Characterization of an arachidonic acid-deficient (Fads1 knockout) mouse model.

Arachidonic acid (20:4Δ5,8,11,14, AA)-derived eicosanoids regulate inflammation and promote cancer development. Previous studies have targeted prostaglandin enzymes in an attempt to modulate AA metabolism. However, due to safety concerns surrounding the use of pharmaceutical agents designed to target Ptgs2 (cyclooxygenase 2) and its downstream targets, it is important to identify new targets upstream of Ptgs2. Therefore, we determined the utility of antagonizing tissue AA levels as a novel approach to suppressing AA-derived eicosanoids. Systemic disruption of the Fads1 (Δ5 desaturase) gene reciprocally altered the levels of dihomo-γ-linolenic acid (20:3Δ8,11,14, DGLA) and AA in mouse tissues, resulting in a profound increase in 1-series-derived and a concurrent decrease in 2-series-derived prostaglandins. The lack of AA-derived eicosanoids, e.g., PGE2, was associated with perturbed intestinal crypt proliferation, immune cell homeostasis, and a heightened sensitivity to acute inflammatory challenge. In addition, null mice failed to thrive, dying off by 12 weeks of age. Dietary supplementation with AA extended the longevity of null mice to levels comparable to wild-type mice. We propose that this new mouse model will expand our understanding of how AA and its metabolites mediate inflammation and promote malignant transformation, with the eventual goal of identifying new drug targets upstream of Ptgs2.

commercial 10% saffl ower oil diet (D03092902R; Research Diets), free of AA. In a separate experiment, the basal 10% saffl ower oil diet was supplemented with ARASCO oil (containing 42% AA, w/w) to determine the effect of dietary AA on the life span of Null mice.

Real-time PCR
Fads1 and Fads2 mRNA expression levels in colon mucosa and liver were determined by real-time PCR on an ABI 7900 instrument. cDNA was synthesized from 2 g total RNA using random hexamers and oligo(dT) primers with Superscript II RT (Invitrogen). PCR was performed (primer sequences available online) using predeveloped Taqman assays (Applied Biosystems). Expression levels were normalized to ribosomal 18S expression using assay kits from Applied Biosystems (cat# Mm00507605 for Fads1 , Mm00517221 for Fads2 , and Mm03928990 for 18S).

Measurement of total phospholipid fatty acid composition
Total lipids in colon mucosa, liver, splenocytes, and serum were extracted by the method of Folch ( 17 ). Total phospholipids were separated by one-dimensional thin-layer chromatography on silica gel 60 G plates using chloroform/methanol/ acetic acid/water (90:8:1:0.8, v/v/v/v) as the developing solvent. Isolated total phospholipids were transesterifi ed in 6% methanolic HCl overnight, followed by GC analysis as previously described ( 18 ).

Eicosanoid analysis
Eicosanoids were extracted from colonic mucosa, small intestine mucosa, and lung using a previously described method ( 19,20 ). Briefl y, snap-frozen tissues were ground to a fi ne powder and homogenized with an Ultrasonic Processor (Misonix). An aliquot of homogenate was subjected to extraction with hexane/ ethyl acetate (1:1). The upper organic layer was collected, and the organic phases from three extractions were pooled and then evaporated to dryness under a stream of nitrogen. All extraction procedures were performed at minimum light levels at 4°C. Samples were then reconstituted in 100 l of methanol/10 mmol/l ammonium acetate buffer, pH 8.5 (50:50, v/v), before liquid chromatography/tandem mass spectroscopic analysis. Protein concentration was determined by the method of Bradford according to the manufacturer's instructions (Pierce). Liquid chromatography/tandem mass spectroscopic analyses were performed using a QuattroUltima mass spectrometer (Waters) equipped with an Agilent 1100 binary pump high-performance liquid chromatography system (Agilent) using a modifi ed version of the method of Yang et al. ( 19 ). Eicosanoids of interest were chromatographically separated using a Luna 3 m phenyl-hexyl 4.6 × 100 mm analytic column (Phenomenex). The mobile phase T-cell subsets involved in adaptively mediated infl ammation ( 8 ). Therefore, PGE 2 has the ability to infl uence the cytokine microenvironment, thereby skewing naïve T-cell differentiation, and ultimately function, toward infl ammatory T-cell subsets. For example, it has been reported that PGE 2 propagates infl ammatory bowel disease (IBD) by enhancing the development and function of IL-17-producing Th17 cells ( 9 ). In addition to its role in cytokine regulation, AA is a major constituent of phosphatidylinositol 4,5bisphosphate (PI(4,5)P 2 ), and via the action of phospholipase C, results in the accumulation of inositol trisphosphate (IP 3 ) and 1-stearoyl-2-AA-diacylglycerol to elicit intracellular Ca 2+ mobilization ( 10 ). The hydrolysis of PI(4,5)P 2 plays an important role in membrane rapid cytoskeletal remodeling ( 11 ) and Ras/Erk, as well as Ca 2+ signaling in lymphocytes ( 12 ).
Most studies have targeted prostaglandin biosynthetic and degradation enzymes in an attempt to suppress AAderived eicosanoid-mediated infl ammation and tumorpromoting action ( 13 ). Surprisingly, no investigators to date have attempted to target AA (substrate levels) as a way of modulating prostaglandin biosynthesis and tumor development. The controversies associated with the role of aspirin and Ptgs inhibitors indicate that more work is needed to elucidate the effects of eicosanoids in colon cancer and infl ammatory diseases ( 3 ). Therefore, we generated a novel genetic model, i.e., the Fads1 ( ⌬ 5 desaturase) knockout mouse, to determine the role of AA-derived 2-series eicosanoids in mucosal physiology and infl ammation. This model allows for the specifi c investigation of AA defi ciency without the underlying complications of essential fatty acid (LA and DGLA) defi ciency.

Generation of Fads1 null mice
Mutant Fads1 mice were generated using a gene-trapping technique ( 14 ). Mice (strain C57BL/6) were cloned from an ES cell line (IST11525H2; Texas Institute for Genomic Medicine, TIGM). The ES cell clone contained a retroviral insertion in the Fads1 gene identifi ed from the TIGM gene trap database, and was microinjected into C57BL/6 host blastocysts to generate germline chimeras using standard procedures. The retroviral Omni-Bank Vector 76 ( Fig. 1 ) contained a splice acceptor sequence (SA) followed by a 5 ′ selectable marker ␤ -geo, a functional fusion between the ␤ -galactosidase and neomycin resistance genes, for identifi cation of successful gene trap events followed by a polyadenylation signal (pA). Insertion of the retroviral vector into the Fads1 gene led to the splicing of the endogenous upstream exons into this cassette to produce a fusion transcript that was used to generate a sequence tag (OST) of the trapped gene by 3 ′ RACE ( 15 ). Chimeric males were bred to C57BL/6 females for germline transmission of the mutant Fads1 allele.

Animals and diets
Three genotypes [wild-type (Wt), heterozygous (Het) and null (Null)] of Fads1 mice were derived from heterozygous males and females. All procedures followed the guidelines approved by Public Health Service and the Institutional Animal Care and Use Committee at Texas A&M University. All animals were fed Biotec). Cells (5 × 10 5 ) were washed with cold 1× PBS, and then pelleted at 4,000 g for 5 min at 4°C. Supernatant was removed, and the pellet was resuspended in 800 l of 1:1 methanol:CHCl 3 . The mixture was vortexed for 1 min, and then centrifuged at 7,500 g for 5 min at 4°C. The supernatant was discarded, and pellets resuspended in 400 l of 80:40:0.3 methanol:CHCl 3 :HCl. The mixture was vortexed for 5 min and subsequently centrifuged at 3,000 g for 1 min at 4°C. An additional 80 l of 1 N HCl was added to the extract and vortexed for 15 s prior to centrifugation at 18,000 g for 15 s at 4°C. The organic layer was collected and dried under a stream of N 2 . The lipid fi lm was dissolved in 1× PBS supplemented with 0.0025% of protein stabilizer (Echelone Biosciences) and used for detection of PI(4,5)P 2 . PI(4,5)P 2 dissolved in 1× PBS supplemented with 0.0025% protein stabilizer was added in duplicate into 96-well fl at-bottom plates and incubated at room temperature for 2 h. The wells were then washed with 1× PBS three times, and then blocked in 5% BSA in PBS overnight at 4°C. The wells were again washed with 1× PBS three times, and then incubated with primary mouse anti-PI(4,5)P 2 (Abcam) in blocking solution at a dilution of 1:2500 for 1.5 h at room temperature. The wells were washed with 1× PBS and incubated with secondary goat anti-mouse IgG labeled with horseradish peroxidase (KPL, Gaithersburg, MD) in blocking solution at a dilution of 1:5,000 for 1 h at room temperature and protected from light. This was followed by incubation in TMB high-sensitivity substrate solution (BioLegend) for 5 min at room temperature and protected from light. The reaction was stopped by the addition of 1 N H 2 SO 4 , and the absorbance was read at 450 nm. A standard curve was generated with known concentrations of PI(4,5)P 2 (Echelon Biosciences). To test the cross-reactivity of the primary antibody, 50 pmol of PI(4)P and PI(3,4,5)P 3 (Avanti) were dissolved in 1× PBS supplemented with 0.0025% protein stabilizer, added into separate wells, and subjected to the same ELISA protocol as described. No detectable signals were observed (data not shown). In addition, ELISA data were validated by mass spectrometry ( 23 ). consisted of 10 mmol/l ammonium acetate (pH 8.5) and methanol using a linear methanol gradient consisting of 50% to 60% in 10 min , and then from 60% to 80% in 4 min. This was then increased to 100% methanol concentration over the next 6 min and kept at this condition for an additional 2 min to achieve chromatographic baseline resolution. The fl ow rate was 0.5 ml/min with a column temperature of 60°C. The mass spectrometer was operated in negative electrospray ionization mode with a cone voltage of 40 V. Source temperatures were 125°C with a desolvation gas temperature at 350°C. Collision-induced dissociation of the eicosanoids was performed using argon gas at a cell pressure of 1.6 × 10 Ϫ 3 Torr. Eicosanoids were detected and quantifi ed by multiple reaction mode monitoring of the transitions m/z 351 → 271 for AA-derived PGE 2 and m/z 353 → 317 for DGLA-derived PGE 1 ( 21 ).

Analysis of colonic cell proliferation
In vivo colonic cell proliferation was determined by immunohistochemical detection. Mice were intraperitoneally injected with 5-ethynyl-2´-deoxyuridine (EdU, 30 mg/kg body weight) 2 h prior to termination. One centimeter of distal colon was removed, fi xed in 4% paraformaldehyde, followed by a series of ethanol washes, and embedded in paraffi n. The incorporation of EdU into DNA of actively dividing cells was determined using a commercially available kit (Click-iT EdU Alexa Fluor 647 Imaging Kit; Invitrogen). Briefl y, after deparaffi nization, samples were washed in 3% BSA in PBS, treated with 0.5% Triton in PBS for 20 min, washed again with 3% BSA in PBS, then incubated with Click-It reaction cocktail for 30 min. Coupling of EdU to the Alexa fl uor substrate was then observed using fl uorescence microscopy ( 22 ).

Targeted deletion of Fads1 in mice
Fads1 knockout mice were generated by a gene-trapping technique described in Materials and Methods. Tail DNA was genotyped by PCR. As shown in Fig. 1 , Fads1 gene product (310 bp) was only detected in Wt and Het animals, whereas the gene trap product (270 bp) was amplifi ed in both Null and Het animals. To confi rm that targeted deletion resulted in the anticipated reduction in ⌬ 5 desaturase expression, Fads1 and Fads2 mRNA expression levels in mouse colon mucosa and liver were measured by RT-qPCR. As shown in Fig. 2 , Fads1 expression levels were altered as expected in haploinsuffi cient Het and Null mice compared with Wt siblings. In comparison, Fads2 expression levels were not different ( P > 0.05) among the three genotypes.

Deletion of Fads1 reduces viability
To limit the exogenous source of AA for the Fads 1 mice, animals were fed an LA-enriched 10% saffl ower oil-based commercial diet (cat# D03092902R; Research Diets) devoid of AA. There was no signifi cant effect of genotype on animal body weights ( P > 0.05) (supplementary Fig. II). With respect to longevity, Null mice began to die gradually starting at 5-6 weeks of age, with no survivors past 12 weeks of age ( Fig. 3 ). The average age at death of male Nulls was 7.6 weeks and of female Nulls, 7.5 weeks. From a gross anatomical perspective, no overt physical differences between Wt/Het and Null mice were observed. Occasionally, Null mice exhibited hip dysplasia at ‫ف‬ 5 weeks of age, but this did not occur in all mice ( ‫ف‬ 20% of the time). Often, there were no visible signs of any illness or infi rmity as close as

Characterization of immune cell populations
A single-cell suspension was produced by combining spleen and mesenteric and inguinal lymph nodes, which were pushed through a sterile stainless-steel wire screen (100 mesh) and resuspended in RPMI 1640 medium with 25 mmol/l HEPES (Irvine Scientifi c), supplemented with 10% fetal bovine serum (FBS, Irvine Scientifi c), 2 mmol/l GlutaMAX (Gibco), penicillin 100 U/ml, and streptomycin 0.1 mg/ml (Gibco); henceforth, "complete medium." Subsequently, a mononuclear cell suspension was produced by density gradient centrifugation using Lympholyte-M (Cedarlane Laboratories). Cell numbers were determined using a hemocytometer, and viability assessed by trypan blue exclusion always exceeded 96% in each genotype. The T-cell compartment was identifi ed by surface expression of CD3, and the antigenpresenting cell (APC) compartment was identifi ed by surface expression of major histocompatibility complex (MHC) class II (i.e., I-A[b]). For this purpose, 10 6 viable mononuclear cells were incubated for 10 min with a Fc ␥ R-blocking monoclonal antibody (1 µg/ml) (2.4G2; BD Pharmingen) on ice and were subsequently stained with either 1 µg/ml of PE-anti-mouse I-A[b] (clone AF6-1201; BD Bioscience) or 1 µg/ml of APC-anti-mouse CD3 ␣ (clone 145-2C11; eBioscience) antibodies for 30 min. Flow cytometric analysis was conducted using an Accuri C6 fl ow cytometer (Accuri Cytometers).

Colitis induction
Animals were exposed to dextran sodium sulfate (DSS, molecular weight, 36,000-50,000; MP Biomedicals) treatment as previously described ( 24,25 ). To induce intestinal infl ammation, 2.5% DSS was administered in the drinking water for 5 days, followed by 14 days of tap water.

Histological scoring
Various tissues (brain, heart, lung, liver, kidney, stomach, colon, and small intestine) were removed, fi xed in 10% neutral buffered formalin and paraffi n embedded. Sections were stained with hematoxylin and eosin. Histological examination was performed in a blinded manner by a board-certifi ed pathologist (B. Weeks ), and the degree of infl ammation (score, 0-3) and epithelial injury (score, 0-3) in microscopic cross-sections of the tissues was graded as described previously ( 20 ).

Statistical analysis
Data were analyzed using two-way ANOVA. Differences of P < 0.05 were considered signifi cant. was signifi cantly ( P < 0.05) different from that in both Wt and Null mice ( Fig. 4 ).

Deletion of Fads1 alters 1-and 2-series eicosanoid levels
DGLA and AA are precursors to 1-and 2-series prostanglandins (PGE 1 and PGE 2 ), respectively. Therefore, we further probed the effect of gene deletion-induced alterations in DGLA and AA levels with respect to the biosynthesis of 1-and 2-series prostaglandins. Prostaglandins were extracted from colonic mucosa, small intestine mucosa, and lung and measured by mass spectrometry ( Fig. 5 ). As expected, PGE 2 levels were signifi cantly lower, whereas PGE 1 levels were higher in Null versus Wt mice ( P < 0.05). Het mice exhibited an intermediate PGE 2 /PGE 1 phenotype.

Fads1 deletion decreases colonic cell proliferation
It has been reported that AA-derived prostaglandins enhance cell proliferation in colon cancer cells ( 26 ). Because intestinal PGE 2 was decreased in Fads1 null mice (described above), we examined the level of cell proliferation in mouse colon crypts. As shown in Fig. 6 , deletion of the Fads1 gene did not alter the number of cells in colonic 12 h prior to death of the Null mice. Null mice did not exhibit lethargy, lack of grooming, skin conditions, or infi rmity of any kind. No visible internal differences appeared during dissection. Histologically, tissues showed no differences between Wt/Het and Null mice (e.g., colon, duodenum, stomach, liver, kidney, lung, brain, and heart) (supplementary Fig. III). Clinical blood biochemistry parameters were also normal (data not shown). Mice were not evaluated reproductively. In an attempt to rescue the Null mice, graded levels of AA (0.1, 0.4, or 2% w/w) were added to the diet. The highest level of supplementation (2%) extended the longevity of Null mice to levels comparable with Wt and Het mice ( Fig. 3 ). To further characterize the phenotype, mice were exposed to DSS to induce acute intestinal infl ammation ( 24,25 ). All Wt and Het mice survived the DSS challenge. In contrast, Null mice exhibited a signifi cant ( P < 0.05) reduction in viability during the 5-day DSS treatment period (supplementary Fig. IV). These data indicate that Fads1 Null mice are unable to tolerate an acute intestinal infl ammatory challenge.

Deletion of Fads1 reciprocally alters the levels of AA and DGLA in mouse tissues
Membrane total phospholipid fatty acid profi les were measured in Fads1 mouse colonic mucosa, liver, splenocytes, and serum ( Fig. 4 ). Differences were primarily observed in the levels of DGLA and AA ( P < 0.05) in Null mice compared with Wt and Het siblings, with AA decreased to nearly undetectable levels in multiple organ sites in Null animals (supplementary Tables I-IV). Although the levels of AA and DGLA in Het mice were comparable with Wt mice, the ratio of AA/DGLA in Het mice  compared basal immune cell populations in Fads1 Wt, Het, and Null mice. Specifi cally, mononuclear cells were isolated, and CD3 + T-cell and MHCII-cell populations were quantifi ed by fl ow cytometry. As shown in Fig. 8 , the total number of mononuclear cells and percentage of CD3 + T cells were signifi cantly ( P < 0.05) reduced in Null mice. In contrast, there was no difference in the level of MHCIIexpressing cell populations across the three genotypes. In complementary experiments, culture supernatants were collected from mononuclear cells and isolated from spleen, and then mesenteric and inguinal lymph nodes were incubated overnight with agents designed to stimulate APCs (LPS, ␣ CD40) or T cells ( ␣ CD3 + ␣ CD28). Following overnight stimulation ex vivo, mixed cultures (i.e., mononuclear cell cultures composed of both T cells and APCs) from both Het and Null mice contained lower levels of IFN ␥ and TNF ␣ following T-cell activation ( Fig. 9 and supplementary Table V).

DISCUSSION
In this study, we determined the utility of antagonizing tissue AA levels as a novel approach to suppressing AAderived eicosanoids (e.g., PGE 2 ) biosynthesis. For this purpose, we have generated a previously unknown genetic model, namely, the Fads1 ( ⌬ 5 desaturase) knockout mouse. In many tissues, LA is converted to AA by an alternating sequence of ⌬ 6 desaturation ( Fads2 gene product), chain elongation, and ⌬ 5 desaturation ( Fads1 gene product), in which hydrogen atoms are selectively removed to create new double bonds, and then two carbon atoms are added to lengthen the fatty acid chain ( 29 ). Dietary AA crypts, although the percentage of EdU -labeled cells was signifi cantly ( P < 0.05) decreased in Null compared with Wt and Het mice. Het mice exhibited an intermediate proliferative phenotype between Wt and Null mice, consistent with the levels of AA and PGE 2 ( Figs. 4 and 5 ).

CD4 + T cell PI(4,5)P 2 levels are reduced in Fads1 null mice
As membrane phospholipid PI(4,5)P 2 is a key node of second messenger metabolism in T cells ( 12 ) and its acyl backbone is predominantly composed of 1-stearoyl-2arachidonoyl species, we examined whether the absence of AA associated with the deletion of the Fads1 gene altered the PI(4,5)P 2 concentration in CD4 + T cells. The basal level of PI(4,5)P 2 was signifi cantly ( P < 0.05) decreased by approximately 24% in Fads1 Null CD4 + T cells compared with Het and Wt siblings ( Fig. 7 ).

Immune cell populations are altered by Fads1 deletion
As AA-derived eicosanoids play an important role in immune regulation ( 27,28 ), we next investigated the immunomodulatory effects associated with a defi ciency in AA-derived 2-series prostaglandins. For this purpose, we  acid analysis also confi rmed that Fads1 null mice have negligible levels of AA ( ‫ف‬ 1 mol%) compared with heterozygous ( ‫ف‬ 14 mol%) and wild-type ( ‫ف‬ 15 mol%) mice ( Fig. 4 and supplementary Tables I-IV). Similar effects were observed in the liver and spleen, indicating a systemic depletion of AA. In contrast, Fads1 ablation resulted in the massive enhancement of dihomo-␥ -linolenic acid (20:3 ⌬ 8,11,14 , DGLA), the 1-series prostaglandin substrate in select tissues ( Fig. 4 ). As anticipated, the alteration in prostaglandin precursor levels was associated with a profound shift in DGLA and AA-derived eicosanoid biosynthesis ( Fig. 5 ). Interestingly, 1-series derived prostaglandins, e.g., PGE 1 , interact with discrete receptors compared with PGE 2 ( 30,31 ). This is noteworthy, because PGE 1 is capable of inhibiting colon cancer cell proliferation ( 32,33 ) as opposed to PGE 2 , which promotes growth ( 1,34 ). Therefore, we are in a unique position to explore the impact of carcinogen exposure on mice that have virtually no 2-series prostaglandin substrate (20:4 ⌬ 5,8,11,14 , AA). Consistent with the loss of AA in membrane phospholipids, it is noteworthy that PIP 2 levels were decreased in T cells isolated from Null mice ( Fig. 7 ). Considering that PIP 2 is made up primarily of AA-containing molecular species ( 35 ), we propose that the lack of AA substrate for the 1-acyl-glycero 3-phosphoinositol acyltransferase-dependent remodeling likely contributed to a reduction in its mass ( 23,33,36 ). Because PIP 2 controls the activity of numerous proteins and serves as a source of second messengers ( 37 ), further work is needed to assess how AA defi ciency affects individual phospholipid classes and PIP 2 -dependent signaling at the plasma membrane. These data will provide insight regarding how different tissues compensate in response to depletion of AA.
Disruption of the Fads1 versus Fads2 genes in mice manifests distinct phenotypic outcomes. For example, deletion of ⌬ 6 desaturase resulted in dermal and intestinal ulceration (38)(39)(40). In contrast, we report that ⌬ 5 desaturase null mice failed to thrive, gradually dying off at 5-6 weeks of age, with no survivors past 12 weeks of age ( Fig. 3A ). In most cases, the pathologies observed in Fads2 null mice were not apparent until approximately 17 weeks of age for females and 11 weeks for males (38)(39)(40). Fads1 Null mice did not live long enough to see if similar adverse events were present. A contributing factor for the survival of the Fads2 null mice was likely the modest amounts of AA present in the "control" diets ( 39,40 ). This would have allowed the mice to survive and develop fertility and spermatogenesis complications. In addition, differences in the genetic background of the mice may have affected the phenotype. All Fads 2 null mice were on a mixed (129S6/ Svev/Tac/C57BL/6) background. Fads 1 mice were on a pure C57BL/6 background. In addition, differences between these metabolically related models may be attributed to the fact that DGLA-derived 1-series eicosanoids are able to compensate for some of the AA-derived eicosanoid functions ( 30 ). In contrast to Fads1 Null mice, Fads2 mice are unable to synthesize 1-series PGs due to the absence of DGLA ( 38 ). Although the precise cause of the loss of viability in Fads1 Null mice is still under investigation, this bypasses the rate-limiting ⌬ 5 desaturation step and can be metabolized by cyclooxygenase enzymes (Ptgs1 and Ptgs2) to generate immunomodulatory, tumor-promoting 2series prostaglandins (e.g., PGE 2 ) (supplementary Fig. I) ( 30 ).
Initial genotyping/phenotyping indicated that we successfully deleted the Fads1 gene. Intestinal mucosa fatty Fig. 7. Levels of PI(4,5)P 2 in Fads1 knockout mouse T cells. Splenic CD4 + T cells were isolated, and the levels of PI(4,5)P 2 were quantifi ed by indirect anti-PIP 2 ELISA. Data were normalized to wild-type (Wt) at time 0; mean ± SEM, n = 3. Values not sharing the same letter indicate signifi cant differences ( P < 0.05). Values not sharing the same letter indicate signifi cant differences ( P < 0.05). address our overall goal, to determine at a mechanistic level, the utility of antagonizing tissue AA levels as a novel approach to suppressing AA-dependent infl ammation and cancer.
phenotype was fully reversible upon supplementation of AA to the diet ( Fig. 3B ). With regard to the underlying mechanisms driving the perturbation of immune cell populations observed in Fads1 knockout mice ( Figs. 8 and 9 ), given the importance of PI(4,5)P 2 and PGE 2 in regulating immune cell function (41)(42)(43), it is possible that the suppression of these critical mediators contributed to a disruption in immune cell homeostasis.
The lack of AA-derived eicosanoid production is consistent with the observed reduction in intestinal crypt proliferation ( Fig. 6 ) and the inability of Fads1 Null mice to tolerate an acute intestinal infl ammatory challenge (supplementary Fig. IV). Similar effects have been reported in microsomal PGE synthase Null mice and following cyclooxygenase-2 defi ciency ( 44,45 ). PGE 2 in the intestine has a protective effect on the integrity of the epithelial intestinal wall, and its loss promotes polymicrobial sepsis, a frequently fatal disease ( 45 ). Clearly, until the full effects of Fads1 deletion on the entire spectrum of AA-derived eicosanoids is known, it is not possible to exclusively link the loss of tissue AA to the specifi c role of PGE 2 .
In summary, we have generated and characterized a novel AA-defi cient ( Fads1 knockout) mouse model. Although young-adult mice die prematurely, this phenotype can be rescued by adding AA to the diet. Hence, in future studies, it will be possible to titrate membrane levels of AA. Lipidomic analyses indicate that DGLA and AA-derived mediators are reciprocally modulated in this model. Given the central role of AA-derived eicosanoids in epithelial and immune cell biology, we are in a unique position to Fig. 9. Levels of T-cell-derived infl ammatory cytokines. T cells were isolated from pooled spleens and mesenteric and inguinal lymph nodes, and then stimulated with 5 g/ml of plate-bound anti-CD3 plus 20 µg/ml of soluble anti-CD28 for 24 h. Culture supernatants were collected, and cytokine levels were measured using a Bio-Plex 200 System (Bio-Rad). Only selected cytokines are shown; mean ± SEM, n = 3. Values not sharing the same letter indicate signifi cant differences ( P < 0.05).