Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Angiogenesis is a fundamental process by which new capillaries are formed from existing blood vessels. This process plays important roles in physiological events such as formation of the corpus luteum, development of the embryo, and wound healing, including recovery from both myocardial ischemia and peptic ulcer (Folkman and Shing, 1992). Unregulated growth of blood vessels can contribute to tissue injury in a large number of diseases such as arthritis, diabetes, and tumor progression (Folkman, 1995). Endothelial cells are normally quiescent and are activated during the angiogenic response. Upon stimulation, endothelial cells can degrade their basement membrane and proximal extracellular matrix, migrate directionally, then divide and organize into functional capillaries invested by a new basal lamina (Arenberg and Strieter, 1999).

There is a growing body of evidence demonstrating that the angiogenic switch is regulated by the net balance between positive and negative regulators of new capillary growth (Folkman, 1995). Persistence of neovascularization requires a proangiogenic environment, with the expression of angiogenic factors outweighing that of angiostatic factors. A range of peptides can influence this balance, including mitogenic factors such as vascular endothelial growth factor (VEGF) (Arenberg and Strieter, 1999), nonmitogenic factors (selected cytokines, CXC chemokines), and internal peptide fragments of angiostatin and endostatin (Arenberg and Strieter, 1999). Certain eicosanoids also display potent angiogenic properties (Ziche and Gullino, 1982; Ben-Av et al., 1995). In rabbits, PGE₂, PGF₂, and prostacyclin (PGI₂) stimulate angiogenesis where prostaglandin E series, in particular PGE₁, is most potent (Ziche and Gullino, 1982). PGE₂ is a potent inducer of VEGF expression in synovial fibroblasts (Ben-Av et al., 1995). In addition to its well known vasodilator and antiplatelet properties, PGI₂ can also induce VEGF gene expression and protein synthesis (Höper et al., 1997). It was recently reported that 12-lipoxygenase activity and one of its products, 12-(S)-HETE, regulate angiogenic responses (Nie et al., 2000), and that cytochrome P450-derived 12-(R)-hydroxyeicosatrienoic acid stimulates angiogenesis via nuclear factor-B (Stoltz et al., 1996). The cyclooxygenase (COX)-2 gene in endothelial cells is rapidly up-regulated by several growth factors as well as inducers of angiogenesis (Hla et al., 1993). Along these lines, results from three different endothelial cell models show that COX-2 is an essential component of angiogenesis, at least in vitro (Jones et al., 1999). Nonsteroidal anti-inflammatory drugs such as aspirin (ASA) have been implicated in the prevention of certain cancers such as lung and colon cancer (Marcus, 1995; Vane, 2000) that might be related to the ability of ASA to reduce angiogenesis (Hla et al., 1993).

Aspirin's well known therapeutic mechanism of action includes inhibition of COX-derived prostanoids (Vane, 2000). After the discovery of COX-2, this laboratory reported that, when acetylated by ASA, the ability of COX-2 to generate prostanoids is blocked, yet this enzyme remains active in endothelial cells, epithelial cells, and mononuclear cells and initiates the biosynthesis of new products of cell-cell interactions or transcellular biosynthesis termed aspirin-triggered-15-epi-lipoxins (ATLs) (Clària and Serhan, 1995). These novel endogenous lipid mediators are the carbon 15 epimers of lipoxin (LX) that carry their 15 alcohol in the R configuration compared with their native LX counterparts and appear to mimic most if not all endogenous LX bioactivities. To date the actions of ATL appear to be most relevant in regulating inflammatory responses, because they are generated during cell-cell interactions that can involve, for example, endothelial cells-neutrophils in vivo (Chiang et al., 1998), and display potent inhibitory actions in several key and strategic events in inflammation (Serhan, 1997; Chiang et al., 1998; Clish et al., 1999). Both LX and ATL actions include inhibiting adhesion and transmigration of neutrophils, and hence can serve as counter-regulatory signals to limit and/or regulate leukocyte accumulation that are potentially operative in the dampening and resolution of inflammatory sites (Serhan, 1997). Because LX are rapidly generated and inactivated with the local microenvironment, stable analogs of both LX and ATL were designed that enhance both their bioavailability and activities compared with their native products (Serhan, 1997). The ATL analogs also proved to be ~100 times the potency of ASA (Clish et al., 1999). In this report, we investigated whether ATL could regulate endothelial responses in vitro and in vivo that are relevant for angiogenesis. Using a metabolically more stable ATL synthetic analog, [15-epi-16-(para-fluoro)-phenoxy-lipoxin A₄ (denoted ATL-1)], ATL proved to be a potent angiostatic eicosanoid in vivo, identifying new activities for LX and ATL as endogenous regulators of endothelial responses that are in sharp contrast to the actions of other eicosanoids and may be relevant in several human diseases.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture. Human umbilical vein endothelial cells (HUVECs) were isolated by 0.1% collagenase digestion (Worthington Biochemical, Bedford, MA) and propagated on gelatin-coated (0.1%) tissue culture plates in medium 199 (Invitrogen, Carlsbad, CA) supplemented with 20% heat-inactivated fetal bovine serum (BioWhittaker, Walkersville, MD), 50 µg/ml of endothelial cell mitogen (Biomedical Technologies, Stoughton, MA), 8 U/ml heparin (American Pharmaceutical Partners, Los Angeles, CA), 50 U/ml penicillin, and 15 µg/ml streptomycin. Only passages 2 and 3 were used in reported experiments.

Endothelial Cell Proliferation. HUVECs (5 × 10³) were plated in 96-well plates coated with 0.1% gelatin for 1 h at room temperature. After 24 h, the medium was removed and replaced with fresh medium 199 supplemented with 5% fetal bovine serum and different concentrations of recombinant human VEGF₁₆₅ (R & D Systems, Minneapolis, MN), LTD₄ or LTB₄ (Cayman Chemical, Ann Arbor, MI). Endothelial cells were enumerated after 72 h by using the MTT assay (Sigma Chemical Co., St. Louis, MO) (Marshall et al., 1995). ATL-1 was prepared as in Clish et al. (1999) and was a gift from Drs. Daniel Perez, John Parkinson, and William Guilford (Berlex Biosciences, Richmond, CA). Percentage of inhibition was evaluated in a similar manner and included a 15-min incubation (37°C) with 15-epi-lipoxin A₄, ATL-1 before the addition of agonists. All incubations were performed in triplicate. Before each experiment the integrity and concentration of ATL-1 was assessed by physical methods, including liquid chromatography/mass spectrometry/mass spectrometry and UV (Clish et al., 1999).

Endothelial Cell Migration. VEGF, ATL-1, or vehicle was added to the lower wells of a 48-well chemotaxis chamber (NeuroProbe, Cabin John, MD). The wells were overlaid with a 10-µm pore size polycarbonate filter coated with 0.1% gelatin. HUVECs (1 × 10⁶) were placed in the upper wells and the chamber was incubated (37°C, 5% CO₂ for 12 h). After incubations, filters were removed, scraped of cells from the upper surface, fixed, and stained with Diff-Quik (Dade Behring, Newark, DE). Cells that migrated across the filter toward the lower surface were enumerated by light microscopy; four fields were counted at high magnification (100×). Incubations were performed in triplicate. To assess inhibition, endothelial cells were suspended in media with vehicle or ATL-1 for 15 min before placement in the chamber.

Quantitative Determination of DNA Fragmentation. DNA fragmentation in individual apoptotic cells was quantitated using a photometric enzyme immunoassay (Apoptosis detection kit; R&D Systems). HUVECs grown in 96-well microtiter plates (5 × 10³ cells/well) were incubated for 3 days, fixed with 3.7% formaldehyde, and permeabilized with proteinase K before the labeling. Biotinylated nucleotides are incorporated onto the DNA fragments and detected by using streptavidin-horseradish peroxidase conjugate followed by the substrate TACS-Sapphire.

Inflammatory Angiogenesis. Angiogenesis was assessed with murine air pouches that were raised via subcutaneous injection of sterile air (3 ml) beneath the dorsal skin of anesthetized mice (BALB/c, male 6-8 weeks old). After 24 h, ATL-1 (10 µg/pouch) or vehicle was delivered locally, immediately before the injection of VEGF (1 µg/pouch). The vascular content was assessed by the formation of vascular casts (as in Colville-Nash et al., 1995). Briefly, mice were anesthetized (at 144 h) and peripheral vasodilation was raised by placing the animals in a heated jacket (40°C, 10 min). Vascular casts were formed by the i.v. injection of 1 ml of 5% carmine red (Sigma Chemical) in 5% gelatin solution warmed to 40°C. Air pouch linings were dissected and weighed. The tissue was then dissolved in 2 ml of 3 N NaOH solution for 0.5 h, 21°C and completely digested in hot water (56°C) for 10 min. Digested samples were centrifuged (2500 rpm, 15 min) and filtered through a 0.45-µm filter. The dye content was quantified using a 96-well plate spectrophotometer at 530 nm with a calibration curve. The results were expressed as vascular index as micrograms of carmine dye per milligram weight of tissue, for n = 4 animals/group. For visualization of the vasculature, the dorsal surface of the pouches was excised and fixed in formalin for 48 h. The tissues were dehydrated with 100% ethanol (5 days, 4°C) and cleared by immersion in cedar wood oil for 2 weeks. In another set of experiments, mice were anesthetized and injected i.v. with 200 µl of 0.05 g/ml fluorescein isothiocyanate-dextran (Sigma Chemical) in phosphate-buffered saline at 144 h immediately before sacrifice. Dissected linings were fixed, mounted on glass slides, and examined for fluorescence (model E600 Nikon Eclipse; Nikon, Melville, NY). In both protocols, the observers were not blinded to the treatments.

Immunohistochemistry. Air pouch membranous tissues were fixed in 10% buffered formalin overnight and processed for paraffin embedding. Five-micrometer paraffin sections of membrane tissue cut en face were used for immunohistochemistry for CD31 expression. Briefly, slides were deparaffinized and pretreated in 0.25% trypsin (Sigma Chemical) for 20 min at 37°C, followed by washing in distilled water. All further steps were performed at room temperature in a hydrated chamber. Slides were pretreated with Peroxidase Block (DAKO, Carpinteria, CA) for 5 min to quench endogenous peroxidase activity, followed by a 1:5 dilution of goat serum in 50 Mm Tris-Cl, pH 7.4, for 20 min to block nonspecific binding sites. Primary rat anti-murine CD31 antibody (BD PharMingen, San Diego, CA) was applied at a 1:100 dilution in 50 mM Tris-Cl, pH 7.4, with 3% goat serum for 1 h. After washing in 50 mM Tris-Cl, pH 7.4, secondary rabbit anti-rat antibody (DAKO) was applied at a 1:200 dilution in 50 mM Tris-Cl, pH 7.4, with 3% goat serum for 30 min. Slides were washed again in 50 mM Tris-Cl, pH 7.4, and goat anti-rabbit horseradish peroxidase-conjugated antibody (Envision detection kit; DAKO) was applied for 30 min. After further washing, immunoperoxidase staining was developed using a DAB chromogen kit (DAKO) per the manufacturer and counterstained with hematoxylin.

Statistical Analysis. Results are presented as means ± S.E.M. Statistical evaluation of the results was performed by analysis of variance, and values of P < 0.05 were taken to represent statistically significant differences between group means.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The aspirin-triggered lipoxin A₄ stable analog (denoted ATL-1) proved a potent inhibitor of VEGF-stimulated proliferation of HUVECs (IC₅₀ of ~3 nM) (Fig. 1). Inhibition was concentration-dependent and maximal at 10 nM and was partially reversed (from 38.0 ± 2.5 to 78.0 ± 6.2% of VEGF-stimulated proliferation, P < 0.05) by incubating the cells with genistein, an inhibitor of tyrosine kinase activity (50 µM, 5 min, 37°C, n = 3). Even at high concentrations (100 nM), the ATL-1 analog alone had no apparent actions on HUVEC proliferation (Fig. 1, inset). In sharp contrast, LXB₄ stable analogs increased proliferation (n = 2; data not shown). Because these are related structures, the separate actions of ATL and LXB₄ analogs with these cells indicate that the ATL-1 response is highly stereoselective. Also, a direct comparison of ATL-1 with native LXA₄ and 15-epi-LXA₄ at equimolar concentrations (10 nM) in a representative experiment showed that ATL-1 > 15-epi-LXA₄ > LXA₄ in rank order of activity (e.g., 63.3 ± 3.3, 59.8 ± 1.8, and 38.1 ± 2.0% inhibition, respectively). After exposure of the cells to ATL-1, ~98% of the HUVECs remained viable, as determined by trypan blue exclusion assay, indicating that the compound was not cytotoxic.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   A, ATL-1 inhibits VEGF-stimulated HUVEC proliferation. HUVECs (5 × 103) were plated in 96-well culture plates and cell proliferation was stimulated with 3 ng/ml VEGF. Three days after treatment, cell numbers were measured using MTT assay. Results are expressed as percentage of inhibition of proliferation relative to vehicle and represent mean ± S.E. for four independent experiments performed in triplicate. B, representative experiment showing the time course of cell proliferation induced by 3 ng/ml () or 10 ng/ml () VEGF. Vehicle () and ATL-1 (100 nM) ().

In addition to being an endothelial cell-specific mitogen, VEGF is also an endothelial cell survival factor, thus promoting angiogenesis not only by stimulating cell proliferation but also by inhibiting endothelial cell apoptosis (Gerber et al., 1998). To determine whether this new inhibitory action of ATL-1 on HUVEC proliferation involved apoptosis, we quantitated DNA fragments (under Materials and Methods). Neither ATL-1 (100 nM) alone nor in combination with VEGF (3 ng/ml) affected DNA fragmentation pattern (n = 2, d = 3), suggesting that the antiproliferative actions of the ATL-1 analog were not a result of induction of apoptosis in endothelial cells (data not shown). Binding of v integrins by endothelial cells is accompanied by a decrease in the tumor-suppressor p53 activity and inhibition of apoptotic pathways, thereby facilitating the formation of new blood vessels (Eliceiri and Cheresh, 1999). Along these lines, Kelavkar and Badr (1999) recently showed that p53 can up-regulate human 15-lipoxygenase promoter activity, providing the first link between this enzyme's activity and an established tumor-suppressor gene. Also of interest, over-expression of 15-lipoxygenase enhances endogenous LXA₄ formation that in turn inhibits progression of glomerulonephritis (Munger et al., 1999). Both ATL and LXA₄ share sites of action that are receptor- (Gronert et al., 2001), cell type-, and tissue-specific (Sodin-Semrl et al., 2000), and together with their stable analogs display potent anti-inflammatory actions in vivo (Serhan, 1997).

Endothelial cell migration is an essential component of the angiogenic process, providing directionality for the budding capillary toward the angiogenic stimulus (Arenberg and Strieter, 1999). Therefore, we next assessed endothelial migration with ATL. VEGF (3 ng/ml) was added to the lower compartment of a chemotaxis chamber and cell migration across a 10-µm pore size, gelatin-coated filter was quantitated (Fig. 2). Results in Fig. 2B show that ATL-1 gave a concentration-dependent inhibition of VEGF-stimulated HUVEC migration with a maximum level of inhibition (~45%) at 10 nM ATL. As observed with proliferation assays, ATL-1 alone, even at higher concentrations (100 nM), did not induce endothelial cell migration (Fig. 2A), findings that suggest that ATL could play a role in blocking the early stages of cell migration to sites of neovascularization. Like VEGF (Fig. 1), LTD₄ also stimulated proliferation of HUVEC (42 ± 1.2%) with a maximum at 10 nM, similar to the response obtained with VEGF (52.7 ± 1.6%). In contrast, LTB₄ at 10 nM did not give a significant response with these cells (Fig. 3B). The mitogenic action of LTD₄ (10 nM) was antagonized by exposure of the cells to ATL-1 (0.1-100 nM) with an IC₅₀ of ~3 nM (Fig. 3A). LTD₄ did not enhance or inhibit the VEGF-stimulated proliferation of HUVECs (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   ATL-1 inhibits endothelial cell chemotaxis. A, chemotaxis was initiated by addition of VEGF (3 ng/ml) or ATL-1 (100 nM) to the lower compartment of a 48-well chemotaxis chamber. Results are expressed as percentage of cell migration compared with vehicle alone and represent mean ± S.E. for three independent experiments performed in triplicate (P < 0.05). B, HUVECs were incubated with vehicle or indicated concentrations of ATL-1 (15 min, 37°C) and added to the upper compartment of the microchamber (1 × 106 cells/well). Chemotaxis was initiated by addition of VEGF (3 ng/ml) to the lower compartment. Results are expressed as percentage of inhibition of migration relative to VEGF for a representative experiment performed in triplicate.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   A, ATL-1 inhibits LTD4-stimulated HUVEC proliferation. HUVECs (5 × 103) were plated in 96-well culture plates, cell proliferation was stimulated with 10 nM LTD4, and cell numbers were determined after 3 days using MTT. B, concentration-dependent cell proliferation induced by LTD4 and LTB4. Results are expressed as mean ± S.E. for four independent experiments performed in triplicate.

The antiproliferative actions of native lipoxins were first found with the human lung adenocarcinoma cell line (Clària et al., 1996) and recently with human renal mesangial cells (McMahon et al., 2000). The present findings, together with our endothelial cell results, draw attention to the potential regulatory role for endogenous ATL in proliferative diseases. The actions of LX, ATL, and stable analogs are transduced by a high-affinity transmembrane receptor (ALXR) identified in several cell types (for review, see Chiang et al., 2000). In mesangial cells, LXA₄ interacts with its own high-affinity receptor (i.e., ALXR) as well as with a subclass of peptido-leukotriene receptors (cysLT₁), where LXA₄ is a partial agonist (McMahon et al., 2000). In this regard, LXA₄ and its bioactive stable analogs effectively displace [³H]LTD₄ specific binding to vascular endothelial cells (Takano et al., 1997). Also, recent findings provide the first evidence that ATL specifically antagonizes LTD₄ specific binding at recombinant human cysLT₁ cloned from endothelial cells, as well as acts at specific LXA₄ receptors (Gronert et al., 2001). Because ATL-1 proved to be a potent inhibitor of HUVEC proliferation (Fig. 3), we tested whether ATL affected angiogenesis in vivo.

During chronic inflammation, new vessels are required not only for the maintenance of tissue perfusion but also to allow increased cellular traffic (Cotran et al., 1999). Therefore, to this end we assessed angiogenesis in vivo by using a well established murine chronic granulomatous air pouch model, and the 6-day time interval was selected because it was shown to give near maximal vascular density (Colville-Nash et al., 1995). ATL-1 injected locally (10 µg/pouch) immediately before the administration of VEGF (1 µg/pouch) gave an ~48% reduction in the vascular index (Fig. 4A). For comparison, ATL-1 (10 µg/mouse or 0.4 mg/kg mouse) proved to be more potent than other described antiangiogenic agents that required much higher doses, including steroids at 1 to 3 mg/kg (Colville-Nash et al., 1995), or the COX-2 inhibitors, which require 1 to 6 mg/kg (Masferrer et al., 1999).

View larger version (57K): [in this window] [in a new window]   Fig. 4.   ATL-1 inhibits angiogenic phenotype in vivo. A, vascular index (micrograms of carmine dye per milligram weight of tissue) in day 6 murine air pouch. Animals received a local injection of ATL-1 (10 µg/pouch) or vehicle immediately before VEGF (1 µg/pouch), 24 h after raising the pouch. Results are expressed as the mean ± S.E. for n = 4 animals/group. , denotes statistically significant difference (P < 0.05) from VEGF alone. B, cedar wood oil histology of air pouch. Carmine dye vascular casts were made in day 6 air pouch from mice treated locally with vehicle, ATL-1 (10 µg), VEGF (1 µg), or VEGF plus ATL-1. Tissue was fixed in ethanol and cleared in cedarwood oil.

Figure 4B shows representative vascular casts from typical day 6 air pouch linings. In the mice given VEGF (Fig. 4B, bottom left), there is an established neovasculature with an extremely high degree of vascular density compared with only slightly dilated capillaries in the ATL-treated animals, where there was routinely clearly reduced vascular density (Fig. 4B, bottom right). In another set of experiments, fluorescein isothiocyanate-dextran was used to visualize the vessels in this region. In sharp contrast to the actions of ATL-1, when LTB₄, another lipoxygenase pathway product, was given alone at the same dose as ATL-1 (10 µg LTB₄/pouch), LTB₄ stimulated neovascularization (n = 2; data not shown). Figure 5 shows photomicrographs of the dorsal linings dissected at day 6. Again, profound angiogenesis was demonstrated with extensive vascular networks in VEGF-treated pouch (Fig. 5, bottom left). Here, too, treatment with ATL-1 (10 µg/pouch) gave striking reduction of VEGF-elicited vasculature, as exemplified by the lack of visible fine capillaries (Fig. 5, bottom right).

View larger version (152K): [in this window] [in a new window]   Fig. 5.   Antiangiogenic action of ATL-1: fluorescence microscopy. Representative fluorescence photomicrographs showing the antiangiogenic action of ATL-1 (10 µg/pouch) in the murine air pouch (under Materials and Methods).

It is important to note that ATL-1 at this dose (10 µg/mouse i.v.) does not evoke apparent changes in mean arterial pressure (Clish et al., 1999, supplemental material), excluding a possible action of ATL-1 at the level of vascular tone. This is particularly noteworthy because, at high doses, LXA₄ can stimulate vasodilation in certain vascular beds (for review, see Serhan, 1997). The present in vivo experiments were performed in a separate series to evaluate histology and the presence of a vascular marker by using immunohistochemical staining of the murine air pouch with the vascular endothelial cell marker CD31 (Fig. 6). Platelet/endothelial cell adhesion molecule-1 (or CD31) is a member of the Ig superfamily that is strongly expressed at the endothelial cell-cell junction, is present on platelets as well as leukocytes, and is held to play a role in angiogenesis and in transendothelial migration of leukocytes (for recent review, see Duncan et al., 1999; Muller and Randolph, 1999). Immunohistochemical staining for CD31 in the pouches showed that, in the VEGF-treated mice, strong specific endothelial cell staining was present and identified a prominent vascular network (Fig. 6C). In contrast, a marked diminution of vessels was observed in VEGF-treated mice that were also treated with the LXA₄ analog (Fig. 6D). The levels of mild nonspecific staining associated with these air pouches were essentially identical to those of the air pouch sections from mice treated with either vehicle alone (Fig. 6A) or with LX analog alone (Fig. 6B), namely, mild nonspecific staining of inflammatory cells, predominantly leukocytes and macrophages, that are known to be associated with these air pouches created from murine skin (Colville-Nash et al., 1995). Taken together, these findings indicate that ATL reduces VEGF-stimulated angiogenesis in vivo, suggesting that LXA₄ and 15-epi-LXA₄ can regulate these actions in vivo.

View larger version (122K): [in this window] [in a new window]   Fig. 6.   Immunohistochemistry for murine air pouch CD31. Paraffin-embedded air pouch sections were stained for CD31 from mice as in Figs. 4 and 5 and treated with vehicle alone (A), analog alone (B), VEGF-treated mice (C); and VEGF plus ATL-treated mice (D). Results are representative of eight separate mice each in duplicate. Magnification is 200× power and horseradish peroxidase with hematoxylin counterstain.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Results from many clinical and laboratory studies have demonstrated protective effects of aspirin in several forms of human cancer, including lung, colon, and breast cancer, yet its potential anticancer mechanism is not clear (Marcus, 1995). ASA is thought to act, in part, via reduction of angiogenesis, which might be related to the ability of ASA to inhibit prostanoid biosynthesis (Hla et al., 1993). More recently, ASA was found to trigger a novel switch in eicosanoid biosynthesis, because the acetylation of COX-2 enables the enzyme to produce 15R-HETE that is converted to 15-epi-lipoxins, also known as ATL, during cell-cell interactions in vitro and in vivo (Clària and Serhan, 1995; Serhan, 1997; Chiang et al., 1998). ATL as well as their stable bioactive analogs are potent inhibitors of several key events in acute inflammation, such as neutrophil chemotaxis and transmigration across both endothelial and epithelial cells, as well as diapedesis from postcapillary venules (Serhan, 1997; Clish et al., 1999). The analogs of ATL mimic both endogenous ATL and LX actions and were designed to resist rapid enzymatic inactivation in vivo. Bioactive analogs of 15-epi-LXA₄ were also found to complete at both the ALXR on leukocytes and the cysLT₁ receptor present on vascular endothelial cells (Chiang et al., 2000; Gronert et al., 2001). These novel aspirin-triggered mediators inhibit cytokine release and can act at the gene transcription level (Gewirtz et al., 1998) to redirect the local cytokine-chemokine axis (Hachicha et al., 1999), actions that are of interest in the angiogenic process (Arenberg and Strieter, 1999). It should be noted that most if not all other eicosanoids examined to date (Stoltz et al., 1996; Masferrer et al., 1999; Nie et al., 2000) are proangiogenic, including leukotrienes (e.g., LTD₄, LTB₄), as found in the present experiments (Fig. 3; under Results).

In summary, our present results demonstrate that an aspirin-triggered-lipoxin, 15-epi-LXA₄ analog, is a potent inhibitor of endothelial cell responses in vitro and in vivo. Together, these results reveal a novel action of 15-epi-lipoxins and suggest a role for the aspirin-triggered lipoxin circuit (Serhan, 1997) as a potential mechanism that may contribute to aspirin's recognized antiangiogenic and anti-inflammatory properties (Hla et al., 1993; Folkman, 1995; Vane, 2000). With increasing insight into the fundamental role of angiogenesis within a broad range of physiological as well as disease processes (Folkman and Shing, 1992; Folkman, 1995; Arenberg and Strieter, 1999), the modulation of vascular growth could be a previously unappreciated and important strategic action(s) for ATL generated locally within the microenvironment, the natural endogenous lipoxin mimetic, and their longer acting synthetic analogs.