a rationale for its high degree of enrichment in sperm cells. Thus, sPLA 2 -III promotes epididymal sperm maturation, allowing enrichment of DPA/DHA-containing PC species in sperm membranes, and then sPLA 2 -X acts on these phospholipids to release DPA for successful fertilization, thereby underscoring an elegant cooperation of the two sPLA 2 s in the process of male reproduction.