During lymphocyte migration, engagement of VCAM-1 stimulates the generation of endothelial cell-derived reactive oxygen species (ROS) and activation of matrix metalloproteinases, facilitating endothelial retraction. Because bilirubin is a potent antioxidant, we examined the hypothesis that this bile pigment inhibits VCAM-1-dependent cellular events. The migration of isolated murine splenic lymphocytes across monolayers of murine endothelial cell lines (which constitutively express VCAM-1) is significantly inhibited by physiological concentrations of bilirubin, in the absence of an effect on lymphocyte adhesion. Bilirubin administration also suppresses VCAM-1-stimulated ROS generation and reduces endothelial cell matrix metalloproteinase activity. In a murine asthma model characterized by VCAM-1-dependent airway inflammation, treatment of C57BL6/J mice with i.p. bilirubin decreases the total leukocyte count in the lung parenchyma and lavage fluid, through specific inhibition of eosinophil and lymphocyte infiltration. Blood eosinophil counts were increased in bilirubin-treated animals, while VCAM-1 expression in the capillary endothelium and cytokine levels in both lung lavage and supernatants from cultured lymph node lymphocytes were unchanged, suggesting that bilirubin inhibits leukocyte migration. Conclusion: bilirubin blocks VCAM-1-dependent lymphocyte migration in vitro and ameliorates VCAM-1-mediated airway inflammation in vivo, apparently through the suppression of cellular ROS production. These findings support a potential role for bilirubin as an endogenous immunomodulatory agent.

Adhesion molecules are transmembrane glycoproteins that mediate an array of cellular interactions and play an essential role in the process of inflammatory cell migration (1, 2). One such molecule, VCAM-1, is expressed on the surface of endothelial cells and promotes the binding and movement of lymphocytes and eosinophils across vascular endothelium (1, 2, 3). VCAM-1-mediated leukocyte recruitment and migration are implicated in the pathogenesis of several disease states, including inflammatory bowel disease (4), conjunctivitis (5), nephropathy (6, 7), and arthritis (8). The essential role of VCAM-1 in tissue injury has been corroborated in a number of animal models. For example, radiation-induced intestinal inflammation, which is characterized by increased endothelial VCAM-1 expression, is attenuated by treatment with anti-VCAM-1 Abs (9). Immunoneutralization of VCAM-1 has also been shown to ameliorate hepatic necrosis and neutrophil infiltration in response to endotoxin (4). Indeed, VCAM-1 appears to be integral to leukocyte recruitment in a variety of tissues, including the brain (10, 11), kidney (12), and intestine (13).

The molecular mechanisms involved in VCAM-1 signaling have recently begun to be elucidated. The binding of lymphocyte VLA-4 (α4β1 integrin) to endothelial cell VCAM-1 triggers the calcium- and Rac1-dependent activation of cellular NADPH oxidase (14), which generates low (micromolar) levels of reactive oxygen species (ROS)3 (3). This is in contrast with signaling by other endothelial adhesion molecules, such as selectins (15) and PECAM (16), which does not involve ROS production. The localized increase in ROS in response to VCAM-1 stimulation induces the activity of matrix metalloproteinase (MMP)-2 and -9. MMP-mediated digestion of the extracellular matrix and intercellular tight junction complexes facilitates transendothelial lymphocyte migration (17, 18). ROS also have been implicated in the induction of the morphologic changes necessary for cellular retraction, through modulation of the actin cytoskeleton (17). Because the generation of ROS appears to be critical to VCAM-1 signaling, oxygen radical scavengers would be expected to block this process. In this regard, antioxidants have been shown to inhibit leukocyte infiltration in murine nephritis (19) and asthma (20) models, and to decrease cytokine production (21) by cultured endothelial cells.

Bilirubin, one of the principal metabolic end-products of heme catabolism and a potent antioxidant (22), is generated by the sequential action of heme oxygenase (HO) and biliverdin reductase (BVR) (Fig. 1). There is a growing body of evidence supporting that bilirubin is cytoprotective, as this bile pigment has been shown to suppress postischemic myocardial dysfunction (23), protect against LPS-induced liver injury (24), and ameliorate local hindpaw inflammation induced by the injection of λ-carrageenan. Because bilirubin is the most abundant endogenous antioxidant (22), with a potency comparable to α-tocopherol (25), we speculate that bilirubin may exert its cytoprotective effects by disrupting VCAM-1-mediated signaling. In support of this hypothesis, bilirubin has been shown to block oxidant-mediated leukocyte adhesion (26) and prevent oxidative injury to vascular smooth muscle cells (27). To determine whether bilirubin acts via a VCAM-1-dependent mechanism, we examined the effect of bilirubin on the migration of lymphocytes across isolated endothelial cell monolayers and on eosinophil migration across the pulmonary vascular endothelium in OVA-primed mice. Bilirubin administration inhibited the transendothelial migration of leukocytes in both of these VCAM-1-dependent model systems, supporting a potentially important regulatory function of this endogenous bile pigment on inflammatory responses.

FIGURE 1.

Pathway of bilirubin metabolism. Heme is serially catabolized by the action of HO and BVR to form biliverdin and bilirubin, respectively. In the liver, bilirubin is conjugated with glucuronic acid by the action of UGT1A1 before secretion in the bile.

FIGURE 1.

Pathway of bilirubin metabolism. Heme is serially catabolized by the action of HO and BVR to form biliverdin and bilirubin, respectively. In the liver, bilirubin is conjugated with glucuronic acid by the action of UGT1A1 before secretion in the bile.

Close modal

Unconjugated bilirubin (bilirubin IXα), biliverdin (biliverdin IXα), and ditaurobilirubin (BDT) were obtained from Porphyrin Products. Bilirubin IXα was further purified according to the method of McDonagh and Assisi (28) to eliminate potential lipid contaminants. Rat anti-mouse VCAM-1 was obtained from BD Pharmingen. Rat anti-mouse major basic protein-1 (MBP-1) was kindly provided by Drs. J. and N. Lee (Mayo Clinic, Scottsdale, AZ). Rat anti-mouse α4 integrin (VLA-4) was purchased from BioDesign International. Dihydrorhodamine 123 (DHR) was obtained from Molecular Probes.

Murine high endothelial venule-like cells, derived from BALB/c mouse axillary murine high endothelial venule (mHEVa) or cervical mHEV (mHEVc) lymph nodes (29), were grown to confluence in the upper chamber of 12-μm transwells (Costar) in RPMI 1640 supplemented with 20% FCS, 2 mM glutamine, 1 mM HEPES, 10 mM sodium bicarbonate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μg/ml gentamicin. Monolayers were preincubated for 1 h in the presence of bilirubin, biliverdin, BDT, or the corresponding vehicle, which were simultaneously added to both the upper and lower chambers. Bile pigments were solubilized in 50 mM potassium phosphate (pH 12) immediately before each experiment (30). The addition of a small aliquot (0.4% v/v) of this stock solution had no effect on the pH of the culture medium. To ensure adequate bilirubin solubilization, all experiments were conducted in the presence of 20% FCS (30). Migration was initiated by the addition of murine splenic lymphocytes (∼4 × 106 cells) and erythrocytes (∼4 × 106 cells; nonmigratory control), to the upper chamber of the transwell (29). Chemotaxis is induced by the constitutive expression of MCP-1 by the endothelial cell lines (31). Bile pigments and vehicle were present in both the upper and lower chambers throughout the duration of the experiment. Monolayer integrity was confirmed by the absence of erythrocytes in the lower chamber of the transwell, as these cells are nonmigratory (14). In the event that erythrocytes were detected in the lower chamber, data from that transwell were discarded, a situation that occurred only rarely and was unrelated to treatment. Transendothelial migration was assessed by counting the number of viable lymphocytes present in the lower chamber at 24 h, with data normalized relative to lymphocyte migration across untreated mHEV monolayers. Cell viability was determined by trypan blue exclusion (3).

For Northern blotting, total cellular RNA was extracted and probed with cDNA encoding murine HO-1 (24). RT-PCR was performed with primers specific for murine UDP-glucuronosyltransferase (UGT1A1) (5′-ctt gaa atg act gct gat ga-3′ and 5′-ttt ctt cac tcg ccc ctt tc-3′) (32) and BVR (5′-ggg acg ctg caa ttt ggg c-3′ and 5′-gcg aca tca acc tct tgg c-3′), with GAPDH primers (5′-tga agg tcg gtg tga acg cat ttg gc-3′ and 5′-cat gta ggc cat gag gtc cac cac-3′) used to control for amplification.

Endothelial cells (mHEVa and mHEVc) were grown to confluence in 96-well tissue culture plates (Fisher Scientific). Monolayers were incubated for 24 h in presence of bilirubin, biliverdin, BDT, or vehicle, after which freshly isolated mouse splenic lymphocytes (∼1 × 106 cells) labeled with calcein acetoxymethyl ester (AM) (1 μM) were added (3). Following a 30-min incubation at 37°C, nonadherent lymphocytes were removed by gently vortexing and aspirating the supernatant (3). Monolayers were then overlaid with 100 μl of PBS containing 0.2 mM CaCl2 and 0.1 mM MgCl2. Calcein fluorescence intensity (excitation: 485 nm; emission: 530 nm), indicative of lymphocyte binding, was measured on a 7620 Microplate Fluorometer (Cambridge Technology International). As a control for binding specificity, monolayers were incubated with blocking anti-VCAM-1 (20 μg/ml) or anti-VLA4 Abs (20 μg/ml) immediately before the addition of lymphocytes. These Ab concentrations have been demonstrated to maximally inhibit lymphocyte adhesion to mHEV monolayers in prior studies (33). Isotype control Abs have previously been shown to have no effect on lymphocyte adhesion to mHEV monolayers (14). Additionally, background adhesion, as determined by lymphocyte binding to a mouse melanoma cell line (B78H1), was found to be negligible (data not shown).

Endothelial cells were grown to confluence on 22-mm glass coverslips. Monolayers were pretreated with bilirubin (or vehicle) for 1 h at 37°C in phenol red-free medium, and 1 μM DHR (Molecular Probes) was added 15 min before stimulation with anti-VCAM-1-coated beads (3). ROS generation results in the oxidation of the nonfluorescent DHR to rhodamine 123, which manifests as an increase in fluorescence intensity (ex: 488 nm, em: 525 nm). Cellular fluorescence was measured at 0 min and 30 min after VCAM-1 stimulation using a Leica TCS4D confocal microscope equipped with an Omnichrome krypton-argon laser. These time points were selected because it has previously been shown that maximal ROS generation occurs at 30 min (3). The field of observation was changed after each reading to avoid laser-induced photoactivation. Data are presented as the change in fluorescence intensity from 0 min and are corrected for background by subtracting the fluorescence intensity of identically treated cells that were not loaded with DHR.

Confluent endothelial cell monolayers were incubated with anti-VCAM-1 Ab-coated beads in the presence of bilirubin or the potassium phosphate vehicle. At specified time intervals, cells were lysed with 120 mM Tris buffer containing 0.25% Triton X-100, 0.001g NaN3 (pH 8.7). Lysates were subjected to overnight electrophoresis (70V, 25 mA, 4°C) on a 7.5% polyacrylamide gel impregnated with gelatin (0.2 mg/ml). The gel was subsequently incubated for 1 h at 25°C in 2.5% Triton X-100 and washed twice for 20 min in distilled deionized water. MMP activity was visualized by incubating the gel for 24 h at 37°C in 0.05 M Tris-HCl, 5 mM CaCl2 (pH 8.0). Control gels were developed in the presence of EDTA (0.05 M Tris-HCl, 0.01 M EDTA (pH 8.0)) to demonstrate cation dependency. Gels were fixed and stained in 50% methanol, 10% acetic acid, and 0.25% Coomassie blue R250. MMP activity appears as clear bands, which were quantified by densitometry using Image J software (National Institutes of Health, 〈http://rsb.info.nih.gov/ij/〉). MMP isoforms were identified on the basis of molecular mobility and by comparison with purified standards (17).

Confluent endothelial cell monolayers treated with bilirubin, biliverdin, BDT, or vehicle were harvested with PBS containing 0.03% EDTA, washed with PBS, and labeled with rat anti-mouse VCAM-1 (1/100 dilution) in PBS containing 0.3% BSA and 0.15% NaN3 for 30 min at 4°C. The cells were labeled with goat anti-rat FITC-conjugated IgG (1/500 dilution; Southern Biotechnology Associates), washed, and fluorescence intensity quantified by flow cytometry (Coulter EPICS-XL). Unlabeled cells, which were incubated with secondary Ab alone, were used as a control for staining, and are comparable to isotype Ab-labeled cells (data not shown).

Adult female C57BL6/J mice were sensitized with two serial i.p. injections of OVA (200 μl of 200-μg grade V OVA adsorbed to 180 mg of aluminum hydroxide in 4 ml of saline for 1 h at room temperature) 1 wk apart (days 0 and 7) (34, 35) (Fig. 2). Control animals received an equivalent volume of aluminum hydroxide in saline. Mice were challenged with 150 μg of OVA in saline intranasally on day 14, with control animals receiving intranasal saline. Rodents in each group were administered six doses of i.p. bilirubin (30 mg/kg) or the potassium phosphate vehicle at 8-h intervals (days 14 and 15), commencing immediately after OVA challenge. Animals were sacrificed by CO2 inhalation and the lungs were rapidly lavaged by injecting 0.5 ml of ice-cold PBS into the trachea using a 16-gauge needle. The lavage fluid was aspirated and the procedure repeated an additional time. Differential leukocyte counts were performed on H&E-stained cytospins, with a total of 200 cells counted from each of the pooled lavage samples. Frozen sections of mouse lung were stained with H&E to assess lung histology. Studies were reviewed and approved by the University of Cincinnati Institutional Animal Care and Use Committee.

FIGURE 2.

Schema for the induction of airway inflammation in mice. Adult C57BL6/J mice were sensitized with i.p. OVA or saline vehicle on days 0 and 7, and challenged with intranasal OVA (or saline) on day 14. Mice subsequently received six doses of i.p. bilirubin (30 mg/kg) or potassium phosphate vehicle at 8-hourly intervals on days 14 and 15.

FIGURE 2.

Schema for the induction of airway inflammation in mice. Adult C57BL6/J mice were sensitized with i.p. OVA or saline vehicle on days 0 and 7, and challenged with intranasal OVA (or saline) on day 14. Mice subsequently received six doses of i.p. bilirubin (30 mg/kg) or potassium phosphate vehicle at 8-hourly intervals on days 14 and 15.

Close modal

Lymphocytes were obtained by gently pressing isolated bronchial lymph nodes through a 100-μm nylon mesh (3). Cells were counted, aliquoted to 96-well plates at a concentration of 5 × 106 per ml, and grown in RPMI 1640, 1 mM HEPES, 10 mM NaHCO3, 2 mM glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamicin, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 50 μM 2-ME, and 5% heat-inactivated FCS (pH 7.2) for 24 h. Samples obtained from each of the two saline control groups were pooled due to low cell yields. Cultured lymphocytes were stimulated with 500 μg/ml OVA for 48 h at 37°C, and the medium harvested and stored at −80°C before analysis.

Cytokine (IL-2, IFN-γ, IL-4, IL-5, IL-6, IL-10, IL-12, TNF-α, and MCP-1) levels in lung lavage fluid and in lymphocyte supernatants were measured using Cytometric Bead Array kits (BD Biosciences), according to the manufacturer’s instructions. Briefly, samples were incubated with cytokine-specific Ab-coated beads for 2 h, washed, and analyzed by flow cytometry. Concentrations were determined from generated standard curves. Eotaxin levels were detected colorimetrically using the Mouse Eotaxin Immunoassay Kit (BioSource International), with data corrected for bilirubin absorbance, as previously described (30).

Frozen tissue sections were prepared and incubated in the presence of the appropriate primary Ab (rat anti-mouse VCAM-1 (1/200); rat anti-mouse MBP-1 (1/500)) or the isotype control Ab (rat anti-mouse Ig; 1/500) at room temperature for 1 h. For detection of VCAM-1, tissue sections were labeled with a FITC-conjugated secondary Ab (goat anti-rat IgG; 1/200) and analyzed by confocal microscopy (ex: 488 nm; em: 530 nm). Lung tissue sections were labeled with anti-MBP-1, a marker specific for eosinophils (36) and stained using an HRP-conjugated secondary Ab (goat anti-rat IgG; 1/500), followed by diaminobenzidine substrate and counterstaining with methyl green. The number of MBP-1-positive cells in 20 separate high-power fields was determined for each specimen.

Data were analyzed using a computer-based statistical package (Statistix 7; Analytical Software) with differences between mean values assessed by ANOVA with Scheffe comparison (30).

To determine the effect of bilirubin on VCAM-1-mediated transendothelial lymphocyte migration, we used two separate endothelial cell lines, mHEVa (derived from murine axillary lymph nodes) and mHEVc (derived from murine cervical lymph nodes), which constitutively express VCAM-1. These cells have been shown to bind resting lymphocytes through VCAM-1 α4 integrin interactions (29) and to promote transmonolayer migration. Because the viability of cultured mHEVa and mHEVc cells and isolated splenic lymphocytes (as determined by trypan blue exclusion) was unaffected by bilirubin levels up to 20 μM, concentrations within this range were used in our experiments. The effect of bilirubin, biliverdin, and BDT on the migration of freshly isolated splenic lymphocytes across confluent endothelial cell monolayers is depicted in Fig. 3. Treatment with either bilirubin or biliverdin was associated with a significant reduction in the number of lymphocytes in the lower chamber of the transwell at 24 h, while the bilirubin conjugate, BDT, had no effect on the transendothelial lymphocyte migration. Monolayer integrity, evidenced by the absence of erythrocyte translocation from the upper to the lower chamber of the transwell, was not altered by any of the treatments. These data suggest that physiologic concentrations of bilirubin, and its metabolic precursor, biliverdin, inhibit the migration of lymphocytes across endothelial cell monolayers. That this inhibitory effect is exerted at the level of the endothelial cell is supported by the finding that preincubation of lymphocytes for 24 h in the presence of bilirubin (20 μM), biliverdin (20 μM), or BDT (20 μM) had no effect on migration when treatments were removed before the addition of these pretreated cells to the transwell (data not shown).

FIGURE 3.

Bilirubin and biliverdin inhibit lymphocyte migration across endothelial cell monolayers. Confluent monolayers of mHEVc (left panel) and mHEVa (right panel) cells were preincubated for 1 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle, at the indicated concentrations. The number of lymphocytes in the lower chamber of the transwell at 24 h is plotted relative to untreated monolayers. Bars reflect the mean (± SEM) of four separate experiments, with three to four replicates performed per experiment. ∗, p < 0.05 vs vehicle and untreated controls.

FIGURE 3.

Bilirubin and biliverdin inhibit lymphocyte migration across endothelial cell monolayers. Confluent monolayers of mHEVc (left panel) and mHEVa (right panel) cells were preincubated for 1 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle, at the indicated concentrations. The number of lymphocytes in the lower chamber of the transwell at 24 h is plotted relative to untreated monolayers. Bars reflect the mean (± SEM) of four separate experiments, with three to four replicates performed per experiment. ∗, p < 0.05 vs vehicle and untreated controls.

Close modal

As outlined in Fig. 2, the sequential activity of HO and BVR is required for the generation of bilirubin from heme, while the bilirubin-specific isoform UGT1A1, catalyzes the main pathway of bilirubin degradation. If bilirubin is an important modulator of inflammation, it follows that the expression of these enzymes in endothelial cells will be tightly regulated. As bilirubin has been shown to stimulate hepatic expression of the bilirubin-conjugating enzyme, UGT1A1 (37, 38), we sought to determine whether bilirubin, or its analogues, modulates endothelial cell expression of the enzymes involved in bilirubin production and metabolism. Northern blotting demonstrated expression of HO-1 mRNA in mHEVc cells, while message for BVR and UGT1A1 was detected by RT-PCR (Fig. 4), supporting the ability of this cell line to generate and metabolize bilirubin. Similar results were obtained with the mHEVa cells (data not shown). The presence of message for BVR, the enzyme responsible for the conversion of biliverdin to bilirubin, also indicates that mHEV cells possess the enzymatic machinery necessary for the redox cycling of bilirubin. None of the bile pigments tested were found to alter cellular expression of the enzymes regulating bilirubin metabolism.

FIGURE 4.

Expression of enzymes involved in bilirubin metabolism in endothelial cells. Monolayers of mHEVc cells were incubated in the presence or absence (N) of bilirubin (BR), biliverdin (BV), BDT, or the potassium phosphate vehicle (V) for 24 h and total cellular RNA extracted. Expression of HO-1 was assessed by Northern blotting (left panel), with SYBr Green II stained ribosomal RNA serving as a loading control. Message for BVR and UGT1A1 was detected by RT-PCR (right panel), with GAPDH serving as a control for amplification. Data are representative of two separate experiments. Total RNA isolated from mouse macrophage-like (RAW 264.7) cells incubated for 24 h in the presence of LPS was used as a hybridization control (24 ).

FIGURE 4.

Expression of enzymes involved in bilirubin metabolism in endothelial cells. Monolayers of mHEVc cells were incubated in the presence or absence (N) of bilirubin (BR), biliverdin (BV), BDT, or the potassium phosphate vehicle (V) for 24 h and total cellular RNA extracted. Expression of HO-1 was assessed by Northern blotting (left panel), with SYBr Green II stained ribosomal RNA serving as a loading control. Message for BVR and UGT1A1 was detected by RT-PCR (right panel), with GAPDH serving as a control for amplification. Data are representative of two separate experiments. Total RNA isolated from mouse macrophage-like (RAW 264.7) cells incubated for 24 h in the presence of LPS was used as a hybridization control (24 ).

Close modal

Because migration across mHEV monolayers is predicated upon the binding of endothelial cell VCAM-1 to leukocyte VLA-4, we assessed whether bilirubin disrupts the interaction between these adhesion molecules. For these studies, the adhesion of calcein AM-labeled splenic lymphocytes to cultured mHEV cells was assessed in the presence or absence of bilirubin, biliverdin, or BDT (Fig. 5). None of these bile pigments was found to alter lymphocyte binding to mHEV monolayers, while in contrast, Abs to VCAM-1 and VLA-4 significantly reduced the number of bound lymphocytes. Similarly, neither bilirubin, biliverdin, nor BDT treatment affected the expression of VCAM-1 by mHEV cells (Fig. 6). Taken together, these data support that bilirubin and its analogues do not block lymphocyte migration by disrupting the cellular expression or interaction of adhesion molecules.

FIGURE 5.

Bilirubin does not alter lymphocyte adhesion to mHEV cells. Confluent mHEVc (left panel) and mHEVa (right panel) monolayers were treated for 24 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle. Endothelial monolayers pretreated with Abs to VLA-4 and VCAM-1 serve as positive controls. The number of adherent calcein-AM-labeled lymphocytes (± SEM) is plotted relative to untreated monolayers. Data reflect the results of three separate experiments (three to four replicates per experiment). ∗, p < 0.05 vs all other treatments.

FIGURE 5.

Bilirubin does not alter lymphocyte adhesion to mHEV cells. Confluent mHEVc (left panel) and mHEVa (right panel) monolayers were treated for 24 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle. Endothelial monolayers pretreated with Abs to VLA-4 and VCAM-1 serve as positive controls. The number of adherent calcein-AM-labeled lymphocytes (± SEM) is plotted relative to untreated monolayers. Data reflect the results of three separate experiments (three to four replicates per experiment). ∗, p < 0.05 vs all other treatments.

Close modal
FIGURE 6.

Bilirubin does not affect VCAM-1 expression by mHEVc cells. Cell monolayers were incubated for 24 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle, and then subjected to immunofluorescence staining for VCAM-1. A, Representative histograms of VCAM-1 labeling in untreated and bilirubin-treated cells (lower panels). Identically treated cells that were not incubated with the primary Ab (unlabeled) serve as controls (upper panels). Mean fluorescence intensity (± SEM) is plotted relative to untreated cells in B.

FIGURE 6.

Bilirubin does not affect VCAM-1 expression by mHEVc cells. Cell monolayers were incubated for 24 h in the presence or absence of bilirubin, biliverdin, BDT, or the potassium phosphate vehicle, and then subjected to immunofluorescence staining for VCAM-1. A, Representative histograms of VCAM-1 labeling in untreated and bilirubin-treated cells (lower panels). Identically treated cells that were not incubated with the primary Ab (unlabeled) serve as controls (upper panels). Mean fluorescence intensity (± SEM) is plotted relative to untreated cells in B.

Close modal

It has previously been demonstrated that lymphocyte migration across mHEV monolayers requires the activation of endothelial NADPH oxidase and the subsequent generation of extracellular superoxide radicals (3). Because bilirubin is a potent antioxidant (22), we speculate that bilirubin inhibits VCAM-1-mediated endothelial cell signaling by scavenging ROS produced by the activation of NADPH oxidase. To determine whether bilirubin inhibits NADPH oxidase-mediated ROS production, mHEVc cells were treated with either bilirubin or vehicle and then labeled with the redox-sensitive dye, DHR. Following stimulation with anti-VCAM-1-coated microbeads, the ROS-mediated conversion of DHR to the highly fluorescent rhodamine 123 was assessed by confocal microscopy (Fig. 7). Bilirubin attenuates the enhancement in rhodamine fluorescence in response to VCAM-1 stimulation, supporting that bilirubin blocks VCAM-1-induced ROS generation. Because increases in NADPH oxidase activity lead to the specific activation of MMP-2 and -9 in mHEV cells, we sought to determine whether bilirubin treatment alters this downstream event in VCAM-1 signaling. Following pretreatment with bilirubin (or vehicle), mHEV monolayers were stimulated with anti-VCAM-1-coated beads, and the MMP activity in cell lysates was assessed by zymography. As shown in Fig. 8, treatment with bilirubin abolishes the time-dependent activation of MMP-2 and MMP-9, providing additional evidence that bilirubin impedes VCAM-1-mediated cellular events.

FIGURE 7.

Bilirubin inhibits the cellular production of ROS in response to VCAM-1 stimulation. Confluent monolayers of mHEVc cells were incubated in the presence of bilirubin (20 μM) or the potassium phosphate vehicle, and then loaded with DHR before stimulation with anti-VCAM-1-coated beads. A, Representative confocal images of anti-VCAM-1-treated mHEVc cells treated with vehicle (left panel) or bilirubin (right panel). B, The increase in rhodamine fluorescence over 30 min in unstimulated (Nonstim) and anti-VCAM-1-stimulated (VCAM) cells in the presence or absence of bilirubin (Bili) or vehicle (Veh). Data reflect the mean (± SEM) of two separate experiments (two replicates per experiment). ∗, p < 0.05 vs anti-VCAM-1-stimulated cells.

FIGURE 7.

Bilirubin inhibits the cellular production of ROS in response to VCAM-1 stimulation. Confluent monolayers of mHEVc cells were incubated in the presence of bilirubin (20 μM) or the potassium phosphate vehicle, and then loaded with DHR before stimulation with anti-VCAM-1-coated beads. A, Representative confocal images of anti-VCAM-1-treated mHEVc cells treated with vehicle (left panel) or bilirubin (right panel). B, The increase in rhodamine fluorescence over 30 min in unstimulated (Nonstim) and anti-VCAM-1-stimulated (VCAM) cells in the presence or absence of bilirubin (Bili) or vehicle (Veh). Data reflect the mean (± SEM) of two separate experiments (two replicates per experiment). ∗, p < 0.05 vs anti-VCAM-1-stimulated cells.

Close modal
FIGURE 8.

Bilirubin blocks VCAM-1-induced activation of MMP-2 and MMP-9. Monolayers of mHEVc cells were incubated with anti-VCAM-1 Ab-coated beads for 0–25 min, in the presence of bilirubin (20 μM) or the potassium phosphate vehicle. Cell lysates were subjected to SDS-PAGE on gelatin-impregnated gels, followed by zymography (upper panel). Bands corresponding to MMP-2 (55 kDa), MMP-9 (84 kDa), and pro-MMP-9 (97 kDa) are identified. The lower panels display densitometry analysis of MMP-2 (A) and MMP-9 (B) activity in the presence of bilirubin (▦) or vehicle (▪). Data are representative of two separate experiments.

FIGURE 8.

Bilirubin blocks VCAM-1-induced activation of MMP-2 and MMP-9. Monolayers of mHEVc cells were incubated with anti-VCAM-1 Ab-coated beads for 0–25 min, in the presence of bilirubin (20 μM) or the potassium phosphate vehicle. Cell lysates were subjected to SDS-PAGE on gelatin-impregnated gels, followed by zymography (upper panel). Bands corresponding to MMP-2 (55 kDa), MMP-9 (84 kDa), and pro-MMP-9 (97 kDa) are identified. The lower panels display densitometry analysis of MMP-2 (A) and MMP-9 (B) activity in the presence of bilirubin (▦) or vehicle (▪). Data are representative of two separate experiments.

Close modal

To evaluate whether bilirubin is able to inhibit VCAM-1-mediated leukocyte migration in vivo, we used a murine model of OVA-induced airway inflammation (34, 35). C57BL6/J mice were administered OVA via i.p. injection. Subsequent exposure of sensitized animals to intranasal OVA produces an asthma-like Th2 immune response. The resultant pulmonary eosinophilia and lymphocytosis has previously been demonstrated to be VCAM-1-dependent (34, 35). Following intranasal OVA challenge, mice were administered a total of six doses of bilirubin (30 mg/kg), or potassium phosphate vehicle, i.p. at 8-h intervals before sacrifice. As expected, OVA-challenged mice exhibit a marked infiltration of inflammatory cells into the lung parenchyma as compared with saline-treated control animals (Fig. 9,A, left panels). Notably, treatment with bilirubin was associated with a significant reduction in the number of inflammatory cells in the lung tissue, while bilirubin administration in the absence of OVA did not alter lung morphology. These findings were confirmed by measuring leukocyte accumulation in lung lavage fluid (Fig. 9,B), which was markedly increased following OVA stimulation. Concomitant bilirubin administration significantly blunted this leukocyte response. When stratified by cell type, bilirubin specifically inhibited the infiltration of eosinophils and lymphocytes into the airways (Fig. 9,C). These findings were corroborated by immunohistochemical staining of lung tissue sections for MBP-1, an eosinophil-specific marker (36), which demonstrates substantially lower numbers of eosinophils in lung tissue from bilirubin-treated vs vehicle-treated animals (Fig. 10). Because, in this model, both eosinophil and lymphocyte migration into the lungs have been shown to be VCAM-1-dependent (34, 35), our data are consistent with the hypothesis that bilirubin blocks pulmonary leukocyte infiltration by inhibiting VCAM-1 signaling.

FIGURE 9.

Bilirubin prevents leukocyte infiltration into the lungs of OVA-treated mice. A, Representative H&E-stained sections of mouse lung tissue. The mean leukocyte counts (± SEM) in lung lavage fluid of C57BL6/J mice primed and challenged with saline or OVA, and then treated with either bilirubin or the potassium phosphate vehicle, are depicted in B. C, Differential counts of lung lavage smears are displayed. Bilirubin treatment caused a significant reduction in OVA-induced infiltration of eosinophils (Eos) and lymphocytes (Lymphs) into the airways. No significant effect of bilirubin on monocyte or neutrophil (PMN) counts was observed. ∗, p < 0.001

FIGURE 9.

Bilirubin prevents leukocyte infiltration into the lungs of OVA-treated mice. A, Representative H&E-stained sections of mouse lung tissue. The mean leukocyte counts (± SEM) in lung lavage fluid of C57BL6/J mice primed and challenged with saline or OVA, and then treated with either bilirubin or the potassium phosphate vehicle, are depicted in B. C, Differential counts of lung lavage smears are displayed. Bilirubin treatment caused a significant reduction in OVA-induced infiltration of eosinophils (Eos) and lymphocytes (Lymphs) into the airways. No significant effect of bilirubin on monocyte or neutrophil (PMN) counts was observed. ∗, p < 0.001

Close modal
FIGURE 10.

Effect of bilirubin on OVA-induced pulmonary eosinophilia. A, Representative frozen sections of lung tissue from mice primed with saline or OVA and subsequently treated with bilirubin or vehicle. Sections are stained for MBP-1 to identify eosinophils (arrows). B, The average number of eosinophils (± SEM) per high power field (hpf). ∗, p < 0.001

FIGURE 10.

Effect of bilirubin on OVA-induced pulmonary eosinophilia. A, Representative frozen sections of lung tissue from mice primed with saline or OVA and subsequently treated with bilirubin or vehicle. Sections are stained for MBP-1 to identify eosinophils (arrows). B, The average number of eosinophils (± SEM) per high power field (hpf). ∗, p < 0.001

Close modal

An examination of peripheral blood smears from OVA- and vehicle-treated mice demonstrates that bilirubin administration increases the number of circulating eosinophils (Fig. 11), indicating that the inhibition of pulmonary eosinophilia by bilirubin is not the result of diminished eosinophil production or release from the bone marrow. Immunohistochemical staining of lung tissue sections further demonstrates that bilirubin treatment does not significantly alter VCAM-1 expression in the pulmonary vasculature, suggesting that the associated decrease in leukocyte infiltration into the airway is not the result of impaired cellular adhesion (Fig. 12). To assess for the possibility that the bilirubin-induced reduction in eosinophil accumulation is due to altered chemotactic signaling, we measured the concentration of eotaxin, a chemokine responsible for eosinophil migration (39), in lung lavage fluid. Consistent with the findings of other investigators (40), eotaxin levels increase following OVA challenge, but are unaffected by bilirubin administration (Fig. 13,A). Concentrations of IL-5, a Th2 cytokine that acts as an eosinophil chemoattractant (41), are also elevated in response to OVA challenge (42) and are also not altered by bilirubin treatment (Fig. 13,C). Indeed, bilirubin had no effect on any of the measured cytokine or chemokine levels in lung lavage fluid when compared with the corresponding vehicle-treated animals (Table I).

FIGURE 11.

Bilirubin treatment increases blood eosinophil levels in mice. The average number of eosinophils (± SEM) in the peripheral blood of mice primed with saline or OVA and treated with bilirubin or vehicle is displayed. ∗, p < 0.001 vs saline + vehicle and OVA + vehicle cohorts.

FIGURE 11.

Bilirubin treatment increases blood eosinophil levels in mice. The average number of eosinophils (± SEM) in the peripheral blood of mice primed with saline or OVA and treated with bilirubin or vehicle is displayed. ∗, p < 0.001 vs saline + vehicle and OVA + vehicle cohorts.

Close modal
FIGURE 12.

Bilirubin administration does not affect VCAM-1 expression in the lungs of OVA-treated mice. A, Representative confocal images of VCAM-1 stained frozen tissue sections from the lungs of saline- and OVA-primed mice treated with bilirubin or vehicle are displayed. B, The results of quantitative analyses (± SEM) of tissue immunofluorescence (▦). Isotype Ab-stained sections (▪) serve as controls for background fluorescence.

FIGURE 12.

Bilirubin administration does not affect VCAM-1 expression in the lungs of OVA-treated mice. A, Representative confocal images of VCAM-1 stained frozen tissue sections from the lungs of saline- and OVA-primed mice treated with bilirubin or vehicle are displayed. B, The results of quantitative analyses (± SEM) of tissue immunofluorescence (▦). Isotype Ab-stained sections (▪) serve as controls for background fluorescence.

Close modal
FIGURE 13.

Bilirubin treatment does not inhibit cytokine production. Levels of eotaxin (A and B), IL-5 (C and D), and IL-2 (E and F) in lung lavage fluid (A, C, and E), and in supernatants from OVA-stimulated lymphocytes isolated from bronchial lymph nodes (B, D, and F) are shown. ∗, p < 0.05 vs saline controls.

FIGURE 13.

Bilirubin treatment does not inhibit cytokine production. Levels of eotaxin (A and B), IL-5 (C and D), and IL-2 (E and F) in lung lavage fluid (A, C, and E), and in supernatants from OVA-stimulated lymphocytes isolated from bronchial lymph nodes (B, D, and F) are shown. ∗, p < 0.05 vs saline controls.

Close modal
Table I.

Effect of ovalbumin and bilirubin on cytokine secretion in lung lavage fluid and activated lymphocytes (lymph node) in picograms per milliliter (± SEM)

CytokinesLymph NodeLung Lavage
Saline + VehicleaOvalbumin + VehicleSaline + BilirubinaOvalbumin + BilirubinSaline + VehicleOvalbumin + VehicleSaline + BilirubinOvalbumin + Bilirubin
Th1         
 IL-2 6.2 4.2 (± 0.6) 7.9 5.6 (± 1.2) 5.9 (± 0.3) 6.6 (± 0.2) 6.8 (± 0.1) 6.6 (± 0.2) 
 IFN-γ 6.0 19.0 (± 8.3) 5.5 45.6 (± 13.8) 3.7 (± 0.7) 5.1 (± 0.8) 3.6 (± 1.7) 3.3 (± 0.4) 
Th2         
 IL-4 7.1 (± 1.9) 7.1 (± 1.9) 8.7 (± 0.2) 9.1 (± 0.2) 9.9 (± 0.9) 8.7 (± 0.2) 
 IL-5 2.6 191.4 (± 79.6) 301.1 (± 119.8) 10.4 (± 0.2) 15.7 (± 1.4)b 11.7 (± 0.7) 23.7 (± 6.6)b 
 IL-6 98.1 661.9 (± 224.8) 26.9 233.1 (± 68.9) 3.4 (± 0.2) 13.9 (± 4.9) 4.6 (± 0.9) 5.1 (± 1.5) 
 IL-10 157.6 (± 37.3) 164.6 (± 65.1) 106.1 (± 8.5) 98.2 (± 16.6) 80.1 (± 18.5) 99.2 (± 18.6) 
Chemokine         
 MCP-1 673.9 2939.1 (± 1479.5) 403.3 3061.1 (± 814.2) 27.1 (± 1.0) 33.2 (± 4.9) 33.2 (± 4.9) 23.6 (± 3.9) 
 Eotaxin 7.5 16.6 (± 1.4) 11.9 21.7 (± 6.3) 112.7 (± 8.3) 146.9 (± 27.1) 110.7 (± 10.2) 179.4 (± 28.9) 
Other cytokines         
 TNF-α 122.7 256.6 (± 64.1) 152.4 329.0 (± 53.5) 9.6 (± 1.2) 11.9 (± 1.4) 12.1 (± 1.5) 10.2 (± 1.2) 
 IL-12 31.0 41.9 (± 3.2) 27.0 66.3 (± 12.6) 32.9 (± 4.1) 23.4 (± 6.2) 42.4 (± 10.6) 24.4 (± 3.5) 
CytokinesLymph NodeLung Lavage
Saline + VehicleaOvalbumin + VehicleSaline + BilirubinaOvalbumin + BilirubinSaline + VehicleOvalbumin + VehicleSaline + BilirubinOvalbumin + Bilirubin
Th1         
 IL-2 6.2 4.2 (± 0.6) 7.9 5.6 (± 1.2) 5.9 (± 0.3) 6.6 (± 0.2) 6.8 (± 0.1) 6.6 (± 0.2) 
 IFN-γ 6.0 19.0 (± 8.3) 5.5 45.6 (± 13.8) 3.7 (± 0.7) 5.1 (± 0.8) 3.6 (± 1.7) 3.3 (± 0.4) 
Th2         
 IL-4 7.1 (± 1.9) 7.1 (± 1.9) 8.7 (± 0.2) 9.1 (± 0.2) 9.9 (± 0.9) 8.7 (± 0.2) 
 IL-5 2.6 191.4 (± 79.6) 301.1 (± 119.8) 10.4 (± 0.2) 15.7 (± 1.4)b 11.7 (± 0.7) 23.7 (± 6.6)b 
 IL-6 98.1 661.9 (± 224.8) 26.9 233.1 (± 68.9) 3.4 (± 0.2) 13.9 (± 4.9) 4.6 (± 0.9) 5.1 (± 1.5) 
 IL-10 157.6 (± 37.3) 164.6 (± 65.1) 106.1 (± 8.5) 98.2 (± 16.6) 80.1 (± 18.5) 99.2 (± 18.6) 
Chemokine         
 MCP-1 673.9 2939.1 (± 1479.5) 403.3 3061.1 (± 814.2) 27.1 (± 1.0) 33.2 (± 4.9) 33.2 (± 4.9) 23.6 (± 3.9) 
 Eotaxin 7.5 16.6 (± 1.4) 11.9 21.7 (± 6.3) 112.7 (± 8.3) 146.9 (± 27.1) 110.7 (± 10.2) 179.4 (± 28.9) 
Other cytokines         
 TNF-α 122.7 256.6 (± 64.1) 152.4 329.0 (± 53.5) 9.6 (± 1.2) 11.9 (± 1.4) 12.1 (± 1.5) 10.2 (± 1.2) 
 IL-12 31.0 41.9 (± 3.2) 27.0 66.3 (± 12.6) 32.9 (± 4.1) 23.4 (± 6.2) 42.4 (± 10.6) 24.4 (± 3.5) 
a

Data obtained from pooled samples.

b

, p < 0.05 versus saline controls.

To determine the effect of bilirubin on cellular cytokine production, the response of activated bronchial lymph node-derived lymphocytes to OVA rechallenge was assessed in vitro. As seen in Fig. 13,B, the production of eotaxin by isolated lymphocytes increases in response to OVA challenge, but remains unaffected by bilirubin treatment. Lymphocytes derived from OVA-treated mice also exhibit specific responses to OVA exposure, with enhanced production of Th2 (e.g., IL-5) but not Th1 (e.g., IL-2) cytokines (Fig. 13, D and F). That bilirubin does not modulate levels of inflammatory stimuli is evidenced by the fact that an identical pattern of cytokine production is observed in lymphocytes derived from bilirubin- vs vehicle-treated animals. These data, along with similar findings for the Th2 cytokines IL-4, IL-6, and IL-10, the Th1 cytokines IL-2 and IFN-γ, and the chemokine MCP-1, are summarized in Table I. Taken together, our findings indicate that bilirubin prevents OVA-induced eosinophil infiltration into the lungs despite the presence of a sufficient circulating pool of available cells and abundant chemotactic stimuli, supporting the conclusion that bilirubin inhibits VCAM-1 signaling pathways responsible for leukocyte migration across the pulmonary vascular endothelium.

The present studies demonstrate that both unconjugated bilirubin and the bilirubin precursor, biliverdin, inhibit transendothelial leukocyte migration in vitro. Because lymphocyte migration across mHEV monolayers is VCAM-1-dependent (3, 14, 17, 18), and we have shown that bilirubin suppresses cellular ROS generation and MMP activation, events which are downstream of VCAM-1 stimulation (3), our findings suggest that bilirubin disrupts endothelial VCAM-1 signaling. It is notable that these effects occur at bilirubin concentrations that are within the normal range present in human serum (≤1.4 mg/dL ≈ 24 μM) (43). Moreover, to ensure that bilirubin was adequately solubilized, experiments were conducted in the presence of 20% FCS, which is optimal for transendothelial lymphocyte migration (data not shown). This level of FCS corresponds to a BSA concentration of ∼90 μM (44), and results in bilirubin:BSA molar ratios of between 0 and 0.2. As a result of the high affinity binding of bilirubin by BSA (45, 46, 47), even at the highest levels of bilirubin used (20 μM), the free bilirubin concentration in the medium is estimated to be very low (∼3 nM), well within the reported solubility at physiologic pH (48). Because the free concentration dictates the cellular uptake and biological activity of bilirubin (49, 50), the observed effects can reasonably be extrapolated to in vivo processes.

The relevance of our in vitro data to physiological events is highlighted by the observation that the i.p. administration of bilirubin blocks the influx of leukocytes into the lungs of mice with OVA-induced asthma. It has previously been shown that Abs to α4 integrins or VCAM-1 inhibit the recruitment of eosinophils and lymphocytes into the airways of OVA-sensitized and OVA-challenged mice (34, 35). Our finding that bilirubin treatment has identical effects on OVA-induced leukocyte recruitment is consistent with a mechanism of action that involves the disruption of VCAM-1 signaling. This conclusion is bolstered by the observation that bilirubin does not alter VCAM-1 expression in the pulmonary vascular endothelium and that bilirubin-treated animals exhibit similar cytokine and chemokine responses as vehicle-treated controls. Moreover, a decrease in eosinophil infiltration into the airways occurs despite an enhanced pool of cells available for migration into the lungs, as evidenced by the significant increase in the number of circulating blood eosinophils following bilirubin administration. The mechanism underlying this bilirubin-associated blood eosinophilia, which was observed in both OVA-sensitized and -nonsensitized mice, is unclear. Although a stimulatory effect of bilirubin on eosinophil production cannot be excluded, because VCAM-1 has been shown to be involved in the trafficking of leukocytes to the intestine (51), bone marrow (52), and lymph nodes (53), we speculate that bilirubin enhances the number of eosinophils in the systemic circulation by blocking the normal migration into these tissues.

That cellular NADPH oxidase activity is integral to VCAM-1 signaling is supported by the observation that lymphocyte migration across mHEV monolayers is abrogated by specific inhibitors of NADPH oxidase, as well as by the ROS scavengers superoxide dismutase and catalase (3). As bilirubin is a potent antioxidant (22, 25), we postulate that this bile pigment inhibits VCAM-1-mediated leukocyte migration by scavenging the low (micromolar) concentrations of ROS generated by VCAM-1-dependent NADPH oxidase activation (18). Additionally, it has been shown that administration of an antioxidant mimetic, AEOL 10113, ameliorates OVA-induced pulmonary eosinophilia and lymphocytosis (20) and is further buttressed by our finding that bilirubin suppresses endothelial cell ROS production in response to VCAM-1 stimulation. Indeed, at 20-μM concentrations, bilirubin blocked lymphocyte migration to a comparable degree as previously shown for the ROS scavengers, superoxide dismutase (500 U/ml) and catalase (5000 U/ml), as well as an inhibitor of NADPH oxidase, apocynin (0.5 mM). However, these data do not preclude the possibility that bilirubin interferes with NADPH oxidase assembly or with upstream signaling events (14, 18).

Baranano et al. (25) recently proposed a cytoprotective mechanism in which bilirubin is oxidized to biliverdin and then recycled back to bilirubin by the action of cellular BVR. That redox cycling of bilirubin may be occurring in vascular endothelium receives support from our demonstration that mHEV cells express message for BVR and that treatment with biliverdin blocks lymphocyte migration to a comparable degree as bilirubin in vitro. In contrast, the ubiquitous nature of BVR activity in mammals, which results in the quantitative and rapid conversion of biliverdin to bilirubin under both physiological and pathophysiological conditions (54), makes it likely that bilirubin is the more relevant inflammatory mediator in vivo. In contrast with biliverdin, treatment of cell monolayers with the stable bilirubin conjugate, BDT, has no discernable effect on transendothelial lymphocyte migration, despite the fact that conjugated bilirubin also functions as an antioxidant (55). Although unconjugated bilirubin (and presumably biliverdin) is capable of diffusing rapidly through cellular membranes (56, 57), bilirubin conjugates, which are more hydrophilic, are membrane impermeant (58). Whether these differences in permeation are related to the inability of BDT to modulate lymphocyte migration is unknown at this time, but remains an intriguing possibility. Alternatively, these findings may be explained by the fact that the potency of BDT as an antioxidant is inferior to that of bilirubin or biliverdin (59).

It is notable that HO, the rate-limiting enzyme in bilirubin synthesis, is up-regulated at sites of inflammation and that the stimulated activity of this enzyme is associated with diminished tissue injury in a host of model systems (60, 61, 62, 63), including allergen-induced airway disorders (64, 65). That bilirubin is a key mediator of the cytoprotective effects of HO is supported by a number of in vitro (26, 66, 67, 68, 69) and in vivo (21, 24, 70) studies demonstrating that bilirubin ameliorates cellular oxidative damage. One of the ways in which bilirubin has been shown to inhibit tissue injury is by preventing the induction of NO synthase and consequent increases in NO levels (24). However, data are conflicting as to whether inducible NO synthase (iNOS) contributes to lung inflammation in the OVA asthma model. For example, it has been reported that pretreatment of mice with specific iNOS inhibitors decreases airway eosinophilia in response to allergen stimulation, either by blocking chemokine expression (71) or by delaying the efflux of eosinophils from the bone marrow (72). In contrast, other investigators have been unable to demonstrate a protective effect of iNOS inhibition on pulmonary leukocytosis (73, 74, 75), and iNOS-deficient mice paradoxically exhibit enhanced lung injury in response to OVA challenge (76). Our demonstration that bilirubin treatment does not alter chemokine levels and is associated with an increase (as opposed to a decrease) in the number of circulating eosinophils makes it unlikely that the effects of this bile pigment on airway inflammation are related to alterations in iNOS expression. Furthermore, mHEV cells do not express iNOS (data not shown), rendering it doubtful that the mechanism underlying bilirubin-mediated inhibition of transendothelial lymphocyte migration is iNOS-dependent.

Our finding that both endothelial and inflammatory cells (24) express all of the necessary enzymes involved in bilirubin synthesis and degradation, implies a high degree of regulation of cellular bilirubin levels and supports a potentially broad role for this endogenous bile pigment as a physiological regulator of inflammation. VCAM-1-mediated leukocyte recruitment is implicated in the pathogenesis of a number of inflammatory conditions, including atherosclerosis (77, 78) and inflammatory bowel disease (4, 9, 13). With regard to the former, VCAM-1 is detectable in atherosclerotic plaques (79) and endothelial expression of this adhesion molecule has been shown to be an early event at sites predisposed to atherosclerosis (80). Moreover, disruption of VCAM-1 expression in the atherosclerosis-prone low density lipoprotein receptor knockout mouse is associated with a significant decrease in the number of vascular lesions (78). Consistent with the hypothesis that bilirubin may modulate the process of atherogenesis by inhibiting VCAM-1 signaling, a host of epidemiological analyses have identified an inverse correlation between serum bilirubin levels and both the risk and severity of cardiovascular disease (81, 82, 83, 84). In summary, our data demonstrate that bilirubin suppresses Th2 cytokine-mediated airway inflammation in response to an allergen challenge by inhibiting endothelial cell VCAM-1 signaling, and that the mechanism underlying these effects appears to involve the scavenging of NADPH oxidase-generated ROS. As the essential role of low concentrations of ROS as cellular second messengers becomes more clearly elucidated (85, 86), our data raise the intriguing possibility that bilirubin may regulate a variety of fundamental cellular processes through BVR-catalyzed redox cycling.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This study was supported by National Institutes of Health Grants HL68171 (J.M.C-M.), DK063954 (S.D.Z), and Rehn Family Research Award (P.K.)

3

Abbreviations used in this paper: ROS, reactive oxygen species; AM, acetoxymethyl ester; BDT, ditaurobilirubin; BVR, biliverdin reductase; DHR, dihydrorhodamine 123; HO, heme oxygenase; iNOS, inducible NO synthase; MBP, major basic protein; mHEV, murine high endothelial venule; mHEVa, axillary mHEV; mHEVc cervical mHEV; MMP, matrix metalloproteinase; Rac, Ras-related C3 botulinum substrate; UGT1A1, UDP-glucuronosyltransferase.

1
Wang, J., T. A. Springer.
1998
. Structural specializations of immunoglobulin superfamily members for adhesion to integrins and viruses.
Immunol. Rev.
163
:
197
.
2
Quarmby, S., P. Kumar, S. Kumar.
1999
. Radiation-induced normal tissue injury: role of adhesion molecules in leukocyte-endothelial cell interactions.
Int. J. Cancer
82
:
385
.
3
Matheny, H. E., T. L. Deem, J. M. Cook-Mills.
2000
. Lymphocyte migration through monolayers of endothelial cell lines involves VCAM-1 signaling via endothelial cell NADPH oxidase.
J. Immunol.
164
:
6550
.
4
Jones, S. C., R. E. Banks, A. Haidar, A. J. Gearing, I. K. Hemingway, S. H. Ibbotson, M. F. Dixon, A. T. Axon.
1995
. Adhesion molecules in inflammatory bowel disease.
Gut
36
:
724
.
5
Ciprandi, G., A. M. Riccio, V. Venturino, M. Rolando, D. Contursi, G. W. Canonica.
2002
. VCAM-1 in conjunctival inflammation.
Allergy
57
:
961
.
6
Mrowka, C., B. Heintz, H. G. Sieberth.
1998
. VCAM-1, ICAM-1, and E-selectin in IgA nephropathy and Schonlein-Henoch syndrome: differences between tissue expression and serum concentration.
Nephron
81
:
256
.
7
Chakravorty, S. J., A. J. Howie, P. Cockwell, D. Adu, C. O. Savage.
1999
. T lymphocyte adhesion mechanisms within inflamed human kidney: studies with a Stamper-Woodruff assay.
Am. J. Pathol.
154
:
503
.
8
Carter, R. A., I. P. Wicks.
2001
. Vascular cell adhesion molecule 1 (CD106): a multifaceted regulator of joint inflammation.
Arthritis Rheum.
44
:
985
.
9
Molla, M., M. Gironella, R. Miquel, V. Tovar, P. Engel, A. Biete, J. M. Pique, J. Panes.
2003
. Relative roles of ICAM-1 and VCAM-1 in the pathogenesis of experimental radiation-induced intestinal inflammation.
Int. J. Radiat. Oncol. Biol. Phys.
57
:
264
.
10
Gimenez, M. A. T., J. E. Sim, J. H. Russell.
2004
. TNFR1-dependent VCAM-1 expression by astrocytes exposes the CNS to destructive inflammation.
J. Neuroimmunol.
151
:
116
.
11
James, W. G., D. C. Bullard, M. J. Hickey.
2003
. Critical role of the α4 integrin/VCAM-1 pathway in cerebral leukocyte trafficking in lupus-prone MRL/faslpr mice.
J. Immunol.
170
:
520
.
12
Shappell, S. B., L. H. Mendoza, T. Gurpinar, C. W. Smith, W. N. Suki, L. D. Truong.
2000
. Expression of adhesion molecules in kidney with experimental chronic obstructive uropathy: the pathogenic role of ICAM-1 and VCAM-1.
Nephron
85
:
156
.
13
Rijcken, E., C. Krieglstein, C. Anthoni, M. G. Laukoetter, R. Mennigen, H. U. Spiegel, N. Senninger, C. F. Bennett, G. Schuermann.
2002
. ICAM-1 and VCAM-1 antisense oligonucleotides attenuate in vivo leucocyte adherence and inflammation in rat inflammatory bowel disease.
Gut
51
:
529
.
14
Cook-Mills, J. M., J. D. Johnson, T. L. Deem, A. Ochi, L. Wang, Y. Zheng.
2004
. Calcium mobilization and Rac1 activation are required for VCAM-1 (vascular cell adhesion molecule-1) stimulation of NADPH oxidase activity.
Biochem. J.
378
:
539
.
15
Wang, Q., C. M. Doerschuk.
2002
. The signaling pathways induced by neutrophil-endothelial cell adhesion.
Antiox. Redox Signal
4
:
39
.
16
Ilan, N., J. A. Madri.
2003
. PECAM-1: od friend, new partners.
Curr. Opin. Cell Biol.
15
:
515
.
17
Deem, T. L., J. M. Cook-Mills.
2004
. Vascular cell adhesion molecule-1 (VCAM-1) activation of endothelial cell matrix metalloproteinases: role of reactive oxygen species.
Blood
104
:
2385
.
18
Cook-Mills, J. M..
2002
. VCAM-1 signals during lymphocyte migration: role of reactive oxygen species.
Mol. Immunol.
39
:
499
.
19
Khachigian, L. M., T. Collins, J. W. U. Fries.
1997
. N-acetyl cysteine blocks mesangial VCAM-1 and NF-κB expression in vivo.
Am. J. Pathol.
151
:
1225
.
20
Chang, L. Y., J. D. Crapo.
2002
. Inhibition of airway inflammation and hyperreactivity by an antioxidant mimetic.
Free Rad. Biol. Med.
33
:
379
.
21
Kunsch, C., J. Luchoomun, J. Y. Grey, L. K. Olliff, L. B. Saint, R. F. Arrendale, M. A. Wasserman, U. Saxena, R. M. Medford.
2004
. Selective inhibition of endothelial and monocyte redox-sensitive genes by AGI-1067: a novel antioxidant and anti-inflammatory agent.
J. Pharmacol. Exp. Ther.
308
:
820
.
22
Stocker, R., Y. Yamamoto, A. F. McDonagh, A. N. Glazer, B. N. Ames.
1987
. Bilirubin is an antioxidant of possible physiological importance.
Science
235
:
1043
.
23
Clark, J. E., R. Foresti, P. Sarathchandra, H. Kaur, C. J. Green, R. Motterlini.
2000
. Heme oxygenase-1-derived bilirubin ameliorates postischemic myocardial dysfunction.
Am. J. Physiol.
278
:
H643
.
24
Wang, W. W., D. L. H. Smith, S. D. Zucker.
2004
. Bilirubin inhibits iNOS expression and nitirc oxide production in response to endotoxin.
Hepatology
40
:
424
.
25
Baranano, D. E., M. Rao, C. D. Ferris, S. H. Snyder.
2002
. Biliverdin reductase: a major physiologic cytoprotectant.
Proc. Natl. Acad. Sci. USA
99
:
16093
.
26
Hayashi, S., R. Takamiya, T. Yamaguchi, K. Matsumoto, S. J. Tojo, T. Tamatani, M. Kitajima, N. Makino, Y. Ishimura, M. Suematsu.
1999
. Induction of heme oxygenase-1 suppresses venular leukocyte adhesion elicited by oxidative stress: role of bilirubin generated by the enzyme.
Cell Regul.
85
:
663
.
27
Clark, J. E., R. Foresti, C. J. Green, R. Motterlini.
2000
. Dynamics of haem oxygenase-1 expression and bilirubin production in cellular protection against oxidative stress.
Biochem. J.
348
:
615
.
28
McDonagh, A. F., F. Assisi.
1972
. The ready isomerization of bilirubin IXα in aqueous solution.
Biochem. J.
129
:
797
.
29
Cook-Mills, J. M., J. S. Gallagher, T. L. Feldbush.
1996
. Isolation and characterization of high endothelial cell lines derived from mouse lymph nodes.
In Vitro Cell. Dev. Biol.
32
:
167
.
30
Keshavan, P., S. J. Schwemberger, D. L. H. Smith, G. F. Babcock, S. D. Zucker.
2004
. Unconjugated bilirubin induces apoptosis in colon cancer cells by triggering mitochondrial depolarization.
Int. J. Cancer.
112
:
433
.
31
Qureshi, M. H., J. M. Cook-Mills, D. E. Doherty, B. A. Garvy.
2003
. TNF-α-dependent ICAM-1- and VCAM-1-mediated inflammatory responses are delayed in neonatal mice infected with Pneumocystis carinii.
J. Immunol.
171
:
4700
.
32
Tanaka, Y., M. Yoshikawa, Y. Kobayashi, M. Kuroda, M. Kaito, A. Shiroi, J. Yamao, H. Fukui, Y. Adachi.
2003
. Expressions of hepatobiliary organic anion transporters and bilirubin-conjugating enzyme in differentiating embryonic stem cells.
Biochem. Biophys. Res. Commun.
309
:
324
.
33
Tudor, K. S. R. S., K. L. Hess, J. M. Cook-Mills.
2001
. Cytokines modulate endothelial cell intracellular signal transduction required for VCAM-1-dependent lymphocyte transendothelial migration.
Cytokine
15
:
196
.
34
Kudlacz, E., C. Whitney, C. Andresen, A. Duplantier, G. Beckius, L. Chupak, A. Klein, K. Kraus, A. Milici.
2002
. Pulmonary eosinophilia in a murine model of allergic inflammation is attenuated by small molecule α4β1 antagonists.
J. Pharmacol. Exp. Ther.
301
:
747
.
35
Chin, J. E., C. A. Hatfield, G. E. Winterrowd, J. R. Brashler, S. L. Vonderfecht, S. F. Fidler, R. L. Griffin, K. P. Kolbasa, R. F. Krzesicki, L. M. Sly, et al
1997
. Airway recruitment of leukocytes in mice is dependent on α4-integrins and vascular cell adhesion molecule-1.
Am. J. Physiol.
272
:
L219
.
36
Stelts, D., R. W. Egan, A. Falcone, C. G. Garlisi, G. J. Gleich, W. Kreutner, T. T. Kung, D. K. Nahrebne, R. W. Chapman, M. Minnicozzi.
1998
. Eosinophils retain their granule major basic protein in a murine model of allergic pulmonary inflammation.
Am. J. Respir. Cell Mol. Biol.
18
:
463
.
37
Munoz, M. E., A. Esteller, J. Gonzalez.
1987
. Substrate induction of bilirubin conjugation and biliary excretion in the rat.
Clin. Sci.
73
:
371
.
38
Li, Y. Q., D. A. Prentice, M. L. Howard, M. L. Mashford, P. V. Desmond.
2000
. Bilirubin and bile acids may modulate their own metabolism via regulating uridine diphosphate-glucuronosyltransferase expression in the rat.
J. Gastroenterol. Hepatol.
15
:
865
.
39
Pease, J. E., T. J. Williams.
2001
. Eotaxin and asthma.
Curr. Opin. Pharmacol.
1
:
248
.
40
Wohllenben, G., J. Muller, U. Tatsch, C. Hambrecht, U. Herz, H. Renz, E. Schmitt, H. Moll, K. J. Erb.
2003
. Influenza A virus infection inhibits the efficient recruitment of Th2 cells into the airways and the development of airway eosinophilia.
J. Immunol.
170
:
4601
.
41
Lampinen, M., M. Carlson, L. D. Hakansson, P. Venge.
2004
. Cytokine regulated accumulation of eosinophils in inflammatory disease.
Allergy
59
:
793
.
42
Wu, A. Y., S. C. Chik, A. W. Chan, Z. Li, K. W. Tsang, W. Li.
2003
. Anti-inflammatory effects of high-dose montelukast in an animal model of acute asthma.
Clin. Exp. Allergy
33
:
359
.
43
Zucker, S. D., P. S. Horn, K. E. Sherman.
2004
. Serum bilirubin levels in the U.S. population: gender effect and inverse correlation with colorectal cancer.
Hepatology
40
:
827
.
44
Horn, K. V., J. A. Singley, W. Chavin.
1975
. Fetal bovine serum: a multivariate standard.
Proc. Soc. Exp. Biol. Med.
149
:
344
.
45
Blauer, G., T. E. King.
1970
. Interactions of bilirubin with bovine serum albumin in aqueous solution.
J. Biol. Chem.
245
:
372
.
46
Blaha, G., M. Siam, H. Lehner.
1997
. A circular dichroism (CD) study of the consecutive binding of serum albumin to bilirubin: possible implications for the bilirubin level.
J. Chem. Soc. Perkin Trans.
2
:
2119
.
47
Siam, M., G. Blaha, H. Lehner.
1998
. Maximum binding capacity of serum albumin for bilirubin is one, as revealed by circular dichroism.
J. Chem. Soc. Perkin Trans.
2
:
853
.
48
Ostrow, J. D., P. Mukerjee, C. Tiribelli.
1994
. Structure and binding of unconjugated bilirubin: relevance for physiological and pathophysiological function.
J. Lipid Res.
35
:
1715
.
49
Pascolo, L., S. del Vecchio, R. K. Koehler, E. J. Bayon, C. C. Webster, P. Mukerjee, J. D. Ostrow, C. Tiribelli.
1996
. Albumin binding of unconjugated [3H]bilirubin and its uptake by rat liver basolateral plasma membrane vesicles.
Biochem. J.
316
:
999
.
50
Ostrow, J. D., L. Pascolo, C. Tiribelli.
2002
. Mechanisms of bilirubin neurotoxicity.
Hepatology
35
:
1277
.
51
Salmi, M., S. Jalkanen.
1991
. Regulation of lymphocyte traffic to mucosa-associated lymphatic tissues.
Gastroenterol. Clin. North Am.
20
:
495
.
52
Koni, P. A., S. K. Joshi, U. A. Temann, D. Olson, L. Burkly, R. A. Flavell.
2001
. Conditional vascular cell adhesion molecule 1 deletion in mice: impaired lymphocyte migration to bone marrow.
J. Exp. Med.
193
:
741
.
53
Faveeuw, C., M. E. Di Mauro, A. A. Price, A. Ager.
2000
. Roles of α4 integrins/VCAM-1 and LFA-1/ICAM-1 in the binding and transendothelial migration of T lymphocytes and T lymphoblasts across high endothelial venules.
Int. Immunol.
12
:
241
.
54
McDonagh, A. F..
2001
. Turning green to gold.
Nat. Struct. Biol.
8
:
198
.
55
Stocker, R., E. Peterhans.
1989
. Antioxidant properties of conjugated bilirubin and biliverdin: biologically relevant scavenging of hypochlorous acid.
Free Radic. Res. Commun.
6
:
57
.
56
Hayward, D., D. Schiff, S. Fedunec, G. Chan, J. P. Davis, M. J. Poznansky.
1986
. Bilirubin diffusion through lipid membranes.
Biochim. Biophys. Acta
860
:
149
.
57
Zucker, S. D., W. Goessling.
2000
. Mechanism of hepatocellular uptake of unconjugated bilirubin.
Biochim. Biophys. Acta
1463
:
197
.
58
Zucker, S. D., W. Goessling, A. G. Hoppin.
1999
. Unconjugated bilirubin exhibits spontaneous diffusion through model lipid bilayers and native hepatocyte membranes.
J. Biol. Chem.
274
:
10852
.
59
Stocker, R., B. N. Ames.
1987
. Potential role of conjugated bilirubin and copper in the metabolism of lipid peroxides in bile.
Proc. Natl. Acad. Sci. USA
84
:
8130
.
60
Otterbein, L. E., B. Y. Chin, S. L. Otterbein, V. C. Lowe, H. E. Fessler, A. M. K. Choi.
1997
. Mechanism of hemoglobin-induced protection against endotoxemia in rats: a ferritin-independent pathway.
Am. J. Physiol.
272
:
L268
.
61
Amersi, F., R. Buelow, H. Kato, B. Ke, A. J. Coito, X. D. Shen, D. Zhao, J. Zaky, J. Melinek, C. R. Lassman, et al
1999
. Upregulation of heme oxygenase-1 protects genetically fat Zucker rat livers from ischemia/reperfusion injury.
J. Clin. Invest.
104
:
1631
.
62
Mosley, K., D. E. Wembridge, V. Cattell, H. T. Cook.
1998
. Heme oxygenase is induced in nephrotoxic nephritis and hemin, a stimulator of heme oxygenase synthesis, ameliorates disease.
Kidney Int.
53
:
672
.
63
Wang, W. P., X. Guo, M. W. Koo, B. C. Wong, S. K. Lam, Y. N. Ye, C. H. Cho.
2001
. Protective role of heme oxygenase-1 on trinitrobenzene sulfonic acid-induced colitis in rats.
Am. J. Physiol.
281
:
G586
.
64
Almolki, A., C. Taille, G. F. Martin, P. J. Jose, C. Zedda, M. Conti, J. Megret, D. Henin, M. Aubier, J. Boczkowski.
2004
. Heme oxygenase attenuates allergin-induced airway inflammation and hyperreactivity in guinea pigs.
Am. J. Physiol.
287
:
L26
.
65
Jia, Y. X., K. Sekizawa, S. Okinaga, R. Lie, H. Sasaki.
1999
. Role of heme oxygenase in pulmonary response to antigen challenge in sensitized rats in vivo.
Int. Arch. Allergy Immunol.
120
:
141
.
66
Hayashi, S., R. Takamiya, T. Yamaguchi, K. Matsumoto, S. J. Tojo, T. Tamatani, M. Kitajima, N. Makino, Y. Ishimura, M. Suematsu.
1999
. Induction of heme oxygenase-1 suppresses venular leukocyte adhesion elicited by oxidative stress: role of bilirubin generated by the enzyme.
Circ. Res.
85
:
663
.
67
Dore, S., M. Takahashi, C. D. Ferris, L. D. Hester, D. Guastella, S. H. Snyder.
1999
. Bilirubin, formed by activation of heme oxygenase-2, protects neurons against oxidative stress injury.
Proc. Natl. Acad. Sci. USA
96
:
2445
.
68
Wu, T. W., D. Carey, J. Wu, H. Sugiyama.
1991
. The cytoprotective effects of bilirubin and biliverdin on rat hepatocytes and human erythrocytes and the impact of albumin.
Biochem. Cell Biol.
69
:
828
.
69
Wu, T. W., J. Wu, R. K. Li, D. A. G. Mickle, D. Carey.
1991
. Alubmin-bound bilirubins protect human ventricular myocytes against oxyradical damage.
Biochem. Cell Biol.
69
:
683
.
70
Kato, Y., M. Shimazu, M. Kondo, K. Uchida, Y. Kumamoto, G. Wakabayashi, M. Kitajima, M. Suematsu.
2004
. Bilirubin rinse: a simple protectant against the rat liver graft injury mimicking heme oxygenase-1 preconditioning.
Hepatology
38
:
364
.
71
Trifilieff, A., Y. Fujitani, F. Mentz, B. Dugas, M. Fuentes, C. Bertrand.
2000
. Inducible nitric oxide synthase inhibitors suppress airway inflammation in mice through down-regulation of chemokine expression.
J. Immunol.
165
:
1526
.
72
de Araujo Ferreira, H. H., R. A. de Oliveira Costa, J. M. Jacheta, A. R. Martins, M. V. Medeiros, M. F. Macedo-Soares, I. M. Silva De Luca, E. Antunes, G. De Nucci.
2004
. Modulation of eosinophil migration from bone marrow to lungs of allergic rats by nitric oxide.
Biochem. Pharmacol.
68
:
631
.
73
Feder, L. S., D. Stelts, R. W. Chapman, D. Manfra, Y. Crawley, H. Jones, M. Minnicozzi, X. Fernandez, T. Paster, R. W. Egan, et al
1997
. Role of nitric oxide on eosinophilic lung inflammation in allergic mice.
Am. J. Respir. Cell Mol. Biol.
17
:
436
.
74
Muijsers, R. B., E. van Den Worm, G. Folkerts, C. J. Beukelman, A. S. Koster, D. S. Postma, F. P. Nijkamp.
2000
. Apocynin inhibits peroxynitrite formation by murine macrophages.
Br. J. Pharmacol.
130
:
932
.
75
Schuiling, M., H. Meurs, A. B. Zuidhof, N. Venema, J. Zaagsma.
1998
. Dual action of iNOS-derived nitric oxide in allergen-induced airway hyperreactivity in conscious, unrestrained guinea pigs.
Am. J. Respir. Crit. Care Med.
158
:
1442
.
76
Kenyon, N. J., K. Gohill, J. A. Last.
2003
. Susceptibility to ovalbumin-induced airway inflammation and fibrosis in inducible nitric oxide synthetase-deficient mice: mechanisms and consequences.
Toxicol. Appl. Pharmacol.
191
:
2
.
77
Blankenberg, S., S. Barbaux, L. Tiret.
2003
. Adhesion molecules and atherosclerosis.
Atherosclerosis
170
:
191
.
78
Cybulsky, M. I., K. Iiyama, H. Li, S. Zhu, M. Chen, M. Iiyama, V. Davis, J. C. Gutierrez-Ramos, P. W. Connelly, D. S. Milstone.
2001
. A major role for VCAM-1, but not ICAM-1, in early atherosclerosis.
J. Clin. Invest.
107
:
1255
.
79
Obrien, K. D..
1993
. Vascular cell adhesion molecule-1 is expressed in human coronary atheroslcerotic plaques: implications for the mode of progression of advanced coronary atherosclerosis.
J. Clin. Invest.
92
:
945
.
80
Nakashima, Y., E. W. Raines, A. S. Plump, J. L. Breslow, R. Ross.
1998
. Upregulation of VCAM-1 and ICAM-1 at atherosclerosis-prone sites on the endothelium in the apoE-deficient mouse.
Arterioscler. Thromb. Vasc. Biol.
18
:
842
.
81
Schwertner, H. A., W. G. Jackson, G. Tolan.
1994
. Association of low serum concentration of bilirubin with increased risk of coronary artery disease.
Clin. Chem.
40
:
18
.
82
Breimer, L. H., G. Wannamethee, S. Ebrahim, A. G. Shaper.
1995
. Serum bilirubin and risk of ischemic heart disease in middle-aged British men.
Clin. Chem.
41
:
1504
.
83
Hopkins, P. N., L. L. Wu, S. C. Hunt, B. C. James, G. M. Vincent, R. R. Williams.
1996
. Higher serum bilirubin is associated with decreased risk for early familial coronary artery disease.
Arterioscler. Thromb. Vasc. Biol.
16
:
250
.
84
Madhavan, M., W. A. Wattigney, S. R. Srinivasan, G. S. Berenson.
1997
. Serum bilirubin distribution and its relation to cardiovascular risk in children and young adults.
Atherosclerosis
131
:
107
.
85
Dröge, W..
2001
. Free radicals in the physiological control of cell function.
Physiol. Rev.
82
:
47
.
86
Forman, H. J., J. M. Fukuto, M. Torres.
2004
. Redox signaling: thiol chemistry defines which reactive oxygen and nitrogen species can act as second messengers.
Am. J. Physiol.
287
:
C246
.