Chemical Probes That Differentially Modulate Peroxisome Proliferator-activated Receptor α and BLTR, Nuclear and Cell Surface Receptors for Leukotriene B4 *

Peroxisome proliferator-activated receptor α (PPARα)is a nuclear receptor for various fatty acids, eicosanoids, and hypolipidemic drugs. In the presence of ligand, this transcription factor increases expression of target genes that are primarily associated with lipid homeostasis. We have previously reported PPARα as a nuclear receptor of the inflammatory mediator leukotriene B4 (LTB4) and demonstrated an anti-inflammatory function for PPARα in vivo (Devchand, P. R., Keller, H., Peters, J. M., Vazquez, M., Gonzalez, F. J., and Wahli, W. (1996) Nature 384, 39–43). LTB4 also has a cell surface receptor (BLTR) that mediates proinflammatory events, such as chemotaxis and chemokinesis (Yokomizo, T., Izumi, T., Chang, K., Takuwa, Y., and Shimizu, T. (1997)Nature 387, 620–624). In this study, we report on chemical probes that differentially modulate activity of these two LTB4 receptors. The compounds selected were originally characterized as synthetic BLTR effectors, both agonists and antagonists. Here, we evaluate the compounds as effectors of the three PPAR isotypes (α, β, and γ) by transient transfection assays and also determine whether the compounds are ligands for these nuclear receptors by coactivator-dependent receptor ligand interaction assay, a semifunctional in vitro assay. Because the compounds are PPARα selective, we further analyze their potency in a biological assay for the PPARα-mediated activity of lipid accumulation. These chemical probes will prove invaluable in dissecting processes that involve nuclear and cell surface LTB4 receptors and also aid in drug discovery programs.

Hormones and nutrient-derived molecules, such as retinoids and fatty acid derivatives, are important signals in many biological processes. Dysregulation or disruption of their signaling pathways can manifest in various ways, with defects that range in severity, rate of onset, and organ systems affected. For instance, a prolonged disturbance of lipid homeostasis is often associated with many late-onset inflammatory conditions, obesity, diabetes, and cardiovascular disease. In order to effi-ciently treat and prevent these prominent metabolic problems, a better understanding of the mechanisms involved in lipid regulation is required.
Intracellular targets for lipid mediators have been postulated for many years (1). However, it is only recently that we have seen the emergence of reports describing nuclear receptors for fatty acids and their derivatives (see Refs. 2 and 3 and references therein). Particular attention has focused on a group of ligand-activated transcription factors called peroxisome proliferator-activated receptors (PPARs). 1 The three PPAR isotypes (␣, ␤/␦, and ␥) form a distinct subclass of the nuclear hormone receptor superfamily (4). The functional complex is a heterodimer of PPAR and the retinoid X acid receptor (RXR) that binds to a consensus sequence in the promoter of target genes and can up-regulate transcription in the presence of a PPAR ligand. Although the PPAR target genes identified so far, are generally associated with lipid homeostasis, the extent of their involvement in biological processes related to disease are yet to be elucidated.
In transient transfection experiments, PPAR␣ activity can be induced by a range of structurally diverse compounds (see Ref. 5 and references therein). Many of these natural and synthetic compounds are bona fide PPAR␣ ligands (6 -11). Although it is now clear that PPAR␣ is a nuclear receptor for various fatty acids (e.g. linoleic and arachidonic acid), eicosanoids (e.g. 8(S)-HETE and LTB 4 ) and hypolipidemic drugs (e.g. fibrates and Wy14,643), we are only beginning to understand the functional relevance of these receptor-ligand interactions.
Our knowledge of the biology of PPAR␣ is still superficial but is increasing rapidly. Extensive analyses of PPAR␣ knockout mice indicate that PPAR␣ can be associated with many homeostatic functions of peroxisomes and also with response to various compounds termed peroxisome proliferators (for review, see Ref. 12). For example, wild-type mice respond to hypolipidemic drugs such as fibrates and Wy14,643; but the PPAR␣(Ϫ/Ϫ) mice display neither lowering of blood lipid levels nor the proliferation of peroxisomes in the liver (13,14). Consistent with this, PPAR␣(Ϫ/Ϫ) mice exhibit normal basal levels of hepatic fatty acidand ␤-oxidation. However, they lack the ability to increase expression of these genes in response to various peroxisome proliferators (13). In the liver, these induc-ible pathways serve not only to regulate normal dietary fatty acids but also as a detoxification or degradation process for xenobiotics and potent eicosanoid mediators, such as the chemoattractant leukotriene B 4 (15,16).
PPAR␣ has also been evaluated for its ability to induce adipogenesis (17). In vitro, NIH 3T3 fibroblasts can be retrovirally infected to express high levels of PPAR␣ and then challenged with potent activators to promote lipid droplet accumulation. High levels of lipid accumulation can in turn potentiate fat cell differentiation, a process that presumably involves a pathway mediated by the adipogenic regulator, PPAR␥.
So far, two studies have reported a role for PPAR␣ in inflammation. Recently, Staels et al. (18) have proposed that in vascular walls, PPAR␣ has an anti-inflammatory action that is mediated via repression of NF-B signaling. We have previously shown, in vivo, that PPAR␣ has a role in inflammation control (6,19). Compared with wild-type mice, the PPAR␣(Ϫ/Ϫ) mouse exhibits a prolonged inflammatory response when challenged by the eicosanoid LTB 4 and its precursor arachidonic acid, but not the phorbol ester 12-O-tetradecanoylphorbol-13acetate. Based on the available data, a potential mechanism proposed for this anti-inflammatory role of PPAR␣ is a negative feedback loop. LTB 4 would effectively control its own degradation by initiating the up-regulation of the fatty acid oxidation pathways. This catabolic inactivation of the eicosanoid is facilitated by direct interaction and activation of PPAR␣, a nuclear receptor for LTB 4 (K D of 60 and 90 nM, determined by two independent groups using fluorometric (20) and radioligand (6) binding assays).
Leukotriene B 4 also binds to a cell surface receptor, BLTR (21). A full-length cDNA for BLTR has recently been isolated from human HL-60 leukemia cells and characterized (22). The predicted protein contains the classical seven membrane-spanning domains but shows little amino acid homology to other known proteins. The proinflammatory effects of LTB 4 are thought to be triggered by high affinity binding to the BLTR on immune cells (K D ϭ 0.15 nM). Analyses of downstream signaling pathways in stably transfected Chinese hamster ovary cells indicate that BLTR potentially couples to different G-proteins.
At the mechanistic level, little is known about the LTB 4 signaling pathways. Elucidation of these pathways would be greatly facilitated by the use of chemical probes that could differentially target cell surface and nuclear LTB 4 receptors. PPAR␣ and BLTR are promising targets in therapeutic invention for lipid-related disorders. Indeed, the pharmaceutical industry has deployed considerable resources to the development of antagonists of BLTR as anti-inflammatory drugs and agonists of PPAR␣ as lipid-lowering compounds. Here, we investigate some of these compounds to find useful probes for BLTR and PPAR␣ function, the reasoning being that if PPAR␣ and BLTR share a ligand (LTB 4 ), then one would expect to find some overlap in recognition of other ligands. We report the effect of some BLTR effectors on transcriptional activation by the three PPAR subtypes. We then established whether the effects are mediated by direct interaction with the PPARs. Finally, we used NIH 3T3-mPPAR␣ cells as a biological assay to evaluate the potential of these activators in the PPAR␣associated activity of lipid accumulation.
Fusion Protein Constructs and Protein Expression-The LBD frag-ments (shown in Fig. 2A) for mPPAR␣ and mPPAR␤ were amplified by PCR from clones containing the full-length sequences, whereas mP-PAR␥ fragment was isolated as a BamHI-XbaI fragment from the vector pSG5-mPPAR␥-stop. The fragments were cloned in-frame into BamHI site of either pGEX-1 (for mPPAR␣) or pGEX-5X-3 (for mP-PAR␤ and mPPAR␥). The GST-PPAR (LBD) fusion proteins were expressed in Escherichia coli BL2 DE3 (pLysS) bacteria. Briefly, a freshly transformed colony was used to start a 2-ml innoculum (in log phase) for a 500-ml culture. After 12 h, cultures were diluted with 500 ml of medium and induced with either 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside for 4 -6 h (for mPPAR␤ and mPPAR␥) or 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 9 h (for mPPAR␣). All cultures were grown at 30°C in LB medium containing 17 g/ml chloramphenicol and 50 g/ml ampicillin. Bacteria pellets were stored at Ϫ70°C as aliquots equivalent to 50 ml of culture.
CARLA-The CARLA was performed as described by Krey et al. (11) with the following modifications. The above bacterial pellets were resuspended in 10 ml of lysis buffer (phosphate-buffered saline A containing 1% Triton X-100 and 0.5 mM phenylmethylsulfonyl fluoride) and lysed by repeated freeze-thaw. The DNA and insoluble matter were removed by centrifugation. Fusion proteins were purified onto glutathione-Sepharose beads at 4°C (Amersham Pharmacia Biotech), washed three times in lysis buffer, and equilibrated in NETN buffer (20 mM Tris, pH 8.0, 100 nM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, and 1 mM dithiothreitol) supplemented with 1% (w/v) of dry 2% milk powder prior to use in CARLA. The amount of protein used per reaction was 1 g for mPPAR␣ and 2-3 g for mPPAR␤ and mPPAR␥. Reactions were performed in NETN buffer supplemented with 1% (w/v) of dry 2% milk powder, and the first wash contained 0.5% milk powder. The exposure times for autoradiography was 12-24 h for mPPAR␣, and 1-4 h for mPPAR␤ and mPPAR␥.
Adipogenesis Assay-Generation of virally infected stable NIH mP-PAR␣ cell lines and adipogenesis assays were performed as described by Brun et al. (17). All test compounds were dissolved in Me 2 SO. Cells were used for either Oil Red O staining of neutral lipids or RNA isolation followed by Northern blot analysis of the aP2 message.
LIC Assay-The ligand-induced complex (LIC) assays were performed similar to Forman et al. (8). Briefly, reaction volumes were 20 l. The two proteins, baculovirus-expressed mRXR␤ (0.4 l) and in vitro translated mPPAR␣ (0.7 l), were incubated for 10 min on ice in 10 mM Tris-HCl, pH 8.5, 100 mM KCl, 2 mM MgCl 2 , 1 mM dithiothreitol, 10% glycerol, 0.1% Nonidet P-40, and 3.5 g of poly(dI-dC). After addition of test compounds (eicosanoids, drugs, or solvent control) and a 30-min incubation on ice, 1 ng of radiolabeled probe (the PPAR response element consensus site) was added, and the reaction was incubated for 20 min at room temperature. Complexes were resolved on 5% nondenaturing polyacrylamide gels in 0.5ϫ TBE. Gels were dried, and results were visualized by autoradiography. 4 acts as a potent chemoattractant for leukocytes (23). This process involves binding and activation of the LTB 4 to the cell surface receptor, BLTR (22). Based on this biological activity, one can evaluate effectors of BLTR by their ability to modulate LTB 4induced neutrophil chemotaxis. Compounds that enhance chemotaxis are classified as BLTR agonists, and those that inhibit chemotaxis are classified as antagonists. In this study, we used four BLTR effectors (Ref. 24 and Table I): two agonists (LTB 4 itself and ZK151657) and two antagonists (ZK158252 and ZK183838). For convenience, the synthetic compounds ZK151657, ZK158252 and ZK183838 are referred to as BLTR agonist 1 (BAg1), BLTR antagonist 1 (BAntag1), and BAntag2, respectively.

BLTR Effectors Are Partial Agonists of PPAR␣-LTB
The effects of the compounds on the three PPAR isotypes were evaluated by transient transfection experiments. In this system, HeLa cells are co-transfected with three constructs: a PPAR expression vector, a reporter construct containing the CAT gene driven by a PPAR-responsive promoter element (see Fig. 1A), and a vector that constitutively expresses ␤-galactosidase to standardize between different samples. After exposure to test compounds for 48 h, cells were harvested and lysed, and the CAT enzymatic activity per ␤-galactosidase unit was measured. A test compound that increases standardized CAT activity in a PPAR-dependent manner was classified as a PPAR activator.
We first determined the effect of 10 M concentrations of the synthetic compounds on the three isotypes of mouse PPARs (mPPARs, see Fig. 1B). Transfection results of empty pSG5 expression vector indicate that in the absence of exogenous PPAR, none of the compounds induced a positive response from the reporter construct. However, when the different mPPARs were cotransfected, isotype-specific responses to the synthetic compounds were observed. BAg1 and BAntag1 are activators that are selective for mPPAR␣. At a 10 M concentration, BAg1 also induced a small response in the presence of mPPAR␤. The third compound, BAntag2, did not induce mPPAR activity. Interestingly, mPPAR␥ activity was not induced by these compounds. Similar activation profiles were observed from PPAR␣ from other species (Xenopus and human) (data not shown).
The level of mPPAR␣ activation observed with 10 M of BAg1 and BAntag1 was 40 -50% that of the standard activator and ligand Wy14,643. This could reflect either that higher doses of the compounds are required for maximal activation or that the compounds are partial agonists. An activation profile showing the mPPAR␣ response in the presence of increasing amounts of compound, helps to differentiate between the two scenarios ( Fig. 1C). In these transient transfection assays, BAg1 and BAntag1 are partial agonists with equal efficacy; at 10 M, they gave quantitatively indistinguishable responses. Log dose curves show that BAg1 (EC 50 ϭ 0.5 M) is more potent than BAntag1 (EC 50 ϭ 8 M) and LTB 4 (EC 50 Ͼ 10 M).
BAg1 and BAntag1 Are PPAR␣ Ligands-The transient transfection assays report the activity of PPARs when cells are challenged with a given compound. This increased transcriptional activity could conceivably be a consequence of different mechanisms: a direct interaction of the compound with the nuclear receptor, an indirect mechanism of activation, such as production of a metabolite, or a combination of both. We used a semifunctional assay to evaluate whether the compounds can directly bind to the PPARs and render them functionally active.
Transcriptional activation by nuclear hormone receptors can be mediated by a class of proteins called co-activators (25). The current model indicates that the binding of a ligand to the nuclear hormone receptor induces a conformational change in the ligand binding domain (LBD) that allows interaction with the co-activator. For example, the steroid receptor coactivator 1 (SRC-1) binds to xPPAR␣ in the presence of its ligand Wy14,643 but not in the presence of a nonligand such as ␤-estradiol. This type of semifunctional assay (CARLA, see Fig.  2A), has been successfully exploited to identify different fatty acids, eicosanoids, and hypolipidemic agents as ligands for PPARs (11).
Studies have indicated that although PPAR response to natural compounds, such as fatty acids and eicosanoids, is conserved between species, response to xenobiotics can be speciesspecific (10). For this reason, GST fusion proteins with the different mouse PPAR LBDs were constructed (Fig. 2, B and C). Expression of soluble fusion protein of the mouse PPAR␣ isotype was not as efficient as for PPAR␤ and PPAR␥. However, all proteins were active, as confirmed by CARLA on known  ligands (Fig. 2D).
CARLA was used to evaluate the three BLTR effectors as ligands of mPPARs (Fig. 3). At 10 M concentrations, BAg1 and BAntag1 were ligands of mPPAR␣, but BAntag2 was not. Similar profiles were obtained with the Xenopus PPAR␣ (data not shown). BAg1, the weak activator of mPPAR␤ in transfection experiments, also bound to mPPAR␤. None of the compounds induced SRC-1 interaction with PPAR␥.
Potency of BLTR Effectors on PPAR␣ Biology-The function of PPAR␥ as a master gene in adipogenesis was first demonstrated by its ability to induce the adipogenic program in fibroblastic cells (for review, see Ref. 26). Different strains of NIH 3T3 cells do not induce lipid accumulation or expression of adipocyte markers when cultured under conditions permissive for adipogenesis. However, when PPAR␥ is ectopically expressed in these cells using a retroviral system, one can stimulate both the accumulation of lipid droplets and gene expression of adipocyte markers, such as the fatty acid-binding protein aP2 (27). Similar functional evaluation of the PPAR␣ and PPAR␤ isotypes have highlighted some important mechanistic differences between members of the PPAR family (17). For instance, the PPAR␣ isotype can prime fibroblastic cells to accumulate lipid droplets and also turn on the adipocyte markers. However, this PPAR␣-mediated process seems to be distinct from that of PPAR␥, and moreover is sensitive to the efficacy of the PPAR␣ activator. We used this PPAR␣ functional assay to evaluate the biological potency of our BLTR effectors.

FIG. 2.
A, schematic of CARLA. Fusion proteins of GST and the LBD of the nuclear hormone receptor were bacterially expressed and partially purified onto glutathione-Sephadex beads (Amersham Pharmacia Biotech). Beads were incubated with test compound (solvent control, ligand, or nonligand) and radiolabeled SRC-1 (produced in vitro using a coupled transcription-translation rabbit reticulocyte lysate system (Promega)). The reaction was incubated to equilibrium, and beads were recuperated by centrifugation. Beads were then washed and analyzed for interaction SRC-1 using SDS-polyacrylamide gel electrophoresis. Coomassie staining of GST-LBD fusions allows standardization between different reactions. The SRC-1 protein is visualized by autoradiography. Because a ligand enhances interaction between the LBD and SRC-1, the amount of SRC-1 pulled down in the presence of a ligand was higher than in the absence of ligand (solvent or nonligand compound). B, GST-mPPAR(LBD) fusions. The amino acids of the different PPARs included in the GST fusion proteins are indicated in the schematic. C, soluble proteins were partially purified on glutathione-Sepharose beads, and the beads were analyzed on a 10% SDS-polyacrylamide gel electrophoresis Coomassie stain. The amounts loaded are equivalent to 0.2 ml of bacterial culture for the GST control, 10 ml for GST-mPPAR␣, and 5 ml for GST-mPPAR␤ and GST-mPPAR␥. The sizes of the Bio-Rad low molecular weight standards (M) are indicated (in thousands), and the proteins of interest are marked to the left by a filled circle. D, evaluation of fusion proteins by CARLA. Autoradiograms showing amount of radiolabeled SRC-1 pulled down by respective GST fusion proteins (either mPPAR␣, ␤, or ␥). The solvent used was ethanol. Synthetic ligands (10 M) used were Wy14,643 for mPPAR␣, Merck A for mPPAR␤ (40), and BRL 49643 for mPPAR␥ (41). Nonligands (10 M) used were BRL 49643 for mPPAR␣ and Wy14,643 for mPPAR␤ and mPPAR␥. LTB 4 the ligand for mPPAR␣ was used at 50 M.
NIH PPAR␣ stable cell lines were created by transduction of NIH 3T3 fibroblasts with packaged pBABE-mPPAR␣ retroviral vector. Cells were cultured under conditions that permit adipogenesis and challenged with the various compounds. The effects of the compounds on cells were analyzed by staining of neutral lipids with Oil Red O or Northern blot analyses of the transcript for adipocyte marker, aP2. Consistent with published data (17), NIH PPAR␣ cells treated with Me 2 SO solvent scored negative for lipid accumulation, whereas those treated with the mPPAR␣ activator Wy14,643 accumulated lipid to a reasonable extent (Fig. 4A). In comparison to these standards, one can rank the potency of the BLTR effectors. Surprisingly, even though BAg1 is a partial agonist in transient transfection assays, it is a more potent inducer of PPAR␣-mediated lipid accumulation than Wy14,643. The staining pattern with BAn-tag1 was positive, but less efficient than the standard Wy14,643. BAntag2 scored negative in this assay.
Depending on their potency in lipid accumulation, activators of PPAR␣ can also trigger the adipogenic program in NIH PPAR␣ cells (17). For instance, Wy14,643 will induce expression of adipogenic genes, whereas a weaker mPPAR␣ activator will not. Analyses of NIH mPPAR␣ cells treated with BLTR effectors were consistent with this (Fig. 4B). The weak inducer of lipid accumulation (BAntag1) did not significantly increase aP2 mRNA, whereas BAg1, the most potent inducer of lipid accumulation, also induced expression of this adipocyte fatty acid-binding protein.
Stabilizing LTB 4 -The adipogenesis assay result with LTB 4 as inducer (Fig. 4A) was surprising, as one would expect that a ligand with reported K D values of 60 and 90 nM as estimated by radioligand and fluorescence-based assays (6,20) would induce lipid accumulation to a significant level at 10 M concentrations. This lack of response could be due to an inefficient uptake of this eicosanoid, although this is probably not the main reason because the two synthetic analogues (BAg1 and BAntag1) were effective in the assay. An alternate explanation is that LTB 4 is unstable under these assay conditions. We further explored this possibility.
The chemical and metabolic stability of LTB 4 (see Fig. 5) can be improved significantly by either stabilizing the reactive triene system (BAg1) or by introducing a trifluoromethyl group at the -position (CF 3 -LTB 4 , Fig. 6A and Ref. 28). If it is the instability of LTB 4 that hinders response in the assays, then modification of the molecule by either of the two methods should improve the outcome.
We have already seen above that BAg1 is more efficacious than LTB 4 in both transient transfections for mPPAR␣ (EC 50 of 0.5 versus Ͼ10 M) and in adipogenesis assays using NIH mPPAR␣ cells. We evaluated whether making LTB 4 stable to -oxidation induces a better response in different assay systems (Fig. 6). In transient transfection assays, the EC 50 for mPPAR␣ shifted from Ͼ10 M LTB 4 to 0.8 M CF 3 -LTB 4 (Fig.  6B). Results from the ligand detection assays are consistent with this (Fig. 6, C and D). In the LIC assay (8), complex FIG. 3. BLTR effectors as ligands for PPARs. CARLAs with GST-LBD fusion proteins of the three PPAR isotypes (␣, ␤, and ␥) were used to test BAg1, BAntag1, and BAntag2 as ligands for the different PPARs. One g of GST-PPAR (LBD) fusion protein was used for ␣, and 3 g was used for ␤ and ␥. Compounds were used at concentration of 10 M. Lanes with solvent alone (Me 2 SO (DMSO)) show the background level of SRC-1 obtained in the assay. Graphical representation of the average intensity of the SRC-1 bands for each duplicate is shown below. Me 2 SO lanes (DMSO) are used as reference point for each protein (␣, ␤, and ␥). BAg1 and BAntag1 were the ligands for mPPAR␣. BAg1 also bound to mPPAR␤. By this assay, none of the BLTR effectors were found to be ligands for mPPAR␥. formation was observed at 50 M LTB 4 , and stabilizing against -oxidation resulted in almost a 100-fold less concentration of CF 3 -LTB 4 required for LIC detection (Fig. 6C). The complexes observed are ligand-specific and do not occur in the presence of 50 or 100 M of non-PPAR␣ ligands, such as ␤-estradiol or the eicosanoid and LTB 4 precursor 5(S)-HETE. Consistent with previous reports (8), the PPAR-RXR complex did not occur at limiting concentrations of mPPAR␣ but could be induced in the presence of 1 M Wy14,643. The apparent affinity of mPPAR␣ for CF 3 -LTB 4 as detected by CARLA (Fig. 6D) is consistent with that observed in the LIC assay (0.5 M). Finally, we evaluated CF 3 -LTB 4 in the adipogenesis assay. When NIH mPPAR␣ cells were induced with as little as 1 M CF 3 -LTB 4 the result was significant lipid accumulation (Fig. 6E). Thus, effectively increasing the lifetime of LTB 4 , by stabilizing against -oxidation, resulted in the expected responses in the different assay systems. DISCUSSION Following its isolation, leukotriene B 4 was recognized as a potent chemoattractant and aggregatory agent in leukocytes (29,21). The realization of LTB 4 as a proinflammatory signal triggered many drug discovery programs for anti-inflammatory agents, including the synthesis of banks of compounds as effectors of LTB 4 -mediated processes. With the help of these chemical probes, rapid progress has been made in understanding the biology of LTB 4 .
It is now clear that the action of LTB 4 is tightly regulated at many levels (for review, see Ref. 30). Below, we outline some key features of regulation of LTB 4 biosynthesis, specificity of response, and inactivation by catabolism (Fig. 7). LTB 4 is a downstream product of the 5-lipoxygenase (5-LO) pathway (31,32). Accumulating evidence suggests that when activated, cytosolic and intranuclear pools of 5-LO translocate to the nuclear membrane, where they channel arachidonic acid to leukotriene biosynthesis (33). This activity is facilitated by FLAP, a nuclear envelope protein that binds the arachidonate released from the nuclear envelope phospholipids (34,35). The translocation of the FLAP/5-LO complex to the nuclear membrane conceivably results in two pools of leukotrienes. The nuclear pool predicts both target(s) and functions for leukotrienes in the nucleus. Consistent with this is the identification of PPAR␣ as a nuclear receptor for LTB 4 (6). Large quantities of LTB 4 are also released to the extracellular milieu to recruit circulating leukocytes (36). For many years, chemotaxis of the immune cells has been postulated to be mediated by a seven-membranespanning receptor. Recently, the full-length cDNA of BLTR from human HL-60 monocytic cell line has been isolated as a G protein-coupled cell surface receptor, which is functionally confirmed by its ability to render Chinese hamster ovary cells responsive to LTB 4 (22). One potential route for the inactivation of LTB 4 is through catabolism via the fatty acidand ␤-oxidation pathways (16). This process can be up-regulated by exposure to polyunsaturated fatty acids; to xenobiotics, such as the hypolipidemic drug clofibrate; and also to LTB 4 itself. These structurally unrelated compounds have recently been shown to be activators and ligands of the nuclear receptor PPAR␣ (for review, see Ref. 5).
Synthesis, uptake, and response to leukotriene B 4 are celltype dependent (for review, see Ref. 37). For instance, polymorphonuclear leukocytes respond to LTB 4 by chemotaxis and hyperadhesion, and once activated, they can both produce and catabolize the eicosanoid. Hepatocytes, on the other hand, are efficient at clearing LTB 4 from the system via uptake and catabolism but they are unable to synthesize LTB 4 from arachidonic acid.
At a simplistic level, one can correlate the tissue expression patterns of the two LTB 4 receptors with what is known about their function thus far. The cell surface receptor, BLTR responds to subnanomolar concentrations of LTB 4 and is expressed primarily in leukocytes, where it mediates chemotaxis and hyperadhesiveness (21). The nuclear receptor PPAR␣ has a 1000-fold lower affinity for LTB 4 and also exhibits a less restricted expression pattern (e.g. liver, kidney, brown adipose tissue, and immune system). Both characteristics of PPAR␣ reflect a broader role in adaptive responses of the organism, including lipid homeostasis and detoxification of xenobiotics and lipid mediators (38). The different affinities of the two receptors for LTB 4 , as well as the overlap in expression patterns, potentially result in complex cell contexts. For instance, polymorphonuclear leukocytes express both cell surface and nuclear receptors for LTB 4 . This poses an interesting problem because polymorphonuclear leukocytes are capable of all aspects of LTB 4 processes (production, response, and catabolism), and the balance of these activities is critical to the final biological outcome of inflammation (Fig. 7). To evaluate the mechanisms of cross-talk, we need tools that will differentially modulate activities of the two LTB 4 receptors.
BLTR Effectors-The structures of the BLTR effectors used are depicted in Fig. 5. Two compounds, the BLTR agonist ZK151657 (BAg1) and BLTR antagonist ZK158252 (BAntag1) are from libraries based on the structure of leukotriene B 4 . For example, BAg1 is a stable BLTR agonist resulting from replacement of the reactive conjugated triene system in LTB 4 with a substituted pyridine ring. The third compound, ZK183838 (BAntag2), can be described as an acetylphenone type derivative or aryl-carboxylic acid derivative. It was isolated from a random compound library, and its structure is not obviously related to LTB 4 .
BLTR and PPAR␣ Are Pharmacologically Distinct-The three synthetic compounds ( Fig. 5 and Table II) cover an interesting spectrum of activity on the two LTB 4 receptors. ZK151657 (BAg1) is an efficacious activator of both nuclear and cell surface receptors. Compared with the natural agonist and ligand LTB 4 , the compound is more potent on PPAR␣ activity. This is reflected in both the transient transfection assay (EC 50 of 0.5 M) and the ability to accumulate lipid droplets in the biological assay. Even though ZK151657 is only a partial agonist of mPPAR␣, the adipogenesis assay indicates that it is potent enough to induce expression of the aP2 gene, which codes for the adipocyte fatty acid-binding protein. ZK158252 (BAntag1) has an interesting profile with opposite effects on the cell surface and nuclear receptors. It is a partial agonist of PPAR␣, with a relatively high EC 50 of 8 M. Evaluation by the biological assay indicates that it is potent enough to induce lipid droplet accumulation, but not to the extent where cells markedly increase expression of the adipocyte-specific genes. ZK183838 (BAntag2) is an antagonist of BLTR that does not induce PPAR␣ activity in transfection experiments. Consequently, it does not induce either coactivator interaction or lipid accumulation in NIH mPPAR␣ cells.
We are only beginning to understand the significance of the sharing of ligands as a means of cross-talk between cell surface and nuclear receptors. In such a scenario, understanding the response of a cell to the shared ligand requires tools that differentiate or uncouple the different receptors. An earlier report by Reginato et al. (39) studies the role of PPAR␥ as a nuclear receptor in cross-talk of signaling by two different eicosanoids. The focus of this study is BLTR and PPAR␣ as receptors for a common eicosanoid ligand, LTB 4 . The three FIG. 6. Stabilizing LTB 4 to -oxidation. A, structure of CF 3 -LTB 4 indicates the position on LTB 4 at which CF 3 was substituted for CH 3 . B, dose-response curve for CF 3 -LTB 4 on mPPAR␣ obtained in HeLa transient transfection experiments using the same system as in Fig. 1. C, LIC assay. The complex of PPAR-RXR heterodimer bound to a PPAR response element DNA consensus site (42) is indicated by the arrow. Concentrations of various compounds are indicated. LTB 4 induced complex formation at 50 and 100 M. This is specific because no complex was observed with equal concentrations of either ␤-estradiol or 5(S)-HETE, the precursor to LTB 4 . Consistent with its stability, CF 3 -LTB 4 is a ligand at lower concentrations than LTB 4 . D, binding of CF 3 -LTB 4 to mPPAR␣ in a CARLA. Increasing concentrations of CF 3 -LTB 4 from left to right: 0 (solvent control), 0.01, 0.1, 1, 2, and 5 M. The GST-mPPAR␣ protein was visualized by Coomassie staining (above) and radiolabeled SRC-1 by autoradiography (below). E, adipogenesis assay on NIH PPAR␣ cells. In the assay system described in Fig. 4A, 1 M of CF 3 -LTB 4 induced a significant amount of lipid accumulation as determined by Oil Red O staining.

FIG. 7. LTB 4 metabolism in an activated neutrophil.
Leukotriene B 4 binds to the cell surface receptor BLTR (K D ϭ 0.15 nM). Upon activation, nuclear and cytoplasmic pools of 5-LO and the 5-lipoxygenase-associated protein (FLAP) translocate to the nuclear membrane, where they channel arachidonic acid into synthesis of leukotrienes. Potentially, two pools of LTB 4 are produced, one in the nucleus and a second (cytoplasmic) that is released to the extracellular milieu. PPAR␣, a nuclear target for LTB 4 (K D ϭ 90 nM), is thought to mediate catabolism or degradation of LTB 4 . Hence, the neutrophil is capable of producing, responding to, and degrading LTB 4 . How the cell integrates the appropriate response at the right time is still unclear.

TABLE II
Summary of effects of synthetic BLTR effectors on mPPAR␣ CARLA and aP2 mRNA accumulation are stated simply as significant (ϩ) or no significant (Ϫ) effect. Lipid accumulation is summarized by the amount of oil red staining ranging from Ϫ (not significant) to ϩϩϩϩ (highest). Ϫ Ϫ Ϫ Ϫ synthetic ligands identified are promising tools to probe for detailed mechanisms of the interplay between nuclear and cell surface LTB 4 receptors. Given the importance and success of BLTR and PPAR␣ ligands as therapeutic targets, this information will no doubt aid in drug discovery for disorders associated with disturbance of lipid homeostasis, including inflammatory disorders, diabetes, obesity, and cardiovascular disease.