Occurrence and biological activity of palmitoleic acid isomers in phagocytic cells

Recent studies have highlighted the role of palmitoleic acid [16:1n-7 (cis-9-hexadecenoic acid)] as a lipid hormone that coordinates cross-talk between liver and adipose tissue and exerts anti-inflammatory protective effects on hepatic steatosis and insulin signaling in murine models of metabolic disease. More recently, a 16:1n-7 isomer, cis-7-hexadecenoic acid (16:1n-9), that also possesses marked anti-inflammatory effects, has been described in human circulating monocytes and monocyte-derived macrophages. By using gas chromatographic/mass spectrometric analyses of dimethyl disulfide derivatives of fatty acyl methyl esters, we describe in this study the presence of a third 16:1 isomer, sapienic acid [16:1n-10 (6-cis-hexadecenoic acid)], in phagocytic cells. Cellular levels of 16:1n-10 appear to depend not only on the cellular content of linoleic acid, but also on the expression level of fatty acid desaturase 2, thus revealing a complex regulation both at the enzyme level, via fatty acid substrate competition, and directly at the gene level. However, unlike 16:1n-7 and 16:1n-9, 16:1n-10 levels are not regulated by the activation state of the cell. Moreover, while 16:1n-7 and 16:1n-9 manifest strong anti-inflammatory activity when added to the cells at low concentrations (10 μM), notably higher concentrations of 16:1n-10 are required to observe a comparable effect. Collectively, these results suggest the presence in phagocytic cells of an unexpected variety of 16:1 isomers, which can be distinguished on the basis of their biological activity and cellular regulation.

Thus, although a unifying mechanistic hypothesis that may help reconcile the variety of effects that 16:1n-7 appears to serve has yet to emerge, the recent finding that innate immune cells contain an isomer of 16:1n-7, namely cis-7-hexadecenoic acid (16:1n-9) that also displays potent anti-inflammatory activity (14), has raised the challenging possibility that this multiplicity of actions and the compartmentalized manner in which they occur could be due, at least in part, to the overlapping actions of such isomers being present at the same or neighboring locations. The 16:1n-9 is synthesized by human phagocytic cells via -oxidation of oleic acid and its levels are elevated in the neutral lipids of lipid droplet-laden monocytes, suggesting the possibility that detection of increased levels of this fatty acid in monocytes may constitute a biomarker for foamy cell formation (14). When added exogenously, 16:1n-9 exhibits potent anti-inflammatory actions in vitro and in vivo that are distinguishable from those of 16:1n-7 and comparable in magnitude to those exerted by n-3 fatty acids (14).
In the current study, we report that human phagocytic cells also express low levels of a third 16:1 isomer, sapienic acid [16:1n-10 (cis-6-hexadecenoic acid)]. Despite the common assumption that 16:1n-10 is restricted to humans (15), we found relatively high levels of this fatty acid in murine macrophages and in murine macrophage-like cell lines also. In these cells, 16:1n-10 levels appear to be controlled by complex mechanisms of regulation, including substrate competition and gene expression. Of note, 16:1n-10 shows anti-inflammatory activity when added to cells, but the activity is observed at higher concentrations (>25 M) than those required for 16:1n-7 and 16:1n-9 to exert similar effects.

Cell isolation and culture conditions
Human monocytes were isolated from buffy coats of healthy volunteer donors obtained from the Centro de Hemoterapia y Hemodonación de Castilla y León. Written informed consent was obtained from each donor. In brief, blood cells were diluted 1:1 with PBS, layered over a cushion of Ficoll-Paque, and centrifuged at 750 g for 30 min. The mononuclear cellular layer was recovered and washed three times with PBS, resuspended in RPMI 1640 medium supplemented with 40 g/ml gentamicin, and allowed to adhere in sterile dishes for 2 h at 37°C in a humidified atmosphere of CO 2 /air (1:19). Nonadherent cells were removed by washing extensively with PBS and the remaining attached monocytes were used the following day (14,16,17). Human macrophages were obtained by incubating plastic-adhered monocytes in RPMI with heat-inactivated 5% human serum and 40 g/ml gentamicin for 2 weeks in the absence of exogenous cytokine mixtures (16). All Mass spectra corresponding to the n-10 (B), n-9 (C), n-7 (D), and n-5 (E) isomers showing the molecular ion (M + ) and the three diagnostic peaks (fragment corresponding to the terminal methyl part of the molecule, fragment containing the carboxyl group, and the latter with the loss of methanol).
procedures involving human samples were undertaken in accordance with the Spanish National Research Council Committee on Bioethics, under the guidelines established by the Spanish Ministry of Health and the European Union.
Murine resident peritoneal macrophages from Swiss mice (University of Valladolid Animal House; 10-12 weeks old) were obtained by peritoneal lavage using 5 ml of cold PBS, as described elsewhere (18,19). The cells were plated at 2 × 10 6 per well (6-well plates) in 2 ml of RPMI 1640 medium with 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin, and allowed to adhere for 20 h in a humidified atmosphere of 5% CO 2 at 37°C. Wells were extensively washed with PBS to The mass spectra corresponding to the n-10 (D, H), n-9 (E, I), n-7 (F, J), and n-5 (G, K) isomers, indicating the molecular ion (M + ) and the three diagnostic peaks (fragment corresponding to the terminal methyl part of the molecule, fragment containing the carboxyl group, and the latter with the loss of methanol) are also shown. remove nonadherent cells. Adherent macrophages were then used for experimentation. All the protocols and procedures were approved by the Institutional Animal Care and Usage Committee of the University of Valladolid and are in accordance with the Spanish and European Union guidelines for the use of experimental animals. The macrophage-like cell lines, RAW264.7, P388D 1 , and THP-1, were maintained in culture and handled exactly as described elsewhere (20)(21)(22)(23)(24)(25)(26)(27).
For experiments, the cells were incubated in fresh serum-free medium for 1 h before addition of the various stimuli for different times. Unstimulated cells received vehicle. Opsonized zymosan was prepared exactly as described (28). No endogenous phospholipase A 2 activity was detected in the zymosan batches used in this study, as assessed by in vitro assay. Total cellular protein was measured utilizing a commercial kit (Bio-Rad Protein Assay). Protein levels did not significantly change over the course of the experiments.

GC/MS of fatty acid methyl esters
The samples, corresponding approximately to 10 7 cells, were washed with PBS, lysed with water, and sonicated in a tip homogenizer twice for 15 s. Afterward, 10 nmol of 1,2-diheptadecanoyl-sn-glycero-3-phosphocholine and 4 nmol of 1,2,3-triheptadecanoylglycerol were added and total lipids were extracted according to Bligh and Dyer (29). Lipid classes were separated by thin-layer chromatography. For neutral lipid separation, a solvent mixture consisting of n-hexane/diethyl ether/acetic acid (70:30:1, v/v/v) was used as the mobile phase (30). For analysis of phospholipid subclasses, the plates were predeveloped with chloroform/ methanol (1:1, v/v) and, after being air-dried, they were impregnated with 1.5% boric acid in ethanol/water (1:1, v/v). Phospholipids were separated twice with chloroform/methanol/28% ammonium hydroxide (60:37.5:4, v/v/v) as the mobile phase (31). The various lipid classes were scraped and extracted from the silica with 1 ml of chloroform/methanol (1:2, v/v) followed by 1 ml chloroform/methanol (2:1, v/v). The lipid extracts thus obtained were treated with 500 l of 0.5 M KOH in methanol for 60 min at 37°C to obtain fatty acid methyl esters. To neutralize, 500 l of 0.5 M HCl were added. Extraction of the fatty acid methyl esters was carried out with 2 ml n-hexane and 1 l was subjected to GC/MS analysis, as previously described (14,16,32), using an Agilent 7890A gas chromatograph coupled to an Agilent 5975C mass-selective detector operated in electron impact mode (EI, 70 eV), equipped with an Agilent 7693 autosampler and an Agilent DB23 column (60 m length × 0.25 mm internal diameter × 0.15 m film thickness). Data analysis was carried out with the Agilent G1701EA MSD Productivity Chemstation software, revision E.02.00.

Dimethyl disulfide derivatization of fatty acid methyl esters
Derivatization of fatty acid methyl esters to dimethyl disulfide (DMDS) adducts was carried out according to the procedure described by Sansone et al. (33), with slight modification. Briefly, the sample containing the fatty acid methyl esters in n-hexane (70-100 l) was placed in a screw amber cap vial with Teflonlined cap. DMDS (100 l) (Sigma-Aldrich) was added, together with 50 l of a 6% solution of iodine in diethyl ether. The vial was capped under nitrogen. The reaction was stirred at 40°C overnight. Afterward, 1 ml of n-hexane and 1 ml of 5% sodium thiosulfate solution were added. After mixing vigorously, the sample was centrifuged and the organic phase was collected, dried with anhydrous sodium sulfate, concentrated, and subjected to GC/MS analysis. DMDS derivatives were analyzed with the same protocol as fatty acid methyl esters, with slight modifications of the oven conditions. The final ramp temperature was set at 10°C/min up to 250°C and this temperature was held for 20 min to allow the DMDS derivatives of 16:1 isomers to elute. Identification of 16:1 isomers was carried out by elution time of the different DMDS adducts and by specific fragments of the mass spectra generated by cleavage of the C-C bond that originally contained the double bond (-fragment, -fragment, and -fragment with the loss of methanol).

LC/MS of 16:1-containing phospholipids
A cell extract corresponding to 5 × 10 6 cells was used for these analyses. The following internal standards were added: 600 pmol each of 1,2-dipentadecanoyl-sn-glycero-3-phosphocholine, 1,2-dilauroyl-sn-glycero-3-phosphoethanolamine, 1,2-dipalmitoylsn-glycero-3-phosphoinositol, and 1,2-dimyristoyl-sn-glycero-3phosphoserine, before lipid extraction according to the method of Bligh and Dyer (29). After evaporation of the organic solvent under vacuum, lipids were redissolved in 100 l methanol/water (9:1, v/v) and injected into a high-performance liquid chromatograph equipped with a binary pump, Hitachi LaChrom Elite L-2130 and a Hitachi Autosampler L-2200 (Merck). The column Human monocytes were treated with 10 M arachidonic acid (A), 100 ng/ml LPS (B), 1 mg/ml zymosan (C), or 2 M calcium ionophore A23187 (D) for the times indicated. Afterward, the cellular content of 16:1n-10 (gray), 16:1n-9 (dark red), and 16:1n-7 (green) was measured by GC/MS as described in the Materials and Methods. Results are shown as the mean ± SE of three independent experiments with incubations in duplicate. *P < 0.05, significantly different from the corresponding fatty acid in the unstimulated state (zero time). Note that in the experiments utilizing zymosan, no data on 16:1n-7 can be shown because zymosan particles contain relatively high amounts of this fatty acid. was a Supelcosil LC-18 (5 m particle size, 250 × 2.1 mm) (Sigma-Aldrich) protected with a Supelguard LC-18 (20 × 2.1 mm) guard cartridge (Sigma-Aldrich). Mobile phase was a gradient of solvent A (methanol/water/n-hexane/32% ammonium hydroxide, 87.5:10.5:1.5:0.5, v/v/v/v) and solvent B (methanol/n-hexane/32% ammonium hydroxide, 87.5:12:0.5, v/v/v). The gradient was started at 100% solvent A; it was decreased linearly to 65% solvent A and 35% solvent B in 20 min, to 10% solvent A and 90% solvent B in 5 min, and to 0% solvent A and 100% solvent B in an additional 5 min. The flow rate was 0.5 ml/min and 80 l of the lipid extract was injected. The LC system was coupled online to a Bruker esquire6000 ion-trap mass spectrometer (Bruker Daltonics, Bremen, Germany). The total flow rate into the column was split and 0.2 ml/min entered into the electrospray interface of the mass spectrometer. Nebulizer gas was set to 30 psi, dry gas to 7 l/ min, and dry temperature to 325°C. Ethanolamine glycerophospholipid (PE), phosphatidylinositol (PI), and phosphatidylserine (PS) were detected in negative ion mode with the capillary current ) and the three diagnostic peaks (fragment corresponding to the terminal methyl part of the molecule, fragment containing the carboxyl group, and the latter with the loss of methanol) are also shown. set at +3,500 V over the initial 25 min as [MH]  ions. Choline glycerophospholipid (PC) species were detected over the elution interval from 25 to 35 min in positive ion mode, as [M+H] + ions, with the capillary current set at 3,500 V. The 16:1-containing PE, PI, and PS species were identified by comparison with previously published data (34)(35)(36)(37)(38)(39). The cut off parameter was set at m/z 150 and fragmentation amplitude at 1 arbitrary unit. Because of the lability of vinyl ether linkages in acid media, plasmanyl (1-alkyl) and plasmenyl (1-alk-1′-enyl) glycerophospholipids were distinguished by acidifying the samples before lipid extraction. For the identification of acyl chains of 16:1-containing PC species, ionization was carried out in negative mode with post-column addition of acetic acid at a flow rate of 100 l/h as [M+CH 3 CO 2 ]  adducts, and acyl chains were identified by MS 3 experiments (34-39).

Real-time PCR
Total RNA was extracted using Trizol reagent (Ambion). The cDNA templates were synthesized using Verso cDNA synthesis kit (Thermo Scientific), following the manufacturer's instructions. Quantitative PCR (qPCR) was carried out with an ABI 7500 machine (Applied Biosystems, Carlsbad, CA) using Brilliant III Ultra-Fast SYBR Green qPCR Master Mix (Agilent Technologies, Santa Clara, CA), as previously described (40).

Immunoblot analysis
After the different treatments, the cells were lysed for 30 min in ice-cold buffer containing 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% Triton X-100, 100 mM Na 3 VO 4 , 1 mM phenyl methyl sulfonyl fluoride, and protease inhibitor cocktail (Sigma). Total protein (50 g) was resolved on 10-12% SDS-PAGE gels and transferred to a PVDF membrane. After transfer, nonspecific binding sites were blocked with 5% BSA in PBS containing 0.1% Tween-20 at room temperature for 1 h. The membranes were then probed with the corresponding antibodies followed by HRP-conjugated secondary antibodies in blocking solution. -Actin was used as a load control. The immunoblots were visualized using enhanced luminescence. Densitometry was performed on scanned images using Quantity One ® software (Bio-Rad Laboratories) and values were normalized for the corresponding controls of each experiment.

Statistical analysis
All experiments were carried out at least three times with incubations in duplicate or triplicate, and the data are expressed as mean ± SE. Statistical analysis was carried out by Student's t-test, with P < 0.05 taken as statistically significant.

RESULTS
We have previously shown that human monocytes and macrophages contain significant amounts of 16:1n-9, an unusual isomer of 16:1n-7 (14). We began the current study by investigating the possible occurrence of other unidentified 16:1 isomers in human phagocytic cells. For these experiments, we prepared fatty acid methyl esters of cellular fatty acids, which were then converted into DMDS derivatives. DMDS derivatization provides unambiguous analytical resolution of positional isomers because methyl sulfide substituents add to the two carbon atoms of the double bond, thus preventing isomerization (41). Figure 1 shows a GC/MS analysis of a mixture of four commercial 16:1 isomers, namely 16:1n-10, 16:1n-9, 16:1n-7, and palmitvaccenic acid (16:1n-5). All four isomers resolved well in GC and each of them produced specific fragments with electron-impact ionization MS, because fragmentation occurred at the level of the C-C bond, where the sulfide groups are attached. In addition, a third fragment was detected, which corresponded to the loss of 32 mass units (methanol) from the fragments containing the methyl ester (Fig. 1). Collectively, these data indicate that analysis of fatty acyl DMDS derivatives by GC (retention time) coupled to electron impact MS (detection of diagnostic peaks) allows unambiguous identification of the various isomers.
Application of this technique to analyze the occurrence of 16:1 isomers in human peripheral blood monocytes confirmed the predominant presence of 16:1n-7 and 16:1n-9, in agreement with our previous estimates (Fig. 2) (14). Interestingly, significant amounts of 16:1n-10 could also be detected, albeit at lower amounts than those of the other isomers ( Fig. 2A). Remarkably, the analysis of human monocyte-derived macrophages showed a different pattern of expression; while 16:1n-7 and 16:1n-9 were readily detected, the levels of 16:1n-10 were much lower in comparison. Traces of 16:1n-5 were also detected in macrophages ( Fig. 2B). In human serum, 16:1n-7 was, by far, the major isomer detected. Lower amounts of 16:1n-9 and even lower amounts of 16:1n-10 were also detected. No 16:1n-5 was found in serum (Fig. 2C). Collectively, these results suggest that production of 16:1n-10 by human cells could be celltype specific, being detectable at significant amounts only in monocytes. On the other hand, because 16:1n-5 was found only in macrophages and at trace levels, this isomer was not considered further.
The occurrence of 16:1 isomers was also studied in monocytes activated with different stimuli. In the first place, we used arachidonic acid, which can be secreted by damaged endothelium and possesses monocyte-activating properties (42). We have recently shown that human monocytes exposed to this fatty acid (10 M, 2 h) acquire a foamy phenotype, which is due to the accumulation of cytoplasmic lipid droplets (14,16). Lipid loading of the monocytes under these conditions is due to activation of the de novo pathway of lipid biosynthesis by arachidonic acid and ultimately results in the accumulation of 16:1n-9 and 16:1n-7 in various lipid fractions (14,16). In keeping with our previous observations, monocytes activated by arachidonic acid demonstrated a time-dependent increase in the levels of both 16:1n-9 and 16:1n-7 (Fig. 3A). In contrast, the cellular levels of 16:1n-10 did not significantly change.
Other monocyte stimuli, such as bacterial lipopolysaccharide (LPS) (Fig. 3B), yeast-derived zymosan (Fig. 3C), or the Ca 2+ ionophore, A23187 (Fig. 3D), all failed to increase the levels of any of the three 16:1 isomers. Thus, while there are activation conditions that result in upregulation of 16:1n-9 and 16:1n-7 levels (i.e., these leading to the formation of lipid-laden monocytes), 16:1n-10 levels appear not to be regulated upon any state of activation of the cells. Note as well that differentiation of monocytes to macrophages also did not upregulate 16:1n-10 levels (compare Fig. 2A and Fig. 2B).
Importantly, 16:1n-10 was found at quite significant levels in cells of murine origin, such as resident peritoneal macrophages (Fig. 4A). Moreover, when we analyzed RAW264.7 cells, a murine cell line that is widely used as a paradigm for studies of macrophage lipid signaling and metabolism (43), 16:1n-10 was detected at levels comparable to those of 16:1n-7 and much higher than those of 16:1n-9 ( Fig. 4B). High levels of 16:1n-10 were also observed in another murine-macrophage-like cell line, P388D 1 (Fig. 4C). When the study was extended to a human cell line, THP-1, 16:1n-10 was also found at high levels (Fig. 4D).
The distribution of 16:1 fatty acids among phospholipids of RAW264.7 cells was measured by LC/MS. The 16:1containing phospholipids were identified by determining the formation of a m/z 253 fragment in MS n experiments, corresponding to a 16:1 acyl chain. Structural identification of the glycerophospholipids was achieved by looking at the fragments and/or neutral losses obtained in MS 2 experiments for PE, PI, and PS, or MS 3 experiments for PC (34)(35)(36)(37)(38)(39). Fatty chains within phospholipids are designated by their number of carbons:double bonds. A designation of O-before the first fatty chain indicates that the sn-1 position is ether-linked, while a P-designation indicates a plasmalogen form (sn-1 vinyl ether linkage) (44).
A total of 15 glycerophospholipid species were found to contain 16:1 acyl chains, with PC species being, by far, the most enriched in these fatty acids. Remarkably, a single PC species, PC(16:0/16:1), comprised almost half of total cellular 16:1. By class, PC contained 70-80% of total 16:1, the remainder being present in PE (8-10%), PS (7-9%), and PI (2%) (Fig. 5A, B). GC/MS was utilized next to differentiate among 16:1 isomers within the different phospholipid classes. Approximately 50% of total 16:1 in PC was found to correspond to the n-7 isomer, while the n-10 comprised about 40%. The 16:1n-7 was also the most abundant isomer in PE (70%) and 16:1n-10 was the most abundant one in PI (60%) and PS (85%). The 16:1n-9 was detected at significant levels only in PC and PE (Fig.  5C-F). This distribution did not appreciably change when the cells were stimulated with bacterial LPS.
As a consequence of long-term culture, many cell lines are known to contain low levels of essential fatty acids compared with primary cells (45), and we confirmed this to be true as well for the cell lines utilized in this study (Fig. 6). While the double bond of both 16:1n-9 and 16:1n-7 is introduced by stearoyl-CoA desaturase ( 9 desaturase), 16:1n-10 is produced by a different enzyme, fatty acid desaturase 2 [FADS2 ( 6 desaturase)]. In keeping with these observations, incubation of the cells with the selective  6 desaturase inhibitor, SC-26196 (46,47), strongly inhibited 16:1n-10 levels, while leaving 16:1n-9 and 16:1n-7 levels unchanged (Fig. 7A).
Because FADS2 also converts linoleic into -linolenic acid within the n-6 fatty acid synthesis pathway, we reasoned that high levels of 16:1n-10 in cultured cells could reflect, at least in part, an increased access of  6 desaturase to palmitic acid due to the relative scarcity of linoleic acid. In keeping with this observation, we noted that the palmitic acid (16:0) to linoleic acid (18:2n-6) molar ratio in cultured cell lines was several fold higher than that of primary cells (Fig. 7B). Further, long-term incubation of the cells with exogenous linoleic acid led to a significant nanomolar reduction of 16:1n-10 levels, which corresponded with a similar nanomolar increase of n-6 products in cells (Fig. 7C). Importantly, analyses of FADS2 expression by qPCR clearly demonstrated higher expression levels of the gene in cultured cell lines with respect to primary cells (Fig. 7D). Thus, high expression levels of FADS2, together with higher 16:0/18:2n-6 ratios (i.e., low 18:2n-6 levels), would explain why cell lines produce more 16:1n-10 than primary cells. Thus, increased 16:1n-10 levels in cultured cells are not only a consequence of favored desaturation of palmitic acid over linoleic acid due to reduced levels of the latter, but also to increased expression of the  6 desaturase.
We have recently shown that both 16:1n-7 and 16:1n-9 possess potent anti-inflammatory activity both in vivo and in vitro (14). To extend these results to 16:1n-10, we prepared cells enriched in the various isomers by incubating them with a low concentration (10 M) of the corresponding fatty acid for 14 h in serum-free medium (14). This procedure results in the cells taking up the fatty acids and accumulating them in various cellular lipids, mainly phospholipids (Fig. 8A). In keeping with our previous data (14), subtle differences were found between 16:1n-7 and 16:1n-9 with regard to their incorporation into cellular lipid classes; the latter manifesting a greater tendency to accumulate into neutral lipids than the former. Significant amounts of 16:1n-10 were also found in neutral lipids (Fig. 8A). Further analyses of 16:1 incorporation into phospholipid classes revealed that, in keeping with the distribution of the endogenous fatty acids, most of the exogenous fatty acids incorporated into PC, with lesser amounts being found in PE. Exogenous 16:1n-7 was found to be a little more enriched in PE than in 16:1n-10 or 16:1n-9; the relative distribution of the latter two being very similar (Fig. 8B).
After the incubations with the fatty acids, the cells were stimulated with bacterial LPS and the effect on the expression of a number of pro-inflammatory genes was investigated. In agreement with our previous data (14), cells enriched with 10 M 16:1n-7 or 16:1n-9 showed significant inhibitory effects on the expression level of a number of inflammatory genes. Remarkably, these effects were not observed in the 16:1n-10-enriched cells at this concentration ( Fig. 9A-C); significantly higher concentrations of 16:1n-10 (>25 M) were necessary to observe a similar anti-inflammatory effect, as shown in dose-response studies (Fig. 9D). At doses at which 16:1n-10 was ineffective (<25 M), it did not alter the anti-inflammatory effect of 16:1n-9 (Fig. 9E). These results show that, although all 16:1 isomers present in macrophages may manifest antiinflammatory activity, there is a marked difference between 16:1n-7 and 16:1n-9 on the one hand, and 16:1n-10 on the other, with regard to the range of concentrations at which an effect is demonstrated. Interestingly, when yeastderived zymosan was used instead of LPS to stimulate the cells, none of the fatty acids exerted significant effects on pro-inflammatory gene expression (Fig. 9F). Thus the anti-inflammatory effects of 16:1 fatty acids may be specific for LPS stimulatory conditions, suggesting an interference with signals arising from TLR4 engagement. In this regard, a hallmark of LPS activation of macrophages is the rapid phosphorylation activation of the extracellularly regulated mitogen-activated protein kinases p42 and p44 (48). Figure 10 shows that 10 M 16:1n-9 blunted the LPS-induced phosphorylation of the two kinases in a timedependent manner, while 16:1n-10 was ineffective.

DISCUSSION
In this study, we show that human and murine phagocytic cells contain, in addition to 16:1n-7 and 16:1n-9, a third 16:1 isomer, namely 16:1n-10. Originally described in human sebum and later in hairs and nails, 16:1n-10 was commonly thought not to be present anywhere else in the human body (15,49,50). However, 16:1n-10 has recently been identified in erythrocyte membranes (41). Thus, our identification in circulating human monocytes was certainly no surprise, although the low levels at which the fatty acid presents in monocytes and macrophages make it difficult to envision a biological role for it in innate immunity and inflammation. It was more unexpected, however, to identify 16:1n-10 at relatively high levels in mouse peritoneal macrophages and, especially, cultured macrophage cell lines, because this fatty acid is conventionally considered to be unique to humans (hence its trivial name, sapienic acid, deriving from the root sapiens).
While the double bond present in 16:1n-7 and 16:1n-9 is introduced by a  9 desaturase enzyme (stearoyl-CoA desaturase), 16:1n-10 is produced by the action of a different enzyme,  6 desaturase. Our results indicate that, consistent with these different metabolic origins, cellular regulation of 16:1n-10 levels differs from that of 16:1n-7 and 16:1n-9 in several key respects. The levels of the latter two fatty acids may be regulated during lipid loading conditions, owing to stimulus-induced activation of de novo fatty acid biosynthesis (14,16,51,52). In contrast, the levels of 16:1n-10 appear not to depend on the activation state of the cells, but rather on complex, possibly interacting, mechanisms involving both substrate competition reactions and increased expression of FADS2.
It has been long thought that desaturation at carbon 6 is greatly favored when the fatty acid substrate already contains a double bond at position 9, as is the case of linoleic acid (53). Accordingly, in cells, desaturation of linoleic acid to -linolenic acid would be preferred over desaturation of palmitic acid to 16:1n-10, which would easily explain why cells containing high endogenous levels of linoleic acid produce 16:1n-10 at very low amounts and vice versa. While this idea is fully consistent with our results using human phagocytic cells and cultured cell lines, it is worth noting that mouse peritoneal macrophages, which contain large amounts of linoleic acid and, in general, of n-6 fatty acids, amounting to approximately 25% of total cellular fatty acids (54,55), also contain relatively high levels of 16:1n-10. Moreover, our studies using macrophage cell lines, which contain low levels of linoleic acid, show much greater expression of FADS2 relative to primary cells, suggesting that not only the palmitic acid to linoleic acid molar ratio but also the amount of enzyme present in the cell constitute another key factor regulating 16:1n-10 levels. Experiments are currently in progress to explore questions such as whether FADS2 operates constitutively at different efficiency depending on cell type or whether there are Fig. 9. Anti-inflammatory effect of 16:1 fatty acids. A-C: The cells were incubated with 10 M of 16:1n-10 (gray bar), 16:1n-9 (dark red bar), 16:1n-7 (green bar), or neither (black bar) for 14 h. Afterward, the cells were stimulated by 100 ng/ml LPS for 6 h and expression of Il6 (A), Ptgs2 (B), and Ccl2 (C) was measured by qPCR. D: Dose-response determinations. The cells incubated with the indicated fatty acid concentrations (16:1n-9, dark red bars; 16:1n-10, gray bars) were stimulated by 100 ng/ml LPS and gene expression (Nos2) was measured by qPCR. Responses of control incubations receiving the different fatty acid concentrations, but not LPS, were no different than control unstimulated cells (open bar) and are omitted for clarity. E: The 16:1n-10 does not alter the effect of 16:1n-9. The cells, incubated with 10 M 16:1n-10 (gray bar), 16:1n-9 (dark red bar), both (pink bar), or neither (black bar), were stimulated by 100 ng/ml LPS for 6 h and gene expression (Ccl2) was measured by qPCR. F: Effect of 16:1 fatty acids on zymosan-induced gene expression. The cells were incubated with 10 M of 16:1n-10 (gray bar), 16:1n-9 (dark red bar), 16:1n-7 (green bar), or neither (black bar) for 14 h. Afterward, the cells were stimulated by 200 g/ml zymosan for 6 h and gene expression (Ccl2) was measured by qPCR. Results are shown as the mean ± SE of three independent experiments with incubations in triplicate. *P < 0.05, significantly different from incubations with LPS without fatty acids.
cell-specific cofactors that may stimulate FADS2's preference for palmitic acid utilization. It is also possible that, depending upon cell type, linoleic acid is diverted to other reactions, i.e., serving as substrate of lipoxygenases for oxylipin formation, thus enabling the enzyme to use palmitic acid to a larger extent.
A striking feature of this work is that, unlike 16:1n-7 and 16:1n-9, 16:1n-10 is devoid of anti-inflammatory activity when used at low concentrations (up to 25 M), as judged by its inability to prevent the increased expression of proinflammatory cytokines by LPS-stimulated macrophages. Higher concentrations of 16:1n-10 are required to observe an effect. It is worth mentioning, in this context, that high concentrations of 16:1n-10 (100 M and above) have been demonstrated to possess antimicrobial activity in skin by a mechanism involving partitioning of the fatty acid into the cell membrane (56,57). While a change in cellular lipid composition due to the incorporation of the various 16:1 isomers may itself have an effect on receptor-mediated responses, our initial studies on the mechanism of regulation of LPS signaling by 16:1n-9 show a selectively reduced phosphorylation of the extracellular signal-regulated kinases p42 and p44 by LPS in the presence of the fatty acid, thus suggesting genuine interaction of the fatty acid with receptor-activated signaling. It is also striking that 16:1 fatty acids blunt LPS-mediated signals only and not those from other innate immune receptors, such as TLR2, dectin-1, or mannose receptors, which recognize yeast-derived zymosan (58), emphasizing selective receptor-mediated effects. This situation is reminiscent of recent work by Snodgrass et al. (59) on the inhibitory effect of docosahexaenoic acid, an n-3 fatty acid, on TLR-mediated responses in monocytes. In this study, the fatty acid was found to impair receptor dimerization, blunting subsequent signaling in this way (59). In our previous studies, we showed that docosahexaenoic acid is as effective as 16:1 fatty acids in inhibiting LPSmediated gene expression responses (14).
Our results also raise provocative questions as to the cellular utilization of the various lipid pools that may function as reservoirs for anti-inflammatory 16:1 fatty acids. In preliminary work, we noted that the anti-inflammatory effects of both 16:1n-7 and 16:1n-9 can be abrogated if the incorporation of the fatty acids into cellular lipids is prevented by the presence of the general acyl-CoA synthetase inhibitor, triacsin C (C. Meana and J. Balsinde, unpublished observations), suggesting that formation of an ester of the corresponding 16:1 fatty acid is necessary for biological activity. Hence, the initial lipid localization where the fatty acid is esterified after incubating it with the cells may determine its subsequent biological activity. It is remarkable, in this regard, that 70-80% of total 16:1 in untreated cells is found in PC and that almost half of it is present in just a single phospholipid species, namely PC(16:0/16:1). These observations become even more relevant in the light of the data showing that, regardless of isomer, the bulk of exogenous fatty acid added to the cells gets incorporated into PC. Given that recent work has demonstrated that certain PC species with defined fatty acyl ester composition have biological activity on their own (60,61), it will be of interest to determine whether pure phospholipid species containing either 16:1n-7, 16:1n-9, or 16:1n-10 as one of the acyl substituents are also biologically active. While it is clear that much work needs to be carried out to unravel the effects of 16:1 fatty acids on innate immune reactions, the elucidation of intracellular metabolic routes utilizing these fatty acids might provide interesting opportunities of intervention to ameliorate the inflammatory response. Afterward, the cells were stimulated by 100 ng/ml LPS for the indicated times and the phosphorylation of extracellular-regulated kinases p44 and p42 was studied by immunoblot. Blots shown are representative of three independent experiments and the figures below the blots show the relative quantifications of p44 and p42 phosphorylation with respect to -actin. Note that the blots corresponding to total p44 and p42 are from the same samples as phosphorylated p44 and p42, but from a different gel run in parallel.