Role of the 5-Lipoxygenase–activating Protein (FLAP) in Murine Acute Inflammatory Responses

A 680-bp fragment of rat Flap cDNA (17) was used as a probe to screen a genomic DNA library in the Lambda FIX II vector generated from the 129 mouse strain (Stratagene Corp., La Jolla, CA). Five positive clones were obtained. One of these genomic clones, Flap 221, was isolated and purified to homogeneity and the insert was subcloned into the NotI site of pBluescript II SK+ (Stratagene Corp.). Regions of the insert containing coding sequence were identified and sequenced by the dideoxy chain termination method (U.S. Biochemical, Cleveland, OH) to verify the identity of the clone and to further localize regions containing coding sequence.

A targeting vector for the Flap gene was constructed from the murine genomic clone. Fragments identical to their corresponding regions in the gene were inserted 5′ and 3′ of the neomycin resistance gene (neo) of pJNS2 (36) such that coding material would be replaced by neo after homologous recombination of the plasmid with the endogenous locus. This construct was electroporated into E14TG2a embryonic stem (ES) cells established from the 129/Ola mouse line (37), and colonies resistant to neomycin and ganciclovir were identified as described previously (36). Genomic DNA was isolated from doubly resistant ES cell lines, digested with EcoRV and analyzed by Southern blot using a 1.2-kb BamHI/NotI fragment downstream of the targeting construct as a probe (Fig. 1 A). Verified recombinant ES cell lines were microinjected into mouse blastocysts, and chimeric mice were generated. Chimeras were bred to either B6D2 or 129/SvEv mice, and offspring were genotyped by Southern blot analysis of tail DNA digested with EcoRV using the 1.2-kb probe described above. All animals used in our studies of inflammation, with the exception of those assayed for platelet-activating factor (PAF) survival, were offspring of chimeras crossed with 129/SvEv mice. Likewise, the 5-LO–deficient mice used as controls for these experiments were on the 129 genetic background.

Macrophages were obtained by peritoneal lavage with 10 ml of DME supplemented with 20 mM Hepes buffer 3 d after mice received an intraperitoneal injection of 1 ml of 5% thioglycollate medium. Cells were cultured in 1.5 ml DME for 30 min. Supernatant was collected and replaced by medium containing A23187 Ca²⁺ ionophore at 5 μg/ml for an additional 30 min. A23187 has been shown to stimulate the release of eicosanoids from macrophages (38). Supernatant was collected, and the number of cells was verified by using a hexosaminidase assay, as is previously described (39). Concentrations of LTC₄, prostaglandin E₂ (PGE₂) and thromboxane B₂ (TXB₂) in the pre- and poststimulation supernatants were determined using enzyme immunoassay (EIA; Cayman Chemical Co. Inc., Ann Arbor, MI).

Macrophages from mice of the indicated genotypes were isolated by peritoneal lavage as described above. Total RNA was extracted from the cells using RNAzol B (Tel-Test, Inc., Friendswood, TX) according to the protocol supplied by the manufacturer and separated by gel electrophoresis (15–20 μg/lane). Northern blot analysis was carried out according to standard techniques using the 680-bp rat Flap cDNA (17) as a probe to detect expression of Flap mRNA. A probe for rat β-actin was used as a positive control. Densitometry was performed to detect any changes in Flap expression (ImageQuaNT v4.1 software; Molecular Dynamics, Sunnyvale, CA).

129 strain Flap (−/−) and (+/+) animals were injected intravenously with 0.5% Evans blue dye (10 ml of dye solution/kg of body weight) dissolved in PBS (2.7 mM KCl, 1.2 mM KH₂PO₄, 138 mM NaCl, 8.1 mM Na₂HPO₄, pH 7.5) and filtered through a 0.2 μm nitrocellulose membrane. The inside of the left ear was painted with 20 μl of AA (100 mg/ml in acetone; Sigma Chemical Co., St. Louis, MO) to induce an inflammatory response, while the right ear received acetone alone on its inner surface. After a time period of 1–2.5 h for edema assays or 7 h for myeloperoxidase (MPO) analysis, mice were killed and an 8-mm-diam disc of tissue was obtained from the center of each ear for analysis. For comparison, experiments were carried out in parallel with 129 5-lo (−/−) mice.

The wet weight of each ear biopsy was determined to assess edema formation. Each ear disc was then incubated in 1 ml of formamide at 55°C for 24 h. Extravasation of Evans blue dye was quantified by spectrophotometric analysis of the formamide extracts at 610 nm (40).

8-mm-diam ear punches were homogenized and analyzed according to a protocol modified from Bradley et al. (41) which has been previously described (42). Lysate MPO was quantified by comparison to a standard curve derived from serial dilution of MPO standard (Calbiochem Corp., San Diego, CA).

To induce peritoneal inflammation in Flap (−/−), Flap (+/+), and 5-lo (−/−) animals, we used a protocol modified from Rao et al. (43). Zymosan A (Sigma Chemical Co.) was dissolved in warm (56°C) PBS (pH 7.5) to a final concentration of 1 mg/ml. The Zymosan A solution was mixed until it passed freely through the 30-gauge delivery needle. Each mouse received 1 ml Zymosan A solution by intraperitoneal injection. After 3 h, mice were euthanized and a peritoneal lavage was performed with 4 ml of ice-cold PBS. The lavage fluid was collected, transferred to microcentrifuge tubes, and centrifuged at 40,000 g for 10 min. Supernatants were collected and LTC₄ levels were determined by EIA (Cayman Chemical Co.). In related experiments, mice received 0.5% Evans blue dye (10 ml of dye solution/kg of body weight) in PBS intravenously immediately before injection of Zymosan A solution. Mice were euthanized after 10, 20, or 30 min and underwent peritoneal lavage with 4 ml of cold PBS as described above. Evans blue dye extravasation was assessed by light spectrophotometry at 610 nm.

Peritoneal inflammation was induced in FLAP-deficient and wild-type animals as described above. Lavage fluid was collected 30 min after Zymosan A delivery and was centrifuged. 12-hydroxyeicosatetraenoic acid (HETE) levels in the supernatants were determined by EIA (PerSeptive Diag., Framingham, MA).

A method adapted from that of Furue and Tamaki (44) was used to analyze the effect of leukotriene deficiency on the development of a delayed-type hypersensitivity (DTH) response. In brief, 129 strain Flap (−/−), 5-lo (−/−), and (+/+) mice homozygous for the p and c ^ch alleles of the 129/Ola line were sensitized by application of a 100 μl epicutaneous dose of 0.5% FITC (Sigma Chemical Co.) in 1:1 acetone/dibutyl phthalate to the abdomen. After 6 d, the ears of each animal were measured using calipers before eliciting a contact hypersensitivity reaction by topical application of 20 μl of 0.5% FITC to the left ear. The right ear of each mouse received vehicle alone. 24 h later, the ear thickness was again measured with calipers. Mice were then killed, and ears were collected and weighed as described above. Caliper measurements, as well as the differences in wet weight between experimental (left) and control (right) ears, were compared in mice of each genotype.

Passive systemic anaphylaxis was induced as previously described (36). Animals received 20 μg of a monoclonal mouse anti–mouse DNP IgE in 200 μl of PBS intravenously. After 24 h, baseline body temperatures were determined using a rectal probe, and mice received an intravenous injection of 1 mg DNP–human serum albumin (HSA) and 0.5% Evans blue in 200 μl PBS to induce anaphylaxis. 30 min after this injection, temperature changes were measured. Mice were then killed, and edema was quantitated in 8-mm-diam ear punches as described above.

Mice received intravenous injections of 50 μg of PAF per kilogram of body weight in BSA at 2.5 mg/ml and propranolol at 0.05 mg/ml in saline. Survival was assessed for 24 h after the injection of PAF.

The Flap targeting vector was constructed using a clone isolated from a library constructed with DNA from the 129 mouse strain (Fig. 1,A). Fragments derived from the mouse Flap clone were used to prepare a plasmid that, upon homologous recombination with the endogenous locus, would result in the disruption of the gene for this protein. The 800-bp XhoI/NheI fragment, which is replaced by the neomycin resistance cassette after homologous recombination (see Fig. 1 A), was subcloned and sequenced. The sequence analysis showed that a 130-bp exon which encodes amino acids 108–151 is eliminated in recombinant clones. These amino acids have been shown to be highly conserved between several species (86% identity with the human amino acid sequence), and their removal would result in the loss of the third putative transmembrane domain of the protein (18). Thus, even if splicing around the missing exon were to take place, a large and conserved portion of the transcript would be absent and an inactive protein would likely result.

The JNS2/Flap1 plasmid was electroporated into E14TG2a ES cells, and potential targeted clones were selected by growth in the presence of G418 and ganciclovir. Resistant colonies were picked, expanded, and screened for homologous recombination by Southern blot analysis of EcoRV digested genomic DNA using a 1.2-kb BamHI/NotI probe downstream of the targeting construct (Fig. 1,A). This probe hybridizes to a fragment of >16 kb from the endogenous Flap locus, while introduction of a new EcoRV site into the targeted locus results in a hybridizing fragment of 8 kb (Fig. 1 B). A targeting frequency of 34% was observed with the JNS2/Flap1 construct. Microinjection of two verified recombinant ES cell lines, 160-9 and 160-37, into mouse blastocysts yielded chimeric animals which could transmit the recombinant Flap allele to their progeny. Each of these cell lines gave rise to mice homozygous for the Flap mutation in the F2 generation. Flap (−/−) mice were present in the litters in expected numbers based on Mendelian ratios for a heterozygous mating. This indicates that disruption of the Flap gene had no effect on the fetal survival or development of the animals. By simple observation, the FLAP-deficient mice were indistinguishable from their Flap (+/−) and Flap (+/+) littermates.

To verify that the predicted coding sequence had been deleted from the targeted Flap locus in these transgenic mice, DNA samples from tail biopsies of Flap (−/−), Flap (+/−), and Flap (+/+) mice were analyzed by Southern blot, using the 800-bp XhoI/NheI fragment replaced by neo in the recombinant locus. This probe hybridized to Flap (+/+) and Flap (+/−) DNA, but showed no hybridization to DNA from a Flap (−/−) animal, verifying the deletion of this critical coding region (data not shown).

Total RNA was isolated from peritoneal macrophages of Flap (−/−), Flap (+/+), and 5-lo (−/−) mice. A Northern blot probed with the rat Flap cDNA (Fig. 1 C) shows that the murine Flap mRNA is present in the Flap (+/+) and 5-lo (−/−) lanes, but there is no hybridization in the Flap (−/−) lanes. Because the full-length cDNA probe contains sequences in addition to those eliminated by targeting, these results support the interpretation that a critical region of the Flap gene has been deleted through recombination and that there is no stable message being synthesized.

Densitometry was performed on the autoradiograph in Fig. 1 C. Comparison of the integrated areas of the Flap mRNA bands relative to the levels of the β-actin mRNA indicated only a slight elevation in Flap expression in macrophages from 5-lo (−/−) mice. In the (+/+) sample, Flap mRNA was 53.53% of β-actin mRNA levels which are an indicator of the amount of RNA loaded in a given lane. In the 5-lo (−/−) lane, Flap mRNA was 59.87% of β-actin mRNA levels, or 1.12-fold of that measured in the wild type. Thus, the absence of 5-LO results in little difference in Flap expression.

Peritoneal macrophages collected from four Flap (−/−) mice and from three (+/+) littermates were stimulated with calcium ionophore to determine the ability of these cells to release prostaglandins, thromboxanes, and leukotrienes. After A23187 activation, a >200-fold increase in LTC₄ was detectable in supernatant from macrophages of wild-type mice, while no LTC₄ was detected in cells from Flap (−/−) animals (Fig. 2,A). The ability of the FLAPdeficient mice to produce prostaglandins and thromboxanes by the cyclooxygenase-dependent pathway remained intact (Fig. 2, B and C). Similar to results obtained with 5-LO–deficient mice (42), levels of PGE₂ and TXB₂ that were released from the Flap (−/−) macrophages after stimulation were greater than those released from their (+/+) counterparts. While both PGE₂ and TXB₂ production by FLAP-deficient mice was increased compared to controls, only the difference in prostaglandin production reached statistical significance (P = 0.028 and 0.078, respectively).

To determine whether a similar defect in leukotriene synthesis occurred in vivo, an experimental peritonitis was induced by the injection of Zymosan A into the peritoneal cavity of Flap (−/−), 5-lo (−/−), and (+/+) mice. In wild-type mice, Zymosan A induced a significant stimulation of LTC₄ (1,580 ± 22 pg/ml). Consistent with the in vitro data (Fig. 2 A), however, no LTC₄ could be detected in either Flap (−/−) or 5-lo (−/−) mice after Zymosan A injection.

This question was further addressed in the accompanying manuscript by performance of HPLC analysis of peritoneal lavage fluids after Zymosan A stimulation of Flap (−/−) and (+/+) animals on the DBA/1 lacJ genetic background. The chromatography verified that no leukotrienes are being synthesized by FLAP-deficient animals. Furthermore, these profiles failed to detect an increase in other AA metabolites of the 5-LO pathway (for example, 5-HETE and the nonenzymatic hydrolysis products of LTA₄) in the absence of FLAP. Surprisingly, a significant decrease in 12-HETE, a product of the 12-LO pathway, was observed in the DBA/1 lacJ Flap (−/−) mice. An ∼54% reduction in 12-HETE compared to wild-type levels was verified in these mutant mice by EIA. We have seen a similar trend in an identical EIA for 12-HETE in 4 ml of stimulated peritoneal lavage fluid from four Flap (−/−) and three (+/+) mice on the 129 genetic background (50.25 ng/ml for FLAP-deficient mice versus 82.16 ng/ml in their normal controls). This reduction in 12-HETE levels, however, did not achieve statistical significance.

AA, when topically applied to the skin of the murine ear, causes an intense acute inflammatory reaction with both vascular and cellular components. Several lines of evidence support a role for eicosanoids in this inflammation, including the recent demonstration that this response is attenuated in 5-LO–deficient and prostaglandin H₂ synthase-1–deficient animals (42, 45, 46). To examine the role of FLAP in the edema and neutrophil recruitment characteristic of this response, the ears of Flap (−/−) and control mice were treated with 2 μg of AA. The ears were later biopsied, and edema was measured by determining the wet weight of each biopsy. In conjunction with the measurement of edema, alteration in plasma exudation was assessed using Evans blue dye. Since Evans blue binds to serum proteins (40), quantification of this dye in each ear biopsy gives a reliable indication of changes in vascular permeability. Edema and vascular permeability in the Flap (−/−) mice were compared to wildtype littermates.

As can be seen in Fig. 3 A, the increase in ear weight after AA treatment was significantly reduced in Flap (−/−) mice compared to controls (P = 8.9 × 10⁻⁹). To determine whether the changes seen in the FLAP-negative animals were due solely to leukotriene deficiency, 5-lo (−/−) mice were examined in the same way. As reported, the AA-induced edema was significantly reduced in 5-lo (−/−) animals compared to control mice (P = 5.9 × 10⁻¹⁶). Surprisingly, however, the reduction in edema in the 5-lo (−/−) was greater than that of the Flap (−/−) animals (P = 0.001). The differences between these groups of mice were replicated in at least two experiments performed on independent days, with a total of 17 5-lo (−/−), 17 Flap (−/−), and 38 (+/+) animals examined.

Changes in vascular permeability and extravasation of serum proteins in response to topical AA followed a similar pattern for mice of these three genotypes (Fig. 3 B). Ears of wild-type mice displayed a deep blue color after topical AA treatment, reflecting leakage of serum proteins into the interstitial space, while Flap (−/−) mice and 5-lo (−/−) mice showed a lighter blue color by observation and after quantitation (P = 1.7 × 10⁻⁷ and 3.5 × 10⁻¹², respectively). Again, the loss of 5-lo gene function had a greater impact on edema formation than did the disruption of the gene for FLAP, and this relationship was consistent in at least two experiments performed on separate days with a total of 16 5-lo (−/−), 17 Flap (−/−), and 38 (+/+) animals. The difference between the two experimental groups was not as pronounced, however, as the difference in their ear weight described above (P = 0.0445).

In addition to their roles in edema formation and vascular changes, the leukotrienes are important in the establishment and maintenance of the cellular response to acute injury. LTB₄ exerts a potent chemotactic effect upon neutrophils, causing the adhesion of PMN to vascular endothelia and extravasation into the tissues in the acute inflammatory response. We therefore examined the ability of FLAP to affect the cellular infiltration elicited by topical AA. Levels of MPO, an enzyme specific for neutrophils, can be used as an indicator of the number of PMN present in inflamed tissue (38). The MPO level for experimental and control ears of mice of each genotype was determined by enzymatic assay 7 h after the topical AA treatment. The MPO content of unstimulated ears was uniformly undetectable by our assay for all mice tested. After AA stimulation, ear biopsies exhibited MPO activity, reflecting the recruitment of neutrophils to the treated tissue. Mean MPO activity for twelve 129 strain (+/+), six Flap (−/−), and five 5-lo (−/−) mice is shown in Fig. 3 C. Neutrophil infiltrate was significantly decreased in the Flap (−/−) and 5-lo (−/−) animals compared to the wild-type controls (P = 10⁻⁵ and 7 × 10⁻⁶ respectively). These results were obtained in two separate experiments. In contrast to the assays of edema and vascular leakage, we observed no difference in MPO levels between FLAP– and 5-LO–deficient mice. Thus, in the Flap (−/−) animals we attribute reduction in recruitment of PMN to the inability of these mice to produce leukotrienes.

To further explore the roles of FLAP, 5-LO, and leukotrienes in inflammatory edema, we evaluated plasma protein extravasation in response to another injurious stimulus, Zymosan A treatment. Peritoneal lavage was performed at 0, 10, 20, and 30 min time points, and the wash fluid was analyzed for Evans blue. As shown in Fig. 4, the absence of FLAP and 5-LO had similar effects on Evans blue extravasation after Zymosan A exposure. In two independent experiments, no difference was observed in the degree of inhibition conferred by FLAP or 5-LO deficiency.

Contact hypersensitivity to FITC is a form of DTH induced by topical application of antigen. It is initiated by T cells and can be quantitated by challenging the ears of sensitized animals and then measuring the swelling of the tissue after 24 h. Previous reports have suggested that leukotrienes may affect T cell cytokine production in vivo (47, 48). Production of leukotrienes by myeloid cells activated in this inflammatory response are likely to contribute further to the pathological changes of the DTH response. To examine a possible role for FLAP and/or leukotrienes in the DTH response, the ability of five Flap (−/−) and five 5-lo (−/−) mice to mount a DTH response to cutaneous FITC in acetone/dibutyl phthalate was compared to that of nine wild-type animals. In wild-type mice, ear weight increased (14.12 ± 2.8 mg) after FITC challenge. Both FLAP– and 5-LO–deficient mice tended to exhibit a smaller increase in ear edema after similar treatment (Fig. 5). The differences between the three groups of mice, however, do not achieve statistical significance. FLAP-deficient animals showed a mean Δ ear weight of 10.42 ± 2.16 mg after treatment, whereas 5-LO–deficient mice had a Δ ear weight of 10.62 ± 2.93 mg (P = 0.16 and 0.23, respectively).

Systemic IgE anaphylaxis in the mouse is characterized by bronchconstriction, plasma extravasation, hypotension with resultant tachycardia, and drop in body temperature. In the passive IgE-induced anaphylaxis model, these changes are dependent on binding of IgE to the high affinity IgE receptor, followed by mast cell degranulation (36). Leukotrienes, which are not stored in granules but are synthesized when the cells are activated, are among several proinflammatory mediators released by stimulated mast cells that can contribute to these physiological effects. To test the role of FLAP in the anaphylactic response, IgE-dependent passive systemic anaphylaxis was induced in five FLAP-deficient and four control mice. Mast cells of the animals were loaded with a monoclonal mouse anti–mouse DNP IgE. 24 h later DNP-HSA and Evans blue were administered. Body temperature was taken using a rectal probe before and 30 min after the DNP-HSA treatment. Finally, mice were killed, and edema was quantitated by measuring the amount of Evans blue dye in 8-mm-diam ear biopsies. Flap (−/−) mice averaged a 3.76 ± 0.35°C drop in body temperature after the induction of passive anaphylaxis that was less extreme than the 4.98 ± 1.04°C drop seen in the (+/+) mice (Fig. 6,A). This reduced response, however, was not statistically significant (P = 0.174). Likewise, the A₆₁₀ of ear extracts from each group showed similar levels of Evans blue dye extravasation (Fig. 6 B). The mean A₆₁₀ of extracts from the ears of FLAP-deficient mice was 0.1006 ± 0.0057. Wild-type mice edema levels reached A₆₁₀ = 0.0859 ± 0.0082, which again indicates no significant difference between Flap (−/−) and (+/+) animals (P = 0.075). Analogous results have been seen on examination of the 5-LO–deficient mice (data not shown).

PAF is a potent endogenous mediator of inflammatory systemic anaphylaxis and shock, causing microvascular leakage, vasodilation, contraction of smooth muscle, endothelial adhesion, and activation of neutrophils, macrophages, and eosinophils. It is thought that many of the physiological changes that result from intravenous doses of PAF could be brought about by the PAF-mediated release of leukotrienes (42, 45, 49–52). Vulnerability of FLAP-deficient mice and their wild-type littermates to PAF-induced anaphalaxis was determined by treating 12 Flap (−/−) and 8 (+/+) animals with intravenous PAF. 75% (9 of 12) of Flap (−/−) animals survived this treatment, compared to only 12.5% (1 of 8) of the control group. Analysis of these results using the χ² test indicates a difference in susceptibility to PAF-induced anaphylaxis between the Flap (−/−) and (+/+) groups with a significance level of 0.0062. This difference is similar to that seen between the 5-LO–deficient and wild-type mice, as reported previously (42, 45). The protection from anaphylactic shock conferred by FLAP or 5-LO deficiency suggests a crucial role for leukotrienes in this process. While other inflammatory mediators are involved in the systemic anaphylactic response, none are able to compensate for the absence of the leukotrienes.

FLAP was discovered by the identification of targets for agents that block leukotriene synthesis but have no effect on 5-LO enzyme activity (22). The protein has been shown to possess significant structural homology with LTC₄ synthase, another enzyme involved in the leukotriene synthetic pathway. In particular, the region of FLAP that is critical for the action of the inhibitory drug MK-886 is conserved in LTC₄ synthase, suggesting that the two proteins may also share some functional homology (19). Here we characterized the role of FLAP by examining the inflammatory responses of mice that are homozygous for a mutant allele of the Flap gene. To determine whether alterations in these responses are due solely to FLAP's facilitation of leukotriene synthesis, parallel experiments were performed with mice carrying a mutation in the 5-lo gene. These mice, like the FLAP-deficient animals, were generated by introducing mutations into the genome of embryonic stem cells by homologous recombination (42).

No difference was observed in the ability of mice lacking FLAP to survive, nor was any alteration in the development of the various populations of myeloid and lymphoid cells apparent. An essential role for FLAP in the production of leukotrienes was suggested by in vitro studies and by work using drugs that were shown to bind specifically to FLAP (17, 22, 23). Here we show that leukotriene synthesis is undetectable in cells or whole animals that fail to express FLAP. This confirms the requirement of functional FLAP for normal production of leukotrienes. Also, the lack of leukotriene production in these mice indicates that the newly isolated 30-kD FLAP-2 (20), which shows homology to both FLAP and LTC₄ synthase, is not a functional complement of FLAP in AA metabolism. In addition, FLAPdeficient mice showed slight decreases in 12-HETE production in response to Zymosan A stimulus. Although we did not see a statistically significant difference in our experiments, examination of a larger group of mice in the accompanying paper confirmed this observation. It is possible, therefore, that whereas not absolutely required, FLAP facilitates the production of 12-HETE in the 12-LO pathway.

To further define the role of leukotrienes in acute inflammatory processes and to continue our exploration of potential additional functions for FLAP, we analyzed FLAP– and 5-LO–deficient mice in a number of experimental models of inflammation. We initially focused on vascular and cellular changes after topical AA administration. Attenuation of this response has been reported previously in mice deficient in the 5-LO protein (42, 45). In our own early studies, we found that edema was decreased in 5-LO–deficient mice, but the difference was not statistically significant. Much of the difficulty in achieving statistically significant differences between the two groups of animals stemmed from the considerable variability in the responses of both wild-type and 5-LO–deficient mice to AA. We reasoned that the inflammatory responses in these mice, which were 129 and B6D2 F2 offspring, were influenced by genetic factors in addition to the 5-lo gene. To examine the roles of FLAP and 5-LO independent of these factors, the in vivo experiments described in this report (with the exception of PAF-induced anaphylaxis) were carried out on mice in which the mutated 5-lo and Flap genes were maintained on the 129 background by breeding the chimeric animals (generated from 129/Ola ES cells) to 129/SvEv mice. When background variability was eliminated, a statistically significant reduction in mean edema was achieved between the 5-LO– deficient and control groups.

The edema and increased vasopermeability characteristic of AA-induced ear inflammation were reduced in the FLAP-deficient mice relative to control animals, but each remained significantly higher than that of the 5-LO–deficient animals. This was, notably, the only assay in which a reproducible and consistent difference was seen between the 5-LO– and FLAP–deficient mice. Even in this inflammatory response, significant differences were seen between these two groups only when inflammation was assessed by measuring changes in the weight of a constant area of tissue. Smaller changes were observed on evaluation of vasopermeability and cellular inflammation. One possible explanation for the difference in response between mice lacking 5-LO and FLAP in this particular assay may be that, in some circumstances of extremely high AA substrate, formation of leukotrienes is possible without the association of the enzyme with FLAP (17). This would be compatible with the finding that in vitro formation of leukotrienes from AA is possible in the absence of FLAP. Experiments that have examined the dependence of leukotriene synthesis on FLAP have shown that leukotriene production is stimulated by the treatment of cells with ionophore. Consistent with the results obtained here by stimulating FLAPdeficient peritoneal macrophages, leukotriene synthesis was found to be inhibited by drugs that interfere with the function of FLAP (23, 53). These results and our finding that FLAP-deficient mice show some response to topical AA support the theory that, under unique conditions of high substrate availability, production of leukotrienes is possible in the absence of FLAP. It is not clear why the positive response of the FLAP-deficient mice to AA was most apparent in measurements of ear weight edema and (to a lesser extent) vascular permeability, whereas levels of cellular infiltration in similarly treated ears were equivalent in the FLAP– and 5-LO–deficient mice. One possible explanation for these results is that the enzymes that carry out subsequent metabolism of LTA₄ differ depending on the intracellular localization of the 5-LO enzyme. Metabolism of AA in the absence of FLAP may occur at intracellular sites other than the nuclear envelope, where 5-LO and FLAP normally interact. The production of LTA₄ at a different intracellular location could alter its availability to LTA₄ hydrolase and LTC₄ synthase and, in turn, affect the ratio of LTB₄ to LTC₄ production in FLAP-deficient cells. Because LTB₄ and LTC₄ differ in their contributions to vascular and cellular events during inflammation, changes in the production of these two leukotrienes in the FLAP-deficient mice could explain the results we see in the edema/plasma extravasation and neutrophil infiltration assays.

Edema formation is also characteristic of the acute inflammatory response seen in IgE-mediated anaphylaxis. The absence of observable differences in edema or vasopermeability between FLAP-deficient mice, 5-LO–deficient mice, and normal controls in these experiments is not surprising given the complexity of this response. In contrast with the relatively straightforward stimulation of a single metabolic pathway by topical AA, a wide variety of mediators that can lead to inflammatory changes similar to those induced by leukotrienes are produced in IgE-mediated anaphylaxis. It is likely that the release of other mediators such as histamine, a potent vasoactive substance, plays a key role. Although no differences were observed between FLAPdeficient and control mice here, under physiological conditions in which IgE receptors are not completely occupied or antigen is limited such that few mast cells are activated, the absence of leukotrienes may possibly have an impact on the extent of the edema formation.

The lack of FLAP and leukotrienes also failed to result in detectable attenuation of the contact hypersensitivity induced by FITC. Past reports suggested that leukotrienes may influence T cell–mediated responses at a number of levels. LTB₄ has been shown to alter IFN-γ, IL-2, and IL-1 production by T cells and monocytes (47, 48, 54). In addition, leukotrienes may contribute to this inflammatory process indirectly since they are likely to be produced by the myeloid cells recruited to sites of inflammation. Our failure to see a difference in this response may again reflect the complexity of the process. Numerous pathways may exist for eliciting the physiological changes that have been observed, and, whereas leukotrienes may contribute to such changes, other mediators may mask their deficiency. It could be that lower levels of stimulation would allow us to observe a more subtle change in this response. Such a situation has been described in an analysis of trinitrochlorobenzene-mediated contact hypersensitivity in IL-1–deficient mice (55). In this case, response variation in mutant mice was apparent only at low antigen concentrations.

In both the 5-LO– and the FLAP–deficient mice, striking alterations were seen in the kinetics of peritoneal edema formation after the injection of Zymosan A. Zymosan A can induce an inflammatory response through a number of pathways including activation of alternate pathway of complement, the generation of anaphylatoxins, and the activation of mast cells and basophils. The response is also characterized by the biosynthesis of arachidonic acid metabolites and by neutrophil and macrophage phagocytosis (43). Examination of the edema formation in the 5-LO– and FLAP–deficient mice shows that leukotrienes are essential mediators of the early reaction to this stimulus. Again, the similar responses seen in the FLAP– and the 5-LO– deficient mice suggest that FLAP has no additional role in inflammation beyond its absolute requirement for leukotriene synthesis.

In conclusion, we have used a genetic approach to show that mice made deficient in FLAP survive and breed normally but are generally unable to metabolize AA to leukotriene products. Further, we show that the absence of FLAP reduces the inflammatory response in a subset of model systems. The mice lacking FLAP, along with a growing number of other animals deficient in immunologically important genes, should eventually allow specific cellular and molecular dissection of the biochemical pathways leading to inflammatory pathologies.

The authors thank Drs. J. Snouwaert and T. Coffman for critically reading the manuscript and for their constructive comments. We also appreciate the expertise of A. Latour and B. Garges in ES cell manipulation and animal husbandry.

These studies were supported by Research grant R01 DK 46003-01 with equivalent cofunding by the National Heart, Lung and Blood Institute, by National Institutes of Health grants PO1-DK38103, and by a grant from Pfizer Central Research to B.H. Koller. R.S. Byrum is a recipient of a University of North Carolina Medical Alumni Endowment Fund Award.